Sniffer Portable
Transcription
Sniffer Portable
EXPERT ALARMS REFERENCE GUIDE Sniffer ® Portable VERSION 4.8 COPYRIGHT © 2005 Network General Corporation. All Rights Reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form or by any means without the written permission of Network General Corporation or its suppliers or affiliate companies. TRADEMARK ATTRIBUTIONS Appera, InfiniStream, Know The Network, Netasyst, Network General, Network Performance Orchestrator, nPO, PrimeSupport, and Sniffer are registered trademarks or trademarks of Network General Corporation and/or its affiliates in the US and/or other countries. All other registered and unregistered trademarks in this document are the sole property of their respective owners. © 2005 Network General Corporation. All Rights Reserved. LICENSE AGREEMENT NOTICE TO ALL USERS: CAREFULLY READ THE APPROPRIATE LEGAL AGREEMENT CORRESPONDING TO THE LICENSE YOU PURCHASED, WHICH SETS FORTH THE GENERAL TERMS AND CONDITIONS FOR THE USE OF THE LICENSED SOFTWARE. IF YOU DO NOT KNOW WHICH TYPE OF LICENSE YOU HAVE ACQUIRED, PLEASE CONSULT THE SALES AND OTHER RELATED LICENSE GRANT OR PURCHASE ORDER DOCUMENTS THAT ACCOMPANIES YOUR SOFTWARE PACKAGING OR THAT YOU HAVE RECEIVED SEPARATELY AS PART OF THE PURCHASE (AS A BOOKLET, A FILE ON THE PRODUCT CD, OR A FILE AVAILABLE ON THE WEB SITE FROM WHICH YOU DOWNLOADED THE SOFTWARE PACKAGE). IF YOU DO NOT AGREE TO ALL OF THE TERMS SET FORTH IN THE AGREEMENT, DO NOT INSTALL THE SOFTWARE. IF APPLICABLE, YOU MAY RETURN THE PRODUCT TO NETWORK GENERAL OR THE PLACE OF PURCHASE FOR A FULL REFUND. March, 2005 / 100421 Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Getting More Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Contacting Network General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii 1 Expert Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 About Expert Alarms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Information Reported with Expert Alarms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 About Expert Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Setting Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Expert Alarms and Thresholds Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Service Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Slow FTP Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Slow HTTP Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Slow Mail Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Slow News Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Slow Oracle SQL Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Slow POP3 Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Slow Sybase/Microsoft SQL Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Slow Telnet Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Application Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 [VOIP] H323 - High Call Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 [VOIP] H323 - Too Many Incomplete Calls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 [VOIP] SCCP - High Call Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 [VOIP] SCCP - Too Many Incomplete Calls . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 [VOIP] SIP - High Call Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 [VOIP] SIP - Too Many Incomplete Calls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Access to Resource Denied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Browser Election Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 DB Connect Request Denied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 DB Security Breach Attempt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 DB Slow Connect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Expert Alarms Reference Guide iii Contents DB Slow Server Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 DB Slow Server Response Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Excessive Failed Resource Login Attempts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Excessive Mailslot Broadcasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 FCIP Special Frame Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 FCIP Invalid Encapsulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 FCIP Repeat Special Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 FCIP Special Frame Echo Has Changed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 FCIP Special Frame Delayed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 FCIP Too Many Special Frame Echoes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 FCIP Transmission Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 FTP Login Attempts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 FTP Slow Connect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 FTP Slow First Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 FTP Slow Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 FTP Slow Transfer Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 HTTP Slow Server GET Response Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 HTTP Slow Server POST Response Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Invalid Network Name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Invalid Resource User Name or Password . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 iSCSI: Authorization Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 iSCSI: Can Not Generate Target Transfer Tag - Out of Resources . . . . . . . . . . 25 iSCSI: Command Not Supported in this Session Type . . . . . . . . . . . . . . . . . . . . 26 iSCSI: Command Retransmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 iSCSI: Data Digest Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 iSCSI: Data or Status again Requested . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 iSCSI: Function Authorization Failed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 iSCSI: Function Rejected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 iSCSI: Immediate Command Reject (Too Many Immediate Commands) . . . . . . 27 iSCSI: Initiator Authentication Failed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 iSCSI: Invalid Data Ack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 iSCSI: Invalid PDU field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 iSCSI: Invalid During Login . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 iSCSI: LUN Does Not Exist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 iSCSI: Miscellaneous Initiator Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 iSCSI: Missing Parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 iSCSI: Service Unavailable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 iSCSI: Session Does Not Exist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 iv Sniffer Technologies Contents iSCSI: Target Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 iSCSI: Target Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 iSCSI: Target Out of Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 iSCSI: Target Removed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 iSCSI: Task Does Not Exist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 iSCSI: Task Management Function Not Supported . . . . . . . . . . . . . . . . . . . . . . 30 iSCSI: Too Many Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 iSCSI: Unsupported Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Missed Browser Announcement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 MS RAP Logon Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 NCP Server Busy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 NFS Retransmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 NNTP Slow Response Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 POP3 Slow Connect Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Protocol Negotiation Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 SAP R3 Slow Server Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 SAP R3 Slow Server Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 SMTP Slow Connect Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Station Not in Domain’s Computer List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Telnet Slow Response to Login . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Session Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 [VOIP] H225 - Abnormal Disconnect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 [VOIP] H245 - Open Logical Channel Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 [VOIP] H245 - Terminal Capability Set Reject . . . . . . . . . . . . . . . . . . . . . . . . . . 36 [VOIP] RAS - Admission Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 [VOIP] RAS - Bandwidth Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 [VOIP] RAS - Disengage Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 [VOIP] RAS - Gatekeeper Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 [VOIP] RAS - Location Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 [VOIP] RAS - Registration Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 [VOIP] RAS - Unregistration Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 [VOIP] RTCP - Report High Jitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 [VOIP] RTP - High Jitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Decreasing the Frequency of High Jitter Alarms . . . . . . . . . . . . . . . . . . . . . . . . 44 Alarm Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 [VOIP] RTP - High Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 [VOIP] RTP - Too Many Dropped Packets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 [VOIP] RTP - Too Many Out of Sequence Packets . . . . . . . . . . . . . . . . . . . . . . 47 Expert Alarms Reference Guide v Contents [VOIP] SCCP - Register Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 [VOIP] SCCP - Station Alarm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 [VOIP] SIP - Client Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 [VOIP] SIP - Global Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 [VOIP] SIP - Server Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 [VOIP] SIP - Server Slow Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 File Retransmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Invalid Presentation Context ID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 IP NETBIOS Session Reject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Local Limit on Defined Context Set Exceeded . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Loops on Same Request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Low Throughput . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Read/Write Overlap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Request Denied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Slow File Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Temporary Congestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 The Operation Number Is Greater than the Number of Operations in the Interface . 57 The Output Parameters of the Operation Exceed Their Declared Maximum Size 57 The Request Is Being Rejected for Unspecified Reasons . . . . . . . . . . . . . . . . . 57 The RPC Client or Server Protocol Has Been Violated . . . . . . . . . . . . . . . . . . . 57 The Server Did Not Support the Requested Authentication Level . . . . . . . . . . . 57 The Server Does Not Export the Requested Interface . . . . . . . . . . . . . . . . . . . . 57 The Server Does Not Implement the Requested Operation for the Type of Requested Object . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 The Server Does Not Support the Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 The Server Does Not Support the Proposed Transfer Syntax . . . . . . . . . . . . . . 58 The Server Does Not Support RPC Protocol Version . . . . . . . . . . . . . . . . . . . . . 58 The Server Is Too Busy To Handle the Call . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 The Server Manager Routine Has Not Been Entered and Executed . . . . . . . . . 58 TNS Connect Refused . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 TNS Security Breach Attempt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 TNS Slow Connect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 TNS Slow Server Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 TNS Slow Server Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Too Many File Retransmissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Too Many Loops on Same Request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Too Many Requests Denied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 WINS Duplicate Name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 vi Sniffer Technologies Contents WINS No Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Connection Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Ack Too Long . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Fast Retransmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Idle Too Long . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Local Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Non-Responsive Station . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Repeat Ack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Retransmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Slow Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 TCP Fast Keepalive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Too Many Retransmissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 UDP Bouncing Frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Window Frozen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Window Size Exceeded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Zero Window Too Long . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Station Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Duplicate Network Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Duplicate PDC Name Registration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 ICMP Destination Unreachable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 ICMP Host Unreachable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 ICMP Inconsistent Subnet Mask . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 ICMP Net Unreachable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 ICMP Parameter Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 ICMP Port Unreachable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 ICMP Redirect for Host . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 ICMP Redirect for Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 ICMP Source Quench . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 IP Fragment Missing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 IP Fragment Out of Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Misdirected Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Multiple Routers to Local . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Multiple Routers to Remote . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Route Flapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Router Not Updating Routes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Router Storm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Router Superseded Too Frequently . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Server Running BDC Shutting Down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Expert Alarms Reference Guide vii Contents Server Running PDC Shutting Down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Server Running WINS Shutting Down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Source Address Is Broadcast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Time-to-live Exceeded in Transmit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Time-to-live Expiring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Zero Broadcast Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 DLC Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Abort Delimiter Transmitted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 AC Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Burst Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Collision after 64 Bytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 DLC Source Address Broadcast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 DLC Source Address Multicast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Duplicate Address Found . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 FCI and ARI Bits Mismatch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 FDDI Ring Wrap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Frame Has Alignment Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Frame-Copied Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Frequency Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 High Rate of Line/Burst Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 High Rate of Physical Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 High Rate of Receiver Congestion Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 High Rate of Ring Purge Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Invalid Frame Status Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Line Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Local Station with Route Designator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Lost Frame Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Multiple Local Ring Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 New Active Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Oversized Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Receiver Congestion Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Returned to Ring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Ring Beaconing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Ring Purge Symptom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Runt Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Same Source and Destination Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Station Internal Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Station Off Ring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 viii Sniffer Technologies Contents Station Removed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Token Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Too Many Removes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Too Many Returns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Zero MAC Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 WAN Connection Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Backward Congestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Exceeded CIR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Forward Congestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 PVC Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 PVC Bandwidth Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 PVC Congestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 PVC Deactivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 PVC Deletion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Traffic on Deleted DLCI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Traffic on Inactive DLCI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Traffic on Unknown DLCI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 WAN Link Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Async Status Change of Unknown DLCI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 DCE Sequence Number Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Delayed STATUS ENQUIRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 DTE Sequence Number Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Irregular Full Status Report Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Link Management Message Format Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Link Management Message Minor Format Error . . . . . . . . . . . . . . . . . . . . . . . . 97 Network Equipment Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 New PVC Configured . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 No STATUS ENQUIRY When Expected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 No STATUS Following STATUS ENQUIRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 PVC Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 PVC Bandwidth Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 PVC Congestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 PVC Deactivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 PVC Deletion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Unexpected Link Management Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Unsolicited STATUS Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 User Equipment Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Wrong Report Type in STATUS Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Expert Alarms Reference Guide ix Contents ATM Application Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Attempt to Bind L3 Address to Multiple MPOA Devices . . . . . . . . . . . . . . . . . . 100 Data Frames Detected Following Data Plane Purge . . . . . . . . . . . . . . . . . . . . 100 DDVCC Idle Period Exceeds VCC Timeout C12 . . . . . . . . . . . . . . . . . . . . . . . 101 DDVCC Not Created in Response to LE ARP . . . . . . . . . . . . . . . . . . . . . . . . . 101 Default MCFVCC Not Created in Response to LE ARP . . . . . . . . . . . . . . . . . . 101 Default MCSVCC Not Created in Response to LE ARP . . . . . . . . . . . . . . . . . . 102 Excessive Unknown Frames Issued by LEC . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Frames Seen on Default Flow After Shortcut Active . . . . . . . . . . . . . . . . . . . . . 102 LANE Config Phase Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 LANE Cfg Restart Delay < Min Recfg Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 LANE Cfg Restart Delay > Max Recfg Delay . . . . . . . . . . . . . . . . . . . . . . . . . . 104 No Resolution Request Issued After Excess Default Flow . . . . . . . . . . . . . . . . 104 Shortcut Create Failed After Resolution Reply . . . . . . . . . . . . . . . . . . . . . . . . . 104 ATM Flow Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 ARE or STE Frames on Data Direct VCC . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 ARP Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Attempt to Bind DLC to MPC and Non-MPOA Device . . . . . . . . . . . . . . . . . . . 105 Attempt to Multiply Bind MPOA Devices to LEC . . . . . . . . . . . . . . . . . . . . . . . . 106 Broadcast/Multicast Data Frames on DDVCC . . . . . . . . . . . . . . . . . . . . . . . . . 106 Cache ID Set to Zero in DLL Header Ext . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Config Fail, Insufficient Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Config Frag Info TLV Not Recognized by LECS . . . . . . . . . . . . . . . . . . . . . . . . 107 Config Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Config Request Response Timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Config Request Retry Frame Mismatch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Config Request Retry Rate Exceeds 1 Frame/sec . . . . . . . . . . . . . . . . . . . . . . 108 Control Request on DDVCC Retry Frame Mismatch . . . . . . . . . . . . . . . . . . . . 108 Control Request on DDVCC Retry Rate > 1 Frame/Sec . . . . . . . . . . . . . . . . . 109 Database Summary Exchange Lockstep Failure . . . . . . . . . . . . . . . . . . . . . . . 109 Database Summary Master/Slave Negotiation Timeout . . . . . . . . . . . . . . . . . . 109 Denied Access to ELAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 DLL Header Extension Not Found in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Duplicate ATM Destination Address Registration . . . . . . . . . . . . . . . . . . . . . . . 110 Duplicate LAN Destination Registration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Egress Cache Tag Was Not Issued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Failed to Issue Resolution Req. Following Trigger . . . . . . . . . . . . . . . . . . . . . . 111 Failure Reported in Reply CIE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 x Sniffer Technologies Contents Flush Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 IG Tag Violation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Insufficient Info from LEC, Service Refused . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Invalid ATM Primary Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Invalid CIE Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Invalid Client ATM Primary Source Address . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Invalid Client MAC Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Invalid Common Header Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Invalid Fixed Header Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Invalid Frame Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Invalid Hello Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Invalid MPOA MPS MAC Address Count in Device Type TLV . . . . . . . . . . . . . 115 Invalid MPOA Packet Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Invalid Tag in Data Frame Header . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Join Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Join Request Received from LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Join Request Transmitted by LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Join Response Issued by LEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Keep Alive Lifetime Set to Zero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Keep Alive Sequence Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Keep Alive Lifetime Ext Not Found in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . 117 LANE Control Request Response Timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 LANE Control Request Retry Frame Mismatch . . . . . . . . . . . . . . . . . . . . . . . . 118 LANE Control Request Retry Rate > 1 Frame/sec . . . . . . . . . . . . . . . . . . . . . . 118 LANE Version Unsupported by ELAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 LE ARP Request Rate Exceeds 1 Frame/Sec . . . . . . . . . . . . . . . . . . . . . . . . . 119 LE_CONFIGURE Params Conflict, Service Refused . . . . . . . . . . . . . . . . . . . . 119 LEC Issued Config Response to LECS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 LEC Not Recognized by LECS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 LEC ID in Frame Mismatch with Client LEC ID . . . . . . . . . . . . . . . . . . . . . . . . . 120 LECID Is Not 0x0000 in Request Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 LECS Issued Config Request to LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Local Segment ID not in Source Routed Frame . . . . . . . . . . . . . . . . . . . . . . . . 121 MAC Address bound to Multiple LECs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Max Transmission Unit Size Undefined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Missing ULIA on Outside Link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 MPOA Capable LEC Did Not Issue Device TLV . . . . . . . . . . . . . . . . . . . . . . . . 122 MPOA Config Parameter Invalid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Expert Alarms Reference Guide xi Contents MPOA Control Request Response Timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 MPOA Control Request Retry Rate > 1 Frame/sec . . . . . . . . . . . . . . . . . . . . . 123 MPOA Error Frame Detected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Multiply Assigned Secondary LEC Address . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 NARP Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 NARP Request Received from LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 No CIEs Were Present in MPOA Reply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 No CIEs Were Present in MPOA Request . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 No Reply Flag Is Not Set in ICache Purge . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 PTSE Violation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Register Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Register Request Received from LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Register Response Issued by LEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Remote Address Issued by Non Proxy LEC . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Request Params Are Incompatible with ELAN . . . . . . . . . . . . . . . . . . . . . . . . . 127 TA List Resources Exhausted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Topology Change Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Topology Mismatch in Data Frame Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Unexpected DLL Header EXT in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Unexpected Egress Cache EXT in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Unexpected Hop Count EXT in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Unexpected IG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Unexpected Keepalive Lifetime EXT in Frame . . . . . . . . . . . . . . . . . . . . . . . . . 129 Unexpected Original Error Code EXT in Frame . . . . . . . . . . . . . . . . . . . . . . . . 129 Unexpected Service Category EXT in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Unexpected TLVs Detected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Unknown Extension in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Unknown IG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Unknown LANE Config/Control Opcode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Unknown Marker/Protocol Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Unknown/Unexpected Authenticate EXT in Frame . . . . . . . . . . . . . . . . . . . . . . 131 Unknown/Unexpected CIE in Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Unregister Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Unregister Request Received from LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Unregister Response Issued by LEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 V2 TLVs Issued in V1 ELAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Verify Request Multi Queued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Verify Request Received from LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 xii Sniffer Technologies Contents Verify Response Issued by LEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 ATM Connection Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 AAL5 CRC Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 AAL5 Length Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 AAL5 Timeout Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Abnormal Disconnect for SVC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 CLP = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Connection Too Short . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Crankback Has Occurred (ATM Connection Layer) . . . . . . . . . . . . . . . . . . . . . 136 EFCI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Excessive Signaling Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Missing DTL or Excessive DTL Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 MPCMPS Flow Not Allowed on this VC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Possible 3.0/3.1 UNI Mismatch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Signaling Frame Sequence Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Unexpected Data for SVC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 ATM Host Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Excessive Abnormal Disconnects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Excessive Abnormal Frame Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Excessive Short Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 High Signaling Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Host Address Cannot Be Mux’ed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Too Many Simultaneous Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Unexpected Data for SVC (ATM Host Layer) . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Wireless Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 ACK Frame Timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Association Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Authentication Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 CTS Frame Timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Deauthentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Disassociation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Mcast/Bcast Fragmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Missing Fragment Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Oversized WLAN Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Reassociation Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 Rogue Access Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 Rogue Mobile Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Runt WLAN Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Expert Alarms Reference Guide xiii Contents Same Transmitter and Receiver Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Transmitter Address Is Broadcast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Transmitter Address Is Multicast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 WEP-ICV Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Global Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Bad CRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Broadcast/Multicast Storm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Broadcast/Multicast Storm Diag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Channel Mismatch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Collisions Over Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 LAN Overload . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 LAN Overload Percentage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 PLCP Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Spanning Tree Topology Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 VLAN Not Operational . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 VLAN Removed from Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 VTP Versions Different . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Route Layer Alarms and Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Nonsense Route . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Route Metric Changed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Route Superseded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Route Validity Changed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 xiv Sniffer Technologies Preface This guide describes the alarms generated by the Expert analyzer, as well as the thresholds used to generate them. The Expert analyzer is included as part of Sniffer® Portable. For information about your Sniffer application and how to use its Expert features, refer to the product User’s Guide and the online help files. Audience This guide is designed and produced for Network Administrators, Network Engineers, and Information Technology Managers who are responsible for monitoring, capturing, and analyzing data from their corporate networks. Getting More Information Source Contents Sniffer Portable Installation Guide Provides the system requirements and installation instructions for Sniffer Portable and Sniffer Portable enhanced drivers. Sniffer Portable User’s Guide Provides a comprehensive overview of all Sniffer Portable features. Sniffer Portable Expert Alarms Reference Guide Describes each of the alarms generated by Sniffer Portable’s Expert analyzer, along with their related thresholds. Switch Expert Guide Describes how to connect and configure Sniffer Portable to use Switch Expert features. Sniffer Mobile Operations Guide Provides information specific to configuring, and operating Sniffer Mobile. Sniffer Mobile provides decodes and Expert Analysis for Mobile IP protocols. Sniffer Reporter Describes how to install and configure Sniffer Reporter to generate a wide variety of reports based on data collected by Sniffer network analysis and monitoring products. Expert Alarms Reference Guide xv Preface xvi Source Contents Sniffer Tool Collection’s Sniffer Focused Analysis Guide Describes how to use Sniffer Focused Analysis to leverage existing Sniffer trace files for additional analysis and troubleshooting. Sniffer Tool Collection’s Sniffer Capture Format Converter Describes how to use Sniffer Capture Format Converter to convert existing third-party trace files to .cap format. Sniffer Wireless Guide Describes how to install, configure, and operate Sniffer Portable with a supported wireless network adapter. Sniffer Voice Operations Guide Provides information specific to configuring and operating Sniffer Voice. Sniffer Voice provides decodes and Expert analysis for Voice over IP (VoIP) protocols. ATM Adapter Reference Guide Describes how to install, connect, and configure Sniffer Portable when using ATM hardware. Describes ATM interface pods. ATMbook Reference Guide Describes how to install, connect, and configure the ATMbook to capture and generate packets using Sniffer Portable. Full Duplex 10/100 Ethernet Reference Guide Describes how to install, connect, and configure Sniffer Portable when using the Full Duplex Ethernet PCI adapter. Upgrading the Full Duplex 10/100 Ethernet PCI Adapter Guide Describes how to use the FlashUpd utility provided with Sniffer Portable to upgrade the FPGA firmware used on the FDX 10/100 Ethernet PCI adapter. Gigabit Ethernet Reference Guide Describes how to install, connect, and configure Sniffer Portable when using the Gigabit Ethernet PCI adapter. WAN Adapter Cards Reference Guide Describes how to install, connect, and configure Sniffer Portable when using the LM2000 or HSSI adapter. Snifferbook Ultra Reference Guide Describes how to install and configure the Snifferbook Ultra unit and optional Phys. Snifferbook Reference Guide Describes how to install, connect, and configure Sniffer Portable when using the Snifferbook. Sniffer Technologies Preface Source Contents Help Product information that is accessed from within the application. Release Notes • The Help system provides high-level and detailed information. Access from either the Help menu option or the Help button in the application. • Context-sensitive (also called What’s This?) Help provides brief descriptions of the selections in the application. Access by right-clicking an option, pressing the [F1] control key, or dragging the question icon to an option. README file. Product information, system requirements, resolved issues, any known issues, and last-minute additions or changes to the product or its documentation. Expert Alarms Reference Guide xvii Preface Contacting Network General Customer Service Get help with license entitlement, registrations, grant number inquiries, tech support validation and more by contacting the Network General Customer Service department at: North America phone: 1-800-764-3337 (1-800-SNIFFER) Email: [email protected] Web: http://www.networkgeneral.com/ContactUs.aspx The department's hours of operation are 7:00 AM to 7:00 PM Central time, Monday through Friday. International phone numbers: http://www.networkgeneral.com/ContactUs.aspx Latin America: +55 (11) 5180-6643 Europe: +44 (0) 1753 217500 Australia/New Zealand: +61 (2) 9761 4200 Asia: +65 6222 7555 Japan: +81 3-5219-1221 Mail: Network General Corporation (North America) Customer Service Department Mail Stop 2S362 5000 Headquarters Drive Plano, TX 75024 USA xviii Sniffer Technologies Preface Customer Service International Address Network General International BV (EMEA) Customer Service Department PO Box 58326 1040 HH Amsterdam The Netherlands Technical Support Visit Network General Technical Support at: • Sniffer University http://www.networkgeneral.com/TechnicalSupport.aspx Sniffer University is a comprehensive educational resource for building and enhancing all network professionals' skills in fault and performance management. Sniffer University has trained over 70,000 network professionals worldwide. The Sniffer Certified Professional Program provides network professionals industry-recognized accreditation as experts in their field. For more information: Consulting Services • Toll-free: 866-764-3337 • Email: [email protected] • Web: http://www.networkgeneral.com/SnifferUniversity.aspx Our consultants provide an expert supplemental resource and independent perspective to resolve your problems. They are ready to assist you during all stages of network growth, from planning and design, through implementation, and with ongoing management. They will help integrate our products into your environment and troubleshoot or baseline network performance. Our consultants also develop and deliver custom solutions to help accomplish your project goals. Currently, custom and product consulting are available. For more information: • http://www.networkgeneral.com/Consulting.aspx Expert Alarms Reference Guide xix Preface xx Sniffer Technologies Exper t Alarms and Thresholds 1 This section describes the alarms generated by the Expert analyzer along with their corresponding thresholds. You can use this section to improve your understanding of Expert alarms, as well as how to configure the thresholds that determine when they are generated. Alarms in this section are organized by the network layer at which they occur. Within each network layer, alarms are organized alphabetically. Each alarm is described along with its corresponding thresholds (if any). About Expert Alarms During Expert analysis, the Expert constructs a database of network objects from the traffic it sees. The Expert protocol interpreters learn all about the network stations, routing nodes, subnetworks, and connections related to the frames in the capture buffer. Using this information, the Expert detects and alerts you to potential problems that may exist on the network by generating Expert alarms and reporting them in the Expert display as either symptoms or diagnoses (depending on their severity). A symptom indicates that a threshold has been exceeded and may indicate a problem on your network. A diagnosis can be several symptoms analyzed together, high rates of recurrence of specific symptoms, or single instances of particular network events that cause the Expert to conclude that the network has a real problem. A diagnosis should be investigated immediately. Expert alarms are reported in the Expert display. An example is shown in Figure 1-1. Expert Alarms Reference Guide 1 Chapter 1 Selecting an alarm in the Summary Pane lets you see detailed alarm information in the Detail Pane (including a “?' link to the Expert Explain file for the alarm). Expert alarm counts are provided at each layer in the Overview Pane of the Expert display. Figure 1-1. Expert Alarms in the Expert Display Information Reported with Expert Alarms Applications supporting Sniffer Reporter can be configured to report Expert alarms: In the Alarm Log To a third party SNMP Console (as SNMP traps) Sniffer Portable enhances and extends the information included with each reported Expert alarm. Expert alarms are reported with the following information: 2 Name of the alarm Client IP address and port number Server IP address and port number IP address of the Sniffer Portable installation reporting the alarm Name of the adapter on the Sniffer Portable installation reporting the alarm Time and date when the alarm was detected Sniffer Technologies Expert Alarms and Thresholds Figure 1-2 shows an example of an enhanced Expert alarm entry in the Alarm Manager. Expert alarms are reported by name with addresses, ports, timestamps, and identifying information for the reporting Appliance. Figure 1-2. Enhanced Information in Reported Expert Alarms (Alarm Manager