IoT Swarm SRAM Dataset for Firmware Attestation

This dataset was created to develop and test firmware attestation techniques for embedded IoT swarms using Static Random Access Memory (SRAM). It contains sequential, synchronous SRAM traces collected from four-node and six-node IoT swarms of devices, each with a 2KB SRAM. Each device is loaded with "normal" or "tampered" firmware to create different network scenarios. Swarm-1 is a four-node network encompassing thirteen scenarios, including two normal network states, two physical twin states, and nine anomalous states. Swarm-2 consists of six nodes, and the corresponding dataset comprises ten scenarios, including four normal states and six anomalous states. Each swarm's data is arranged as a dictionary of scenarios and compiled into a .pkl file. In addition, each SRAM trace is an array of 2048 integer equivalent values (ranging from 0 to 255) of the 2048 hexadecimal bytes stored in the devices' SRAM. The dataset captures normal swarm behavior, physical twin networks, malicious firmware, propagated anomalies, faulty/tampered data, abnormal peripheral control, malicious peripherals, and out-of-sync network states.