IT Security - Software and Systems Engineering
Transcription
IT Security - Software and Systems Engineering
Workshop Hot-Spots der Software-Entwicklung IT Security 19. Februar 2013 Technische Universität München Institut für Informatik Software & Systems Engineering Prof. Dr. Dr. h.c. Manfred Broy Florian Deißenböck Daniel Méndez Fernández BICC-NET Bavarian Information and Communication Technology Cluster Inhaltsverzeichnis 1 Einleitung 3 2 Teilnehmerliste 4 3 Programm 5 4 Das Sicherheitsnetzwerk München Peter Möhring 6 5 Sicherer Browser – Schutz des Einfallstors Reto Weber 21 6 Einsatz von Zertifikatssystemen im Internet Jamshid Shokrollahi 42 7 Chipkartenbetriebssysteme – Gefahrenpotentiale und Gegenmaßnahmen Helmut Scherzer 51 8 Cybersecurity-as-a-Service: strategische und technische Herausforderungen Philipp Müller, PeterRehäusser 66 2 1 Einleitung Durch die fortschreitende Digitalisierung und Vernetzung von Software-intensiven Systemen und Diensten sowie die daraus resultierende Bedrohung durch Angriffe nimmt das Thema IT Security eine immer zentralere Rolle ein. Ein besonderes Augenmerk gilt dabei der Sicherstellung der Verfügbarkeit, Integrität und Vertraulichkeit von Software-Systemen und deren Infrastruktur. Diese Aspekte müssen in allen Phasen des Software-Entwicklungsprozesses unter Einsatz geeigneter Methoden und Verfahren berücksichtigt werden. Betroffen hiervon sind nicht nur klassische betriebliche Informationssysteme, wie sie beispielsweise im Finanzdienstleistungssektor vorzufinden sind, sondern auch eingebettete Systeme, z.B. in den Bereichen Automobil und Avionik. Ziel dieses Workshops ist es, zum besseren grundsätzlichen Verständnis des Themas IT Security beizutragen und konkrete Erfahrungen aus der Praxis bzgl. des erfolgreichen Einsatzes unterschiedlicher Verfahren auszutauschen, aber auch über Erfahrungen mit dem Einsatz von Verfahren, die sich als weniger geeignet erwiesen. Themen sind unter anderem: Strategische und technische Herausforderungen der IT Security Überblick über Konzepte und Methoden zur Lösung derselben Einsatz von Zertifikatssystemen Cybersecurity & Zugangskontrolle 3 2 Teilnehmerliste Josef Wernke, Eurocopter Deutschland GmbH Helmut Scherzer, Giesecke & Devrient GmbH Stefan Finkenzeller, BLB Dr. Jamshid Shokrollahi, Bosch GmbH Thomas Mey, Münchner Rück Peter Möhring, BICCNET-Clusterbüro I&K Dr. Florian Deißenböck, Technische Universität München Dr. Philipp Müller, CSC Reto Weber, Consecom AG Thomas Schön, Software Tomography Dr. Claudia Salazar Dorn, NTT Data Deutschland GmbH Martin Luy, ESG GmbH Roman Kochanek, Audi Helga Stephan-Dreinhoff, IBM Stefan Kassal, MaibornWolff et al GmbH Stefan Prechtl, ESG GmbH Jakob Tewes, MaibornWolff et al GmbH Alexander Bluhm, Gesellschaft fŸr Netzwerk- und Unix-Administration mbH Kurt Meindl, Lorenz Software GmbH Norman Thomson, ATOSS Olaf Kaudelka, EADS Gabriele Käsberger-Hoschek, EADS Dr. Heinrich Hördegen Ingenieurbüro Guttenberg & Hördegen Dr. Martin Wechs, BMW Group Jan Philipps, Validas AG Manuel Then, Technische Universität München Bertram Janositz, CIBOteam eSolutions AG Ovidiu Stan, ATOSS Dr. Philipp Guttenberg, Ingenieurbüro Guttenberg & Hördegen Bernhard Weber, msg systems ag Carsten Genth, ASM Assembly Systems Michael Spreng, Arcor AG & Co. KG Prof. Dr. Reiner Hüttl, Fachhochschule Rosenheim Michael Schulz, EADS Walter Trapa, BMW Group Michael Greulich, Interface AG Rainer Bitzer, Bosch GmbH Dr. Oscar Slotosch, Validas AG Ümit Kusdogan, ABSC GmbH Klaus Lochmann, Technische Universität München Carsten Tauss, ABSC GmbH Nils Oppermann, Audi Electronics Venture GmbH Peter Rehäusser, CSC Dr. Daniel Méndez, Technische Universität München Christine Rittinger, Münchner Rück Christopher Schulz, SYRACOM Consulting AG Tomas Benes, OSD Open Systems Design GmbH 4 3 Programm 13:30 Begrüßung Manfred Broy, Technische Universität München 13:45 Das Sicherheitsnetzwerk München Peter Möhrung, Sichernetzwerk München 14:30 Sicherer Browser – Schutz des Einfallstors Reto Weber, Consecom AG 15:15 Kaffee-Pause 15:30 Einsatz von Zertifikatssystemen im Internet Jamshid Shokrollahi, Robert Bosch GmbH 16:15 Chipkartenbetriebssysteme – Gefahrenpotentiale und Gegenmaßnahmen Helmut Scherzer, Giesecke & Devrient 17:00 Kaffee-Pause 17:15 Cybersecurity-as-a-Service: strategische und technische Herausforderungen Philipp Müller, PeterRehäusser, CSC 18:00 Abschlussdiskussion 18:30 Empfang 5 4 Das Sicherheitsnetzwerk München Peter Möhring Sicherheitsnetzwerk München Peter Möhring 19. Februar 2013, TU München, Garching 6 Vorgeschichte • • • • • • • • BMBF Spitzenclusterwettbewerb 2011 AISEC & G&D initiieren Netzwerk Skizze und Strategie Nominierung scheitert Januar 2012 Positive Netzwerkeffekte, Relevanz der Thematik Fortsetzung des Clusters Unterstützung durch Bay. Wirtschaftsministerium Einrichtung einer Geschäftsstelle am 1. Oktober 2012 Sicherheitsnetzwerk München, 19. Februar 2013 Forschung Sicherheitsnetzwerk München, 19. Februar 2013 7 Forschung Industrie Sicherheitsnetzwerk München, 19. Februar 2013 Forschung Industrie Anwender Sicherheitsnetzwerk München, 19. Februar 2013 8 Bündelung von Innovationskompetenzen Sicherheitsnetzwerk München, 19. Februar 2013 Warum dieses Netzwerk? • IKT als Innovationstreiber • IT-Sicherheit ist ein Wirtschaftsfaktor und schafft neue Märkte • IT-Sicherheit