Providing fault tolerance for peer-to-peer streaming service

Abstract

Because each node in the peer-to-peer network is autonomous, it is difficult to provide deterministic service on the network. In this paper, we propose a framework for a fault tolerant streaming service on a peer-to-peer network. The proposed framework provides a service migration-based fault tolerance mechanism to cope with both node/link failures and link state changes. Simulation results show that using a small-sized buffer, the proposed framework can provide hiccup-free streaming service and the overhead required to maintain the up-to-date node state information is negligible. © Springer-Verlag Berlin Heidelberg 2003.