Industrial Internet of Things: Sifting Reality From Hype
| By: Kristin Lewotsky, Contributing Editor
Broad industrial connectivity holds potential of enormous value, but only if enterprises do it right.
There was a time that connectivity was location oriented, with stationary computers communicating with one another across Internet. With the advent of smart phones, the Internet tied together people, not places as users accessed the Internet on the move. Now, the next great shift is underway to the Internet of Things (IoT), in which not just people but machines and devices will communicate with one another over the net. “IoT is not a specific component or technology, it’s a conceptual framework,” says Bill Morelli, Director, Internet of Things, M2M & Digital ID at IHS (Austin, Texas). “It’s about enabling connectivity and embedded intelligence in devices.” Much of the buzz about the Internet of Things focuses on the consumer space, but the industrial IoT promises to dwarf the consumer market share within a decade. Indeed, IHS predicts that by the year 2025, the number of industrial IoT devices shipped annually will reach 5.4 billion, for a total installed base of 48.2 billion.
The industrial IoT is being heralded as the next industrial revolution, and not without reason. The first Industrial Revolution replaced human sweat with machine power. The second used mass production to bring quality goods to the masses. The third introduced intelligence in a wide range of devices and machines to dramatically boost functionality and efficiency. With the industrial IoT, those devices are being linked together to communicate with operators, maintenance, management, and other machines. To some extent, these capabilities exist today but typically in a silo environment. One of the core value propositions of the industrial IoT is the broad degree of connectivity. It is that aspect in particular that makes people talk about industrial revolutions.
By harvesting and processing the data gathered to generate insights, the industrial IoT could be used to improve efficiency, minimize downtime, shrink leadtime, reduce inventory levels and allow manufacturers to deliver personalized products, all while reducing cost. Some extravagant predictions have been thrown out for the influence of the industrial IoT Internet of things, with GE predicting that it could boost global GDP by some $15 trillion in the coming decades.1
The industrial IoT today
Currently the industrial IoT suffers from empty dance floor syndrome. Companies want to see use cases before they invest, so many are remaining on the sidelines, but that means that there are limited use cases to review. The mining industry provides one example. In a presentation at the recent Internet of Things World Forum, John McGagh, head of innovation at Rio Tinto, discussed the company’s project to network its processes and equipment to increase efficiency, maximize safety, minimize staffing, and optimize output.2 Part of the project involved automating their fleet of roughly 900 dump trucks with some 92 sensors each divided among the engines, drivetrains, and wheels. The sensors track condition, speed, location, etc., allowing the trucks to operate without human drivers.
In aggregate, the fleet generates about 4.9 Tb/day of data. The information not only controls operation, but also enhances efficiency. “Because they’ve implemented all these sensors on their vehicles, they’re able to do preventative maintenance and ensure that they’re getting maximum life out of their equipment,” says Tom Moore, Analyst, Discrete & Process Automation at IHS (Wellingborough, UK). The position sensors on the trucks can also ensure the vehicles take the shortest possible route to minimize fuel consumption. “There’s just so much potential to the technology, and so many little gains in efficiency that all add up to big benefits,” he adds.
Of course, those big benefits also require big investments. Both IHS analysts estimate that the project cost in the neighborhood of hundreds of millions of dollars. That’s far outside the reach of any but the largest organizations. That said, experience has demonstrated time and again that the price of a technology drops over time once it reaches the point of broad adoption. And make no mistake, there will be broad adoption. In fact, most of the required technology already exists in the form of smart sensors, intelligent components, connectivity protocols, and software expertise. As with many things, however, it must be done properly to deliver benefit. Let’s take a closer look.
It’s not surprising that the industrial IoT is lagging behind the consumer version here at the beginning. Given the money involved, industrial automation in general tends to be risk-averse. Meanwhile, implementing the industrial IoT involves a number of challenges. First among them is developing the conductivity that yields the data sharing that fuels the value proposition of the IoT.
Machine-based networks of intelligent devices have become relatively common on the factory floor. It’s not sufficient to realize the promise of the industrial IoT, however. True collectivity requires the ability to funnel data from the shop floor to the top floor and elsewhere outside the plant.
For starters, it’s a question of bandwidth. Will the factory network support all of those connected devices or will it become the limiting factor? This is important point, especially for companies upgrading existing plants and machines to the IoT. Having spent the money upgrading to devices with enough intelligence to connect to the network, will the asset owners find that the network itself becomes a bottleneck in the enterprise?
“Just because you can connect doesn’t’ mean you should,” says Mike Hannah, commercial program manager for the connected enterprise, Rockwell Automation (Cleveland, Ohio). “You really have to think about the value of connecting that device because every time you put something on your network, you’ve got to make sure your network and support policy can handle it.” To help with this portion of the process, a number of industry consortia have sprung up around the globe to tackle the industrial IoT, with a special focus on connectivity – designing interoperability frameworks, open architectures, and Internet standards.
Is it safe?
As stated, the industrial IoT truly shines when the shop floor next to the top floor to deliver data that can be mined for high level insights. Unfortunately, the results of a survey conducted in 2014 by Rockwell and Industry Week revealed that only about 14% of manufacturing facilities have integrated their ERP systems with their plant operations. The biggest concern? Security.