unterstützt relevante Exportindustrien • Vertrauenswürdigkeit, Manipulationsschutz, Wahrung der Privatsphäre, Verläßlichkeit in den Anwendungen notwendig • München hat höchstes Wirtschafts- und Forschungsniveau in Deutschland und Europa im Bereich IT-Sicherheit • Sicherheitstechnologien „Made in Germany“ als langfristiger Wettbewerbsvorteil deutscher Anbieter Sicherheitsnetzwerk München, 19. Februar 2013 9 Sicherheitsnetzwerk München, 19. Februar 2013 Warum dieses Netzwerk? • IKT als Innovationstreiber • IT-Sicherheit ist ein Wirtschaftsfaktor und schafft neue Märkte • IT-Sicherheit unterstützt relevante Exportindustrien • Vertrauenswürdigkeit, Manipulationsschutz, Wahrung der Privatsphäre, Verläßlichkeit in den Anwendungen notwendig • München hat höchstes Wirtschafts- und Forschungsniveau in Deutschland und Europa im Bereich IT-Sicherheit • Sicherheitstechnologien „Made in Germany“ als langfristiger Wettbewerbsvorteil deutscher Anbieter Sicherheitsnetzwerk München, 19. Februar 2013 10 Sicherheitsnetzwerk München, 19. Februar 2013 Industriegetrieben mit wirtschaftlichem Fokus • Integrierte Erforschung, Entwicklung und schnelle Vermarktung innovativer Sicherheitstechnologien und Produkte „Made in Germany“ • Ausbau von Weltmarktstellung der beteiligten Unternehmen • Deutschland mit München zum weltweit führenden Standort im Bereich IT-Sicherheit entwickeln Sicherheitsnetzwerk München, 19. Februar 2013 11 Sicherheitsnetzwerk München, 19. Februar 2013 Umfeld des Clusters • Politik • IT Gipfel • Underground economy • Digitale Gesellschaft • Mobilgeräte und Vernetzung • Neue Geschäftsmodelle • Sicherheitskonferenz Sicherheitsnetzwerk München, 19. Februar 2013 12 Geschäftsstelle: Schwerpunktziele • Aufstellung und Koordinierung von F&E Kooperationsprojekten • Standort- und Branchenstärkung • Netzwerkarbeit • Übergreifende Themen: Forschung, Fachkräfte, Trends,… Sicherheitsnetzwerk München, 19. Februar 2013 Maßnahmen 2013 • Clusterkonferenz: Ermittlung neuer Kooperationspotenziale • Kompetenzübersicht aller Mitglieder • Anbahnung von Fördervorhaben, Konsortienbildung • Platzieren von Themen in Förderprogrammen • Bildung von Arbeitskreisen • Schaffung einer Kommunikationsplattform • Neue Partnerschaften (auch international) • Schaffung informeller Austauschmöglichkeiten • Beeinflussung politischer Gestaltungsaufgaben • Gewinnung neuer Mitglieder Sicherheitsnetzwerk München, 19. Februar 2013 13 Projektvorhaben Sicherheitsnetzwerk München, 19. Februar 2013 Verbundprojekte 1. 2. 3. 4. 5. 6. 7. SIBASE ICEMAN SIKOMFAN Ambient security (neu) Trust ME Secure Appstore (neu) Lagebild (neu) Sicherheitsnetzwerk München, 19. Februar 2013 14 Schwerpunkt Mobile Endgeräte Sicherheitsnetzwerk München, 19. Februar 2013 Mobile werthaltige Dienste Sicherheitsnetzwerk München, 19. Februar 2013 15 Eisattacke (cold boot) Sicherheitsnetzwerk München, 19. Februar 2013 TEE Sicherheitsnetzwerk München, 19. Februar 2013 16 Sichere eingebettete Systeme Sicherheitsnetzwerk München, 19. Februar 2013 Schutz kritischer Infrastrukturen Sicherheitsnetzwerk München, 19. Februar 2013 17 Sicheres Cloud Computing Sicherheitsnetzwerk München, 19. Februar 2013 Anwendungsorientierte Technologien Sicherheitsnetzwerk München, 19. Februar 2013 18 Verbundprojekte: Zukünftige Fördermöglichkeiten LANDESEBENE • Hoher Freiheitsgrad, Direktbeantragung auch ohne calls • Kleine, effiziente Konsortien, geringeres Projektvolumen • Industrieorientiert (wenn von Bay. WiMi gefördert) BUNDESEBENE • Geringerer Freiheitsgrad durch vorgebene calls • Größere Konsortien mit höherem Projektvolumen möglich • Eher forschungsorientiert (speziell BMBF) EU-EBENE • Horizon 2020 • Internationale Konsortien, hohes Projektvolumen möglich Sicherheitsnetzwerk München, 19. Februar 2013 Verbundprojekte: Organisation, Rolle der Geschäftsstelle ORGANISATION • Konsortien mit eigener Projektkoordination • Konsortien sind für Beantragung und Durchführung der Verbundprojekte selbst verantwortlich ROLLE DER GESCHÄFTSSTELLE • Unterstützung in Anbahnung, Beantragung und Durchführung • Schaffung von Projekttransparenz für alle Mitglieder • Plattform und Drehscheibe für Mitgliederkoordination, sowohl inhaltlich als auch organisatorisch Sicherheitsnetzwerk München, 19. Februar 2013 19 Fragen Sicherheitsnetzwerk München, 19. Februar 2013 20 5 Sicherer Browser – Schutz des Einfallstors Reto Weber Bleicherweg 64a, CH-8002 Zürich, +41-44-515-0000 21 Experienced IT Risk/Security Professional Consecom AG Senior Security Consultant Reto Weber [email protected] +41 44 515 00 03 Employment History IT Risk Officer, Credit Suisse AG CERT Analyst, Credit Suisse AG IT Security Engineer, UBS AG Education CAS in Risk Management, University of Zürich eMBA in International Management, Kalaidos Zürich Master in Information Technology, Bond University Australia 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 2 Consecom – your partner for securing technology Clients Main clients: SME, and major Swiss/international enterprises and organization of all sectors. Location of work: primarily in Switzerland and neighboring countries. Design Services History 19.02.2013 Build Review concepts, strategies, policies, organization, processes, secure solutions programming, integration, special engineering audits, security reviews, risk assessments, penetration