Improving the WAF with Machine Learning

Cloudflare has introduced a new Machine Learning WAF detection system that complements its existing Web Application Firewall (WAF) features. The new tool identifies bypasses and malicious payloads without human intervention, providing better protection against a broader range of old and new attacks. This machine learning-based detection system runs on all traffic, improves detection rates based on past traffic and feedback, and offers a new definition of performance by identifying bypasses and anomalies before they are exploited or identified by human researchers. The secret sauce is a combination of innovative machine learning modeling, a vast training dataset built on the attacks blocked daily, data augmentation techniques, an evaluation and testing framework based on behavioral testing principles, and cutting-edge engineering that allows for negligible latency in assessing each request.