Shown) About Expert Thresholds Expert thresholds determine whether the Expert generates a symptom or a diagnosis (also called an alarm) based on a given network event. The default thresholds supplied with the Expert have been carefully calculated to ensure accurate and informative symptom and diagnosis detection. Although this manual will provide you with valuable information on understanding Expert thresholds, you should be sure to understand your network before changing any of the thresholds. Setting Thresholds To set Expert thresholds: 1 Select Expert Options from the Tools menu. 2 Click the Alarms tab. The Alarms tab lists all Expert alarms by layer, along with their corresponding thresholds. The Alarms tab and further instructions are shown in Figure 1-3. Expert Alarms Reference Guide 3 Chapter 1 Click to expand/collapse all Expert layers. Click to set the severity of this alarm. Different severities can be associated with different actions. Click the + to open an Expert layer and display all symptoms and diagnoses (alarms). Click to specify whether the alarm should appear in the alarm log. Click in the Threshold Value cell and type the new threshold value. Click the + to display the settings for this alarm. The thresholds display at the end of the settings list. Click to reset the selected value to the factory default. Figure 1-3. Setting Expert Thresholds 4 Sniffer Technologies Click to reset all settings for all layers to the factory defaults. Expert Alarms and Thresholds Expert Alarms and Thresholds Descriptions This section describes each Expert alarm generated by the Expert analyzer, along with its corresponding threshold (if any). Alarms are organized by Expert layer, from the highest layer (Service) to the lowest (Route). Within each layer, alarms are organized alphabetically. This follows the same organization shown in the Alarms tab of the Expert UI Object Properties dialog box shown in Figure 1-5. Table 1-1 summarizes this organization Table 1-1. Location of Alarm Descriptions at each Layer Alarms At This Layer... ...Are Described Starting Here Service Layer page 6 Application Layer page 12 Session Layer page 34 Connection Layer page 65 Station Layer page 72 DLC Layer page 82 WAN Cnx Layer page 92 WAN Link Layer page 95 ATM Application Layer page 100 ATM Flow Layer page 105 ATM Connection Layer page 134 ATM Host Layer page 140 Wireless Layer page 143 Global Layer page 152 Route Layer page 158 Expert Alarms Reference Guide 5 Chapter 1 Service Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Service layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Slow FTP Server The Expert generates this alarm when the average transaction time observed for an FTP server exceeds the Slow FTP Server Time threshold in the Alarms tab of the Expert UI Object Properties dialog box. At the Service layer, the Expert observes transactions for individual servers (in this case, an FTP server). For each transaction the server takes part in (for example, logins, file requests, and so on), the Expert measures the time it takes to complete the transaction. Transactions are measured from the first frame of a request to the last frame of the corresponding response. When the average time of all such transactions observed for a single FTP server exceeds the Slow FTP Server Time threshold, the Expert generates this alarm. Possible causes: There is a problem with the FTP server. There is excessive network congestion. Slow HTTP Server The Expert generates this alarm when the average transaction time observed for an HTTP server exceeds the Slow Internet Server Time threshold in the Alarms tab of the Expert UI Object Properties dialog box. At the Service layer, the Expert observes transactions for individual servers (in this case, an HTTP server). For each transaction the server takes part in (for example, GETs, POSTs, and so on), the Expert measures the time it takes to complete the transaction. Transactions are measured from the first frame of a request to the last frame of the corresponding response. When the average time of all such transactions observed for a single HTTP server exceeds the Slow Internet Server Time threshold, the Expert generates this alarm. Possible causes: 6 The network is experiencing congestion. The web server is overloaded. Sniffer Technologies Expert Alarms and Thresholds Slow Mail Server The Expert generates this alarm when the average transaction time observed for a SMTP server exceeds the Slow Mail Server Time threshold in the Alarms tab of the Expert UI Object Properties dialog box. At the Service layer, the Expert observes transactions for individual servers (in this case, an SMTP server). For each transaction the server takes part in (for example, logins, get mail commands, and so on), the Expert measures the time it takes to complete the transaction. Transactions are measured from the first frame of a request to the last frame of the corresponding response. When the average time of all such transactions observed for a single SMTP server exceeds the Slow Mail Server Time threshold, the Expert generates this alarm. Possible causes: There is a problem with the SMTP server. There is excessive network congestion. Slow News Server The Expert generates this alarm when the average transaction time observed for an NNTP server exceeds the Slow News Server Time threshold in the Alarms tab of the Expert UI Object Properties dialog box. At the Service layer, the Expert observes transactions for individual servers (in this case, an NNTP server). For each transaction the server takes part in (for example, logins, ARTICLE commands, and so on), the Expert measures the time it takes to complete the transaction. Transactions are measured from the first frame of a request to the last frame of the corresponding response. When the average time of all such transactions observed for a single NNTP server exceeds the Slow News Server Time threshold, the Expert generates this alarm. Possible causes: There is a problem with the NNTP server. There is excessive network congestion. Slow Oracle SQL Server The Expert generates this alarm when the average transaction time observed for an Oracle SQL server exceeds either the Slow Oracle Server Time w/Cursors threshold (for transactions using cursors) or the Slow Oracle Server Time threshold (for transactions that do not use cursors) in the Alarms tab of the Expert UI Object Properties dialog box. Expert Alarms Reference Guide 7 Chapter 1 At the Service layer, the Expert observes transactions for individual servers (in this case, an Oracle SQL server). For each transaction the server takes part in, the Expert measures the time it takes to complete the transaction. Transactions using cursors are measured from the start of an Open Cursor statement to the cursor's close. Transactions that do not use cursors are measured from the first frame of a request to the last frame of the corresponding response. When the average time of either type of transaction observed for a single Oracle SQL server exceeds the corresponding threshold (Slow Oracle Server Time w/Cursors for transactions using cursors; Slow Oracle Server Time for transactions that do not use cursors), the Expert generates this alarm. Possible causes: 8 Sniffer Technologies The problem is network related. Determine if your network is overloaded. Increase your network's bandwidth. Determine if your network is optimally configured. Minimize frame fragmentation by using larger frame sizes or more efficient protocols. Reduce polling frequency of polled devices in your network. The problem is in the Multiprotocol Interchange. Configure your MPI for better data throughputs rather than for connections. Decrease the number of pumps per MPI. Add more MPIs. The problem is in the server. Assess the current utilization of your server. Determine if you have enough CPU power to support the number of users in this server. Optimize the file distribution of your database. Place the tables and their associated indexes in separate disk drives. Tables that are most frequently referenced should be placed in separate disk drives. Determine if your server has enough memory to support the applications and services that have to run concurrently. Install faster disk drives. Check for runaway (orphan) processes and clean up after them. The problem is in the database. Perform database tuning suggested by Oracle and other experts in this area. For example: Expert Alarms and Thresholds Make sure that the shared pool is large enough to reduce reparsing. Optimize the size of the DB_BLOCK_BUFFERS. Increase the value of DML_LOCKS to allow more locks placed on an object at one time. Determine your library cache misses and set the CURSOR_SPACE_FOR_TIME according to the recommendation in the Oracle7 Server Administrator's Guide. Check if indexed clusters can speed up your access to tables that are commonly joined together on a standard set of matching columns. The problem is in the client workstation. Use optimized query statements. The “explain plan” command will show the access paths to the data. Remove all unnecessary “lock table” statements. Do not use “for update” clause in select statements if there is no need to update immediately. Issue commit statements immediately after “update”, “delete”, and “insert” statements to release locked resources. Use explicit cursors in all of your PL/SQL blocks. Identical query statements in your application must match character-by-character to use the same shared pool area. Slow POP3 Server The Expert generates this alarm when the average transaction time observed for a POP3 server exceeds the Slow POP Server Time threshold in the Alarms tab of the Expert UI Object Properties dialog box. At the Service layer, the Expert observes transactions for individual servers (in this case, a POP3 server). For each transaction the server takes part in (for example, logins, send mail commands, and so on), the Expert measures the time it takes to complete the transaction. Transactions are measured from the first frame of a request to the last frame of the corresponding response. When the average time of all such transactions observed for a single POP3 server exceeds the Slow POP Server Time threshold, the Expert generates this alarm. Possible causes: There is a problem with the POP3 server. Expert Alarms Reference Guide 9 Chapter 1 There is excessive network congestion. Slow Sybase/Microsoft SQL Server The Expert generates this alarm when the average transaction time observed for a Sybase or Microsoft TDS server exceeds either the Slow TDS Server Time w/Cursors threshold (for transactions using cursors) or the Slow TDS Server Time threshold (for transactions that do not use cursors) in the Alarms tab of the Expert UI Object Properties dialog box. At the Service layer, the Expert observes transactions for individual servers (in this case, a TDS server). For each transaction the server takes part in, the Expert measures the time it takes to complete the transaction. Transactions using cursors are measured from the start of an Open Cursor statement to the cursor's close. Transactions that do not use cursors are measured from the first frame of a request to the last frame of the corresponding response. When the average time of either type of transaction observed for a single TDS server exceeds the corresponding threshold (Slow TDS Server Time w/Cursors for transactions using cursors; Slow TDS Server Time for transactions that do not use cursors), the Expert generates this alarm. Possible causes: The problem is network related. Determine if your network is overloaded. Increase your network's bandwidth. Determine if your network is optimally configured. Minimize frame fragmentation by using larger frame sizes or more efficient protocols. Reduce polling frequency of polled devices in your network. The problem is in the Multiprotocol Interchange. Configure your MPI for better data throughputs rather than for connections. Decrease the number of pumps per MPI. Add more MPIs. The problem is in the server. 10 Sniffer Technologies Assess the current utilization of your server. Determine if you have enough CPU power to support the number of users in this server. Expert Alarms and Thresholds Optimize the file distribution of your database. Place the tables and their associated indexes in separate disk drives. Tables that are most frequently referenced should be placed in separate disk drives. Determine if your server has enough memory to support the applications and services that have to run concurrently. Install faster disk drives. The problem is in the client workstation. Use optimized query statements. The “explain plan” command will show the access paths to the data. Remove all unnecessary “lock table” statements. Do not use “for update” clause in select statements if there is no need to update immediately. Issue commit statements immediately after “update”, “delete”, and “insert” statements to release locked resources. Use explicit cursors in all of your PL/SQL blocks. Identical query statements in your application must match character-by-character to use the same shared pool area. Check for runaway (orphan) processes and clean up after them. The problem is in the database. Perform database tuning suggested by experts in this area. For example: Make sure that the shared pool is large enough to reduce reparsing. Check if indexed clusters can speed up your access to tables that are commonly joined together on a standard set of matching columns. Slow Telnet Server The Expert generates this alarm when the average transaction time observed for an Telnet server exceeds the Slow Telnet Server Time threshold in the Alarms tab of the Expert UI Object Properties dialog box. At the Service layer, the Expert observes transactions for individual servers (in this case, a Telnet server). For each transaction the server takes part in (for example, logins, file requests, and so on), the Expert measures the time it takes to complete the transaction. Transactions are measured from the first frame of a request to the last frame of the corresponding response. When the average time of all such transactions observed for a single Telnet server exceeds the Slow Telnet Server Time threshold, the Expert generates this alarm. Expert Alarms Reference Guide 11 Chapter 1 Possible causes: There is a problem with the Telnet server. There is excessive network congestion. Application Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Application layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). [VOIP] H323 - High Call Volume The Expert generates the H323 – High Call Volume alarm when the number of simultaneous H.323 calls on the network exceeds the Max concurrent calls threshold. This threshold is found under the [VOIP] H323 – High call volume alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Expert Options from the Tools menu). During capture, the Expert maintains a counter of the number of H.323 calls currently open. Each time a new H.323 call is opened, the Expert increments the counter. Each time a call is released (that is, the Expert sees the H.225 Release Complete message for the call), the counter is decremented. Every fifteen seconds, the Expert checks this counter to see if it exceeds the value specified in the Expert UI Object Properties dialog box for the Max concurrent calls threshold. If the number of open H.323 calls exceeds the threshold, the Expert generates the H323 – High Call Volume alarm. NOTE: Be sure to set the Max concurrent calls threshold to a value that makes sense for your network. If the Expert is generating multiple instances of this alarm but your network is not experiencing any other difficulties, you may want to increase the value of the threshold so that the alarm is only generated when the call volume becomes problematic for your network. [VOIP] H323 - Too Many Incomplete Calls The Expert generates the H323 – Too Many Incomplete Calls alarm when the number of incomplete H.323 calls observed by the Expert exceeds the Max incomplete calls threshold. This threshold is found under the [VOIP] H323 – Too many incomplete calls alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Tools > Expert Options). 12 Sniffer Technologies Expert Alarms and Thresholds The Expert considers an H.323 call to be incomplete if more than three minutes elapses after the last frame observed for the call and it doesn't see a call termination message (H.225 Release Complete) for the call. The Expert keeps a counter of these incomplete calls. Every fifteen seconds, the Expert checks the incomplete calls counter to see if it exceeds the value specified in the Expert UI Object Properties dialog box for the Max incomplete calls threshold. If the number of incomplete H.323 calls exceeds the threshold, the Expert generates the H323 – Too Many Incomplete Calls alarm. Possible cause: Incomplete calls can be caused by unexpected shutdowns of network equipment (for example, because a device was turned off), abnormal channel disconnections, or other network problems. [VOIP] SCCP - High Call Volume The Expert generates the SCCP – High Call Volume alarm when the number of simultaneous calls on the network using Cisco's Skinny Client Control Protocol (SCCP) exceeds the Max concurrent calls threshold. This threshold is found under the [VOIP] SCCP – High call volume alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Tools > Expert Options). During capture, the Expert maintains a counter of the number of SCCP calls currently open. Each time a new SCCP call is opened, the Expert increments the counter. Each time a call is released (that is, the Expert sees the SCCP StationStopMediaTransmission message for the call), the counter is decremented. Every fifteen seconds, the Expert checks this counter to see if it exceeds the value specified in the Expert UI Object Properties dialog box for the Max concurrent calls threshold. If the number of open SCCP calls exceeds the threshold, the Expert generates the SCCP – High Call Volume alarm. NOTE: Be sure to set the Max concurrent calls threshold to a value that makes sense for your network. If the Expert is generating multiple instances of this alarm but your network is not experiencing any other difficulties, you may want to increase the value of the threshold so that the alarm is only generated when the number of calls becomes problematic for your network. Expert Alarms Reference Guide 13 Chapter 1 [VOIP] SCCP - Too Many Incomplete Calls The Expert generates the SCCP – Too Many Incomplete Calls alarm when the number of incomplete calls using Cisco's Skinny Client Control Protocol (SCCP) observed by the Expert exceeds the Max incomplete calls threshold. This threshold is found under the [VOIP] SCCP – Too many incomplete calls alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Tools > Expert Options). The Expert considers an SCCP call to be incomplete if more than three minutes elapses after the last frame observed for the call and it doesn't see a call termination message (SCCP StationStopMediaTransmission) for the call. The Expert keeps a counter of these incomplete calls. Every fifteen seconds, the Expert checks the incomplete calls counter to see if it exceeds the value specified in the Expert UI Object Properties dialog box for the Max incomplete calls threshold. If the number of incomplete SCCP calls exceeds the threshold, the Expert generates the SCCP – Too Many Incomplete Calls alarm. Possible cause: Incomplete calls can be caused by unexpected shutdowns of network equipment (for example, because a device was turned off), abnormal channel disconnections, or other network problems. [VOIP] SIP - High Call Volume The Expert generates the SIP – High Call Volume alarm when the number of simultaneous SIP calls on the network exceeds the Max concurrent calls threshold. This threshold is found under the [VOIP] SIP – High call volume alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Tools > Expert Options). During capture, the Expert maintains a counter of the number of SIP calls currently open. Each time a new SIP call is opened, the Expert increments the counter. Each time a call is released (that is, the Expert sees the SIP BYE message for the call), the counter is decremented. Every fifteen seconds, the Expert checks this counter to see if it exceeds the value specified in the Expert UI Object Properties dialog box for the Max concurrent calls threshold. If the number of open SIP calls exceeds the threshold, the Expert generates the SIP – High Call Volume alarm. NOTE: Be sure to set the Max concurrent calls threshold to a value that makes sense for your network. If the Expert is generating multiple instances of this alarm but your network is not experiencing any other difficulties, you may want to increase the value of the threshold so that the alarm is only generated when the number of calls becomes problematic for your network. 14 Sniffer Technologies Expert Alarms and Thresholds [VOIP] SIP - Too Many Incomplete Calls The Expert generates the SIP – Too Many Incomplete Calls alarm when the number of incomplete SIP calls observed by the Expert exceeds the Max incomplete calls threshold. This threshold is found under the [VOIP] SIP – Too many incomplete calls alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Tools > Expert Options). The Expert considers a SIP call to be incomplete if more than three minutes elapses after the last frame observed for the call and it doesn't see a call termination message (SIP BYE) for the call. The Expert keeps a counter of these incomplete calls. Every fifteen seconds, the Expert checks the incomplete calls counter to see if it exceeds the value specified in the Expert UI Object Properties dialog box for the Max incomplete calls threshold. If the number of incomplete SIP calls exceeds the threshold, the Expert generates the SIP – Too Many Incomplete Calls alarm. Possible cause: Incomplete calls can be caused by unexpected shutdowns of network equipment (for example, because a device was turned off), abnormal channel disconnections, or other network problems. Access to Resource Denied The Expert generates the Access to Resource Denied alarm when a station fails to establish an SMB session due to access restrictions. The name of the station is listed in the corresponding summary row. Possible cause: The station has attempted to connect to a system resource that is either not sharable or that has reached its limit of simultaneously sharing users. Check with the user/administrator of the resource to gain access rights. A typical situation in which this alarm might occur is when a user is scanning the Network Neighborhood, attempting to see what resources outside of their domain they might be able to access. Browser Election Force The Expert generates the Browser Election Force alarm when a node sends a Browser Force Election request. Periodic election forces are normal in Microsoft networks. The node has not seen a local master browser and is initiating the process to become one itself. The name of the station is listed in the corresponding summary row. Expert Alarms Reference Guide 15 Chapter 1 Possible causes: The listed node's local network currently has no master browser to service its browser requests. Check the current settings for the node's subnet mask, IP address, and source-level routing to ensure that its network interface is properly configured. No other nodes running the browser service are configured to run as a master browser. In this case, the node sending the Browser Force Election request should become the local Master Browser. Some nodes (such as LINUX boxes running SAMBA) can configure Browser Roles. DB Connect Request Denied The Expert generates the DB Connect Request Denied alarm when the number of SQL Server connect requests denied exceeds the DB Connection Failed threshold. A request to login is refused by the server, because it is unable to satisfy the options asked for in the login record. Examine the service options in the connect packet and determine whether they are options that are allowed in the target server. Possible causes: Invalid username. Invalid password. Invalid server. DB Security Breach Attempt The Expert generates the DB Security Breach Attempt alarm when the number of login failures exceeds the DB Security Breach Attempt threshold. Possible causes: Invalid username. Invalid password. Invalid server. DB Slow Connect The Expert generates the DB Slow Connect alarm when the time it takes a SQL Server to respond to a request to login to a SQL Server exceeds the DB Slow Connect Time threshold. This may or may not be normal, depending upon the application and the setting of the DB Slow Connect Time threshold. 16 Sniffer Technologies Expert Alarms and Thresholds Possible causes: The server is too busy. If this occurs frequently in a server with several users already active, consider taking performance tuning steps to improve the server response. There is a lot of traffic on your network. Other users may be trying to connect to the server at the same time, thereby reducing the response time. The DB Slow Connect Time threshold may be set too low for your environment. DB Slow Server Response The Expert generates the DB Slow Server Response alarm when the time it takes for a server to respond to one or more SQL commands exceeds the DB Slow Server Response Time threshold. Possible causes: The problem is network related. Determine if your network is overloaded. Increase your network's bandwidth. Determine if your network is optimally configured. Minimize frame fragmentation by using larger frame sizes or more efficient protocols. Reduce polling frequency of polled devices in your network. The problem is in the Multiprotocol Interchange. Configure your MPI for better data throughputs rather than for connections. Decrease the number of pumps per MPI. Add more MPIs. The problem is in the server. Assess the current utilization of your server. Determine if you have enough CPU power to support the number of users in this server. Optimize the file distribution of your database. Place the tables and their associated indexes in separate disk drives. Tables that are most frequently referenced should be placed in separate disk drives. Expert Alarms Reference Guide 17 Chapter 1 Determine if your server has enough memory to support the applications and services that have to run concurrently. Install faster disk drives. The problem is in the client workstation. Use optimized query statements. The “explain plan” command will show the access paths to the data. Remove all unnecessary “lock table” statements. Do not use “for update” clause in select statements if there is no need to update immediately. Issue commit statements immediately after “update”, “delete”, and “insert” statements to release locked resources. Use explicit cursors in all of your PL/SQL blocks. Identical query statements in your application must match character-by-character to use the same shared pool area. Check for runaway (orphan) processes and clean up after them. The problem is in the database. Perform database tuning suggested by experts in this area. For example: Make sure that the shared pool is large enough to reduce reparsing. Check if indexed clusters can speed up your access to tables that are commonly joined together on a standard set of matching columns. DB Slow Server Response Diagnosis The Expert considers a server's response to an SQL request to be slow when the time between request and response is longer than the DB Slow Server Connect Time threshold. The Expert counts these slow responses. When the total number of slow responses seen during the time period specified by the DB Slow Server Response Interval threshold exceeds the DB Slow Server Response Count threshold, the Expert generates the DB Slow Server Response Diagnosis. Possible causes: The problem is network related. 18 Sniffer Technologies Determine if your network is overloaded. Increase your network's bandwidth. Expert Alarms and Thresholds Determine if your network is optimally configured. Minimize frame fragmentation by using larger frame sizes or more efficient protocols. Reduce polling frequency of polled devices in your network. The problem is in the Multiprotocol Interchange. Configure your MPI for better data throughputs rather than for connections. Decrease the number of pumps per MPI. Add more MPIs. The problem is in the server. Assess the current utilization of your server. Determine if you have enough CPU power to support the number of users in this server. Optimize the file distribution of your database. Place the tables and their associated indexes in separate disk drives. Tables that are most frequently referenced should be placed in separate disk drives. Determine if your server has enough memory to support the applications and services that have to run concurrently. Install faster disk drives. The problem is in the client workstation. Use optimized query statements. The “explain plan” command will show the access paths to the data. Remove all unnecessary “lock table” statements. Do not use “for update” clause in select statements if there is no need to update immediately. Issue commit statements immediately after “update”, “delete”, and “insert” statements to release locked resources. Use explicit cursors in all of your PL/SQL blocks. Identical query statements in your application must match character-by-character to use the same shared pool area. Check for runaway (orphan) processes and clean up after them. The problem is in the database. Perform database tuning suggested by experts in this area. For example: Make sure that the shared pool is large enough to reduce reparsing. Expert Alarms Reference Guide 19 Chapter 1 Check if indexed clusters can speed up your access to tables that are commonly joined together on a standard set of matching columns. Excessive Failed Resource Login Attempts The Expert generates the Excessive Failed Resource Login Attempts alarm when the number of consecutive login attempts with an incorrect password or user name for the listed station exceeds the Excessive Failed Resource Login Attempts threshold. The address of the offending station is listed in the Summary row. Possible causes: A user has forgotten their password and continues to try different dimly recalled passwords. An attempt to guess passwords for a given user ID has occurred. Repeated alarms may indicate an attempted security breach (for example, a program that tries to guess user names and passwords for a given node may be running). Excessive Mailslot Broadcasts The Expert generates the Excessive Mailslot Broadcasts alarm when the number of mailslot broadcast messages sent by a station in the last second exceeds the Excessive Mailslot Broadcasts threshold. The address of the offending station is listed in the Summary row. Excessive mailslot broadcasts can indicate poorly-written software and should be investigated. Possible causes: Faulty software resulting in a burst of broadcast messages. Typically, there should be a delay between consecutive broadcasts so the network is not flooded with mostly useless information. Excessive broadcasts to networks reachable through a router can cause the router to drop packets. FCIP Special Frame Discovery The Expert generates the FCIP Special Frame Discovery alarm when it detects an FCIP Special Frame (FSF) with the field for the destination entity's World Wide name set to zero. Source FCIP entities send FSF frames such as this to discover the World Wide name of the destination FCIP entity. This is not a serious alarm. However, you should keep in mind that FCIP connections cannot be set up in response to the receipt of an FSF with the destination entity's World Wide name field set to zero. 20 Sniffer Technologies Expert Alarms and Thresholds Possible causes: The source entity does not know the destination entity's World Wide Name. FCIP Invalid Encapsulation The Expert generates the FCIP Invalid Encapsulation alarm when it detects an FCIP frame with errors in its encapsulation. The Expert detects encapsulation errors by examining the 1's Complement fields in FCIP frames. Any FCIP frame with errors in any of the 1's complement fields will be dropped by the receiving FCIP entity. In an FCIP frame, the 1's complement values are used to ensure the consistency of the data being transmitted. This is irrespective of whether a CRC is present or not. Possible causes: There were transmission errors. The packet was encoded incorrectly in duplicated or 1’s Complement fields. FCIP Repeat Special Frame The Expert generates the FCIP Repeat Special Frame alarm when it detects more than one FCIP Special Frame (FSF) sent in either direction on an FCIP connection. Typically, an FSF is transmitted only once for each successful FCIP connection established. A repeated FSF may cause the connection setup between the two FCIP entities to fail or the connection to flap. Possible causes: The FCIP source entity may be doing discovery using the FCIP Special Frame. This may be a Replay Attack to disrupt the FCIP connection. FCIP Special Frame Echo Has Changed The Expert generates the FCIP Special Frame Echo Has Changed alarm when it detects that an FCIP entity echoed back an FCIP Special Frame (FSF) that was different than the FSF it originally received. In FCIP, for a connection to be set up correctly, the FCIP destination entity has to echo back the FSF to the source entity without any changes. A changed Echo sent in response to an FSF will cause the FCIP connection to break. Expert Alarms Reference Guide 21 Chapter 1 Possible causes: The FCIP destination entity cannot accept any more connections. The connection parameters in the FSF are not acceptable to the FCIP destination entity. This is a reply to FCIP Special Frame Discovery for the destination entity’s World Wide Name. FCIP Special Frame Delayed The Expert generates the FCIP Special Frame Echo Delayed alarm when the FCIP Special frame echo has not been received within the amount of time specified by the Special Echo Frame Delayed threshold. Possible causes: The FCIP destination Entity is too busy. The echo FCIP SF has been dropped somewhere in the network. FCIP Too Many Special Frame Echoes The Expert generates the FCIP Too Many Special Frame Echoes alarm when the number of FCIP Special Frame (FSF) echoes transmitted by the FCIP Destination entity exceeds the Frames Contain Errors threshold. The presence of more than one FSF echo will cause immediate tear down of the FCIP connection. Possible causes: The destination FCIP entity needs to close the FCIP connection. This may be caused by a replay attack. FCIP Transmission Errors The Expert generates the FCIP Transmission Errors alarm when the number of transmission errors detected for an FCIP connection exceeds the Frames Contain Errors threshold. Too many transmission errors on an FCIP connection will cause loss of throughput for the connection. Possible causes: 22 Sniffer Technologies Causes will depend on the physical media. For example, on a wireless network, this can be caused by signal attenuation. Expert Alarms and Thresholds FTP Login Attempts The Expert generates the FTP Login Attempts alarm when the number of FTP login attempts for a station exceeds the FTP Connect Tries threshold. Possible causes: The login ID is incorrect. The login password is incorrect. There is excessive network congestion. FTP Slow Connect The Expert generates the FTP Slow Connect alarm when the amount of time it takes an FTP server to acknowledge a login request exceeds the FTP Connect Time threshold. Possible causes: There is a problem with the FTP server. There is excessive network congestion. FTP Slow First Response When an FTP session is established between a server and client, the server's first response to an FTP command is usually the slowest. This is because the server typically has to open or create files in response to the first command. Subsequent responses typically do not take as much time. The FTP Slow First Response alarm is generated when a response to a client's first command exceeds the FTP Login Time threshold. FTP Slow Response When an FTP session is established between a server and client, the server's first response to an FTP command is usually the slowest. This is because the server typically has to open or create files in response to the first command. Subsequent responses typically do not take as much time. The FTP Slow Response alarm is generated when a response to a command after the first one exceeds the FTP Interframe Time threshold. Expert Alarms Reference Guide 23 Chapter 1 FTP Slow Transfer Diagnosis The Expert considers an FTP server's response to an FTP request to be slow when the time between request and response is longer than either the FTP Login Time threshold (for the response to the first request in the session) or the FTP Interframe Time threshold (for responses to all subsequent requests during the session). The Expert counts these slow responses. When the total number of slow responses seen during the time period specified by the FTP Interframe Interval threshold exceeds the FTP Interframe Count threshold, the Expert generates the FTP Slow Transfer Diagnosis. HTTP Slow Server GET Response Time The Expert generates the HTTP Slow Server GET Response Time alarm when an HTTP server's response to a GET request takes longer than the HTTP Slow Server GET Response Time threshold to reach the requesting station. When a station sends an HTTP GET request, the Expert starts a counter. If the counter exceeds the HTTP Slow Server GET Response Time threshold before the HTTP server returns a response, the Expert generates the alarm. This alarm can be useful when measuring the responsiveness of a web server. Possible causes: The network is experiencing congestion. The web server is overloaded. HTTP Slow Server POST Response Time The Expert generates the HTTP Slow Server POST Response Time alarm when an HTTP server's response to a POST message takes longer than the HTTP Slow Server POST Response Time threshold to reach the sending station. When a station sends an HTTP POST message, the Expert starts a counter. If the counter exceeds the HTTP Slow Server POST Response Time threshold before the HTTP server returns a response, the Expert generates the alarm. This alarm can be useful when measuring the responsiveness of a web server. Possible causes: 24 The network is experiencing congestion. The web server is overloaded. Sniffer Technologies Expert Alarms and Thresholds Invalid Network Name The Expert generates the Invalid Network Name alarm when a station fails to establish an SMB session because it is specifying a network resource name that does not exist on the station with which it is attempting to communicate. Possible cause: The client has specified a network resource name that does not exist on the server. Invalid Resource User Name or Password The Expert generates the Invalid Resource User Name or Password alarm when a station has fails to establish an SMB session due to an incorrect user name or password. Possible cause: A client running Windows 95 or Windows 98 attempted to connect to a resource using an invalid user name or password. In this case, the client found the resource and attempted to connect to it using an invalid password or user ID. Repeated alarms may indicate an attempted security breach (for example, a program which tries to guess user names and passwords for a given node may be running). iSCSI: Authorization Failure The Expert generates the iSCSI: Authorization Failure alarm when it detects an iSCSI Login Response message with the Status Code set to 0202 Authorization Failure. This status code indicates that the initiator is not allowed access to the target (the initiator’s login request has been rejected by the target). Possible cause: The initiator does not have sufficient privileges to access the target. iSCSI: Can Not Generate Target Transfer Tag - Out of Resources The Expert generates the iSCSI: Can Not Generate Target Transfer Tag (Out of Resources) alarm when it detects an iSCSI Reject message with the Reason Code set to 0x0a - Can’t Generate Target Transfer Tab - Out of Resources. This Reason Code indicates that the target is out of resources and cannot generate a new Target Transfer Tag. Because of this, the target rejects the command from the initiator. Expert Alarms Reference Guide 25 Chapter 1 iSCSI: Command Not Supported in this Session Type The Expert generates the iSCSI: Command Not Supported in this Session Type alarm when it detects an iSCSI Reject message with the Reason Code set to 0x05 - Command Not Supported. This Reason Code indicates that the SCSI command from the initiator is not supported by the target for the specified session type of the connection. Because of this, the target has rejected the initiator command request. iSCSI: Command Retransmission The Expert generates the iSCSI: Command Retransmission alarm when it detects an initiator resending the same iSCSI command PDU (“retry”) in the absence of a command acknowledgement (by way of an ExpCmdSN update) or a response. As described in RFC 3720, this is normal iSCSI behavior when an initiator tries to patch what it thinks are discontinuities in CmdSN ordering on the target end of the connection. This is an informational alarm and does not indicate an error condition. iSCSI: Data Digest Error The Expert generates the iSCSI: Data Digest Error alarm when it detects an iSCSI Reject message with the Reason Code set to 0x02 - Data Digest Error. Data digest errors indicate that data integrity has been compromised. The data (payload) digest protects the integrity of the data. iSCSI: Data or Status again Requested The Expert generates the iSCSI: Data or Status again Requested alarm when it detects an initiator requesting the retransmission of either data or status using a SNACK request. For this to occur successfully, the implementation must support an ErrorRecoveryLevel greater that zero. This is an informational alarm and does not indicate an error condition. iSCSI: Function Authorization Failed The Expert generates the iSCSI: Function Authentication Failed alarm when it detects an iSCSI Task Management Function Response with a response code set to 6 - Function Authorization Failed. This response code may appear if the initiator does not have sufficient privileges to request the specified task management function on the target. 