Saving energy consumption for mixed workloads in cloud platforms

Abstract

Virtualisation technology has been widely applied in cloud systems; however it also introduces many energy-efficiency losses especially when the I/O virtualisation mechanism is concerned. In this paper, we present an energy-efficiency enhanced virtual machine (VM) scheduling policy, namely share-reclaiming with collective I/O (SRC-I/O), with the aim to reduce the energy-efficiency losses caused by the I/O virtualisation. The SRC-I/O scheduler allows running VMs to reclaim extra CPU shares in certain conditions so as to increase the CPU utilisation. Meanwhile, SRC-I/O policy separates I/O-intensive VMs from CPU-intensive ones and schedules them in a batch manner, so as to reduce the context-switching costs of scheduling mixed workloads. Extensive experiments are conducted on various platforms by using different benchmarks to investigate the performance of the proposed policy. The results indicate that when the virtualisation platform is in presence of mixed workloads, SRC-I/O scheduler outperforms existing VM schedulers in terms of energy-efficiency and I/O responsiveness.