Cluster-level feedback power control for performance optimization

Power control is becoming a key challenge for effectively operating a modern data center. In addition to reducing operation costs, precisely controlling power consumption is an essential way to avoid system failures caused by power capacity overload or overheating due to increasing high-density. Control-theoretic techniques have recently shown a lot of promise on power management thanks to their better control performance and theoretical guarantees on control accuracy and system stability. However, existing work over-simplifies the problem by controlling a single server independently from others. As a result, at the cluster level where multiple servers are correlated by common workloads and share common power supplies, power cannot be shared to improve application performance. In this paper, we propose a cluster-level power controller that shifts power among servers based on their performance needs, while controlling the total power of the cluster to be lower than a constraint. Our controller features a rigorous design based on an optimal multi-input-multi-output control theory. Empirical results demonstrate that our controller outperforms two state-of-the-art controllers, by having better application performance and more accurate power control.