Traditionally, allocation of data in distributed database management systems has been determined by o�-line analysis and optimization. This technique works well for static database access patterns, but is often inadequate for frequently changing workloads. In this paper we address how to dynamically reallocate data for partionable distributed databases 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. Based on arti�cial wide area network delays, we show that dynamic reallocation can improve system throughput by a factor of two and a half for wide area networks. We also show that individual site load must be taken into consideration when reallocating data, and provide a simple policy that incorporates load in the reallocation decision.

We examine load balancing in a simple pipeline computation, in which a large number of data sets is pipelined through a series of tasks and load balancing is performed by distributing several available processors among the tasks. We compare the performance of the optimal static processor assignment to the performances of three dynamic processor assignment algorithms. Models are derived which allow us to approximate the performance of the dynamic algorithms theoretically. The relative performances of the algorithms are investigated for various amounts of overhead, using a combination of modeling and simulation. We indicate that an appropriate dynamic algorithm can improve performance even when the overhead induced by the algorithm is relatively high. Keywords: pipeline computation, load balancing, performance evaluation A condensed version of this paper was presented at the IASTED Sixth International Conference on Parallel and Distributed Computing and Systems, 1994. This research was...