Processed Behavioural Execution-Response Dataset for Continuous Device-to-Device Authentication (CD2A)

CD2A Behavioural Execution-Response Dataset for Continuous Device Authentication This dataset contains behavioural execution-response traces collected from 11 Raspberry Pi devices used to evaluate the Continuous Device-to-Device Authentication (CD2A) framework. The data captures workload-dependent timing behaviour influenced by crystal oscillator imperfections and micro-architectural execution dynamics. These measurements are used to model device-specific behavioural fingerprints for identity establishment and continuous authentication. The dataset includes features derived from GPU workload execution, CPU core isolation experiments, and frequency-dependent behavioural measurements. Statistical feature engineering has been applied to capture temporal variability, distributional characteristics, and execution response dynamics. Each record corresponds to a processed behavioural observation linked to an anonymised device identifier and execution core. This dataset supports research in device fingerprinting, continuous authentication, behavioural security modelling, and hardware-induced variability analysis. It enables the evaluation of machine learning-based identity establishment and continuous trust assessment mechanisms under realistic embedded system conditions. The dataset was used in the experimental validation of the CD2A framework presented in: Selvam, M., Khanam, Z., Singh, A. K., Cui, Z., & Rajarajan, M. (2025).Continuous Device-to-Device Authentication Exploiting Crystal Oscillator Impurities.12th IFIP International Conference on New Technologies, Mobility and Security (NTMS).DOI: 10.1109/NTMS65597.2025.11076929 Device identifiers have been anonymised to preserve experimental integrity while maintaining reproducibility of behavioural modelling research.