1 / 24

Non-monotonic Properties for Proving Correctness in a Framework of Compositional Logic

Non-monotonic Properties for Proving Correctness in a Framework of Compositional Logic. Koji Hasebe Mitsuhiro Okada (Dept. of Philosophy, Keio University). Purposes. Make more explicit compositionality of the original compositional logic

corbett
Download Presentation

Non-monotonic Properties for Proving Correctness in a Framework of Compositional Logic

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Non-monotonic Properties forProving Correctness in a Framework of Compositional Logic Koji Hasebe Mitsuhiro Okada (Dept. of Philosophy, Keio University)

  2. Purposes • Make more explicit compositionality of the original compositional logic (Durgin-Mitchell-Pavlovic 2001, Datta-Derek-Mitchell-Pavlovic 2003) • Divide an honest principal's role into primitive actions • Simplify the inferences of compositional logic • Do not use , , temporal operators • Distinguish the monotonic properties and the non-monotonic ones • Give a semantics which is sound for our system 1

  3. Review of Compositional logic Durgin-Mitchell-Pavlovic (2001), Datta-Derek-Mitchell-Pavlovic (2003) • Inference system based on Floyd-Hoare style logical framework to prove a protocol correctness : "after a protocol action , holds from P 's view" • An advantageous point: For proving correctness of a compound protocol, we can reuse properties of its components. 2

  4. New idea of ours • Divide an honest principal's role into primitive actions (sending, receiving, etc.) • Formalize honesty assumptions with explicit reference to a role-component : "if Q honestlyfollows his/her role-component , then holds". (cf.) : "a principal Q is honest, then holds. " 3

  5. The language (1): formulas • Atomic formulas: • atomic action formulas: (denoted by ) (with n m) • atomic non-action formulas: • A sequence of actions: (described by a non-commutative conjunct of atomic action formulas) 4

  6. : Q 's role-component • : a sequence of actions performed by P • : Q honestly follows a role-component • : a property (a sequence of atomic action formulas or a non-action formula) • : a set of properties If Q honestly follows his/her role-components , and if holds, after P performs a sequence of action holds from P 's view. The language (2): basic form of assertion 5

  7. Weakening rule and monotonicity • is a monotonic property if we can freely apply the weakening rule. weakening e.g. • Receives, Fresh : monotonic properties • Firstly Sends : non-monotonic properties • To include non-monotonic properties • Require some restrictions on the weakening rule • However, provide us more powerful derivations 6

  8. Axioms and inference rules • Logical inferences with equality • Action properties axioms • axiom about actions • axioms for relationship between properties • Honesty inferences • Weakening rule 7

  9. 1. Examples of Logical inference rules • Cut • Equality • Inference rules for non-commutative conjunction ( ; ) 8

  10. 2. Action properties axioms (1) • Nonce verification 1: • Axiom about actions: (for each i=1,...,n) • Examples of axioms relationship between properties: • Freshness 1: 9

  11. 2. Action properties axioms (2)(related to the non-monotonic property "firstly sends") • Firstly Sends: • Ordering of actions: (Here is an action including .) These are useful to derive ordering of actions. 10

  12. Idea of the Honesty Inference One can derive some performance of actions by a principal different from the viewer. (e.g.) from P's view: • P receives a message . • is a secret part of Q's public key. • contains a fresh value. Therefore, P knows that Q sends . But, this is not enough. We need some inferences using assumptions about a principal's honesty. We introduce the following three types of honesty inferences. 11

  13. 3. Honesty inferences (1) • Substitution (sending): • receiving 12

  14. 3. Honesty inferences (2) • Matching: (where m m') • Q honestly follows Q sends m'. • Q does not follows Q sends m'' with m m'', m'' m'. Condition: does not appear below this inference. 13

  15. 3. Honesty inferences (3) • Deriving another action (receiving): • sending • generating 14

  16. Other possibilities of combination: A composing process of honesty assumptions 15

  17. Examples of correctness proofs • Proof of the agreement property for the Needham-Schroeder public key protocol. • Proof of the matching conversations for the Challenge Response protocol: 16

  18. If the initiator (say, A) communicates with the responder (say, B) using the concrete values of nonces and , then there exists B actually performing the responder's role with the same nonces and . Example 1: Needham-Schroeder protocol (1) (Needham-Schroeder, 1978) initiator's concrete actions responder's role Agreement Property from A’s view: 17

  19. A’s role Q’s role Example 1: Needham-Schroeder protocol (2) A's view: by the information about key and nonce , with by an equality inference, with by the honesty inference (matching), 18

  20. A’s role Q’s role Example 1: Needham-Schroeder protocol (3) On the other hand, by the information about key and nonce , by the honesty inference (substitution), 19

  21. A’s role Q’s role A’s role Q’s role Example 1: Needham-Schroeder protocol (4) Then by composition of honesty assumptions, Cut Comp(Hon) (Here .) Finally, Honest(Role) Comp(Hon) 20

  22. 1. Following sequents are provable: 2. By “firstly sends” order 3. Finally, we get Example 2: CR protocol 21

  23. Soundness theorem Trace Semantics • Primitive state: • P has information m: • Message m is transmitted through the network: • State: a multiset of primitive states • Trace: a finite sequence of states Theorem.If a sequent S is provable in our system, then S is true for any trace s which includes no duplicated atomic actions. 22

  24. Conclusions and future work • Made more explicit the compositionality of compositional logic • Simplified the inference rules • Gave a trace semantics • Extend by adding , , temporal operators for more powerful derivations 23

More Related