Automated cyber threat analysis and specified process using vector relational data modeling
Ryan F. Kelly · Calhoun: The Naval Postgraduate School Institutional Archive (Naval Postgraduate School) · 2014
Computer network defense systems should be sufficiently integrated to pull data from any information source, model an expert cyber analyst’s decision process, continuously adapt to an evolving cyber threat environment, and amalgamate with industry standard network hardware. Unfortunately, cyber defense systems are generally stovepipe solutions that do not natively integrate disparate network systems. Correlation engines are generally limited in capability, extensibility, and do not evolve with a dynamic cyber threatscape. Current network defense systems mitigate known vulnerabilities, but effective methods of traffic analysis capable of detecting unknown exploits and identifying advanced persistent threats have yet to be developed. Expert analysts can isolate threats by manually aggregating data sources and distinguishing patterns that indicate a compromise, but there are insufficient skilled analysts available to combat the problem. This work demonstrates a process control configuration that can emulate the investigative process of a human cyber security expert into a pseudo cognitive apparatus capable of accessing several network available data sources, determining a network threat, and terminating a connection, in minutes. The investigative process to detect a PHPMyAdmin attack and issue a response was entirely configured in a vector relational data modeling environment. The configuration could detect and respond to multi-part threat specifications.