DDoS attack detection in smart grid network using reconstructive machine learning models | JoVE Visualize

Area of Science:

Cybersecurity
Artificial Intelligence
Smart Grid Technology

Background:

Smart grid networks face significant challenges from network attacks, particularly distributed denial of service (DDoS) attacks.
DDoS attacks disrupt smart grid operations by flooding networks with false data, impacting services for end-users.
Existing machine learning methods for DDoS detection require complete model retraining for new attack classes, causing operational disruptions.

Purpose of the Study:

To propose and evaluate reconstructive deep learning techniques for detecting DDoS attacks in smart grids.
To address the challenge of model retraining for new attack classes without disrupting normal smart grid operations.
To enhance the security, stability, and reliability of smart grid networks against evolving cyber threats.

Main Methods:

Deployment of reconstructive deep learning models (deep and shallow) for DDoS attack detection.
Training separate models to learn representations for individual attack types.
Class-specific reconstruction error-based classification for attack detection.
Rigorous evaluation using two standard DDoS attack databases and comparative analysis against six existing methods.

Main Results:

The proposed reconstructive deep learning technique achieves higher accuracy in DDoS attack detection.
The method effectively eliminates the need for complete model retraining when new attack classes are introduced.
Demonstrated minimum disruption during the introduction of new attack classes, even post-deployment.
Validated through extensive experiments on standard DDoS attack datasets.

Conclusions:

Reconstructive deep learning offers a robust and adaptive solution for securing smart grids against DDoS attacks.
The proposed technique enhances smart grid security while ensuring the stability and reliability of normal operations.
This approach provides a practical and efficient way to manage evolving network attack challenges in critical infrastructure.
The method boosts the overall resilience of smart grid networks against sophisticated cyber threats.