Performance Evaluation of Lookup Mechanisms in Structured and Unstructured P2P Networks under the Impacts of Churn
Adam Chang · UC Research Repository (University of Canterbury) · 2010
This Master thesis investigates the performance of the lookup mechanisms in structured and unstructured Peer-to-Peer (P2P) networks under the impacts of churn. The considered performance metrics include the distance of matched lookups, bandwidth consumption and latencies. We have selected to study Chord, Kademlia and GIA, because these selected P2P networks possess distinctive characteristics. Chord is one of the first structured P2P networks that implements Distributed Hash Tables (DHTs). Kademlia is a more recent structured P2P network, with the XOR mechanism for improving distance calculation. GIA is an unstructured P2P network which has been proposed as an alternative to structured P2P networks based on Distributed Hash Tables. It implements a dynamic topology adaptation algorithm, flow control and biased random walks. The hypothesis was that unstructured P2P network such as GIA could operate better than structured networks such as Chord and Kademlia. Our research involved a series of simulation studies using two network simulators OverSim and OMNeT++. The simulation was also used to investigate the influence of iterative and recursive routing methods, scalability, and resiliency issues, on the performance of the file lookup mechanisms in the selected P2P networks. The results show that, in considered scenarios, GIA performs better than structured P2P networks, especially in resolving file lookups with a match that is closer to the source peer, thus conserving bandwidth.