Integration of Zero Trust Architecture and Machine Learning for Improving the Security of Software Defined Networking: A Review
Manar Hamza Bashaa, Wesam S. Bhaya, Nabeel H. Al-A’araji
Abstract
Many add new networks, but management has a lot of work to do as well. Software Defined Networking (SDN) was conceived to address these challenges in a more structured way, SDN allows centralized management of the network and provision of software based traffic understanding making it relatively easier to manage large scale networks. The downside to SDN is its vulnerability to cyber attacks. The more centralized the structure the more efficient it is, however the more specific weaknesses it possesses such as DoS attacks for example. The ``perimeter'' approach to security is outdated with today's security technology. Zero Trust Architecture (ZTA) emphasizes network holistics and continuous verification. Its unique capacity to solve today's security issues makes it interesting. The integration of features such as centralized control from a general view of the SDN and modification of forwarding rules allows easier use of Machine Learning (ML) models. The purpose of this work is to provide a comprehensive analysis of the existing literature that takes into account ZTA concepts and machine learning techniques for the purpose of protecting SDN environments. This work goes on to explore a variety of difficulties that are associated with the successful integration of ZTA and ML within an SDN framework. Finally, the paper finishes by identifying potential approaches that could be pursued in the future for research.