Autonomic Service Grids

This project involves investigating how to develop software components that monitor and, eventually, manage network resources that form part of an infrastructure supporting a service grid. The advantages of monitoring network resources and their external environment, and constructing and executing plans based on an analysis of this information, relieves humans of the responsibility of directly managing a resource.

The starting idea of the work will be to experiment with an existing reference model based to a control-loop often referred to as MAPE-K (Monitor, Analyze, Plan, Execute, Knowledge) loop, which has been proposed by IBM as part of its autonomic computing initiative. In this context, the work should study the advantages and disadvantages of the current techniques and propose a new model for managing efficiently network resources. In particular, the work will further investigate the use of high-level policies employed by components, and how they can be employed to constantly analyze and optimize the status of MAPE-K components to automatically adapt themselves to the continuous changing conditions of their environments.

The successful candidate should have experience in software agents and multi-agent systems and an interest in artificial intelligence, computational logic, and autonomic systems. He/she will join the DICE Lab in the Department of Computer Science (http://www.dice.cs.rhul.ac.uk).