Botnets embody the ultimate blended threat. Botnet code carries almost every conceivable form of malware — from spyware to downloaders, rootkits, spam engines and more. To answer like with like, defenders must employ multiple layers of security.
The good news is that time-honored techniques are still surprisingly effective against botnets. Taking the following countermeasures can greatly mitigate the likelihood of a bot infection operating from your network.
1. Patch Promptly
Bots can draw upon a wide variety of exploits in order to infect victims. However, the biggest and most successful bots have relied upon exploiting vulnerabilities that the vendor patched six to eighteen months earlier. In the most extreme cases, we’ve seen bots attempting exploits against vulnerabilities that were patched as long as four years earlier.
We cannot account for why bot communications and back-end systems innovate at a breathtaking pace, while the bot uses exploits that are known and old. Our best guess is that botmasters find exploits by waiting for vendors to patch a vulnerability, then reverse-engineer the patch to find out what the flaw was.
We expect that exploiting more of the recent flaws will be one of the next areas for botmasters to improve upon. For now, though, it is good news for the average network administrator. If you patch promptly when vendors release fixes for software that you run on your network, you can move faster than the botmasters and resist their exploits.
We highly recommend having users rely on Firefox as their primary browser, using the NoScript plug-in to prompt whenever a script tries to execute.
3. Watch Those Ports
This is a two-part recommendation:
1) Even though the latest bots can communicate over ports every administrator must leave open, the vast majority of bots still communicate using IRC (port 6667) and other odd, high-numbered ports (such as 31337 and 54321). All ports above 1024 should be set to block both inbound and outbound unless your organization has a custom application or special need to open a given port.
Even then, you can open a port carefully, implementing policies such as “open only during business hours” or “deny all, except traffic from the following list of trusted IP addresses.” This simple measure prevents the garden variety and slow-adopter bots from reaching their Command and Control Center (C&C) for instructions and updates, essentially killing such bots on arrival.
2) Botnet traffic that travels over needed ports such as 80 or 7 often gives itself away by generating traffic when there should be none. Commonly, botmasters update their zombies between 1:00 a.m. and 5:00 a.m., when they assume no one is watching. Make a habit of checking your server logs in the morning. If you see Web browsing activity when no one was there to do the browsing, that’s your cue to investigate.
In addition to deep packet inspection and stateful packet inspection, administrators should look to their firewalls to provide proxy technology for additional botnet protection. Proxies stop nonstandard traffic attempting to run on standard ports. For example, the spamming botnet Mega-D runs nonstandard, homebrew traffic over HTTP port 80. The Firebox’s HTTP Proxy would spot and block such traffic instantly, by default.
4. Redouble User-Awareness Training
Some bots perform mass scans of the Internet, find vulnerable machines, and infect them. This practice is actually diminishing. More often now, bots “social engineer” their way onto your network by enticing victims to click a link or open a file. These bots have the same restrictions as certain legendary vampires: They can’t cross your threshold unless you invite them in.
This “luring” approach has gradually constricted itself even further. Attackers used to send malicious executable code as an attachment to an email. This practice has also fallen into decline. Most of the action now is Web-based. Malicious emails that would have contained an attachment two years ago now contain a link to a malicious site instead. Innocuous Web sites can be infected by Mpack or other surreptitious malware to infect visitors who arrived by users clicking recklessly.
That’s your cue to explain to users, in terms they can understand, why they should never invite the vampire in. Tell them not to open attachments that arrive unsolicited and unexpected; why they shouldn’t click links in email; and why they must think twice about any unusual links they click.
5. Stay Vigilant
This recommendation seems too obvious to mention, almost like “Try not to get infected!” Yet IT administrators who spend so much time putting out fires and maintaining an understaffed help desk, often never look at their system logs. They never monitor bandwidth usage. They cannot tell you who is connecting to what from their network. They have devices connected to their network that they don’t even know about.
If this describes you, all we can say is, you are begging for trouble. You might even have bots on your network as you read this. If you are an administrator who rarely checks your logs, you must start reading them. Today. Once you learn what “normal” looks like on your network, 30 minutes a day is all you need for a spot check.
Don’t Be a Victim
We believe the recent unprecedented bot breakthroughs merely foreshadow innovations to come. As never before in our years in Internet security, each month seems to bring newly discovered exploits that researchers cannot fully explain.
It turns out that botnets have been blended threats, but they have not been ultimate blended threats. Botmasters now freely supplement the traditional botnet architecture with added components that enhance automation, administration and evasion. These combinations of technologies are frighteningly sophisticated and surprisingly polished. Any observer can safely predict that this trend will not only continue, but grow.
A simple way to undermine a botmaster’s power is to make it difficult for bot code to recruit victims. The key to winning this war is using security appliances that utilize numerous layers of security, which are intelligently applied across many protocols, with powerful proxy technology that scrubs both inbound and outbound traffic to keep business networks safe.
Joe Wang is CEO of WatchGuard, a provider of business security solutions.