Compression and Quantitative Analysis of Buffer Map Message in P2P Streaming System

BM compression is a straightforward and operable way to reduce buffer message length as well as to improve system performance. In this paper, we thoroughly discuss the principles and protocol progress of different compression schemes, and for the first time present an original compression scheme which can nearly remove all redundant information from buffer message. Theoretical limit of compression rates are deduced in the theory of information. Through the analysis of information content and simulation with our measured BM trace of UUSee, the validity and superiority of our compression scheme are validated in term of compression ratio. In recent years, P2P streaming media system has gotten explosive development. As an important and hot section in P2P world, the P2P streaming system is attracting a lot research eyes. Many works such as [4]-[13] try to characterize the system's features in nature based on network measurement. With deeper understanding the system, some system models are proposed to improve the system performance in terms of information broadcast speed, network sharing efficiency and startup delay, etc. Although some of the models, which are based on real P2P streaming applications, are engineering feasible, majority of them stay on the authors' desks due to too far away to the reality. In 2008, Libo [1] found that increasing the exchange rate of peers' buffer map message (BM) could help chunk diffusion. This find opens another convenient door for us to improve the P2P streaming system's performance. Naturally, decreasing the cycle of BM exchange will increase the BM protocol overhead in term of network traffic, which will lead to uncertain negative effects on system performance. Nearly at the same time, we found that some mainstream ASPs including PPLive and UUSee have reduced their BM length while kept buffer width and BM exchange cycle unchanged. Our findings confirms that even under the same BM exchange cycle as before, the original BM of more than 80 bytes length has brought non-ignorable overhead. It gives us the BM overhead baseline, which any performance optimization solution must be kept below. Under this principle, the original BM must be scaled down. In general, there are several ways to shorten the BM, such as decreasing the buffer width, enlarge the chunk size, and compressing the bitmap. Both buffer width and chunk size are the system design parameters, and any slight adjust on them may affect the whole situation and bring systematic challenges. In practice, …