AnonFog-PQ Simulation Data and Reproducibility Code for Intermittent IoT Networks

This dataset contains the Python discrete-event simulation outputs and reproducibility code supporting the evaluation of AnonFog-PQ, a fog-coordinated architecture for post-quantum authentication and segmented recovery in intermittent Internet of Things (IoT) networks. AnonFog-PQ combines standard ML-DSA-44 signatures, epoch-rotating pseudonyms and token references, work-conserving fog-side verification scheduling, Merkle-based audit receipts, and segmented blockchain synchronization. The experiments model 500 report instances per scenario and random seed, with verification-window sizes of 8, 16, 32, and 64 reports, 30 paired random seeds, packet-loss rates from 0% to 30%, and both narrowband and Wi-Fi-assisted recovery modes. Three matched configurations are evaluated: PQ-Static+Full, PQ-Static+SBS, and AnonFog-PQ. All configurations use the same ML-DSA-44 profile, four verification workers, workload, and underlying random events, enabling paired comparisons without changing the cryptographic security assumptions. The deposited files include message-level and epoch-level simulation outputs, run-level summaries, aggregated results with 95% confidence intervals, paired statistical comparisons, scheduler-sensitivity results, simulation configuration files, protocol implementation modules, figure-generation inputs, and unit tests. Ledger application bytes are accounted for once per explicit epoch rather than once per report. All numerical results in this record were produced by Python simulation. The dataset does not contain physical-device measurements, real network traffic traces, or hardware energy measurements. The reported completion rates therefore represent modeled task completion rather than experimentally measured device reliability. Similarly, privacy is evaluated from the epoch-pseudonym construction and the stated threat-model boundaries; no empirical unlinkability score is reported. This record supports reproduction of the verification-scaling, synchronization-overhead, recovery-time, completion-rate, and sensitivity results reported in the associated manuscript.