We present a framework for concurrency control and avail-ability in multi-datacenter datastores. While we consider Google's Megastore as our motivating example, we define general abstractions for key components, making our solu-tion extensible to any system that satisfies the abstraction properties. We first develop and analyze a transaction man-agement and replication protocol based on a straightforward implementation of the Paxos algorithm. Our investigation reveals that this protocol acts as a concurrency prevention mechanism rather than a concurrency control mechanism. We then propose an enhanced protocol called Paxos with Combination and Promotion (Paxos-CP) that provides true transaction concurrency while requiring the same per in-stance message complexity as the basic Paxos protocol. Fi-nally, we compare the performance of Paxos and Paxos-CP in a multi-datacenter experimental study, and we demon-strate that Paxos-CP results in significantly fewer aborted transactions than basic Paxos.
ASJC Scopus subject areas
- Computer Science (miscellaneous)
- Computer Science(all)