Bridging the host-network divide: survey, taxonomy, and solution

  • Authors:
  • Glenn A. Fink;Vyas Duggirala;Ricardo Correa;Chris North

  • Affiliations:
  • Virginia Polytechnic Institute and State University;Virginia Polytechnic Institute and State University;University of Pennsylvania;Virginia Polytechnic Institute and State University

  • Venue:
  • LISA '06 Proceedings of the 20th conference on Large Installation System Administration
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a new direction in security awareness tools for system administration-the Host-Network (HoNe) Visualizer. Our requirements for the HoNe Visualizer come from needs system administrators expressed in interviews, from reviewing the literature, and from conducting usability studies with prototypes. We present a tool taxonomy that serves as a framework for our literature review, and we use the taxonomy to show what is missing in the administrator's arsenal. Then we unveil our tool and its supporting infrastructure that we believe will fill the empty niche. We found that most security tools provide either an internal view of a host or an external view of traffic on a network. Our interviewees revealed how they must construct a mental end-to-end view from separate tools that individually give an incomplete view, expending valuable time and mental effort. Because of limitations designed into TCP/IP [RFC-791, RFC-793], no tool can effectively correlate host and network data into an end-to-end view without kernel modifications. Currently, no other visualization exists to support end-to-end analysis. But HoNe's infrastructure overcomes TCP/IP's limitations bridging the network and transport layers in the network stack and making end-to-end correlation possible. The capstone is the HoNe Visualizer that amplifies the users' cognitive power and reduces their mental workload by illustrating the correlated data graphically. Users said HoNe would be particularly good for discovering day-zero exploits. Our usability study revealed that users performed better on intrusion detection tasks using our visualization than with tools they were accustomed to using regardless of their experience level.