Visualizing web server attacks: patterns in PHPIDS logs‡

The prevalence and severity of application-layer vulnerabilities increase dramatically their corresponding attacks. In this paper, we present an extension to PHPIDS, an open source intrusion detection and prevention system for PHP-based web applications, to visualize its security log. Our usage of security data visualization is motivated by the fact that most security defense systems are mainly based on text-based logs for recording security-related events, which are difficult to analyze and correlate. The proposed extension analyzes PHPIDS logs, correlates these logs with the corresponding web server logs, and plots the security-related events. We use a set of tightly coupled visual representations of hypertext transfer protocol server requests containing known and suspicious malicious content, to provide system administrators and security analysts with fine-grained visual-based querying capabilities. We present multiple case studies to demonstrate the ability of our PHPIDS visualization extension to support security analysts with analytic reasoning and decision making in response to ongoing web server attacks. Experimenting the proposed PHPIDS visualization extension on real-world datasets shows promise for providing complementary information for effective situational awareness. Copyright © 2014 John Wiley & Sons, Ltd.