IEEE P2668-Compliant Multi-Layer IoT-DDoS Dataset (IEEE P2668-MLIDD)
The IEEE P2668-compliant multi-layer IoT-DDoS dataset (IEEE P2668-MLIDD) is composed of numerous kinds of DDoS attacks generated from multiple leading protocols applied in IoT. The dataset includes the TCP SYN flood, UDP flood, ICMP flood, MQTT Publish flood, HTTP flood, and CoAP flood. The testbed of IEEE P2668-MLIDD complies to the IEEE P2668-defined framework. The volume size of IEEE P2668-MLIDD is around 22GB which includes over 58 million data samples. In addition, IEEE P2668-MLIDD includes 36 malicious IPs and 12 benign IPs. The adoption of IEEE P2668-MLIDD helps researchers or engineers tostudy the patterns of different IoT-DDoS attacks,as well as the potential security challenges in IoT systems. Finally, IEEE P2668-MLIDD facilitates researchers and developers to develop advanced machine learning or deep learning algorithms for the detection and/or defense of IoT-DDoS attacks.
1. Construction of the Dataset:
(1) Main Folder: “IoT-DDoS Dataset”
(2) Testbed: “IoT-DDoS Testbed.xlsx”
(3) Two Sub Folders: “training” and “testing”
(4) In “training” Folder: There are six .pacp files named as “training_XX.pcap”.
(5) In “testing” Folder: There are six .pcap files named as “testing_XX.pcap”.
2. Utilization of the Dataset:
(1) Use the “Wireshark” tool or “Pyshark” tool to open the .pcap files you would like to check;
(2) Follow the official website of “Wireshark” or “Pyshark” to set up the monitoring rules if you would like to do;
(3) If you want to develop supervised learning algorithms, you can open the “IoT-DDoS Dataset” file to set up label for each IP.
3. Range of Utilization of Dataset:
(1) Industry: The dataset of IEEE P2668-MLIDD helps users in industry to develop industrial level IoT-DDoS defense and/or detection solutions. Users should follow the standardization framework of the IEEE P2668 to construct standardized IoT networks which nurtures and proliferates mature, reliable, and secure IoT networks/products.
(2) Academics: The dataset of IEEE P2668-MLIDD may also serve as a testbed to explore the potential cyber vulnerabilities of IoT networks, and facilitates researchers to develop advanced machine learning or deep learning based algorithms, schemes, or models.
(3) Industry and Academics: Users from both industry and academic sectors may supplement new data sets and/or extend the existing dataset of IEEE P2668-MLIDD by following the attributes in the standardization framework of IEEE P2668 to nurture IoT engineering development as well as related researches.