skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Firewall Architectures for High-Speed Networks: Final Report

Abstract

Firewalls are a key component for securing networks that are vital to government agencies and private industry. They enforce a security policy by inspecting and filtering traffic arriving or departing from a secure network. While performing these critical security operations, firewalls must act transparent to legitimate users, with little or no effect on the perceived network performance (QoS). Packets must be inspected and compared against increasingly complex rule sets and tables, which is a time-consuming process. As a result, current firewall systems can introduce significant delays and are unable to maintain QoS guarantees. Furthermore, firewalls are susceptible to Denial of Service (DoS) attacks that merely overload/saturate the firewall with illegitimate traffic. Current firewall technology only offers a short-term solution that is not scalable; therefore, the \textbf{objective of this DOE project was to develop new firewall optimization techniques and architectures} that meet these important challenges. Firewall optimization concerns decreasing the number of comparisons required per packet, which reduces processing time and delay. This is done by reorganizing policy rules via special sorting techniques that maintain the original policy integrity. This research is important since it applies to current and future firewall systems. Another method for increasing firewall performance is with newmore » firewall designs. The architectures under investigation consist of multiple firewalls that collectively enforce a security policy. Our innovative distributed systems quickly divide traffic across different levels based on perceived threat, allowing traffic to be processed in parallel (beyond current firewall sandwich technology). Traffic deemed safe is transmitted to the secure network, while remaining traffic is forwarded to lower levels for further examination. The result of this divide-and-conquer strategy is lower delays for legitimate traffic, higher throughput, and traffic differentiation (a key component for maintaining QoS). Furthermore, the distributed design is scalable to traffic loads and is less susceptible to DoS attacks. Simulation and analytical results show these new architectures out-perform any current firewall system, providing higher throughput, lower delays, and predictable traffic differentiation.« less

Authors:
Publication Date:
Research Org.:
Wake Forest Univrsity
Sponsoring Org.:
USDOE Office of Science (SC)
OSTI Identifier:
924750
Report Number(s):
DOE/ER/25581-1
TRN: US201006%%824
DOE Contract Number:
FG02-03ER25581
Resource Type:
Technical Report
Country of Publication:
United States
Language:
English
Subject:
99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; DESIGN; OPTIMIZATION; PERFORMANCE; PROCESSING; SECURITY; SIMULATION; SORTING; Firewalls, high-speed, networks, security policy

Citation Formats

Errin W. Fulp. Firewall Architectures for High-Speed Networks: Final Report. United States: N. p., 2007. Web. doi:10.2172/924750.
Errin W. Fulp. Firewall Architectures for High-Speed Networks: Final Report. United States. doi:10.2172/924750.
Errin W. Fulp. Mon . "Firewall Architectures for High-Speed Networks: Final Report". United States. doi:10.2172/924750. https://www.osti.gov/servlets/purl/924750.
@article{osti_924750,
title = {Firewall Architectures for High-Speed Networks: Final Report},
author = {Errin W. Fulp},
abstractNote = {Firewalls are a key component for securing networks that are vital to government agencies and private industry. They enforce a security policy by inspecting and filtering traffic arriving or departing from a secure network. While performing these critical security operations, firewalls must act transparent to legitimate users, with little or no effect on the perceived network performance (QoS). Packets must be inspected and compared against increasingly complex rule sets and tables, which is a time-consuming process. As a result, current firewall systems can introduce significant delays and are unable to maintain QoS guarantees. Furthermore, firewalls are susceptible to Denial of Service (DoS) attacks that merely overload/saturate the firewall with illegitimate traffic. Current firewall technology only offers a short-term solution that is not scalable; therefore, the \textbf{objective of this DOE project was to develop new firewall optimization techniques and architectures} that meet these important challenges. Firewall optimization concerns decreasing the number of comparisons required per packet, which reduces processing time and delay. This is done by reorganizing policy rules via special sorting techniques that maintain the original policy integrity. This research is important since it applies to current and future firewall systems. Another method for increasing firewall performance is with new firewall designs. The architectures under investigation consist of multiple firewalls that collectively enforce a security policy. Our innovative distributed systems quickly divide traffic across different levels based on perceived threat, allowing traffic to be processed in parallel (beyond current firewall sandwich technology). Traffic deemed safe is transmitted to the secure network, while remaining traffic is forwarded to lower levels for further examination. The result of this divide-and-conquer strategy is lower delays for legitimate traffic, higher throughput, and traffic differentiation (a key component for maintaining QoS). Furthermore, the distributed design is scalable to traffic loads and is less susceptible to DoS attacks. Simulation and analytical results show these new architectures out-perform any current firewall system, providing higher throughput, lower delays, and predictable traffic differentiation.},
doi = {10.2172/924750},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon Aug 20 00:00:00 EDT 2007},
month = {Mon Aug 20 00:00:00 EDT 2007}
}

Technical Report:

Save / Share: