An execution prototype of mobile agent-based peer-to-peer systems

Abstract

Peer-to-peer networks are one of the trends in the field of internetworking. Mobile agent-based technology is a newly proposed technology that can be used in peer-to-peer networks. Mobile agents are small software entity that can migrate freely from node to node in heterogeneous networks and has attracted considerable interest in both academia and industry in recent years. In this paper, we focus on the problem of using mobile agents executing in peer-topeer networks. In our proposed mobile agent-based execution model, mobile agents are classified into three kinds: inner agents, outer agents, and worker agents. Inner agents roam inside peer groups, outer agents routing among peer groups, and outer agents execute users’ tasks on behalf of the users. During agents routing process, the network traffic is considered. For each step, mobile agents select their targets based on a probability distribution. We proved that this probability distribution is the only distribution that both makes inference on the known traffic information and approximates to a unbiased distribution. We also provide some properties of this probability distribution.