First-Class Protocols for Agent-Based Coordination of Scientific Instruments

The coordination of distributed processing is of great interest to a number of research communities. However these research communities, such as those involved in e-science GRIDs and multi-agent systems, view this problem from disparate viewpoints. This paper aims to contribute to the necessary reconciliation of these perspectives. By demonstrating the coordination of processes whether reactive (such as web services) or proactive (such as autonomous agents) can be done with a single representation using a protocol language, RASA The multi-agent paradigm introduces, into any model of distributed systems, flexibility and autonomy that can be daunting and intimidating to scientists accustomed to more orthodox approaches. It is for this reason that it is important that the model for coordination of this system is reliable, verifiable, inspectable, referable, composable and executable. The language RASA provides this functionality, which we extend its use for not only agent interaction protocols but also to express workflows. The language for expression then becomes the domain of discourse as well as potentially the language for the workflow´s execution.