Building Of Semantic Web Service Indices And QoS-Aware Service Discovery

Great amounts of available web services have emerged with the appearances of maturedmodels during many years of technology accumulation.With the help of the semantic web tech-nology,the descriptions of web sevices now are provided with rich semantics,which become se-mantic web services and ontology reasoning tools are widely used in service discovery,whichhas highly improved the precision of the discovery algorithms nevertheless slowing down thespeed.Moreover,the increase of the number of the advertisement services highlights service se-curity problem.Users raise concern about the dependability of service,reliability and other QoSfactors,and the discovery algorithms return more and more functionally equivalent services amongwhich service requesters are required to select on QoS properties.Hence,service discovery modelsshould provide further support in both description and matching of such QoS properties as secu-rity,dependability to make the service discovery and invocation secure,reliable and controllable asmuch as possible.In this thesis,we first discuss an index-based method for semantic web service discovery andapply it to the service functionally matchup,expecting to solve the performance problem causedby ontology reasoning.We use ontology concepts to describe each functionally parameter of theservice and use the inclusion relation between these concepts which is determined by their positionin the ontology concept tree to define the essential conditions of service matchup.Next,we employan interval encoding of these concepts and build indicies(B+-tree or R-tree),based on which thesemantic matchup of service functionally parameters could be implemented by the rang-query inthe indices.At last of this discussion,we present the concrete matchup algorithms.In respect of security and dependability properties,we first discuss the characteristics of webservice QoS and develope an ontology description model—QoSOL,which is used to describe thesemantic information of QoS properties such as service security,dependability,performence andnetwork.Then we devise a matchup model for semantic web service QoS parameters.The modelwould first inspect whether the QoS parameters of both sides were semantically matching.If itwas,the model would furtherly use constraint programming method to inspect whether the con-straints of those parameters were matching,then classify the results in detail so as to give servicerequestermore feedback.In addition,we discuss the device of the service qualiy evaluation functionand service selection based on QoS.Finally,weprovedwithaseriesofexperimentsthat,inthematchupprocessoflarge-scaleofse-mantic web services,using index to implement the semantic matchup could prune the search space and minimize the number of comparisons,which consequently quickened the matchup process andguaranteed the accuracy as well,and it was possible to describe and matchup QoS semantics infor-mation such as service security,dependability,performence and network so as to make the servicediscovery and invocation more secure,reliable and flexible.