A bandwidth-aware algorithm for solving topology mismatch problem in peer-to-peer networks utilizing the combination of learning automata and X-BOT algorithm

Abstract

Peer-to-peer networks construct an overlay network above underlying networks. These networks are classified as structured and unstructured networks. Unstructured overlay networks are characterized by a relaxed topology where neighboring associations are random in nature and oblivious to the properties of the underlying network which leads to topology mismatch. This usually results in the overlay networks owing several sub-optimal links leading to latency, redundant traffic, wasting bandwidth and finally degrading the performance of the network. So designing an appropriate algorithm for topology adapting can considerably affect the efficiency as much as performance. In this paper, an adaptive and smart algorithm for solving topology mismatch will be proposed by considering the bandwidth of each peer and combining learning automata and X-BOT algorithm. In proposed algorithm, neighboring association constitutes based on bandwidth and by utilizing the information provided by learning automata. The information about peer's bandwidth is kept in its local table used periodically for making decision about appropriate links. Simulations has demonstrated that the proposed algorithm would improve the efficiency through decreasing end to end latency, traffic and appropriate use of bandwidth.