Using Machine Learning for Detecting Timing Side-Channel Attacks in SDN

Abstract

Software-Defined Networking (SDN) is a networking technology that allows for the programming and efficient management of networks. Due to the separation of the data plane and the control plane, SDN is prone to timing side-channel attacks. The adversary can use timing information to obtain data about the network such as flow tables, routes, controller types, ports, and so on. The focus of current mitigation strategies for timing side-channel attacks is largely on minimizing them through network architectural changes. This adds considerable overhead to the SDNs and makes establishing the origin of the attack a challenge. In this paper, we propose a machine learning-based approach for detecting timing side-channel attacks and identifying their source in SDNs. We adopt the machine learning methodology for this solution since it delivers faster and more accurate output. As opposed to conventional methods, it can precisely detect timing side-channel activity in SDN and determine the attacker’s origin. Because this security solution is intended to be used in association with SDN, its architecture ensures that it has a low impact on network traffic and resource consumption. The overall design findings indicate that our method is effective in detecting timing side-channel attacks in SDN and accurately identifying the attacker’s machine.