tests, technology assessments, organizational and process reviews Founded in 2007 as a management buy out. Privately owned, substantial growth to seven employees today. Consecom AG -- We Secure Your Solutions Slide 3 22 Agenda Problem Method Analysis Solution Experience Objectives Present lesson learned from a joining development with a University Exchange of experience of product development lifecycle Show result: a secure browser solution 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 4 Global news about virus infections Unbekannte haben einem Bericht von Amorize zufolge zahlreiche Onlineshops mit einer veralteten Version von osCommerce zur Verbreitung von Schadcode missbraucht. Die Angreifer nutzten mindestens drei bekannte Schwachstellen in der Version 2.2. heise.de 08/2011 Earlier this month, security researchers discovered a new piece of malware had infected more than half a million Apple computers in what was the largest-scale attack on Apple’s Mac OS X operating system to date. Nytimes, 04/2012 A serious flaw in the Java software found on most personal computers could expose the machines to being taken over by malicious attacks over the internet, the US agency responsible for policing such vulnerabilities warned on Thursday. ft, 01/2013 A new piece of Mac malware has been discovered on a Web site linked to the Dalai Lama, using a well-documented Java exploit to install a Trojan on visitors' computers and steal personal information. cnet.com 12/2012 Hackers are increasingly targeting childfocused gaming websites, according to a leading anti-virus firm. Avast says it detected malware threats at more than 60 sites that contained "game" or "arcade" in their title, in the 30 days running up to 12 January. 01/2012 bbc.co.uk It’s a scenario security researchers have long worried about, a man-in-the-middle attack that allows someone to impersonate Microsoft Update to deliver malware disguised as legitimate Microsoft code to unsuspecting users. wired.com 06/2012 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 5 23 Malware is today one of the major threats and frequently used to attack Threat Agent: No clear pictures of attacking agents. Educated guesses possible e.g. profit driven, organized crime. Threat Facts: Malware shows an exponential growth since years. Various method of infection vectors: mail, USB, social engineering, remote exploit, web-browsing. Impact: Massive infection rate on end-user desktop environments. Key question, are you at risk or not (large unreported cases)? Results: Breaches in Confidentiality, Integrity and Availability Multiple second order effects 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 6 Most infections come through the browser channel Infected systems: Web technologies are mostly used to place malware. Source Microsft.com 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 7 24 Drive – By – Infection is a common issue 1 Open page 2 Send Exploit 3. Compromise 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 8 25 Problem Methods Analysis Solution Experience 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 10 Multiple areas with limited influence Measure Feasibility / Applicability Reduce complexity in web-pages No governance; not applicability Improve security with web-pages No governance; limited reach Extend to cloud-AV solutions Partial success; confidentiality breach Change user-behavior Only partially applicable Reduce EuP flexibility and usability Change in working model; limited reach Myth: Keyboard encryption There is no keyboard encryption! EuP: End-user Platform 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 11 26 Where can we start Conclusion: Start with the web-browser Browsers, an integral part of the operating system. Provide a platform based on international standards/languages (HTML, http, CSS, JavaScript, …). Having a super secure browser would reduce the risk of infection dramatically. 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 12 EuP: End-user Platform 27 Problem Methods Analysis Solution Experience 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 14 The security product has multiple dimentions Dimensions Criteria Trust Administrator and manufacturer trust (prerequisite) Life-cycle trust (deployment and update) Threat Coverage On-line (in-bound): Vulnerabilities in applications Off-line threats: Malicious host At run-time At rest Protection Methods to protect from infected hosts Isolation Methods to isolate applications from infected hosts Integration Capabilities to integrate into standard work process Deployment Method and scalability Maintenance Patch and update capabilities, and scalability Usability Technological user support Weaknesses Limitations and weaknesses 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 15 28 EuP: End-user Platform Overview Secure Browser Browser on USB Stick, Hardened Web-browsers Browser Sandbox, Hardened Web-browsers Browser on Native Boot-Platform Browser in VM VM 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 16 A USB stick or «wrapper» approaches Design Paradigm Run on shared commodity platform Hardened Web-Browser Interrupt Vectors Memory Map Exemplary implementation variants Application Commodity Operating System Share of central operating system tables Application Hardened Web-Browser