IoT Device RAM Traces For Firmware Attestation

Citation Author(s):: Muhammad Naveed Aman (University of Nebraska-Lincoln)

Haroon Basheer (National University of Singapore)

Asif Iqbal (National University of Singapore)

Biplab Sikdar (National University of Singapore)
Submitted by:: Muhammad Naveed Aman
Last updated:: Mon, 07/01/2024 - 16:50
DOI:: 10.21227/0nze-r023
Data Format:: Serialized Pickle File
Research Article Link:: https://doi.org/10.1109/JIOT.2022.3176530

179 views

Categories:

Keywords:

machine learning; malware detection; IoT security; embedded devices; random access memory; dynamic analysis

ACCESS DATASET CITE

Abstract

This dataset is generated for the purpose of developing and testing attestation techniques for IoT devices. The dataset consists of RAM traces for eight different firmwares including traces for running the legitimate firmware as well as tampered versions of the firmwares. we upload the firmware onto the IoT device and allow it to operate for a predefined time period of 300 seconds. Throughout the device's normal operation, we utilize the gateway node to collect numerous RAM trace samples, each comprising 2048 bytes, with randomized intervals between consecutive samples. These samples are labeled as genuine and preserved for subsequent analysis. Once genuine RAM traces are obtained for all eight device firmware, we proceed by uploading each attack variant of a particular firmware to obtain their corresponding tampered RAM traces. Consequently, we obtain three distinct sets of tampered RAM traces (A_1, A_2, A_3) for each firmware variant.