Title :
Network reduction for coded multiple-hop networks
Author :
Du, Jinfeng ; Sweeting, Naomi ; Adams, David C. ; Medard, Muriel
Author_Institution :
Research Lab of Electronics, Massachusetts Institute of Technology, Cambridge, 02139, USA
Abstract :
Data transmission over multiple-hop networks is impaired by random deleterious events, and characterizing the probability of error for the end-to-end transmission is challenging as the size of networks grows. Adams et al. showed that, when re-encoding at intermediate nodes is enabled, coded transmission over tandem/parallel links can be reduced to a single equivalent link with a specified probability function. Although iterative application of the tandem/parallel reduction techniques in alternation can simplify the task, they are generally not sufficient to reduce an arbitrary network to a single link. In this paper, we propose upper- and lower- bounding processes to bound the end-to-end probability distribution of a network by combining the parallel/tandem link reduction with the structure of flows over the network. We evaluate the performance of the proposed bounding methods at the 99% success rate of end-to-end data transmission over randomly generated acyclic networks. The numerical results demonstrate that our bounding approaches enable us to characterize a network by a single probability function to a very good precision.
Keywords :
Data communication; Delays; Packet loss; Throughput; Unicast; Upper bound; coded transmission; multi-hop networks; network coding; network reduction;
Conference_Titel :
Communications (ICC), 2015 IEEE International Conference on
Conference_Location :
London, United Kingdom
DOI :
10.1109/ICC.2015.7249034