基于贝叶斯模型的云服务服务质量预测

发布时间：2017-12-28 14:30

本文关键词：基于贝叶斯模型的云服务服务质量预测　出处：《计算机应用》2016年04期 　论文类型：期刊论文

【摘要】：针对如何分配一个未来一段时间内满足QoS要求的云服务和感知可能将要发生的QoS违规的问题,提出一种基于时间序列预测方法的云服务QoS预测方法。该预测方法利用改进的贝叶斯常均值(IBCM)模型,能够准确地预测云服务未来一段时间内的QoS状态。实验通过搭建Hadoop集群模拟云平台并收集了响应时间和吞吐量两种QoS属性的数据作为预测对象,实验结果表明:相比自回归积分滑动平均(ARIMA)模型和贝叶斯常均值折扣模型等时间序列预测方法,基于改进的贝叶斯常均值模型的云服务QoS预测方法的平方和误差(SSE)、平均绝对误差(MAE)、均方误差(MSE)和和平均绝对百分比误差(MAPE)均比前两者小一个数量级,因此具有更高的预测精度;同时预测结果对比图说明提出的预测方法具有更好的拟合效果。
[Abstract]:Aiming at the problem of how to allocate a QoS service violation that may happen in the future to meet the QoS requirements and perceive the problem of QoS violation, a prediction method of cloud service based on time series prediction method is proposed. This prediction method uses an improved Bayesian constant mean value (IBCM) model to accurately predict the QoS state in the future period of cloud services. Through the experiment of building a Hadoop cluster simulation of cloud platform and collect the response time and throughput of two QoS attribute data as forecasting object. The experimental results show that compared with the autoregressive integrated moving average (ARIMA) model and Bayesian constant mean time discount model series prediction method, improved BCM model of cloud service QoS based on the method of prediction error sum of squares (SSE), the mean absolute error (MAE), mean square error (MSE) and the mean absolute percentage error (MAPE) than the former two one order, so it has higher prediction accuracy; and prediction results illustrated the fitting effect of the proposed method has better.
【作者单位】：江苏大学计算机科学与通信工程学院;
【分类号】：TP393.09;TP18
【正文快照】： 0引言近年来,随着云计算的发展和成熟,用户对云服务的服务质量(Quality of Service,Qo S)越来越重视,并且希望云服务提供商能够保证分配给他们的云服务满足云任务的Qo S要求。云计算主要有三种服务模式[1],它们都是通过动态的Internet网来提供服务的,因而服务质量会随着时间的

【相似文献】