Refresco: Improving Query Performance Through Freshness Control in a Database Cluster
Résumé
We consider the use of a cluster system for managing autonomous databases. In order to improve the performance of read-only queries, we strive to exploit user requirements on replica freshness. Assuming mono-master lazy replication, we propose a freshness model to help specifying the required freshness level for queries. We propose an algorithm to optimize the routing of queries on slave nodes based on the freshness requirements. Our approach uses non intrusive techniques that preserve application and database autonomy. We provide an experimental validation based on our prototype Refresco. The results show that freshness control can help increase query throughput significantly. They also show significant improvement when freshness requirements are specified at the relation level rather than at the database level.