Admission Control for WebServices in Enterprise Systems Using Expert Systems

摘要

Presently, in order to make connections among applications in heterogeneous environments, webservices are utilized. QoS is one of the nonfunctional topics of webservices being the subject matter of a lot of research works. Considering the multiplicity of systems which are connected together, and on the other hand, notifying the shared resources, a system may face congestion problem if there is no admission control applied. In this article to solve the problem, we will propose a mechanism covering admission control for webservices in enterprise systems. In this mechanism, for each webservice, the rate of resource consumption and the time taken for its execution are evaluated. Then, by receiving requests, functionality of webservice is investigated and stored as learning data. In case the webservice cannot respond in the allocated time, the system will sense the probability of congestion. This mechanism uses Naive Bayesian method to predict next status of system. Admission control mechanism, considering the present status of the system and learning data, will distinguish what kind of response should be given to the requester of the service in order to protect the system from congestion. Consequently, by using this method, we can tackle with congestion problem when we use webservices in enterprise systems.