A 2016 BDO Manufacturing RiskFactor Report shows that more than 9 in 10 manufacturers have concerns over cyber security. There is a good reason for that concern considering IBM tells us manufacturing was the second most targeted industry for cyber attacks in 2015.

To prevent an attack from wreaking havoc, we put up firewalls, install virus protection on computers, and limit the privileges users have when accessing certain systems. However, we still have a vulnerable entry point, and that vulnerability is in the sensors at the heart of IIoT.

Sensors access the internet to transfer data, and engineers use the internet to update the software on their sensors. The ability to connect is a tremendous benefit, but that ability is where the risk lies. Every time that connection occurs, there is the potential for a malicious attack.

While we have software to protect our computers, such protection is, in some cases not possible and in others not practical. Sensors have very simple processors and very little memory as they are designed to execute one task and nothing else. If you were to put software protection on the sensor, it could not do its job as all its effort would be focused on protecting itself.

Alenka Zajic, an assistant professor in the School of Electrical and Computer Engineering at the Georgia Institute of Technology said the challenge for manufacturers is, “How do I keep hardware simple enough, so it’s not too costly but also make it secure?”

Zajic is sympathetic to the plight of the manufacturer. She says that while solutions do exist, they are not designed for industrial control systems. Currently, if you want to monitor a network you need to install software. When you do that, you are affecting what the system is supposed to be doing, and the setup has to change. That can be costly and time consuming.

There may still be hope. Zajic is the principal investigator on a project to develop a new technique for wirelessly monitoring IoT devices for malicious software – without affecting the operation of the equipment.

Zajic and her team are attempting to wirelessly monitor electromagnetic emanations from electronics and assign the relationship between software activity and the signals they are observing. Once they have measured enough emanations from the normal activity, they can spot when there is a change.

It is that difference between normal and abnormal that is key. It is impossible to pattern match based on an intrusion because they continue to change, so you have no idea what you are trying to identify. Instead, you profile the intended program and detect when there is a change in the signal.

While Zajic’s team has been working on this technology for about eight years, the program got a boost when they were awarded a $9.4 million grant this past June from the Defense Advanced Research Projects Agency (DARPA).

“We were not expecting to see such a difference between normal and abnormal,” said Zajic. She added that they also didn’t expect they would be able to measure signals as far as 3 meters away from the device. “We thought it would be much weaker, that we’d have to be right on top.”

Aside from the obvious value in protecting the manufacturer from cyber attack, there is an additional side benefit. Technology has a tendency to break down. Because of the ability to recognize an abnormal pattern, you can quickly identify when a program is not doing what it is supposed to be doing and fix it.

The hope is that they can develop one centralized device that will monitor multiple sensors, but currently, they are able to monitor just one and that is still costly.

Costly but still valuable when you consider the damage an attack can rain down on your business. Intellectual property can be stolen, and control of a device can be taken over by an attacker.