Paul Rubel
BBN Technologies
USA

Michael Ihde
University of Illinois at Urbana-Champaign
USA

Steven Harp
Adventium Labs
USA

Charles Payne
Adventium Labs
USA

Coordinating multiple overlapping defense mechanisms, at differing
levels of abstraction, is fraught with the potential for
misconfiguration. This paper presents the process we used to minimize
these risks and the lessons learned as we developed and validated
security policies for a system that withstood sustained red team
attack. Our process began with application level policy templates,
which were instantiated with system-specific values. These application
policy templates were used to directly generate host-level firewall
rules and policy visualizations. They also provided useful guidance
during the manual construction of process-level security policies for
SELinux and CSA. Following policy generation, the policies were
enforced on the system and subsequently validated with both
application- and network-level testing tools.

Keywords: defense in depth, policy generation, layered security, validation

Read Paper (in PDF)