Random Walk Search on Concatenated Hop-Limited Trees Embedded into Unstructured P2Ps

Abstract

We propose a random-walk-based file search for unstructured P2P networks. In the proposal, each node keeps two pieces of information, one is the hop-limited routing tables, storing the route information to nodes within a small number (say n) hops from each node. The hop-limited routing table is substantially equivalent to the hop-limited shortest path tree, which is made of nodes within limited hop distance from the root. The other is the file list, which is an index over files kept by nodes within n hops from each node. To find a file, a node first checks its file list. If the requested file is found in the list, the node sends a file request message to the file owner, otherwise, it sends a file-search message to a randomly-selected leaf node on the hop-limited shortest path tree rooted at itself. Numerical examples show that our proposal has much smaller search latency and lower file-search failure ratio than the primitive random-walk-based search, while it wastes much less network bandwidth than network-broadcast-based searches.