基于M/M/c排队模型的云计算中心能耗管理策略

Energy Consumption Strategy for Cloud Computing Center Based on M/M/c Queuing Model

由于云计算中心在降低能耗的同时还需要保证服务质量(QoS),针对用户访问云计算中心的排队机制,给出一种云计算任务排队模型,在此基础上提出一种基于M/M/c排队过程的云计算中心能耗管理算法,通过求解该模型获得了平均等待时间、阻塞概率等性能指标进而建立系统的能耗模型;同时用参量ERP (Energy-Response time Product)作为排队网络的反馈量,引入反馈策略及服务器休眠预留机制,动态调整云计算中心服务器服务数;仿真结果表明,与其他策略进行比较该策略能够在保证QoS值的情况下,有效降低系统的能耗,避免了服务器资源浪费。

Owing to the cloud computing center needs to guarantee the quality of service(QoS)while reducing the energy consumption,presenting aqueuing model of cloud computing tasks based on the queuing mechanism of users accessing the cloud computing center.Based on this model,an energy consumption management algorithm of cloud computing center based on M/M/C queuing process is proposed.By solving the model,the average waiting time,blocking probability and other performance indexes are obtained.Then the energy consumption model of the system is established.At the same time,the parameter ERP(energy response time product)is used as the feedback quantity of the queuing network,and the feedback strategy and the server sleep reservation mechanism are introduced to dynamically adjust the number of server services in the cloud computing center.The simulation results show that compared with other strategies,this strategy can effectively reduce the energy consumption of the system and avoid the waste of server resources while ensuring the QoS value.