USB-stick based web-browser Application Application Windows Hardened web-browser by platform extension, e.g. Trusteer Rapport® 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 17 29 Main weakness is a shared platform Dimensions Criteria Details Trust Life-cycle trust High trust for stick rollout, common otherwise Threat Coverage On-line (in-bound) Off-line threats: Malicious host At run-time At rest Gradual improvement against known threads (Low) gradual improvement Only if stored on read-only medium Protection Capabilities Application internal only Isolation Method None Integration Capabilities Stick: limited; Platform: full Deployment Method and scalability Stick: shipment; Platform: software Maintenance Method and scalability Stick: replacement, both: incremental updates Usability Support by technology Stick: non-IE; browser independent Weaknesses Limitations Shared platform Web-browser only protection 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 18 A USB stick or CD boot systems Design Paradigm Requires respective interface Isolation Controlled Application 1 Controlled Application 2 Isolation LPS Boot-off clean OS from dedicated media Controlled Application 3 Hardened OS User Space SELinux Mandatory Access Control Provides conceptual improvement only with trusted read-only media Minimal Linux Kernel stateful firewall Exemplary implementation variants c’t bankix: Linux Knoppix (Debian)-based boot-CD Lightweight Portable Solution (LPS) by US DoD: Hardened Linux for remote access 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 19 30 Usability and mobile technology makes it harder to use Dimensions Criteria Details Trust Life-cycle trust No protection of image at download Threat Coverage On-line (in-bound) Off-line threats: Malicious host At run-time At rest Platform hardening Protection Capabilities Hardened Linux platform Firewall Isolation Method Native boot Integration Capabilities None (remote access only) Deployment Method and scalability CD image download Maintenance Method and scalability None, full image deployment Usability Support by technology Base-installation too complicated Weaknesses Limitations No integration, no incremental maintenance no image-protection 19.02.2013 No exposure. If stored on read-only medium Consecom AG -- We Secure Your Solutions Slide 20 A VM system to browse resolves a lot problems. Design Paradigm Run web-browser in a VM Application 1 (Browser) Isolate web-browser from host Application 2 (PDF-Reader) Application 3 (Flashplayer) Debian Linux Virtual Machine constrained disk access Virtualization layer Isolation Windows Separate OS Exemplary implementation BitBox – Browser in the Box – by German BSI/Sirrix AG 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 21 31 Generally OS independent Dimensions Criteria Details Trust Life-cycle trust Update-only Threat Coverage On-line (in-bound) Off-line threats: Malicious host At run-time At rest Technical improvement (Standard Linux) Gradual improvement Only if stored on read-only medium Protection Capabilities Isolation by VM, controlled read-write Isolation Method Platform virtualization by VMs Integration Capabilities Integrated for default-browser launch Constrained data exchange Deployment Method and scalability Full image deployment Maintenance Method and scalability Standard Debian update mechanisms Usability Support by technology Base-installation too complicated Weaknesses Limitations Protection by standard Linux combined with Oracle VirtualBox abstraction 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 22 32 Problem Methods Analysis Solution Experience 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 24 Security Enhanced – Linux the answer to the web-infection problem Objective Protect a system against unauthorized access/execution Function A technical policy enforcement framework. Any operation executed is validated by a security filter prior to execution. A kernel module is compiled into the machine with predefined rules. The Implementation kernel’s security filter preforms the checks. No one can overwrite the rules at run-time (since denied by the security filter) and compiled into the system. History SELinux was originally a development project by the National Security Agency (NSA). It is an implementation of the Flask operating system security architecture. The Flask architecture defines MAC with focus on providing an administratively-defined security policy that can control all subjects and objects, basing decisions on all security-relevant information. Flask was then renamed to SELinux. Ref www.nsa.gov/research/selinux/ 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 25 33 A combination of methods into a browsing platform SEBPS Controlled Application 2 (Acroread) Isolation SEBPS Controlled Application 1 (Firefox) Isolation Trusted Update Server SEBPS Controlled Application 3 (Flashplayer) Scalable Trusted Maintenance Hardened OS User Space SELinux Mandatory Access Control Minimal Linux Kernel stateful firewall Virtual Machine constrained disk access Isolation Commodity Operating System 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 26 Multiple benefits and usability for end users are important Deployment Web-Installer abstracts installation complexity Prepare the platform: pre-required tools