A generic middleware-based platform for scalable cluster computing

In this paper, we address the design of a generic and scalable platform for cluster computing. The architecture of the platform is based on middleware technology in order to ensure easy distribution of the software components along the participating workstations and to exploit advanced communication techniques, such as event notification and object registration. The computational tasks are referred to as intelligent agents, which are software agents that are capable of executing particular algorithms on input data. The developed platform offers advanced features such as transparent load balancing, task scheduling, run time compilation of agent code and migration of tasks. The architecture of the platform will be outlined from a computational point of view and each component will be described in detail. Furthermore, the engineering aspects of the platform will be covered. In addition, a sample scenario for computational tasks in the area of telecommunication network design and management will be described, together with a performance evaluation of the platform. (C) 2002 Published by Elsevier Science B.V.