Reducing asynchrony to synchronized rounds

Abstract

Synchronous computation models simplify the design and the verification of fault-tolerant distributed systems. For efficiency reasons such systems are designed and implemented using an asynchronous semantics. In this paper, we bridge the gap between these two worlds. We introduce a (synchronous) round-based computational model and we prove a reduction for a class of asynchronous protocols to our new model. The reduction is based on properties of the code that can be checked with sequential methods. We apply the reduction to state machine replication systems, such as, Paxos, Zab, and Viewstamped Replication.