Implementation of Distributed Variable Chunk based Switching in Peer to Peer Network

Abstract

The Peer-to-peer (P2P) computing has been one of the emerging technologies, particularly contributing in distributed file sharing. Experimental studies show that for a file download, network congestion or service capacity fluctuation takes minutes to several hours. For a P2P n/w one of the fundamental performance metrics is the average download time. The common approach to analyse the average download time is average service capacity. Heterogeneity and fluctuation have significant impact on service capacity and hence averages the download time. Random Chunk Based Switching is one of the file downloading scheme where, the file to be downloaded is divided into many chunks but the shortcomes of this scheme are mentioned below. User sequentially downloads one chunk at a time. If user gets stuck in a low service capacity peer, downloading a fixed amount of bytes from that peer may take a long time. Chunk size remains fixed and it does not change with time. We have designed and implemented a new scheme Distributed Variable Chunk Based Switching where chunk size changes with time. The downloader will be downloading the file from different peers. If bandwidth available is increased then downloading can complete before specified time. If bandwidth available is decreased then downloader will search another peer with good bandwidth and get it replaced. Our new scheme removes heterogeneity and fluctuation. It also provides a distributed approach to a sequential Random Chunk Based Switching.