Evaluating the coordination overhead of replica maintenance in a cluster of databases

Abstract

We investigate the design of a coordinator for a cluster of databases. We consider the following alternatives: TP-Heavy using the TUXEDO TP-monitor, TP-Lite with the ORACLE8 database system, and a TP-Less coordinator implemented in Embedded SQL/C++. In particular, we investigate the scalability of full replication. We assume that update actions on all replica are executed either synchronously or asynchronously. It turns out that the TP-Less approach outperforms commercial TP-middleware for small cluster sizes already. Another observation is that asynchronous updates are the preferred option compared to synchronous updates. The conclusion is that a transaction protocol at the second layer must be replication-aware. © Springer-Verlag Berlin Heidelberg 2000.