Network coding and MPTCP: Enhancing security and performance in an SDN environment

Abstract

Recently, several attempts have been made to jointly secure end-to-end network communication systems against various malicious attacks, and enhance its performance level. Network coding, which is a step forward towards realizing these systems, promises many advantages such as increased throughput and enhanced reliability. However, several security concerns remain an open issue. A major limitation is the unreliable broadcast nature of network channels. One possible solution is to leverage MPTCP (Multi-Path Transmission Control Protocol), and to utilize a secure variant of the Network Coding (NC) technique to secure reliable data transmission over multiple paths. Such a system is possible when integrated in a Software Defined Networking (SDN) environment. In this paper, the conventional Random Linear Network Coding (RLNC) technique is modified and employed in a secure MPTCP-enabled end-to-end system. This variant is referred to as secret coding, where communicating entities multiply the transmitted data with invertible secret coding matrices to achieve data confidentiality and data availability, while maintaining the desired system performance. Each user will separately derive the coding coefficients of the secret coding matrices from the channel and a common secret, which significantly reduces the communication overhead. The proposed solution exhibits several properties such as high efficiency, flexibility, low error propagation and low execution time. In order to further improve the security and performance of the proposed scheme, the SDN controller is programmed to provide independent paths between the communicating MPTCP users. The obtained results prove that the proposed solution is able to achieve both the desired security level and performance. © 2022 Elsevier Ltd