Adaptive Web Application Firewall for Multi-Threat Detection

Cyber threats targeting web applications have increased in frequency and sophistication, thus the need of such applications to be protected with advanced mechanisms. We introduce in this paper an Adaptive Web Application Firewall WAF that is capable of protecting applications from a wide range of threats that includes but are not limited to SQL injection, DDoS attacks, directory traversal, and CSRF. Using machine learning and traffic analysis, the proposed ABS adapts in real-time to the changing attack vectors and axes of engagement against web applications that are already compromised. With the ever-growing amount of incoming traffic, it is the Adaptive WAF that will enable web applications to maintain various forms of attacks and in the process develop varied defensive approaches. This paper explores how the Adaptive WAF can be used to enhance AWS security, integrity, and availability through the reduction of available cyber risk. The KNN model, with its introduction, strengthens the system further by highly increasing the precision and accuracy of the classification of web attacks to a level of 80.3% in terms of accuracy and 74.8% in terms of precision. This model enables the reliable detection of threats through distinguishing between malicious patterns with fewer false positives, thus further enhancing the ability of the system to detect and respond accordingly to the various web-based threats.