Design and Develop an IBM ODM application to optimize Performance
Transcription
Design and Develop an IBM ODM application to optimize Performance
3224 - Design and Develop an IBM ODM application to optimize Performance Pierre-André Paumelle ODM Performance Architect [email protected] © 2015 IBM Corporation IBM Operational Decision Manager Manage Govern Business Console Enterprise Console Access and Control Decision Artifacts Decision Center Versioned Assets Decision Server Decision Execution Decision Monitoring Web Services – API – GUI – Execution REST API Rule Execution Server Console Rule Designer Design POS Enterprise Application Monitor BPM CRM Mobile 1 Decision Service architecture Rule Engine Execution IN Ruleflow IN Rules OUT BOM to XOM XOM ODM Performance team role • The performance team • Works on the evolution of the product – To improve performance – To improve scalability • • • • Measures the performance Publishes the performance results Publishes Red papers on tuning and performance of ODM Follow up of customers – Performance and scalalbility questions – Migration from older versions – Tuning of architecture 3 Performance Evolution of ODM Performance evolution HTDS SOAP and REST JSON Sequential execution XML/JAVA XOM Total factor improvement 11x 16000 22x 20x 19x SOAP XML XOM V7.1.1 (CRE) 14485 14000 SOAP Java XOM v7.5 (CRE) 12663 SOAP Java XOM V8.0 (CRE) 12000 SOAP JAVA XOM V8.6 (DE) TPS 10000 REST JSON V8.7 (DE) 8203 8000 7180 6353 6000 4898 3670 4000 2000 1376 570 1489 2400 2642 1658 686 234 442 0 300 4588 2912 8736 137 294 356 14560 Ruleset size 4 Performance Evolution of ODM Performance evolution POJO RES Sequential execution JAVA XOM Total factor improvement 3x 9x 10x V7.1.1.1 (CRE) 35000 30000 7x 31818 V7.5 (CRE) 29679 V8.0 (CRE) V8.6 (DE) 25000 V8.7 (DE) TPS 20000 15000 13723 12115 11216 12864 12643 10000 5000 1817 2309 2627 3869 4677 526 689 746 0 300 2912 8736 266 338 504 2380 2692 14560 Ruleset size 5 Characteristics of an ODM application with good performance • Application meets the current SLA • Application should be ready to meet the SLA with the evolution of rules • The main evolutions could be: • Increasing number of rules (~20% per year) • Addition of new decision services on a shared Rule Execution Server • Changes in the structure of the decision services 6 Rule Project Design Rule Project design • Choose the XOM type: • JAVA XOM – Provide better performance and scalability. – Now accessible with HTDS • XML XOM easy to use • XOM Import : Import what is necessary. • Ex. Do not import all WAS classes if not necessary. • BOM/VOC/B2X: • Limit your BOM and VOC, to limit the size of the vocabulary when you edit the rules. • Limit the usage of “Update object state” property ( RetePlus case) 8 Rule Project Design • Divide your rule application in a set of rule projects with dependencies. • Enable Decision Validation Services before Rule Team Server Publication. • Ruleflow: • With Classic Rule Engine ruleflows are interpreted – limit the size and the complexity of the ruleflow • With Decision Engine Ruleflows are compiled • Limit the usage of dynamic filter in rule tasks, use static list of rules as much of possible • Limit the number of engine algorithms used to save memory 9 Rule Project design • Choose the correct engine algorithm depending on your Decision Service. • RetePlus – Rule chaining application – May be useful in the case of many objects and a small number of rules – Performance impacted by the number of rules • Sequential – Application with many rules and few objects – Really efficient in multi-thread environment. • Fastpath (default mode in ODM 8.7) – Application with a large number of rules and many objects – Faster at runtime. – Really efficient in multi-thread environment – Much more scalable when the rule number is growing – Optimal for Decision Table 10 Decision Engine • Massive investment and innovation on a new rule engine: • • • • • • • • Full rewrite of compilation and execution pipeline Faster execution under any circumstances Improved scalability Far less memory consumption Significantly reduced ruleset loading time Designed for concurrent execution Architecture open for extensions Migration path trivial for the vast majority of cases • On a next generation rule engine now called “Decision Engine” • Compatible with the ‘classic rule engine’ 11 Decision Engine benefit • Faster ruleset Loading Time Loading time for sequential execution (JSE) Java XOM - Heap 