Unlocking the real value of the IoT
By Anne-Lindsay Beall, SAS Insights Editor
When you hear “Internet of Things,” you naturally think smart homes, connected cars, wearable devices and other consumer applications, all of which will have an incredible impact on daily life. But according to analysts, the Internet of Things’ biggest economic impact is expected to come on the factory floor, on the energy grid and in virtually any large-scale industrial setting.
It’s the Industrial Internet of Things, and the potential is enormous: McKinsey & Co. estimates that more than half the nearly $6 trillion in economic value that the IoT could bring by 2025 is related to the industrial sector. The value is tied to the roughly 50 billion devices that will be deployed by 2020, most of which will be devices such as motors, controllers, pumps and chillers.
“Most of the devices are already there. It’s just a matter of connecting them to the Internet – and that will change everything,” says Shahram Mehraban, a Marketing Director in Intel’s Internet of Things group.
“Connect them and suddenly you’ve got 44 zettabytes of data,” Mehraban says. “But having the data and making sense of it are two different things.” Improved productivity, better efficiency and lower costs only come if companies can figure out how to transmit the data securely and then quickly use the data.
“With the industrial IoT, there needs to be clear ROI to invest in connecting the devices,’’ Mehraban says. “Margins are razor thin in manufacturing, so the investment has to yield a quick payback.”
The perfect use of the IoT
Mehraban says the most obvious cases for IoT in the industrial settings are increasing production yield, and reducing waste. Anything that affects the yield, such as downtime or equipment degradation, is a blow to the bottom line. Data that can alert organizations to problems with yield, preventively alert workers to fix equipment that is about to fail, or otherwise highlight pending issues is the perfect use of the IoT.
“It’s not hard to convince factory managers,” Mehraban says. “They get it.’’ In deploying the industrial IoT, companies will have to make decisions about when to use data in motion versus data at rest. “If you want to look at data across multiple factories to understand trends in yield, then you can use stored data that is analyzed after it is gathered,’’ Mehraban says.
To catch parts before they fail (such as a wonky vibration on a motor), “You need to be able to analyze in real time,’’ Mehraban says. For that, companies need to use a technology like event stream processing. “Real time is so critical.’’
How Intel is using the IoT
One of Mehraban’s roles at Intel is to figure out how to securely and efficiently move data from legacy systems into the IoT world. Intel’s gateways provide a secure bridge from legacy to IoT, and they’ve employed them in their own factories.
Applying analytics to the data that comes off the equipment has helped Intel lower operating costs, increase yields and reduce the test times.
Intel pulls as much as five terabytes of data an hour from the entire factory network and analyzes the information to quickly determine when one of its manufacturing processes starts to deviate from normal tolerances. The analytics helps predict parts that are about to fail, allowing production to run for longer periods because planned maintenance takes less time.
“For example, when one of our testing tools breaks down, it rejects good units,” says Mehraban. “With predictive maintenance analytics, breakdowns are avoided and good units aren’t scrapped. Analytics helps us proactively detect processes that are getting out of control.”
And the industrial IoT extends into the business side. “Supply chains benefit from their ability to dynamically react to before-and-after sales events,” says Mehraban. “Retailers receive product more quickly, get better prices from manufacturers, and better-quality products to sell. Plus, they can adjust supply based on rapidly changing market demands. It’s a win for everyone.”