Skip to main content
Article
Efficient Peer Assignment for Low-Latency Transmission of Scalable Coded Images
Journal of Communication and Networks (2008)
  • Xiao Su, San Jose State University
  • Tao Wang, San Jose State University
Abstract

In this paper, we propose efficient peer assignment algorithms for low-latency transmission of scalable coded images in peer-to-peer networks, in which peers may dynamically join and leave the networks. The objective of our algorithm is to minimize the transmission time of a requested image that is scalable coded. When an image is scalable coded in different bit rates, the bit stream encoded in a lower bit rate is a prefix subset of the one encoded in a higher bit rate. Therefore, a peer with the same requested image coded in any bit rate, even when it is different from the requested rate, may work as a supplying peer. As a result, when a scalable coded image is requested, more supplying peers can be found in peer-to-peer networks to help with the transfer. However, the set of supplying peers is not static during transmission, as the peers in this set may leave the network or finish their transmission at different times. The proposed peer assignment algorithms have taken into account the above constraints. In this paper, we first prove the existence of an optimal peer assignment solution for a simple identity permutation function, and then formulate peer assignment with this identity permutation as a mixed-integer programming problem. Next, we discuss how to address the problem of dynamic peer departures during image transmission. Finally, we carry out experiments to evaluate the performance of proposed peer assignment algorithms.

Keywords
  • coded,
  • low latency
Publication Date
March, 2008
Publisher Statement
SJSU users: use the following link to login and access the article via SJSU databases
Citation Information
Xiao Su and Tao Wang. "Efficient Peer Assignment for Low-Latency Transmission of Scalable Coded Images" Journal of Communication and Networks Vol. 10 (2008)
Available at: http://works.bepress.com/xiao_su/6/