Multiple Granularity Control Scheme for System Utility Optimization in Grid Environments

Abstract

In complex grid environment, a control system should consider all applications and coordinate all layers of grid architecture upon any changes in the system. However, this brings large overhead because any changes will invoke a global coordination. The paper proposes a multiple granularity control scheme in grid computing, which balances control scope and control frequency to improve system performance. Multiple granularity control policies are deployed at different levels: system level control at coarse time granularity and application level control at fine time granularity. System level control considers all applications and coordinates three layers of grid architecture in response to large system changes at coarse time granularity; it exploits the interlayer coupling of fabric layer, collective layer, and application layer to achieve a system-wide optimization based on the user’s preferences. Application level control adapts a single application to small changes at fine time granularity. The paper presents a multiple granularity control algorithm (MGCA). Simulations are conducted to test the performance of the control algorithm.