Integrating QoS And Reputation For Services Discovery In Web Services Market

In recent years, with the development of Web services and its related technologies,the number of Web services on the Internet increased rapidly. Against this background,online Web service portals and service markets are playing an important role in Webservices search and discovery. What’s more, how to search the services which meetuser’s requirements effectively, is one of the issues received much attention now andstill unresolved. At the same time, due to the development of portals and markets, thesolution to the problem not only has practical significance but also can greatlycontribute the development of the service markets forward.The properties of Web services can be divided into two categories: functionalrequirements and non-functional requirements. There are many methods andtechnologies about Web service discovery based on functional requirements. Thefocus of research in recent years for service discovery mainly based on non-functionalrequirements, it is the quality of Web services. If exist many similar functional Webservices, non-functional attributes will be the most important factor for serviceselection undoubtedly. However, the quality of service may be untrustworthy andinaccurate. Because QoS generally declared by Internet service provider(ISP), notnecessarily credible, and it may vary due to the conditions of different user’s locationand network. Therefore, lots of users involved in the consumption and evaluationservices, to reflect the service’s reputation. It will be a very important indicator forservice discovery. But only relying on reputation to choose service is not feasible,since some evaluations of users is incredible and the results can not meet thenon-functional requirements of the user(e.g., such as price, response time, reliability).As discussed above, this paper proposes an algorithm integrating QoS andreputation for services and users ranking and discovery in Web services market. Weimplemented it in Web Services Supermarket, which can improve the accuracy andeffectiveness of service discovery and meet user’s requirements.Firstly, we introduce the services markets and developed a distributed Webservice testing system in order to collect the QoS data of Web services. A prototypeimplementation of this algorithm and examples will be given later to show the search results. Finally, the experiment about the services search results of prototype systemwill be obtained and experimental data shows that it meets the user’s requirementsthan other models which only consider QoS or reputation.