Using EM Side-Channels Near a Bluetooth Server Implementation to Monitor Bit-Level Leakages in BLE Communication Channels
V Iyer, Jacob D. Rezac · 2024
Unintentional electromagnetic (EM) emanations from general purpose embedded systems have been comprehensively studied as data recovery and monitoring tools for hardware security. While the potency of these physical side channels has been repeatedly demonstrated on embedded cryptosystems and microprocessors, limited work has been performed on judging their effectiveness in surveilling wireless communication between interconnected devices. This work uses EM side channels near a Bluetooth Low Energy (BLE) server to recover data received from a client. The method uses near-field EM scans to isolate on-chip locations susceptible to data leakage. At these locations, information leakage from a p-bit data value is partitioned into p bit-wise components. These components are individually and independently classified using SVMs trained on a novel side-channel metric. Only O(p2) measurements and p bit-wise classifications are needed to identify which of the 2p possible values has been received. The performance of the method is evaluated using the average recovery rate across bits in several unknown test data packets—a recovery rate of Math 1 was observed when the method was tested on a commercial BLE module.