Reduce the installation to the minimally needed steps: “Two Clicks to secureBrowse” Use Support the user Launch the Web-browser Support system hibernate and standby Logout shuts down secureBrowse Administration Relieve the user from interfering with administrative tasks 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 27 34 secureBrowse -- the answer to the web-infection problem Dimensions Criteria Details Trust Life-cycle trust • Threat Coverage On-line (in-bound) Off-line threats: Malicious host At run-time At rest • Platform hardening (MAC) Protection Capabilities • Hardened Linux platform (MAC) Isolation Method • Virtualized platform Integration Capabilities • Cut-and-Paste between host and platform Deployment Method and scalability • Trusted web-based Installer Maintenance Method and scalability • Trusted incremental updates Usability Support by technology • Straightforward web-supported installer Standard applications Weaknesses Limitations • Limited integration 19.02.2013 Full trust: Deployment, Maintenance • Separate OS-addr- space, separate IVT • If stored on read-only medium Consecom AG -- We Secure Your Solutions Slide 28 35 36 37 Problem is reality, solution is around. Companies are daily in the news because of incidents with malware. Moreover, old browsers are still used. Where and when infections happen really, remains undiscovered. The main entry gate, the Web-Browser, can be secured. Non of the frequently used standard products proves complete protection. A VM with Mandatory Access Control Browser is very secure approach. A solution is around and free to use. http://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 35 38 Problem Methods Analysis Solution Experience 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 36 Solving a problem doesn't make you millionaire but the knowledge brings you forward Product never reached the mass market (by now) Generally users and companies still accept the risk by drive-by infections. Windows / iOS look and feel is absolutely essential. Usability change require a large investment and are not accepted. Nish market Forensic Investigators, Fraud Analyzers (Analyze Dangerous Websites). Selective usage for untrusted web-pages (i.e. all). Highly secured areas (Nuclear Power Plants, Military). New business by knowledge gain Expert in secure platforms. Providers of secure platform for Web Entry Systems. Provider of secure appliances for major financial institutes. 19.02.2013 Consecom AG -- We Secure Your Solutions Slide 37 39 40 41 6 Einsatz von Zertifikatssystemen im Internet Jamshid Shokrollahi Certificates: How to Build Trust in the Internet J Jamshid hid Sh Shokrollahi, k ll hi Corporate Research (CR/AEA) Robert Bosch GmbH 1 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 42 Certificates: How to Build Trust in the Internet Overview 2 Solved problem (Public Key Infrastructure and Certificates) Symmetric Key Cryptography Public P bli kkey C Cryptography t h Man in the Middle attack X.509 Certificate Not Completely solved problem (Secure Deployment) Different levels of realization Potential Vulnerabilities Department | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Mission Alice and Bob want to securely communicate in the presence of Eve and Mally! Eve: eavesdropping Mally: read, write, and modify the messages 3 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 43 Certificates: How to Build Trust in the Internet Bob shops in the Internet! What can go wrong? 4 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Eavesdropping Eve sees Bob’s credit card number and uses it next time for shopping! What is the solution? 5 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 44 Certificates: How to Build Trust in the Internet Symmetric Encryption is a solution Enc. / Enc. Dec.. Dec 103 40... Enc.. 1034 Enc Dec.. 0... /Dec Communication partners have the secret common key 6 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Public Key Cryptography (PKC) Bob and the Server can establish the secret Key using asymmetric Encryption Bob can encrypt its symmetric key using Server’s public key, but only the server has the private key to decrypt itit. In practical realizations there are more details which are ignored here 7 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 45 Certificates: How to Build Trust in the Internet Verifying Public Keys How can Bob be sure that Mally is not performing “man in the middle” attack? Mally’s public key 8 Server’s public key Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Public Key Infrastructure CA(Certification ( Authority) CA’s public key is embedded into most of browsers Secure Connection • Signed by CA Server’s private ke ey • Server’s identity • Server’s S ’ public bli kkey: Certificate, e.g., X.509 User‘s browser 9 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 46 Certificates: How to Build Trust in the Internet Example: X.509 Certificate Structure (1) • Version • ... • Serial Number • Certificate Signature Algorithm • Algorithm Al ith ID •C Certificate tifi t Signature Si t • Issuer (E.g. Certificate Authority) • Validity • Not Before • Not After • Subject (E.g., server) • Subject Public Key Info • Public Key Algorithm • Subject Public Key • Issuer Unique Identifier (optional) • Subject Unique Identifier (optional) • Extensions (optional) 1) Source: Wikipedia 10 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Server Authentication and Key Exchange (simplified) 8) Ensures the security by seeing https 11 Browser 2) Provides the certificate 4) Generates a random 128 bit key, encrypts t using i the th public bli kkey iin th the certificate and send to the server 7) XORes the two sequences to generate the symmetric key 5) Generates a random 128 bit key and sends to the browser (client) in plain 6) De ecrypts the e received packet, XOR Res the two o sequences to gene erate the syymmetric key 1) https: request a secure connection 3) Verifies the signature of the certificate, and if the subject matches the server Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 47 Certificates: How to Build Trust in the Internet Realization Aspects Issuer’s Infrastructure API Level Digital Signatures 12 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Digital Signature Hash Function • Nonlinear mapping pp g to compress large messages • Must be collision resistant Padding md mod N Adding g specific patterns to the compressed message The fundamental and most time consuming operation Nonlinearity cancels the multiplicative property of the exponentiation 13 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 48 Certificates: How to Build Trust in the Internet Digital Signatures, lessons learned • • • • 14 Do not use hash functions like MD4 with known collisions Always use large modulo numbers N which are generated according to th standards the t d d The parameters and functions get obsolete. Never issue the certificate for very long time Always choose the functions and parameters according to the most recent standards, e.g., BSI – Technische Richtlinie, BSI TR-02102, Version 2013.02 https://www.bsi.bund.de/SharedDocs/Downloads/DE/BSI/Publikationen/Te chnischeRichtlinien/TR02102/BSI-TR02102_pdf.pdf?__blob=publicationFile Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Vulnerabilities in using API • Always verify if the certificate matches the identity of the provider (server)? Even consider the possibility of having '\0' in the identity of the provider id off certificate tifi t who h wants t to t impersonate i t a famous f identity id tit • SSL and several other software also provide the PKI verification as f functions. Always read the documentation carefully f to enable the verification of the signatures Several examples of to dos and not to dos can be found in „Georgiev et al., The Most Dangerous Code in the World: Validating SSL Certificates in Non-Browser Software, In proceedings of ACM CCS '12, pp. 38-49, 2012“ 15 Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 49 Certificates: How to Build Trust in the Internet Issuer's Issuer s Infrastructure • • • • 16 Certificates are your identity and their security depends on the private key of the issuer If th the private i t kkey iis nott stored, t d or used, d iin th the right i ht way, attackers tt k can impersonate you Revocating certificates and the lost reputation can cause high costs! Always think about the reputation off the issuer and iff possible ensure that they have enough security in the processes Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. Certificates: How to Build Trust in the Internet Conclusion • • • 17 For the selection and the realization of digital signatures always use the most recent standards and follow them carefully Wh verifying When if i signatures, i t always l b be careful f l about b t th the id identities, titi and d also variables and flags if you use verification functions from others' libraries When buying the certificates f think about the reputation and history off the provider! Jamshid Shokrollahi, CR/AEA3 | 2/19/2013 | © Robert Bosch GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights. 50 7 Chipkartenbetriebssysteme – Gefahrenpotentiale und Gegenmaßnahmen Helmut Scherzer SmartCard Operating Systems Potential Risks and Security Measures Helmut Scherzer Giesecke & Devrient [email protected] 51 SmartCard Operating Systems Standards and Specifications P Standards < ISO 7816/1..9 < CEN TC224/EN726 < GSM 11.xx < ETSI < JAVA 2.x P Specifications < JAVA < EMV < Mondex < Multos < SECCOS(ZKA) < ...and others P Platform < ST Microelectornics < Infenion < Philips < Samsung < Sharp < Renesas < Atmel < ... and others P “Hidden Agenda” < Security Features < Attacks < Countermeasures < Programming Tricks < Performance Optimization < Memory Optimization < Algorithms PIN Attack Attack P Cut Off Power before updating the PIN error counter Correct PIN- Entry Wrong PIN-Entry PIN - Check Update of PIN error counter Attack: Detection and interruption 52 PIN Attack Countermeasures P Update PIN error counter prior to PIN verification Cnt = 2 PIN - Check Cnt = 3 Power Break Attack or Accident ... P Power Drop in the SmartCard < Unintentionally (Accident) – – – – Withdraw SmartCard Functional misbehaviour in Terminal Bad Contacts Environmental Factors (Vibration in Car etc...) < Intentionally (Attack) – Data manipulation 53 Power Break Intention of the attack EEPROM Write Cycle EEPROM 1.) Delete EEPROM cell EEPROM ................. ................. 0010 0110 ................. ................. ................. 2.) Write new value ................. ................. 