Software Security Erik Poll Digital Security group Radboud

(Software) Security Erik Poll Digital Security group Radboud University Nijmegen

A brief history of security problems • attacks on multi-user UNIX systems for fun • viruses & worms attacking operating systems – due to buffer overflow, format string attacks, integer overflows, . . . – response: firewalls & better operating system code – variant: bot-nets • attacks on web-applications – due to SQL injection, XSS, session riding, . . – response: intrusion detection systems & application firewalls • attacks on web-browsers – man-in-the-browser attacks • . . . 2

A brief history of security problems • Trojans & spyware – installed via OS security hole, or silly user • spam – response: spam filters – variants: phishing, spear-phishing & whaling, spit • . . . 3

Trend • move away from hacking for 'fun' (ie vandalism) to hacking for profit – zero-day exploits are worth a lot of money 4

Causes Important causes of these security problems • bad software – buggy OS software, web-application software, . . . – solution: patching? • gullible & stupid users – too easy to just blame users: if they are gullible and stupid, they should be protected – solution: user education? 5

Security is about people (users, employees, sys-admins, programmers, . . . ), incompetence, confusion, lethargy, stupidity, mistakes, access control, passwords, biometrics, cryptology, protocols, policies & their enforcement, monitoring, auditing, legislation, persecution, liability, risk management, complexity, software, bugs, verification, hackers, viruses, hardware, operating systems, networks, databases, public relations, public perception, conventions, standards, physical protection, data protection, . . . 6

Security is about people (users, employees, sys-admins, programmers, . . . ), incompetence, confusion, lethargy, stupidity, mistakes, access control, passwords, biometrics, cryptology, protocols, policies & their enforcement, monitoring, auditing, legislation, persecution, liability, risk management, complexity, software, bugs, verification, hackers, viruses, hardware, operating systems, networks, databases, public relations, public perception, conventions, standards, physical protection, data protection, . . . 7

Cryptography is usually not the problem "Using encryption on the Internet is the equivalent of arranging an armored car to deliver credit card information from someone living in a cardboard box to someone living on a park bench. " -- Gene Spafford 8

Root cause analysis • Security is always a secondary concern – primary goal of software is to provide some functionality or services – managing the associated risks is only a derived or secondary concern • There is often a trade-off/conflict between – security – functionality & convenience 9

"The only system which is truly secure is one which is switched off and unplugged, locked in a titanium -lined safe, buried in a concrete bunker, and surrounded by nerve gas and very highly paid armed guards. Even then, I wouldn't stake my life on it” -- Gene Spafford 10

What goes wrong? • at a high-level – not understanding the security requirements and threats • at a low-level – silly little program bugs or flaws that introduce vulnerabilities 11

How can we improve things? • at a high-level – seriously looking at security requirements and threats when (prior to) building systems • at a low-level – train programmers to be aware of pitfalls associated with programming language, OS, platform – improve these languages, OSs, platform • making them less error-prone 12

low-level

Security as secondary concern • in programming languages – Algol 60 introduces array bound checks, in 1960 – C doesn't use this, in 1970 s. . . –. . . 3 decades later, we're still trying to get rid of buffer overflows – early 2000 s: people start using safestr. h 14

Security as secondary concern • in training – many students learn programming in C(++) – nobody tells them about buffer overflows or safestring libraries – a case of criminal negligence? 15

Functionality vs security Lost battles? • operating systems – huge OS, with huge attack surface (API), • programming languages – buffer overflows, format strings, . . . in C – lots of things in PHP • webbrowsers – plug-ins for various formats, javascript, VBscript, . . . • email clients –. . . 16

Software infrastructure Applications are built on top of very complex "infrastructure" consisting of • operating system • programming language/platform/APIs/middleware • other applications & utilities – eg SQL database This infrastructure provides security mechanisms, but is also a source of security pitfalls. You need to understand both to develop secure applications on a give infrastructure. 17

Threats & vulnerabilities • Knowledge about threats & vulnerabilities crucial – my first dynamic webpage, with a cgi-bin script, was hopelessly insecure • Vulnerabilities can be specific to programming language, operating system, database, . . . used, and are continuously evolving, so we cannot hope to cover all vulnerabilities. • “Fortunately” people keep making the same mistakes – and the same security problems keep reappearing 18

The bad news: people keep making the same mistakes • when writing programs & when designing programming languages or other infrastructure • SQL injection – string concatenation is a convenient way to build SQL queries. . . • will we be chasing SQL injection faults in 30 years, just as we're still chasing buffer overflow attacks? • insist on use of eg. Prepared. Statement? • PHP – PHP is a convenient way to quickly build a website. . . 19

Functionality vs security : PHP ". . . , I've come to the conclusion that it is basically impossible for normal programmers to write secure PHP code. It takes far too much effort. . . PHP's raison d'etre is that it is simple to pick up and make it do something useful. There needs to be a major push. . . to make it safe for the likely level of programmers - newbies. Newbies have zero chance of writing secure software unless their language is safe. " [http: //www. greebo. cnet] 20

The "good" news: people keep making the same mistakes • We can make checklists for common mistakes • We can teach students about common mistakes • We can implement tools that check them – source code analysers aka static analysis tools • Static analysis tools for C(++) – Coverity, Fortify, j. Test, Poly. Space, PREfast, PREfix, . . . • C/C++ checkers focus on memory-related issues and for Java – Check. Style, Findbugs, PMD, Fortify, j. Test, Intelli. J, . . . 21

