Web Service Composition Research In Cloud Environment

The rapid development of cloud computing has promoted the coordination and integration of resources in various industries.More and more individuals and organizations migrate their local application resources to the cloud computing resource pool in the form of Web services for users to choose and employ.This service deployment model accelerates the update of the quantity and quality of Web service resources,and web services in the cloud resource pool exploded.Individuals or enterprises continue to create new Web services.While satisfying the needs of users,in order to pursue the maximization of Web service utilization and good user experience,a single-function Web service is combined to form a more valuable Web service combination.Web service composition in cloud computing environment brings new challenges to service computing due to the complexity of the ecosystem.In this paper,we discuss the Web service composition in cloud computing environment.Combining the PSO algorithm which can be parallel computing,for different solution scale,respectively,using serial and parallel computing technology to solve the problem of Web service composition.The following is the main research tasks of this article:(1)Considering the Web service composition problem as a NP problem with diversified and diversified solutions,the defects are improved based on the analysis of standard particle swarm optimization.Mainly divided into three parts,first of all,the use of activity detection mechanism increase the diversity of solutions.Secondly,learn the principle of fireworks explosion mechanism and enhance the search ability of population particles during particle migration.Then,through the reverse learning process of some particles,the traction objects are changed,the flight inversion of the population particles is adjusted,the local optimal problem is exited,and the global optimization ability is improved.The experimental results show that the proposed service composition optimization algorithm can effectively improve the particle optimization ability,and has better performance than the existing methods.(2)Aiming at the quantity and heterogeneity of Web services in cloud computing environment,this paper proposes a improved PSO service composition method based on Spark distributed environment.Firstly,clustering Web services using parallel covering algorithm,the multiple cluster centers as the starting point of particle to improve the regional dispersion of initial population particle distribution.Then,according to the parallel data coding rules of RDD,a large-scale composite service is generated by the parallel migration of sub-populations.Finally,using the particle elitism selection strategy,inert particles are removed to optimize the performance of the combined service selection.In this paper,a real WS-Dream dataset is used to prove that the proposed method can effectively solve the problem of mass Web service composition through a large number of experimental results.