A case for taxation in peer-to-peer streaming broadcast

Abstract

Most existing research on peer-to-peer (p2p) has been on file sharing applications. In this paper, we focus on p2p streaming applications. In particular, we argue that the Bit-for-Bit model, widely adopted in p2p file sharing, is not applicable in p2p streaming. In p2p streaming, the bottleneck resource is the upstream bandwidth capacity. Our empirical experience with p2p streaming indicates that a large percent of peers on the Internet have limited upstream bandwidth capacity, and the Bit-for-Bit model severely limits the amount of bandwidth these resource-poor peers can receive. To address this issue, we propose a taxation model. In the taxation model, resource-rich peers contribute more bandwidth to the system, and subsidize for the resource-poor peers. This redistribution of wealth improves social welfare. Such a model is applicable in the streaming context because the publisher of the video stream has the means to enforce taxation on peers and the will to maximize their collective social welfare. We design a simple linear taxation scheme and incorporate it in a distributed streaming protocol. Our simulation results indicate that taxation can significantly improve social welfare without incurring a significant overhead to the system.