The Performance Analysis For Web Services Composition Based On Queueing Petri Net

Web Services achieve interaction with the data from different platforms by itsstrict protocol standard, unified description language and flexible communicationmode, and the technology advantage solved the problem of waste of resources aboutsoftware service caused by platform independent. The emergence of the Web Serviceto promote the service oriented architecture concept of SOA rapid development, asingle Web services is difficult to meet the needs of users, so the industry puts forwardto Web Service composition technology, will compose more powerful combination ofservices from a single Web Service through a certain layout rules to meet the needs ofthe user changeable. Web services has to not only meet the need of users to function,but also to improve the performance of the composition, so the performance analysisof the combination of services has become the research focus.WS-BPEL belongs to process description language of Web service composition,which used to describe the orchestrated processes of Web service composition, firstwe need to establish formal model to describe the flow of service compositions, inorder to analysis the performance of the combined service. Users need to wait in linewhen access to any service in the combination system, however, in the current formalmodeling methods, such as process algebra method, Coloured Petri nets andStochastic Petri nets etc are not introduced queuing theory to analyze this problem,This paper present an application of Queueing Petri nets to BPEL process modeling,through the simulation model QPME to achieve the performance analysis of the Webservices composition.First we introduced the Queueing Petri nets, it derive from the Colored Petrinets and Stochastic generalized Petri nets, and join the queue concept in queueingplace，so the Queueing Petri nets has a stronger ability to build a model.And then describes the language feature and basic structure of BPEL, andprovides the mapping rules and methods to the transform from WS-BPEL toQueueing Petri nets t, realize the WS-BPEL atomic activities and structural activitiesto Queueing Petri nets model conversion.Finally provides the mapping rules and methods in accordance with a booksmanagement system management system of WS-BPEL into Queueing Petri netsexample model, using QPME simulation tools for the models performance simulation and analysis, mainly through the user to control interval time distribution parametersand node service time length parameter analysis each service node mean responsetime, service utilization as well as the mean total tokens three measure the change ofattributes, so as to optimize the system, so that the whole Web service compositionmodel to achieve the best performance.