An advanced feature engineering approach for sybil attack detection in fog computing environment

Abstract

Sybil attacks gravely impair the integrity and dependability of fog computing environments, especially when operating in IoT networks. These attacks consist of malicious entities creating multiple identities to disrupt authentic operations of a network. Traditional detection mechanisms have been known to report very high false-positive rates along with latency issues. This paper is an introduction to an advanced feature engineering strategy focused on Sybil attack detection in fog computing environments. The proposed strategy, when experimenting with a balanced and engineered dataset, achieved an accuracy of 86%, which is an improvement over the result gotten from the original dataset. The proposed approach uses the Synthetic Minority Over-sampling Technique (SMOTE) to address class imbalance problems in sybil attack simulation datasets. The results illustrate the promise of advanced feature engineering on datasets to further protect fog computing infrastructure from Sybil attacks especially when integrated with Federated Learning