DETECTION AND PREVENTION OF MAN-IN-THE-MIDDLE SPOOFING ATTACKS IN MANETS USING PREDICTIVE TECHNIQUES IN ARTIFICIAL NEURAL NETWORKS (ANN)

Abstract

Detection and Prevention of Man-In-The-Middle Spoofing Attacks in MANETs Using Predictive Techniques in Artificial Neural Networks (ANN)

A Mobile Ad-Hoc Network (MANET) is a convenient wireless infrastructure which presents many advantages in network settings. There are however a lot of challenges faced on such networks; with their relatively much more flexible setups. They are comparatively more susceptible to attacks of various kinds as opposed to their wired cousins. This is due to the non-centralization of their overall architecture. Among these numerous attacks are spoofing, blackhole and man-in-the-middle (MITM) attacks. Each of the identified attack methods, have through numerous researches and discussions been tackled to a great degree. The focus of this particular research was on the man-in-the-middle attacks; presenting a possible way of simulating such attacks on the NS2 platform, and after that, using the results to create an Artificial Neural Network(ANN) based software for attack detection, blacklisting, and node reconfiguration; preventing future attacks in the process. The ANN generated provided, generated an average detected rate of 88.235 for packets. With an average true positive reading of 0.25 as oppose to an average false positive reading of 0.0399 from the confusion matrix of 12 test runs on the final detection model, the research not only offered a productive and less expensive way to perform mitm attacks on simulation platforms, but also identified time as a crucial factor in determining such attacks. It is expected that the findings of this research would present a viable way of porting know MITM attack software to the realized software solution in future; for attack signature creation. This is to train the ANN to offer better attack detection. Systems administrators would find such software useful and easily configurable on their networks making them more secure. The research is also intended to be an opening for future malicious software signature creation to supplement existing Intrusion Detection Systems (IDSs).