Presentation is loading. Please wait.

Presentation is loading. Please wait.

May 9, 2008IPA Lentedagen, Rhenen1 Dynamic Fault Tree analysis using Input/Output Interactive Markov Chains Hichem Boudali 1, Pepijn Crouzen 2, and Mariëlle.

Similar presentations


Presentation on theme: "May 9, 2008IPA Lentedagen, Rhenen1 Dynamic Fault Tree analysis using Input/Output Interactive Markov Chains Hichem Boudali 1, Pepijn Crouzen 2, and Mariëlle."— Presentation transcript:

1 May 9, 2008IPA Lentedagen, Rhenen1 Dynamic Fault Tree analysis using Input/Output Interactive Markov Chains Hichem Boudali 1, Pepijn Crouzen 2, and Mariëlle Stoelinga 1. 1 Formal Methods and Tools group CS, University of Twente, NL. 2 Dependable Systems and Software group, CS, Saarland University, Germany

2 May 9, 2008IPA Lentedagen, Rhenen2 Introduction: Dependability Dependability: The trustworthiness of a computer system such that reliance can justifiably be placed upon the service it delivers. Reliability: The probability that a computer system does not fail within a given time bound.

3 May 9, 2008IPA Lentedagen, Rhenen3 Introduction: Formal dependability  Continuous-time Markov chains (CTMC)  States and Markovian transitions  Probability of traversing a λ- transition within t time-units is: 1-e -λt  Tools: Reachability analysis (among others) λ μ μ λ

4 May 9, 2008IPA Lentedagen, Rhenen4 Introduction: CTMC characteristics  CTMCs describe probability distributions (phase-type distributions)  Phase-type distributions can approximate any arbitrary distribution arbitrarily closely  Goal: Find a CTMC which describes the probability of system failure within t time- units (i.e. the unreliability of the system)  Problem: Difficult to find the CTMC that models a large system λ μ μ λ

5 May 9, 2008IPA Lentedagen, Rhenen5 Introduction: Engineering dependability  Fault Trees (1960’s)  Graphical  Easy to use  Syntax:  Basic events  Gates  Semantics: logical formula  Problem: Not expressive enough Mem1 fails CPU fails Workstation fails OR AND Mem1 fails

6 May 9, 2008IPA Lentedagen, Rhenen6 Introduction: Engineering dependability  Dynamic Fault Trees (1992)  Extension of classic fault trees  Additions:  Use of spares  Dependencies  Order-based failure  Tools:  Convert to CTMC P2P2 System failure P1P1 SPARE OR S

7 May 9, 2008IPA Lentedagen, Rhenen7 But… DFT Drawbacks  Scalability  Ambiguous syntax and semantics  Lack of modularity:  Dynamic modules can not be reused  Restrictions on spares and dependencies  Existing analysis technique is hard to extend or modify

8 May 9, 2008IPA Lentedagen, Rhenen8 Outline  Case study: FTPP system  DFT approach  Formalizing DFTs  DFT semantics in I/O-IMCs  Deep compositionality  Extending the DFT formalism  Conclusion  Future work

9 May 9, 2008IPA Lentedagen, Rhenen9 Case study: FTPP A D C B A D C B ADCB ADCB NE1 NE3 NE2NE2 NE4NE4  16 processors divided into 4 groups  4 network elements connect the processors  Per group 2 processors must be operational  Different configurations are possible

10 May 9, 2008IPA Lentedagen, Rhenen10 D D D D S S S S C B AA C B A B C C A B NE1 NE3 NE2NE2 NE4NE4 Case study: FTPP  16 processors divided into 4 groups  4 network elements connect the processors  Per group 2 processors must be operational  Different configurations are possible  Dynamic redundancy management is possible How reliable is each configuration?

11 May 9, 2008IPA Lentedagen, Rhenen11 FTPP DFT S S S S C B AA C B A B C C A B NE1 NE3 NE2NE2 NE4NE4

12 May 9, 2008IPA Lentedagen, Rhenen12 C AB 0.2 0.4 Failure rate: 0.2 f/h Failure rate: 0.4 f/h AND-gate Starting state: A is operational B is operational A has failed B is operational Pr(A fails in T hours) = 1 – e -0.2T A’s Mean time to failure = 1/0.2 = 5 hours A is operational B has failed A has failed B has failed  For static fault trees binary decision diagrams can be used!  Otherwise: Convert the DFT into a CTMC.  Analyze CTMC using standard solution techniques. Existing DFT analysis [Dugan et al. 1992] Unreliability = Prob[Reaching in time T] But…  State space explosion: CTMC grows exponentially  FTPP difficult to analyze

13 May 9, 2008IPA Lentedagen, Rhenen13 FTPP Results Group 1 Failure 2/3 S CBA Group 2 Failure 2/3 S CBA Group 3 Failure 2/3 S CBA Group 4 Failure 2/3 S CBA System Failure FDEP NE1 AAAA FDEP NE2 BBBB FDEP NE3 CCCC FDEP NE4 SSSS Analysis method Max number of states Max number of transitions Unreliability (T=10) Standard327574268262.55479 · 10 -8 Compositional1325141532.55479 · 10 -8 S S S S C B AA C B A B C C A B NE1 NE3 NE2NE2 NE4NE4

