Systems need to run a larger and more diverse set of applications, from real time to interactive to batch, on uniprocessor & multiprocessor platforms. The tasks that are scheduled using proportional share concept, assigns a weight randomly. This paper suggests a mechanism which assigns a weight based on the resource requirement of an application. The problem of inferring application resource requirements is difficult because the relationship between application performance and  resource requirement is complex and workload dependent. We present a measurement-based approach to resource-inference employing online measurements of workload characteristics and system resource usage to estimate application resource requirements. These requirements are translated to appropriate weights and to modify these weights dynamically by employing weight readjustment algorithm.

Key words: Multiprocessor, Operating Systems, Proportional Share Schedulers, Resource Inference.