Industrial IoT Zero-Day Telemetry Dataset

This dataset is derived from industrial IoT telemetry observed in a Zurich smart hospital and its internal logistics infrastructure for experimental assessment. Operational monitoring systems in automated medical equipment, internal transport units, storage facilities, and edge coordination platforms support continuous healthcare and logistical operations.

The released dataset represents a reconstructed telemetry benchmark that preserves the statistical properties, temporal dynamics, and imbalance characteristics observed in real industrial environments, including heterogeneous device behavior, heavy-tailed communication patterns, non-stationary operating modes, and rare security anomalies.

Multivariate time-series records are provided at 8-minute resolution from 01 March 2022 to 30 April 2025, resulting in 208,260 timestamped observations. Features span system and device state, network communication behavior, temporal statistics, manufacturing and logistics operations, cross-entity coordination, and security and policy compliance indicators.

The target variable, ZeroDay_Attack_Label, denotes rare zero-day attack events under severe class imbalance. The dataset contains no personal or patient-identifiable information and is suitable for research on zero-day attack detection in industrial IoT systems.

本数据集源自苏黎世智能医院及其内部物流基础设施中采集的工业物联网（Industrial IoT, IIoT）遥测数据，用于实验评估。自动化医疗设备、内部运输单元、仓储设施及边缘协同平台中的运行监控系统，为持续的医疗与物流运维提供支撑。
本次发布的数据集为经重构的遥测基准数据集，保留了真实工业环境中观测到的统计特性、时序动态性与类别不平衡特征，涵盖异构设备行为、重尾通信模式、非平稳运行模式以及罕见安全异常等特性。
数据集提供2022年3月1日至2025年4月30日期间、采样分辨率为8分钟的多变量时序记录，共计208260条带时间戳的观测样本。特征维度覆盖系统与设备状态、网络通信行为、时序统计量、制造与物流运维、跨实体协同以及安全与政策合规性指标。
目标变量'ZeroDay_Attack_Label'（零日攻击标签）用于标记严重类别不平衡场景下的罕见零日攻击（Zero-day Attack）事件。本数据集未包含任何个人或患者可识别信息，适用于工业物联网系统中的零日攻击检测研究。