Power Optimization with Performance Assurance for Multi-tier Applications in Virtualized Data Centers

摘要: Modern data centers must provide performance assurance for complex system software such as multi-tier web applications. In addition, the power consumption of needs to be minimized reduce operating costs and avoid overheating. Various power-efficient management strategies have been proposed based on dynamic voltage frequency scaling (DVFS). Virtualization technologies also made it possible consolidate multiple virtual machines (VMs) onto a smaller number active physical servers even greater savings, but at cost higher overhead. This paper proposes performance-controlled optimization solution virtualized with While existing work relies either DVFS or server consolidation in separate manner, our utilizes both maximized savings by integrating feedback control strategies. At application level, multi-input-multi-output controller is designed achieve desired applications spanning VMs, short time scale, reallocating CPU resources DVFS. center optimizer incrementally VMs most longer scale. Empirical results hardware testbed demonstrate that can effectively performance-assured savings. Extensive simulation results, trace file 5,415 real servers, efficacy large-scale centers.