Machine Learning-Powered DNS Firewall
Asma Ahmed A. Mohammed · Nanotechnology Perceptions · 2024
With the quick development in landscape of cybersecurity, the importance of DNS firewall solutions has been recently pronounced. Such solutions work as building blocks in forming inoficial access to various domains, suggesting real-time protection and gretaly unclear communications. The conventional paradigm depends heavily on preprepared lists of known malicious domains, necessitating frequent updates to maintain relevance. However, this method shows inadequate in yet-to-be-cataloged malicious or domains identifying emerging, leading to potential vulnerabilities. Throughout this paper, a creative research endeavor is discussed to shed lights on presenting a cutting-edge DNS firewall solution that proves the power of Machine Learning (ML) techniques. The major purpose is to use the real-time detection of malicious domain requests, thereby critically enhancing cybersecurity protocols. A reasonable assembled dataset, incorporating 34 intricate features and meticulously recorded instances totaling 90,000, was critically chosen from genuine DNS logs. Similarly, it becomes more riched through the careful integration of Open-Source Intelligence (OSINT) sources. The set goal includes the empowerment of precise in addition to rapid classification of domain requests as either malicious or benign.