It's hard to imagine a manufacturing business that doesn't have full-time Internet access, a local network and...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
most likely at least a few sensors or devices connected to that network. Larger plants have hundreds or even thousands of sensors and devices feeding data through the local network. Simply put, that's the Industrial Internet.
According to MIT researcher Ed Schuster:
"The Industrial Internet seeks to improve manufacturing and supply chain efficiency via data, information, mathematical modeling and greater coordination. The components of the Industrial Internet include: a) machine learning, b) big data, c) Internet of Things, and machine-to-machine communication."
There's no doubt that there has been an incredible proliferation of devices, sensors and computer chips in the world at large that are able to communicate through Wi-Fi, cellular networks, Bluetooth, Near Field Communication and broadband cabling -- from cell phones to GPS location services, smart appliances, smart cards and many more.
The same phenomenon is happening in the plant and the supply chain. We have had programmable logic controllers and SCADA devices communicating through plant networks for decades, of course, but more and more "things" have an IP address and are able to send data into the network. Many of these can also accept information and commands from central computers or other devices. This industrial Internet of Things is fundamentally changing the way we manage and track manufacturing. We have more visibility of more items and activities in real time and should be able to manage them more effectively as a result. These devices are also creating a very detailed audit trail, with data that is more timely and accurate than would be possible with manual reporting. And that potentially offers a level of traceability not previously available.
Quality problems can be detected earlier, resulting in faster remediation and fewer bad parts or products. If a defective product does get out into the market, we should be able to isolate the problem to a much smaller population, because of the visibility, and be able to recall all the affected items with far fewer products and customers affected.
Schuster's definition also includes machine learning as a part of the Industrial Internet. This implies that these collections of computing intelligence in plant networks will be capable of some level of analysis and self-modification (learning) that will make them able to adapt to changing conditions and get better at it as time and experience accumulate. This should result in higher quality and better performance -- higher productivity and throughput at lower cost -- with less need for human supervision and intervention. Smart machines and networks will certainly be able to detect developing issues sooner and more precisely, and make fine-grained adjustments to keep quality levels high and the number and severity of defects low.
Developments in sensor technology, data collection and management, machine-to-machine communications, analytics and machine learning, and the ubiquity and capability of the Internet are truly transforming manufacturing. Machines are taking over the repetitive, boring, physically difficult and dangerous tasks in the factory, and technology is helping to manage those machines and improve their performance and flexibility.
As a consequence of this growing level of automation and technology, manufacturing jobs are also changing dramatically. There are fewer requirements for direct production workers and more need for programmers, electronics technicians, computer scientists and other technically adept individuals. That at least partially explains the so-called skills gap currently existing in manufacturing as the industry transitions from the old way of making things to the new, Industrial Internet-enabled factory environment.
Read a definition of the Internet of Things (IoT)
Understand how M2M (machine-to-machine) differs