An Efficient Search Scheme in Unstructured Peer-to-Peer Networks

Abstract

In peer-to-peer (P2P) networks, nodes are quite different from each other in many aspects, such as sharing resources, online time and bandwidth. Some approaches have been introduced to take advantage of the query forwarding and answering heterogeneity such that the high bandwidth and query answering capability of nodes can be fully utilized to improve the system performance. In this paper, we suggest using the online time heterogeneity to improve the search efficiency of P2P networks. In our proposed Differentiated Index (Diff-Index) algorithm, the nodes with long online time will have higher priority to be queried. Because the online time is quite different among nodes, much search traffic can be saved by querying only a small portion of a network. The query success rate can be kept high because the nodes sharing a great amount of resources tend to have long online time. Our simulation results show that the Diff-Index algorithm can save 66 percent of search traffic.

Bacterial foraging search in unstructured P2P networks

“…In this method, the metric of selecting ultrapeers is the number of shared files in each node, and this causes ultrapeer overlay contain the main portion of shared files. In [20], the authors proposed formation of an overlay based on online time heterogeneity of nodes, and performing search in that layer.…”

“…The aim of most of these algorithms is to introduce at least one source to the requesting node, regardless of the properties of the found source such as upload bandwidth and length of presence time in the network. In recent years, different search algorithms have been proposed to improve search efficiency by adapting overlay connections according to nodes' capacities [10,16], nodes' interests and similarities [17, 18, and 19], nodes' online time [20], and nodes' query answering capabilities [21,22]. Yet, to the best of our knowledge, no search algorithm has been proposed so far to take peers' upload bandwidth, processing capacity and online duration heterogeneity into account simultaneously in a completely unstructured overlay in which joining peers connect to absolutely random nodes and without any imposed adaptation.…”