Document Type


Publication Date

February 2014


Video distribution over the Internet has become a popular service because of technological advances in internet (e.g., higher network bandwidth) and video coding (e.g., H.264/SVC). In this and other similar media distribution applications, a server or distribution center sends a media/video to a group peers with different bandwidth resources and display capacities. In one of the approaches, the peer-to-peer approach, the server sends only one copy of the media over Internet, and each peer receives one segment of the media and exchanges his/her segment with other peers to receive the complete media. A key design issue in this approach is deciding the sizes of the segments delivered to individual peer which affect the time of complete media distribution. Equal sized segmentation does not always result in the least distribution time. In this paper, we study the problem of how to distribute non-scalable and scalable coded media from a server in closed peer-to-peer based IPTV networks. We propose a new distribution algorithm to find the media segment sizes optimized for the bandwidths of participating peers in order to minimize the time it takes to distribute the entire media to all end subscribers. First, we focus on finding the optimal solution in non-scalable media distribution. Then, we extend our method to scalable media distribution to find optimal segment sizes for all media layers. Simulations are conducted by varying the number of peers and media sizes to investigate the impact of these parameters on both non-scalable and scalable video distribution. The experimental results have demonstrated the scalability and efficiency of the proposed distribution algorithm.


This is an Author's Accepted Manuscript of an article originally published in Computer Networks Volume 60, Issue 26 by Elsevier on February 2014, DOI: 10.1016/j.bjp.2013.11.007. This work is licensed under a Creative Commons CC_BY-NC-ND International License.
The Version of Record can also be found online at this link.