26 Sniffer Technologies Expert Alarms and Thresholds iSCSI: Function Rejected The Expert generates the iSCSI: Function Rejected alarm when it detects an iSCSI Task Management Function Response with a response code set to 255 - Function rejected. This response code indicates that the requested task management function can not be executed on the target and has therefore been rejected. iSCSI: Immediate Command Reject (Too Many Immediate Commands) The Expert generates the iSCSI: Immediate Command Reject (Too Many Immediate Commands) alarm when it detects an iSCSI Reject message with the Reason Code set to 0x06 - Immediate Command Reject. This reason code indicates that the initiator has marked a command as immediate but the target cannot process it immediately because it already has too many immediate commands waiting in the queue to be processed. iSCSI: Initiator Authentication Failed The Expert generates the iSCSI: Initiator Authentication Failed alarm when it detects an iSCSI Login Response message with the Status Code set to 0201 - Authentication Failure. This status code indicates that the initiator could not be successfully authenticated or that target authentication is not supported. Possible causes: Wrong login parameters given by the initiator in login request PDU. Authentication not supported at target. iSCSI: Invalid Data Ack The Expert generates the iSCSI: Invalid Data Ack alarm when it detects an iSCSI Reject message with the Reason Code set to 0x08 - Invalid Data ACK. Rejects with this reason code can occur when an initiator replies to a target’s request for positive acknowledgement of received data with a SNACK using a DataAck type that is invalid for sessions with an ErrorRecoveryLevel of 1 or higher. iSCSI: Invalid PDU field The Expert generates the iSCSI: Invalid PDU field alarm when it detects an iSCSI Reject message with the Reason Code set to 0x09 - Invalid PDU Field. Rejects with this reason code indicate that a request PDU contained wrong or invalid values for fields that are meant to describe a task, a response, or a data transfer (for example, a Buffer offset or a Target Transfer Tag value). Expert Alarms Reference Guide 27 Chapter 1 iSCSI: Invalid During Login The Expert generates the iSCSI: Invalid During Login alarm when it detects an iSCSI Login Response message with the Status Code set to 020b - Invalid Request Type During Login. This status code indicates that the target is rejecting the login request from the initiator because it is invalid in the negotiation phase (that is, it can only be valid after entering the full feature phase). Because of this, the Login Request has been rejected by the target. iSCSI: LUN Does Not Exist The Expert generates the iSCSI: LUN Does Not Exist alarm when it detects an iSCSI Task Management Function Response with a response code set to 2 - LUN does not exist. This response code indicates that the LUN on which the task management function has been requested does not exist. iSCSI: Miscellaneous Initiator Error The Expert generates the iSCSI: Miscellaneous Initiator Error alarm when it detects an iSCSI Login Response message with the Status Code set to 0200 - Miscellaneous iSCSI initiator error. iSCSI: Missing Parameter The Expert generates the iSCSI: Missing Parameter alarm when it detects an iSCSI Login Response message with the Status Code set to 0207 Missing Parameter. This status code indicates that the target is rejecting the login request from the initiator because some parameter is missing from the initiator login request PDU (for example, the iSCSI Initiator or Target Name). Because of this, the target cannot process the Login Request. iSCSI: Service Unavailable The Expert generates the iSCSI: Service Unavailable alarm when it detects an iSCSI Login Response message with the Status Code set to 0301 - Service Unavailable. This status code indicates that the target is rejecting the login request from the initiator because the iSCSI service or target requested by the initiator command PDU is not available to service the request. This is a target side error. iSCSI: Session Does Not Exist The Expert generates the iSCSI: Session Does Not Exist alarm when it detects an iSCSI Login Response message with the Status Code set to 020a - Session Does Not Exist. This status code indicates that the initiator has requested login to a non-existent session. 28 Sniffer Technologies Expert Alarms and Thresholds iSCSI: Target Error The Expert generates the iSCSI: Target Error alarm when it detects an iSCSI Login Response message with the Status Code set to 0300 - Target Error. This status code indicates that the target is rejecting the login request from the initiator because either a hardware or software error has occurred on the target side of the connection making it impossible for the target to accept connections. iSCSI: Target Failure The Expert generates the iSCSI: Target Failure alarm when it detects an iSCSI Service Response Code set to 0x01 - Target Failure in a SCSI Response. This response code can indicate that there is a SCSI service failure and the target is not able to proceed with the processing of initiator commands. This can occur if the SCSI response PDU does not arrive before the session is terminated. iSCSI: Target Out of Resources The Expert generates the iSCSI: Target Out of Resources alarm when it detects an iSCSI Login Response message with the Status Code set to 0302 - Out of Resources. This status code indicates that the target is rejecting the login request from the initiator because it has insufficient resources to service any new request from the initiator. This is a target side error. iSCSI: Target Removed The Expert generates the iSCSI: Target Removed alarm when it detects an iSCSI Login Response message with the Status Code set to 0204 - Target removed. This status code indicates that the target is rejecting the login request from the initiator because the requested iSCSI Target Name has been removed and no forwarding address was provided. iSCSI: Task Does Not Exist The Expert generates the iSCSI: Task Does Not Exist alarm when it detects an iSCSI Task Management Function Response with a response code set to 1 - Task Does Not Exist. This response code indicates that the task management function has been requested by the initiator on a task which does not exist on the target. The Task Management functions provide an initiator with a way to explicitly control the execution of one or more tasks (SCSI and iSCSI tasks). Expert Alarms Reference Guide 29 Chapter 1 iSCSI: Task Management Function Not Supported The Expert generates the iSCSI: Task Management Function Not Supported alarm when it detects an iSCSI Task Management Function Response with a response code set to 5 - Task management function not supported. This response code indicates that the task management function requested by the initiator is not supported by the target. iSCSI: Too Many Connections The Expert generates the iSCSI: Too Many Connections alarm when it detects an iSCSI Login Response message with the Status Code set to 0206 - Too Many Connection. This status code indicates that the target is rejecting the login request from the initiator because the maximum number of connections possible on the requested Session ID has been reached; no more connections are possible on this Session ID. iSCSI: Unsupported Version The Expert generates the iSCSI: Target Moved alarm when it detects an iSCSI Login Response message with the Status Code set to 0205 Unsupported Version. This status code indicates that the target is rejecting the login request from the initiator because the iSCSI version range in the login request PDU is not supported by the target. Missed Browser Announcement The Expert generates the Missed Browser Announcement alarm when a node has not sent a browser announcement frame within its stated interval multiplied by the Missed Browser Announcement threshold (by default, three). The offending station is listed in the Object column of the summary row. Client systems on a Microsoft network send browser announcement frames at regular intervals. These frames announce that the client system is still on the network and identify the announcing system's resources. One field in the announcement frame specifies the number of milliseconds that will elapse before the system will send the next announcement frame. If the next announcement frame does not reach the Master Browser within the stated number of seconds multiplied by the Missed Browser Announcement threshold, the Expert generates the Missed Browser Announcement alarm. Possible causes: 30 The node has been shut down. The local network is very busy, resulting in a high loss of frames. Sniffer Technologies Expert Alarms and Thresholds MS RAP Logon Failure The Expert generates the MS RAP Logon Failure alarm when a Microsoft station fails to log in to a server. Possible causes: The user has insufficient privileges to log in to this server. Contact the server administrator to change the user's privileges, if necessary. An error occurred while loading or running a login script. Contact the server administrator. The login was not validated by a server. Contact the server administrator. The login server is running an older software version the client. Because of this, the server cannot validate the login. Contact the server administrator. The user is not allowed to log in from this computer. Contact the server administrator. The user is not allowed to log in at this time. Contact the server administrator. NCP Server Busy The Expert generates the NCP Server Busy alarm when the total of the following types of frames seen by the analyzer exceeds the Server Busy Count threshold: NCP 0x9999 (server busy) frames Novell Burst Mode frames with the communication control flag bit set for server busy. Possible causes: Too many clients using the same server. Server is overloaded with NCP requests. Network is busy. Expert Alarms Reference Guide 31 Chapter 1 NFS Retransmission The Expert generates the NFS Retransmission alarm when it sees a retransmitted NFS request. NFS retransmissions are detected by comparing the current RPC Transaction Identifier with the previous value for a connection. The summary row for this alarm will have identifying information for the offending application. An NFS retransmission may result in the retransmission of multiple frames, because many NFS transactions involve up to 8K of data spanning several frames. Possible cause: NFS retransmissions are often due to missing IP fragments. Check the network symptoms to see if there are any IP fragment problems. NNTP Slow Response Time The Expert generates the NNTP Slow Response Time alarm when the time it takes for a server's first response to an NNTP ARTICLE command exceeds the NNTP Slow Connect Time threshold. Possible causes: There is a problem with the NNTP server. There is excessive network congestion. POP3 Slow Connect Time The Expert generates the POP3 Slow Connect Time alarm when the time it takes to connect to a POP3 server exceeds the POP3 Slow Connect Time threshold. Possible causes: There is a problem with the POP3 server. There is excessive network congestion. Protocol Negotiation Failure The Expert generates the Protocol Negotiation Failure alarm when a station fails to negotiate a SMB protocol. 32 Sniffer Technologies Expert Alarms and Thresholds Possible cause: The versions of SMB offered by the client are not supported by the Server. Check the configurations of the SMB client and server. This could be an attempt by the client to circumvent security improvements in the SMB protocol by attempting to negotiate an older, “weaker” dialect. SAP R3 Slow Server Connection The amount of time it took a SAP R3 server to acknowledge a login request exceeded the SAP Slow Server Connection Time threshold. Possible causes: There is a problem with the SAP R3 server. There is excessive network congestion. Since SAP R3 servers are primarily used for e-business automation, this alarm may occur during periods of high business activity, such as an end of quarter close, or a holiday sales peak. SAP R3 Slow Server Response The time it took for a SAP R3 server's first response to a request other than a login request exceeded the SAP Slow Server Response Time threshold. Possible causes: There is a problem with the SAP R3 server. There is excessive network congestion. Since SAP R3 servers are primarily used for e-business automation, this alarm may occur during periods of high business activity, such as an end of quarter close, or a holiday sales peak. SMTP Slow Connect Time The Expert generates the SMTP Slow Connect Time alarm when the time it takes for a server's first response to an SMTP request exceeds the SMTP Slow Connect Time threshold. Possible causes: There is a problem with the SMTP server. There is excessive network congestion. Expert Alarms Reference Guide 33 Chapter 1 Station Not in Domain’s Computer List The Expert generates the Station not in Domain’s Computer List alarm when it sees a station that is not a member of the computer list maintained on a particular domain controller/server. This can cause problems if the node is a local domain controller used by clients for pass-through authentication to remotely-maintained domains. It can also prevent a client from being able to log on to a domain and access domain resources. Possible causes: The computer list on the remote domain controller does not contain the station listed in the alarm. If you need to, add the node to the remotely located domain controller's computer list. This will allow that node to communicate with the domain and its resources. A new computer has been added to the network and it has not yet been added to the domain's computer list. This node will not be allowed to log in to that domain and access domain resources. Telnet Slow Response to Login The Expert generates the Telnet Slow Response to Login alarm when the time it takes for a server to respond to a Telnet login exceeds the Telnet Connect threshold. Possible causes: There is a problem with the Telnet server. There is excessive network congestion. Session Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Session layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). [VOIP] H225 - Abnormal Disconnect The Expert generates the [VOIP] H225 – Abnormal Disconnect alarm when it observes an H.225 Release Complete message with either the ReleaseCompleteReason or Cause information element set to one of the following: 34 noBandwidth — Bandwidth taken away or ARQ denied gatekeeperResources — Gatekeeper exhausted Sniffer Technologies Expert Alarms and Thresholds gatewayResources — Switching equipment congestion adaptiveBusy — Call is dropping due to LAN crowding [VOIP] H245 - Open Logical Channel Reject The Expert generates the [VOIP] H245 – Open Logical Channel Reject alarm when it observes an H.245 Open Logical Channel Reject message. One of the purposes of H.245 is to govern the opening and closing of logical channels and ensure that receiving stations have the capability to decode the data to be sent to them before they actually receive any data. H.245 Open Logical Channel messages include an indication of the type of data to be sent over the proposed logical channel (as well as the speed at which it will be sent). If a receiving station will be unable to receive and decode the proposed data properly, it sends an Open Logical Channel Reject message rejecting the request to open a logical channel. The Open Logical Channel Reject message includes a cause field indicating the reason why the request was rejected. The possible values for the cause field are as follows: unspecified – No cause for rejection was specified. unsuitableReverseParameters – The requested reverse logical channel parameters for a bi-directional logical channel request were not appropriate. dataTypeNotSupported – The data type indicated in the Open Logical Channel message was not supported by the receiving station. dataTypeNotAvailable – The receiving station is incapable of supporting the data type indicated in the Open Logical Channel message in combination with data types already in use on other open logical channels. unknownDataType – The receiving station did not understand the data type indicated in the Open Logical Channel message. dataTypeALCombinationNotSupported – The receiving station is incapable of supporting the data type indicated in the Open Logical Channel message in combination with the Adaptation Layer type indicated in the H223 Logical Channel Parameters field. multicastChannelNotAllowed – A multicast channel could not be opened. insuffientBandwdith – A channel could not be opened because permission to use the bandwidth requested in the Open Logical Channel message was not granted. Expert Alarms Reference Guide 35 Chapter 1 separateStackEstablishmentFailed – It was not possible to run the data portion of a call on a separate stack. invalidSessionID – The slave attempted to set the Session ID when opening the logical channel to the master. masterSlaveConflict – The slave attempted to open a logical channel on which the master has identified a possible conflict. waitForCommunicationMode – There was an attempt to open the logical channel before the Multipoint Control Entity transmitted the Communication Mode Command. invalidDependentChannel – An invalid dependent channel was specified for the attempted logical channel. replacementForRejected – The type of logical channel attempted can not be opened with the replacementFor parameter. [VOIP] H245 - Terminal Capability Set Reject The Expert generates the [VOIP] H245 – Terminal Capability Set Reject alarm when it observes an H.245 Terminal Capability Set Reject message. H.245-capable terminals exchange H.245 terminal capability messages to determine the different transmit and receive modes of which each is capable (for example, CIF H.263 video, G.723.1 audio, and so on). H.245 Terminal Capability Set messages include a capability table listing these transmit and receive modes. H.245-capable stations reject Terminal Capability Set messages by sending the Terminal Capability Set Reject message. The reject message includes a cause field indicating the reason for the rejection. The possible values for the cause field are as follows: undefinedTableEntryUsed – One of the capability descriptors in the Terminal Capability Set message referred to an undefined capability table entry. descriptorCapacityExceeded – The receiving station could not store all the information in the Terminal Capability Set message. tableEntryCapacityExceeded – Either the receiving station could not store more capability table entries than indicated in the highestEntryNumberProcessed field or could not store any entries at all. [VOIP] RAS - Admission Reject The Expert generates the [VOIP] RAS – Admission Reject alarm when it observes a RAS Admission Reject (ARJ) message. 36 Sniffer Technologies Expert Alarms and Thresholds In the H.323 protocol stack, the Registration, Admission, and Status (RAS) protocol provides H.225 terminal to gatekeeper signalling services. H.225 terminals send RAS Admission Request messages (ARQs) to request access to a packet-based network from the gatekeeper. In turn, gatekeepers respond to ARQs with either an Admission Confirm (ACF) message granting the request or an Admission Reject (ARJ) message denying the request. The ARJ message includes a rejectReason field indicating the reason why the ARQ was denied. The possible values for the rejectReason field for an ARJ are as follows: Called Party not Registered (address cannot be translated) Invalid Permission (permission expired) Request Denied (no bandwidth available) Undefined Reason Caller not Registered Route Call to Gatekeeper Invalid Endpoint Identifier Resource Unavailable Security Denial QOS Control Not Supported Incomplete Address Aliases Inconsistent (there are multiple aliases in the request) Route Call to SCN (the ARJ will include a list of alternate telephone numbers to which the gateway can redirect the call, if supported by the gateway). Exceeds Call Capacity (the destination does not have the capacity to accept the call at this time). Collect Destination (gatekeeper requests that the gateway collect the final destination address). Collect PIN (gatekeeper requests that the gateway collect an authorization code or PIN). Generic Data Reason (there may be additional information in the genericData field of the ARJ message; examine the Decode display for details) Needed Feature Not Supported Expert Alarms Reference Guide 37 Chapter 1 [VOIP] RAS - Bandwidth Reject The Expert generates the [VOIP] RAS – Bandwidth Reject alarm when it observes a RAS Bandwidth Reject (BRJ) message. In the H.323 protocol stack, the Registration, Admission, and Status (RAS) protocol provides H.225 terminal to gatekeeper signalling services. H.225 terminals and gatekeepers send RAS Bandwidth Request messages (BRQ) to request changes in allocated bandwidth from one another. When a terminal or gatekeeper receives a BRQ, it can respond with either a Bandwidth Confirm (BCF) message granting the request or a Bandwidth Reject (BRJ) message denying the request. The BRJ message includes a rejectReason field indicating the reason why the BRQ was denied. The possible values for the rejectReason field for a BRJ are as follows: Not Bound (discovery permission aged out) Invalid Conference ID (possible revision) Invalid Permission (true permission violation) Insufficient Resources Invalid Revision Undefined Reason Security Denial [VOIP] RAS - Disengage Reject The Expert generates the [VOIP] RAS – Disengage Reject alarm when it observes a RAS Disengage Reject (DRJ) message. In the H.323 protocol stack, the Registration, Admission, and Status (RAS) protocol provides H.225 terminal to gatekeeper signalling services. H.225 gatekeepers send RAS Disengage Reject messages (DRJs) to reject Disengage Requests (DRQs) from unregistered endpoints. [VOIP] RAS - Gatekeeper Reject The Expert generates the [VOIP] RAS – Gatekeeper Reject alarm when it observes a RAS Gatekeeper Reject (GRJ) message. 38 Sniffer Technologies Expert Alarms and Thresholds In the H.323 protocol stack, the Registration, Admission, and Status (RAS) protocol provides H.225 terminal to gatekeeper signalling services. H.225 terminals send RAS Gatekeeper Request messages (GRQs) to request permission to register with any gatekeeper receiving the message. In turn, gatekeepers respond to GRQs with either a Gatekeeper Confirm (GCF) message granting the request or a Gatekeeper Reject (GRJ) message denying the request. RAS GRJ messages are an indication that the H.225 terminal should look for another gatekeeper with which to register. The GRJ message includes a rejectReason field indicating the reason why the GRQ was denied. The possible values for the rejectReason field for a GRJ are as follows: Resource Unavailable Terminal Excluded (permission failure; not a resource failure) Invalid Revision Undefined Reason Security Denial Generic Data Reason (there may be additional information in the genericData field of the GRJ message; examine the Decode display for details) Needed Feature Not Supported [VOIP] RAS - Location Reject The Expert generates the [VOIP] RAS – Location Reject alarm when it observes a RAS Location Reject (LRJ) message. In the H.323 protocol stack, the Registration, Admission, and Status (RAS) protocol provides H.225 terminal to gatekeeper signalling services. H.225 terminals send RAS Location Request messages (LRQs) to request address translation services from the gatekeeper. In turn, gatekeepers respond to LRQs with either a Location Confirm (LCF) message containing the transport address of the requested destination or a Location Reject (LRJ) message denying the request. The LRJ message includes a rejectReason field indicating the reason why the LRQ was denied. The possible values for the rejectReason field for an LRJ are as follows: Not Registered (requesting terminal is not registered with gatekeeper) Invalid Permission (exclusion by an administrator or feature) Request Denied (location cannot be found) Undefined Reason Expert Alarms Reference Guide 39 Chapter 1 Security Denial Aliases Inconsistent (there are multiple aliases in the request) Route Call to SCN (the LRJ will include a list of alternate telephone numbers to which the gateway can redirect the call, if supported by the gateway) Resource Unavailable Generic Data Reason (there may be additional information in the genericData field of the LRJ message; examine the Decode display for details) Needed Feature Not Supported [VOIP] RAS - Registration Reject The Expert generates the [VOIP] RAS – Registration Reject alarm when it observes a RAS Registration Reject (RRJ) message. In the H.323 protocol stack, the Registration, Admission, and Status (RAS) protocol provides H.225 terminal to gatekeeper signalling services. H.225 terminals send RAS Registration Request messages (RRQs) to gatekeepers to request permission to register. In turn, gatekeepers respond to RRQs with either a Registration Confirm (RCF) message granting registration or a Registration Reject (RRJ) message denying registration. If the terminal receives an RCF, it uses the responding gatekeeper for its calls; if it receives an RRJ, it must find a different gatekeeper with which to register. The RRJ message includes a rejectReason field indicating the reason why the RRQ was denied. The possible values for the rejectReason field for an RRJ are as follows: 40 Discovery Required Invalid Revision Invalid Call Signal Address Invalid RAS Address (the supplied address is invalid) Duplicate Alias (the provided alias is registered to another terminal) Invalid Terminal Type Undefined Reason Transport not Supported Transport Quality of Service (QOS) not Supported (the endpoint's QOS is not supported) Resource Unavailable (gatekeeper resources are exhausted) Sniffer Technologies Expert Alarms and Thresholds Invalid Alias (the provided alias is not consistent with gatekeeper rules) Security Denial Full Registration Required (registration permission expired) Additive Registration Not Supported Invalid Terminal Aliases Generic Data Reason Needed Feature Not Supported [VOIP] RAS - Unregistration Reject The Expert generates the [VOIP] RAS – Unregistration Reject alarm when it observes a RAS Unregistration Reject (URJ) message. In the H.323 protocol stack, the Registration, Admission, and Status (RAS) protocol provides H.225 terminal to gatekeeper signalling services. H.225 terminals send RAS Unregistration Request messages (URQs) to gatekeepers to request that they no longer be associated with the gatekeeper. In turn, gatekeepers respond to URQs with either an Unregistration Confirm (UCF) message granting registration or an Unregistration Reject (URJ) message denying unregistration. The Expert generates this alarm when it sees that a URQ has been rejected with a URJ. [VOIP] RTCP - Report High Jitter The Expert generates the [VOIP] RTCP – Report High Jitter alarm when the interarrival jitter reported in either an RTCP Sender Report or Receiver Report packet exceeds the Max jitter threshold. This threshold is found under the [VOIP] RTCP – Report high jitter alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Tools > Expert Options). The Real-Time Transport Control Protocol (RTCP) is used to monitor and report statistics on the quality of an ongoing RTP transaction between two or more endpoints. Each of the endpoints in an RTP transaction periodically issues RTCP report packets providing various statistics measuring the quality of the RTP data stream it is receiving. RTCP report packets (both Sender Reports and Receiver Reports) always include an interarrival jitter statistic. Essentially, interarrival jitter measures the mean difference in interpacket spacing between a sending station and a receiving station. For example, if Station A sends packets x and y at a spacing of 50 milliseconds and Station B receives the same packets x and y at a spacing of 80 milliseconds, interarrival jitter for this pair of packets is 30 milliseconds (the difference in packet spacing from sender to receiver). The exact calculation performed by an RTP endpoint Expert Alarms Reference Guide 41 Chapter 1 is somewhat more complicated than this since packets are continuously arriving. Each RTP endpoint maintains an ongoing calculation of interarrival jitter so that the value becomes a statistically smoothed mean for all packets received (see Annex A of ITU-T Recommendation H.225 for complete details). Whenever an RTCP Report packet is issued, the RTP endpoint includes its current value for interarrival jitter. The Expert captures all RTCP Report packets. If the value reported for interarrival jitter in one of these Report packets exceeds the Max jitter threshold (expressed in milliseconds), the Expert generates the RTCP – Report High Jitter alarm for the RTP session. The alarm display provides a breakdown of the individual jitter alarms for a given connection, indicating for which side of the connection different alarms have been generated, as well as the source of the last RTCP Report packet the Expert has seen for this RTP connection. The alarm display provides the following information: 42 Last Report Source – The source of the last RTCP Report packet for this RTP connection. In addition to the source's address, the type of report is also indicated (Sender Report or Receiver Report). Threshold – The value of the Max jitter threshold when this alarm was generated. Reported Jitter – The value indicated for interarrival jitter in the last received RTCP Report packet for this RTP connection. # of Net Station 1 Send Jitter Alarms – The number of jitter alarms indicated in Sender Reports issued by Net Station 1 on this RTP session. # of Net Station 1 Recv Jitter Alarms – The number of jitter alarms indicated in Receiver Reports issued by Net Station 1 on this RTP session. # of Net Station 2 Send Jitter Alarms – The number of jitter alarms indicated in Sender Reports issued by Net Station 2 on this RTP session. # of Net Station 2 Recv Jitter Alarms – The number of jitter alarms indicated in Receiver Reports issued by Net Station 2 on this RTP session. Sniffer Technologies Expert Alarms and Thresholds [VOIP] RTP - High Jitter The Expert generates the [VOIP] RTP –High Jitter alarm when the jitter measured by the Expert for an RTP session exceeds the Max jitter (ms) threshold at least once during the interval specified by the Period (ms) threshold. These thresholds are found under the [VOIP] RTP – High jitter alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Expert Options from the Tools menu). NOTE: By default, the Period (ms) threshold is set to zero – High Jitter alarms are generated each time the ongoing jitter value calculated by the Expert for an RTP stream exceeds the threshold. However, if you find that you are receiving too many RTP - High Jitter alarms, you can enable the Period threshold to decrease the number of alarms generated. With the Period threshold enabled, the Expert will only generate one alarm per Period. Keep in mind, however, that if you do enable a Period, flags in the Decode display for this Expert alarm will not correspond to the exact packet that caused the alarm to be generated. Instead, they will correspond to the end of the Period window in which the alarm was generated. NOTE: Essentially, jitter measures the mean difference in inter-packet spacing between a sending station and a receiving station. For example, if Station A sends packets x and y at a spacing of 50 milliseconds and the Expert captures the same packets x and y at a spacing of 80 milliseconds, the Expert calculates jitter for this pair of packets as 30 milliseconds (the difference in packet spacing from the sending station to the Expert). The exact calculation performed by the Expert is somewhat more complicated than this since packets are continuously arriving. See the Sniffer Voice Operations manual for details. The Expert maintains an ongoing measurement of jitter for each detected RTP session on the network so that the value becomes a statistically smoothed mean for all packets received. If the measured value for jitter exceeds the Max jitter threshold (expressed in milliseconds) over the specified period, the Expert generates the RTP – High Jitter alarm for the offending RTP session. NOTE: This alarm is only generated if the Enable Calculation option for RTP Inter-Packet Variation is disabled in the Tools > Expert Options > VoIP Options tab. Otherwise, the RTP – High Variation alarm is used. Expert Alarms Reference Guide 43 Chapter 1 How the "RTP – High Jitter" Alarm Differs from the "RTP – High Variation" Alarm Both the RTP – High Jitter and the RTP – High Variation alarms measure essentially the same thing – the mean difference in inter-packet spacing between a sending station and a receiving station. The crucial difference lies in how jitter is calculated for the two alarms. The RTP – High Jitter alarm relies on the timestamps in the RTP packets themselves to calculate inter-packet spacing. In contrast, the RTP – High Variation alarm uses a custom Codec Packet Interval specified in the Tools > Expert Options > VoIP Options tab for RTP packets in the equations used to calculate and average the deviation from the expected inter-packet spacing. See the Sniffer Voice Operations manual for details. How the "RTP – High Jitter" Alarm Differs from the "RTCP – Report High Jitter" Alarm Because the RTP – High Jitter alarm is based on the difference in inter-packet spacing between a sending station and the Expert, the frequency of these alarms will depend in part on how far physically the Expert PC is from the sending station. This is also why the Expert generates the RTCP – Report High Jitter and the RTP – High Jitter alarms differently. The RTCP – Report High Rate alarm is based on RTCP Report packets issued by the endpoints of an RTP connection. RTCP Report packets report the jitter seen between the endpoints of the RTP connection. In contrast, the RTP – High Jitter alarm is based on the jitter measured by the Expert between the sending station and itself – a distance shorter than that between the two endpoints of the RTP connection. For example, suppose the Expert PC is placed halfway between RTP Station A and RTP Station B. Theoretically, the jitter measured for an RTP session between these two stations will be less at the Expert PC than at either of the endpoints (since the distance is less). Decreasing the Frequency of High Jitter Alarms In addition to setting the Max jitter (ms) and Period (ms) thresholds differently, you can also prevent false positive High Jitter alarms by setting the RTP Packet Gap (ms) option in the Tools > Expert Options > VoIP Options tab to a lower value. This option tells the Expert to ignore inter-packet variations that exceed the stated value when making jitter calculations. This way, statistical anomalies are ignored, resulting in fewer false positives. Alarm Display The alarm display provides the following information: 44 Sniffer Technologies Direction – The direction of the connection on which the high jitter was observed (for example, [168.34.123.1] -> [168.34.123.2]. Expert Alarms and Thresholds Threshold – The value of the Max jitter threshold when this alarm was generated. RTP Jitter – The last value measured for jitter by the Expert for this RTP connection. # of Net Station 1 -> Net Station 2 Alarms – The number of High Jitter alarms generated for traffic from Net Station 1 to Net Station 2. # of Net Station 2 -> Net Station 1 Alarms – The number of High Jitter alarms generated for traffic from Net Station 2 to Net Station 1. [VOIP] RTP - High Variation The Expert generates the [VOIP] RTP – High Variation alarm when the jitter measured by the Expert for an RTP session exceeds the Max variation (ms) threshold at least once during the interval specified by the Period (ms) threshold. These thresholds are found under the [VOIP] RTP – High variation alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Expert Options from the Tools menu). NOTE: By default, the Period (ms) threshold is set to zero – High Variation alarms are generated each time the ongoing jitter value calculated by the Expert for an RTP stream exceeds the threshold. However, if you find that you are receiving too many RTP - High Variation alarms, you can enable the Period threshold to decrease the number of alarms generated. With the Period threshold enabled, the Expert will only generate one alarm per Period. Keep in mind, however, that if you do enable a Period, flags in the Decode display for this Expert alarm will not correspond to the exact packet that caused the alarm to be generated. Instead, they will correspond to the end of the Period window in which the alarm was generated. NOTE: The Expert maintains an ongoing measurement of variation for each detected RTP session on the network. If the measured value for inter-packet variation exceeds the Max variation threshold (expressed in milliseconds) over the specified interval, the Expert generates the RTP – High Variation alarm for the offending RTP session. How the "RTP – High Variation" Alarm Differs from the "RTP – High Jitter" Alarm Both the RTP – High Variation and the RTP – High Jitter alarms measure essentially the same thing – the mean difference in inter-packet spacing between a sending station and a receiving station. The crucial difference lies in how jitter is calculated for the two alarms. The RTP – High Variation alarm uses a custom Codec Packet Interval specified in the Tools > Expert Expert Alarms Reference Guide 45 Chapter 1 Options > VoIP Options tab for RTP packets in the equations used to calculate and average the deviation from the expected inter-packet spacing. In contrast, the RTP – High Jitter alarm relies on the timestamps in the RTP packets themselves to calculate inter-packet spacing. See the Sniffer Voice Operations manual for details. NOTE: This alarm is only generated if the Enable Calculation option for RTP Inter-Packet Variation is enabled in the Tools > Expert Options > VoIP Options tab. Otherwise, the RTP – High Jitter alarm is used. The alarm display provides the following information: Direction – The direction of the connection on which the high percentage of dropped frames was observed (for example, [168.34.123.1] -> [168.34.123.2]. Threshold – The value of the Percentage of dropped packets threshold when this alarm was generated. RTP Variation (ms) – The measured variation that caused this alarm to be generated. Because an alarm was generated, this value will be higher than the Threshold value. # of Net Station 1 -> Net Station 2 Alarms – The number of High Variation alarms generated for traffic from Net Station 1 to Net Station 2. # of Net Station 2 -> Net Station 1 Alarms – The number of High Variation alarms generated for traffic from Net Station 2 to Net Station 1. [VOIP] RTP - Too Many Dropped Packets The Expert generates the [VOIP] RTP – Too Many Dropped Packets alarm when the percentage of dropped packets in either direction on an RTP connection exceeds the Percentage of dropped packets threshold at least once during the interval specified by the Period (ms) threshold. These thresholds are found under the [VOIP] RTP – Too Many Dropped Packets alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Expert Options from the Tools menu). NOTE: By default, the Period (ms) threshold is set to zero – Too Many Dropped Packets alarms are generated each time the dropped packet percentage calculated by the Expert for an RTP stream exceeds the threshold. However, if you find that you are receiving too many RTP - Too Many Dropped Packets alarms, you can enable the Period threshold to decrease the number of alarms generated. With the Period threshold enabled, the Expert will only generate one alarm per Period. 