Transaction Management using Causal Snapshot Isolation in Partially Replicated Databases

Abstract

We present here a transaction management protocol using snapshot isolation in partially replicated multi-version databases. We consider here replicated databases consisting of multiple disjoint data partitions. A partition is not required to be replicated at all database sites, and a site may contain replicas for any number of partitions. Transactions can execute at any site, and read or write data from any subset of the partitions. The updates in this model are performed using asynchronous update propagation. The protocol ensures that the snapshot observed by a transaction contains data versions that are causally consistent. In developing this protocol, we address the issues that are unique in supporting transactions with causal consistency together with the snapshot isolation model in partially replicated databases. Such a model is attractive in geo-scale systems because of its support for partial replication, use of causal consistency model which does not require a global sequencer, and asynchronous propagation of updates.