OWASP Top 10 of web application vulnerabilities 1. Unvalidated Input 2. Broken Access Control Anyone building a web 3. Broken Authentication and Session Management application should know 4. Cross Site Scripting (XSS) Flaws about these vulnerabilities: 5. Buffer Overflows • how they work, 6. Injection Flaws • how they arise, and 7. Improper Error Handling • how they can be avoided 8. Insecure Storage 9. Denial of Service 10. Insecure Configuration Management Lots of info on web-application security on OWAPS. org 22

19 Deadly sins of software security 1. 2. 3. 4. 5. 6. 7. 8. buffer overruns format string problems integer overflows SQL injection command injection failing to handle errors XSS failing to protect network traffic 9. use of magic URLs or hidden form fields 10. improper use of TLS, SSL [Howard, Le. Blanc, Viega] 11. weak passwords 12. failing to store & protect data securely 13. information leakage 14. improper file access 15. trusting network name resolution 16. race conditions 17. unauthenticated key exchange 18. weak random numbers 19. poor usability 23

Classification of Software Security Errors 1. Input Validation and Representation 2. API Abuse 3. Security Features 4. Time and State 5. Errors 6. Code Quality 7. Encapsulation *. Environment • [Katrina Tsipenyuk, Brian Chess, Gary Mc. Graw, Seven Pernicious Kingdoms: A Taxonomy of Software Security Errors] 24

Two sides to software security: do’s & dont’s • What are the methods and technologies, available to us if we want to provide security? – security in the software development lifecycle – engineering & design principles – security technologies • What are the methods and technologies available to the enemy who wants to break security ? ie. what are threats and vulnerabilities we’re up against 25

How to improve software security? • Awareness & knowledge of vulnerabilities (don'ts) – general (input validation, . . . ) – specific to a kind of application (SQL injection, XSS, . . . ), or – specific to a kind of programming language (buffer overflows, . . . ) • Awareness & knowledge of countermeasures (do's) at different points in the development lifecycle at level of application, programming language, or platform Eg security technologies such as – access control, session management – untrusted code security • type-safe languages, sandboxing, code-based access control – runtime monitoring – program analyses: typing, information flow, static analysis, verification But beware that security software ≠ software security 26

high-level

Security concepts availability/ usefulness want to maximise owners impose [as used in Common Criteria] want to minimise countermeasures reduce may have require attackers give rise to increase vulnerabilities exploit threats lead to increase to of risks assets want to abuse 28

Starting point for ensuring security • Any discussion of security should start with an inventory of – the stakeholders, – their assets, and – the threats to these assets by possible attackers – employees, clients, script kiddies, criminals • Any discussion of security without understanding these issues is meaningless 29

Threat Modeling Aka security/risk/requirements analysis A first step, not just for software • Identify assets & stakeholders • Consider architecture of application & its environment • Brainstorm about known threats • Define security assumptions • Rank threats by risk ≈ impact x likelihood • Decide which threats to respond to • Decide how to migitate these threats – which techniques & technologies 30

Countermeasures and more vulnerabilities Countermeasures can lead to new vulnerabilities • eg. if we only allow three incorrect logins, as a countermeasure to brute-force attacks, which new vulnerability do we introduce? If a countermeasure relies on software, bugs in this software may mean • that it is ineffective, or • worse still, that it introduces more weaknesses 31

STRIDE Simple approach to threat modelling developed at Microsoft, as part of SDL (Secure Development Lifecycle process) • Spoofing Identity • Tampering with Data • Repudiation • Information Disclosure • Denial of Service • Elevation of Privilige These and other checklists can be useful See also http: //msdn 2. microsoft. com/en-us/library/aa 302419. aspx 32

Techniques to migitate these threats • Spoofing Identity – authentication, protect keys & passwords, . . . • Tampering with Data – access control, hashes, digital signatures, MACs (message authentication codes), write-once storage. . . • Repudiation – logging, audit trails, digital signatures, . . . • Information Disclosure – access control, encryption, not storing secrets, . . . • Denial of Service – graceful degradation, filtering • Elevation of Privilige – access control, esp. least privilige, sandboxing, . . . 33

Security in Software Development Life Cycle [Source: Gary Mc. Graw, Software security, Security & Privacy Magazine, IEEE, Vol 2, No. 2, pp. 80 -83, 2004. ] 34

Security principles • keep it simple • secure the weakest link • defence in depth • principle of least privilige • minimise attack surface • compartementalise • secure defaults • identify & question your assumptions • fail securely • promote privacy • hiding secrets is hard • use community resources – ie find out about attacks and countermeasures • be reluctant to trust • . . . 35

Conclusions

Conclusions • Security problems growing, because of – more networks – more software – growing importance of services provided by these • more money to be made by attacker! • People will always concentrate on functionality first, and worry about security later (if at all) 37

Conclusions To prevent security problems, you must • understand the security requirements & the threats that attackers pose – aka threat modeling – think like an attacker! • know the technologies you use – operating system – programming language & platform –. . . and their security problems (do's & don'ts) – buffer overflows, SQL injection, . . . 38

Questions?