Service identification by enhanced K-mean algorithm in service-oriented architecture

Abstract

Service identification with right level of granularity is the most critical aspect in service oriented architecture. Service identification is challenging due to several reasons such as lack of reuse of services, lack of right decision to choose the appropriate service. SOA-based project uses large repository, where in services are stored randomly in the repository. This causes considerable amount of search time when service is searched from the database. In case of trigger-based application, current service storage process will not be much effective due to inadequacy in service level agreement. In this regard, the author(s) explored the possibilities of service identification using k-means clustering which will reduce search time, service replication and increase the performance and reliability of the service. Proposed model has been experimentally validated and author(s) found significant decrease in service search time. This model will be helpful in building the applications with minimum service level agreement.