Dataset for Empirical Evaluation of a Layered Edge–Cloud IoT Security Architecture.

This dataset contains empirical measures obtained from a 14-day longitudinal assessment of an ESP32-based biometric and RFID access control system. The study evaluates a stratified edge–cloud framework designed to maintain security and operational continuity during network disruptions and power fluctuations in emerging urban settings. Data Structure for Replication: The primary raw data is provided in the file "14-Day Longitudinal IoT Smart Lock Performance Log Dataset 50", which encompasses the variables necessary to reproduce all study findings: Cycle_ID: Sequential identifier for each of the 50 authentication test cycles. Date & Time: Timestamps for each access event. Auth_Method: Indicates whether Fingerprint (Biometric) or RFID was used. Attempt_Type: Categorized as "Authorized" (registered user) or "Unauthorized" (intruder) to calculate FAR and FRR. Result: A Binary outcome (Success/Fail) used to derive the 98.0% overall accuracy. Latency_ms: The raw verification time in milliseconds for each attempt (used to calculate the mean latency of 0.85s and 1.21s reported in the study). Network_Status: Logs indicating "Online" vs "Offline" states to validate the 100% edge-autonomy resilience. Cloud_Sync: Verification of successful asynchronous log synchronization via MQTT protocols upon reconnection. Replication Transparency: This repository offers the data underlying the means, standard deviations, and metrics presented in the publication. It encompasses the precise numerical frequency counts utilized to create Figure 4 (Authentication Success Distribution) and the temporal intervals employed for Figure 5 (Network Resilience Timeline).