基于系统响应时间的云服务质量评估模型

摘要

响应时间是云服务系统的主要质量衡量指标之一,过长的系统响应时间意味着用户的流失.以队列等待、服务器处理时间和网络延迟构成云访问的总体响应时间,采用嵌入马尔可夫链的方法生成观察点的马尔可夫链.借助马尔可夫链的无后效特性来模拟云访问的随机到达,以此为基础构建云服务排队模型,用以评估云服务质量.通过解析仿真的方法验证了所提出的云服务质量评估模型的有效性.%The response time is one of the key quality indexes of cloud service system,which means that the longer of system response time,the more of losing customers.Queuing waiting,server processing time and network delay constitute the overall response time of cloud service access,and the Markov chain of observation points is generated by embedded Markov chain.The random arrival of cloud access is simulated by the non-aftereffect property of Markov chain,and the queuing model of cloud service is constructed to evaluate the quality of cloud service.At the same time,the validity of the proposed cloud service quality evaluation model is verified by analytic simulation.