QoS-Based Web Services Selection Using a Hidden Markov Model

Abstract

Web services are a technology that has been growing rapidly since itsinception earlier this century. They are so popular nowadays that theyare being used for numerous purposes, spanning business, healthcare,education and government applications. They are used for accessingdiverse types of data in domains such as the financial, climateforecasting and sports, among others. As a result of this increasedpopularity, it is very common to find a large number of applicationsproviding exactly the same service. Having a large number ofapplications to choose from, one that integrates a selection mechanismusing quality criteria for selecting the best choice, has an advantageover other applications that do not provide this feature. However, theselection of the best Web service, among various services with the samefunctionality, is not an easy task. Therefore, the design of aworthwhile technique to determine the best Web service by assessing itsquality of service is important. Hidden Markov models are probabilisticmethods that allow us to build behavior models of the Web services. Suchmodels can be used to predict their behavior in the near future. In thispaper, we propose a method for selecting Web services based on theresponse time QoS parameter and hidden Markov models.