I recently came across a discussion group thread bemoaning the demise of host-based intrusion prevention systems...
(HIPS). One contributor heard a "current trend in security [is] to move away from host-based intrusion prevention/detection systems" and went on to seek advice regarding the role of host-based and network-based intrusion prevention systems in the enterprise. This kind of talk scares me. Why? Because it may get others thinking that this is a good idea. It's not.
However, before we start reviewing host-based intrusion prevention systems benefits, let's first take a look at host-based intrusion prevention system technology and its proper role in the enterprise. As many readers know, host-based intrusion prevention systems allow the protection of individual systems against attacks. When they detect a potential attack using either signature-based or anomaly-based algorithms, they can ignore the detection, alert administrators to the event or block the traffic altogether. Additionally, host-based intrusion prevention systems are beneficial because:
- Host-based intrusion prevention systems can detect attacks that network-based systems simply can't. Host-based systems have access to detailed data about system processes, resource utilization and device activity that network-based sensors lack. In some cases, analysis of this data may be the only clue that an attack is underway. Consider the case of an insider attack. If the attacker resides on the same local network as the target system, traffic between the two may never come under the scrutiny of a network-based intrusion prevention systems sensor.
- Believe it or not, there are attack vectors other than the network. The movie WarGames may be from 1983, but computers are still attached to telephone lines and the war-dialing technique made famous by Matthew Broderick can be just as dangerous as a network-based incursion. Similarly, an attacker can walk up to the console of your system and begin a brute-force password attack. Network-based intrusion prevention systems are useless in these scenarios.
- Anomaly-based detection techniques are often more effective at the host than on the network. Anyone who's experimented with anomaly-based intrusion detection knows that they can be finicky, at best. These systems rely upon a baseline of "normal" activity and then attempt to detect deviations from that norm. The problem is that, for many networks, there's no such thing as "normal" activity. Individual hosts, on the other hand, may have very clearly defined roles on the network, making network anomaly detection feasible.
And finally, for those who believe host-based intrusion prevention systems are difficult to manage and monitor, that's also largely untrue. Modern host-based intrusion prevention system implementations often use an agent/manager approach that allows centralized enterprise management of deployed systems.
Hopefully, I've convinced you that host-based intrusion prevention systems do indeed play an important role in the enterprise. More importantly, keep this general lesson in mind the next time you hear about a "trend" or "fad" in information security: Defense-in-depth means instituting layers of protection to safeguard your data and systems against many different types of attacks. After all, ivory tower security analysts who declare trends like "the perimeter is dead" often make those pronouncements from behind a strongly defended perimeter!
About the Author:
Mike Chapple, CISA, CISSP is an IT Security Professional with the University of Notre Dame. He previously served as an information security researcher with the National Security Agency and the U.S. Air Force. Mike is a frequent contributor to SearchSecurity, a technical editor for Information Security magazine and the author of several information security titles, including the CISSP Prep Guide and Information Security Illuminated.