Self-Sustaining Platforms: a Semantic Workflow Engine

[email protected] Abstract. In this paper, we provide a novel semantic workflow system, based on semantic functional service descriptions and a rule file. The workflow engine follows a three-step process. First, it determines for all the resources in its knowledge base the functionality they need to progress in the workflow. This uses a phase-functionality rule file which binds phases of the workflow to functionalities. During a second phase, the functionalities are mapped to REST service calls using RESTdesc functional descriptions. During the third step, the engine executes the generated service calls and pushes the resource it acted on to the next phase in the workflow using a phase-transition rule file. The main advantage of this approach is that each step can be influenced by external information from the Linked Open Data cloud. It exploits the fact that Linked Open Data and RESTful Web services and APIs are resource-oriented. Moreover, the workflow rule file makes the system easily adaptable and extensible to achieve new functionalities or to obey changing company policies. Finally, the separation between functional descriptions and service descriptions supports easy management over the fast-changing services at hand.