A Balanced Tree-Based Strategy for Unstructured Media Distribution in P2P Networks

Most research on P2P multimedia streaming assumes that users access video content sequentially and passively. Unlike P2P live streaming in which the peers start playback from the current point of streaming when they join the streaming session, in P2P video-on-demand streaming VCR-like operations such as forward, backward, and random-seek have to be supported. Providing this level of interactive streaming service in a P2P environment is a significant challenge. This paper proposes a balanced binary tree-based strategy for unstructured video-on- demand distribution in P2P networks (BBTU). BBTU assumes videos can be divided into several segments which can be fetched from different peers. BBTU involves two steps: 1) balance binary tree construction based on a prefetching algorithm in order to support interactivity; 2) unstructured video dissemination over network based on gossip protocol, which is the overlay for video distribution. Analysis and simulation show how BBTU is an efficient interactive streaming solution in P2P environment.