批量到达下的IaaS云计算中心服务性能评价

Service Performance Evaluation of Iaa S Cloud Computing Center Under Batch Arrivals

针对请求批量到达下基础设施即服务(Iaa S)云计算中心性能分析问题,提出基于排队系统的云计算中心分析模型,并获得平稳状态时重要的服务性能参数:阻塞概率、立即服务概率、响应时间百分比、平均队长等。通过数值仿真实验分析了缓冲区和批量大小变化对系统性能的影响。数值仿真结果表明:同等排队强度下,缓冲区的增加对批量到达系统性能的改善优于单个到达系统;每批到达请求数的突发度越大,系统性能越差。

An analytical model based on queue system is proposed to deal with performance analysis of cloud center under batch arrivals. Some important performance indicators are acquired at steady status;these indicators include blocking probability, instance service probability, percentile response time, average queue length, and so on. The system performance influenced by changing buffer size and batch arrivals size is analyzed through numerical simulation. The numerical simulation results indicate that when buffer size is increased, the system performance with batch arrivals is better than single arrival under the same queuing intensity, and the system performance decreases as the burstiness of the number of every batch arrivals increase.