0000 0000 ................. ................. ................. EEPROM ................. ................. 1111 1110 ................. ................. ................. Best possible moment of power cut Power Break Pseudo Random Attack EEPROM EEPROM Random No. EEPROM 00 00 00 00 00 00 00 00 New Random No. Random No. 35 BC 01 A7 48 D5 3B 1C Old Random No. DES New Random No. is written to EEPROM EEPROM Random No. AB 33 2F 8E 46 7A 29 FD New Random No. 54 Power Break Countermeasures P Limited Transaction Protection < Recognition and Indication of Power Breaks < Protection of sensitive data < Card Blocking P Full Transaction Protection < Backtrace / Write Ahead Buffer < Atomized Transaktionens < Data Committment P Countermeasures : “Absolutely MANDATORY” Power Break Command Message “.. ..03 45 .. ..” Target Data Backtrace Buffer 01 23 01 23 01 23 01 23 Full Transaction Protection ?? ?? 01 23 00 00 03 45 01 23 01 23 01 23 03 45 03 45 01 23 55 Memory Defragmentation Defragmentation by multiple file deletion EEPROM P Creation of files by EF < Additional Applications < JAVA - Applets < Version Control/Update < Application Deletion < Application Buy and Run < Temporary Files EF EF EF EF EF EF EF EF EF EF EF 1.) Initialization Free Memory EF 2.) Erase EEPROM EF EF EF EF EF EF EF EF Free Memory EF EF 3.) New layout EEPROM EF EF Memory Full ! EF EF EF EF EF EF EF EF Memory Defragmentation Defragmentation Process EEPROM EEPROM EF EF EF EF EF EF EF EF EF Defrag ... EF EF EF EF EF EF EF EF EF EF EF P Difficulties < References must be recorded < Defragmentation (up to 10 Sec.) must be 100 % Power Break resistant P Solution < Atomic operations with wear-leveling mechanism – Defragmentation possible very effectively – New Application Perspectives for SmartCards Smart Card Development P Auto-Defrag < 100 % Power brk.Prot. < High Performance ! < “on Auto-Demand” only P 100 % Fail resistant < Theoretically proovable 56 EEPROM Errors Few Bits dropping P Chip Mfg.Guaranty : 10 Years Data retension EEPROM < Credible and proven value < Very seldom, accidential drops have been reported 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1101 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 11111 1110111 11111 P Bit Drop Situation < Most embarassing situation as no obvious reason available < Murphy’s Law: “Always the most crucial bit drops" P Countermeasures < CheckSum on any EEPROM boundary < Update of Checksum must also be power break resistant Timing Attack PIN/Key Verification P Time pattern of current samples // Compare PIN for (i = 0; i<length(PIN); i++) { if (PIN[i] != Correct_PIN[i]) return(1); } return(0); Reference Point ) t Difference Measurement 57 Timing Attack The Square-Multiply 'always' problem P New programming styles rlc r2 ; r2 = exponent jnc NoSwap xchg r0,r1 NoSwap: ret SmartCard StoneAge 2013 Coding // evaluate exponent bit if (ExpBit == 1) Swap(Source,Target); else Keep(Source,Target); return(0); push push push mov r0 r1 r0 bp,sp rlc addc mov inc mov r2 ; r2 = exponent bp,#0 r0,[bp] bp r1,[bp] add ret sp,#3 ; <- bp ; ; remove '3 x push' Differential Fault Analysis The “DES” Bellcore Attack P Assumption: < “Bits in RAM may be altered intentionally” P Attack by comparison of output 0 Key: 0100 1010 1110 0111 1011 1111 0011 0110 ... .. ... “Alice is pretty” 1.) “Xy202 01aM b201” 2.) “Xy202 01aM b201” DES “a839 x15k b7fm” 58 Reaction of the market Decrease of Orders since June 1998 !!! 100 80 DPA-Attack ! 60 40 20 0 January March May July September November SmartCard Orders Single Power Analysis Direct Evaluation of Current Samples P Direct evaluation of Current Samples < Insider knowledge required < Program Code must be locally known P Countermeasures < No bit operation with sensitive data < Source Code “ CONFIDENTIAL “ < Support by Chip Hardware 59 Differential Power Analysis Attack Scenario P Statistical Attack Key Shift < Many DES Calculations required ! < No Source Code Knowledge required ! S-Box Output Perm. DES Round # n The Final Attack Correlate Signed Samples P If the hypothized key was correct, each calculation will contribute a deterministic part to the final signal P If the hypothized key was wrong, only 'noise' will be added and no singularity will be found Signal P For each n of 64 possible subkeys a particular hypothesis on the signal signs exist. P For each n of 64 possible subkeys the addition (correlation) of the N samples will be performed + - + Signal Signal + + + Noise + - = N Noise Noise = 3 = /N 60 Differential Power Analysis Attack Scenario High number of samples required ! Key Shift S-Box Output Perm. DES Round # n Subkey Subkey Subkey Subkey Subkey Subkey Subkey Subkey Key : 48 bits 010110 011101 110110 001011 010110 101110 001111 110001 Data 0110 S-Box S-Box S-Box S-Box S-Box S-Box S-Box S-Box 1101 0010 1011 0010 0101 1001 1000 0110 S-Box being attacked Attacked bit Finding the Key ... Maximum Evaluation P In one of 64 correlation signals we will find a significant maximum P This maximum confirms our 'guess' for the subkey. P We may confirm the guess by evaluation of the other three bits on the same S-Box .... .... Subkey 17/64 : wrong hypothesis Subkey 18/64 : wrong hypothesis Subkey 19/64 : wrong hypothesis Subkey 20 : CORRECT HYPOTHESIS ! 