using strategy objectives for network security analysis
play

Using Strategy Objectives for Network Security Analysis Elie - PowerPoint PPT Presentation

Network Security Game Strategy Automated Analysis Conclusion Using Strategy Objectives for Network Security Analysis Elie Bursztein Stanford University / LSV, Ens-Cachan Inscrypt 2009 Elie Bursztein Using Strategy Objectives for Network


  1. Network Security Game Strategy Automated Analysis Conclusion Using Strategy Objectives for Network Security Analysis Elie Bursztein Stanford University / LSV, Ens-Cachan Inscrypt 2009 Elie Bursztein Using Strategy Objectives for Network Security Analysis 1 / 48

  2. Network Security Game Strategy Automated Analysis Conclusion Introduction Work purpose Analyzing and anticipating computer networks attacks. Elie Bursztein Using Strategy Objectives for Network Security Analysis 2 / 48

  3. Network Security Game Strategy Automated Analysis Conclusion Network complexity: The Pentagon Case Huge network ◮ 15 000 LAN Networks ◮ 7 000 000 Computers Huge Security problems ◮ Flash Drive banned due to a virus spread (Nov 2008). ◮ 1500 computers taken (Jun 2007) Elie Bursztein Using Strategy Objectives for Network Security Analysis 3 / 48

  4. Network Security Game Strategy Automated Analysis Conclusion Attack Complexity Elie Bursztein Using Strategy Objectives for Network Security Analysis 4 / 48

  5. Network Security Game Strategy Automated Analysis Conclusion Some Epic Failures ◮ 2004 Bouygues Telecom: 2 servers downs → 3 200 000 cellphones down ◮ 2005 Japan Mitsubishi: 1 computer infected → 40 MB of confidential reports leaked on a P2P network ◮ 2007 Apple: 1 computer in the production line infected → 150 000 ipods infected by the trojan RavMonE.exe Elie Bursztein Using Strategy Objectives for Network Security Analysis 5 / 48

  6. Network Security Game Strategy Attacks Automated Analysis Conclusion Outline Network Security Attacks Game Strategy Automated Analysis Conclusion Elie Bursztein Using Strategy Objectives for Network Security Analysis 6 / 48

  7. Network Security Game Strategy Attacks Automated Analysis Conclusion Vulnerabilities ◮ A vulnerability is a software bug that can be exploited by attacker to gain privilege. ◮ An exploit is the piece of software that takes advantage of a software bug. ◮ A 0day exploit is an exploit for an undisclosed vulnerability. Elie Bursztein Using Strategy Objectives for Network Security Analysis 7 / 48

  8. Network Security Game Strategy Attacks Automated Analysis Conclusion Vulnerabilities as Step stones ◮ Large networks may suffer multiple vulnerabilities ◮ Patches and counter-measures need to be prioritized ◮ A minor vulnerability can turn into a major hole when used as a step-stone Elie Bursztein Using Strategy Objectives for Network Security Analysis 8 / 48

  9. Network Security Game Strategy Attacks Automated Analysis Conclusion Illustration of a Complex attack Exploit a bug in Firefox Stealth the web Install a trojan server password Stealth all user Upload a rogue page password Elie Bursztein Using Strategy Objectives for Network Security Analysis 9 / 48

  10. Network Security Game Strategy Attacks Automated Analysis Conclusion The Need for Automation Attack analysis can’t be done by hand: network and attack are just too complex and big for that. We need models and tools for this ! Elie Bursztein Using Strategy Objectives for Network Security Analysis 10 / 48

  11. Network Security Game Strategy Attacks Automated Analysis Conclusion Attack Graph Frameworks ◮ 1998: Use of model-checking for host security [RS98] ◮ 2000: Use of model-cheking for network [RA00] ◮ 2004: First complete framework that constructs the attack scenario [SW04] ◮ 2005: Mulval [Ou05] a framework based on Datalog. ◮ 2006: NetSpa [ALI06] a framework that scale up to 50 000 nodes. Elie Bursztein Using Strategy Objectives for Network Security Analysis 11 / 48

  12. Network Security Game Strategy Attacks Automated Analysis Conclusion Time is the Essence Network security is a race between Intruder and Administrator. Windows of vulnerability Exploit Patch Released Released Windows of Vulnerability Time Elie Bursztein Using Strategy Objectives for Network Security Analysis 12 / 48

