Eavesdropping Minimization via Transmission Power Control in Ad-Hoc Wireless Networks

Abstract

Reconnaissance activity is the most frequent incident on computer networks since 2002. In fact, most attacks (including DoS attacks) are usually preceded by reconnaissance activity. In order to defend against reconnaissance activity in ad-hoc wireless networks, we propose to use transmission power control as an effective mean to minimize the eavesdropping risk. Our main contributions are as follows: first, we cast the w-th order eavesdropping risk as the maximum probability of packets being eavesdropped when there are w adversarial nodes in the network. Second, we derive the closed-form solution of the 1st order eavesdropping risk as a 3rd-order polynomial function of normalized transmission radius. This derivation is based on the recently proposed model by El Gamal which assumes a uniform distribution of user nodes. Then we generalize the model to allow arbitrary user nodes distribution and prove that the uniform user distribution actually minimizes the 1st order eavesdropping risk. This result plays an essential role in deriving the first analytical bounds for the eavesdropping risk given arbitrary user distribution. Our simulation results show that for a wide range of non-uniform traffic patterns, the eavesdropping risk has the same order of magnitude as the corresponding uniform traffic cases