Service Location Protocol: Automatic Discovery of IP Network Services
IEEE Internet Computing
Making gnutella-like P2P systems scalable
Proceedings of the 2003 conference on Applications, technologies, architectures, and protocols for computer communications
Algorithms for advanced packet classification with ternary CAMs
Proceedings of the 2005 conference on Applications, technologies, architectures, and protocols for computer communications
Understanding churn in peer-to-peer networks
Proceedings of the 6th ACM SIGCOMM conference on Internet measurement
ATEC '06 Proceedings of the annual conference on USENIX '06 Annual Technical Conference
Proceedings of the 7th ACM SIGCOMM conference on Internet measurement
Proceedings of the 7th ACM SIGCOMM conference on Internet measurement
A generic, self-organizing, and distributed bootstrap service for peer-to-peer networks
IWSOS'07 Proceedings of the Second international conference on Self-Organizing Systems
A survey on automatic configuration of virtual private networks
Computer Networks: The International Journal of Computer and Telecommunications Networking
The Journal of Supercomputing
Hi-index | 0.00 |
So far, bootstrapping constitutes the only centralized task in otherwise decentralized peer-to-peer (P2P) systems. As a contribution to the development of generally applicable decentralized bootstrap mechanisms, in this paper we analyze two decentralized approaches from a practical point of view. We consider local host caches and random address probing for bootstrapping into the BitTorrent DHT as an example for a widely deployed P2P system. Based on the results of an extensive measurement study we show that local host caches allow rejoining the P2P system quickly after short times of disconnection, but are impracticable for infrequent or first-time users. Furthermore, random address probing is feasible using a direct Internet connection with high bandwidth, but is subject to practical issues raised by typical NAT routers and the distribution of ports used by BitTorrent clients. We propose two mechanisms for increasing the performance of random address probing: (1) probing multiple ports per host and (2) hash-based filter-resistant port selection, making distributed bootstrapping feasible even from a practical point of view.