Decentralized Load Balancing on Unstructured Peer-2-Peer Computing Grids
Raymond Hu · 2006
We present, analyze, and measure a new algorithm to perform decentralized load balancing of computing jobs in computational Peer-2-Peer (P2P) grids. The unstructured P2P network consists of nodes which only know about their nearest neighbors as defined by a one hop communication path. With limited local information, our algorithm converges to a global load balancing state. We demonstrate the algorithm convergence both by simulations and by theoretical analysis.