Analysis of Self-⋆ and P2P Systems Using Refinement

Velisarios: Byzantine Fault-Tolerant Protocols Powered by Coq

“…Here n is the sequence number of the last executed request and d is the digest of the state. Once a replica received f + 1 different checkpoint messages 4 (possibly including its own) for the same n and d, it holds a proof of correctness of the log corresponding to d, which includes messages up to sequence number n. The checkpoint is then called stable and all messages lower than n (except view-change messages) are pruned from the log. View change.…”

“…The Rodin [2] platform for Event-B provides support for refinement, and automated and interactive theorem proving. Both have been used in a number of projects, such as: to prove the safety and liveness of self-systems [4]; to prove the agreement and validity properties of the synchronous crash-tolerant Floodset consensus algorithm [57]; and to prove the agreement and validity of synchronous Byzantine agreement algorithms [50]. In [50], the authors assume that messages cannot be forged (using PBFT, at most f nodes can forge messages), and do not verify implementations of these algorithms.…”

“…5). (4) We implemented a runtime environment to run the OCaml code we extract from Coq (see Sec. 6).…”

EventML: Specification, verification, and implementation of crash-tolerant state machine replication systems

“…Bryans proved that Floodset satisfies agreement and validity, and in addition he showed using ProB-a model-checker for Event-B-that the algorithm terminates. Event-B has also been used to model and verify safety and liveness properties of self-⋆ systems [6] (including self-healing, self-stabilizing, self-organizing, etc.). The authors illustrate their methodology using a P2P-based self-healing protocol.…”

“…5, and generate a verified implementation in Sec. 6, all of that in a single formal method tool. Sec.…”