Research On QoS-aware Service Composition And Load Balancing Strategy

Web service technology is an effective way to call resources in Service-Oriented Architecture(SOA).Although Web services encapsulate all kinds of resources in the network,single Web service functions are often limited and can only accomplish specific tasks,which can not satisfied the designated businesses,which do not conform to the real users's requests.According to different requests,Service Composition can dynamically form a set of Web services in the network into effective call sequences,so as to achieve more powerful functions.With the rapid development of mobile networks,especially the 5G technology,the number of Web services and user's requests are increasing rapidly in the cloud environment.The method of manual service selection and service composition by human experts has been unable to adapt to the current cloud environment,and the technology of automatic service composition has become a hot research topic.This paper focuses on several key issues of automatic service composition in the cloud environment,taking semantic analysis of web service parameters as the starting point,aiming at improving service composition efficiency,optimizing service composition performance,ensuring system stability and load balancing.The research content of this paper is as follows.To solve the problem of Service Composition in cloud environment,a service matching method based on input and output parameters is proposed.Through semantic analysis of Web services,a logical relationship is built between services to generate a relationship dependency graph,and a service composition scheme that satified the functional requirements of users is found by graph theory.Aiming at the similar or replaceable problems of massive services,a service class generation method based on parameter specification is proposed.The service specification with the same input and output parameters is put into the same service class to support the subsequent service selection and multipaths service composition scheme.In view of the huge number of candidate services,a service filtering algorithm is proposed to preprocess the candidate service set,which starts with the service request,traverses the service resource pool layer by layer,keeps the effective service and parameter matching table,reduces the generation scale of dependency graph,and improves the efficiency of service compositionAiming at the problem that a single optimal path leads to the unbalanced load of system service nodes,a multipaths service composition scheme based on QoS(quality of service)is proposed.This method establishes a mathematical model for request traffic and system load information,and proposes an algorithm to ensure the implementation of multipaths to satified the load balancing requirements of the system.The algorithm consists of two parts.The first part is to find the feasible path based on the relation dependence graph,that is,to find multiple end-to-end service composition schemes satisfying QoS constraints;The second part is to allocate the request trafic to these feasible paths,that is,to allocate the appropriate request traffic according to the load capacity of the current service node in the feasible service composition scheme,so as to avoid node overload and satisfied the load balancing requirements of the system at the same time.In order to solve the problem that there is no uniform standard and specification in the existing QoS model,which can not provide users with real satisfactory service composition scheme accurately,an improved and expandable QoS model is proposed.The authenticity of the service and the correlation between the services are verified by the historical call records of the service,and the calculation rules of the QoS value of the service composition are improved.This paper put real QoS dataset QWS on four testsets which with different scales to verfiy the effectiveness of the proposed method,the simulation results show that the method proposed in this paper can not only improve the success rate of service composition,but also find more execution options of service composition schemes.