earticle

영어

As a common delivery model in cloud computing, SaaS applications are becoming increasingly popular. Multi-tenancy is a key characteristics of SaaS applications. Service composition plays a main role in SaaS applications because of frequent composability and reusability of software services. With the development of cloud computing, one side, there are an increasing number of available services, other side, tenant’ individual and diverse requirement become more intense, which makes service composition tend to be more complex. This paper proposes an effective infrastructure for service composition, MultiGranular, which provides a semantic basis for multi-tenant service composition. The MultiGranular supports the characteristic of hierarchy, uncertain correlation, inheritance and versioning, effectively responding to the requirements of complexity of service composition for cloud computing SaaS applications. The computing process of MultiGranular includes three steps: basic service clustering, correlation mining and hierarchy building. MultiGranular makes it possible to change disorder services into the hierarchy and ordered clustering services, and makes it easy to develop multi-tenant SaaS applications. The experimental results demonstrate that by utilizing the proposed method, complexity of multi-tenant service composition are solved and the efficiency of service composition has been improved greatly.