Instrumentation of Open-Source Software for Intrusion Detection

Abstract

A significant number of cyber assaults and intrusion attempts are made against open source software written in C, C++, or Java. Detecting all flaws in a large system is still a daunting, unrealistic task. The information assurance area known as ”intrusion detection” (ID) senses unauthorized access attempts by monitoring key pieces of system data. There is a desire to at least detect intrusion attempts in order to stop them while in progress, or repair the damage at a later date. Most ID systems examine system log files, or monitor network traffic. This research presents a new approach to generating records for intrusion detection by means of instrumentation. Open source code such as a web server can be compiled and the execution path of the server can be observed externally in near real-time. This method thus creates a new data source for ID which can be incorporated into a discovery system.