Improving content retrieval recall in KAD

KAD is the largest DHT system with several million simultaneous users. The dynamic of peer participation which is called peer churn affects the performance of content retrieval in KAD. In this paper, we performed a simple way to improve content retrieval recall in KAD, by taking highly available contacts as lookup entries. We track highly available peers in KAD by the ID appearance ranks in routing tables from an running crawler designed by ourselves. The measurement study, compared with the normal content retrieval process, shows that the improved way is much effective.

[1]  Moritz Steiner,et al.  Faster Content Access in KAD , 2008, 2008 Eighth International Conference on Peer-to-Peer Computing.

[2]  Zhoujun Li,et al.  ID repetition in Kad , 2009, 2009 IEEE Ninth International Conference on Peer-to-Peer Computing.

[3]  John Kubiatowicz,et al.  Handling churn in a DHT , 2004 .

[4]  Daniel Stutzbach,et al.  Understanding churn in peer-to-peer networks , 2006, IMC '06.

[5]  Taoufik En-Najjary,et al.  Long Term Study of Peer Behavior in the kad DHT , 2009, IEEE/ACM Transactions on Networking.

[6]  David Mazières,et al.  Kademlia: A Peer-to-Peer Information System Based on the XOR Metric , 2002, IPTPS.