Dissertation/Thesis Abstract

A quantitative experimental study of the effectiveness of systems to identify network attackers
by Handorf, C. Russell, Ph.D., Capella University, 2016, 102; 10252986
Abstract (Summary)

This study analyzed the meta-data collected from a honeypot that was run by the Federal Bureau of Investigation for a period of 5 years. This analysis compared the use of existing industry methods and tools, such as Intrusion Detection System alerts, network traffic flow and system log traffic, within the Open Source Security Information Manager (OSSIM) against techniques that were used to prioritize the detailed analysis of the data which would aid in the faster identification of attackers. It was found that by adding the results from computing a Hilbert Curve, Popularity Analysis, Cadence Analysis and Modus Operandi Analysis did not introduce significant or detrimental latency for the identification of attacker traffic. Furthermore, when coupled with the traditional tools within OSSIM, the identification of attacker traffic was greatly enhanced. Future research should consider additional statistical models that can be used to guide the strategic use of more intense analysis that is conducted by deep packet inspection software and broader intelligence models from reviewing attacks against multiple organizations. Additionally, other improvements in detection strategies are possible by these mechanisms when being able to review full data collection.

Indexing (document details)
Advisor: Livingood, Richard, Brown, Steven
School: Capella University
Department: School of Business and Technology
School Location: United States -- Minnesota
Source: DAI-B 78/08(E), Dissertation Abstracts International
Subjects: Information Technology, Computer science
Keywords: Attacker attribution, Information security, Intrusion detection
Publication Number: 10252986
ISBN: 9781369484632