A QoS Guarantee Framework for Cloud Services Based on Bayesian Prediction

Abstract

With the quality of service (QoS) of cloud services becoming increasingly concerned, how to ensure that the QoS of cloud services can meet the users' QoS requirement has become a focus of the study on cloud computing. However, the QoS of cloud services is dynamic regularly, we propose a QoS guarantee framework for cloud services and the Bayesian prediction method is used to predict QoS of cloud service. In our framework, cloud system can monitor and predict the QoS of cloud services in real-time, not only during the selection of cloud services, but also during the execution of cloud services. Once QoS prediction results show that some QoS violations will occur, cloud system will take measures to avoid the occurrence of QoS violations. We use the cloud simulation software called CloudSim for the experiment and the results demonstrate that compared with ARIMA and other prediction methods, our Bayesian prediction method has higher accuracy. Moreover, our QoS guarantee framework can greatly reduce the probability of QoS violation.