MQTT DoS DDoS IoT Attack

The data collected on a raspberry pi 3B+ as a mosquitto brocker, and as mosquittos clients, from 12 sessions with terminator on an ubuntu 20 (4 publishers and 8 subscribers). The data collection was done with wireshark. The normal traffic was done during 3 days and produced 142.000 data entries. The attack traffic lasted 2 to 3 minutes for hundreds of thousands of data with the most popular and useful tools like hping3 and LOIC. The complete dataset is 424.716 harvested entries. The choice of the dataset fields was made in relation to the parameters most affected by a DDos attack. The dataset can be used for research purposes and on the other hand with machine learning techniques and implementation in IDS or IPS.本数据集采集自一台作为Mosquitto消息代理（Mosquitto broker）与Mosquitto客户端的树莓派3B+（Raspberry Pi 3B+），实验基于Ubuntu 20系统下12个使用Terminator终端的会话完成，其中包含4个数据发布者与8个数据订阅者。数据采集工作采用Wireshark网络分析工具完成。正常流量采集周期为3天，共生成142000条数据条目。攻击流量采集使用hping3、LOIC等主流实用工具开展，持续时长2至3分钟，生成数十万条数据。完整数据集总计包含424716条采集得到的数据条目。数据集字段的选取依据分布式拒绝服务（DDoS）攻击影响最为显著的参数进行确定。本数据集可用于学术研究，同时可结合机器学习技术，应用于入侵检测系统（IDS）或入侵防御系统（IPS）的研发与部署。