Energy-Saving I/O software for Cloud Computing

在这项工作中，我们评估了现有I/O调度器在虚拟化环境中的性能。我们的评估表明，现有的I/O调度器无法在虚拟化环境中提供足够的性能。为了分析I/O调度器的行为，我们开发了I/O调度器分析系统。系统分析清楚地揭示了其性能不足的原因。在此基础上，提出了一种提高虚拟环境下I/O性能的方法。我们的性能评估表明，所提出的方法已经能够提高性能。

In this work, we have evaluated performance of the existing I/O scheduler in virtualized environment. Our evaluation has demonstrated that the existing I/O scheduler have not been able to provide enough performance in virtualized environment. In order to analyze the behavior of the I/O schedulers, we have developed I/O scheduler analyzing system. Analyses with the system have clearly shown the reason of their insufficient performance.After the analyses, we have proposed a method for improving I/O performance in virtualized environment. Our performance evaluation has demonstrated that the proposed method has been able to improve performance.