Security

Four Tips for Securing a Network DMZ

Follow this advice to lock down high-risk servers.

Mike Chapple is associate teaching professor of IT, analytics and operations at the University of Notre Dame.

A network DMZ likely houses some of the highest-risk servers in an organization: those that provide direct connections to the Internet and are at significant risk of attack. An organization should do everything it can to lock down the DMZ and protect it from threats.

The term “DMZ” comes from the military concept of a demilitarized zone, a neutral area that separates warring parties. Instead of separating armies, a network DMZ is designed to separate the general public — and hackers — from an internal network. In the most common DMZ scenario, a firewall separates the network into three segments: the internal network housing critical resources, the DMZ and the Internet. Any communication between servers in different zones must pass through the firewall and is subject to network security policies.

The typical DMZ houses web servers, e-mail servers, DNS servers and other systems that must have some level of accessibility from the outside world. The DMZ is set up so that an attacker who is able to compromise one of these servers is able to leverage that server to gain access only to other systems in the DMZ, isolating the internal network from the attack. For this reason, it’s critical to design added layers of security control around the DMZ.

Here are four tips to help ensure that a DMZ is secure:

1. Preserve isolation as much as possible.

Keep the rules that allow traffic between the DMZ and an internal network as tight as possible. Too often, administrators seeking to troubleshoot a problem create a rule allowing full access between a DMZ system and a back-end server on the internal network (or the entire internal network). This defeats the purpose of the DMZ and effectively merges it with the internal network. Instead, create specific firewall rules that allow communication only between specific servers on specific ports required to meet business requirements.

2. Practice good vulnerability management.

DMZ servers are exposed to the world, so take extra steps to ensure that they are fully patched to deal with the latest security vulnerabilities. Many security professionals recommend daily, automated vulnerability scans of DMZ systems that provide rapid alerts of newly detected vulnerabilities. In addition, consider patching DMZ systems on a much more frequent basis than protected systems to reduce the window of vulnerability between the time when a patch is released and its application to DMZ servers.

3. Use application layer defenses for exposed services.

Choose a network firewall that has strong application layer protection, rather than just a port filter. A firewall should have the ability to inspect the content of traffic and block malicious requests. One common example of this is screening inbound web requests for signs of embedded SQL injection attacks, preventing them from even reaching the web server.

4. Monitor, monitor, monitor.

The DMZ should be one of the major focuses of an organization’s network monitoring efforts. Use intrusion detection systems, security incident and event management systems, log monitoring and other tools to remain vigilant for signs of an attack.

DMZ systems are at the pointy end of the network security spear and are subject to external attack on a daily basis. For this reason, it’s important to take the time to ensure that they are among the most secure servers in an organization and are rigorously maintained.