Vector Network Coding Based on Subspace Codes Outperforms Scalar Linear Network Coding

TL;DR

This paper demonstrates that vector network coding using subspace and rank-metric codes can significantly reduce the required alphabet size compared to scalar linear solutions in multicast networks, especially as the number of messages increases.

Contribution

It introduces new vector network coding solutions based on subspace codes that outperform scalar linear coding in terms of alphabet size for multicast networks.

Findings

Vector solutions reduce alphabet size compared to scalar solutions.

Achieved gap in alphabet size grows exponentially with message number and vector length.

New family of subspace codes from rank-metric codes is proposed.

Abstract

This paper considers vector network coding solutions based on rank-metric codes and subspace codes. The main result of this paper is that vector solutions can significantly reduce the required alphabet size compared to the optimal scalar linear solution for the same multicast network. The multicast networks considered in this paper have one source with $h$ messages, and the vector solution is over a field of size $q$ with vectors of length~$t$. For a given network, let the smallest field size for which the network has a scalar linear solution be $q_s$, then the gap in the alphabet size between the vector solution and the scalar linear solution is defined to be $q_s-q^t$. In this contribution, the achieved gap is $q^{(h-2)t^2/h + o(t)}$ for any $q \geq 2$ and any even $h \geq 4$. If $h \geq 5$ is odd, then the achieved gap of the alphabet size is $q^{(h-3)t^2/(h-1) + o(t)}$. Previously, only a gap of size size one had been shown for networks with a very large number of messages. These results imply the same gap of the alphabet size between the optimal scalar linear and some scalar nonlinear network coding solution for multicast networks. For three messages, we also show an advantage of vector network coding, while for two messages the problem remains open. Several networks are considered, all of them are generalizations and modifications of the well-known combination networks. The vector network codes that are used as solutions for those networks are based on subspace codes, particularly subspace codes obtained from rank-metric codes. Some of these codes form a new family of subspace codes, which poses a new research problem.