Skip to main content

Linux security: Intrusion detection and prevention

If an ounce of prevention is worth a pound of cure, an ounce of detection is only worth an ounce.
Image
Torn fence

"Intruder alert!  Strace to PID 45555, log and store."
"Roger that, overseer. Target acquired."
"Status report!"
"Suspicious activity verified, permission to terminate?"
"Permission granted, deploy Sigkill."
"Process terminated at 2146 hours."

Intrusion Prevention Systems, or IPS, are tools designed to detect and stop intrusions in their tracks. They come two basic flavors, network-based and host-based. As you may suspect, a network-based IPS is meant to be deployed to monitor the network and a host-based IPS is deployed on a host with the intention of monitoring just a single host.

How these tools work varies from vendor to vendor, but the basics are the same. The network-based tool monitors traffic on the network and matches it to a long list of known signatures. These signatures describe a variety of attacks ranging from simple corrupt packets to more specific attacks such as SQL injection.

Host-based tools tend to have more capabilities as they have access to the entire host. A host-based IPS can look at network traffic as well as monitor files and logs. One of the more popular tools, OSSEC-HIDS, monitors traffic, logs, file integrity, and even has signatures for common rootkits.

More advanced tools have additional detection capabilities such as statistical anomaly detection or stateful protocol inspection. Both of these capabilities use algorithms to detect intrusions. This allows detection of intrusions that don't yet have signatures created for them.

IDS vs IPS

Unlike it's predecessor, the Intrusion Detection System, or IDS, when an IPS detects an intrusion it moves to block the traffic and prevent it from getting to its target. As you can imagine, ensuring that the system blocks only bad traffic is of utmost importance. Deploying a tool that blocks legitimate traffic is a quick way to frustrate users and get yourself in trouble. An IDS merely detects the traffic and sends an alert. Given the volume of traffic on a typical network or host today, an IDS is of limited use.

The Redhat Enterprise Packages for Enterprise Linux repository includes two great HIDS apps that every admin should look into. HIDS apps are a bit of a hybrid solution in that they're both an IDS and an IPS. That is, they can be configured to simply alert when they see issues or they can often be configured to run scripts to react to scenarios that trigger alerts.

Tripwire

First up from EPEL is Tripwire, a file integrity monitoring tool, which Seth Kenlon wrote about for Enable Sysadmin back in April. Tripwire's job in life is to monitor files on the host and send an alert if they change in ways they're not supposed to. For instance, you can monitor a binary such as /bin/bash and send an alert if the file changes in some way such as permissions or other attributes of the file. Attackers often modify common binaries and add a payload intended to keep their access to the server active. Tripwire can detect that.

fail2ban

The second EPEL package is fail2ban. Fail2ban is more of an IPS style tool in that it monitors and acts when it detects something awry. One common implementation of fail2ban is monitoring the openssh logs. By building a signature that identifies a failed login, fail2ban can detect multiple attempts to login from a single source address and block that source address. Typically, fail2ban does this by adding rules to the host's firewall, but in reality, it can run any script you can come up with. So, for instance, you can write a script to block the IP on the local firewall and then transmit that IP to some central system that will distribute the firewall block to other systems. Just be careful, however, as globally blocking yourself from every system on the network can be rather embarrassing.

OSSEC

OSSEC-HIDS, mentioned previously, is a personal favorite of mine. It's much more of a swiss army knife of tools. It combines tools like tripwire and fail2ban together into a single tool. It can be centrally managed and uses encrypted tunnels to communicate with clients. The community is very active and new signatures are created all the time. Definitely worth checking out.

Snort

Snort is a network-based IDS/IPS (NIDS/NIPS). Where HIDS are installed on servers with the intention of monitoring processes on the server itself, NIDS are deployed to monitor network traffic. Snort was first introduced in 1998 and has more recently been acquired by Cisco. Cisco has continued to support Snort as open source while simultaneously incorporating it into their product line.

Snort, like most NIDS/NIPS, works by inspecting each packet as it passes through the system. Snort can be deployed as an IDS by mirroring traffic to the system, or it can be deployed as an IPS by putting it in-line with traffic. In either case, Snort inspects the traffic and compares it to a set of signatures and heuristic patterns, looking for bad traffic. Snort's signatures are updated on a regular basis and uses libpcap, the same library used by many popular network inspection tools such as tcpdump and wireshark.

Snort can also be configured to capture traffic for later inspection. Be aware, however, that this can eat up disk space pretty rapidly.

Suricata

Suricata is a relatively new IDS/IPS, released in 2009. Suricata is designed to be multi-threaded, making it much faster than competing products. Like Snort, it uses signatures and heuristic detection. In fact, it can use most Snort rules without any changes. It also has it's own ruleset that allows it to use additional features such as file detection and extraction.

Zeek (Formerly Bro)

Bro was first release in 1994, making it one of the oldest IDS applications mentioned here. Originally named in reference to George Orwell's book, 1984, more recent times have seen a re-branding to the arguably less offensive name, Zeek.

Zeek is often used as a network analysis tool but can also be deployed as an IDS. Like Snort, Zeek uses libpcap for packet capture. Once packets enter the system, however, any number of frameworks or scripts can be applied. This allows Zeek to be very flexible. Scripts can be written to be very specific, targeting specific types of traffic or scenarios, or Zeek can be deployed as an IDS, using various signatures to identify and report on suspect traffic.

Wrap-Up

One final word of caution. IPS tools are powerful and extremely useful in automating the prevention of intrusions. But like all tools, they need to be maintained properly. Ensure your signatures are up to date and monitor the reports coming from the system. It's still possible for a skilled attacker to slip by these tools and gain access to your systems. Remember, there is no such thing as a security silver bullet.

[ Free online course: Red Hat Enterprise Linux technical overview. ]

Topics:   Security  
Author’s photo

Jason Frisvold

Jason is a 25+ year veteran of Network and Systems Engineering.  He spent the first 20 years of his career slaying the fabled lag beast and ensuring the passage of the all important bits.  For the past 5 years he has transitioned into the DevOps world, doing the same thing he used to, but now wit More about me

Related Content

OUR BEST CONTENT, DELIVERED TO YOUR INBOX