Detecting and forecasting cryptojacking attack trends in Internet of Things and wireless sensor networks devices

Area of Science:

• Cybersecurity
• Network Security
• Machine Learning

Background:

• Cryptojacking poses a significant threat to wireless sensor networks (WSN) and Internet of Things (IoT) devices by exploiting computational resources for unauthorized cryptocurrency mining.
• Existing security measures often struggle to proactively detect and predict the evolving trends of these attacks.

Purpose of the Study:

• To develop and evaluate an innovative approach for forecasting and detecting cryptojacking attack trends in WSN and IoT environments.
• To enhance early detection capabilities and provide predictive insights into emerging cryptojacking patterns.

Main Methods:

• Integration of time series analysis techniques (e.g., ARIMA) with Graph Neural Networks (GNNs).
• Utilized the 'Cryptojacking Attack Timeseries Dataset' for training and validation.
• Employed an ensemble approach combining individual model predictions to improve robustness.

Main Results:

• Individual models achieved high accuracy, with ARIMA reaching 99.98% and GNN reaching 99.99% on specific metrics.
• The ensemble approach demonstrated an overall accuracy of 90.97%, showcasing enhanced predictive robustness and adaptability.
• The proposed method effectively identifies emerging cryptojacking trends under varying network conditions.

Conclusions:

• The integrated time series analysis and GNN approach offers a robust and proactive defense mechanism against cryptojacking in WSN and IoT.
• While ensemble accuracy was slightly lower, it provides superior adaptability for detecting novel attack patterns.
• This research significantly contributes to strengthening cybersecurity against the pervasive threat of cryptojacking.