Anomaly detection for data streams in large-scale distributed heterogeneous computing environments
Yue Dang, Bin Wang, Ryan Brant +3 more
Counteracting cyber threats to ensure secure cyberspace faces great challenges as cyber-attacks are increasingly stealthy and sophisticated; the protected cyber domains exhibit rapidly growing complexity and scale. It is important to design big data-driven cyber security solutions that effectively and efficiently derive actionable intelligence from available heterogeneous sources of information using principled data analytic methods to defend against cyber threats. In this work, we present a scalable distributed framework to collect and process extreme-scale networking and computing system traffic and status data from multiple sources that collectively represent the system under study, and develop and apply real-time adaptive data analytics for anomaly detection to monitor, understand, maintain, and improve cybersecurity. The data analytics will integrate multiple sophisticated machine…