Enhancing cloud security with intelligent load balancing and malicious request classification

Enhancing cloud security with intelligent load balancing and malicious request classification

연구 분야: Analysis

학회: Cluster Computing

The cloud computing landscape presents a critical intersection of security and performance. To address this, an intelligent load-balancing system is proposed coupled with a malicious request classification approach. This research tackles the growing threat of malicious requests, which pose a significant risk to cloud systems. By integrating a novel classification mechanism within the load-balancing framework, we can pre-emptively identify and mitigate potential security breaches. The approach combines Intelligent Load Balancing with blockchain technology to enhance cloud security and performance. Users, or clients interacting with cloud-based services, access these systems through the Internet. The proposed system leverages the golden eagle optimizer (GEO), a metaheuristic optimization algorithm, to optimize quality of service (QoS) parameters while managing dynamic workloads. To accurately classify malicious requests, we employ a hybrid graph neural network (GNN) and logistic regression (LR) model. The GNN captures complex relationships among request features (e.g., IP addresses, URLs, user-agent strings) to identify patterns indicative of malicious activity. The LR model then makes the final classification decision based on the GNN’s output. Implemented and evaluated using Jupyter Notebook, the system demonstrates an impressive 98% accuracy in classifying malicious requests, highlighting its effectiveness in safeguarding cloud environments.