Network Communication with operators in Dedekind Finite and Stably Finite Rings

TL;DR

This paper extends linear network coding to infinite and continuous message spaces using modules over rings, identifying conditions on rings that ensure realistic network capacities and comparing digital and analogue communication performance.

Contribution

It introduces a framework based on Dedekind and stably finite rings to model infinite message spaces in network coding, enabling comparison of digital and analogue methods.

Findings

Modules over Dedekind finite rings prevent unrealistic infinite capacities.

Modules over stably finite rings correspond to realistic capacity constraints.

Digital can outperform analogue, and vice versa, depending on network configuration.

Abstract

Messages in communication networks often are considered as "discrete" taking values in some finite alphabet (e.g. a finite field). However, if we want to consider for example communication based on analogue signals, we will have to consider messages that might be functions selected from an infinite function space. In this paper, we extend linear network coding over finite/discrete alphabets/message space to the infinite/continuous case. The key to our approach is to view the space of operators that acts linearly on a space of signals as a module over a ring. It turns out that modules over many rings $R$ leads to unrealistic network models where communication channels have unlimited capacity. We show that a natural condition to avoid this is equivalent to the ring $R$ being Dedekind finite (or Neumann finite) i.e. each element in $R$ has a left inverse if and only if it has a right inverse. We then consider a strengthened capacity condition and show that this requirement precisely corresponds to the class of (faithful) modules over stably finite rings (or weakly finite). The introduced framework makes it possible to compare the performance of digital and analogue techniques. It turns out that within our model, digital and analogue communication outperforms each other in different situations. More specifically we construct: 1) A communications network where digital communication outperforms analogue communication. 2) A communication network where analogue communication outperforms digital communication. The performance of a communication network is in the finite case usually measured in terms band width (or capacity). We show this notion also remains valid for finite dimensional matrix rings which make it possible (in principle) to establish gain of digital versus analogue (analogue versus digital) communications.