Beyond Network Equivalence

In earlier work, we described an equivalence result for network capacity. Roughly, that result is as follows. Given a network of noisy, memoryless, point-to-point channels, replace each channel by a noiseless, memoryless bit pipe of the same capacity yields a new network such that any collection of...

Full description

Bibliographic Details
Main Authors: Koetter, Ralf (Author), Effros, Michelle (Author), Medard, Muriel (Contributor)
Other Authors: Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science (Contributor)
Format: Article
Language:English
Published: Institute of Electrical and Electronics Engineers, 2010-10-22T14:39:17Z.
Subjects:
Online Access:Get fulltext
LEADER 02113 am a22002293u 4500
001 59467
042 |a dc 
100 1 0 |a Koetter, Ralf  |e author 
100 1 0 |a Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science  |e contributor 
100 1 0 |a Medard, Muriel  |e contributor 
100 1 0 |a Medard, Muriel  |e contributor 
700 1 0 |a Effros, Michelle  |e author 
700 1 0 |a Medard, Muriel  |e author 
245 0 0 |a Beyond Network Equivalence 
260 |b Institute of Electrical and Electronics Engineers,   |c 2010-10-22T14:39:17Z. 
856 |z Get fulltext  |u http://hdl.handle.net/1721.1/59467 
520 |a In earlier work, we described an equivalence result for network capacity. Roughly, that result is as follows. Given a network of noisy, memoryless, point-to-point channels, replace each channel by a noiseless, memoryless bit pipe of the same capacity yields a new network such that any collection of demands can be met on the network of noisy links if and only if the same demands can be met on the network of noiseless links. We here expand on these ideas to provide a framework for studying more general networks, including networks containing memoryless broadcast, multiple access, and interference channels. For each network in this broader class, we design two corresponding networks of noiseless, memoryless point-to-point links. These two networks provide upper and lower bounds in the following sense. Fix an arbitrary collection of demands. If the given demands can be met on the lower bounding network, then they can also be met on the original network. Likewise, if the given demands can be met on the original network, then they can also be met on the upper bounding network. 
520 |a California Institute of Technology. Lee Center for Advanced Networking 
520 |a United States. Defense Advanced Research Projects Agency. Information Theory for Mobile Ad-Hoc Networks Program (Flows project under Grant W911NF-07-10029) 
546 |a en_US 
655 7 |a Article 
773 |t 47th Annual Allerton Conference on Communication, Control, and Computing, 2009. Allerton 2009