Improving the Performance and Robustness of Kademlia-Based Overlay Networks

Structured peer-to-peer (p2p) networks are highly distributed systems with a potential to support business applications. There are numerous different suggestions on how to implement such systems. However, before legal p2p systems can become mainstream they need to offer improved efficiency, robustness, and stability. While Chord is the most researched and best understood mechanism, the Kademlia algorithm is widely-used in deployed applications. There are still many open questions concerning the performance of the latter. In this paper we identify the main problems of Kademlia by large scale simulations and present modifications which help to avoid those problems. This way, we are able to significantly improve the performance and robustness of Kademlia-based applications, especially in times of churn and in unstable states. In particular, we show how to increase the stability of the overlay, make searches more efficient, and adapt the maintenance traffic to the current churn rate in a self-organizing way.

[1]  Robert Morris,et al.  Chord: A scalable peer-to-peer lookup service for internet applications , 2001, SIGCOMM 2001.

[2]  Krzysztof Pawlikowski,et al.  On credibility of simulation studies of telecommunication networks , 2002, IEEE Commun. Mag..

[3]  Seif Haridi,et al.  A Statistical Theory of Chord Under Churn , 2005, IPTPS.

[4]  Daniel Stutzbach,et al.  Improving Lookup Performance Over a Widely-Deployed DHT , 2006, Proceedings IEEE INFOCOM 2006. 25TH IEEE International Conference on Computer Communications.

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

[6]  Dirk Staehle,et al.  On the stability of chord-based P2P systems , 2005, GLOBECOM '05. IEEE Global Telecommunications Conference, 2005..

[7]  Youki Kadobayashi.  Achieving heterogeneity and fairness in Kademlia , 2004, 2004 International Symposium on Applications and the Internet Workshops. 2004 Workshops..

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

[9]  Robert Tappan Morris,et al.  Comparing the Performance of Distributed Hash Tables Under Churn , 2004, IPTPS.

[10]  David R. Karger,et al.  Chord: A scalable peer-to-peer lookup service for internet applications , 2001, SIGCOMM '01.