On Source-Channel Separation in Networks

We consider the source-channel separation architecture for lossy source coding in communication networks. It is shown that the separation approach is optimal when the memoryless sources at source nodes are arbitrarily correlated, each of which is to be reconstructed at possibly multiple destinations within certain distortions, and the channels in this network are synchronized, orthogonal and memoryless (i.e., noisy graphs). The extracted pure network source-coding problem has to incorporate user interactions and the corresponding causality constraints, which suggests a distinct research direction into interactive network source coding that has not received much attention in the literature. The separation result in this paper is a companion to results of a similar flavor established recently in [11] for the case of independent sources over communication networks with more general multiuser channels.