On the Capacity of p2p Multipoint Video Conference

TL;DR

This paper analyzes the maximum achievable video delivery rate in peer-to-peer multipoint video conferences, proposing a capacity measure and demonstrating a lower bound that exceeds previous approaches, with constructive proofs applicable to real systems.

Contribution

It introduces a formal capacity measure for P2P multipoint video conferences and provides a constructive proof establishing a new lower bound on conference capacity.

Findings

Capacity lower bound of 5/6 for single video scenarios

Constructive proofs suitable for real implementation

Capacity exceeds previous chained approach rate of 1/2

Abstract

In this paper, The structure of video conference is formulated and the peer-assisted distribution scheme is constructed to achieve optimal video delivery rate in each sub-conference. The capacity of conference is proposed to referee the video rate that can be supported in every possible scenario. We have proved that, in case of one user watching only one video, 5/6 is a lower bound of the capacity which is much larger than 1/2, the achievable rate of chained approach in [2]. Almost all proofs in this paper are constructive. They can be applied into real implementation directly with a few modifications.