Deauthentication and disassociation detection and mitigation scheme using artificial neural network

Abstract

Wireless local area networks (WLAN) are increasingly deployed and widespread worldwide due to the convenience and the low cost that characterized it. However, due to the broadcasting and the shared nature of the wireless medium, WLANs are vulnerable to many kinds of attacks. Although there are many efforts to improve the security of a wireless network, some attacks are inevitable. Attackers can send fake de-authentication or disassociation frames to end the session a victim leading to a denial of service, stolen passwords, and leaks of sensitive information among many other cybercrimes. Effectively detecting such attacks is crucial in today’s critical applications. However, the extant security standards are vulnerable to such an attack, and it is still an open research problem. In this paper, a scheme called D3MS is proposed to detect and mitigate de-authentication and disassociation attack effectively. The aim is to construct a model that can distinguish between benign and fake frames by recognizing the normal behavior of the wireless station before sending the authentication and de-authentication frames. The hypothesis is that the emulating the normal behavior of a benign station prior to the authentication and de-authentication attack is useless. The experimentation results showed the effectiveness of the proposed detection technique. The proposed scheme has improved the detection performance by 64.4% comparing to the related work.