Experimental evaluation of dynamic data allocation strategies in a distributed database with changing workloads

Abstract

Traditionally, allocation of data in distributed database management systems has been determined by off-line analysis and optimization. This technique works well for static database access patterns, but is often inadequate for frequently changing workloads. This paper addresses the problem of dynamically reallocating data in a partionable distributed database with changing access patterns. Rather than complicated and expensive optimization algorithms, a simple heuristic is presented and shown, via an implementation study, to improve system throughput by 30% in a local area network based system. For a wide area network the performance gain is expected to be even larger. It is also shown that individual site load must be taken into consideration when reallocating data. A a simple policy that incorporates load in the reallocation decision is provided.