ozmos: bio-inspired load balancing in a chord-based P2P grid

Abstract

Load balancing in distributed computing systems is an important requirement to make efficient use of all available resources. Envisioning a increase in the scale and dynamicity of future grid systems, fully distributed autonomic solutions are required to address this problem. In this regard, we introduce a load balancing mechanism, called ozmos, that follows the principle of osmosis to relocate tasks between nodes in a P2P based grid. Our solution is based on a Chord overlay upon which bio-inspired agents are deployed to share information about the status of the grid as well as to reschedule tasks between nodes. The key based routing capabilities of Chord are exploited to discover other nodes in the overlay, and to efficiently support relocation of incompatible tasks in heterogeneous grids. By means of a simulation study conducted in various scenarios, we highlight the efficacy of the proposed algorithm in achieving system-wide load balance in grids of different scales, and with both homogenous and heterogeneous resources.