46 Sniffer Technologies Expert Alarms and Thresholds Keep in mind, however, that if you do enable a Period, flags in the Decode display for this Expert alarm will not correspond to the exact packet that caused the alarm to be generated. Instead, they will correspond to the end of the Period window in which the alarm was generated. NOTE: The Real-Time Transport Protocol (RTP) provides end-to-end delivery services for time-sensitive data, such as interactive audio and video. A high number of dropped packets on an RTP connection can indicate congestion and may negatively affect end-user performance of audio or video applications. The alarm display provides the following information: Direction – The direction of the connection on which the high percentage of dropped packets was observed (for example, [192.168.1.1] -> [192.168.1.2]. Threshold – The value of the Percentage of dropped packets threshold when this alarm was generated. Percentage of Drop Packets Count – The percentage of dropped packets measured by the Expert for this session. Because an alarm was generated, this value will be higher than the Threshold value. # of Net Station 1 -> Net Station 2 Alarms – The number of Too Many Dropped Packets alarms generated for traffic from Net Station 1 to Net Station 2. # of Net Station 2 -> Net Station 1 Alarms – The number of Too Many Dropped Packets alarms generated for traffic from Net Station 2 to Net Station 1. [VOIP] RTP - Too Many Out of Sequence Packets The Expert generates the [VOIP] RTP – Too Many Out of Sequence Packets alarm when the number of out of sequence packets in either direction on an RTP connection exceeds the Max # of out of sequence packets threshold at least once during the interval specified by the Period (ms) threshold. These thresholds are found under the [VOIP] RTP – Too Many Out of Sequence Packets alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Expert Options from the Tools menu). NOTE: By default, the Period (ms) threshold is set to zero – Too Many Out of Sequence Packets alarms are generated each time the number of out of sequence packets seen by the Expert for an RTP stream exceeds the threshold. However, if you find that you are receiving too many RTP - Too Expert Alarms Reference Guide 47 Chapter 1 Many Out of Sequence Packets alarms, you can enable the Period threshold to decrease the number of alarms generated. With the Period threshold enabled, the Expert will only generate one alarm per Period. Keep in mind, however, that if you do enable a Period, flags in the Decode display for this Expert alarm will not correspond to the exact packet that caused the alarm to be generated. Instead, they will correspond to the end of the Period window in which the alarm was generated. NOTE: The Real-Time Transport Protocol (RTP) provides end-to-end delivery services for time-sensitive data, such as interactive audio and video. However, it does not guarantee delivery of packets in the same order as they were sent. Instead, each RTP packet includes a sequence number so that a receiving station can reconstruct the sender's sequence of packets, if necessary. The Expert examines the sequence numbers of all packets on a given RTP connection. If it detects that the number of out of sequence RTP packets on a given connection (packets without consecutive sequence numbers) exceeds the Max # of out of sequence packets threshold, it generates this alarm. Since RTP typically relies on UDP for its efficient transport, you may want to examine the underlying UDP packets to understand why RTP packets are being delivered out of sequence. The alarm display provides the following information: 48 Direction – The direction of the connection on which the offending out of sequence packets were observed (for example, [168.34.123.1] -> [168.34.123.2]. Threshold – The value of the Max # of out of sequence packets threshold when this alarm was generated. Out of Seq. Frame Count – The number of out of sequence packets measured by the Expert for this session. Because an alarm was generated, this value will be higher than the Threshold value. # of Net Station 1 -> Net Station 2 Alarms – The number of Too Many Out of Sequence Packets alarms generated for traffic from Net Station 1 to Net Station 2. # of Net Station 2 -> Net Station 1 Alarms – The number of Too Many Out of Sequence Packets alarms generated for traffic from Net Station 2 to Net Station 1. Sniffer Technologies Expert Alarms and Thresholds [VOIP] SCCP - Register Reject The Expert generates the [VOIP] SCCP – Register Reject alarm when it observes a Skinny Client Control Protocol (SCCP) Station Register Reject message. The SCCP protocol allows simple IP telephony devices to operate without implementing the entire H.323 specification. Instead, they can operate as SCCP “skinny” clients and communicate with H.323 proxies (called Call Managers in SCCP) to interact with H.323-compliant devices. When an SCCP skinny client comes online, it sends a registration message (StationRegister) to the Call Manager to announce its existence. The Call Manager can reject the client's registration attempt with a Station Register Reject message. When the Expert observes an SCCP Station Register Reject message, it generates this alarm. Possible cause: The Station Register Reject message includes a short textual description (up to 33 bytes) indicating both the reason the Call Manager rejected the registration attempt and the ID of the requesting station. [VOIP] SCCP - Station Alarm The Expert generates the [VOIP] SCCP – Station Alarm alarm when it observes a Skinny Client Control Protocol (SCCP) Station Alarm message. The SCCP protocol allows simple IP telephony devices to operate without implementing the entire H.323 specification. Instead, they can operate as SCCP “skinny” clients and communicate with H.323 proxies (called Call Managers in SCCP) to interact with H.323-compliant devices. When a skinny client has an error condition to report to the Call Manager, it does so using the SCCP Station Alarm message. The Station Alarm message includes a text string of up to 80 characters indicating the nature of the error condition on the skinny client. Possible cause: The Station Alarm message includes a short textual description (up to 80 characters) indicating the nature of the error condition being reported to the Call Manager. [VOIP] SIP - Client Error The Expert generates the [VOIP] SIP – Client Error alarm when it detects a failure response from a SIP server to a SIP client's request with a 4xx error code. 4xx error codes indicate that the client should not retry the same request without modification. Expert Alarms Reference Guide 49 Chapter 1 The exact 4xx error code causing the alarm was one of the following: 50 (400) Bad Request – The request was not understood because of bad syntax. (401) Unauthorized – The request requires authentication. (402) Payment Required – (Reserved for future use.) (403) Forbidden – The server understood the request but refuses to fulfill it, regardless of authentication. (404) Not Found – The server does not believe that the user exists in the domain specified in the Request-URI. (405) Method Not Allowed – The specified method is not allowed for the specified address. (406) Not Acceptable – The resource identified in the request can only generate responses unacceptable to the requester. (407) Proxy Authentication Required – The request requires prior authentication with a proxy. (408) Request Timeout – The server could not provide a response before the expiration indicated in the request. (409) Conflict – The request could not be processed because of a conflict with the requested resource. (410) Gone – The requested resource is no longer available on the server. (411) Length Required – The server requires a length field for the request. (413) Request Entity Too Large – The request entity is larger than the server will process. (414) Request-URI Too Large – The Request-URI is larger than the server will process. (415) Unsupported Media Type – The server is refusing the request because it is in a format not supported by the request resource. (420) Bad Extension – The server did not understand a protocol extension in the request. (480) Temporarily Not Available – The called party's end system was contacted but the called party is not currently available. (481) Call Leg/Transaction Does Not Exist – Either the server received a BYE request not matching any existing call leg or the server received a CANCEL request not matching any existing transaction. Sniffer Technologies Expert Alarms and Thresholds (482) Loop Detected – The request included a Via path containing itself (that is, a loop). (483) Too Many Hops – The request contained more hops than allowed by the Max-Forwards field in the request header. (484) Address Incomplete – The request contained an incomplete To address or Request-URI. (485) Ambiguous – The called party address indicated in the request was ambiguous. (486) Busy Here – The called party's end system was contacted but the called party is not currently accepting additional calls. The alarm display indicates the exact error code that caused the alarm. [VOIP] SIP - Global Error The Expert generates the [VOIP] SIP – Global Error alarm when it detects a global failure response from a SIP server. Global failure responses include 6xx error codes. SIP servers send global failure responses to indicate that they have authoritative information that a particular user cannot be located or contacted. The exact 6xx error code causing the alarm was one of the following: (600) Busy Everywhere – The called party's end system was contacted but the called party is busy and will not take the call now. This response is returned only if the client knows that no other endpoint will answer the request. Otherwise (486) Busy Here is returned (causing the Expert to generate the SIP – Client Error alarm). (603) Decline – The called party's machine was contacted successfully, but the user is explicitly declining participation. (604) Does Not Exist Anywhere – The server has definitive information that the called party does not exist anywhere. (606) Not Acceptable – The called party's machine was contacted successfully but some aspects of the session description were not acceptable (for example, the requested media or bandwidth). The alarm display indicates the exact error code that caused the alarm. Expert Alarms Reference Guide 51 Chapter 1 [VOIP] SIP - Server Error The Expert generates the [VOIP] SIP – Server Error alarm when it detects a server failure response from a SIP server. Server failure responses include 5xx error codes. SIP servers send server failure responses to indicate that they have experienced some sort of error. Client's receiving SIP server failure responses typically reissue their requests to other SIP servers. The exact 5xx error code causing the alarm was one of the following: (500) Internal Server Error – The server experienced an internal error that prevented it from servicing the request. (501) Not Implemented – The server can not fulfill the request because it does not provide the requisite functionality. (502) Bad Gateway – The server received an invalid response from a downstream server while attempting to service the request. (503) Service Unavailable – The server is unable to service the request due to temporary overloading or maintenance on the server. (504) Gateway Timeout – The server did not receive a timely response from a downstream server it accessed while attempting to service the request. (505) SIP Version not Supported – The server does not support the version of SIP indicated in the request. The alarm display indicates the exact error code that caused the alarm. [VOIP] SIP - Server Slow Response The Expert generates the [VOIP] SIP – Server Slow Response alarm when the time it takes for a SIP server to respond to a client's request exceeds the SIP Slow response time threshold. This threshold is found under the [VOIP] SIP – Server Slow Response alarm entry in the Alarms tab of the Expert UI Object Properties dialog box (accessed by selecting Expert Options from the Tools menu). Because all responses to a SIP request must use the same values in the Call-ID, CSeq, To, and From fields, the Expert is able to match responses from a SIP server to the corresponding request from a SIP client. When the Expert sees a SIP request, it stores the timestamp when it captured the frame. Then, when the Expert sees a corresponding response from the SIP server, it calculates the delay between the last request from the client and the first response from the server. If the value for this delay is greater than the SIP Slow response time threshold (expressed in milliseconds), the Expert generates the SIP – Server Slow Response alarm. The alarm display includes the following information: 52 Sniffer Technologies Expert Alarms and Thresholds Threshold – The value of the SIP Slow response time threshold when this alarm was generated. Response Time – The response time measured by the Expert that caused this alarm to be generated. This value will be higher than the Threshold value. Command – The type of SIP client request to which the server was slow in responding. For example, INVITE. Response code – The numerical response code returned by the SIP server to the client's request. For description of all SIP response codes, see RFC 2543 describing the SIP protocol. Description – A short textual description of the response code returned by the SIP server to the client's request. File Retransmission The Expert generates the File Retransmission alarm when it detects that an entire file, or a subset of a file has been retransmitted. This differs from a transport layer retransmission because the application is purposely retransmitting the file. Each retransmission is counted as a separate alarm. Possible causes: The application is not using the network efficiently. Possibly, the application was not designed to be used over networks. It is possible that different data was written by the application with each transmission. The data within the file is not checked. Invalid Presentation Context ID The Expert generates this alarm when it detects an invalid presentation context ID on an RPC connection. IP NETBIOS Session Reject The Expert generates the IP NETBIOS Session Reject alarm when an attempt to initiate a NETBIOS Session is rejected by the other half of the connection. Possible causes: The receiving node is shutting down and cannot create a new NetBIOS connection. The receiving node was too busy to accept additional connections. Expert Alarms Reference Guide 53 Chapter 1 Local Limit on Defined Context Set Exceeded The Expert generates this alarm when it detects that the local limit on a defined context set has been exceeded on an RPC connection. Loops on Same Request The Expert generates the Loops on Same Request alarm when it sees the application process on one station send out the same request even though it has already received the appropriate reply from the destination station. Only requests that are sent within the time specified by the Filter time threshold are considered “repeated.” Possible causes: This may be legitimate, depending upon the application. For example, some print servers check for print requests every few seconds, resulting in many repetitions of the same request. You can eliminate this situation by lowering the Filter time threshold. Particular applications may be inefficient. For example, some X Windows applications continually check the location of the mouse cursor, which can cause network traffic of this type. Excessive repeated requests are characteristic of applications that do not use the network efficiently, or that are not designed to be used over networks. If the request is a Novell NDS verb request, it may indicate the need to partition the NDS tree among more servers. However, the verbs DSV_Resolve_Name, DSV_Read, and DSV_Get_Server_Address are often sent out as part of the NDS protocol. Low Throughput The Expert generates the Low Throughput alarm when the rate of throughput (in KBytes/s) on a session is less than the minimum rate specified by the Local Xfer threshold (for local transfers) or Remote Xfer threshold (for remote transfers). The expert system assumes that a transfer is local unless it has received a Routing Information Protocol (RIP) packet indicating a non-zero hop count. Some transfers that are indeed remote may be tentatively labeled as local, because no RIP packets have been received by the analyzer since the beginning of the current capture. Under these conditions the expert system will use the Local xfer rather than the Remote xfer threshold. Since the Local xfer threshold is characteristically set to a faster rate than the Remote xfer 54 Sniffer Technologies Expert Alarms and Thresholds threshold, it will be possible for a remote transfer to trigger this symptom falsely. If you suspect that this is the case, then notice whether the symptom has occurred within 30 seconds of the start of capture. RIP packets are sent at 30 second intervals, and therefore such false symptoms will not generally occur after 30 seconds of capture. Occasional instances of low throughput are generally not considered to be a problem. When low throughput occurs often, you should try to determine the cause. Possible causes: If more than one application has low throughput, and if this symptom occurs for exchanges between different pairs of stations, then there is a network problem, such as heavy traffic going through a router. If more than one application has low throughput, and if this symptom occurs primarily for exchanges between a particular pair of stations, then there is a connection problem – the transport level communication software is not properly configured. If only one application has low throughput, then the problem is with that application. An excessive number of slow file transfers may be due to an overloaded server. You should examine the traffic to your server to see if the traffic is justified. If so, you should consider getting a faster server or distributing the load across multiple servers. Read/Write Overlap The Expert generates the Read/Write Overlap alarm when read or write requests to a file overlap each other. For example, if an application first wrote bytes 50 through 700 and then wrote bytes 200 through 800, there would be a write overlap because bytes 200-700 were written twice. Request Denied The Expert generates the Request denied alarm when the number of denied requests on a session exceeds the Denied Count threshold. A request is only counted as denied if it occurs within the time specified by the Filter time threshold. NOTE: The Filter Time threshold is located under the Alarm entry for Loops on same request at the Session Layer. These alarms share the same threshold. Expert Alarms Reference Guide 55 Chapter 1 Possible causes: This may be normal depending upon the application. An excessive number of error responses may indicate a misconfigured application. The application may not have been designed for use over networks. A password is wrong or the requester does not have the necessary access privileges. If the network is running Novell NDS, then the error response may be an NDS verb error reply. Validate the NDS tree. Slow File Transfer The Expert generates the Slow File Transfer alarm when the percentage of slow file transfers on a given session is greater than the Slow File Xfer % threshold and the number of requests has exceeded the Min appl req threshold. Occasional instances of slow file transfer are generally not considered to be a problem. If slow file transfers occur often, the cause should be determined. Possible causes: If more than one application exhibits slow file transfer and if this symptom occurs for exchanges between different pairs of stations, then there is a network problem, such as heavy traffic through a router. If more than one application exhibits slow file transfer and if this symptom occurs primarily in exchanges between a particular pair of stations, then there is a problem with the connection – the transport level communication software is not properly configured. If only one application exhibits slow file transfer, then there is a problem with that application. An excessive number of slow file transfers may be due to an overloaded server. You should examine the traffic to your server to see if the traffic is justified. If so, you should consider getting a faster server or distributing the load across multiple servers. Temporary Congestion The Expert generates this alarm when it detects that an RPC server is experiencing temporary congestion. 56 Sniffer Technologies Expert Alarms and Thresholds The Operation Number Is Greater than the Number of Operations in the Interface The Expert generates this alarm when it detects that operation number passed in the request PDU on an RPC connection is greater than or equal to the number of operations in the interface. The Output Parameters of the Operation Exceed Their Declared Maximum Size The Expert generates this alarm when it detects that the output parameters of an RPC operation exceed their declared maximum size. The Request Is Being Rejected for Unspecified Reasons The Expert generates this alarm when it detects that a request on an RPC connection was rejected for unspecified reasons. The RPC Client or Server Protocol Has Been Violated The Expert generates this alarm when it detects that the RPC client or server protocol was violated. The Server Did Not Support the Requested Authentication Level The Expert generates this alarm when it detects an RPC server does not support the authentication level requested by an RPC client. The Server Does Not Export the Requested Interface The Expert generates this alarm when it detects that an RPC server does not support a requested interface. The Server Does Not Implement the Requested Operation for the Type of Requested Object The Expert generates this alarm when it detects that an RPC server does not implement an operation requested by a client for the type of requested object. The Server Does Not Support the Interface The Expert generates this alarm when it detects that an RPC server does not support a requested interface. Expert Alarms Reference Guide 57 Chapter 1 The Server Does Not Support the Proposed Transfer Syntax The Expert generates this alarm when it detects that an RPC server does not support the transfer syntax proposed by a client. The Server Does Not Support RPC Protocol Version The Expert generates this alarm when it detects that a server does not support the RPC protocol version specified in the bind PDU. The Server Is Too Busy To Handle the Call The Expert generates this alarm when it detects that an RPC server rejects a request because it is too busy to handle the call. The Server Manager Routine Has Not Been Entered and Executed The Expert generates this alarm when it detects that the server manager routine on an RPC server is not entered and executed when it should be. TNS Connect Refused The Expert generates the TNS Connect Refused alarm when the number connect requests refused by an Oracle server exceeds the Oracle Connection Failed threshold. A request to connect is refused by the server because it is unable to satisfy the options asked for in the connect packet. Examine the service options in the connect packet and determine whether they are options that are allowed in the target server. Check the data in the refuse packet and look for the error code given. Refer to the Oracle manual containing the SQL*Net error codes. Possible causes: Invalid username. Invalid password. Invalid server. TNS Security Breach Attempt The Expert generates the TNS Security Breach Attempt alarm when the number of failed login attempts exceeds the Oracle Security Breach Attempt threshold. This may not be a problem if it does not continue. 58 Sniffer Technologies Expert Alarms and Thresholds An Oracle7 application usually terminates when it gets three occurrences of this error type. However, this may not apply to user-written applications. Two or more occurrences of this alarm should be investigated. TNS Slow Connect The Expert generates the TNS Slow Connect alarm when the time a server or a Multiprotocol Interchange (MPI) takes to respond to a TNS Connect Request or a login request exceeds the Oracle Slow Connect Time threshold. This may or may not be a problem, depending upon the application environment. The Oracle Slow Connect Time threshold should be adjusted to the desired or typical connect response time in your application environment. Possible causes: The server is overloaded. The MPI is too busy. The server or MPI is misconfigured. TNS Slow Server Diagnosis The Expert considers an Oracle server's response to an SQL request to be slow when the time between request and response is longer than the Oracle Slow Server Response Time threshold. The Expert counts these slow responses. When the total number of slow responses seen during the time period specified by the Oracle Slow Server Response Interval threshold exceeds the Oracle Slow Server Response Count threshold, the Expert generates the TNS Slow Server Diagnosis alarm. Possible causes: The problem is network related. Determine if your network is overloaded. Increase your network's bandwidth. Determine if your network is optimally configured. Minimize frame fragmentation by using larger frame sizes or more efficient protocols. Reduce polling frequency of polled devices in your network. The problem is in the Multiprotocol Interchange. Expert Alarms Reference Guide 59 Chapter 1 60 Sniffer Technologies Configure your MPI for better data throughputs rather than for connections. Decrease the number of pumps per MPI. Add more MPIs. The problem is in the server. Assess the current utilization of your server. Determine if you have enough CPU power to support the number of users in this server. Optimize the file distribution of your database. Place the tables and their associated indexes in separate disk drives. Tables that are most frequently referenced should be placed in separate disk drives. Determine if your server has enough memory to support the applications and services that have to run concurrently. Install faster disk drives. Check for runaway (orphan) processes and clean up after them. The problem is in the database. Perform database tuning suggested by Oracle and other experts in this area. For example: Make sure that the shared pool is large enough to reduce reparsing. Optimize the size of the DB_BLOCK_BUFFERS. Increase the value of DML_LOCKS to allow more locks placed on an object at one time. Determine your library cache misses and set the CURSOR_SPACE_FOR_TIME according to the recommendation in the Oracle7 Server Administrator's Guide. Check if indexed clusters can speed up your access to tables that are commonly joined together on a standard set of matching columns. The problem is in the client workstation. Use optimized query statements. The “explain plan” command will show the access paths to the data. Remove all unnecessary “lock table” statements. Do not use “for update” clause in select statements if there is no need to update immediately. Issue commit statements immediately after “update”, “delete”, and “insert” statements to release locked resources. Expert Alarms and Thresholds Use explicit cursors in all of your PL/SQL blocks. Identical query statements in your application must match character-by-character to use the same shared pool area. TNS Slow Server Response The Expert generates the TNS Slow Server Response alarm when the time it takes for a server to respond to one or more SQL commands exceeds the Oracle Slow Server Response Time threshold. Possible causes: The problem is network related. Determine if your network is overloaded. Increase your network's bandwidth. Determine if your network is optimally configured. Minimize frame fragmentation by using larger frame sizes or more efficient protocols. Reduce polling frequency of polled devices in your network. The problem is in the Multiprotocol Interchange. Configure your MPI for better data throughputs rather than for connections. Decrease the number of pumps per MPI. Add more MPIs. The problem is in the server. Assess the current utilization of your server. Determine if you have enough CPU power to support the number of users in this server. Optimize the file distribution of your database. Place the tables and their associated indexes in separate disk drives. Tables that are most frequently referenced should be placed in separate disk drives. Determine if your server has enough memory to support the applications and services that have to run concurrently. Install faster disk drives. The problem is in the client workstation. Use optimized query statements. The “explain plan” command will show the access paths to the data. Expert Alarms Reference Guide 61 Chapter 1 Remove all unnecessary “lock table” statements. Do not use “for update” clause in select statements if there is no need to update immediately. Issue commit statements immediately after “update”, “delete”, and “insert” statements to release locked resources. Use explicit cursors in all of your PL/SQL blocks. Identical query statements in your application must match character-by-character to use the same shared pool area. Check for runaway (orphan) processes and clean up after them. The problem is in the database. Perform database tuning suggested by Oracle and other experts in this area. For example: Make sure that the shared pool is large enough to reduce reparsing. Optimize the size of the DB_BLOCK_BUFFERS. Increase the value of DML_LOCKS to allow more locks placed on an object at one time. Determine your library cache misses and set the CURSOR_SPACE_FOR_TIME according to the recommendation in the Oracle7 Server Administrator's Guide. Check if indexed clusters can speed up your access to tables that are commonly joined together on a standard set of matching columns. Too Many File Retransmissions The Expert generates the Too Many File Retransmissions alarm when the percentage of file transfers that are retransmissions on a session exceeds the File Retrans % threshold. Possible cause: 62 Sniffer Technologies This alarm indicates excessive retransmissions of the same file request, or overlap between two file requests. This usually means that the application is not using the network efficiently. If possible, the application should be reconfigured (or rewritten) to make it more efficient. Expert Alarms and Thresholds Too Many Loops on Same Request The Expert generates the Too Many Loops on Same Request alarm when the ratio of repeated requests (loops) to normal requests (expressed as a percentage) on a session exceeds the Loop % threshold. This alarm is only generated after the total number of application requests seen by the analyzer exceeds the Min Appl Req threshold. A repeated request is one that is sent out even though the proper response to the previous request has already been seen by the analyzer. The analyzer does not count certain repeated requests. For example, if a repeated request is sent after the Filter time threshold expires, the analyzer “forgets” the previous request and does not consider it to be repeated. A normal request is any request that is not considered repeated. Possible causes: This may be legitimate, depending upon the application. For example, some print servers check for print requests every few seconds, resulting in many repetitions of the same request. You can eliminate this situation by lowering the Filter time threshold. Particular applications may be inefficient. For example, some X Windows applications continually check the location of the mouse cursor, which can cause network traffic of this type. Excessive repeated requests are characteristic of applications that do not use the network efficiently, or that are not designed to be used over networks. If the request is a Novell NDS verb request, it may indicate the need to partition the NDS tree among more servers. However, the verbs DSV_Resolve_Name, DSV_Read, and DSV_Get_Server_Address are often sent out as part of the NDS protocol. Too Many Requests Denied The Expert generates the Too Many Requests Denied alarm when the percentage of denied versus successful application requests on a session object exceeds the Denied request %threshold. This alarm is only generated after the total number of application requests seen by the analyzer exceeds the Min Appl Req threshold. Possible causes: The application was not designed for use over networks. Expert Alarms Reference Guide 63 Chapter 1 The application doesn't take into account that the responding application is on a different station. The application is waiting for a semaphore on a server that has not been released. WINS Duplicate Name The Expert generates the WINS Duplicate Name alarm when it detects a node attempting to register a duplicate node name with the WINS Server. Only unique names can exist at a time on any given network. Possible cause: An existing machine's name has been changed or a new machine has been added with a conflicting computer name. Use the Network control panel to change one of the conflicting stations' computer name. WINS No Response The Expert generates the WINS No Response alarm when a station does not respond to a WINS Request within the time specified by the WINS No Response threshold. Possible causes: 64 The Expert did not see the WINS response because the Expert is connected to a switched network environment. Switches have the effect of segmenting network traffic, preventing the Expert from seeing all network traffic unless port mirroring is set up effectively. Because the Expert never saw the response from the WINS Server, this alarm was generated. Connectivity to the WINS Server (the recipient of the WINS Request) has temporarily been lost, or the WINS Server has shut down. WINS packets use UDP as a transport. During periods of high network activity, UDP packets can be lost. If there are multiple alarms naming the same WINS Server, check the named WINS Server to make sure it is properly connected to the network. Then, check the state of the WINS Service by running WINS Manager. The WINS configuration on the requesting client specifies an invalid address for the WINS Server. Check the TCP/IP Settings in the requesting client's Network control panel and correct them, if necessary. Sniffer Technologies Expert Alarms and Thresholds Connection Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Connection layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Ack Too Long The Expert generates the Ack Too Long alarm when the time that it has taken to acknowledge data on a connection exceeds the Long Ack Time threshold plus three times the average acknowledgment time for this connection. Possible causes: The recipient of the original data frame was temporarily busy, and could not process the frame as quickly as usual. The ACK arrived late because a server had to look up and/or process data before responding with an ACK. The path changed in a way that increased the time between the request and its acknowledgement. There were multiple paths between the two stations, and the time to acknowledgement was longer for some paths than for others. Fast Retransmission The Expert generates the Fast Retransmission alarm when both of the following conditions for a connection are true: The TCP sequence number is less than or equal to its previous value, indicating that this frame contains data that has already been transmitted. The time between the duplicate transmissions is less than the Fast retransmission time threshold. Possible cause: If the total amount of data to be retransmitted is relatively small, some TCP applications will retransmit data that has not yet been acknowledged. If this happens frequently, network bandwidth is being wasted, and you should change to software that does not perform unnecessary retransmissions. Expert Alarms Reference Guide 65 Chapter 1 Idle Too Long The Expert generates the Idle Too Long alarm when it sees a TCP or NFS connection has been idle (no frames in either direction) for longer than the Idle Time threshold. If the idle connection is still alive, it is reserving system resources at both ends of the connection without using them. Possible causes: A station is inoperative. A router has lost the connection. The transport software has hung. An application has hung. An application is open but not in use. Local Routing The Expert generates the Local Routing alarm when it detects two DLC stations on the same segment communicating via a router (packets are being transmitted via the router rather than directly from one DLC station to the other). Possible causes: The router table may be improperly configured. If this is the case, reconfigure the router table so that the stations can communicate with each other directly. The router may be used for security purposes, to control access to particular parts of the network. The router may be performing gateway type functions, such as protocol translation at the application level. A device that is new to the network is sending frames through a router because it has not yet become aware of the direct route to a local station, or because it is configured to use that router. Non-Responsive Station The Expert generates the Non-Responsive Station alarm when the number of consecutive retransmissions without a response on a connection exceeds the No Response Count threshold. 66 Sniffer Technologies Expert Alarms and Thresholds Possible causes: Retransmissions often indicate a problem with a server. Some examples are: Improperly configured software. An overloaded server. Faulty hardware. Repeat Ack The Expert generates the Repeat Ack alarm when it sees a station send a frame on a connection with the acknowledgment number less than its previous value. Possible cause: Some implementations do this purposely to keep the circuit “alive”. However, this is not good practice because it causes unnecessary network traffic. Retransmission The Expert generates the Retransmission alarm when it sees a frame with a TCP sequence number less than or equal to its previous value, indicating that it contains data that has already been transmitted. The retransmission occurred because the transmitting station did not receive an acknowledgment (ACK) from the receiving station, and therefore retransmitted the frame. Possible causes: The receiving station was overloaded. A router was overloaded, or there was some other propagation delay. The network traffic load was very high. The ACK was sent but lost. Either the original packet or its ACK was damaged. The ACK was transmitted via a slower path. An IP fragment was lost. There is a network integrity problem. Expert Alarms Reference Guide 67 Chapter 1 Slow Server The Expert generates the Slow Server alarm when the ratio of slow responses to total responses for a particular station exceeds the Slow Response % threshold. A slow response is one in which the time between request and response is longer than the Response Time threshold. Possible cause: This usually means that your server is overloaded. You should examine the traffic to your server to see if the traffic is justified. If so, you should consider getting a faster server or distributing the load across multiple servers. TCP Fast Keepalive The Expert generates the TCP Fast Keep-Alive alarm when the delta time between the last packet seen on a TCP connection and a subsequent keepalive is less than the TCP Fast Keep-Alive (msec) threshold. In some legacy TCP implementations, stations send keepalives to the opposite end of a connection when the connection has been idle for a set amount of time. The purpose of sending a keepalive is to elicit a response from the opposite end of the connection, thereby ensuring that the connection stays alive. This alarm can help detect stations that are sending keepalives too quickly, resulting in an inefficient use of network bandwidth. However, per RFC 1122, if you are still using a legacy implementation of TCP, you may want to consider updating it because of the following drawbacks associated with keepalives: Established connections can break during short-lived network failures (for example, due to temporary congestion) if a keepalive is sent during the problematic period and the sending station receives no response. Bandwidth can be consumed unnecessarily by keepalives. On network paths that charge by the packet, excessive keepalives can result in increased costs. Most modern implementations of TCP do not use keepalives to keep connections open. Instead, connections are closed when they are idle and re-opened when data needs to be sent. 68 Sniffer Technologies Expert Alarms and Thresholds How The Expert Detects Keepalives There is no explicit way to detect a TCP keepalive – they carry no telltale flag or message ID. Instead, the Expert assumes that a TCP packet with a payload of 0 or 1 bytes and the same sequence number as the last packet seen on the connection (that is, a retransmission) is a keepalive. In earlier versions of the Expert, the Expert flagged keepalives as retransmission alarms. In this version, keepalives will not cause retransmission alarms, but they may result in Fast Keep-Alive alarms, depending on their timing. Disabling the TCP Fast Keep-Alive Alarm If you find that too many TCP Fast Keep-Alive alarms are being generated on your network, you can either increase the TCP Fast Keep-Alive (msec) threshold to a value that makes more sense for your network or disable the alarm entirely by setting the threshold to zero. Too Many Retransmissions The Expert generates the Too Many Retransmissions alarm when the number of retransmissions on a connection (expressed as a percentage of good transmissions versus retransmissions on the connection) exceeds the Retransmission % threshold. Excessive retransmissions can result in a noticeable reduction in response time. Retransmissions occur when the transmitting station does not receive an acknowledgment (ACK) from the receiving station, and therefore retransmits the frame. Possible causes: The receiving station was overloaded. This is particularly likely if the receiving station was a server. A router was overloaded, or there was some other source of propagation delay. The network traffic load was very high. The ACKs were being transmitted via a slower path. There is a network integrity problem. UDP Bouncing Frames The Expert generates the UDP Bouncing Frames alarm when it detects UDP bouncing frames on a connection. A bouncing frame is one in which the IP layer's time-to-live field has decreased by one from its previous value. It usually indicates that a frame has entered the local network through a router, but the intended recipient of the frame cannot be found. Expert Alarms Reference Guide 69 Chapter 1 If this happens very often, the cause should be discovered and eliminated. The intended recipient, probably on another network, may not be operating properly as a result of this routing error. Possible cause: There is a misconfigured router or routers. Examine the routing table(s) and correct them if necessary. Window Frozen The Expert generates the Window Frozen alarm when it sees a connection using a windowing protocol (such as DECnet or TCP) for which the window size has been stuck at some number of bytes for longer than the Window Frozen Time threshold. When a window is stuck, data flow will not be as efficient because the transmitter will not send data in excess of the current window size. Possible causes: There is insufficient buffer space, because the recipient's memory allocation process is not providing enough buffer memory. There is insufficient buffer space, because there are new open connections, so that the buffer space must be shared. Window Size Exceeded The Expert generates the Window Size Exceeded alarm when it sees a frame whose TCP data size exceeds the current TCP window size of the receiving end of this connection. Most TCP implementations will ignore this frame. However, if this happens very often, network bandwidth is being wasted. Possible causes: This situation may occur occasionally if the window has just decreased before the transmitter has had a chance to notice. An incorrect implementation may be ignoring the window size or the maximum segment size. Zero Window Too Long The Expert generates the Zero Window Too Long alarm when it detects a connection using a windowing protocol (such as DECnet or TCP) where the window size for a station has been zero longer than the time specified by the Zero Window Time threshold. 70 Sniffer Technologies Expert Alarms and Thresholds Possible causes: The receiving station is very busy. There are insufficient network buffers in the receiving station. This is an indirect result of an application program behavior. For example, the application may be waiting for some other event to happen, or the application might not be releasing the frame buffer. An application is very slow. A fast station is overwhelming a slow one. Expert Alarms Reference Guide 71 Chapter 1 Station Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Station layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Duplicate Network Address The Expert generates the Duplicate Network Address alarm when it frames from more than one DLC address with the same network address. If the DLC source addresses are not routers, there is a serious error condition that could cause unpredictable and potentially destructive results. Check the configuration of the network stations and change one of the network addresses. Possible causes: Someone has assigned a new network address without checking the existing network addresses to assure that the assignment is unique. The network addresses are being dynamically allocated, or in some other way are being chosen by software, and the software is not assigning them properly. Someone has copied another station's configuration file, and has neglected to change the network address. If the DLC addresses differ in only one or two bit positions, and the number of frames associated with one of the DLC address is very small, then there may be a defective network adapter. If the DLC addresses are routers, this is not a problem. This alarm was triggered because the expert system has not yet identified the duplicate addresses as routers. The expert system learns about routers with the help of Routing Information Protocol (RIP) packets. If the expert system has not seen any RIP packets from a given DLC address since the beginning of the capture, it will assume that the address is that of an ordinary station rather than a router. RIP packets occur at 30 second intervals. If a RIP packet is not detected, it is either because 30 seconds have not elapsed, or because this segment of the network does not broadcast RIP packets. 72 Sniffer Technologies Expert Alarms and Thresholds Duplicate PDC Name Registration The Expert generates the Duplicate PDC Name Registration alarm when it detects a node attempting to initialize using the name of a Primary Domain Controller that is already running. Microsoft networks can have only a single Primary Domain Controller in a given domain. However, there can be multiple Backup Domain Controllers in a single domain. Possible cause: A PC has been added to the network that was configured offline or that was cloned from an existing PC. Change the listed PC to become a BDC for the domain or a PDC for another domain. ICMP Destination Unreachable The Expert generates the ICMP Destination Unreachable alarm when a station receives an ICMP destination unreachable message. Possible causes: This message is usually sent by a router, and may indicate a routing table problem. Examine the RIP information. If the RIP packets are not occurring at 30 second intervals, there may be an unstable condition on the network. The destination network does not exist. The request is “administratively disallowed” because the requester does not have the necessary access privileges. The station sending this packet may have a misconfigured netmask. ICMP Host Unreachable The Expert generates the ICMP Host Unreachable alarm when a station receives an ICMP host unreachable message. Possible causes: The station that sent the message has received a frame with an unknown destination address. The requested service is not available from the station that sent the message. The request is “administratively disallowed” because the requester does not have the necessary access privileges. Expert Alarms Reference Guide 73 Chapter 1 The situation can be investigated further by PINGing another host on the same segment as the one that sent the 'host unreachable' message. ICMP Inconsistent Subnet Mask The Expert generates the ICMP Inconsistent Subnet Mask alarm when the subnet mask in an ICMP address mask reply message does not match any of the subnet masks known to the Expert. Possible causes: The subnet mask is legitimate, but is not known to the Expert. Select the Expert Options command from the Tools menu. In the dialog box that appears, use the Subnet Masks tab to add known subnet masks to the Expert. The mask is not legitimate. This problem should be resolved immediately. Otherwise serious routing problems can result. ICMP Net Unreachable The Expert generates the ICMP Net Unreachable alarm when a station receives an ICMP network unreachable message. Possible causes: This message is usually sent by a router, and may indicate a routing table problem. Examine the RIP information. If the RIP packets are not occurring at 30 second intervals, there may be an unstable condition on the network. The destination network does not exist. The request is “administratively disallowed” because the requester does not have the necessary access privileges. The station sending this packet may have a misconfigured netmask. ICMP Parameter Problem The Expert generates the ICMP Parameter Problem alarm when it sees a station send an ICMP message indicating a parameter problem. 