A highly scalable model for network attack identification and path prediction

The rapid growth of the Internet has triggered an explosion in the number of networked applications that leverage its capabilities. Unfortunately, many of them are intentionally designed to burden or destroy the capabilities of their peers and the supporting network infrastructure. Hence, considerable effort has been focused on detecting and predicting the breaches in security propagated by these malicious applications. However, the enormity of the Internet poses a formidable challenge to representing and analyzing such attacks on it using scalable models. Furthermore, the unavailability of complete information on network vulnerabilities makes the task of forecasting the systems that are likely to be exploited by such applications in the future even harder. This paper presents a technique to identify attacks on large networks using a highly scalable model, while filtering for false positives and negatives. It also forecasts the propagation of the security failures proliferated by attacks over time and their likely targets in the future.