Distributed Load Balancing in Peer-to-Peer Computing

In this paper, we address the load balancing problem in the context of peer-to-peer computing environments. The key challenge to employ peer-to-peer networks for distributed computing is to exploit the heterogeneous processing capability of the participating hosts as well as the diverse network conditions. The contribution of our work is twofold. First, we model the load balance problem as an optimization problem with the objective of minimizing the system response time. This modeling considers not only the current loading of hosts, but also the fluctuation of network delay, which completely captures the characteristics of the P2P systems. Second, we propose a gradient projection algorithm to solve the optimization problem, which is fully distributed and easy for implementation. Simulation results demonstrate that our scheme has satisfied performance in terms of convergence, response time and load distribution.