61 Differential Power Analysis Countermeasures P System Level < Limited Usage of Error Counters < Logging of Error Counters in Host System P Hardware Level < Bus Scrambling < Power Noise < Redundant Clock Cycles P Software Level < Relative Protection – Make the Alignment 'impossible' – Nonsens Statements < Absolute Protection – Theoretical Proove of Effectiveness – ITSEC Evaluation possible ! – IBM DPADES9 : “Excellent Protection !” P Critical Factor: “PERFORMANCE” Table compression Partitioning attack ROM CPU 256 Bytes page 256 Bytes page 256 Bytes page ... ; 512 db db db byte table 01,45,62,F3, 8F,7B,2A,3F, .... .... P Address and value of a (EEP)ROM table may leak DPA information P Addressing another (EEP)ROM page may leak SPA information P Countermeasures (srambled RAM table) requires too much memory (here 512 bytes) 62 Table compression Partitioning attack ROM CPU 256 Bytes page 256 Bytes page 256 Bytes page ... P Address and value of a (EEP)ROM table may leak DPA information P Addressing another (EEP)ROM page may leak SPA information P Countermeasures (srambled RAM table) requires too much memory (here 512 bytes) Table compression Large Table attack countermeasures ROM 128 Bytes page data + index masking RAM 128 Bytes 128 Bytes page 128 Bytes page ... P Overlay ROM table entries in RAM P index/value mask 'on the fly' P Decoding can only be done with help of the original ROM table, but this can be achieved in a well protected way. 63 Table compression Large Table attack countermeasures ROM Encode Generate rand1..rand4 [each 0..127] for (i = 1 to 128) { RAM[i] = ROM[i r rand1] r ROM[i r rand2 + 128] r ROM[i r rand3 + 256] r ROM[i r rand4 + 384] } 256 Bytes page 256 Bytes page 256 Bytes page ... Decode (e.g. value from 0..127) j = i r rand1 Decode (e.g. value from 128.255) j = (i-128) r rand2 ROM[i] = RAM[j] r ROM[j r rand2 + 128] r ROM[j r rand3 + 256] r ROM[j r rand4 + 384] ROM[i] = RAM[j] r ROM[j r rand1] r ROM[j r rand3 + 256] r ROM[j r rand4 + 384] RAM 128 Bytes End SmartCard Operating Systems Potential Risk and Security Measures 64 65 8 Cybersecurity-as-a-Service: strategische und technische Herausforderungen Philipp Müller, PeterRehäusser Cybersecurity-as-a-Service Dr. Philipp Müller Peter Rehäußer CSC Proprietary and Confidential 66 Hackerangriffe sind allgegenwärtig. Was nun? CSC Proprietary and Confidential March 25, 2013 2 Unsere Welt heute Zunehmende Digitalisierung von Geschäftsprozessen Die neue Art von Bedrohungen: Gezielter, strukturierter und mit uneingeschränkten Ressourcen für Angriffe Private und geschäftliche IT verschmelzen: Cloud, ByoD, Social Media, … Wachsende Anzahl an Regularien und Compliance-Vorschriften CSC Proprietary and Confidential March 25, 2013 3 67 Wie denken wir Sicherheit auf der Vorstandsebene? Wie denken wir Sicherheit gesamtgesellschaftlich? CSC Proprietary and Confidential March 25, 2013 4 March 25, 2013 5 Sicherheit ganzheitlich angehen: der Cybersecurity Stack Ebene 4: Nationale Sicherheitsstrategie Bedrohungsalarm Nation/Staat Ebene 3: Situationsbewusstsein EventKorrelation Organisation Ebene 2: Sicherheitsschicht Klassisches Vorgehen: Vorbeugen-AufdeckenReagieren Perimeter & Gateway Ebene 1: Sichere Infrastruktur LAN-WANApplikationen und -Daten Einzelne Systeme CSC Proprietary and Confidential 68 Den Security-Live-Cycle komplett abdecken Awareness Trainings IT Security Principles Methodology Coaching Security Policy IT Security Concepts IT Security Organisation IT Risk Management Implementation Standards Business Continuity Planning Cyberconfidence Check Controlling Coordination Communication Compliance with Security Policies Identity Management IT Security for Outsourcing Comprehensive Audits Managed Security Services Penetration Test Common Criteria Evaluations IdM Prequalification Checks CSC Proprietary and Confidential March 25, 2013 6 March 25, 2013 7 Warum Cybersecurity-as-a-Service? Cost CSC Proprietary and Confidential 69 The CSC Cybersecurity Demonstration Center CSC Proprietary and Confidential March 25, 2013 8 March 25, 2013 9 Storyline der Live-Demo Der Feind beauftragt einen Hacker, … … die Daten eines Großprojektes zu löschen und dem Unternehmen damit Schaden zuzufügen … CSC Proprietary and Confidential … und fortwährend zu überprüfen, ob die Organisation am Ende des Projektes in der Lage war, die Informationen wieder herzustellen. 70 Wie gehen Hacker vor? von Netzwerken 1 Analysieren • Potentielle Opfer? der potenziellen Opfer 2 Beobachtung • Wo kann man am einfachsten eindringen? 3 Angriff • Ausnutzen von Unachtsamkeit 4 Illegale Tätigkeit • Löschen von Projektdaten • Installation eines maßgeschneiderten Trojanischen Pferdes, das alle neu erstellten Dokumente sofort an den Hacker sendet 5 Escape • Der Remote-Zugriff wird unterbrochen CSC Proprietary and Confidential March 25, 2013 10 March 25, 2013 11 The CSC Cybersecurity Demonstration Center • Global Learning • Situational Awareness Threat Alerting Nation / States Event Correlation Organization • Real-time Detect and Respond • Data-loss prevention • End-point security Classical Prevent- Detect-Respond Perimeter & Gateway • Ongoing Vulnerability Analysis • Identity Management LAN-WAN-Applications and Data CSC Proprietary and Confidential Single Systems 71 VIELEN DANK FÜR IHRE AUFMERKSAMKEIT! HABEN SIE FRAGEN? CSC Proprietary and Confidential March 25, 2013 12 CSC Proprietary and Confidential March 25, 2013 13 72