74 Sniffer Technologies Expert Alarms and Thresholds ICMP Port Unreachable The Expert generates the ICMP Port Unreachable alarm when a station receives an ICMP port unreachable message. The message indicates that this station previously transmitted a frame to a port. The station that sent the port unreachable message is saying that the port cannot be accessed by the requesting station. Possible causes: The requested service is not implemented or not active on the station that sent the message, or is in use. The request is “administratively disallowed” because the requester does not have the necessary access privileges. ICMP Redirect for Host The Expert generates the ICMP Redirect for Host alarm when a station receives an ICMP redirect message with the Code value set to 1 (redirect datagrams for the host). The Code value specifies to which datagrams this redirect message replies. The value of 1 indicates that datagrams sent to this host in the future should use the new route suggested by the router in the redirect message. Possible cause: A router may have sent the message to inform this station that a better route exists to its intended destination. This is not usually a problem, but it might be helpful to understand why one route is better than another. ICMP Redirect for Network The Expert generates the ICMP Redirect for Network alarm when station receives an ICMP redirect message with the Code value set to 0 (redirect datagrams for the network). The Code value specifies to which datagrams this redirect message replies. The value of 0 indicates that datagrams sent to this network in the future should use the new route suggested by the router in the redirect message. NOTE: Because subnets are extensively used in most networks today, code 0 is not often used anymore (subnets create an ambiguity about the subnet mask to be used to interpret a redirect for network message). Most hosts receiving an ICMP redirect with the Code value set to 0 simply treat the redirect as if it included the Code value 1 (redirect datagrams for host). Expert Alarms Reference Guide 75 Chapter 1 Possible cause: A router may have sent the message to inform this station that a better route exists to its intended destination. This is not usually a problem, but it might be helpful to understand why one route is better than another. ICMP Source Quench The Expert generates the ICMP Source Quench alarm when a station receives an ICMP source quench message. Possible causes: A congested router may have sent the message to inform this station to reduce its rate of frame transmission. The station that sent the source quench message may have a full buffer or some other kind of processing bottleneck. Transmit more slowly. The station that sent the message may be down or rebooting. The station that sent the message may have received an excessive number of bad frames. IP Fragment Missing The Expert generates the IP Fragment Missing alarm when it sees that an IP fragment on an NFS connection has been lost. This will usually result in an NFS layer retransmission. When the maximum transmission size (MTS) on one side of a router is different from the MTS on the other side, the router sometimes has to break the frame up into fragments. In such a case, the router assigns sequential numbers to the fragments – for example, 1 of 3, 2 of 3, 3 of 3. The “IP fragment missing” alarm is generated when one or more of these fragments fails to arrive at its destination. Possible causes: Heavy LAN traffic. Multiple paths between the endpoint stations. IP Fragment Out of Order The Expert generates the IP Fragment Out of Order alarm when it sees an IP layer fragment on an NFS connection that is not in sequential order. 76 Sniffer Technologies Expert Alarms and Thresholds When the maximum transmission size (MTS) on one side of a router is different from the MTS on the other side, the router sometimes has to break the frame up into fragments. In such a case, the router assigns sequential numbers to the fragments – for example, 1 of 3, 2 of 3, 3 of 3. The “IP fragment out of order” symptom is generated when one or more of these fragments arrive at their destination in a different sequence. This is not necessarily a problem, but it does indicate a condition that could become a problem under heavy LAN traffic conditions. Possible cause: There may be multiple paths between the endpoint stations. Misdirected Frame The Expert generates the Misdirected Frame alarm when it sees a station send a large number of frames in which the IP destination address is not on the local subnet and the DLC destination address is not that of a router advertising a route to the IP destination. Possible causes: This station does not know what subnet it is attached to. This station is using expired or canceled routes. A protocol other than RIP is being using to circulate routing information. The analyzer has not been configured properly to know its own subnet address or addresses, and the automatic mechanism that the analyzer uses to find these addresses has not been able to correct the problem. Multiple Routers to Local The Expert generates the Multiple Routers to Local alarm when the number of routers used to access a local station exceeds the Multiple Local Routers threshold. NOTE: The characterization of the network station as local is tentative. The expert system assumes that a station is local unless it has received a Routing Information Protocol (RIP) packet indicating a non-zero hop count to that station. Expert Alarms Reference Guide 77 Chapter 1 Possible causes: The router configurations may be incorrect. Resolve this by checking the router tables. The existing configuration may be legitimate. Multiple routers are sometimes used for security purposes, or to provide fault-tolerant network connections. There may be a duplicate network address shared by one of the routers and a local station. For VINES-IP networks this is not a problem. This symptom can help you see just how many paths are being used to reach a local node. If, after running the analyzer for several hours, the number of routers suddenly increases, then the network could be reacting to change. Multiple Routers to Remote The Expert generates the Multiple Routers to Remote alarm when the number of routers used to access a remote station exceeds the Multiple Remote Routers threshold. Possible causes: The router configurations may be incorrect. Resolve this by checking the router tables. The existing configuration may be legitimate. The redundant paths may have been set up deliberately to increase fault tolerance or to improve response time. NOTE: For VINES-IP networks this is not a problem. This alarm can help you see just how many paths are being used to reach a remote node. If, after running the analyzer for several hours, the number of routers suddenly increases, then the network could be reacting to change. Route Flapping The Expert generates the Route Flapping alarm when it sees a router change one or more routes from valid to invalid and back too frequently. This alarm is generated if more than four such changes are made to any route in one minute. Possible cause: 78 Sniffer Technologies There is an intermittent problem with the communication line. This causes links to fail and come back up. Expert Alarms and Thresholds Router Not Updating Routes The Expert generates the Router Not Updating Routes alarm when it sees a router let more than half of its advertised routes expire without confirmation. Possible causes: The router has crashed or reset. The router has been turned off. The router has been disconnected from this subnet. This router was reconfigured and no longer knows about its connection to this subnet. Router Storm The Expert generates the Router Storm alarm when it sees a router reconfirming one or more of its routes more frequently than the Router Storm threshold. Possible causes: The router software is sending its routing table more often than it should, possibly in response to external changes in the topology of the network. The routing updates are legitimate, but the routing software happened to pick several random update times that were at the minimum end of the range. This cause is likely if this diagnosis is isolated and of short duration. Router Superseded Too Frequently The Expert generates the Router Superseded Too Frequently alarm when it sees a router changing the metrics on one or more routes too frequently. The analyzer generates this alarm if any route changes between being the best route to its destination and not being the best route too frequently. Possible causes: There is an intermittent problem with the communication line. This causes links to fail and come back up. The network topology is changing, causing the metric of several routes to change. Expert Alarms Reference Guide 79 Chapter 1 Server Running BDC Shutting Down The Expert generates the Server Running BDC Shutting Down alarm when it sees a node acting as a Backup Domain Controller shut down. This alarm is only generated if a WINS Release packet is seen from the node that is shutting down. Possible cause: The server operating as the BDC was shut down. This can affect the ability of clients to log in to the network using accounts managed by this controller. Both accounts in the local domain and accounts in remote domains managed by the controller using Pass-Through Authentication could be affected. Server Running PDC Shutting Down The Expert generates the Server Running PDC Shutting Down alarm when it sees a node acting as a Primary Domain Controller shut down. Possible cause: The server operating as the PDC was shut down. If there is no other local Domain Controller (for example, a backup domain controller), the ability of clients to log in to the network using accounts managed by this controller could be affected. Both accounts in the local domain and accounts in remote domains managed by the controller using Pass-Through Authentication could be affected. Server Running WINS Shutting Down The Expert generates the Server Running WINS Shutting Down alarm when it sees a WINS Server shut down. This alarm is only generated if a WINS Release packet is seen from the node that is shutting down. WINS service for clients will be disrupted as they transition from the named WINS Server to an optionally configured backup. Backup WINS Servers for clients are configured using the properties sheet for the TCP/IP protocol in the Network Control Panel. This alarm is only generated if a WINS Release packet is seen from the node that is shutting down. It is normal to see a burst of WINS No Response alarms after this alarm. Possible cause: 80 Sniffer Technologies The server running the WINS Service shut down. Expert Alarms and Thresholds Source Address Is Broadcast The Expert generates the Source Address Is Broadcast alarm when it detects a frame with the source address field set to a broadcast address. A source address should not be a broadcast address. Time-to-live Exceeded in Transmit The Expert generates the Time-to-live Exceeded in Transmit alarm when a station receives an ICMP Time Exceeded message. The time-to-live field of the IP layer has reached zero (or the fragmentation reassembly time has been exceeded). Therefore, the frame has been discarded. Possible cause: If the time-to-live field has reached zero, it usually indicates that there is a routing table error somewhere in the network, but not necessarily on this local network. Try to locate the source of the original frame. If the fragmentation reassembly time has been exceeded, it usually indicates that a frame has been lost. Time-to-live Expiring The Expert generates the Time-to-live expiring alarm when it sees an IP frame with a time-to-live field set to 0 or 1, indicating that the frame is about to expire. Possible cause: If the time-to-live field has reached zero, it usually indicates that there is a routing table error somewhere in the network, but not necessarily on this local network. Try to locate the source of the original frame. Zero Broadcast Address The Expert generates the Zero Broadcast Address alarm when it sees an IP frame with an all-zero broadcast address sent. Possible cause: This is an obsolete form of TCP/IP broadcast address. It should not be used, because it can cause compatibility problems with stations that use all-ones broadcast addressing. Expert Alarms Reference Guide 81 Chapter 1 DLC Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the DLC layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Abort Delimiter Transmitted The Expert generates the Abort Delimiter Transmitted alarm when it detects an “Abort delimiter transmitted” MAC condition. Possible cause: This alarm can be caused by a faulty adapter. If this occurs frequently, check the adapter on the reporting station. AC Errors The Expert generates the AC Errors alarm when the NAUN (Nearest Active Upstream Neighbor) is unable to set the FCI (frame copied indicator) and ARI (address recognized indicator) bits in a frame it has copied. Burst Errors The Expert generates the Burst Errors alarm when it detects a signaling error in the cable. Note that this error is not caused by a station entering or leaving the ring – although the MAC protocol reports burst errors that occur when a station enters or leaves the ring, such errors are filtered by the Expert system, and do not trigger this alarm. Possible cause: There is probably a cabling problem. Check the fault domain. Collision after 64 Bytes The Expert generates the Collision after 64 Bytes alarm when it sees a collision taking place after 64 bytes of a frame have already been received. DLC Source Address Broadcast The Expert generates the DLC Source Address Broadcast alarm when it sees a frame with the source DLC address field set to a broadcast address. 82 Sniffer Technologies Expert Alarms and Thresholds Possible causes: The frame has been corrupted, possibly as a result of a collision or a malfunctioning station. If the problem recurs, look for a defective cable, transceiver, or network interface board. Source routing is being used, but the Expert is currently configured to disable interpretation of the RI bit. The DLC address is a mistake. Examine the surrounding frames for clues as to which DLC station sent this frame. Isolate the offending station by eliminating other known stations. When you have located the offending station, change the DLC address or remove that station from the network. Changing the DLC address may require replacing the network adapter. DLC Source Address Multicast The Expert generates the DLC Source Address Multicast alarm when it sees a frame with the source DLC address field set to a multicast address. Possible causes: The frame has been badly garbled, possibly as a result of a collision or a malfunctioning station. If the problem recurs, look for a defective cable, transceiver, or network adapter board. The DLC address is a mistake. Examine the surrounding packets for clues as to which DLC station sent this frame. Isolate the offending station by eliminating other known stations. When you have located the offending station, change the DLC address or remove that station from the network. Changing the DLC address may require replacing the network adapter. Duplicate Address Found The Expert generates the Duplicate Address Found alarm when it sees an affirmative response to the Duplicate Address Test MAC frame (another station on the ring has the same address). This is a token ring alarm. When the DLC address of the station trying to enter the ring is the same as the DLC address of a station already in the ring, the station trying to enter is automatically disconnected. That station should change its DLC address, and then try again to enter the ring. Changing the DLC address may require changing the network adapter. Expert Alarms Reference Guide 83 Chapter 1 FCI and ARI Bits Mismatch The Expert generates the FCI and ARI Bits Mismatch alarm when it sees a frame whose FCI (frame copied indicator) and ARI (address recognized indicator) bits have contradictory values. They should both be either 0 or 1. Possible cause: The destination station is overloaded and is losing frames. The destination station's adapter board has a problem. FDDI Ring Wrap The Expert generates the FDDI Ring Wrap alarm when it sees an SMT Neighbor Information Frame (NIF) with both the Rooted Station and the Peer Wrap bits in the Station State parameter set to true. All nodes on an FDDI ring transmit NIF frames to their downstream neighbors at regular intervals, enabling all nodes to determine the MAC addresses of their upstream and downstream neighbors. NIF frames include various parameters identifying the sending station, including the Station State parameter. This parameter includes two bits crucial to the generation of the FDDI Ring Wrap alarm, as follows: Rooted Station – If the Rooted Station bit is set to true, the sending station is a dual-attached part of the main FDDI ring and not a single-attached branch of a tree topology connected via a concentrator. Peer Wrap – If the Peer Wrap bit is set to true, the sending station is currently wrapping the primary and secondary rings due to a fault in the network. FDDI networks use two counter-rotating rings (called primary and secondary). This design increases reliability by providing a certain amount of fault tolerance. Dual-attached nodes connect to both rings. The primary ring carries the data traffic while the secondary ring serves as the redundant path. When a dual-attached station detects a fault on the primary ring, it attempts to route the data around the fault by wrapping the primary and secondary rings. This provides a way to get packets to a destination beyond the point of the fault by sending them in the reverse direction. The Expert generates the FDDI Ring Wrap alarm when this condition is detected. Possible causes: 84 Sniffer Technologies The adjacent station is off. Expert Alarms and Thresholds The station adjacent to the wrapped station was previously active, but now the status is “Removed.” Check the adjacent station to verify that it has been powered off. If the station is on, then it is possibly a link problem (see below). The adjacent link is down. If the station adjacent to the wrapped station is active, the problem is probably with the link. If the adjacent station is functioning normally, check both ends of the link for problems. It may be necessary to replace the link with a known good cable. SMT reporting incorrect information. Although the station is reporting that it is wrapped, if the adjacent station and the link are operating normally, then it may be an SMT problem on the station. You may have a faulty SMT implementation. Frame Has Alignment Problem The Expert generates the Frame Has Alignment Problem alarm when it captures a frame with an alignment error. A frame has an alignment error when its length is not a multiple of 8 bits, and therefore cannot be unambiguously resolved into bytes. Frame-Copied Errors The Expert generates the Frame-Copied Errors alarm each time the token ring adapter (while in receive/repeat mode) receives a frame addressed to it with the address recognized bits (ARI) not set equal to 00. Possible causes: A line hit has occurred. There may be a duplicate address on the network. Frequency Errors The Expert generates the Frequency Errors alarm when a station on the ring detects a frequency error. Frequency errors occur when the token ring adapter detects that its ring clock and the crystal clock differ by an excessive amount. The error causes the adapter to enter the monitor contention process. High Rate of Line/Burst Errors The Expert generates the High Rate of Line/Burst Errors alarm when the number of ring line errors plus burst errors (reported by the MAC protocol) for a particular station exceeds the Line/Burst Errors Count threshold. Expert Alarms Reference Guide 85 Chapter 1 Possible cause: A large number of these errors may indicate a hardware problem. Check the fault domain. High Rate of Physical Errors The Expert generates the High Rate of Physical Errors alarm when the rate of physical errors per second detected for a particular station exceeds the rate specified by the Physical Errors threshold. Physical level errors can include: Short frames (runts) Bad CRCs Lost frames Oversize frames Possible causes: A defective cable An excessively long cable Electrical noise on a cable Crosstalk A defective board (if the error is on the “out” counter) A faulty router, repeater, or gateway device. Examine the detail screen and look for patterns. Make notes as to which DLC stations are involved, and how frequently the errors occur. Compare this data with previous records if any are available. High Rate of Receiver Congestion Errors The Expert generates the High Rate of Receiver Congestion Errors alarm when the number of Receiver Congestion MAC frames sent in one minute (as reported by the MAC protocol) exceeds the Receiver Congestion Error Count threshold. Possible cause: 86 Sniffer Technologies The transport level software in the receiving station is not providing its receive buffers fast enough. Expert Alarms and Thresholds High Rate of Ring Purge Diagnosis The Expert generates the High Rate of Ring Purge Diagnosis alarm when the rate of physical ring purges per minute detected exceeds the rate specified by the Ring Purge Diag Count threshold. Possible causes: A lost token is usually the result of a station entering or leaving the ring, or some other physical configuration change. Too many ring purges may indicate faulty hardware. Check the fault domain. Invalid Frame Status Field The Expert generates the Invalid Frame Status Field alarm when the FCI (frame copied indicator) and ARI (address recognized indicator) bits occur twice in the Frame Status field (because this field is not protected by the frame check sequence). This alarm occurs when 2 ARI or 2 FCI bits are not equal. An alarm is also generated when the FCI is set to 1 and the ARI to 0 because this combination indicates that no ring station recognized the address but the frame was copied. Line Errors The Expert generates the Line Errors alarm when it detects an invalid character in a frame or token, or a Frame Check Sequence (FCS) error in a frame. Possible causes: Unless these errors are frequent, they are most likely caused by a reconfiguration of the ring. No remedial action is required. There is a hardware problem. Check the fault domain. There is crosstalk from voice traffic. A bridge copied the data incorrectly. Local Station with Route Designator The Expert generates the Local Station with Route Designator alarm when it detects a frame that indicates source routing is being used, but the source and destination stations are both on the same local ring. Expert Alarms Reference Guide 87 Chapter 1 Lost Frame Errors The Expert generates the Lost Frame Errors alarm each time the token ring adapter (while in transmit/stripping mode) transmits a frame and fails to receive the end of the transmitted frame before the physical trailer timer expires. As a result, the active monitor issues a new token. If this error occurs, there should be Soft Error MAC frames from the other station indicating Line errors, Burst errors, or Lost frame errors Multiple Local Ring Definition The Expert generates the Multiple Local Ring Definition alarm when it detects that source routing is being used. The routing information in two or more frames is contradictory – that is, two frames have different definitions of the local ring number. Possible cause: There is a configuration problem. For example, a device may have assumed that the local ring number is 1, rather than asking the ring parameter server. New Active Monitor The Expert generates the New Active Monitor alarm when the active monitor on the ring changes. Oversized Frame The Expert generates the Oversized Frame alarm when it captures an Ethernet packet whose length is greater than 1518 bytes. Receiver Congestion Errors The Expert generates the Receiver Congestion Errors each time it sees a Receiver Congestion MAC frame. Receiver congestion occurs when a station does not have enough room in its buffer to copy a frame for which it is the destination. Possible cause: 88 Sniffer Technologies The transport level software in the receiving station is not providing the receive buffers fast enough. Expert Alarms and Thresholds Returned to Ring The Expert generates the Returned to Ring alarm when it sees a station reenter the ring after having left the ring. A new alarm is generated each time the station returns to the ring. Ring Beaconing The Expert generates the Ring Beaconing alarm when it sees a station transmit a Beacon MAC frame in an attempt to bring the ring back into normal operation. Possible causes: There is a physical break in the ring (signal loss). Check the fault domain. A timeout has occurred during monitor contention. A station with a card set to the wrong data rate has attempted to enter the ring. Ring Purge Symptom The Expert generates the Ring Purge Symptom alarm when the rate of physical ring purges per minute detected exceeds the rate specified by the Ring Purge Symptom Count threshold. The active monitor sends a ring purge frame when it detects that the token has been lost. They are part of normal token ring behavior. However, an excessive amount of ring purge frames on the network can indicate potential hardware problems. NOTE: When the Ring Purge Symptom Count threshold is exceeded, the Expert generates a symptom. The Expert generates a diagnosis when the Ring Purge Diag threshold is exceeded. Possible causes: A lost token is usually the result of a station entering or leaving the ring, or some other physical configuration change. Too many ring purges may indicate faulty hardware. Check the fault domain. Expert Alarms Reference Guide 89 Chapter 1 Runt Frame The Expert generates the Runt Frame alarm when it detects a frame that ends short of the DLC address fields. A short frame can propagate through concentrators and produce more short frames through collisions. If you see a lot of short frames, all without source addresses, try to isolate the source. It may be necessary to physically unplug stations that you suspect are giving rise to these short frames. Possible causes: A damaged or defective cable. A bent or stapled cable that is producing reflections. Electrical interference. For example, a cable may be lying on a fluorescent fixture or close to an electrical motor. Same Source and Destination Address The Expert generates the Same Source and Destination Address alarm when it detects a frame with the same DLC source and destination address. Station Internal Errors The Expert generates the Station Internal Errors alarm when a station experiences a recoverable internal error. If the station is reporting multiple internal errors, the station is marginal. Station Off Ring The Expert generates the Station Off Ring alarm when it sees a station leave the ring, either purposely or due to a ring problem. If it is due to a ring problem, there will be other ring errors. This symptom is most significant if it occurs for a highly utilized station. A new symptom is generated each time the station leaves the ring. Station Removed The Expert generates the Station Removed alarm when the configuration report server asks a station to leave the ring. 90 Sniffer Technologies Expert Alarms and Thresholds Token Errors The Expert generates the Token Errors alarm when the token has been lost. This error should only be generated by the active monitor when it recognizes the need to create a new token. If this error occurs, there should be Soft Error MAC frames from the other station indicating Line errors, Burst errors, or Lost frame errors. Too Many Removes The Expert generates the Too Many Removes alarm when the number of “remove ring station” requests made by the configuration report server (reported by the MAC protocol) exceeds the Station Removed Request Count threshold. Too Many Returns The Expert generates the Too Many Returns alarm when the number of ring reentries for a particular station in a minute exceeds the Station Ring Entry Count threshold. Zero MAC Address The Expert generates the Zero MAC Address alarm when it detects a frame with a MAC address of all zeroes in either the source or destination address field. Examine the Decode display to determine whether the offending address was in the source or destination field. Possible causes: There is a hardware problem, possibly related to an adapter whose flash memory has been zeroed. There is a device driver defect. Some PC card drivers allow the MAC addresses of their adapters to be customized. A user may have replaced the original MAC address of the offending adapter with a custom address of all zeroes. The frame containing the offending address has been corrupted, possibly as a result of a collision or a malfunctioning station. If the problem recurs, look for a defective cable, transceiver, or network interface board. Expert Alarms Reference Guide 91 Chapter 1 The DLC address is a mistake. Examine the surrounding frames for clues as to which DLC station sent this frame. Isolate the offending station by eliminating other known stations. When you have located the offending station, change the DLC address or remove that station from the network. Changing the DLC address may require replacing the network adapter. WAN Connection Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the WAN Connection layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Backward Congestion The Expert generates the Backward Congestion alarm when the percentage of the total bytes sent from the network to the user in the last second in frames with the BECN bit set to true exceeds the FR Backward Congestion % threshold in the Expert UI Object Properties dialog box. Exceeded CIR The Expert generates the Exceeded CIR alarm when the number of bits transmitted from the user to the network within the last second exceeds the Committed Information Rate (CIR) for this PVC. This alarm is only generated if the CIR for the PVC is known and is not set to zero. Forward Congestion The Expert generates the Forward Congestion alarm when the percentage of the total bytes sent from the network to the user in the last second in frames with the FECN bit set to true exceeds the FR Forward Congestion % threshold in the Expert UI Object Properties dialog box. PVC Activation The Expert generates the PVC Activation alarm when the PVC Status information element's Active bit changes from false to true (0 to 1). PVC Bandwidth Change On some Frame Relay networks, the bandwidth assigned to each PVC is reported in the PVC Status information element. The Expert generates the PVC Bandwidth Change alarm when this reported value changes. 92 Sniffer Technologies Expert Alarms and Thresholds PVC Congestion The Expert generates the PVC Congestion alarm when it sees a PVC Status Information element sent by the network with the Receiver Not Ready bit set to true. Some Frame Relay networks do this to inform the user side of congestion. PVC Deactivation The Expert generates the PVC Deactivation alarm when the PVC Status information element's Active bit changes from true to false (1 to 0). PVC Deletion The Expert generates the PVC Deletion alarm in either of the following cases: A full status report was seen without an entry for the indicated DLCI (and there was an entry for the DLCI in the previous full status report). In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals. This can be a simple “keep alive” type of enquiry, in which the subscriber tells the WAN that it is still active, and receives a confirmation from the WAN. Alternatively, it can be a “full status” request, to which the WAN responds with a full status report giving the status of all the virtual connections attached to the local Frame Relay link. Typically, connected user equipment issues STATUS ENQUIRY messages every ten seconds with every sixth message requesting the full status report (though the exact interval may be different on some Frame Relay links). The Expert learns about the DLCIs on the Frame Relay link by monitoring the STATUS ENQUIRY requests and the status reports sent by the network in response to these requests and stores what it learns about the network in memory. A PVC Status information element for this DLCI with the Deleted bit set to true appears in a single PVC status report or UPDATE STATUS message. Traffic on Deleted DLCI The Expert generates the Traffic on Deleted DLCI alarm when it sees a frame with a header containing a DLCI that the link management protocol has reported as deleted. Expert Alarms Reference Guide 93 Chapter 1 Traffic on Inactive DLCI The Expert generates the Traffic on Inactive DLCI alarm when it sees a frame with a header containing a DLCI that the link management protocol has reported as inactive. Traffic on Unknown DLCI The Expert generates the Traffic on Unknown DLCI alarm when it sees a frame with a header containing a DLCI not known to the Expert. In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals (usually every 10 seconds). Typically, every sixth STATUS ENQUIRY message from user equipment requests a full status report from the network (though the exact interval may be different on some Frame Relay links). The Expert learns about the DLCIs on the Frame Relay link by monitoring the full status reports sent by the network in response to these requests and stores what it learns about the network in memory. 94 Sniffer Technologies Expert Alarms and Thresholds WAN Link Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the WAN Link layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Async Status Change of Unknown DLCI The Expert generates the Async Status Change of Unknown DLCI alarm when it sees a Single PVC status report or an UPDATE STATUS message with a PVC Status information element that refers to an unknown DLCI. These report/message types should only be used with existing (non-new) PVCs (PVCs which would also be known to the Expert from full status reports). DCE Sequence Number Error The Expert generates the DCE Sequence Number Error alarm in either of the following cases: The Send Sequence Number generated by the network equipment is not next in sequence after the previous Send Sequence Number. Sequence numbers on a Frame Relay link increment in ones from 1 to 255 and then restart again at 1. The Receive Sequence Number sent by the network equipment is not equal to the last Send Sequence Number sent by the user equipment. Delayed STATUS ENQUIRY In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals (usually every 10 seconds). At the beginning of capture, the Expert discovers this interval. Thereafter, if a STATUS ENQUIRY message is sent after the regular interval plus the time specified by the FR Status Enquiry Delay threshold expires, the Expert generates the Delayed STATUS ENQUIRY alarm. For example, if the Expert discovers the regular interval between STATUS ENQUIRY messages to be 10 seconds and the FR Status Enquiry Delay threshold is set to 1000ms (that is, one second), the Expert will generate the Delayed STATUS ENQUIRY alarm after 11 seconds have passed between STATUS ENQUIRY messages. DTE Sequence Number Error The Expert generates the DTE sequence number error alarm in either of the following cases: Expert Alarms Reference Guide 95 Chapter 1 The Send Sequence Number generated by the user equipment is not next in sequence after the previous Send Sequence Number. Sequence numbers on a Frame Relay link increment in ones from 1 to 255 and then restart again at 1. The Receive Sequence Number sent by the user equipment is not equal to the last Send Sequence Number sent by the network equipment. Irregular Full Status Report Requests In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals (usually every 10 seconds). Typically, every sixth STATUS ENQUIRY message from user equipment requests a full status report from the network (though the exact interval may be different on some Frame Relay links). The Expert learns the interval for the connected link at the beginning of capture. Thereafter, the Expert monitors the frequency of full status report requests and generates the Irregular full status report requests alarm when the number of STATUS ENQUIRY messages between full status requests is different from the learned interval. The Expert also generates this alarm if a STATUS ENQUIRY does not contain a full status report request when expected. Link Management Message Format Error The Expert generates the Link Management Message Format Error alarm when it detects any of the following problems in a link management message: 96 The Unnumbered Information byte (03 hex) contains incorrect information. There is a change in the value of the protocol discriminator. There is an unknown protocol discriminator. The Dummy Call Reference byte (00 hex) contains incorrect information. The Locking Shift information element is missing (this applies only to the ANSI link management protocol). An unsupported message type is seen (that is, a message type other than STATUS ENQUIRY, STATUS, or UPDATE STATUS). An unknown information element is seen (that is, an information element not used in Frame Relay link management messages). An information element is seen that is incompatible with the type of message containing it or with the current link management protocol. Sniffer Technologies Expert Alarms and Thresholds An information element is seen in incorrect order. A known information element is seen with an incorrect length. A report type which is unknown or unsupported by the current link management protocol is seen. An information element incompatible with the report type containing it is seen. A STATUS message containing a full status report lists DLCIs in the wrong order (DLCIs should be listed from the least to the greatest). The same DLCI is listed more than once in a STATUS message containing a full status report. More than one PVC Status information element appears in a STATUS message carrying a Single PVC Status report type. Link Management Message Minor Format Error The Expert generates the Link Management Message Minor Format Error (minor) alarm when it detects either of the following problems in a link management message: A full status report was seen with a PVC Status information element whose Deleted bit was set to true, indicating the PVC was deleted. This indicates a minor error since if a PVC is deleted, it should not be listed at all in the full status report. A Single PVC status report or an UPDATE STATUS message was seen with a PVC Status information element whose New bit was set to true. This indicates a minor error since these report/message types should only be used with existing (non-new) PVCs. Network Equipment Reset The Expert generates the Network Equipment Reset alarm when it sees a STATUS message sent by the network side of the link with a Send Sequence Number set to 1. The sequence number of 1 indicates that this is the first STATUS message sent by the network since a reset. Note that the Expert only generates this alarm if 1 is not the correct sequential value for this message. Sequence numbers on a Frame Relay link increment in ones from 1 to 255 and then restart again at 1. New PVC Configured The Expert generates the New PVC Configured alarm when it sees a full status report listing a previously unknown DLCI. Expert Alarms Reference Guide 97 Chapter 1 No STATUS ENQUIRY When Expected In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals (usually every 10 seconds). At the beginning of capture, the Expert discovers this interval. Thereafter, if a STATUS ENQUIRY message is sent after the regular interval plus the time specified by the FR Status Enquiry Waiting Time threshold expires, the Expert generates the No STATUS ENQUIRY When Expected alarm. For example, if the Expert discovers the regular interval between STATUS ENQUIRY messages to be 10 seconds and the FR Status Enquiry Waiting Time threshold is set to 5000ms (that is, five seconds), the Expert will generate the No STATUS ENQUIRY When Expected alarm after 15 seconds have passed between STATUS ENQUIRY messages. No STATUS Following STATUS ENQUIRY The Expert generates the No STATUS Following STATUS ENQUIRY alarm when the network side of the link does not respond to a STATUS ENQUIRY request with a STATUS message before the user side of the link sends the next STATUS ENQUIRY message. In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals. This can be a simple “keep alive” type of enquiry, in which the subscriber tells the WAN that it is still active, and receives a confirmation from the WAN. Alternatively, it can be a “full status” request, to which the WAN responds with a full status report giving the status of all the virtual connections attached to the local Frame Relay link. PVC Activation The Expert generates the PVC Activation alarm when the PVC Status information element's Active bit changes from false to true (0 to 1). PVC Bandwidth Change On some Frame Relay networks, the bandwidth assigned to each PVC is reported in the PVC Status information element. The Expert generates the PVC Bandwidth Change alarm when this reported value changes. PVC Congestion The Expert generates the PVC Congestion alarm when it sees a PVC Status Information element sent by the network with the Receiver Not Ready bit set to true. Some Frame Relay networks do this to inform the user side of congestion. 98 Sniffer Technologies Expert Alarms and Thresholds PVC Deactivation The Expert generates the PVC Deactivation alarm when the PVC Status information element's Active bit changes from true to false (1 to 0). PVC Deletion The Expert generates the PVC Deletion alarm in either of the following cases: A full status report was seen without an entry for the indicated DLCI (and there was an entry for the DLCI in the previous full status report). In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals. This can be a simple “keep alive” type of enquiry, in which the subscriber tells the WAN that it is still active, and receives a confirmation from the WAN. Alternatively, it can be a “full status” request, to which the WAN responds with a full status report giving the status of all the virtual connections attached to the local Frame Relay link. Typically, connected user equipment issues STATUS ENQUIRY messages every ten seconds with every sixth message requesting the full status report (though the exact interval may be different on some Frame Relay links). The Expert learns about the DLCIs on the Frame Relay link by monitoring the STATUS ENQUIRY requests and the status reports sent by the network in response to these requests and stores what it learns about the network in memory. A PVC Status information element for this DLCI with the Deleted bit set to true appears in a single PVC status report or UPDATE STATUS message. Unexpected Link Management Message The Expert generates the Unexpected Link Management Message alarm when a link management message is seen emanating from the wrong side of the link (for example, a STATUS message coming from the user side of the link). Unsolicited STATUS Message The Expert generates the Unsolicited STATUS Message alarm when it sees a STATUS message sent from the network without a corresponding STATUS ENQUIRY message sent from the user. Note that this alarm is not generated when an unsolicited STATUS messages containing a Single PVC Status report is seen. Expert Alarms Reference Guide 99 Chapter 1 In a correctly operating Frame Relay link, connected user equipment issues STATUS ENQUIRY messages at regular intervals. This can be a simple “keep alive” type of enquiry, in which the subscriber tells the WAN that it is still active, and receives a confirmation from the WAN. Alternatively, it can be a “full status” request, to which the WAN responds with a full status report giving the status of all the virtual connections attached to the local Frame Relay link. User Equipment Reset The Expert generates the User Equipment Reset alarm when it sees a STATUS ENQUIRY message sent by the user side of the link with a Send Sequence Number set to 1 and a Receive Sequence Number set to 0. These sequence numbers indicate that this is the first STATUS ENQUIRY message sent since the link was reset. Wrong Report Type in STATUS Message The Expert generates the Wrong Report Type in STATUS Message alarm when it sees a STATUS message sent in response to a request for a full status report that contains only sequence number exchange information (and does not include the requested full status report). ATM Application Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the ATM Application layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Attempt to Bind L3 Address to Multiple MPOA Devices The Expert generates the Attempt to Bind L3 Address to Multiple MPOA Devices alarm when it detects an attempt to bind a single Layer Three address to more than one MPOA device. MPOA devices (that is MPCs and MPSs) serve Layer Three address. Because of this, it is a violation of the MPOA specification for a single L3 address to be bound to more than one MPOA device. Data Frames Detected Following Data Plane Purge The Expert generates the Data Frames Detected Following Data Plane Purge alarm when it detects data frames on an MPC-MPC data flow after a data plane purge was sent. Data frames should not be sent across an MPC-MPC data flow following a data plane purge. This connection should be dropped. 