  13. Network Security Game Strategy Attacks Automated Analysis Conclusion The Need for Time Without time meaningless actions are allowed in the model. ◮ Administrator can patch 1000 services instantly. ◮ Intruder can compromise 1000 services before the administrator have a chance to react. Without time concurrent actions can’t be modeled. Ex: Administrator may patch a service while Intruder tries to exploit it. Elie Bursztein Using Strategy Objectives for Network Security Analysis 13 / 48

  14. Network Security Game Strategy Attacks Automated Analysis Conclusion Time and Game Model Timed automaton game [AFHMS]. Property Property can be written in Timed Alternating-Time Temporal Logic [AHK06]. Elie Bursztein Using Strategy Objectives for Network Security Analysis 14 / 48

  15. Network Security Game Strategy Attacks Automated Analysis Conclusion Collateral Effects Dommage Dommage Email DNS Web collatéral collatéral DDOS Attack Internet Elie Bursztein Using Strategy Objectives for Network Security Analysis 15 / 48

  16. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Outline Network Security Game Structure Rules Strategy Automated Analysis Conclusion Elie Bursztein Using Strategy Objectives for Network Security Analysis 16 / 48

  17. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Dual layer structure The Upper-layer is the timed automaton game, the Lower-layer represents the network state. Elie Bursztein Using Strategy Objectives for Network Security Analysis 17 / 48

  18. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Dual layer structure The Upper-layer is the timed automaton game, the Lower-layer represents the network state. Elie Bursztein Using Strategy Objectives for Network Security Analysis 17 / 48

  19. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Lower-layer: the network state The lower layer is composed of ◮ The dependency graph ◮ A set of states (atomic proposition) Elie Bursztein Using Strategy Objectives for Network Security Analysis 18 / 48

  20. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Web Service Receipt To build a web service you need: ◮ A HTTP frontend to serve the data Elie Bursztein Using Strategy Objectives for Network Security Analysis 19 / 48

  21. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Web Service Receipt To build a web service you need: ◮ A HTTP frontend to serve the data ◮ A SQL backend to store the data Elie Bursztein Using Strategy Objectives for Network Security Analysis 19 / 48

  22. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Web Service Receipt To build a web service you need: ◮ A HTTP frontend to serve the data ◮ A SQL backend to store the data ◮ A way to administrate the service Elie Bursztein Using Strategy Objectives for Network Security Analysis 19 / 48

  23. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Web Service Receipt To build a web service you need: ◮ A HTTP frontend to serve the data ◮ A SQL backend to store the data ◮ A way to administrate the service Elie Bursztein Using Strategy Objectives for Network Security Analysis 19 / 48

  24. Network Security Game Structure Strategy Rules Automated Analysis Conclusion The Dependency graph SQL SSH HTTP HTTP2 Elie Bursztein Using Strategy Objectives for Network Security Analysis 20 / 48

  25. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Set of States SSH SQL HTTP1 HTTP2 Vulnerable ⊤ ⊥ ⊥ ⊥ Compromised ⊥ ⊥ ⊥ ⊥ Elie Bursztein Using Strategy Objectives for Network Security Analysis 21 / 48

  26. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Rule Syntax ◮ ϕ pre : Preconditions. Rule syntax: Γ : Pre ϕ pre − → ∆ , p , a , c Effect ϕ eff Elie Bursztein Using Strategy Objectives for Network Security Analysis 22 / 48

  27. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Rule Syntax ◮ ϕ pre : Preconditions. ◮ ∆ : Time required to Rule syntax: complete the action. Γ : Pre ϕ pre − → ∆ , p , a , c Effect ϕ eff Elie Bursztein Using Strategy Objectives for Network Security Analysis 22 / 48

  28. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Rule Syntax ◮ ϕ pre : Preconditions. ◮ ∆ : Time required to Rule syntax: complete the action. ◮ p : The player that Γ : Pre ϕ pre executes the rule. − → ∆ , p , a , c Effect ϕ eff Elie Bursztein Using Strategy Objectives for Network Security Analysis 22 / 48

  29. Network Security Game Structure Strategy Rules Automated Analysis Conclusion Rule Syntax ◮ ϕ pre : Preconditions. ◮ ∆ : Time required to Rule syntax: complete the action. ◮ p : The player that Γ : Pre ϕ pre executes the rule. − → ∆ , p , a , c Effect ϕ eff ◮ a : Rule name. Elie Bursztein Using Strategy Objectives for Network Security Analysis 22 / 48

Recommend


More recommend