Doctor of Philosophy (PhD)
A wireless sensor network (WSN) is a collection of wireless nodes, usually with limited computing resources and available energy. The medium access control layer (MAC layer) directly guides the radio hardware and manages access to the radio spectrum in controlled way. A top priority for a WSN MAC protocol is to conserve energy, however tailoring the algorithm for this purpose can create or expose a number of security vulnerabilities. In particular, a regular duty cycle makes a node vulnerable to periodic jamming attacks. This vulnerability limits the use of use of a WSN in applications requiring high levels of security. We present a new WSN MAC protocol, RSMAC (Random Sleep MAC) that is designed to provide resistance to periodic jamming attacks while maintaining elements that are essential to WSN functionality. CPU, memory and especially radio usage are kept to a minimum to conserve energy while maintaining an acceptable level of network performance so that applications can be run transparently on top of the secure MAC layer. We use a coordinated yet pseudo-random duty cycle that is loosely synchronized across the entire network via a distributed algorithm. This thwarts an attacker's ability to predict when nodes will be awake and likewise thwarts energy efficient intelligent jamming attacks by reducing their effectiveness and energy-efficiency to that of non-intelligent attacks. Implementing the random duty cycle requires additional energy usage, but also offers an opportunity to reduce asymmetric energy use and eliminate energy use lost to explicit neighbor discovery. We perform testing of RSMAC against non-secure protocols in a novel simulator that we designed to make prototyping new WSN algorithms efficient, informative and consistent. First we perform tests of the existing SMAC protocol to demonstrate the relevance of the novel simulation for estimating energy usage, data transmission rates, MAC timing and other relevant macro characteristics of wireless sensor networks. Second, we use the simulation to perform detailed testing of RSMAC that demonstrates its performance characteristics with different configurations and its effectiveness in confounding intelligent jammers.
Document Availability at the Time of Submission
Release the entire work immediately for access worldwide.
Trammell, David, "Cluster based jamming and countermeasures for wireless sensor network MAC protocols" (2012). LSU Doctoral Dissertations. 950.