Research About Selection Of Web Service Based On QoS In SOA

Web Service is the main approach to SOA(Service-Oriented Architecture)presently. With the wide application of SOA system, Web Service tends to be moreand more widely used in the internet. Therefore some Web Services come into beingfor similar or even the same purposes. Faced these services, user can make theydecision to choose which one by evaluate the QoS(Quality of Service) ofservices.Web Service selection algorithm based on QoS is of primary importance inthe Web Service selection, and the algorithm based on the evaluation model of QoS.Based on the profound research in techniques and methodology related to SOAand Web Service, this paper summarizes on their relationship. As a discussion focusin recent scientific study, the Web Service selection algorithm and the evaluationmodel of QoS are also discussed in profound details in this paper, with the mainfinding and accomplishment listed below:Firstly, based on the profound study of existing QoS model and focusing on theircurrent deficiency, a QoS evaluation model for complex services composition issuggested in this paper. Following this, a description string of composition isproposed, based on which a QoS evaluation algorithm for complex servicescomposition is also provided. At last, an empirical emulation of the QoS evaluationmodel for both basal services composition and complete sevices composition provesthat the algorithm is useful.Web Service selection based on QoS plays a key role in the SOA system based onWeb Service. In this paper, current deficiencies of the relevant Generation algorithmis analyzed before suggesting an improved and QoS-based generation algorithm onWeb Service selection issues. Through the ways of improving the encoding manner ofchromosomes, the evaluation of adoptability, the crossover-ratio as well as themutation rate, an empirical emulation proves that the improved algorithm proposed inthis paper is multi-route in description, efficient in operating, and powerful infull-scale research.