14 May 9, 2008IPA Lentedagen, Rhenen14 What’s behind it?  Model local behavior  We need compositional Markov chains  Combination of LTS and CTMC, with I/O automata features  Markovian transitions (CTMC)  Interactive transitions (LTS)  Action signature (IOA)  ? - Input actions  ! - Output actions  ; - Internal actions λ failed! I/O-IMC for Basic event Input/Output Interactive Markov Chains (I/O-IMC)

15 May 9, 2008IPA Lentedagen, Rhenen15  Properties of IMCs:  Combines stochastic behavior and interactive behavior orthogonally  CSP-style synchronization + interleaving semantics  Maximal progress for internal transitions  Properties of IOIMCs:  Unique outputs  Input enabledness  Outputs cannot be blocked!  Maximal progress for output transitions Input/Output Interactive Markov Chains λ τ

16 May 9, 2008IPA Lentedagen, Rhenen16 f(C)! f(A)? f(B)? f(A)? f(C)! f(A)? f(B)? DFT semantics DFT gate to I/O-IMC

17 May 9, 2008IPA Lentedagen, Rhenen17 What is deep compositionality? Group 1 Failure 2/3 S CBA  Semantics of a DFT arises naturally as composition of the semantics of its building blocks  But: This may lead to huge models. f(G1) f(NE1)f(NE4)… f(NE1)f(NE4) f(G1) f(NE2)f(NE3)

18 May 9, 2008IPA Lentedagen, Rhenen18 Why use deep compositionality?  Formally define semantics  Many useful techniques  Combining models: Composition  Refining models: Hiding  Minimizing models: Bisimulation  Reusing models: Renaming  Well supported by CADP toolset (VASY/INRIA) Combat State-space explosion

19 May 9, 2008IPA Lentedagen, Rhenen19 Compositional Aggregation Translation Composition + Abstraction Aggregation (minimization) Repeat Aggregated system CTMC (CTMDP) Result: System failure probability Analysis

20 May 9, 2008IPA Lentedagen, Rhenen20 Compositional Aggregation Example f(C)! f(A)? f(B)? f(A)? Failure rate: 0.2 f/h Failure rate: 0.4 f/h f(A)! 0.2 f(B)! 0.4

21 May 9, 2008IPA Lentedagen, Rhenen21 Compositional Aggregation Parallel Composition 1 2 3 1 2 3 4 5 1||1 0.2 f(A)! f(A)? f(B)? f(C)! 0.2 f(B)? f(A)! f(C)! 1||2 2||3 3||1 f(B)? 0.2 f(A)! 3||2 4||35||3 Inputs: f(A)? and f(B)? Outputs: f(C)! Inputs: none Outputs: f(A)! C A C||A Synchronize on f(A)

22 May 9, 2008IPA Lentedagen, Rhenen22 f(A); f(A)! Compositional Aggregation Abstraction (hiding) 1||1 0.2 f(B)? 0.2 f(C)! 1||2 2||3 3||1 3||2 4||35||3 C AB Abstraction (hiding): Makes signal internal

23 May 9, 2008IPA Lentedagen, Rhenen23 f(A); Compositional Aggregation Aggregation (weak bisimulation) 1||1 0.2 f(B)? 0.2 f(C)! 1||2 2||3 3||1 3||2 4||35||3 Weak bisimulation: Disregard internal steps Aggregation: Finding a smaller model equivalent (behaviorally) to the original

24 May 9, 2008IPA Lentedagen, Rhenen24 Compositional Aggregation Example (continued) 1 2 3 1 2 3 4 5 1||1 0.2 f(B)! 0.2 f(B)? f(C)! 0.2 0.4 2||1 1||2 0.4 0.2 2||2 C||A B C||A||B f(B)! 4||3 3||3 0.2 f(C)! 5||3

25 May 9, 2008IPA Lentedagen, Rhenen25 Compositional Aggregation Example (continued) 0.2 0.4 0.2 C||A||B f(C)!

26 May 9, 2008IPA Lentedagen, Rhenen26 DFT extensions  Extensions:  Inhibition  Repair-policies  Complex spares  Complex dependencies ……  Adding extensions in the compositional framework is easy:  Modify translation of DFT building blocks  Compositional aggregation algorithm is unaltered Free! DSN07

27 May 9, 2008IPA Lentedagen, Rhenen27 Extension: Repair f(C)! f(A)? f(B)? f(A)? r(C)! r(A)? r(B)? r(C)! r(A)? r(B)? r(A)? λ f(A)! µ r(A)! AND-gate C Basic event A

28 May 9, 2008IPA Lentedagen, Rhenen28 Conclusion: How we tackled drawbacks  State-space explosion.  Ambiguous syntax and semantics.  Lack of modularity:  Dynamic modules can not be reused.  Restrictions on spares and dependencies.  Existing analysis technique is hard to extend and/or modify. Compositional Aggregation DAG Extensions at the lowest level I/O-IMC Formal translation Renaming! Lifted!

29 May 9, 2008IPA Lentedagen, Rhenen29 Future work  Fully automated tool (CORAL)  More aggressive state reduction  Recent work: specialized acyclic algorithm  Apply deep compositionality to more advanced engineering formalisms! (see Boudali et al., DSN08)  Extend DFT formalism  Repair  Failure modes  Non-exponential failure distributions  Sophisticated dependencies

30 May 9, 2008IPA Lentedagen, Rhenen30 The end! Questions?


Download ppt "May 9, 2008IPA Lentedagen, Rhenen1 Dynamic Fault Tree analysis using Input/Output Interactive Markov Chains Hichem Boudali 1, Pepijn Crouzen 2, and Mariëlle."

Similar presentations


Ads by Google