100 Sniffer Technologies Expert Alarms and Thresholds Egress MPCs send data plane purge messages to Ingress MPCs to purge ingress cache entries (that is, to tell the Ingress MPC that a previously-cached shortcuts are no longer valid). Because these ingress cache entries are no longer valid, there should not be data frames on them. DDVCC Idle Period Exceeds VCC Timeout C12 The Expert generates the DDVCC Idle Period Exceeds VCC Timeout C12 alarm when a LEC fails to release a Data Direct Virtual Circuit Connection that has been idle for longer than the time specified by its C12 VCC Time-out Period initial state parameter. The Expert uses timeout values as follows: The Expert uses the timeout value specified by the C12 VCC Time-out Period initial state parameter if it has been able to learn it from network traffic. The ATM Forum specifies a default value of 20 minutes for this timeout, with unlimited minimum and maximum values. If the Expert has not been able to learn the C12 VCC Time-out Period parameter, it uses the DDVCC VCC Timeout Period C12 threshold in the Alarms tab of the Expert UI Object Properties dialog box. By default, NAI sets this threshold to 20 minutes (the same default as the ATM Forum). DDVCC Not Created in Response to LE ARP When a data frame arrives at the LEC, it sends an LE_ARP request to the LES to get the ATM address of the LEC representing the destination LAN MAC address. Once it receives the ATM address of the destination LEC, it sets up the data direct VCC to the destination LEC to carry the unicast data between the LAN MAC address pair. The Expert generates the DDVCC Not Created in Response to LE ARP alarm when it sees an LE_ARP sent to the LES without a corresponding data direct VCC being set up within the time limit specified by the DDVCC Setup in Response to LE ARP threshold in the Alarms tab of the Expert UI Object Properties dialog box. Default MCFVCC Not Created in Response to LE ARP The Expert generates the Default MCFVCC Not Created in Response to LE ARP alarm when the analyzer sees a LEC send an LE_ARP to the LES for the destination MAC address FF FF FF FF (that is, broadcast) without a corresponding Multicast Forward VCC being set up by the BUS within the time limit specified by the Default MCFVCC Setup in Response to LE_ARP threshold in the Alarms tab of the Expert UI Object Properties dialog box. Expert Alarms Reference Guide 101 Chapter 1 The Multicast Forward flow is set up by the BUS with the LEC once the LEC has established the MCS with the BUS. The MCF is used to carry all broadcast, multicast, and unknown data from the BUS to the connected LECs. At least one MCF flow remains active for each LEC during its participation in the ELAN. The MCS is unidirectional and can be either point-to-point or point-to-multipoint. There are different MCF flows for both 802.3 and 802.5 networks. Default MCSVCC Not Created in Response to LE ARP This alarm is generated when the analyzer sees a LEC send an LE_ARP to the LES for the destination MAC address FF FF FF FF (that is, broadcast) without a corresponding multicast send VCC being set up with the BUS within the time limit specified by the Default MCSVCC Setup in Response to LE_ARP threshold in the Alarms tab of the Expert UI Object Properties dialog box. The Multicast Send data flow is set up by the LEC with the BUS to carry all broadcast, multicast, and unknown data from the LEC to the BUS (and in some cases, from the BUS to the LEC). When a LAN MAC frame with a broadcast destination address arrives at the LEC, it sends it to the BUS. At least one MCS flow remains active from each LEC during its participation in the ELAN. The MCS is point-to-point and bidirectional. There are different MCS flows for both 802.3 and 802.5 networks. Excessive Unknown Frames Issued by LEC The Expert generates the Excessive Unknown Frames Issued by LEC alarm when it sees a LEC send too many unknown frames without sending an LE_ARP to the LES. The LEC should issue an LE_ARP when unknown frames (with unicast MAC addresses) do not have a data direct VCC. Frames Seen on Default Flow After Shortcut Active The Expert generates the Frames Seen on Default Flow After Shortcut Active alarm when it detects traffic on a default flow for which an MPOA shortcut has already been created. Once a shortcut for a flow has been created, its traffic must be sent over the shortcut. LANE Config Phase Failure During the Configuration phase of LAN Emulation communications, a LEC sends a Configure Request frame to the LECS in order to obtain configuration information, including the address of a LES for an ELAN that meets the LEC’s needs. In response, the LECS sends the Configure Response frame indicating whether the candidate LEC will be allowed to attempt the Join phase with a LES. 102 Sniffer Technologies Expert Alarms and Thresholds The Expert generates this alarm when the Configure Response frame sent from the LECS to the LEC indicates failure (that is, the LEC will not be allowed to attempt the Join phase). Possible causes: There is a version incompatibility between the LEC and the LECS. The LECS denied the Configure Request for security reasons. The LEC supplied conflicting parameters in the Configure Request frame. Look for instances of alarms related to this failure – for example, Configuration Fail, Insufficient Resources, Denied Access to ELAN, Insufficient Info from LEC, LANE Protocol Version Unsupported by ELAN, and so on. The data provided for these alarms will help you understand the reasons for the Configuration Phase Failure. LANE Cfg Restart Delay < Min Recfg Delay During the Configuration phase of LAN Emulation communications, a LEC sends a Configure Request frame to the LECS in order to obtain configuration information, including the address of a LES for an ELAN that meets the LEC’s needs. If the Configure Request is denied (or if no response is returned at all), the LEC must wait for an interval before restarting the Configuration process. This interval can be any random value between the two LEC initial state parameters C38 (Maximum Reconfigure Delay) and C37 (Minimum Reconfigure Delay). Unfortunately, these two initial state parameters are not carried in the data stream, so the Expert cannot learn them from analyzing traffic. Because of this, it uses the Minimum Reconfigure Delay C37 threshold in the Expert Options dialog box to specify when the LANE Config Restart Delay < Minimum Reconfig Delay C37 alarm should be generated. If a LEC restarts the configuration phase before the Minimum Reconfigure Delay C37 timer has expired, the Expert will generate this alarm. By default, the Minimum Reconfigure Delay C37 threshold is set to the standard value for C37 specified by the ATM Forum – one millisecond. Expert Alarms Reference Guide 103 Chapter 1 LANE Cfg Restart Delay > Max Recfg Delay During the Configuration phase of LAN Emulation communications, a LEC sends a Configure Request frame to the LECS in order to obtain configuration information, including the address of a LES for an ELAN that meets the LEC's needs. If the Configure Request is denied (or if no response is returned at all), the LEC must wait for an interval before restarting the Configuration process. This interval can be any random value between the two LEC initial state parameters C38 (Maximum Reconfigure Delay) and C37 (Minimum Reconfigure Delay). Unfortunately, these two initial state parameters are not carried in the data stream, so the Expert cannot learn them from analyzing traffic. Because of this, it uses the Maximum Reconfigure Delay C38 threshold in the Expert UI Object Properties dialog box to specify when the LANE Config Restart Delay > Max Reconfig Delay C38 alarm should be generated. If a LEC waits longer than the time specified by this threshold, the Expert will generate the LANE Cfg Restart Delay > Max Recfg Delay alarm. By default, the Maximum Reconfigure Delay C38 threshold is set to the standard value for C38 specified by the ATM Forum – five seconds. No Resolution Request Issued After Excess Default Flow The Expert generates the No Resolution Request Issued After Excess Default Flow alarm when the number of data frames sent by an MPC on a default flow without sending a Resolution Request exceeds its Shortcut Setup Frame Count (MPC-p1) initial state parameter within the time specified by the Shortcut Setup Frame Time (MPC-p2) initial state parameter. The MPC-p1 and MPC-p2 initial state parameters are used to ensure that an MPC attempts to create a shortcut (by sending a Resolution Request) if a large amount of traffic is being sent from an MPC to an MPS (that is, on a default flow; a default flow is any LANE flow whose target L2 address is bound to an MPS). Shortcut Create Failed After Resolution Reply The Expert generates the Shortcut Create Failed After Resolution Reply alarm when it detects that an MPC has received an MPOA Resolution Reply from the MPS but has not attempted to send data over a corresponding MPC-MPC shortcut within the time specified by the Shortcut-Setup following Res Reply Time threshold in the Alarms tab of the Expert UI Object Properties dialog box. A shortcut must be attempted following the receipt of a successful MPOA Resolution Reply. 104 Sniffer Technologies Expert Alarms and Thresholds ATM Flow Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the ATM Flow layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). ARE or STE Frames on Data Direct VCC When a LEC is emulating a token ring (that is, an 802.5 LEC) that uses source routing, it will occasionally need to send Token Ring All Routes Explorer (ARE) and Spanning Tree Explorer (STE) frames. Since these frames typically need to be sent to more than one destination, LAN Emulation requires that they be sent to the ELAN via the BUS (over a multicast send VCC). The Expert generates the ARE or STE Frames on Data Direct VCC alarm when it sees an ARE or STE frame on a data direct VCC (a violation of the LAN Emulation standard). ARP Request Multi Queued The Expert generates the ARP Request Multi Queued alarm when it sees a LAN Emulation entity resend an ARP request with the same Transaction ID as the original ARP request. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Attempt to Bind DLC to MPC and Non-MPOA Device The Expert generates the Attempt to Bind DLC to MPC and Non-MPOA Device alarm when it detects an attempt to bind a DLC address to both an MPC and a non-MPOA device. All MPOA devices have layer two (L2) DLC addresses assigned to them. Although an MPS can share its L2 DLC address with a representative LEC, an MPC cannot share its L2 DLC address with any other device (LANE or otherwise). Expert Alarms Reference Guide 105 Chapter 1 Attempt to Multiply Bind MPOA Devices to LEC The Expert generates the Attempt to Multiply Bind MPOA Devices to LEC alarm when it detects one of the following situations: An attempt to bind a LEC to both an MPC and an MPS. An attempt to bind a LEC to more than one MPC or MPS. A LEC can only be bound to a single MPC or MPS. Broadcast/Multicast Data Frames on DDVCC LAN Emulation requires that all broadcast and multicast frames be sent to the ELAN via the BUS (over a multicast send VCC). The Expert generates the Broadcast/Multicast Data Frames on DDVCC alarm when it sees a broadcast or multicast frame on a data direct VCC (a violation of the LAN Emulation standard). Cache ID Set to Zero in DLL Header Ext The Expert generates the Cache ID Set to Zero in DLL Header Ext alarm when it detects an MPOA Egress Cache Purge Request control frame sent from an MPC without the Cache ID included in the DLL Header extension. The MPOA specification requires that the Cache ID field be included in the DLL Header extension of an Egress Cache Purge Request so that the MPS knows which egress cache entry to invalidate. Config Fail, Insufficient Resources LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). The Expert generates the Config Fail, Insufficient Resources alarm when either a Configure, Join, or Register request was made to a LAN Emulation entity that was unable to grant the request due to a lack of resources (for example, an inability to create new VCCs or a lack of table space). Possible cause: 106 Sniffer Technologies The LES, LECS, or BUS is temporarily overloaded due to excessive demands from LECs. Expert Alarms and Thresholds Config Frag Info TLV Not Recognized by LECS During the Configuration phase of LAN Emulation communications, a LEC sends a Configure Request frame to the LECS in order to obtain configuration information, including the address of a LES for an ELAN that meets the LEC's needs. The Configure Request frame can optionally contain the Config Frag Info TLV. The Config Frag Info TLV tells the LECS that it can send TLV data to this LEC that will not fit into a single frame over multiple frames (that is, the LECS can fragment the TLV data). The Expert generates the Config Frag Info TLV Not Recognized by LECS alarm when the LECS sends a Configure Response frame with the Status field set to 24 (decimal), indicating that it does not recognize the Config Frag Info TLV sent by the LEC. Possible cause: A Version 2.0 LEC sent a Configure Response frame to a Version 1.0 LECS with the Config Frag Info TLV present. The Config Frag Info TLV is only supported in Version 2.0. Config Request Multi Queued The Expert generates the Config Request Multi Queued alarm when it sees a A LAN Emulation entity resend a Configure Request frame with the same Transaction ID as the Configure Request frame. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Config Request Response Timeout The Expert generates the Config Request Response Timeout alarm when a LEC sends a Configure Request frame to the LECS but does not receive a Configure Response frame before the timeout value specified by the C7 Control Time-out initial state parameter. The ATM Forum specifies a minimum value of 10 seconds for this timeout, with a default value of 30 seconds and a maximum of 300 seconds. Expert Alarms Reference Guide 107 Chapter 1 The C7 Control Time-out is sent by the LEC as part of the Configure Request frame. The Expert watches for the timeout value and then uses it to measure the responses to the Configure Request. Possible causes: The network is experiencing congestion. The LECS has received a large number of simultaneous requests from various LECs. Config Request Retry Frame Mismatch The Expert generates the Config Request Retry Frame Mismatch alarm when a LEC resends a Configure Request frame with a field changed other than the Transaction ID. This is a violation of the LAN Emulation standard. Whenever a LEC needs to resend a Configure Request frame (perhaps because there was no response from the LECS), it must send the exact same original request frame with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Config Request Retry Rate Exceeds 1 Frame/sec The Expert generates the Config Request Retry Rate Exceeds 1 Frame/sec alarm when it detects a LEC resending the same Configure Request frame more than once per second. Control Request on DDVCC Retry Frame Mismatch The Expert generates the Control Request on DDVCC Retry Frame Mismatch alarm when a LEC resends a Control Request frame on a data direct VCC with a field changed other than the Transaction ID. This is a violation of the LAN Emulation standard. The exact type of Control Request frame could have been Flush Request, Ready Indication, or Ready Query. Whenever a LEC needs to resend a Control Request frame (perhaps because there was no response), it must send the exact same original request frame with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. 108 Sniffer Technologies Expert Alarms and Thresholds Control Request on DDVCC Retry Rate > 1 Frame/Sec The Expert generates the Control Request on DDVCC Retry Rate > 1 Frame/Sec alarm when it detects a LEC resending the same Control Request frame more than once per second on a data direct VCC. The exact type of Control Request frame resent could have been Flush Request, Ready Indication, or Ready Query. Database Summary Exchange Lockstep Failure The Expert generates the Database Summary Exchange Lockstep Failure alarm if the rules governing the request \ response sequence between two PNNI nodes exchanging Database Summary packets are violated. Neighboring peers in a PNNI domain exchange database information through a precisely governed sequence of Database Summary packets. During the initial sequence (called the “Negotiation State”), the two nodes establish master and slave roles. Following this sequence, the “Exchanging State” begins. During the Exchanging State, only a single Database Summary packet can be outstanding at one time. A Database Summary packet is considered to be outstanding until the corresponding acknowledgment is sent from the receiving node, or the DsRxmtInterval timer in the packet expires (at which point a new Database Summary packet can be sent). The Expert generates this alarm if the master sends the same Database Summary packet more than once (as determined by the packet's sequence number), or if the slave fails to acknowledge a packet. Database Summary Master/Slave Negotiation Timeout The Expert generates the Database Summary Master/Slave Negotiation Timeout alarm when the time it takes for the initial master/slave negotiation between two PNNI nodes exceeds the Max time for DB Summary Master/Slave Negotiation threshold in the Alarms tab of the Expert UI Object Properties dialog box. PNNI nodes exchange information from their routing databases to determine which switches belong to the same peer group and to allow efficient routing of packets outside the peer group. During the initial Database Summary exchange between two PNNI nodes, the node pair must determine which is the master and which is the slave. If this negotiation sequence takes longer than the threshold, the Expert generates the alarm. Denied Access to ELAN The Expert generates the Denied Access to ELAN alarm when either a Configure Request control frame or a Join Request control frame is sent to the ELAN by the LEC and is denied for security reasons. Expert Alarms Reference Guide 109 Chapter 1 Possible cause: The Configure Request control frame is sent over a Configuration Direct VCC. The ATM address used by the requesting station to establish this VCC may be unknown to the ELAN and therefore considered a security risk. DLL Header Extension Not Found in Frame The Expert generates the DLL Header Extension Not Found in Frame alarm when it detects either an MPOA Resolution Reply or MPOA Egress Cache Purge Request control frame that does not include the DLL Header extension. The MPOA specification requires that both of these control frame types include this extension. Duplicate ATM Destination Address Registration The Expert generates the Duplicate ATM Destination Address Registration alarm when a LEC attempts to register an ATM address with the LES that is already in the LES' cache of addresses. The registration attempt could be in either a Join Request message or a Registration Request message. Duplicate LAN Destination Registration The Expert generates the Duplicate LAN Destination Registration alarm when it sees a Join Response or Register Response frame sent from the LES to the LEC with the Status Code 4 (decimal). This status code indicates that the LAN destination address that the LEC attempted to register with the LEC was already in the LES's cache of addresses.LEC Possible cause: A LEC attempted to register a LAN address with the LES that was already in the LES' cache of addresses with a mapping to a different LEC. The registration attempt could be in either a Join Request message or a Registration Request message. In either case, the duplication may cause the Register Request or Join Request to fail. Egress Cache Tag Was Not Issued The Expert generates the Egress Cache Tag Was Not Issued alarm when it detects an MPOA Resolution Request control frame that does not include the Egress Cache Tag extension. The MPOA specification requires that the MPOA Resolution Request control frame include the Egress Cache Tag extension. 110 Sniffer Technologies Expert Alarms and Thresholds Failed to Issue Resolution Req. Following Trigger The Expert generates the Failed to Issue Resolution Req. Following Trigger alarm when an MPC fails to issue an MPOA Resolution Request in response to an MPOA Trigger message sent to it by an MPS. Ingress MPSs send MPOA Triggers to Ingress MPCs to cause them to send Resolution Requests. Failure on the part of the MPC to respond with a Resolution Request is a violation of the specification. Failure Reported in Reply CIE The Expert generates the Failure Reported in Reply CIE alarm when it detects an MPOA control response frame with the client information element (CIE) code (the first octet of the CIE) set to something other than 0x00, indicating a failure (0x00 is the only CIE code value indicating success). Flush Request Multi Queued The Expert generates the Flush Request Multi Queued alarm when a LAN Emulation entity resends a Flush request with the same Transaction ID as the original Flush request. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. IG Tag Violation The Expert generates the IG Tag Violation alarm if the Information Group (IG) Tag bits are set to a non-legal value. PNNI routing frames use nested TLVs known as Information Groups (IG) to encode PNNI information. When a PNNI frame is sent, all IGs must be set to one of the legal values. These values are: Optional Summarizable Expert Alarms Reference Guide 111 Chapter 1 Non-transitive The Expert generates this alarm if a PNNI 1.0 frame is seen with an IG set to something other than one of these values. The exception to this rule is the Transit Network IG. The legal values for this IG are: Optional Summarizable Transitive The Expert generates this alarm for a Transit Network IG if its value is set to something other than the preceding three values. Insufficient Info from LEC, Service Refused The Expert generates the Insufficient Info from LEC, Service Refused alarm when a LEC sends a Configure Request frame to the LECS in order to be assigned to an ELAN. However, the LECS was unable to assign the LEC to an ELAN because the LEC did not provide sufficient information in its Configure Request frame. Invalid ATM Primary Address The Expert generates the Invalid ATM Primary Address alarm when it sees a Control Request frame that is only supposed to be sent with the source address set to the LEC's primary source address (for example, a Join Request) sent with some other non-zero value. Possible cause: A LEC mistakenly set the source address in a Join Request to one of its secondary ATM addresses. Invalid CIE Contents The Expert generates the Invalid CIE Contents alarm when it sees an MPOA control message with one of the fields in the client information element not filled in as defined in section 5.3 of the MPOA specification. Invalid Client ATM Primary Source Address The Expert generates the Invalid Client ATM Primary Source Address alarm when the LE Service sends a control response frame with the Status field set to 10 (decimal), indicating that the source ATM address was invalid. 112 Sniffer Technologies Expert Alarms and Thresholds Possible cause: A LEC sent a Configure, Join, Register, ARP, Flush, or Verify Request with an invalid value in the SOURCE-ATM-ADDRESS field. The format of the ATM address was not recognizable or was invalid for some other reason. Invalid Client MAC Address The Expert generates the Invalid Client MAC Address alarm when the LE Service sends a control response frame with the Status field set to 9 (decimal), indicating one of the following problems: A LEC sent a Configure, Join, or Flush Request that indicated a multicast address in the SOURCE-LAN-DESTINATION field. This field is used to list unicast MAC addresses represented by the LEC. It is a violation of the LAN Emulation standard for multicast addresses to be listed there. A LEC that indicated an ELAN type of 802.3\Ethernet sent a Configure, Join, Register, ARP, or Flush Request with a Route Descriptor in the SOURCE-LAN-DESTINATION field. Route Descriptors are only used in 802.5\Token Ring ELANs. A LEC sent an ARP Request to resolve multicast MAC addresses. Invalid Common Header Contents The Expert generates the Invalid Common Header Contents alarm if it sees an MPOA control message with an illegal value in one of the fields of the common header portion of the message. Figure 1-4 shows the fields in the common header portion of an MPOA control message. Figure 1-4. Common Header of an MPOA Control Message Expert Alarms Reference Guide 113 Chapter 1 Invalid Fixed Header Contents The Expert generates the Invalid Fixed Header Contents alarm if it sees an MPOA control message with an illegal value in one of the fields of the fixed header portion of the message. Figure 1-5 shows the fields in the fixed header portion of an MPOA control message. Figure 1-5. Fixed Header of an MPOA Control Message Invalid Frame Contents The Expert generates the Invalid Frame Contents alarm when it detects a PNNI frame which exhibits one or more of the following errors: The packet length in the PNNI header exceeds the received frame length. The version indicated in the Protocol Version field of the PNNI header is unsupported. One or more parameters in the PNNI frame were outside of the valid range. Invalid frame type. This occurs when the value of the Packet Type field in the PNNI packet header is not set to one of the five valid PNNI frame types (Hello, PTSP, PTSE Acknowledgment, Database Summary, or PTSE Request). The originating Peer Group ID or Node ID in the PNNI frame does not match that of the node issuing the frame. Invalid Hello Frame The Expert generates the Invalid Hello Frame alarm when it detects a PNNI Hello Frame exhibiting one or more of the following problems: 114 Sniffer Technologies Expert Alarms and Thresholds A “top level unknown TLV” with its mandatory bit set to true. As described in Section 5.6.2.3 of Version 1.0 of the ATM Forum's PNNI specification, such Hello frames must be discarded by the receiving station. The “Hello Interval” value is set to zero The Port ID is set to zero The Peer Group ID, if known, does not match that of the other nodes on the link. This is only true for Hello packets seen on SVCC RCCs (routing control channels set up at the parent level of the PNNI hierarchy between Logical Group Nodes). Invalid MPOA MPS MAC Address Count in Device Type TLV The Expert generates the Invalid MPOA MPS MAC Address Count in Device Type TLV alarm when the MAC Address Count in a Device Type TLV indicates a violation of the MPOA specification. MPOA uses the Device Type TLV to enable MPOA entities (MPCs and MPSs) to discover one another. Its transmission is required with LANE control messages, including LE_REGISTER and LE_ARP requests and responses. The first octet of every Device Type TLV indicates the type of device that sent it (for example, an MPC, an MPS, and so on). Depending on the value of this first octet, the second octet of the TLV (which indicates the number of MAC addresses) must be a certain value. If the detected value of the second octet is not valid given the value of the first octet, this alarm is generated. Possible causes: The Device Type TLV indicated a Non-MPOA device type (value of the first octet is 0), but the MAC address counts is non-zero. The Device Type TLV indicated an MPC device type (value of the first octet is 2), but the MAC address count is non-zero. The Device Type TLV indicated an MPS and MPC device type (value of the first octet is 3), but the MAC address count is zero. A non-zero MPS MAC address count for device types including Non MPOA (0) or MPC (2) Invalid MPOA Packet Size The Expert generates the Invalid MPOA Packet Size alarm when it detects an MPOA packet whose advertised size does not match the actual size of the captured packet. Expert Alarms Reference Guide 115 Chapter 1 Invalid Tag in Data Frame Header The Expert generates the Invalid Tag in Data Frame Header alarm when it detects a data frame using MPOA tagged encapsulation without a valid, non-zero tag in the frame. MPOA can use either LLC encapsulation or MPOA tagged encapsulation for data frames. A valid, non-zero tag must be included as part of each data frame that uses MPOA tagged encapsulation. Join Request Multi Queued The Expert generates the Join Request Multi Queued alarm when a LAN Emulation entity resends a Join request with the same Transaction ID as the original Join request. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Join Request Received from LES The Expert generates the Join Request Received from LES alarm when a LEC receives a Join Request frame from the LES. This is the opposite of standard operating procedure in an ELAN. Join Request frames are ordinarily sent by the LEC to the LES to request that they be allowed to join the ELAN. The LES then responds to the LEC by sending the Join Response frame. Join Request Transmitted by LES The Expert generates the Join Request Transmitted by LES alarm when a LES sends a Join Request frame to a LEC. This is the opposite of standard operating procedure in an ELAN. Join Request frames are ordinarily sent by the LEC to the LES to request that they be allowed to join the ELAN. The LES then responds to the LEC by sending the Join Response frame. 116 Sniffer Technologies Expert Alarms and Thresholds Join Response Issued by LEC The Expert generates the Join Response Issued by LEC alarm when a LEC sends a Join Response frame to the LES. This is the opposite of standard operating procedure in an ELAN. Join Response frames are ordinarily sent by the LES to the LEC in response to receiving the Join Request frame. Keep Alive Lifetime Set to Zero The Expert generates the Keep Alive Lifetime Set to Zero alarm when the value of the Keep Alive Lifetime field in the Keep Alive Lifetime Extension contained in an MPOA Keep Alive control message is zero. MPSs periodically send MPOA Keep Alive control messages to MPCs to keep the MPC-MPS flow active. The MPOA Keep Alive control message includes the Keep Alive Lifetime extension. The Keep Alive Lifetime extension includes a Keep Alive Lifetime field that indicates the amount of time that its Keep Alive message can be considered valid. When this field reaches zero, the keep alive message expires before it can be acknowledged and the Expert generates this alarm. Keep Alive Sequence Failure The Expert generates the Keep Alive Sequence Failure alarm when an MPC receives a Keep Alive frame from an MPS with a request ID that did not increase from its previous value. MPSs periodically send MPOA Keep Alive control messages to MPCs to keep the MPC-MPS flow active. Each time an MPS sends a Keep Alive message to a given station, it increases the request ID found in the message by at least one (the first request ID sent to a given station is set to zero and increases from there). Keep Alive Lifetime Ext Not Found in Frame The Expert generates the Keep Alive Lifetime Ext Not Found in Frame alarm when it detects an MPOA control message that is required to include the Keep Alive Lifetime Extension but does not have it. MPOA Keep Alive messages are required to include the Keep Alive Lifetime EXT. LANE Control Request Response Timeout The Expert generates the LANE Control Request Response Timeout alarm when a LEC sends a Control Request frame to the LE Service, but does not receive a corresponding response frame before the timeout value expires. The Expert uses timeout values as follows: Expert Alarms Reference Guide 117 Chapter 1 The Expert uses the timeout value specified by the LEC's C7 Control Time-out initial state parameter if it has been able to learn it from network traffic. The C7 Control Time-out is sent by the LEC as part of the Configuration or Join phase of LAN Emulation communications. The ATM Forum specifies a minimum value of 10 seconds for this timeout, with a default value of 30 seconds and a maximum of 300 seconds. If the Expert has not been able to learn the C7 Control Time-out parameter, it uses the Control Timeout C7 threshold in the Alarms tab of the Expert UI Object Properties dialog box. By default, NAI sets this threshold to 30 seconds (the same default as the ATM Forum). The exact type of Control Request frame resent could have been a Join Request, Register Request, Register Request, ARP Request, Flush Request, or a Verify Request. The alarm display will indicate the exact request type(s). Possible Cause: The network is experiencing congestion. The LE Service has received a large number of simultaneous requests from various LECs. LANE Control Request Retry Frame Mismatch The Expert generates the LANE Control Request Retry Frame Mismatch alarm when a LEC resends a Control Request frame in which a field was changed other than the Transaction ID. This is a violation of the LAN Emulation standard. The exact type of Control Request frame could be a Join Request, Register Request, Unregister Request, ARP Request, or a Verify Request. Whenever a LEC needs to resend a Control Request frame (perhaps because there was no response from the LE Service), it must send the exact same original request frame with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. LANE Control Request Retry Rate > 1 Frame/sec The Expert generates the LANE Control Request Retry Rate Exceeds 1 Frame/sec alarm when it detects a LEC resending the same Control Request frame more than once per second. The exact type of Control Request frame resent could have been a Join Request, Register Request, Register Request, ARP Request, Flush Request, or a Verify Request. 118 Sniffer Technologies Expert Alarms and Thresholds LANE Version Unsupported by ELAN The Expert generates the LANE Version Unsupported by ELAN alarm when a requesting station sends a control frame to the ELAN with a VERSION field indicating a higher value than that supported by the ELAN. Possible control frame types sent by the requesting station include Configure Request, Join Request, Register Request, Unregister Request, ARP Request, Flush Request, and Verify Request. Possible Cause: A Version 2.0 LANE client is attempting to operate on a Version 1.0 ELAN. LE ARP Request Rate Exceeds 1 Frame/Sec The Expert generates the LE ARP Request Rate Exceeds 1 Frame/Sec alarm when it detects a LEC sending LE_ARP requests more than once per second. LE_CONFIGURE Params Conflict, Service Refused A LEC sends a Configure Request to the LECS in order to receive the address of the LES for an ELAN meeting its requested operating parameters. The Expert generates the LE_CONFIGURE Params Conflict, Service Refused alarm when the LECS responds with a Configure Response frame with the Status field set to 21 (decimal), indicating a conflict in the parameters specified in the Configure Request frame. A status field set to 21 can also indicate that the LECS has denied a request for unspecified reasons. LEC Issued Config Response to LECS The Expert generates the LEC Issued Config Response to LECS alarm when a LEC sends a Configure Response frame to a LECS. This the opposite of standard operating procedure in an ELAN. Configure Response frames are ordinarily sent by the LECS to the LEC in response to receiving the Configure Request frame. They indicate whether the candidate LEC will be allowed to attempt the Join phase with a LES. LEC Not Recognized by LECS During the Configuration phase of LAN Emulation communications, a LEC sends a Configure Request frame to the LECS in order to obtain configuration information, including the address of a LES for an ELAN that meets the LEC's needs. In response, the LECS sends the Configure Response frame indicating whether the candidate LEC will be allowed to attempt the Join phase with a LES. Expert Alarms Reference Guide 119 Chapter 1 The Expert generates the LEC Not Recognized by LECS alarm when the Configure Response frame sent from the LECS to the LEC indicates failure (that is, the LEC will not be allowed to attempt the Join phase) with a Status code of 20 (decimal). This status code indicates that the LEC is not recognized by the LECS. Possible causes: There is a version incompatibility between the LEC and the LECS. The LECS denied the Configure Request for security reasons. The LEC supplied conflicting parameters in the Configure Request frame. Look for instances of alarms related to this failure – for example, Configuration Fail, Insufficient Resources, Denied Access to ELAN, Insufficient Info from LEC, LANE Protocol Version unsupported by ELAN, and so on. The data provided for these alarms will help you understand the reasons for the failure. LEC ID in Frame Mismatch with Client LEC ID The Expert generates the LEC ID in Frame Mismatch with Client LEC ID alarm when a LEC sends a Register, Unregister, or ARP request with the LECID field set to something other than the LEC's ID. LECID Is Not 0x0000 in Request Frame The Expert generates the LECID Is Not 0x0000 in Request Frame alarm when a LEC sends a Register Request or a Join Request to the LE Service with the LECID field set to something other than zero. LAN Emulation requires that Register and Join requests be sent with the LECID field set to zero. This is because one of the purposes of the Register and Join phases of LAN Emulation communications is for the LEC to get its LECID from the LES. LECS Issued Config Request to LES The Expert generates the LECS Issued Config Request to LES alarm when a LECS sends a Configure Request frame to a LEC. This the opposite of standard operating procedure in an ELAN. Configure Request frames are ordinarily sent by the LEC to the LECS in order to obtain configuration information, including the address of a LES for an ELAN that meets the LEC's needs. In response, the LECS sends the Configure Response frame indicating whether the candidate LEC will be allowed to attempt the Join phase with a LES. 120 Sniffer Technologies Expert Alarms and Thresholds Local Segment ID not in Source Routed Frame The Expert generates the Local Segment ID not in Source Routed Frame alarm when it sees a source routed frame with a route indicator (RI) field that does not contain a route descriptor (RD) matching the local ELAN's segment ID. Source routing is typically used in token ring (802.5) networks. Source routing use the RI field to identify each of the bridges that have forwarded a frame. As each bridge retransmits a frame, it appends its own RD to the RI field of the frame. As a result, the ultimate recipient has a record of all intermediate stations that forwarded the frame. Since LECs operate as bridges, they add their own RDs to the RI field. Frames seen on the network without an RD matching the local ELAN's segment ID are considered invalid. They may be discarded by a LEC. MAC Address bound to Multiple LECs The Expert generates the MAC Address bound to Multiple LECs alarm when it sees the same LAN MAC address bound to multiple LECs. During capture, the Expert examines the control request and response messages sent by LAN Emulation entities and keeps track of the pairings of MAC addresses with LECs it has been able to learn from these messages. Each time the Expert sees a control message with a MAC address/LEC address pairing, it compares the pairing against the pairings it has already stored in its buffer. The Expert generates this alarm when this comparison reveals a MAC address bound to multiple LECs. Max Transmission Unit Size Undefined The Expert generates the Max Transmission Unit Size Undefined alarm when the maximum transmission unit (MTU) field in the Client Information Element (CIE) of an MPOA Cache Imposition Request or Cache Imposition Reply control frame is not defined. This is a violation of the MPOA specification. Missing ULIA on Outside Link The Expert generates the Missing ULIA on Outside Link alarm when it detects a border node sending a PNNI Hello packet to a border node in another peer group without including the Uplink Information Attribute (ULIA) Information Group (IG), even though the sending station was in a Hello state that required the inclusion of the ULIA with the packet. Information Groups are a form of TLV used to encode PNNI information. Expert Alarms Reference Guide 121 Chapter 1 “Outside links” are links between PNNI nodes belonging to different peer groups (border nodes). PNNI nodes determine that they are in different peer groups from one another through the exchange of PNNI Hello packets. The ULIA provides complete information to the receiving station on the resources it must advertise in the uplink advertisement that it sends in response. MPOA Capable LEC Did Not Issue Device TLV The Expert generates the MPOA Capable LEC Did Not Issue Device TLV alarm when it sees a LEC that has been configured as an MPOA device (either an MPC or an MPS) send one of the following message types without the Device Type TLV: LE_REGISTER_REQUEST LE_REGISTER_RESPONSE LE_ARP_REQUEST LE_ARP_RESPONSE Tragedies LE_ARP_REQUEST MPOA uses the Device Type TLV to enable MPOA entities (MPCs and MPSs) to discover one another. Its transmission is required with the LANE control messages listed above. MPOA Config Parameter Invalid The Expert generates the MPOA Config Parameter Invalid alarm when it detects a configuration parameter that is outside of the valid ranges found in the MPOA specification. The Expert learns this information from the frames exchanged between an MPOA device and the LECS to obtain configuration information for the MPOA device. The valid ranges for MPC and MPS configuration parameters are found in sections 4.1.1.1 and 4.1.2.1 of the MPOA specification. MPOA Control Request Response Timeout The Expert generates the MPOA Control Request Response Timeout alarm when An MPOA entity (either an MPC or an MPS) sends a Control Request frame, but does not receive a corresponding response frame before the timeout value expires. The Expert uses timeout values as follows: 122 Sniffer Technologies Expert Alarms and Thresholds The Expert uses the timeout values specified by the sending entity's MPC Initial Retry Time (MPC-p4) and MPC Retry Time Maximum (MPC-p5) initial state parameters if it has been able to learn them from network traffic. The ATM Forum specifies a minimum value of 1 second, a default value of 5 seconds, and a maximum of 300 seconds for the MPC Initial Retry Time (MPC-p4) parameter. The values for the MPC Retry Time Maximum (MPC-p5) parameter are 10 (minimum), 40 (default), and 300 (maximum). If the Expert has not been able to learn the initial state parameters, it uses the MPC Initial Retry Time-p4 threshold and the MPC Retry Time Maximum-p5 threshold in the Alarms tab of the Expert UI Object Properties dialog box. By default, NAI sets these thresholds to 5 seconds and 40 seconds, respectively (the same defaults as the ATM Forum). How MPOA Times Out Requests Using These Values MPOA uses both the Initial Retry Time and the Retry Time Maximum values, in addition to the Retry Time Multiplier (MPS-c1) to time out requests. The Retry Time Multiplier is a constant and has a value of 2. When an entity sends its first request, it must wait MPC-p4 seconds (the initial retry time) before resending the request. Each time the entity resends the request, the sending entity multiplies the prior retry time by the Retry Time Multiplier and waits that long for a response. When the retry time exceeds the Retry Time Maximum (MPC-p5), the request is considered to have timed out. For example, suppose an MPC has an Initial Retry Time of 5 seconds and a Retry Time Maximum of 39 seconds. On the first sending of a control message, the MPC will wait five seconds for a response. If it doesn't receive one, it multiplies the prior retry time (5 seconds) by the multiplier (2) and waits that long (10 seconds) for a response to its next attempt at sending the message. If it doesn't receive a response, it does the same thing again, multiplying 10 (the current retry timer) by 2 and waiting that long for the response to its third message. When the retry timer multiplied by 2 finally exceeds the Retry Time Maximum, it times out the message. In this example, that will happen after the third attempt when the retry timer reaches 40 (one more than the Retry Time Maximum). Possible Cause: The network is experiencing congestion. MPOA Control Request Retry Rate > 1 Frame/sec The Expert generates the MPOA Control Request Retry Rate > 1 Frame/Sec alarm when it detects an MPOA entity (either an MPC or an MPS) resending the same Control Request message more than once per second. Expert Alarms Reference Guide 123 Chapter 1 The exact type of Control Request frame resent could have been Resolution Request, Cache Imposition Request, Egress Cache Purge Request, or an NHRP Purge Request. MPOA Error Frame Detected The Expert generates the MPOA Error Frame Detected alarm when it detects an MPOA control frame with the ar$op.type fixed header field set to 0x88. A value of 0x88 in the ar$op.type fixed header field indicates an error. Multiply Assigned Secondary LEC Address The Expert generates the Multiply Assigned Secondary LEC Address alarm when it sees more than one LEC using the same secondary ATM address. A LEC can have multiple ATM addresses in addition to its primary ATM address. The primary ATM address is used for connections to the LES and BUS. It is used for all exchanges during the Configuration and Join phases of LANE communications. However, a LEC can also have secondary ATM addresses that can be used for Data Direct VCCs. ATM addresses are reported in the C1n variable for the LEC. The first address in the C1n variable is the primary ATM address. Following addresses are considered secondary. During capture, the Expert examines the control request and response messages sent by LAN Emulation entities and keeps track of the secondary LEC addresses used by each LEC. Each time the Expert sees a control message with a secondary LEC address, it compares the address against the addresses used by other LECs it has already stored in its buffer. The Expert generates this alarm when this comparison reveals a secondary LEC address used by multiple LECs NARP Request Multi Queued The Expert generates the NARP Request Multi Queued alarm when a LAN Emulation entity resends a NARP request with the same Transaction ID as the original Join request. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). 124 Sniffer Technologies Expert Alarms and Thresholds Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. NARP Request Received from LES The Expert generates the NARP Request Received from LES alarm when a LES sends a NARP Request frame. This is the opposite of standard operating procedure in an ELAN. NARP Request frames are ordinarily sent by the LEC to inform the LE Service of changes in the bindings of remote addresses. No CIEs Were Present in MPOA Reply The Expert generates the No CIEs Were Present in MPOA Reply alarm when it detects an MPOA response frame that is missing a mandatory Client Information Element (CIE). The exact MPOA reply can be any one of the following MPOA reply control frame types: MPOA Resolution Reply MPOA Cache Imposition Reply MPOA Egress Cache Purge Reply No CIEs Were Present in MPOA Request The Expert generates the No CIEs Were Present in MPOA Request alarm when it detects an MPOA request frame that is missing a mandatory Client Information Element (CIE). The exact MPOA request can be any one of the following MPOA request control frame types: MPOA Cache Imposition Request MPOA Egress Cache Purge Request NHRP Purge Request No Reply Flag Is Not Set in ICache Purge The Expert generates the No Reply Flag Is Not Set in ICache Purge alarm when it detects an NHRP Ingress Cache Purge Request control message with the No Reply flag in the common header portion of the message set to something other than one. The MPOA specification requires that the No Reply flag in an NHRP Ingress Cache Purge Request control message must be set to one. Expert Alarms Reference Guide 125 Chapter 1 PTSE Violation The Expert generates the PTSE Violation alarm if it detects a PNNI Topology State Element (PTSE) with an illegal Information Group (IG). PNNI nodes exchange PNNI Topology State Packets (PTSPs) amongst themselves to distribute routing information throughout a peer group. Each PTSP packet contains multiple PNNI Topology State Elements (PTSEs), each with its own header. The PTSE header includes a PTSEType field indicating the restricted IGs which can legally appear in the PTSE (IGs are TLV used to encode PNNI information). The Expert generates this alarm if a restricted IG other than those indicated in the PTSEType field is found in a PTSE. Register Request Multi Queued The Expert generates the Register Request Multi Queued alarm when aLAN Emulation entity resends a Register request with the same Transaction ID as the original Register request. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Register Request Received from LES The Expert generates the Register Request Received from LES alarm when a LES sends a Register Request frame to a LEC. This is the opposite of standard operating procedure in an ELAN. Register Request frames are ordinarily sent by the LEC to the LES in order to register a LAN destination\ATM address pair. The LES responds by sending the Register Response frame. Register Response Issued by LEC The Expert generates the Register Response Issued by LEC alarm when a LEC sends a Register Response frame to a LES. This is the opposite of standard operating procedure in an ELAN. Register Response frames are ordinarily sent by the LES to the LEC in response to receiving the Register Request frame. 126 Sniffer Technologies Expert Alarms and Thresholds Remote Address Issued by Non Proxy LEC The Expert generates the Remote Address Issued by Non Proxy LEC alarm when it sees a LEC that is known not to be a proxy sending an ARP Response with a flag indicating that the ATM address represents a remote (proxied) MAC address. The Expert uses the C4 Proxy variable sent in Configure Request messages to learn whether a given station is a proxy. When the Expert sees a Configure Request message with the C4 Proxy variable set to false, it records the sending station as a non-proxy station. If a non-proxy station subsequently sends an ARP Response with a flag indicating that the ATM address represents a remote (proxied) MAC address, the Expert generates this alarm. Request Params Are Incompatible with ELAN The Expert generates the Request Parameters Are Incompatible with ELAN alarm when the LE Service sends a control response frame with the Status field set to 2 (decimal), indicating that the parameters in a control request frame sent by a LEC were incompatible with the parameters supported by the ELAN. Possible causes: A LEC sent a control request frame with the REQUESTER-LED-ID field or the SOURCE-ATM-ADDRESS field empty. A LEC sent a Join Request to the LES that was missing some of the required parameters (for example, C1 LE Client's Non-Multiplexed ATM Addresses, C2 LAN Type, C3 Maximum Data Frame Size, C4 Proxy, or C5 ELAN Name). A LES sent a Join Response to a LEC that was missing some of the required parameters (for example, C2 LAN Type, C3 Maximum Data Frame Size, or C5 ELAN Name). An 802.3 data frame was seen on an 802.5 data direct VCC. Source routed frames were seen with an invalid RI (route indicator) field. An RI field is invalid if it is of an odd length or has a Route Descriptor that does not match the ELAN's Segment ID TA List Resources Exhausted This is an internal alarm generated by the Expert when there is no longer enough buffer space to store an incoming Control Frame. Expert Alarms Reference Guide 127 Chapter 1 Possible cause: The Expert is running out of memory. You may want to save and then restart your capture. Topology Change Request Multi Queued The Expert generates the Topology Change Request Multi Queued alarm when LAN Emulation entity resends a Topology request with the same Transaction ID as the original Topology request. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Topology Mismatch in Data Frame Types The Expert generates the Topology Mismatch in Data Frame Types alarm when it detects either of the following: An 802.3 frame on a data direct VCC in an 802.5 ELAN. An 802.5 frame on a data direct VCC in an 802.3 ELAN. Unexpected DLL Header EXT in Frame Each MPOA control message has a set of extensions that can legally appear in a message. The Expert generates the Unexpected DLL Header EXT in Frame alarm when it sees the DLL Header extension in an MPOA control message for which it is not legal. Unexpected Egress Cache EXT in Frame Each MPOA control message has a set of extensions that can legally appear in a message. The Expert generates the Unexpected Egress Cache EXT in Frame alarm when it sees the Egress Cache Tag extension in an MPOA control message for which it is not legal. 128 Sniffer Technologies Expert Alarms and Thresholds Unexpected Hop Count EXT in Frame Each MPOA control message has a set of extensions that can legally appear in a message. The Expert generates the Unexpected Hop Count EXT in Frame alarm when it sees the Hop Count extension in an MPOA control message for which it is not legal. Unexpected IG The Expert generates the Unexpected IG alarm if it detects a PNNI frame containing a legal but unexpected IG. PNNI routing frames use nested TLVs known as Information Groups (IG) to encode PNNI information. Different IGs are expected in different PNNI packet types. If an IG appears in a packet for which it is non-standard, the Expert generates this alarm. For example, a PNNI PTSE ACK packet can legally contain the Nodal PTSE ACK IG or the System capabilities IG. If a PTSE ACK packet is detected with an IG other than these two (for example, the Nodal Hierarchy List IG), the Expert would generate the Unexpected IG alarm. The mapping of valid IGs to PNNI packet types is found in Table 5-19 of Version 1.0 of the ATM Forum's PNNI specification. Unexpected Keepalive Lifetime EXT in Frame Each MPOA control message has a set of extensions that can legally appear in a message. The Expert generates the Unexpected Keepalive Lifetime EXT in Frame alarm when it sees the Keep Alive Lifetime extension in an MPOA control message for which it is not legal. Unexpected Original Error Code EXT in Frame Each MPOA control message has a set of extensions that can legally appear in a message. The Expert generates the Unexpected Original Error Code EXT in Frame alarm when it sees the Original Error Code extension in an MPOA control message for which it is not legal. Unexpected Service Category EXT in Frame Each MPOA control message has a set of extensions that can legally appear in a message. The Expert generates the Unexpected Service Category EXT in Frame alarm when it sees the ATM Service Category extension in an MPOA control message for which it is not legal. Expert Alarms Reference Guide 129 Chapter 1 Unexpected TLVs Detected The Expert generates the Unexpected TLVs Detected alarm when it sees a TLV it does not expect in a frame. Each LAN Emulation control frame type supports a different subset of the total set of LANE TLVs. The Expert compares the TLVs seen in each detected control frame against the TLVs that are actually supported by that frame type. When an unsupported TLV is detected, the Expert generates this alarm. An example is provided in the Possible Cause section, below. Possible cause: A Configuration Request was seen with the LLC-Muxed-ATM-Address TLV. This TLV is not supported for Configuration Requests. Unknown Extension in Frame Any MPOA control frame can use one of a set of valid MPOA Extensions. These MPOA Extensions range in value from 0x1000 to 0x1006. The Expert generates the Unknown Extension in Frame alarm when it detects the presence of an extension in a control frame whose value does not fall within this range. Unknown IG The Expert generates the Unknown IG alarm if it detects a PNNI frame containing an unknown IG. PNNI routing frames use nested TLVs known as Information Groups (IG) to encode PNNI information. Different versions of the PNNI protocol support different sets of IGs. This alarm is likely to be generated when a PNNI node receives an IG from a second node running a newer version of the PNNI protocol. Possible cause: Two PNNI nodes are running different versions of the PNNI protocol. Unknown LANE Config/Control Opcode The Expert generates the Unknown LANE Config/Control Opcode alarm when it detects the presence of an opcode that is unknown or unassigned by the ATM Forum. 130 Sniffer Technologies Expert Alarms and Thresholds Possible cause: Some manufacturers of ATM networking equipment have implemented their own proprietary opcodes Unknown Marker/Protocol Values The Expert generates the Unknown Marker/Protocol Values alarm when it detects invalid values in a frame. Possible causes: An unused field in a control frame is set to an invalid value. Backward compatibility with the LAN Emulation Version 1 specification requires that unused fields in control messages be set to a known value. The OP-CODE identifying a control frame is not valid for the type of flow on which the frame is seen. Unknown/Unexpected Authenticate EXT in Frame Each MPOA control message has a set of extensions that can legally appear in a message. The Expert generates the Unknown/Unexpected Authenticate EXT in Frame alarm when it sees the Authentication extension in an MPOA control message for which it is not legal. Unknown/Unexpected CIE in Frame Each type of MPOA control frame has a set of Client Information Elements that can legally be found within it. The Expert generates the Unknown/Unexpected CIE in Frame alarm when it detects the presence of one or more CIEs that are not valid for the detected control frame type. Unregister Request Multi Queued The Expert generates the Unregister Request Multi Queued alarm when a LAN Emulation entity resends an Unregister request with the same Transaction ID as the original Unregister request. This is a violation of the LAN Emulation standard. LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Expert Alarms Reference Guide 131 Chapter 1 Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Unregister Request Received from LES The Expert generates the Unregister Request Received from LES alarm when a LES sends an Unregister Request frame to a LEC. This is the opposite of standard operating procedure in an ELAN. Unregister Request frames are ordinarily sent by the LEC to the LES in order to remove a registered LAN destination\ATM address pair from the LES' cache of registered addresses. The LES responds by sending the Unregister Response frame. Unregister Response Issued by LEC The Expert generates the Unregister Response Issued by LEC alarm when a LEC sends an Unregister Response frame to a LES. This is the opposite of standard operating procedure in an ELAN. Unregister Response frames are ordinarily sent by the LES to the LEC in response to receiving the Unregister Request frame. V2 TLVs Issued in V1 ELAN The Expert generates the V2 TLVs Issued in V1 ELAN alarm when a control frame with TLVs supported only in LANE Version 2.0 is seen on a Version 1.0 ELAN. Possible cause: A Version 2.0 LEC sent a Configure Request frame including the Config-Frag-Info TLV. This TLV (used so a LECS can use multiple frames to return TLV information that will not fit into a single frame) is supported only in LANE Version 2.0. Verify Request Multi Queued The Expert generates the Verify Request Multi Queued alarm when a LAN Emulation entity resends a Verify request with the same Transaction ID as the original ARP request. This is a violation of the LAN Emulation standard. 132 Sniffer Technologies Expert Alarms and Thresholds LAN Emulation communications are established and managed using a set of standard request and response control frames (such as Configure, Join, Register, Unregister, ARP, Flush, and Verify, among others). Each of these requests has a Transaction ID field. When the station receiving the request returns its response, it preserves this Transaction ID. This way, the requesting station can tell which responses correspond to which of its requests (by matching the Transaction IDs). Whenever a LAN Emulation entity needs to resend a request (perhaps because there was no response), it must send the exact same original request with the exception of the Transaction ID. The Transaction ID must be different so that if the responding station does respond, the requesting station can determine to which request it is responding. Verify Request Received from LES The Expert generates the Verify Request Received from LES alarm when a LES sends a Verify Request frame to a LEC. This is the opposite of standard operating procedure in an ELAN. Verify Request frames are ordinarily sent by the LEC to the LES in order to verify the ATM address of the BUS for the ELAN of which the LEC is a member. The LES responds by sending the Verify Response frame. Verify Response Issued by LEC The Expert generates the Verify Response Issued by LEC alarm when a LEC sends a Verify Response frame to a LES. This is the opposite of standard operating procedure in an ELAN. Verify Response frames are ordinarily sent by the LES to the LEC in response to receiving the Verify Request frame. Expert Alarms Reference Guide 133 Chapter 1 ATM Connection Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the ATM Connection layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). AAL5 CRC Errors The Expert generates the AAL5 CRC Errors alarm when the percentage of frames on a VPI.VCI with CRC errors exceeds the Max Percentage of Frames with CRC Errors threshold in the Alarms tab of the Expert UI Object Properties dialog box. For each VPI.VCI on the network, the Expert maintains counts of frames with CRC errors and frames without CRC errors. When the percentage of frames with CRC errors first exceeds the threshold, the Expert generates this alarm. Each time the percentage of frames with CRC errors on the VPI.VCI re-exceeds the threshold, the Expert either reopens the previous alarm or creates a new one depending on how much time has elapsed between alarms. If the previous alarm occurred less than an hour ago, the Expert reopens the previous alarm with updated frame counts. If the previous alarm occurred more than an hour ago, the Expert creates a new alarm. AAL5 Length Errors The Expert generates the AAL5 Length Errors alarm when the percentage of frames on a VPI.VCI with length errors exceeds the Max Percentage of Frames with Length Errors threshold in the Alarms tab of the Expert UI Object Properties dialog box. For each VPI.VCI on the network, the Expert maintains counts of frames with length errors and frames without length errors. When the percentage of frames with length errors first exceeds the threshold, the Expert generates this alarm. Each time the percentage of frames with length errors on the VPI.VCI re-exceeds the threshold, the Expert either reopens the previous alarm or creates a new one depending on how much time has elapsed between alarms. 134 If the previous alarm occurred less than an hour ago, the Expert reopens the previous alarm with updated frame counts. If the previous alarm occurred more than an hour ago, the Expert creates a new alarm. Sniffer Technologies Expert Alarms and Thresholds AAL5 Timeout Errors The Expert generates the AAL5 Timeout Errors alarm when the percentage of frames on a VPI.VCI with timeout errors exceeds the Max Percentage of Frames with Timeout Errors threshold in the Alarms tab of the Expert UI Object Properties dialog box. For each VPI.VCI on the network, the Expert maintains counts of frames with timeout errors and frames without timeout errors. When the percentage of frames with timeout errors first exceeds the threshold, the Expert generates this alarm. Each time the percentage of frames with timeout errors on the VPI.VCI re-exceeds the threshold, the Expert either reopens the previous alarm or creates a new one depending on how much time has elapsed between alarms. If the previous alarm occurred less than an hour ago, the Expert reopens the previous alarm with updated frame counts. If the previous alarm occurred more than an hour ago, the Expert creates a new alarm. Abnormal Disconnect for SVC The Expert generates the Abnormal Disconnect for SVC alarm when it detects an SVC that was disconnected for one of the following reasons: If the Specific Disc Cause for Alarm threshold in the Alarms tab of the Expert UI Object Properties dialog box is set to 0, the Expert will generate this alarm for any disconnect cause code other than 16 or 31 (normal disconnects). Alternatively, the Specific Disc Cause for Alarm threshold can be used to specify a disconnection cause code whose detection will cause this alarm to be generated. If the threshold is set to a positive integer, this alarm will be generated only if an SVC is disconnected with the specified cause code. The alarm display shows you the exact disconnect cause code that caused the alarm. Extra Cause Codes Used by the Expert In addition to the standard disconnect cause codes, you can also use this additional Expert-specific cause code for the Specific Disc Cause for Alarm (conn level) threshold: 127 – A SETUP message using the same call reference number as this connection was detected without the analyzer having seen an intervening RELEASE message (or any other disconnection message) to drop the prior connection. Expert Alarms Reference Guide 135 Chapter 1 CLP = 1 The Expert generates the CLP = 1 alarm when the percentage of frames on a VPI.VCI with at least one cell's Cell Loss Priority bit set to true (CLP=1) exceeds the Max Percentage of Frames with CLP=1 threshold in the Alarms tab of the Expert UI Object Properties dialog box. For each VPI.VCI on the network, the Expert maintains counts of frames containing at least one cell with the CLP bit set to true and frames without any cells with the CLP bit set to true. When the percentage of CLP=1 frames first exceeds the threshold, the Expert generates this alarm. Each time the percentage of CLP=1 frames on the VPI.VCI re-exceeds the threshold, the Expert either reopens the previous alarm or creates a new one depending on how much time has elapsed between alarms. If the previous alarm occurred less than an hour ago, the Expert reopens the previous alarm with updated frame counts. If the previous alarm occurred more than an hour ago, the Expert creates a new alarm. Connection Too Short The Expert generates the Connection Too Short alarm if it detects an SVC whose duration was shorter than the Min Duration of session in msecs threshold in the Alarms tab of the Expert UI Object Properties dialog box. The analyzer detects short connections using the Q.2931 messages used on the signaling channel to set up and tear down connections. Crankback Has Occurred (ATM Connection Layer) The Expert generates the Crankback Has Occurred alarm when the number of hops on which a PNNI connection has failed exceeds the Max number of crank backs threshold in the Alarms tab of the Expert UI Object Properties dialog box. The PNNI system reacts to a failed hop by rerouting (“cranking back”) the connection via a different path from the Peer Group preceding the failure. The Crankback Has Occurred alarm is call-specific at the ATM Connection layer – the crankback occurred due to a fault on one of the links or nodes on which the call has taken place. PNNI is used to set up multi-hop connections (calls) between two ATM hosts connected to switches which are members of the same PNNI domain. PNNI domains consist of multiple peer groups – groups of switches that exchange detailed routing database information with one another through the exchange of PNNI Hello, Database Summary, and PTSE Request frames. Each peer 136 Sniffer Technologies Expert Alarms and Thresholds group in the PNNI domain receives a summary of the detailed information exchanged within the other peer groups. In this way, each switch within a peer group can set up and route calls destined for a different peer group without needing to know the detailed information from that peer group – the routing within a destination peer group is taken care of by that peer group. Because of this, a call spanning several peer groups can fail at any peer group border along the way. The crankback mechanism is used to reroute calls around a failed node or link and is indicated in Q.2931 Release, Release Complete, or Drop Party frames. Crankback occurs gradually – one link at a time. In the worst case, it is possible for a call to “crank back” all the way to the originating node for the call. EFCI The Expert generates the EFCI alarm when the percentage of frames on a VPI.VCI with congestion (EFCI) indicated in at least one cell exceeds the Max Percentage of Frames with EFCI threshold in the Alarms tab of the Expert UI Object Properties dialog box. Congestion is indicated in the three bit payload type identifier (PTI) in cell headers. For each VPI.VCI on the network, the Expert maintains counts of frames with EFCI indicated and frames without EFCI indicated. When the percentage of EFCI frames first exceeds the threshold, the Expert generates this alarm. Each time the percentage of EFCI frames on the VPI.VCI re-exceeds the threshold, the Expert either reopens the previous alarm or creates a new one depending on how much time has elapsed between alarms. If the previous alarm occurred less than an hour ago, the Expert reopens the previous alarm with updated frame counts. If the previous alarm occurred more than an hour ago, the Expert creates a new alarm. Excessive Signaling Recovery The Expert generates the Excessive Signaling Recovery alarm if the number of one of the following signaling messages seen per second on the signaling channel(s) exceeds the corresponding threshold in the Alarms tab of the Expert UI Object Properties dialog box. Max Q2931 STATs/sec – Alarm is generated if the number of Q.2931 STATUS and STATUS ENQUIRY messages per second exceeds this threshold. Max SSCOP Begin/sec – Alarm is generated if the number of SSCOP Begin messages per second exceeds this threshold. Max SSCOP Err/sec – Alarm is generated if the number of SSCOP Error messages per second exceeds this threshold. Expert Alarms Reference Guide 137 Chapter 1 Max SSCOP RS/sec – Alarm is generated if the number of SSCOP Resynch messages per second exceeds this threshold. Max SSCOP retrns/sec. – Alarm is generated if the number of SSCOP retransmissions per second exceeds this threshold. Retransmissions are counted when a receiving SSCOP entity sends a “long” status message including list elements indicating that it did not receive all messages from a sending station. When the sending station receives this message, it is supposed to retransmit the unreceived messages. The alarm display shows you the number of each type of message detected. Missing DTL or Excessive DTL Size The Expert generates the Missing DTL or Excessive DTL Size alarm if a PNNI Call Setup message is detected which violates the PNNI standards. Examples of this include: Missing DTL The DTL exceeds the maximum legal length of 20 node/port pairs. The DTL is the Designated Transit List. It provides a complete path across a PNNI peer group (a group of PNNI-enabled switches which have exchanged detailed routing information) in the form of an ordered list of node/port pairs across the peer group. Its presence is required in a PNNI Call Setup message. MPCMPS Flow Not Allowed on this VC The Expert generates the MPCMPS Flow Not Allowed on this VC alarm when it detects an attempt to attach an MPC-MPS flow to a VCC which already has an MPC-MPC flow on it (or vice-versa). It is illegal to multiplex both an MPC-MPC flow and an MPC-MPS flow onto the same VCC. Possible 3.0/3.1 UNI Mismatch The Expert generates the Possible 3.0/3.1 UNI Mismatch alarm when it detects a possible mismatch in signaling versions being used on the UNI. Version 3.0 of the UNI specification does not provide explicit support for the use of the N(SQ) field in the SSCOP BEGIN message, whereas Version 3.1 does. When SSCOP BEGIN messages are seen both using and not using this field, the Expert generates this alarm. Signaling Frame Sequence Problem The Expert generates the Signaling Frame Sequence alarm in the following situations: 138 Sniffer Technologies Expert Alarms and Thresholds A sequence of signaling frames was detected that did not match the standard signaling rules (for example, a station receives a Call Setup message and responds with the Connect Acknowledge message instead of the Call Proceed or Call Connect message). A sequence of signaling messages was not completed before a timeout. The alarm display provides a suspected sequence of messages leading up to the alarm. Unexpected Data for SVC The Expert generates the Unexpected Data for SVC alarm when the number of unexpected data packets seen for an SVC exceeds the Number of unexpected data packets between alarms threshold in the ATM Connection section of the Alarms tab of the Expert UI Object Properties dialog box. The Expert considers a data packet to be unexpected when it is seen on an SVC before the requisite signaling information for the SVC has been exchanged. The analyzer counts these unexpected packets. When the total for a given SVC exceeds the corresponding threshold, it generates the Unexpected Data for SVC alarm. The Expert displays the protocol sequence leading up to the unexpected data packets so you can see what was missing. Consider the example shown in Figure 1-6: Figure 1-6. Sample Unexpected Data for SVC Alarm Display In this example, an unexpected data packet was seen on the 0.216 SVC. The Alarm Display shows the sequence of events that led up to the unexpected data. First, the Call Setup message was sent from the DCE side to initiate a call. The DTE side responded with the Call Proceed message, indicating that call establishment had been initiated. Then, the DTE sent the Call Connect message, indicating that the call from the DCE had been accepted. Expert Alarms Reference Guide 139 Chapter 1 Unfortunately, as the display shows, the DCE side went ahead and sent data frames without first sending the Connect Acknowledge frame. The analyzer considered this data to be unexpected since the Connect Acknowledge message had not yet been sent. Accordingly, the alarm was generated. ATM Host Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the ATM Host layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Excessive Abnormal Disconnects The Expert generates the Excessive Abnormal Disconnects alarm when the number of connections disconnected abnormally for a host (either as a source or as a destination) exceeds the Number of Disconnections between Alarms (host level) threshold in the Alarms tab of the Expert UI Object Properties dialog box. A connection is considered to have been disconnected abnormally if it is disconnected with any disconnect cause code other than 16 or 31 (normal disconnects). Each time a host experiences an abnormal disconnect, the Expert increments a counter. When this counter exceeds the Number of Disconnections between Alarms (host level) threshold, the Expert generates the Excessive Abnormal Disconnects alarm for this host and resets the counter to 0 so that counting can begin anew. The Expert display shows you the disconnection that caused the alarm to be generated (that is, the disconnection that put the counter over the threshold). Excessive Abnormal Frame Sequences The Expert generates the Excessive Abnormal Frame Sequences alarm when the number of abnormal frame sequences for a given host (either as a source or as a destination) exceeds the Number of Protocol Problems between Alarms (host level) threshold in the ATM Host section of the Alarms tab of the Expert UI Object Properties dialog box. The Expert considers the following to be abnormal frame sequences: 140 A sequence of signaling frames was detected that did not match the standard signaling rules (for example, a station receives a Call Setup message and responds with the Connect Acknowledge message instead of the Call Proceed or Call Connect message). A sequence of signaling messages was not completed before a timeout. Sniffer Technologies Expert Alarms and Thresholds The analyzer counts these abnormal frame sequences. Each time the number of sequences for a given host exceeds the threshold, the Expert generates the Excessive Abnormal Frame Sequences alarm and resets the counter to zero. The alarm display lists each abnormal sequence of messages that caused the alarm to be generated. Excessive Short Connections The Expert generates the Excessive Short Connections alarm for any host (either as a source or as a destination) involved in a connection whose duration is shorter than the Min Duration of session in msecs (host level) threshold in the Alarms tab of the Expert UI Object Properties dialog box. The analyzer detects short connections using the Q.2931 messages used on the signaling channel to set up and tear down connections. The alarm display provides the name (if known) and ATM address of the source and destination hosts involved in the connection, as well as the duration of the offending connection. High Signaling Activity The Expert generates the High Signaling Activity alarm if the number of one of the following signaling messages seen per second for a particular host (either as a source or as a destination) exceeds the corresponding threshold in the Alarms tab of the Expert UI Object Properties dialog box. Max Setups/sec (host level) – Alarm is generated if the number of Q.2931 SETUP messages per second exceeds this threshold. Max Q2931 STATs/sec – Alarm is generated if the number of Q.2931 STATUS messages per second exceeds this threshold. Max Q2931 STAT ENQs/sec – Alarm is generated if the number of Q.2931 STATUS ENQUIRY messages per second exceeds this threshold. The alarm display shows you the number (and rate) of each type of message detected. Host Address Cannot Be Mux’ed The Expert generates the Host Address Cannot Be Multiplexed (“Mux’ed”) alarm when it detects a LEC sharing an ATM address with some other LANE entity (either a LES, a BUS, a LECS, or another LEC). This is a violation of the LAN Emulation specification. The LES and the BUS can share the same ATM address (along with the LECS), but the LEC cannot. Expert Alarms Reference Guide 141 Chapter 1 Too Many Simultaneous Connections The Expert generates the Too Many Simultaneous Connections alarm when the number of simultaneous connections for a given host (either as a source or as a destination) exceeds the Max simultaneous sessions (host level) threshold in the Alarms tab of the Expert UI Object Properties dialog box. This alarm is a useful reminder when keeping tabs on stations that may potentially be overloaded. Unexpected Data for SVC (ATM Host Layer) The Expert generates the Unexpected Data for SVC alarm when the number of unexpected data packets for a given host (either as a source or as a destination) exceeds the Number of unexpected data packets between alarms threshold in the ATM Host section of the Alarms tab of the Expert UI Object Properties dialog box. The Expert considers a data packet to be unexpected when it is seen on an SVC before the requisite signaling information for the SVC has been exchanged. Each time the analyzer sees such a packet, it increments a separate counter for the source and the destination host on the connection. When the total for a given host exceeds the corresponding threshold, the analyzer generates the Unexpected Data for SVC alarm at the host level. It also resets the counter for the named host so that counting can begin anew. The Expert displays the protocol sequence leading up to the last unexpected data packet that caused the alarm to be generated (for example, if the threshold is set to 5, the sequence leading up to the fifth detected unexpected data packet is shown). Consider the example shown in Figure 1-7: Figure 1-7. Sample Unexpected Data for SVC Alarm Display (ATM Host Layer) In this example, an unexpected data packet was seen on an SVC set up between the named Source and Destination Hosts. The Alarm Display shows the sequence of events that led up to the unexpected data. 142 Sniffer Technologies First, the Call Setup message was sent from the DCE side to initiate a call. Expert Alarms and Thresholds The DTE side responded with the Call Proceed message, indicating that call establishment had been initiated. Then, the DTE sent the Call Connect message, indicating that the call from the DCE had been accepted. Unfortunately, as the display shows, the DCE side went ahead and sent data frames without first sending the Connect Acknowledge frame. The analyzer considered this data to be unexpected since the Connect Acknowledge message had not yet been sent. Accordingly, the alarm was generated. Wireless Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Wireless layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). ACK Frame Timeout The Expert generates the ACK Frame Timeout alarm when it does not see an acknowledgment to a unicast management or data frame within the time specified in the Duration field of the original management or data frame. When this happens, the sending station will resend the original frame and wait for another ACK. Unicast management and data frames include a Duration field indicating the amount of time within which a receiving station should return an ACK frame. The value of this field is typically equal to the amount of time required to send an ACK frame plus one short interframe space (SIFS). The Duration field lets other stations on the network know that during this period, the medium is reserved for the response to the frame. The Expert stores the value specified in the Duration field in a buffer. If it does not see the corresponding ACK to the frame (identified by matching sequence numbers) within the value specified by the Duration field, it generates this alarm. Association Failure The Expert generates the Association Failure alarm when it detects an 802.11 Association Response frame with a value other than zero in the Status Code field. A non-zero value in the Status Code field indicates that the access point sending the Association Response is denying the requested association. Expert Alarms Reference Guide 143 Chapter 1 To be a member of an infrastructure 802.11 wireless network, wireless stations must be associated with an access point. Wireless stations send Association Request frames to become associated with an access point. In turn, access points reply to Association Requests with Association Responses indicating the success or failure of the request. In this case, the access point denied the association request. The exact reason for the denial is found in the Status Code field of the Association Response. The Expert reports both the address of the access point denying the Association Request, as well as the reason for the denial indicated in the Status Code field, as follows: 1 — Unspecified failure. 10 — Cannot support all requested capabilities in the Capability Information field. 12 — Association denied due to reason outside the scope of the 802.11 standard. 17 — Association denied because the access point is unable to handle additional associated stations. 18 — Association denied due to requesting station not supporting all of the data rates in the BSSBasicRateSet parameter. Authentication Failure The Expert generates the Authentication Failure alarm when it detects an 802.11 Authentication frame with a value other than zero in the Status Code field. A non-zero value in the Status Code field indicates that the access point sending the Authentication frame is denying the requested authentication. Wireless stations exchange Authentication frames with access points to authenticate themselves with the network, thereby providing security and privacy. The authentication sequence for 802.11 networks consists of the exchange of either two authentication frames (for open system authentication) or four authentication frames (for shared key authentication), each identified by a transaction sequence number. The extra two authentication frames for shared key authentication are for the exchange of a string of challenge text, first sent in the clear by the access point and then returned in encrypted format by the wireless station. The Expert generates this alarm when the access point refuses to authenticate the requesting wireless station. The exact reason for the denial is found in the Status Code field of the Authentication frame. The Expert reports both the address of the access point denying the Authentication, as well as the reason for the denial indicated in the Status Code field, as follows: 144 1 — Unspecified failure. 13 — Responding station does not support the specified authentication algorithm. Sniffer Technologies Expert Alarms and Thresholds 14 — Received an Authentication frame with authentication transaction sequence number out of expected sequence. 15 — Authentication rejected because of challenge failure. 16 — Authentication rejected due to timeout waiting for next frame in sequence. CTS Frame Timeout The Expert generates the CTS Frame Timeout alarm when it does not see a clear to send (CTS) frame sent in response to a request to send (RTS) frame within the time specified in the Duration field of the original RTS frame. RTS frames include a Duration field indicating the amount of time within which a receiving station should return a CTS frame. The value of this field is typically equal to the amount of time required to send the CTS frame, one ACK frame, and three short interframe spaces (SIFS). The Duration field lets other stations on the network know that during this period, the medium is reserved. When the Expert sees an RTS frame, it stores the value specified in the Duration field in a buffer. If it does not see the corresponding CTS frame within the value specified by the Duration field, it generates this alarm. Deauthentication The Expert generates the Deauthentication alarm when it detects an 802.11 Deauthentication frame. Occasionally, wireless stations need to terminate secure communications with one another or with an access point. To do so, they send Deauthentication frames. Deauthentication frames are a part of normal 802.11 network operations. A relatively small number of these alarms is no cause for concern. However, a large number of Deauthentication frames may indicate a potential authentication denial attack on the wireless network. The alarm display includes the following information: The destination address of the Deauthentication frame (that is, the station with which the sending station want to terminate secure communications). The Reason Code indicating the reason the Deauthentication frame was sent. Possible values for the Reason Code field are as follows: 1 — Unspecified reason. 