Verifying eventual consistency of optimistic replication systems

Abstract

We address the verification problem of eventual consistency of optimistic replication systems. Such systems are typically used to implement distributed data structures over large scale networks. We introduce a formal definition of eventual consistency that applies to a wide class of existing implementations, including the ones using speculative executions. Then, we reduce the problem of checking eventual consistency to reachability and model checking problems. This reduction enables the use of existing verification tools for message-passing programs in the context of verifying optimistic replication systems. Furthermore, we derive from these reductions decision procedures for checking eventual consistency of systems implemented as finite-state programs communicating through unbounded unordered channels.

Making Weak Memory Models Fair

“…Several papers, e.g., [5,6,9], have studied declarative formulations of transactional consistency with prefix-finiteness constraints to ensure that a transaction is never preceded by an infinite set of other transactions. In particular, Gotsman and Burckhardt [9] established a connection between declarative presentations that include fairness constraints and operational presentations for models in their "Global Operation Sequencing" framework.…”

“…In these models prefixfiniteness of mo and fr does not suffice for prefix-finiteness of the appropriate "happens-before" relation. For instance, under ARMv8 [8], assuming prefix-finiteness of mo and fr does not forbid the out-of-thin-air read of the value 5 in the , following example (with an unbounded address domain): 5) . .…”

Eventual Consistency Formalized

“…Bouajjani, Enea and Hamza [13,14] define eventual consistency as a property over traces observed by an external witness. Eventual consistency is grounded in the notions of safety and liveness, and is defined in terms of finite prefixes of a global interpretation of method calls in a system where the result of a call is welldefined (safety), and where there exists a global interpretation of all the method calls in an infinite trace.…”