Just when you think you've mastered the art of intrusion protection, the cyber-criminals discover new some techniques for slipping through your security. Attackers use every possible advantage to stay hidden and gain control. Shouldn't you use every trick to keep them out?

This month's cover story focuses on techniques for keeping intruders off your system. In our lead-off article, "Who's There: Single Packet Port Knocking with fwknop," we study a powerful technique that lets you keep your firewall ports closed to all unauthorized users – but still open to traffic from friends. The next article, "Closing the Book: Fighting Dictionary Attacks with Sshutout and Fail2ban," discusses a pair of tools that will help you keep intruders from guessing your passwords.

Next, we show you how to apply more flexible and precise permissions to files and other objects using Access Control Lists (ACLs). We end with a workshop on the powerful security tool known as SELinux.

Getting In

Read on for more about locking down your Linux system – but first we take a quick look at exactly why this elaborate dance of security is even necessary. The rest of this introduction examines rootkits in the Linux environment.

What if someone broke into your Linux system and replaced the login program with a malignant variant. The new login finds out your username and password and sends the data it collects through a hole it has punched in your firewall to a server somewhere in the wild. No one suspects the new login, although the intruder suspects that an attentive system administrator might eventually wonder about the change in the file size.

But Internet criminals have a way of covering their tracks. Along with the manipulated login tool, the attacker smuggles in a modified ls program. This new version of ls masks any changes to the size or modification date for the login program.

The attacker also decides to replace several other system programs that work together to gather information and hide any evidence of the intrusion. Even anti-virus programs are powerless because the manipulated tools fool them, too.

This scenario is by no means fiction. Attackers often bring along a collection of tools that captures information, opens backdoors, and conceals their activities. This bundle of weapons is commonly known as a rootkit.

A rootkit typically comprises multiple components that fill various needs:

• A trojan drops the rootkit and anchors it on the system.
• A sniffer analyzes the network traffic and retrieves the access credentials.
• In some cases, keyloggers log keyboard input to capture passwords or PINs before the system can encrypt them.
• A backdoor gives the attacker access to the system.

All of these activities are camouflaged by replacing system files and, for current rootkits, redirecting API calls. Other components then put the computer to use – possibly for distributing spam or performing denial of service attacks.

Innovation

In the early days, rootkits simply replaced popular system tools such as ls, passwd, or ps. Security experts soon learned how to detect these early rootkits, and malware programmers soon learned to target the kernel itself.

If an attacker manages to inject malicious code into the kernel, the offending kernel code can capture and redirect any request.

Rootkits running in kernel space are particularly hard to discover. On Linux, kernel rootkits are often injected through a kernel module, which explains why they are known as LKM (for Loadable Kernel Module) rootkits.

Rootkit developers use various approaches to infest the kernel. One option is to manipulate the memory directory via /dev/kmem.

Firmware

Firmware rootkits provide an alternative attack vector. They infect the PC's firmware and survive a reboot. Some rootkits feel quite at home in the ACPI firmware routines. A clean rescue disk is little help against this kind of threat.