ACI Generated or Modified Data Sets

Document Type


Contributing USMA Research Unit(s)

Army Cyber Institute, Cyber Research Center, Electrical Engineering and Computer Science

Publication Date

Spring 4-2014


Unlabeled network traffic data is readily available to the security research community, but there is a severe shortage of labeled datasets that allow validation of experimental results. The labeled DARPA datasets of 1998 and 1999, while innovative at the time, are of only marginal utility in today’s threat environment. In this paper we demonstrate that network warfare competitions can be instrumented to generate modern labeled datasets. Our contributions include design parameters for competitions as well as results and analysis from a test implementation of our techniques. Our results indicate that network warfare competitions can be used to generate scientifically valuable labeled datasets and such games can thus be used as engines to produce future datasets on a routine basis.

Type of Data


Data File Format