Security at the device level
Cyber security is war. You have to defend your systems from all sorts of outside attackers, and if one that’s skilled and determined gets you in his sights, defending yourself may be tougher than you think. In traditional warfare, as an invading army moves into a new territory, logistics become more difficult the farther the army moves from its home bases and resistance usually stiffens. The opposite is true in most cyber defenses. Once an attacker breaks through a hardened perimeter, moving around inside is usually pretty easy. That’s why defense in depth with incident detection, response, and attribution is so important.
Field devices, meaning individual sensors, transmitters, actuators, motor controllers, and the like, are considered the bottom of industrial networks. However, they can still be the target of cyber attacks if you have a sophisticated attacker. Some recent incidents with water utilities attributed to failed sensors or a bad pump resulted in releases of large amounts of water or even destroyed a pump. Some see these as clear cyber attacks.
If your technicians can configure a field device through the control system, dedicated handheld tools, or by plugging in a laptop to the network, an attacker can do the same thing if he follows the right path. If configuration information is accessible, a device can be changed or simply turned off. If enough strategic devices are manipulated in a production unit, all sorts of bad things could result.
Making this happen requires intimate knowledge of your systems as the hacker has to know two things: which devices to get at, and how to get at them. An insider will know this information, but someone outside your plant probably does not. With lots of time, patience, and maybe some luck, a hacker might be able to follow enough trails and eventually find what he’s looking for, but this is a very tedious process if you have a few thousand I/O tags.
A more practical alternative might be to get some help from an insider. If the hacker can identify one or several individuals in the plant who work in the strategic area, he might try to break into those individuals’ e-mail accounts. If a technician uses his or her personal Gmail account to discuss work-related topics, that is probably the easiest attack vector. The hacker may also follow the technician home and break in via the home wireless network. Reading some e-mails or using e-mail as the path to get into the company network can yield a bounty of information that could help the hacker focus his efforts.
If the PLC or I/O section of the control system attached to the strategic devices is not sufficiently protected, it will be little trouble for the attacker to do whatever he wants with those sensors. Even if there is defense in depth, there have to be approved pathways for the desired information to move up and down. If the hacker can sift through enough data and follow those conduits through the defenses, the defenses won’t be very effective. Some security analysts believe the people who deployed Stuxnet used another malware program to gather information on the targeted systems to help find the best ways to get to the desired devices.
In many respects, the best defense against this sort of attack is maintaining tight internal practices. A common factor in most of these situations is that people are part of the attack vector. Getting company data by breaking into personal e-mail accounts, employees uploading malware via phishing attacks, and sloppy practices with passwords often figure into the analysis after incidents surface. When a hacker gets no help from inside, the job is much tougher.
Matt Luallen is founder of Cybati, a cyber security training and consulting organization, and is a frequent contributor to CFE Media.