2 — Previous authentication no longer valid. 3 — Deauthenticated because sending station is leaving (or has left) the network. Expert Alarms Reference Guide 145 Chapter 1 6 — Class 2 frame received from nonauthenticated station. Disassociation The Expert generates the Disassociation alarm when it detects an 802.11 Disassociation frame. Wireless stations and access points send Disassociation frames to terminate associations with one another. For example, an access point may terminate an association with a station because it is unable to handle any more associations. Similarly, a wireless station may terminate an association if it is leaving the network. Disassociation frames are a part of normal 802.11 network operations. A relatively small number of these alarms is no cause for concern. However, a large number of Disassociation frames may indicate a potential denial of service attack on the wireless network. The alarm display includes the following information: The destination address of the Disassociation frame (that is, the station with which the sending station want to terminate its association). The Reason Code indicating the reason the Disassociation frame was sent. Possible values for the Reason Code field are as follows: 1 — Unspecified reason. 4 — Disassociated due to inactivity. 5 — Disassociated because the access point is unable to handle all currently associated stations. 7 — Class 3 frame received from nonassociated station. 8 — Disassociated because sending station is leaving (or has left) the network. 9 — Station requesting (re)association is not authenticated with responding station. Mcast/Bcast Fragmentation The Expert generates the Mcast/Bcast Fragmentation alarm when it detects an 802.11 frame with a multicast or broadcast destination address and fragmentation indicated in the MAC header. This is a violation of the 802.11 specification. Wireless networks commonly implement the fragmentation and defragmentation services provided by the 802.11 MAC layer to increase transmission reliability. However, the 802.11 specification does not allow fragmentation for broadcast or multicast frames because of the overhead this would cause for the network as a whole. 146 Sniffer Technologies Expert Alarms and Thresholds Missing Fragment Number The Expert generates the Missing Fragment Number alarm when it detects a jump in the fragment number of an 802.11 frame, indicating that a portion of a fragmented data unit is at least temporarily missing. Wireless networks commonly implement the fragmentation and defragmentation services provided by the 802.11 MAC layer to increase transmission reliability. When a unicast frame’s length exceeds an internal threshold in the MAC’s MIB, the MAC will break up the frame into smaller constituent frames — fragments — with the same sequence number. Each fragment of a larger data unit is identified with a fragment number indicating its intended ordered position within the reassembled data unit at the receiving station. The Expert observes each transmitted fragment and stores the fragment numbers. If it observes a jump in the fragment number for the transmission of fragments with the same sequence number, it generates this alarm. Possible cause: A relatively small number of these alarms is no cause for concern. 802.11 guarantees the sequential arrival of fragments at a receiving station, but occasionally fragments may be missing due to interference or other network problems. This is why the fragment number exists — so that receiving stations can reassemble data units in the intended order regardless of the sequence in which they arrive. Because each fragment must be positively acknowledged by the receiving station, 802.11 provides a mechanism to ensure that all fragments eventually do arrive. If a sending station does not receive the ACK for a fragment, it simply resends the fragment after an internal timer expires. If the receiving station receives multiple copies of the same fragment, it discards the excess copies of the fragment. With this in mind, you can see that a large number of Missing Fragment Number alarms may indicate significant interference on the network. You should check the Dashboard to see if there are also a large number of CRC errors on the network. If this is true, you may want to adjust the fragment size used by the MAC to use smaller fragments and see if this reduces the number of CRC errors on the network (and, correspondingly, the amount of Missing Fragment Number alarms generated). Oversized WLAN Frame The Expert generates the Oversized WLAN Frame alarm when it detects an 802.11 MAC frame longer than the maximum acceptable length dictated by the 802.11 specification. Expert Alarms Reference Guide 147 Chapter 1 The maximum acceptable length for an 802.11 MAC frame is 2346 bytes. Reassociation Failure The Expert generates the Reassociation Failure alarm when it detects an 802.11 Reassociation Response frame with a value other than zero in the Status Code field. A non-zero value in the Status Code field indicates that the access point sending the Reassociation Response is denying the requested association. Wireless stations send Reassociation Request frames to become associated with a different access point within the same network as its current access point (for example, because the station has moved and is now out of range of its old access point and within range of another). In turn, access points reply to Reassociation Requests with Reassociation Responses indicating the success or failure of the request. In this case, the access point denied the Reassociation Request. The exact reason for the denial is found in the Status Code field of the Reassociation Response. The Expert reports both the address of the access point denying the Reassociation Request, as well as the reason for the denial indicated in the Status Code field, as follows: 1 — Unspecified failure. 10 — Cannot support all requested capabilities in the Capability Information field. 11 — Reassociation denied due to inability to confirm that association exists. 12 — Association denied due to reason outside the scope of the 802.11 standard. 17 — Association denied because the access point is unable to handle additional associated stations. 18 — Association denied due to requesting station not supporting all of the data rates in the BSSBasicRateSet parameter. Rogue Access Point The Expert generates the Rogue Access Point alarm when it detects a wireless access point on the network whose MAC address is not found in its list of known access points. You can view the Expert's list of known access points in the Known Access Points in the Network listbox in the 802.11 Options tab of the Expert Properties dialog box. You access this tab by selecting Expert Options from the Tools menu and clicking on the 802.11 Options tab in the dialog box that appears. 148 Sniffer Technologies Expert Alarms and Thresholds The Rogue Access Point alarm provides you with a convenient means of detecting access points on the network of which you were previously unaware. To use this alarm effectively, you must add the MAC addresses of the known access points on the network to the Expert's list. You can add access points to the Expert's list in any of the following ways: Automatically in the real-time Host Table by selecting entries in the table, right-clicking, and selecting the Add to Known Mobile Unit List command. Automatically in the postcapture display's Expert tab by clicking the Wireless Unit List button and using the options in the dialog box that appears. Automatically in the Address Book by clicking the Export AP button. Manually in the Tools > Expert Options > 802.11 Options tab. In addition, you must also have enabled the Enable Rogue AP Lookup option on the 802.11 Options tab. When the Enable Rogue AP Lookup option is enabled, each time the Expert discovers a new access point, it will compare its MAC address to those in its list of known access points. If the discovered address is not found, the Expert generates the Rogue Access Point alarm. In addition, the Expert displays will identify the offending access point as a rogue (the word “Rogue” will appear in parentheses following the station's entries in Expert Summary and Detail displays). Possible Cause: In most cases, this is a relatively minor alarm, probably indicating nothing more than that you neglected to add the address of a known access point to the Expert's list. However, you may want to examine the address of the access point indicated in the alarm to make sure that it is not an intruder. Rogue Mobile Unit The Expert generates the Rogue Mobile Unit alarm when it detects a mobile unit on the wireless network whose MAC address is not found in its list of known mobile units. You can view the Expert's list of known mobile units in the Known Mobile Units in the Network listbox in the 802.11 Options tab of the Expert Properties dialog box. You access this tab by selecting Expert Options from the Tools menu and clicking on the 802.11 Options tab in the dialog box that appears. The Rogue Mobile Unit alarm provides you with a convenient means of detecting mobile units on the network of which you were previously unaware. To use this alarm effectively, you must add the MAC addresses of the known mobile units on the network to the Expert's list. You can add mobile units to the Expert's list in any of the following ways: Expert Alarms Reference Guide 149 Chapter 1 Automatically in the real-time Host Table by selecting entries in the table, right-clicking, and selecting the Add to Known Mobile Unit List command. Automatically in the postcapture display's Expert tab by clicking the Wireless Unit List button and using the options in the dialog box that appears. Manually in the Tools > Expert Options > 802.11 Options tab. In addition, you must also have enabled the Enable Rogue Mobile Unit Lookup option on the 802.11 Options tab. When the Enable Rogue Mobile Unit Lookup option is enabled, each time the Expert discovers a new mobile unit, it will compare its MAC address to those in its list of known mobile units. If the discovered mobile unit is not found, the Expert generates the Rogue Mobile Unit Detected alarm. In addition, the Expert displays will identify the offending mobile unit as a rogue (the word “Rogue” will appear in parentheses following the station's entries in Expert Summary and Detail displays). Possible Cause In most cases, this is a relatively minor alarm, probably indicating nothing more than that you neglected to add the address of a known mobile unit to the Expert's list. However, you may want to examine the address of the mobile unit indicated in the alarm to make sure that it is not an intruder. Runt WLAN Frame The Expert generates the Runt WLAN Frame alarm when it detects an 802.11 MAC frame shorter than the minimum acceptable length dictated by the 802.11 specification. The minimum acceptable length for an 802.11 MAC frame is 34 bytes. However, some control and management frames are inherently smaller than 34 bytes. The Expert does not generate alarms for these frames. Same Transmitter and Receiver Address The Expert generates the Same Transmitter and Receiver Address alarm when it detects an 802.11 MAC frame in which the transmitter and receiver addresses indicated in the frame are the same. 150 Sniffer Technologies Expert Alarms and Thresholds Wireless LAN frames include up to four addresses in the standard 802.11 MAC format depending on the type of frame. In addition to the Source and Destination addresses (which refer to the original source of and the final destination for the protocol data in the frame body field), 802.11 MAC frames include Transmitter and Receiver addresses. These addresses are those of the wireless stations responsible for transmitting the frame onto the wireless medium (transmitter address) and the next recipient of the frame on the wireless medium (receiver address). The Expert generates this alarm if the transmitter and receiver addresses within the frame are the same. If, however, the source and destination addresses within the same frame are also the same, the Expert will also generate the Same Source and Destination Address alarm at the Expert DLC layer. Transmitter Address Is Broadcast The Expert generates the Transmitter Address Is Broadcast alarm when it detects an 802.11 MAC frame with a broadcast address (all 1s) indicated in the Transmitter Address field. Transmitter Address Is Multicast The Expert generates the Transmitter Address Is Multicast alarm when it detects an 802.11 MAC frame with a multicast address indicated in the Transmitter Address field. WEP-ICV Error The Expert generates the WEP-ICV Error alarm when it detects a WEP-encrypted packet with an Integrity Check Value (ICV) which does not match the ICV calculated by the Expert using its own WEP keys. This usually happens when Sniffer® Technologies is configured with an incorrect set of WEP keys. In a wireless network using shared key authentication, each station on the network is programmed with the same four WEP keys (1-4). Wireless stations send WEP-encrypted packets with header fields indicating which of the four shared WEP keys was used to encrypt the data. Receiving stations use the shared key indicated in the packet’s header (1-4) for decryption and calculate an expected Integrity Check Value (like a checksum for the encrypted data) to compare against the ICV included in the received packet. When Sniffer® Technologies detects a WEP-encrypted packet, it attempts to decrypt the data using its own shared WEP keys specified on the 802.11 tab of the Options dialog box (accessed from the Tools > Options menu). If the ICV it calculates using its WEP keys does not match the ICV included in the packet, the Expert generates this alarm. Expert Alarms Reference Guide 151 Chapter 1 Possible causes: The Expert is configured with WEP keys which do not match those in use on the wireless network being analyzed. Go to the 802.11 tab of the Options dialog box (accessed from the Tools > Options menu), and make sure that the WEP keys specified there match those in use on the network. The station that sent the offending packet is configured with the wrong WEP keys for the network. Make sure its keys are programmed correctly. Global Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Global layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Bad CRC The Expert generates the Bad CRC alarm when a frame with a bad Cyclic Redundancy Check (CRC) is detected. The CRC is a checksum calculated from the contents of a packet. A CRC value is calculated at the source node and appended to the packet. It is then recalculated from the received packet by the destination node. If the two values are different, it indicates an error. Possible causes: When capturing from a wireless network, CRC errors appear routinely because of multipass and attenuation distortions. This is a normal part of wireless LAN operations and is not ordinarily a cause for concern. A faulty device on the network. A bad network interface card in a machine on the network. Broadcast/Multicast Storm The Expert generates the Broadcast/Multicast Storm alarm when the number of broadcast or multicast frames per second exceeds the Broadcast Frames/sec threshold under the Broadcast/Multicast Storm entry in the Alarms tab of the Expert UI Object Properties dialog box, indicating that a broadcast storm has been detected. Possible causes: 152 Sniffer Technologies This could be a temporary condition resulting from a valid broadcast or multicast operation. Expert Alarms and Thresholds A workstation that does not have an adequate host table is sending repeated RWHO packets. Change the workstation to assure that it maintains a host table and stops sending RWHO packets. The broadcast storm indicates some other underlying configuration problem. If broadcast storms occur frequently, their causes should be determined. A network can slow down considerably during such storms. Broadcast/Multicast Storm Diag The Expert generates the Broadcast/Multicast Storm Diag alarm when the number of broadcast or multicast frames per second exceeds the Broadcast Frames/sec threshold under the Broadcast/Multicast Storm Diag entry in the Alarms tab of the Expert UI Object Properties dialog box, indicating that a severe broadcast storm has been detected. If a broadcast storm diagnosis occurs, then a broadcast storm symptom will also have occurred. Refer to the detailed information for that symptom. Possible causes: This could be a temporary condition resulting from a valid broadcast or multicast operation. A workstation that does not have an adequate host table is sending repeated RWHO packets. Change the workstation to assure that it maintains a host table and stops sending RWHO packets. The broadcast storm indicates some other underlying configuration problem. If broadcast storms occur frequently, their causes should be determined. A network can slow down considerably during such storms. Channel Mismatch The Expert generates the Channel Mismatch alarm in the following situations: In an infrastructure wireless network (a wireless network with access to a distribution system), the Expert generates this alarm when it receives beacon and/or probe response frames from an access point on a channel other than the channel on which the access point is configured to operate. In an ad hoc wireless network (a wireless network with no access to a distribution system), the Expert generates this alarm when it receives beacon and/or probe response frames from a wireless station on a channel other than the channel on which the station is operating. Expert Alarms Reference Guide 153 Chapter 1 In an 802.11 infrastructure wireless network, access points send beacon frames at a regular interval. In addition, they send probe response frames in response to probe request frames sent from wireless stations wanting to join the network. In an ad hoc network, the stations themselves send beacon and probe request frames. Among other parameters, beacon frames and probe requests specify the wireless channel on which the basic service set (BSS) is operating. The wireless stations in a single BSS can only operate on one channel at a time — the channel on which the BSS is operating. However, due to adjacent channel interference, wireless stations can occasionally receive frames from stations operating on a different channel. The Expert generates the Channel Mismatch alarm when this happens. Collisions Over Threshold The Expert generates the Collisions Over Threshold alarm when the number of collisions detected exceeds the Collision Count threshold. A collision is the result of two or more nodes attempting to use the medium at the same time. Collisions are part of normal Ethernet behavior. However, a large number of collisions can indicate a problem. Possible cause: Too many devices on the segment. LAN Overload The Expert generates the LAN Overload alarm when the data rate observed on the network exceeds the percentage of available network bandwidth specified by the LAN Load threshold. Possible causes: Multiple large file transfers are occurring simultaneously. Many users are doing high volume work at the same time. This is especially likely if it occurs at about the same time each day. There is a defect that is causing a single station to do something repetitive. A router has recently been rebooted. The remedy is approximately the same for all of these causes. Investigate which devices are involved, and try to surmise whether the overload has resulted from normal activity or from an error condition. Interpret this symptom on a case-by-case basis. 154 Sniffer Technologies Expert Alarms and Thresholds LAN Overload Percentage The Expert generates the LAN Overload Percentage alarm when during any one-minute interval, the percentage of time the LAN is in an overloaded state (the data rate was greater than the LAN Load threshold) exceeds the LAN Overload % threshold. Possible causes: Multiple large file transfers are occurring simultaneously. Many users are doing high volume work over the network at the same time. This is especially likely if it occurs at about the same time each day. A single station is doing something repetitive. A router has been rebooted recently. There is a repeater, bridge, or router loop. This condition causes severe LAN overload (for example, LAN utilization exceeding 80% of bandwidth for more than 10 seconds). The remedy is similar for all of these causes. Investigate which devices are involved, and try to surmise whether the overload resulted from normal activity or from an error condition. Interpret this diagnosis on a case-by-case basis. If the network is chronically overloaded, and if this overload condition results in unacceptable delay times, split your network into segments, or take other measures to reduce network traffic. PLCP Error The Expert generates the PLCP Error alarm when a wireless station receives a Physical Layer Convergence Protocol header with an invalid checksum. Before frames are sent between wireless stations, the physical layer (PHY) sends a PLCP header to a receiving station to negotiate the size of the frames to be sent, the speed at which they should be sent, and so on. This PLCP header includes a checksum that the receiving station uses to validate that the received PLCP header is not corrupt. The Expert generates this alarm if it detects that a station has received a PLCP header in which the checksum is corrupt. Spanning Tree Topology Change The Expert generates the Spanning Tree Topology Change alarm when it sees that a spanning tree bridge calculation has been done. Expert Alarms Reference Guide 155 Chapter 1 Possible causes: A bridge has gone up or come down somewhere on the logical LAN. Someone has reconfigured a port cost or the priority of a bridge somewhere on the logical LAN. If changes are occurring at regular intervals, there is probably a misconfigured bridge on the logical LAN. Spanning tree bridges are also called “transparent” or “learning” bridges because they take responsibility for learning where to forward traffic. Stations on the network can be completely unaware of the bridges that connect several physical segments into one logical LAN, yet everything still works. The bridges also communicate among themselves, with BPDU messages. These messages ensure that no disastrous “routing loops” exist, through which bridges would endlessly forward the same frame from segment to segment on the logical LAN. The bridges preclude such loop formation by shutting off some of the ports on some of the bridges, as necessary, to eliminate loops, while still maintaining full connectivity. The bridges select one “root bridge” for the entire logical LAN, as a point of reference, and one “local designated bridge” for each physical segment. Typically, the root bridge is chosen on the basis of priority. The bridge configured with the highest priority becomes the root. Ties are broken by bridge id. A local designated bridge is chosen based on the lowest “root path cost” – the “cost” of sending a frame from that bridge to the root. A transmission cost is associated with each port on each bridge. A high cost normally means a slow link (for example, a 56 Kbps WAN connection inside the bridge to that port). Conversely, a low cost usually means a fast link. These costs are typically configurable. The “root path cost” for any bridge is the cost for that bridge to send a frame from that bridge toward the root, plus the sum of the costs to forward the frame for all the bridges between that bridge and the root bridge. A root path cost of 0 implies that a bridge is the root. Once a spanning tree has been done, the network settles into a steady state in which the only BPDU messages being sent are “hello” messages advertising the current spanning topology. These are sent periodically by each local designated bridge (the root bridge specifies the “hello time” parameter, which determines how often the hello messages are sent). These hellos are for the benefit of any bridges that might be added to the network. Also, if the local designated bridge goes down, the absence of hellos will enable the other bridges on the segment to detect it. They would then initiate a new spanning tree calculation, which would have the effect of working “around” the failed bridge. 156 Sniffer Technologies Expert Alarms and Thresholds Once in the steady state only the local designated bridge will be sending BPDU messages. Also, BPDU messages are never forwarded beyond the local segment. Therefore, the analyzer will normally be aware of only one bridge – the designated bridge on the segment to which it is attached – even if there are other bridges on that segment. However, if any significant bridge activities take place (such as a bridge being added or going down) anywhere on the logical LAN (even if outside the local segment), the Expert system will be aware of it. This is because the LAN-wide spanning tree calculation must be done. The Expert analyzer presents a complete record of these changes under the Alarms display at the Global layer. The “local initiator” of a change is the sender of the packet on the local segment that first made the Expert system aware that a spanning calculation was being done. This could be: A CHANGE type BPDU message. A hello packet containing new topology information. A hello packet sent by someone other than the local designated bridge. This may or may not be the actual bridge that caused the spanning tree calculation, because the calculation itself may have been initiated outside the local segment. VLAN Not Operational The Expert generates the VLAN Not Operational alarm when it sees a VTP frame with a Status field set to 1, 2, or 3, indicating that the VLAN is not operational. The meanings of these status fields are as follows: 1 – Suspended 2 – MTU (Maximum Transmission Unit) Too Big For Device 3 – MTU (Maximum Transmission Unit) Too Big For Trunk The alarm display shows the value of the Status field that caused the alarm. VLAN Removed from Domain VTP advertisements announce the presence of all VLANs in a domain. The Expert uses the information in these packets to maintain a list of VLANs it has seen advertised in each domain. If a subsequent VTP advertisement includes only a subset of the previously advertised VLANs, the Expert generates the VLAN Removed from Domain alarm, indicating that a VLAN has been removed from the domain. Expert Alarms Reference Guide 157 Chapter 1 For example, suppose the Expert sees a VTP advertisement indicating the presence of VLAN 1, VLAN 2, VLAN 3, and VLAN 4 in the Domain named Perignon. If the Expert sees a subsequent VTP advertisement for Domain Perignon with only VLAN 1 and VLAN 2, it will generate the VLAN Removed from Domain alarm, indicating that VLAN 3 and VLAN 4 are no longer seen in the domain. The alarm display provides the VLAN ID of the VLAN that was removed. VTP Versions Different The Expert generates the VTP Versions Different alarm when it sees two stations on the same VLAN send VTP frames with different VTP version fields. To ensure proper operation, all stations on the same VLAN should use the same version of VTP. Route Layer Alarms and Thresholds This section describes Expert alarms and thresholds at the Route layer. Alarms are organized alphabetically and are described with their corresponding thresholds (if any). Nonsense Route The Expert generates the Nonsense Route alarm when a router advertises a route that requires sending frames to a next hop address that is not on the local subnet. Possible causes: This router has been configured with an IP address that is not valid on the subnet. This router is inappropriately forwarding routing information frames from other routers onto this subnet. The analyzer has not been configured properly to know its own subnet address or addresses, and the automatic mechanism that the analyzer uses to find these addresses has not been able to correct the problem. Route Metric Changed The Expert generates the Route Metric Changed alarm when it sees a router broadcast a route indicating a metric change. 158 Sniffer Technologies Expert Alarms and Thresholds Possible cause: A previously operating router or link in the path has failed, or a previously failed router or link in the path has come back on line. Route Superseded The Expert generates the Route Superseded alarm when it sees a route become the best route to a given destination or be superseded by another, better route. Possible causes: The router changed the metric associated with this route, making it now the best (or no longer the best) route. The metric associated with another route to the same location changed, making this route now the best (or not longer the best) route. Note that if this is the case, then the entry on the history screen will indicate that the change in this route's status was “dynamic.” Route Validity Changed The Expert generates the Route Validity Changed alarm when it sees a route change from valid to invalid or vice versa. Possible causes: The router advertised this route as available but later advertised that the route was unavailable (or vice versa). The router advertised this route but later failed to confirm it before it expired. Expert Alarms Reference Guide 159 Chapter 1 160 Sniffer Technologies Index A AAL5 CRC Errors, 134 AAL5 Length Errors, 134 AAL5 Timeout Errors, 135 Abnormal Disconnect for SVC, 135 Abort Delimiter Transmitted, 82 AC Errors, 82 Access to Resource Denied, 15 ACK Frame Timeout, 143 Ack Too Long, 65 Alarm Expert thresholds, 3 ALM_ApplicationH323IncompleteCall, 12 ALM_ApplicationSCCPIncompleteCall, 14 ALM_ApplicationSCCPTooManyCalls, 13 ALM_ApplicationSIPIncompleteCall, 15 ALM_ApplicationSIPTooManyCalls, 14 ALM_SesSlowFileTransfer, 56 ALM_SessTempCongestion, 56 Application Layer Alarms and Thresholds, 12 ARE or STE Frames on Data Direct VCC, 105 ARP Request Multi Queued, 105 Association Failure, 143 Async Status Change of Unknown DLCI, 95 ATM Application Layer Alarms and Thresholds, 100 ATM Connection Layer Alarms and Thresholds, 134 ATM Flow Layer Alarms and Thresholds, 105 ATM Host Layer Alarms and Thresholds, 140 Attempt to Bind DLC to MPC and Non-MPOA Device, 105 Attempt to Bind L3 Address to Multiple MPOA Devices, 100 Attempt to Multiply Bind MPOA Devices to LEC, 106 audience for this manual, xv Authentication Failure, 144 B Backward Congestion, 92 Bad CRC, 152 Broadcast/Multicast Data Frames on DDVCC, Broadcast/Multicast Storm, 152 Broadcast/Multicast Storm Diag, 153 Browser Election Force, 15 Burst Errors, 82 106 C Cache ID Set to Zero in DLL Header Ext, 106 CLP = 1, 136 Collision after 64 Bytes, 82 Collisions Over Threshold, 153 Config Fail, Insufficient Resources, 106 Config Frag Info TLV Not Recognized by LECS, Config Request Multi Queued, 107 Config Request Response Timeout, 107 Config Request Retry Frame Mismatch, 108 Config Request Retry Rate Exceeds 1 Frame/sec, 108 Connection Layer Alarms and Thresholds, 65 Connection Too Short, 136 contacting Network General, xv Control Request on DDVCC Retry Frame Mismatch, 108 Control Request on DDVCC Retry Rate > 1 Frame/Sec, 109 Crankback Has Occurred, 136 CTS Frame Timeout, 145 107 D Data Frames Detected Following Data Plane Purge, 100 Database Summary Exchange Lockstep Failure, 109 Database Summary Master/Slave Negotiation Timeout, 109 DB Connect Request Denied, 16 DB Security Breach Attempt, 16 DB Slow Connect, 16 Expert Alarms Reference Guide 161 Index DB Slow Server Response, 17 DB Slow Server Response Diagnosis, 18 DCE Sequence Number Error, 95 DDVCC Idle Period Exceeds VCC Timeout C12, 101 DDVCC Not Created in Response to LE ARP, 101 Deauthentication, 145 Default MCFVCC Not Created in Response to LE ARP, 101 Default MCSVCC Not Created in Response to LE ARP, 102 Delayed STATUS ENQUIRY, 95 Denied Access to ELAN, 109 Diagnosis in Expert analysis, 1 Disassociation, 146 DLC Layer Alarms and Thresholds, 82 DLC Source Address Broadcast, 82 DLC Source Address Multicast, 83 DLL Header Extension Not Found in Frame, 110 DTE Sequence Number Error, 95 Duplicate Address Found, 83 Duplicate ATM Destination Address Registration, 110 Duplicate LAN Destination Registration, 110 Duplicate Network Address, 72 Duplicate PDC Name Registration, 73 E EFCI, 137 Egress Cache Tag Was Not Issued, 110 Exceeded CIR, 92 Excessive Abnormal Disconnects, 140 Excessive Abnormal Frame Sequences, 140 Excessive Failed Resource Login Attempts, 20 Excessive Mailslot Broadcasts, 20 Excessive Short Connections, 141 Excessive Signaling Recovery, 137 Excessive Unknown Frames Issued by LEC, 102 Expert diagnoses, 1 symptoms, 1 thresholds, 3 162 Sniffer Technologies F Failed to Issue Resolution Req. Following Trigger, 111 Failure Reported in Reply CIE, 111 Fast Retransmission, 65 FCI and ARI Bits Mismatch, 84 FCIP Invalid Encapsulation, 21 FCIP Repeat Special Frame, 21 FCIP Special Frame Delayed, 22 FCIP Special Frame Discovery, 20 FCIP Special Frame Echo Has Changed, 21 FCIP Too Many Special Frame Echoes, 22 FCIP Transmission Errors, 22 FDDI Ring Wrap, 84 File Retransmission, 53 Flush Request Multi Queued, 111 Forward Congestion, 92 Frame Has Alignment Problem, 85 Frame-Copied Errors, 85 Frames Seen on Default Flow After Shortcut Active, 102 Frequency Errors, 85 FTP Login Attempts, 23 FTP Slow Connect, 23 FTP Slow First Response, 23 FTP Slow Response, 23 FTP Slow Transfer Diagnosis, 24 G Global Layer Alarms and Thresholds, 152 H H225 - Abnormal Disconnect, 34 H245 - Open Logical Channel Reject, 35 H245 - Terminal Capability Set Reject, 36 H323 - High Call Volume, 12 H323 - Too Many Incomplete Calls, 12 High Rate of Line/Burst Errors, 85 High Rate of Physical Errors, 86 High Rate of Receiver Congestion Errors, 86 High Rate of Ring Purge Diagnosis, 87 High Signaling Activity, 141 Host Address Cannot Be Mux’ed, 141 Index HTTP Slow Server GET Response Time, 24 HTTP Slow Server POST Response Time, 24 Function Rejected, 27 Immediate Command Reject (Too Many Immediate Commands), 27 Initiator Authentication Failed, 27 Invalid Data Ack, 27 Invalid During Login, 28 Invalid PDU field, 27 LUN Does Not Exist, 28 Miscellaneous Initiator Error, 28 Missing Parameter, 28 Service Unavailable, 28 Session Does Not Exist, 28 Target Error, 29 Target Failure, 29 Target Out of Resources, 29 Target Removed, 29 Task Does Not Exist, 29 Task Management Function Not Supported, Too Many Connections, 30 Unsupported Version, 30 I ICMP Destination Unreachable, 73 ICMP Host Unreachable, 73 ICMP Inconsistent Subnet Mask, 74 ICMP Net Unreachable, 74 ICMP Parameter Problem, 74 ICMP Port Unreachable, 75 ICMP Redirect for Host, 75 ICMP Redirect for Network, 75 ICMP Source Quench, 76 Idle Too Long, 66 IG Tag Violation, 111 Insufficient Info from LEC, Service Refused, 112 Invalid ATM Primary Address, 112 Invalid CIE Contents, 112 Invalid Client ATM Primary Source Address, 112 Invalid Client MAC Address, 113 Invalid Common Header Contents, 113 Invalid Fixed Header Contents, 114 Invalid Frame Contents, 114 Invalid Frame Status Field, 87 Invalid Hello Frame, 114 Invalid MPOA Packet Size, 115 Invalid Network Name, 25 Invalid Presentation Context ID, 53 Invalid Resource User Name or Password, 25 Invalid Tag in Data Frame Header, 116 IP Fragment Missing, 76 IP Fragment Out of Order, 76 IP NETBIOS Session Reject, 53 Irregular Full Status Report Requests, 96 iSCSI Authorization Failure, 25 Can Not Generate Target Transfer Tag - Out of Resources, 25 Command Not Supported in this Session Type, 26 Command Retransmission, 26 Data Digest Error, 26 Data or Status again Requested, 26 Function Authorization Failed, 26 30 J Join Request Multi Queued, 116 Join Request Received from LES, 116 Join Request Transmitted by LES, 116 Join Response Issued by LEC, 117 K Keep Alive Lifetime Ext Not Found in Frame, Keep Alive Lifetime Set to Zero, 117 Keep Alive Sequence Failure, 117 117 L LAN Overload, 154 LAN Overload Percentage, 155 LANE Cfg Phase Failure, 102 LANE Cfg Restart Delay, 103 LANE Cfg Restart Delay > Max Recfg Delay, 104 LANE Control Request Response Timeout, 117 LANE Control Request Retry Rate > 1 Frame/sec, 118 LANE Control Request Retry Retry Frame Mismatch, 118 LANE Version Unsupported by ELAN, 119 Expert Alarms Reference Guide 163 Index LE ARP Request Rate Exceeds 1 Frame/Sec, 119 LE_CONFIGURE Params Conflict, Service Refused, 119 LEC ID in Frame Mismatch with Client LEC ID, 120 LEC Issued Config Response to LECS, 119 LEC Not Recognized by LECS, 119 LECID Is Not 0x0000 in Request Frame, 120 LECS Issued Config Request to LES, 120 Line Errors, 87 Link Management Message Format Error, 96 Link Management Message Minor Format Error, 97 Local Limit on Defined Context Set Exceeded, 54 Local Routing, 66 Local Segment ID not in Source Routed Frame, 121 Local Station with Route Designator, 87 Loops on Same Request, 54 Lost Frame Errors, 88 Low Throughput, 54 NCP Server Busy, 31 Network Equipment Reset, 97 New Active Monitor, 88 New PVC Configured, 97 NFS Retransmission, 32 NNTP Slow Response Time, 32 No CIEs Were Present in MPOA Reply, 125 No CIEs Were Present in MPOA Request, 125 No Reply Flag Is Not Set in ICache Purge, 125 No Resolution Request Issued After Excess Default Flow, 104 No STATUS ENQUIRY When Expected, 98 No STATUS following STATUS ENQUIRY, 98 Non-Responsive Station, 66 Nonsense Route, 158 O Oversized Frame, 88 Oversized WLAN Frame, 147 M MAC Address bound to Multiple LECs, 121 Max Transmission Unit Size Undefined, 121 Mcast/Bcast Fragmentation, 146 Misdirected Frame, 77 Missed Browser Announcement, 30 Missing DTL or Excessive DTL Size, 138 Missing Fragment Number, 147 Missing ULIA on Outside Link, 121 MPCMPS Flow Not Allowed on this VC, 138 MPOA Capable LEC Did Not Issue Device TLV, 122 MPOA Config Parameter Invalid, 122 MPOA Control Request Response Timeout, 122 MPOA Control Request Retry Rate > 1 Frame/sec, 123 MPOA Error Frame Detected, 124 MS RAP Logon Failure, 31 Multiple Local Ring Definition, 88 Multiple Routers to Local, 77 Multiple Routers to Remote, 78 Multiply Assigned Secondary LEC Address, 124 N NARP Request Multi Queued, 124 NARP Request Received from LES, 164 Sniffer Technologies 125 P PLCP Error, 155 POP3 Slow Connect Time, 32 Possible 3.0/3.1 UNI Mismatch, 138 Protocol Negotiation Failure, 32 PTSE Violation, 126 PVC Activation, 92, 98 PVC Bandwidth Change, 92, 98 PVC Congestion, 93, 98 PVC Deactivation, 93, 99 PVC Deletion, 93, 99 R RAS - Admission Reject, 36 RAS - Bandwidth Reject, 38 RAS - Disengage Reject alarm, 38 RAS - Gatekeeper Reject, 38 RAS - Location Reject, 39 RAS - Registration Reject, 40 RAS - Unregistration Reject alarm, 41 Read/Write Overlap, 55 Reassociation Failure, 148 Receiver Congestion Errors, 88 Index Register Request Multi Queued, 126 Register Request Received from LES, 126 Register Response Issued by LEC, 126 Remote Address Issued by Non Proxy LEC, 127 Repeat Ack, 67 Request Denied, 55 Request Params Are Incompatible with ELAN, 127 Retransmission, 67 Returned to Ring, 89 Ring Beaconing, 89 Ring Purge Symptom, 89 Rogue Access Point, 148 Rogue Mobile Unit, 149 Route Flapping, 78 Route Layer Alarms and Thresholds, 158 Route Metric Changed, 158 Route Superseded, 159 Route Validity Changed, 159 Router Not Updating Routes, 79 Router Storm, 79 Router Superseded Too Frequently, 79 RTCP - Report High Jitter, 41 RTP - High Jitter, 43 RTP - High Variation, 45 RTP - Too Many Dropped Frames, 46 RTP - Too Many Out of Sequence Frames, 47 Runt Frame, 90 Runt WLAN Frame, 150 Session Layer Alarms and Thresholds, 34 Shortcut Create Failed After Resolution Reply, Signaling Frame Sequence Problem, 138 SIP - Client Error, 49 SIP - Global Error, 51 SIP - High Call Volume, 14 SIP - Server Error, 52 SIP - Server Slow Response, 52 SIP - Too Many Incomplete Calls, 15 Slow File Transfer, 56 Slow FTP Server, 6 Slow HTTP Server, 6 Slow Mail Server, 7 Slow News Server, 7 Slow Oracle SQL Server, 7 Slow POP3 Server, 9 Slow Server, 68 Slow Sybase/Microsoft SQL Server, 10 Slow Telnet Server, 11 SMTP Slow Connect Time, 33 Source Address Is Broadcast, 81 Spanning Tree Topology Change, 155 Station Internal Errors, 90 Station Not in Domain’s Computer List, 34 Station Off Ring, 90 Station Removed, 90 Symptom in Expert analysis, 1 104 T S Same Source and Destination Address, 90 Same Transmitter and Receiver Address, 150 SAP R3 Slow Server Connection, 33 SAP R3 Slow Server Response, 33 SCCP - High Call Volume, 13 SCCP - Register Reject, 49 SCCP - Register Reject alarm, 53 SCCP - Station Alarm, 49 SCCP - Too Many Incomplete Calls, 14 Server Running BDC Shutting Down, 80 Server Running PDC Shutting Down, 80 Server Running WINS Shutting Down, 80 Service Layer Alarms and Thresholds, 6 TA List Resources Exhausted, 127 TCP Fast Keepalive, 68 Telnet Slow Response to Login, 34 Temporary Congestion, 56 The Operation Number Is Greater than the Number of Operations in the Interface, 57 The Output Parameters of the Operation Exceed Their Declared Maximum Size, 57 The Request Is Being Rejected for Unspecified Reasons, 57 The RPC Client or Server Protocol Has Been Violated, 57 The Server Did Not Support the Requested Authentication Level, 57 Expert Alarms Reference Guide 165 Index The Server Does Not Implement the Requested Operation for the Type of Requested Object, 57 The Server Does Not Support RPC Protocol Version, 58 The Server Does Not Support the Interface, 57 The Server Does Not Support the Proposed Transfer Syntax, 58 The Server Does Not Support the Requested Interface, 57 The Server Is Too Busy To Handle the Call, 58 The Server Manager Routine Has Not Been Entered and Executed, 58 Thresholds Expert, 3 Time-to-live Exceeded in Transmit, 81 Time-to-live Expiring, 81 TNS Connect Refused, 58 TNS Security Breach Attempt, 58 TNS Slow Connect, 59 TNS Slow Server Diagnosis, 59 TNS Slow Server Response, 61 Token Errors, 91 Too Many File Retransmissions, 62 Too Many Loops on Same Request, 63 Too Many Removes, 91 Too Many Requests Denied, 63 Too Many Retransmissions, 69 Too Many Returns, 91 Too Many Simultaneous Connections, 142 Topology Change Request Multi Queued, 128 Topology Mismatch in Data Frame Types, 128 Traffic on Deleted DLCI, 93 Traffic on Inactive DLCI, 94 Traffic on Unknown DLCI, 94 Transmitter Address Is Broadcast, 151 Transmitter Address Is Multicast, 151 U UDP Bouncing Frames, 69 Unexpected Data for SVC, 139 Unexpected Data for SVC (ATM Host Layer), 142 Unexpected DLL Header EXT in Frame, 128 Unexpected Egress Cache EXT in Frame, 128 Unexpected Hop Count EXT in Frame, 129 166 Sniffer Technologies Unexpected Keepalive Lifetime EXT in Frame, 129 Unexpected Link Management Message, 99 Unexpected Original Error Code EXT in Frame, 129 Unexpected Service Category EXT in Frame, 129 Unexpected TLVs Detected, 130 Unknown Extension in Frame, 130 Unknown IG, 130 Unknown LANE ConfigControl Opcode, 130 Unknown Marker/Protocol Values, 131 Unknown/Unexpected Authenticate EXT in Frame, 131 Unknown/Unexpected CIE in Frame, 131 Unregister Request Multi Queued, 131 Unregister Request Received from LES, 132 Unregister Response Issued by LEC, 132 Unsolicited STATUS Message, 99 User Equipment Reset, 100 V V2 TLVs Issued in V1 ELAN, 132 Verify Request Multi Queued, 132 Verify Request Received from LES, 133 Verify Response Issued by LEC, 133 VLAN Not Operational, 157 VLAN Removed from Domain, 157 , 12 to 15, 34 to 36, 38 to 41, 43, 45 to 47, 49, 51 to 52 VTP Versions Different, 158 W WAN Connection Layer Alarms and Thresholds, WAN Link Layer Alarms and Thresholds, 95 WEP-ICV Error, 151 Window Frozen, 70 Window Size Exceeded, 70 WINS Duplicate Name, 64 WINS No Response, 64 Wireless Layer Alarms and Thresholds, 143 Wrong Report Type in STATUS Message, 100 Z Zero Broadcast Address, 81 Zero MAC Address, 91 Zero Window Too Long, 70 92