Deakin IoT Traffic Dataset

This dataset comprises network traffic collected from 24 Internet of Things (IoT) devices over a span of 119 days, capturing a total of over 110 million packets. The devices represent 19 distinct types and were monitored in a controlled environment under normal operating conditions, reflecting a variety of functions and behaviors typical of consumer IoT products (<b>pcapIoT</b>). The packet capture (pcap) files preserve complete packet information across all protocol layers, including ARP, TCP, HTTP, and various application-layer protocols. Raw pcap files (<b>pcapFull</b>) are also provided, which contain traffic from 36 non-IoT devices present in the network. To facilitate device-specific analysis, a CSV file is included that maps each IoT device to its unique MAC address. This mapping simplifies the identification and filtering of packets belonging to each device within the pcap files. 3 extra CSV (<b>CSVs</b>) files provide metadate about the states that the devices were in at different times. Additionally, Python scripts (<b>Scripts</b>) are provided to assist in extracting and processing packets. These scripts include functionalities such as packet filtering based on MAC addresses and protocol-specific data extraction, serving as practical examples for data manipulation and analysis techniques. This dataset is valuable for researchers interested in network behavior analysis, anomaly detection, and the development of IoT-specific network policies. It enables the study and differentiation of network behaviors based on device functions and supports behavior-based profiling to identify irregular activities or potential security threats.

本数据集涵盖24台物联网（Internet of Things, IoT）设备在为期119天的周期内采集的网络流量，累计捕获超过1.1亿个数据包。这些设备涵盖19种不同品类，在受控环境下的正常运行状态中完成监测，可精准反映消费级物联网产品典型的多元功能与行为特征（<b>pcapIoT</b>）。
数据包捕获（packet capture, pcap）文件完整保留了所有协议层级的数据包信息，涵盖地址解析协议（Address Resolution Protocol, ARP）、传输控制协议（Transmission Control Protocol, TCP）、超文本传输协议（Hypertext Transfer Protocol, HTTP）及各类应用层协议。同时提供原始pcap文件（<b>pcapFull</b>），其中包含网络中36台非物联网设备产生的全部流量。
为支持设备级专项分析，数据集附带一份CSV文件，用于建立每台物联网设备与其唯一媒体访问控制（Media Access Control, MAC）地址的映射关系，可大幅简化pcap文件中各设备对应数据包的识别与筛选流程。另有3份额外的CSV（<b>CSVs</b>）文件，收录了设备在不同时段的运行状态元数据。
此外，数据集配套提供Python脚本（<b>Scripts</b>）以辅助数据包的提取与处理工作。这些脚本涵盖基于MAC地址的数据包筛选、协议专属数据提取等功能，可作为数据处理与分析技术的实操范例。
本数据集面向关注网络行为分析、异常检测及物联网专属网络策略研发的研究人员，具备极高的科研价值。其支持基于设备功能的网络行为研究与差异化分析，可助力开展行为画像分析，以识别异常活动与潜在安全威胁。