A Machine Learning centric NIDS Architecture for SDN-based Cloud IoT Networks

Abstract

The exponential development in smart gadgets with all-round connection has hugely reduced the flow inside the cloud Internet of Things (IoT) and generated possible cyber-attack surfaces. Traditional security techniques to handle security risks in cloud-based IoT networks are insufficient and ineffective. Software Defined Networking (SDN), Network Function Virtualization (NFV), and Machine Learning Technologies provide several advantages to tackle cyber security problems for fog IoT devices. This article proposes collaboration and smart network-based design for IoT networks, called NIDS, for access control based on SDNs. It consists of a hierarchic level of smart IDS nodes that work together to identify abnormalities and create policies in SDN-based integrated application devices to block malicious traffic at the quickest opportunity. First we outline a novel NIDS architecture with an extensive small network and track decision evaluation. Next, the logic of the control system is explored comprehensively by the major sequential procedures comprising initialization, realtime operations and database updates. Then we build the developed model in complete in an SDN-based ecosystem and undertake a number of tests. Finally, the NIDS architecture assessment findings offer great results in anomalous identification and reduction and the treatment of problem bottleneck mostly in SDN-based cloud Iot systems compared with existing alternatives.