Defending Cyber-Physical Systems from Attack Chains

If criminals breach data security and steal credit card numbers or personal information, your company suffers loss of reputation and potentially significant intangible costs. If they breach security of your cyber-physical control systems, you could be facing damage or destruction of physical property and significant tangible costs.

Cyber-physical systems, where computers and the Internet meet the real world, cover a wide range of devices. Industrial automation, home control, smart grids, and medical devices are just a few examples. These machines make decisions and take actions based on inputs from physical readings. Cybersecurity for these systems is an extension of reliability, protecting them from faults or damage introduced by cyberattacks.

These attacks follow a similar attack-chain pattern to non-physical attacks, until the final stages. In the initial reconnaissance, they will research the types of equipment you use that could be compromised and then try to find a weakness in your defenses, whether it is digital, physical, or social. Building a weapon that can get through this weakness comes next, followed by attempted delivery. If delivery is successful, the weapon will exploit the security breach to download and install malware targeting the physical system or device.

Once the malware is installed, the attackers can command and control the compromised device, and this is where the game changes. With access to the device, they can observe your normal operations, query sensors, and run test probes to determine what effect they can have.

Nefarious Objectives

The objectives of a cyber-physical attack are usually not data exfiltration, at least not in the large amounts seen in other attacks. Instead, the attackers could be targeting corporate espionage, denial of control, disablement of alarms, manipulation of sensors or actions to adversely affect output, or physical damage. Overt control could be deferred for a long time while they watch, waiting for the right opportunity to execute or to coordinate with other actions.

A carefully researched and executed series of phishing emails gave attackers access to and control of the production systems in a German steel mill in 2014. Disabling various alarms and safety mechanisms, attackers instigated equipment failures that triggered an emergency shutdown of a blast furnace, causing a massive amount of damage.

In another attack in 2013, snipers shot at and damaged 17 electrical transformers in California, causing them to leak coolant, overheat, and shut down. Just before the attack, they cut the phone and data cables in an attempt to disable the alarms. While there was no cyber component to this attack, it provides an example of the potential of a coordinated cyber and physical attack on vulnerable physical systems.

Defending cyber-physical systems from attack and compromise involves three layers: hardening the devices, securing communications, and monitoring behavior. Older devices can be protected by hardened gateways with a tamper-resistant operating system and strong application execution controls, while new ones should have these functions designed in. Communications between all processes, devices, and systems, should be encrypted in virtual private network tunnels to keep them secure from unauthorized interception or modification. And monitoring of the system and all its components needs to be automated, based on clearly defined policies, to quickly distinguish between normal and suspicious behavior and to catch threats as early as possible.

Sharing intelligence on threats and attacks, with industry partners, government agencies, and security companies is another important step in moving up the attack chain. Given the importance of cyber-physical systems to our lives and communities, it is imperative that we secure them from attacks, and I am confident that we have the resolve and ability to do so.