A Distributed Minimum-Distortion Routing Algorithm with In-Network Data Processing

—In many wired and wireless networks, nodes process input traffic to satisfy a network constraint (e.g. capacity constraints) and to increase the utility of data in the output flows given these constraints. In this paper we focus on the special case in which data processing is applied to satisfy capacity constraints. This occurs when the sum of the rate of the input traffic at a node exceeds the sum of the capacity of its output links, or in a more general case, when the sum of the input rates is larger than any cut capacity in the network. In this case, nodes process data to decrease the output flow rate. This decrease from input rate to output rate distorts the transmitted data, which we characterize by a distortion metric. We first show that if links that connected to the data sink have the same capacity limits, then the distortion cost of distributively processing input traffic in a network is equal to the sum of the distortion increments at individual nodes. We then presen...

• Type: Conference
• Year: 2008
• Where: INFOCOM
• Authors: Ramin Khalili, James F. Kurose
• Keywords: Capacity Constraints, Communications, INFOCOM 2008, Input Traffic, Nodes Process