Achieving dynamic load balancing through mobile agents in small world P2P networks

Abstract

Peer-to-Peer (P2P) networks are a class of distributed networking and are being deployed in a wide range of applications. Besides such an importance, P2P networks still incur complexities in the resource location policies and in the load balancing techniques of the nodes, especially in unstructured P2P networks. One potential solution to resolve such issues is to enable the P2P networks to evolve into a self-optimizing overlay network topology by identifying the overloaded peers promptly. This paper introduces a new load balancing method in unstructured P2P networks based on mobile agents and resource grouping techniques. We firstly propose a resource grouping strategy to cluster the nodes which have same set of resources, thereby balancing the load among inter-group nodes. On the other hand, load balancing among intra-group nodes is achieved by using the mobile agents monitoring technique. By using this technique, the mobile agents migrate through the nodes in the same group, for the purpose of identifying the possible network congestion. Thus, queries can reach the desired resources more quickly while congested nodes can be identified promptly. The simulation results show that our proposed network evolves into a group-based small world network significantly. The evolved network exhibits robustness and adaptability under external attacking, high query workload, and higher network churns. The simulation results also illustrate that the proposed model achieves better search performance than the DANTE system.