Title :
A "Chicken & Egg" Network Coding Problem
Author :
Harvey, N.J.A. ; Kleinberg, R. ; Nair, C. ; Yunnan Wu
Author_Institution :
MIT, Cambridge, MA
Abstract :
We consider the multi-source network coding problem in cyclic networks. This problem involves several difficulties not found in acyclic networks, due to additional causality requirements. This paper highlights the difficulty of these causality conditions by analyzing two example cyclic networks which are structurally similar. Both networks have an essentially identical network code which appears to transmit all information from the sources to the sinks; however, this network code is invalid since it violates causality. We show that, in one of the networks, the invalid code can be modified to obey causality, whereas in the other network this is impossible. This unachievability result is proven by a new information inequality for causal coding schemes in a simple cyclic network.
Keywords :
source coding; causal coding schemes; cyclic networks; information inequality; multisource network coding problem; Art; Computer networks; Cramer-Rao bounds; Decoding; Ethics; Network coding; Technological innovation;
Conference_Titel :
Information Theory, 2007. ISIT 2007. IEEE International Symposium on
Conference_Location :
Nice
Print_ISBN :
978-1-4244-1397-3
DOI :
10.1109/ISIT.2007.4557215