Font Size: 

In this paper, we investigate the effects of load balancing in a distributed Web caching system. Our investigation is focused specifically on adaptive load sharing: an approach that reacts to the current state of the system. Load balancing has been shown to improve system performance in other applications and in this paper, we investigate it in a distributed Web caching
environment using both a unified and partitioned approach. The goal of this work is threefold: (1) to determine the conditions under which load balancing can be beneficial in a distributed Web caching system, (2) to compare load balancing in a unified and partitioned Web caching system, and (3) to determine how much state information is required to achieve any benefit.
Discrete-event simulation is used as the tool to generate results for these different environments.

modelling, simulation, programming