It’s time to move to real-time regulation

The Internet of Things allows for real-time data monitoring, which is crucial to regulatory reform.

One under-appreciated aspect of the changing relationship between the material world and software is that material goods can and will fail — sometimes with terrible consequences.

What if government regulations were web-based and mandated inclusion of Internet-of-Things technology that could actually stop a material failure, such as a pipeline rupture or automotive failure, while it was in its earliest stages and hadn’t caused harm? Even more dramatically, what if regulations could even prevent failures from happening at all?

With such a system, we could avoid or minimize disasters — from Malaysia Airlines Flight 370’s disappearance to the auto-safety debacles at GM to a possible leak if the Keystone XL pipeline is built — while the companies using this technology could simultaneously benefit in a variety of profitable ways.

Let’s call the concept “real-time regulation.” It will be possible in the near future (and, in many cases, already is) because of the Internet of Things. While the IoT offers many benefits, the one that’s relevant to regulation is that it makes it possible for many people and/or devices to share data in real time about how something is actually operating — that’s crucial to regulatory reform.

In the past, regulation has been limited in its ability to achieve its major goal: avoid future incidents where individuals, the environment, or other public goods suffer because of some sort of action, from a pipeline leak to an airplane crash. That’s because the responsible companies or regulators could only intervene after a catastrophe had happened. Equally important, once a product had left the factory, a manufacturer had no idea how it actually operated in the field — until it broke down or malfunctioned.

Because of the IoT’s ability to share data instantly and its ability to report on operating problems in their earliest stages, it is technically possible for companies and regulators to have real-time access to data about a possible violation while it is in progress. In some cases, such as with pipeline leaks or potential crises on off-shore oil rigs, that would make it possible to actually intervene before there’s a violation, in time to take corrective action.

For example, a sensor on an off-shore oil rig might detect a stress fracture in a section before there was visible evidence. That same sensor could automatically trigger a shutdown: not a drop of oil would be spilled, there would be no environmental impact, and the company would know exactly where to go, speeding repairs and reducing downtime. In the case of flight 370, if the airline had only paid $10 more per flight, it would have received real-time access to the data stream from sensors on the engines. The airline could have realized immediately after the voice transmissions ended that the plane was still in flight, in time to have scrambled interceptor planes, and possibly have avoided the disappearance. Or — my favorite example because it involves such an exemplar of 19th-century industry — after implementing a track-sensor program, Union Pacific has cut its maintenance costs and reduced bearing-related derailments by 75%.

These and many other types of real-time monitoring by companies and regulators alike are possible today, using sensor and communications technologies that are increasingly installed in a wide range of devices. Companies using the devices use real-time data for everything from “predictive maintenance” that keeps repair costs low to the ability to optimize a devices’ operating efficiency. In other words, connected devices create opportunities for companies and the public interest.

Many devices, such as thousands of jet engines, already have Internet-of-Things capacity built in, and declining prices and smaller sizes of sensors and communications equipment will speed their adoption in other trouble-prone products, such as cars and medical devices.

Regulators should begin now to study the feasibility of real-time regulation, and to phase it in as soon as the IoT-equipped products become widespread, requiring companies to have sensors installed and processes in place to monitor results in real time. The result will be a fundamental transformation in regulation: from a system that motivates safety improvements primarily by punishing past infractions to one in which the same technology that helps optimize efficiency and cut operating costs simultaneously protects the public interest by avoiding, or at least minimizing, costly — or deadly — infractions.

tags: , ,

Get the O’Reilly Hardware Newsletter

Get weekly insight and knowledge on how to design, prototype, manufacture, and market great connected devices.

  • SPT

    I don’t see how you draw the conclusion that RT regs would change anything. At best, it would be near real-time, as an event would have to occur somewhere before a regulation rule could be applied. Further, we (already) monitor the things that can/will fail – regulations are generally developed in response to a ‘new’ failure. As to the scenarios provided, GM would still interpret the data, correct? Perhaps that interpretation, either knowingly or unknowingly, wasn’t deemed a life threatening issue. Oil rig – certainly some level of non-uniformity must exist in the surface, as would some variation in the composition of the component. Unless we expect to construct an oil rig of perfect components, you still have risk of failure. ML370 – Again, someone must interpret the data, and it must be paired against expected values from other data – in this case, the flight plan. Remember, ML370 was expected to be out of radio contact over the ocean. By the time contact was expected to be reestablished, what could have been done about the situation? Follow the aircraft until it impacted the sea? That would have saved costs on the recovery effort, but the net result is still the same.

  • Nikunj Mehta

    Operations can and will make decisions by taking into account current and likely future behavior. To the extent that real-time processing can factor additional information into that decision making process, we will get superior results.

    Since there is typically separation between regulator and operator (unlike air traffic control), regulators are bound to come in after the fact. We are certainly seeing a lot more operations centers (and no path to an oil pumping regulator) which are staffed with experts and equipped with technology to be able to synthesize a lot of information to make quick and good decisions. In this regard, I am with SPT.