The Limitation of Random Network Coding

TL;DR

This paper explores the capabilities and limitations of random linear network coding in complex multi-source multi-sink networks with diverse demands, providing a simple maximum flow condition to characterize achievable rates.

Contribution

It extends the understanding of random linear network coding from multicast to general multi-source multi-sink networks, offering a new characterization of achievable rates.

Findings

Characterizes all achievable rates using a maximum flow condition.

Shows limitations of random network coding in multi-source multi-sink scenarios.

Provides a framework for analyzing network coding performance in complex networks.

Abstract

It is already known that in multicast (single source, multiple sinks) network, random linear network coding can achieve the maximum flow upper bound. In this paper, we investigate how random linear network coding behaves in general multi-source multi-sink case, where each sink has different demands, and characterize all achievable rate of random linear network coding by a simple maximum flow condition.