Modeling and Detection of Flooding-Based Denial-of-Service Attack in Wireless Ad Hoc Network Using Bayesian Inference

Abstract

Wireless ad hoc networks are widely useful in locations where the existing infrastructure is difficult to use, especially during the situations like flood, earthquakes, and other natural or man-made calamities. Lack of centralized management and absence of secure boundaries make these networks vulnerable to various types of attacks. Moreover, the mobile nodes used in these networks have limited computational capability, memory, and battery backup. Flooding-based denial-of-service (DoS) attack, which results in denial of sleep attack, targets the mobile node's constrained resources which results in excess consumption of battery backup. In SYN flooding-based DoS attack, the attacker sends a large number of spoofed SYN packets which not only overflow the target buffer but also creates network congestion. The present article is divided into three parts: 1) mathematical modeling for SYN traffic in the network using Bayesian inference; 2) proving the equivalence of Bayesian inference with exponential weighted moving average; and 3) developing an efficient algorithm for the detection of SYN flooding attack using Bayesian inference. Based on the comprehensive evaluation using mathematical modeling and simulation, the proposed method can successfully defend any type of flooding-based DoS attack in wireless ad hoc network with higher detection accuracy and extremely lower false detection rate.