1GB CRE DE Intermediate code DE Java bytecode 50 45 43 40 37 Loading time (s) 35 35 33 30 27 26 25 21 21 20 15 12 11 10 7 5 4 5 4 7 5 0.5 1.6 1.3 0.8 0.6 0.5 2.0 2.3 0 300 500 1000 2912 5824 8736 11648 14560 Ruleset size Smaller is better 12 Decision Engine benefit • Reduce memory usage Minimum Heap Size (MB) for sequential execution (JSE) Java XOM CRE DE Intermediate code DE Java bytecode 700 640 600 512 Heap Size (MB) 500 384 400 300 256 256 256 192 200 128 100 32 16 32 4 64 18 128 64 32 4 16 8 4 16 20 16 0 300 500 1000 2912 Ruleset size 5824 8736 11648 14560 Smaller is better 13 Decision Engine benefit • Better Scalability Average Response Time for sequential execution POJO RES on Java XOM (2912 rules) V8.5.0 (CRE) vs V8.7.0 (DE Java bytecode) V8.5.0 (CRE) V8.7.0 (DE Java bytecode) 50 44 Average Response Time (ms) 45 40 35 30 23 25 20 15 11 10 10 8 7 6 5 4 2 1 2 4 1 1 1 1 2 1 0 1 5 10 15 20 25 Number of concurrent executions 30 60 120 Smaller is better 14 Topology choice Topology Choices based on your objectives • Smallest memory consumption? JSE Rule Execution Server with Decision Engine Rule Execution Server with Decision Engine on Liberty • No concurrent execution? JSE Rule Execution Server • Concurrent executions on distributed? JSE Rule Execution Server JEE Rule Execution Server • Execute on Cloud? Rule Service On BlueMix 16 Topology Choices based on your objectives • Integration with JEE products? JEE Rule Execution Server • Call a Web Service (SOAP)? HTDS • Call as a RESTful? HTDS REST XML HTDS JSON ( Java XOM) • Call from z/OS? zRES Rule Execution Server on WAS on zOS 17 Tuning of the Rule Execution Server Rule Execution Server best practices • Tune the usage of the execution trace and the Decision Warehouse by filtering Limit the size of your XOM to useful classes • A ruleset on an XML XOM should be configured to run in multiple simultaneous executions by configuring the pool of XML document drivers. This is configured with the ruleset.xmlDocumentDriverPool.maxSize property. • Decision Engine uses java assertions so you should check that assertions are not enable especially when you run inside a development environment (Eclipse) • Use the ruleset caching information (XU dump) from the Rule Execution Server console to diagnose the status of the internal pools 19 Rule Execution Server best practices • How can you minimize the performance overhead of ruleset parsing? • Parsing at Rule Execution Server launch or first invocation of the ruleset. In this case, the ruleset should be parsed before the first execution. You can force it with a dummy execution or a call of the loadUptodateRuleset (Management Session API). • Update of a ruleset already in use. This can be fixed by using the asynchronous parsing. • A ruleset is used infrequently once a day, for example, and removed from the pool running other rulesets. This can be fixed by setting ruleset.maxIdleTime on this ruleset. • Use Decision Engine (Java bytecode) to minimize the parsing duration 20 Performance testing Performance testing • The performance cost for a Decision Service may looks something like Execution Time XOM Ruleflow Functions B2X Misc. Rule engine JVM (GC) 22 Performance testing • Spend time identifying the bottleneck in your Decision Service. • Is it the rule engine execution, the database or others modules • Recommendations for avoiding performance issues in your rule projects include: • Consider putting continuous performance measurement in place to provide you with a safety net as you modify your XOM and rules. It is much easier to see the impact of adding a synchronized method call to the XOM, or adding a rule with a complex join if you are receiving performance or throughput reports every morning.. • Spend time optimizing your XOM • Spend time getting to know the features of your profiler. 23 Take away Take Away • • • • Use the engine algorithm adapted to your case Choose the topology adapted to the application type Use Decision Engine to increase the scalability Put performance tests in place to verify the evolution of the application • Leverage recent ODM and Java versions. Try to stay current on ODM version. 25 Notices and Disclaimers Copyright © 2015 by International Business Machines Corporation (IBM). No part of this document may be reproduced or transmitted in any form without written permission from IBM. U.S. Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM. Information in these presentations (including information relating to products that have not yet been announced by IBM) has been reviewed for accuracy as of the date of initial publication and could include unintentional technical or typographical errors. IBM shall have no responsibility to update this information. THIS DOCUMENT IS DISTRIBUTED "AS IS" WITHOUT ANY WARRANTY, EITHER EXPRESS OR IMPLIED. IN NO EVENT SHALL IBM BE LIABLE FOR ANY DAMAGE ARISING FROM THE USE OF THIS INFORMATION, INCLUDING BUT NOT LIMITED TO, LOSS OF DATA, BUSINESS INTERRUPTION, LOSS OF PROFIT OR LOSS OF OPPORTUNITY. IBM products and services are warranted according to the terms and conditions of the agreements under which they are provided. Any statements regarding IBM's future direction, intent or product plans are subject to change or withdrawal without notice. Performance data contained herein was generally obtained in a controlled, isolated environments. Customer examples are presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual performance, cost, savings or other results in other operating environments may vary. References in this document to IBM products, programs, or services does not imply that IBM intends to make such products, programs or services available in all countries in which IBM operates or does business. Workshops, sessions and associated materials may have been prepared by independent session speakers, and do not necessarily reflect the views of IBM. All materials and discussions are provided for informational purposes only, and are neither intended to, nor shall constitute legal or other guidance or advice to any individual participant or their specific situation. It is the customer’s responsibility to insure its own compliance with legal requirements and to obtain advice of competent legal counsel as to the identification and interpretation of any relevant laws and regulatory requirements that may affect the customer’s business and any actions the customer may need to take to comply with such laws. IBM does not provide legal advice or represent or warrant that its services or products will ensure that the customer is in compliance with any law. Notices and Disclaimers (con’t) Information concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products in connection with this publication and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. IBM does not warrant the quality of any third-party products, or the ability of any such third-party products to interoperate with IBM’s products. IBM EXPRESSLY DISCLAIMS ALL WARRANTIES, EXPRESSED OR IMPLIED, INCLUDING BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. The provision of the information contained herein is not intended to, and does not, grant any right or license under any IBM patents, copyrights, trademarks or other intellectual property right. • IBM, the IBM logo, ibm.com, Bluemix, Blueworks Live, CICS, Clearcase, DOORS®, Enterprise Document Management System™, Global Business Services ®, Global Technology Services ®, Information on Demand, ILOG, Maximo®, MQIntegrator®, MQSeries®, Netcool®, OMEGAMON, OpenPower, PureAnalytics™, PureApplication®, pureCluster™, PureCoverage®, PureData®, PureExperience®, PureFlex®, pureQuery®, pureScale®, PureSystems®, QRadar®, Rational®, Rhapsody®, SoDA, SPSS, StoredIQ, Tivoli®, Trusteer®, urban{code}®, Watson, WebSphere®, Worklight®, X-Force® and System z® Z/OS, are trademarks of International Business Machines Corporation, registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at: www.ibm.com/legal/copytrade.shtml. Thank You Your Feedback is Important! Access the InterConnect 2015 Conference CONNECT Attendee Portal to complete your session surveys from your smartphone, laptop or conference kiosk. Backup slides How to measure the gains • Configuration: • Server • Client IBM x3550 M2 - Intel ® Xeon ® CPU X5570 @ 2.93 Ghz 2 x 4 cores x 2 threads ( 16 execution threads) 52 GB RAM. Microsoft ® Windows® Server 2008 WebSphere Application Server 8.5.5.03(64 bit) IBM JVM Java 7.1 SR1 JVM Heap size from 1GB to 8GB • Rule Benchmark • • • • • • Eight sizes of ruleset from 300 to 14560 rules From 80 to 2000 rules fired per execution. Decision tables only Ruleflow of 5 rule tasks Java or XML XOM JSE, POJO (JEE), Web Service or REST invocation 30 Acronyms for this presentation • CRE: Classic Rule Engine • DE: Decision Engine • RES: Rule Execution Server • XOM: eXecutable Object Model • HTDS: Hosted Transparent Decision Service • As web service or REST service 31