Files
Abstract
A critical enabler of mass scale software deployment on the cloud will be the ability to compose applications from different providers to create composedapplications. While a number of composite applications are prevalent as manually created mashups, we contend that more automation is needed for composing applications and mediating among them. However, cloud application descriptions currently lack semantic descriptions, making the task of providing any automation infeasible. We present a framework called SemCloud, which uses a specification called SAWADL (Semantically Annotated Web Application Description Language) to provide automated composition and mediation of cloud applications. In our framework, all the applications are semantically described using SAWADL and an extended Artificial Intelligence(AI) graph planner is used to suggest compositions based on the user's goals. The planner also provides support for data mediation. We present an evaluation over real world cloud applications. Based on the annotations, SemCloud was able to automatically generate output for mashups or applications, except for any user interface(UI) code.