Resource allocation with load balancing consideration in distributed computing systems

Abstract

A new resource allocation model is presented in which a given number of copies of a single resource are allocated to the processing sites in such a way that the total communication cost incurred is minimized. The accessing scheme considers both the communication costs and the load levels at the resource sites. Load leveling is imposed as a constraint. This improves system throughput as well as response time. The allocation provided by the model reflects the realistic environment more accurately and therefore gives a better dynamic performance. It is shown that the model can be extended to include other costs such as installation cost and communication cost due to 'write' accesses.<>