Skip to main content
eScholarship
Open Access Publications from the University of California

Dynamic load balancing algorithm complexity

Abstract

This paper presents a theoretical analysis of the asymptotic complexity inherent in a load balancing algorithm in a loosely-coupled network, where processor communication is achieved by message passing. The load balancing complexity depends on the network topology and the overhead of processor communication for each polling strategy. The best, worst, and average case analysis of the load balancing algorithms for the various polling topologies are presented. The polling strategies considered are local, global, and random polling. The complexity is presented as a function of the number of processors in the network.

Main Content
Current View