Intelligent jamming in wireless networks with applications to 802.11b and other networks

It has long been recognized that complete jamming of wireless networks can be realized by generating continuous noise with sufficient power in the vicinity of the wireless network. There are many disadvantages of this approach including high energy requirements and a high probability of detection. The purpose of this paper is to show that similar jamming effectiveness can be achieved with very low energy requirements and low probability of detection. We discuss various measures of performance for jamming and the role of authentication in denial of service attacks. Then we study and simulate, using OPNET 11.5, the effect of periodic jamming on throughput for an 802.11b network. We add intelligence to the jammer by using knowledge of the protocol and exploiting crucial timings and control packets. Intelligent jamming is shown to be more efficient than continuous jamming in terms of signal duration. The next approach is to use a node or two to exploit the backoff timer to create a denial of service attack. Finally, we discuss how these attacks can be applied to networks with protocols such as MILSTD- 188-220D.