A Population Dynamics Model for Data Streaming over P2P Networks

Data streaming (DS) over peer-to-peer (P2P) networks has been intensively studied in recent years and there have been various schemes proposed already. To evaluate these schemes, either measurement in experimental implementations, or simulation and theoretical analysis have been used. The former is inadequate as data are collected from different experiments, while the latter lacks a proper theoretical dynamics model. Our research aims at providing a general theoretical model to evaluate DS over P2P systems and analyze their dynamic behaviors. In this paper, with the analysis and abstraction of the characteristics of peers and their organization in DS over P2P, we propose a general population dynamics model for DS over P2P with fixed population. The model depicts the dynamic distribution of peers as a closed Markov queuing network. In particular, the model is scheme-independent and can be used with various schemes. Through theoretical analysis, we prove the model has equilibrium and only one closed-form solution. Besides, we verify the model through simulations, and show that it is a helpful analytical tool with a case study.