Node clustering based on link delay in P2P networks

  • Authors:
  • Wei Zheng;Sheng Zhang;Yi Ouyang;Fillia Makedon;James Ford

  • Affiliations:
  • Dartmouth College, NH;Dartmouth College, NH;Dartmouth College, NH;Dartmouth College, NH;Dartmouth College, NH

  • Venue:
  • Proceedings of the 2005 ACM symposium on Applied computing
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Peer-to-peer (P2P) has become an important computing model because of its adaptation, self-organization and autonomy etc. But efficient organization of the nodes in P2P networks is still a challenge needs to be addressed. Node clustering is a mechanism that aims to provide an optimal infrastructure to organize the nodes in a P2P network. This paper describes an approach to implement node clustering based on link delay of node communications in the P2P network. This approach is completely distributed, in which each node only depends on its neighbors to implement node clustering. In this approach, we propose two distributed algorithms: T-closure algorithm and hierarchical node clustering algorithm to find node clusters automatically in a P2P network. We explore the node connectivity together with the connection quality. As a result, the link delay of communication between the super-node and the peer-node in node clustering can be limited, which will improve the overall performance of P2P networks.