Abstract
Abstract: Particularly in cloud and IoT systems, this work investigates how the covariance matrix might improve Distributed Denial of Service (DDoS) detection in challenging network situations. Still a constant difficulty, DDoS attacks take advantage of the linked character of contemporary networks. Improving detection accuracy, reducing false positives, and allowing real-time harmful traffic identification is the aim. The suggested approach analyses interactions between several network traffic characteristics using the covariance matrix, therefore spotting abnormalities suggestive of DDoS activity. This method captures subtle, multi-dimensional patterns usually missed by volume-based or signature-based techniques, therefore transcending conventional approaches. In cloud and IoT environments, comprehensive datasets encompassing both valid & hostile traffic assist the models to be trained. Dynamic adaptation to shifting assault pathways and advanced machine learning methods serve to enhance the detection process. Comparative analysis reveals that by basically achieving superior accuracy and resilience, the covariance matrix-based approach overcomes low-rate and zero-day threats. Strong network security is thus guaranteed even in multi-tenant, high-traffic environments since the results demonstrate a clear decrease in false positives and improved detection rates. This work offers a foundation for sophisticated, customizable anomaly detection techniques, therefore offering a possible way to guard critical infrastructure from sophisticated DDoS attacks.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call