At the end of the day, many equipment owners remain reluctant to connect their assets to the Internet because of the fear of hackers accessing their data or worse, their equipment. The idea of connecting so comprehensively, even with the benefits it confers, can be offputting. The developer of an application called Shodan, for example, was able to use the Internet to access the equipment interfaces for dams, water treatment facilities, and even crematoria.3
“It’s not that organizations don’t see the value, they just don’t see enough value to offset the risk,” says Hannah. The benefits are worth the risks, he maintains, which can in any case be managed with proper techniques. Part of the solution lies in the approach used to address security issues. “You have to take the discussion of security away from just strictly a network security piece and start thinking of it as a holistic approach, an in depth approach,” he says. “Believe it or not, you can actually become more secure if you have a connected device than if it’s just sitting out there unknown. If you don’t know about a device than you can’t protect it from internal or external threats.”
That said, it’s never possible to eliminate risk entirely. To paraphrase an old saw about hard disk drives, when it comes to getting hacked, it’s not a matter of if it’s a matter of when. Interestingly, the best solution to that problem is also one of the drivers for the IoT to begin with – data and data analysis. “One of the best ways to minimize any damage or destruction from a hack is going to be better sensing and more connectivity and more intelligent monitoring of what’s happening on those networks,” says Morelli. “If your network is compromised, it becomes a question of you finding out about it in hopefully hours or days instead of weeks or months.”
Making it happen
In terms of motion control, the IoT would involve implementations like smart motors and drives that can be monitored from a remote location for preventative maintenance. Of course, as we discussed in last month’s feature, such technology is already available. The difference with the industrial IoT is connecting not just a few things but so many things that a critical mass of data can be mined to deliver insights much deeper than simply preventing downtime on one machine in one location.
Of course, that’s all easier said than done. Greenfield builds will have an easier time of implementing the technology. Companies operating with legacy systems will have a more challenge. Success lies with having a carefully thought-out plan. Organizations need to develop a strategy for evolving to the connected enterprise. Done properly, it can deliver big benefits.
Rockwell, for example, recently implemented a connected enterprise in its own manufacturing organization, which involves 20 plants scattered across the globe. They identified a team and put the process in motion. It wasn't immediate but it provided concrete results.4 "Five years ago we had an inventory of 120 days and only achieved an 82% time-to-want customer experience [essentially, on time delivery],” Hannah says. “When we got through with our journey to our own connected enterprise, we had reduced that inventory to 82 days and increased our time to want to 98%.” They halved the number of rejected parts and saved 30% in capital avoidance. “We realized huge benefits for making this connection ourselves.”
It's reminiscent of the maturing of machine design. There was a time that machine design took place serially, going from the mechanical engineers to the electrical engineers to the controls specialists. At each step, team members added value but also found themselves solving problems introduced by the previous team. Today, the process tends to be more collaborative. Similarly, building a connected enterprise requires collaborative upfront planning.
One route to success is to divide the process into stages: an assessment, the development of connectivity, then building of the infrastructure, followed by designing methods for harvesting the data and converting it into actionable information. “It’s like laying out the blueprint when you’re going to develop a house,” Hannah says. “You’ve got to make sure you have a fundamental infrastructure in place.” Once again heralding back to electronic approach, bringing together key stakeholders like the operations group and the IT team is essential to success. “IT folks and plant folks think differently,” he says. “What they each consider critical is different. The question is how do you get them both together to start developing that shared strategy or objective to realize the benefit?”
To remain effective, connectivity must always be in service of some greater goal. After all, connecting devices, maintaining them, and storing and processing the data takes time and resources. It has to deliver value. Success starts with having a plan. What is the purpose of gathering the data? How do stakeholders plan to use it? From a practical standpoint, who owns it and who is responsible for maintaining it? All of this needs to be hashed out upfront. Organizations need to determine the why before they tackle the how.
On the upside, during that exercise, the magic frequently begins to happen. “When you’re collaborating together is when you start to really see the value of connectivity, you really start to realize the value that you can get out of it,” Hannah adds.
Potential aside, the journey to the industrial IoT is not a process that will take place overnight. We’re coming off of a global recession, with limited investment in industrial capital expenditures. Organizations need to truly understand the value proposition and have it articulated in a way that allows them to make the business case and the use case--after all, the most comprehensively connected system in the world will do no good if the owners don’t know how to leverage it.
Currently, Hannah sees uptake beginning in a few key industries: oil and gas, food and beverage, and pharmaceutical. Once early adopters begin showing value, other companies will jump on the bandwagon. The overall cost of implementation will drop. The biggest advantages go to the manufacturers that lead the charge, however.
There’s still plenty of opportunity to be in that group. “I don’t think we’ve gotten past the point of early adopters yet, especially in an industrial automation," says Moore. "The capital expenditure of some of the implementations is just enormous. It’s not something most companies are particularly interested in unless it has a good use case to justify it, but they’re just not there at the moment.”
As with many changes, the biggest issue does not lie with the hardware and software but with the people ware. At a recent presentation given at an industrial automation show, Morelli reviewed the benefits of the industrial IoT. He did a quick straw poll on attendees as far as whether they considered the biggest challenge to be technology or a matter of corporate culture. “For technology, I don’t think anyone responded,” he says. “When I said culture, virtually the entire audience raised their hands.”
1. P. Evans and M. Annunziata, “Industrial Internet: pushing the bounds of minds and machines,” http://www.ge.com/docs/chapters/Industrial_Internet.pdf, November, 2012.
2. J. McGagh, “Rio Tinto: Mine of the future,” October, 2014.
3. S. Clements, “The Inventor of Shodan Will Help You Hack Our Internet-Enabled, Security-Free Infrastructure,” April, 2013.
4. The Connected Enterprise Maturity Model