Self-Rectifying Memristor-Based Reservoir Computing for Real-Time Intrusion Detection in Cybersecurity.

Area of Science:

• Materials Science and Engineering
• Cybersecurity
• Computer Science

Background:

• Cybersecurity threats are evolving rapidly due to big data and the Internet of Things (IoT).
• Existing intrusion detection systems (IDS) require advanced solutions for real-time threat identification.
• Memristor-based computing offers potential for efficient hardware implementations.

Purpose of the Study:

• To develop a novel intrusion detection system (IDS) using memristors within a reservoir computing (RC) framework.
• To leverage the unique properties of NiO-doped WO3-/ZnO bilayer self-rectifying memristors (SRMs) for enhanced cybersecurity.
• To address challenges in large-scale integration, such as sneak path currents.

Main Methods:

• Integration of NiO-doped WO3-/ZnO bilayer self-rectifying memristors (SRMs) into a crossbar array architecture.
• Application of a reservoir computing (RC) framework for data processing and classification.
• Utilizing the dynamic properties of SRMs for real-time intrusion detection.

Main Results:

• Achieved a classification accuracy of 93.07% on the CSE-CIC-IDS2018 dataset.
• Demonstrated ultrahigh information-processing efficiency through the proposed memristor-based RC system.
• Successfully addressed sneak path current issues in large-scale memristor integration.

Conclusions:

• The proposed memristor-based RC approach offers a robust and scalable solution for next-generation intrusion detection systems (IDS).
• Emerging electronic devices like memristors can significantly enhance cybersecurity capabilities.
• This work highlights the potential of hardware implementations for advanced cybersecurity applications.