Scientific workflow interoperability framework

This paper presents a scientific workflow interoperability framework (SWIF), based on a publish/subscribe asynchronous messaging system, aiming to achieve workflow interoperability. Using the SWIF approach, interoperability can be achieved among workflow management systems that operate remotely. SWIF realises asynchronous processing which reduces dependencies between processes in a workflow management system and presents a decoupled approach, highly suitable to a coarse-grained distributed environment. The design of SWIF is based on a set of web services that follow WS-Eventing specifications, enabling the scientists using the workflow management system to make processes on a workflow engine available to other workflows, so users can register their interests in this event. When this event occurred, all registered workflows receive a notification message. A scientific application example, using Triana, Taverna and Kepler workflow management systems, is demonstrated to verify the approach and evaluate system performance