Applying Self-Aggregation to Load Balancing: Experimental Results

One of the today issues in software engineering is to find new effective ways to deal intelligently with the increasing complexity of distributed computing systems. In this context a crucial role is played by the balancing of the work load among all nodes in the system. So far load balancing approaches have been designed for networks with fixed or dynamic topologies. These approaches work well in the case each node knows its similes and is able to contact them to delegate tasks. However, they do not address the needs of more dynamic systems where nodes are able to process different types of jobs and have limited knowledge about their neighbors and the whole system. To address these issue, we are experimenting with the usage of autonomic self-aggregation techniques that rewire such highly dynamic systems in groups of homogeneous nodes that are then able to balance the load among each others. We present our approach and show through simulation that it provides significant advantages under the circumstances described before.