Toby McClean, Director of IoT Technology and Innovation, ADLINK Technology
We’ve heard all the benefits that Internet of Things (IoT) solutions can deliver for industries. And these benefits—effective use of resources, performance gains, cost avoidance, new business models, etc.—are real. But, we’re still in the early stages of the IoT and there appears to be a belief that it’s difficult to establish. This isn’t true, but let’s look at some of the reasons why there is an expectation of complications.
The factory CEO is demanding: “You better have a digital experiment underway next year or you’re out of here.” The CIO is saying: “I just want a big red button to make this work.” The GM on the factory floor is asking: “How do I pull data off a machine that was never meant to share it and without voiding any warranties?
It can appear a challenging endeavor. But the IoT doesn’t have to be difficult, and edge computing and data connectivity play major roles in any successful IoT architecture.
A problem is industrial players don’t have anything in place for the edge. Many may say they have an edge solution, but they don’t a majority of the time, because it’s a new concept.
The challenge is the IT personnel are accustomed to big servers, the cloud and data centers. The OT team is used to working with closed systems and keeping the manufacturing process humming. Both have viewed edge computing – pushing processing closer to the device creating the data – from vastly different perspectives.
Another obstacle in the IT space is the tendency to pick a single vendor and hope it provides IoT solutions as the technology evolves.
But, when you walk onto a factory floor, there could be assets from 15 different vendors in one production line.
how does the IoT become simple? Three steps: connect, stream and control
When you’re collecting data from them and trying to analyze it all, the problem is very different than IT is accustomed to. You have an integration issue that IT folks have never dealt with before.
IT-designed edge solutions often fail because of this. They are not designed for success at the edge. So, manufacturers are turning to OEMs or software vendors to solve edge connectivity.
So, how does the IoT become simple? Three steps: connect, stream and control.
Connecting assets doesn’t require adding hardware to your machine, if it has a human-machine interface. That interface can be turned into a stream of data using data extraction that can be analyzed. And we don’t even have to touch the asset. We can take the existing data flowing into the database, capture it and turn it into a data stream like it was coming directly off the asset.The data moves where it needs to (control), and the client is not locked into one vendor.This allows the client to run a digital experiment to determine added value without a large investment.
It also addresses the IT or OT divide with a secure approach to bridging data from the OT network to the IT network on up to the cloud, allowing the different managers to control what data flows in whatdirection. Because there is so much data available from so many sources, we can show an easy way to uncover what data is available to them and turn that into information that can be acted upon through analytics. The amount of data and how to use it are major pain points for manufacturers.
Also, consider that the IoT is a fast-moving space, with startups coming and going, and industrial players failing. How do you know who’s going to win at the end? Is it worth locking yourself in with a potential failure?
If you have a video cable, we can turn that into IoT solution. You don’t have to invest in sensors. We can scrap data that appears to have come off the assets directly. How you approach this at the edge is you don’t have to lock into it today. That reduces your risk of making the wrong choice, which is extremely important to decision makers.
We recently worked with a large food processing company and connected one of its production lines in its factory and streamed the data to a cloud. The company went from zero instrumentation to real-timedata about its machines in fewer than 14 days. The company could evaluate that data, whether it be predictive analytics or whatever, to see if there is value because it can now collect it in the cloud or data center. It allows companies to focus on value-added activities with the data.