Experimental Evaluation of Network-Controlled Physical-Layer Security Through Friendly Jamming

In this paper, we present a real-life, cost-effective implementation of physical layer security (PLS) using friendly jamming (FJ) and the IEEE 802.11 technology. Our approach is based on a recent development in software-defined networking (SDN) called spectrum programming, where a network controller can execute an intelligent access point (AP) selection algorithm to connect the user station to the AP that provides the most secrecy while exploiting idle APs as jammers. Considering a system with two APs, our first contribution is a theoretical optimization of the power of FJ based on system power parameters, as well as distances between the two APs and the user station and the eavesdropper station. Our second contribution is demonstrating not only that PLS can be implemented with commercial-off-the-shelf Wi-Fi devices, but also that our theoretical network-centric approach allows for a significant increase of secrecy capacity and secrecy coverage by applying FJ. Our experiments show that the theoretical optimization of the transmit power of the jamming AP is valid in practice, effectively maximizing the throughput gap between the user and the eavesdropper. To our best knowledge, this is the first work linking information-theoretic optimization of FJ in PLS to a real-world implementation, which is compatible with Wi-Fi standards and devices.