A multicast-based bootstrap mechanism for self-organizing P2P networks

  • Authors:
  • Simone Cirani;Luca Veltri

  • Affiliations:
  • Department of Information Engineering, University of Parma, Parma, Italy;Department of Information Engineering, University of Parma, Parma, Parma

  • Venue:
  • GLOBECOM'09 Proceedings of the 28th IEEE conference on Global telecommunications
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The peer-to-peer (P2P) network paradigm has been introduced in order to overcome some shortcomings of the client-server architecture by providing such features as decentralization, self-organization, scalability, and fault-tolerance. Bootstrapping is the initial process through which new nodes can join an existing P2P overlay network. Typically, a joining peer must first contact a bootstrap peer, which is a peer already enrolled in the overlay. The bootstrap peer is responsible for admitting the new peer by passing information about other peers so that the new peer can actively participate in the overlay. Finding a suitable bootstrap peer is therefore a critical issue. Although different P2P systems have been defined and deployed, the problem of bootstrapping has usually been solved by introducing such mechanisms as the use of a pre-configured list of nodes, caching, or server-based discovery. Unfortunately, although they work in P2P applications running over the Internet, they show some problems when applied to very dynamic and self-organizing intranet or enterprise network scenarios. In fact, in these cases all nodes may join and leave the network very dynamically, without the possibility of guaranteeing any sort of permanent centralized service as current bootstrap solutions may require. In this paper, we propose a multicast-based bootstrapping mechanism for dynamic and self-organized P2P networks that allows a joining peer to discover a proper bootstrap peer in a real distributed manner. The proposed mechanism uses an unsolicited approach and performs well in terms of scalability, load-balancing, and a mean frequency of information exchange. The paper defines the algorithm and proposes an implementation of a suitable communication protocol.