Enhancing Cybersecurity with Trust-Based Machine Learning: A Defense against DDoS and Packet Suppression Attacks

Abstract

As technology becomes more intertwined with our daily lives, it is increasingly important to protect our data from attackers. Cyber security has become a top priority for individuals, businesses, and governments, as the threat of cybercrime is constantly evolving and becoming more sophisticated. With the rapid increase in cyberattacks, it has become tricky and cumbersome for cybersecurity experts to react to them all, predict new attacks and analyze the impact of damage being done to business. Traditional security measures such as firewalls, anti-virus software, and intrusion detections are no longer adequate in protecting against new vulnerabilities, especially insider and misbehavior attacks. Recently, Artificial Intelligence based techniques have brought tremendous improvements in cybersecurity with the integration of machine learning (ML) algorithms. ML methods have been built upon large volumes of real-time network data to deploy automated security and threat detection systems. Nonetheless, various cyber-attacks still circumvent traditional security mechanisms deployed to detect those attacks. To address the challenge, in this paper, we propose a machine learning-enabled trust-based routing protocol (TrustML-RP) that identifies the attacking nodes responsible for Distributed Denial of Service (DDoS) and packet suppression attacks. The proposed TrustML-RP scheme first adopts a distributed trust model for establishing trust factor among the participating nodes and later employs an effective combination of ML algorithms e.g., Artificial Neural Network (ANN) and Support Vector Machine (SVM) to find an optimal and secure route and identify attacker nodes. A comprehensive performance evaluation of the proposed scheme is carried out to demonstrate the efficiency on a reasonably sized network containing mixed nodes. The results demonstrate the effectiveness of the proposed scheme in building a trusted network environment and improving network security. The research findings suggest that the integration of a trust-based model and ML techniques can improve traditional cybersecurity methods thereby enabling cybersecurity professionals to design more effective cybersecurity systems.