A Bayesian network based Qos assessment model for web services

Quality of service (QoS) plays a key role in Web services. In an open and volatile environment, a provider may not deliver the QoS it declared. Hence, it's necessary to provide a QoS assessment model to determine the likely behavior of a provider. Although many researches have been done to develop models and techniques to assist users in QoS assessment, most of them ignore various QoS requirements of users, which are great important to evaluate a provider adopting the policy based on service differentiation. In this paper, we propose an approach, called Bayesian network based QoS assessment model, to QoS assessment. Through online learning, it supports to update the corresponding Bayesian network dynamically. The salient feature of this model is that it can correctly predict the provider's capability in various combinations of users' QoS requirements, especially to the provider with different service levels. Experimental results show that the proposed QoS assessment model is effective.