| Service composition technology builds a composite service that meets users’ complex service needs by aggregating simple services from multiple domains.User complex service demands usually contain multiple sub-service demands,which(for simplicity,in this article called service demands)are met by individual component services in the composite service.At present,more and more users tend to use composite services to accomplish their daily and work tasks.However,in a dynamic and complex application environment,the General Quality of Service(GQoS),Domain Quality of Service(DQo S),and user Service Demands(SD)of component services are dynamic.The changes in these factors often lead to the failure of composite services at runtime,which greatly affects user satisfaction and service resource utilization.Therefore,it is meaningful to study dynamic reconfiguration of composite services in large data environment.At present,domestic and foreign scholars have carried out preliminary studies on dynamic reconfiguration of composite services and achieved some results.However,the existing research mainly focuses on dynamic reconfiguration of composite services when GQoS fails.The proposed methods cannot solve the dynamic reconfiguration of composite services caused by DQo S failures and changes in service demands.In addition,the current research work lacks the active perception of the change of composite service failure factors,and can not accurately predict the failure of composite services,which seriously affects the user experience and service resource utilization.Therefore,how to actively perceive component service quality of service(GQoS or DQo S)failures and changes in user SD and dynamically reconstruct failed composite services according to specific changes has become an urgent problem in the field of service computing and modern services.For this reason,this paper conducts in-depth research on dynamic reconfiguration of composite services caused by quality of service(GQoS or DQo S)failure and changes in user SD.The main contents of this paper are as follows:(1)A dynamic reconfiguration method for composite services based on quality of service failure is presented.This method first obtains the criteria for determining GQoS and DQo S failures of each component service based on the global GQoS constraint decomposition and DQo S constraint statistics.Then,during the execution of the composite service,the GQoS and DQo S values of the component service that will be executed are proactively perceived based on the improved Case Based Reasoning method,and whether the component service will fail in quality of service(GQoS or DQo S).Finally,when the component service to be executed is about to fail in quality of service,based on the result of service clustering,a similar service that best matches the failed service can be quickly discovered and replaced with the failed service,thus completing the dynamic reconfiguration of the composite service oriented to quality of service failure.(2)A dynamic reconfiguration method for composite services is presented,which is oriented to the change of service demands.Based on the scenario information users are in,this method first implements the active perception of users’ current service needs based on the Case Chain Based on Reasoning proposed in this paper.Then,compare the perceived current service needs with the users’ original demands to determine if the users’ service demands have changed.If the service demands change,the areas that need to be reconstructed for composite services are determined according to the Qo S constraints(New Qo S Constraint,NQC)that the new service demands to satisfy and the Qo S constraints(Original Qo S Constraint,OQC)that the original service demands to satisfy.Finally,an adaptive reconfiguration strategy for composite services is designed according to the different service sizes required for reconfiguration,and the dynamic reconfiguration of composite services is completed quickly according to the strategy. |
PDF Full Text Request