Detecting Attacks in Wireless Mesh Networks

Abstract

With the advent of various types of wireless networks in the next generation to provide users with better services and faster technology, wireless mesh networks have been gaining immense popularity. With their unique architectural features and fast connectivity wireless mesh networks promise to extend performance beyond what is obtained with the current WI-FI based infrastructure. Wireless mesh networks not only provide users with unmatched connectivity and performance but also has features like dynamically self-organized and self-configuring. This feature brings many advantages to WMNs such as low-upfront cost, easy network maintenance, robustness and reliable service coverage. Despite all the benefits, wireless mesh networks are highly susceptible to internal as well as external attacks due to the open nature of the wireless medium. Although external attacks can be mitigated using a combination of cryptography based and robust communication techniques, the internal attacks, which exploit the knowledge of the network secrets and protocol semantics, are difficult to detect and require protocols with built-in security measures. Through this project I try to address the problem of Jamming attacks in wireless mesh networks. In these type of attacks the attacker takes advantage of the network secrets and then classifies the packets which are of high importance .To perform Selective jamming, the attacker must be capable of classifying the transmitted packets and corrupting them before the end of their transmission. Through this project I try to detect the various kinds of security threats that hinder the performance of wireless mesh networks. Also, try to formulate new strategies and also improve existing ones to mitigate these attacks on WMNs. Furthermore I am interested in identifying the loopholes of the existing strategies to make them more robust and make them efficient to detect and mitigate attacks more effectively. Hence, to provide a reliable communication network.