CICIoT2023: A Real-Time Dataset and Benchmark for Large-Scale Attacks in IoT Environment

Area of Science:

• Cybersecurity
• Computer Networks
• Data Science

Background:

• The Internet of Things (IoT) is rapidly expanding, integrating into critical sectors like transportation and healthcare.
• Increased IoT adoption presents significant challenges in ensuring efficient and secure operations, including interoperability, security vulnerabilities, and standardization issues.
• Existing IoT attack datasets often lack comprehensiveness, failing to cover a wide range of attacks or utilize extensive network topologies with real devices.

Purpose of the Study:

• To propose a novel and extensive dataset of Internet of Things (IoT) attacks.
• To facilitate the development and improvement of security analytics applications for real-world IoT operations.
• To address the limitations of existing datasets by including a broader spectrum of attacks and a realistic network environment.

Main Methods:

• Execution of 33 distinct cyberattacks against a simulated IoT network.
• The IoT topology comprised 105 devices, simulating a realistic operational environment.
• Attacks were categorized into seven types: DDoS, DoS, Recon, Web-based, brute force, spoofing, and Mirai, with malicious IoT devices targeting other devices.

Main Results:

• A novel and extensive IoT attack dataset has been generated, encompassing 33 attack types.
• The dataset reflects attacks executed within a complex IoT network of 105 devices.
• The dataset includes a wide array of attack vectors, crucial for training robust security systems.

Conclusions:

• The developed IoT attack dataset provides a valuable resource for advancing cybersecurity research and development.
• This dataset enables the creation of more effective security analytics tools capable of detecting and mitigating diverse IoT threats.
• The availability of this dataset on the CIC Dataset website will foster innovation in IoT security.