Experimental Design and Analysis of an Enhanced SDN-Based Model for Intrusion Detection

The diversification and growth of cyberspace in the world always increases exponentially. This Antecedent increased in network hurdle and complexity of managing associated network traffic. Traditionally, organizations amplify their network bandwidth with centering on purchasing more hardware to patch up such difficulty. This approach is a drawback ever since it could be an expensive lapse if the supplementary network resources are not fully utilized. Furthermore, it is habitually sometimes unfeasible to manage the traditional network in a timely manner to respond to errors and malicious activities. Software-Defined Network (SDN) has been emerging technology developed to reduce network complexity and enhanced management of the entire network effectively. SDN is an efficient and robust architecture capable to control the network from centralized point unlike the traditional network architecture. SDN discrete the data and control activity of data communications devices, with an Application programming Interface (APT). The API is a set of rules and protocols that operate as intermediaries allowing communication between several applications without revealing the inner workings of the network. Nowadays, different data center’s network broadly recognizes and implemented SDN, yet, it is facing many threats which are still challenging on securing the SDN architecture. Security problems related to SDN design include, securing network devices communication, controlling the capacity of applications and risk of a compromised SDN controller threat. Therefore, deploying Intrusion Detection Systems (IDSs) to observe and manage malicious activities is an essential factor of the network design. In this work we developed an enhanced, secured and robust SDN model for intrusion detection using snort protection mechanism. The proposed SDN network is tested for quality of service (QoS) under three different network scenarios and measure the performance based on bandwidth, throughput, latency, packet loss and CPU usage. The simulation involved generating normal and various attack traffic with a different packet sizes ranging from 128 bytes to 2,048 bytes. Comparative analysis with the existing model prove the improve quality of service of the proposed SDN network.