Atomic Broadcast and Quorum-based Replication in Asynchronous Crash-Recovery Distributed Systems

Luı́s Rodrigues, Michel Raynal · 1999

This paper describes a solution to the Atomic Broadcast problem in asynchronous distributed systems in which processes can crash and recover. Our solution is based on a transformation of any Consensus algorithm into an Atomic Broadcast algorithm. Thus, our work can benet from recent results in the study of the Consensus problem in the crashrecovery model. The paper also shows how to extend the Atomic Broadcast primitive to support the implementation of Quorum-based replica management in crash-recovery systems. The proposed technique makes a bridge between established results on Weighted Voting and recent results on the Consensus problem. Keywords: Asynchronous Distributed Systems, Atomic Broadcast, Consensus, Crash/Recovery, Quorum, Replica Management, Weighted Voting. 1 Introduction Atomic Broadcast is one of the most important agreement problems encountered in the design and in the implementation of fault-tolerant distributed systems. This problem consists in providing processes w...

Read the paper · More papers on PaperTik