The application of techniques from the multiagent systems community
to the Grid is a relatively unexplored research area. Both the agent
and Grid communities develop concepts for distributed computing,
however they do so from from differing points of view [1]. However in
practice, few steps have been taken towards the vision of uniting these
two worlds of distributed systems research. My PhD applies the flexible
coordination technique of interaction protocols, from the field of
multiagent communication, to the problem of Scientific Workflow
modelling, found in the Grid community. This has allowed the typical
features and requirements of an e-Science experiment, to be
understood
in terms of pure coordination and executed in an agent-based,
decentralised, peer-to-peer architecture.
[1] I. Foster, N. Jennings, and C. Kesselman. Brain meets brawn:
Why Grid and Agents need each other, 2004.