Integrating Machine Learning for Network Threat Detection with SmartX Multi-Sec Framework

Abstract

Security is an essential element for the effective operation of modern network infrastructures. OF@TEIN playground over TEIN (Trans-Eurasia Information Network), is a multi-site cloud with distributed edge nodes has brought the security challenges. In order to address the security challenges of distributed edge nodes, recently, SmartX Multi-Sec framework has been launched. However, it is based on a signature-based network threat detection mechanism. Depending on the current complex and heterogeneous nature of network traffic which is continuously increasing in scale, machine learning capabilities are more effective to recognize hidden and complex patterns of network traffic than traditional signature-based methods. Moreover, in response to the weaknesses of current security solutions, the zero-trust model manipulates that no network is reliable. Therefore, cloud operators should monitor and verify all network traffic continuously. So, by considering the above challenges, we are going to enhance the SmartX Multi-Sec framework by focusing on zero trust and incorporating machine learning techniques for network threat detection (specifically DDoS attacks). Moreover, its prototype version has been realized on the OF@TEIN playground to show its effectiveness.