The authors' distributed cooperative-control framework uses concepts from optimal control theory to adaptively manage the performance of computer clusters operating in dynamic and uncertain environments. Decomposing the overall performance-management problem into smaller sub problems that individual controllers solve cooperatively allows for the scalable control of large computing systems. The framework also adapts to controller failures and allows for the dynamic addition and removal of controllers during system operation. This article presents a case study showing how to manage the dynamic power consumed by a computer cluster processing a time varying Web workload.
All Science Journal Classification (ASJC) codes
- Computer Networks and Communications
- Autonomic computing
- Distributed and cooperative control
- Optimal control
- Power management