Transaction Reordering for Epidemic Quorum in Replicated Databases

Abstract

AbstractEpidemic model gives an efficient approach to transaction processing of replication systems in weakly connected environments. The approach has the advantages of high adaptation, support for low-bandwidth network, committing updates in an entirely decentralized control fashion, and so on. Optimistic voting protocol of epidemic model, which introduces condition and order vote in the voting process of transactions, takes an optimistic approach in conflict reconciliation. Though it brings down transaction abort rate dramatically when compared to other implementing protocols, optimistic voting protocol also rollbacks transactions unnecessarily. In this paper, transaction reordering technique is proposed in optimistic voting protocol to reduce the probability of transaction aborts. The protocol ensures one-copy serializability and does not lead to any deadlock situation. Experimental results suggest that the transaction reordering technique decreases abort rate and improves average response time of transactions markedly.KeywordsData ItemEpidemic ModelAverage Response TimeVersion NumberVote ProcessThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.