混合云资源调度策略性能模拟研究  

Simulations and Experiments on Scheduling Strategies for Hybrid Clouds

作　　者：潘世东[1] 屠成宇[1] 刘镇江[1] 陈卫国[1] 

机构地区：[1]江南计算技术研究所,无锡214000

出　　处：《高性能计算技术》2012年第3期58-63,共6页

摘　　要：资源调度策略与算法对混合云的服务性能具有至关重要的作用，然而构建真实的云计算环境对其进行深入研究不仅开销大，而且不具有可重复性，目前的研究多采用环境模拟的方法。CloudAnalyst是目前比较完善的大规模云模拟研究工具，其提供了能够使得研究者专注于调度策略与算法的模型。本文利用CloudAnalyst工具，并针对特定云计算应用场景进行了调度策略与算法的模拟实验，最终得出：（1）CloudAnalyst工具包适合进行大规模云计算环境调度算法研究的模拟仿真；（2）针对不同的需求以及应用，可以进行扩展以实现性能更优的调度策略与算法。Scheduling strategy and algorithms plays a prominent role for the performance of the service of Hybrid Clouds, however, it is extraordinarily expensive to deploy real Cloud computing environment based on which to make thorough research with no repeatability, therefore more and more researchers choose simulating tools. CloudAnalyst is a rather advanced research tool for large-scaled Clouds recently, and it provides models for enabling researchers concentrating on scheduling strategies and algorithms without considering so much programming technologies. The paper stimulate and experiment on simple scheduling strategies and algorithms for specific application scenarios with CloudAnalyst and make these： （a）it is suitable for CloudAnalyst to simulate research on scheduling strategies and algorithms in large-scaled Cloud Computing environment; （b） it is feasible to extend for new scheduling strategies and algorithms with better performance for various requirements and applications.

关 键 词：云计算 模拟工具 CloudAnalyst 调度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]