Novel compressed linear network coding vectors for multihop communication networks

Abstract

Random Linear Network Coding (RLNC) is well-known to provide high throughput and low latency for vast communication networks. However, RLNC often suffers from high coefficients overhead, specifically, when it’s applied to limited resource or short-packet networks. Herein, the problem of RLNC coefficients vector overhead is revisited. A novel framework, based on modular arithmetic and prime numbers, and influenced by the Chinese remainder theorem (CRT), is proposed to reduce the coefficients overhead by augmenting only a tiny one item coefficient instead of the entire coefficients vector. The proposed method successfully addresses all the shortcomings of previous methods, including restrictions on generation size and packet density, recoding on intermediate nodes, and creating innovative coding vectors. Theoretical analysis and experimental demonstrate the superior performance of the proposed scheme in terms of coefficients overhead ratio, download time, throughput, and packet drop rate. This evaluation has considered two types of networks: wireless sensors network for Internet of things, and conventional wireline Ethernet.