Enterprise IoT adoption progressing but challenges remain
SAN FRANCISCO–As enterprises of all stripes look to cut costs, automate processes, increase efficiency and agility and otherwise successfully make good use of their data, adoption of internet of things technologies has markedly accelerated in the past year, according to Lakshmi Mandyam, vice president of product management for VMware’s Edge/IoT business.
Speaking to Enterprise IoT Insights this week at VMworld, Mandyam said proofs of concept and limited pilots have given way to scaled deployments. As businesses work to execute on digital transformation strategies, she called out manufacturing as a stand out vertical in terms of IoT, and later explored that in more depth during a session with Pete Tapley, VP of global workplace and mobility sales for DXC Technology. He detailed an IoT solution DXC provided to a manufacturing customer that provides bullets to the military. More on that later.
“I think this is where we’ve been thinking about the fact that you can’t have IoT without IT,” Mandyam said. (She opened her session later noting you can’t spell “IoT” without “IT and “OT”). “In that paradigm, one of the things is how do we simplify and provide that unified layer that allows you to…accelerate to business value.”
Securing a growing attack surface
At it core, the IoT is all about effectively using data to make better use of assets, but to get that data you’ve got to create it. This is where connected sensors come in–like a lot of sensors. In its June Mobility Report, Ericsson researchers project 4.1 billion cellular IoT connections by 2024, up from 1 billion in 2018. And that’s just cellular IoT connections–Bluetooth, Zigebee, Wi-Fi, industrial Ethernet, etc…can all be used to connect the broad array of devices under the IoT umbrella. And more connected endpoints means more potential sources of a cyberattack.
So now you’ve got a rapidly increasing threat surface comprised of devices often installed with their factory settings intact. Mandyam said that means this device now becomes an asset a company’s IT organization needs to securely onboard, monitor and manage throughout its lifecycle, which could be decades, particularly when you consider retrofits to legacy equipment. Again, more on that later.
With these devices, “We need to be able to figure out what their security state is,” Mandyam said. “The manageability of these things is really important.” To that end, VMware announced during the show its Pulse IoT Center 2.0, which lets users bring what was previously SaaS solution on-premises if so desired. And there’s a number of security-focused upgrades Mandyam detailed in a blog post:
- Support for Intel’s Trusted Platform Module which “enables hardware-based root of trust on selected gateways, for remote validation of the devices identity and software stack;”
- And a new app “that enables technicians to scan a QR code and whitelist a gateway during the registration process. This gives enterprises better access control over device enrollment.”
The 2.0 update “really [advances] the security capabilities of this management solution,” Mandyam said.
IT/OT integration
Now, to attempt a segue. IoT security requires a level of trust in the systems from both information technology and operational technology teams within an enterprise. Pulling back the lens, a successful enterprise IoT project requires trust and collaboration between those same IT and OT teams.
Mandyam said IoT, which has been a thing well before the term IoT itself became a thing, has historically been handled by OT and that was fine. But, “What we’re seeing is the IT folks are now being asked to deliver services that include IoT use cases. It’s mainly around how do we extend IT.” People on the OT side know VMware, as do people on the IT side. And here we are back at trust.
The Industrial Internet Consortium continues to publish comprehensive and constructive guides on how to design, deploy, and manage digital change systems in highly individualistic verticals. Among its most recent drops is Managing and Assessing Trustworthiness for IIoT in Practice, which describes the intricate technical balance enterprises must strike between IT and OT systems to have “confidence” (trust?) in conjoined industrial IoT setups. The “office floor” and “shop floor” have to be reconciled, as it were.
The white paper, written by IIC members including im Morrish from Machina Research, Marcellus Buchheit from Wibu Systems, and Frederick Hirsch and Jacques Durand from Fujitsu, argues “confidence” in IIoT hinges on the “trustworthiness” of core IT and OT arrangements, and the “interactions and tradeoffs” between them. Trust seems like an abstract concept, right? Well, for IoT at least, we can establish it based on five dimensions, according to the authors:
- Security of data and equipment;
- Safety of people and assets;
- Reliability of systems and operations;
- Resilience of systems and operations;
- And data privacy.
“VMware is in the perfect position to satisfy both the OT and the IT folks,” Mandyam said.
Physical and digital intersect at the edge
In the telco world, the network edge (definitions may vary) is a red-hot topic. This is because 5G can significantly reduce latencies as compared to LTE, getting down into single-digit milliseconds. This opens up all kinds of new possibilities around real-time data collection and analysis being translated into automated action; an autonomous anything, for example. The sensor takes care of the data collection but the data analysis goes on in the cloud. But not the cloud hosted 100s of miles away in someone’s data center. The powers of 5G notwithstanding, light speed and physics are such that the compute power needs to live closer to the device or application in order to translate the latency improvements into something meaningful. That means you’ve got to “extend IT” out to the edge as Mandyam put it.
When it comes to developing an edge strategy, it’s all about prioritizing what data needs to be processed where. As Mandyam’s colleague Shekar Ayyar,executive vice president and general manager of the Telco and Edge Cloud business, put it, “This ends up being maybe six or seven points between the sort of farthest point to the enterprise to the closest point to the enterprise. The idea would be that on the device, on the premise of the enterprise, you start with compute that needs the lowest latency. Then you start moving farther and farther out.”
Mandyam described a hypothetical manufacturing environment with multiple production lines with a bunch of programmable logic controllers running in some type of larger SCADA system and outfitted with things like vibration and temperature sensors. “All those devices connect to a local aggregation point like a gateway that manages all those sensors and things on the production line. It will then be communicated back into…a local compute node at the edge. In some cases people may want to connect directly from the gateway to their corporate data center. They may want it to go to the public cloud. VMware really has that management infrastructure. This is where our any device, any application in any cloud” strategy comes into play.
Taking it to the factory floor
Two hours after our sit-down, Mandyam took the stage elsewhere in the Moscone Center for a session titled “Unlocking business value with the unified management of OT and IT.” That talk featured Tapley from DXC who you may remember from about 1,100 words ago.
Tapley described an existing engagement with an aerospace and defense industry contractor wherein DXC provides device management for things like laptops, phones, printers and the like. “There was a use case that came up about the factory floor and putting sensors on machines. The company is making bullets…and it’s a lot of bullets.” Per the terms of that unidentified company’s multi-year arrangement with their customer, DXC’s customer needed a “smart factory framework” to meet productivity and pricing requirements while maintaining quality.
He said gathering machine data was previously a paper-based task. “Some of these machines date back to WWII. Some of them are much newer. What we ended up doing was helping them first figure out what’s the strategy to get your data together? What are the sensors we need to put into these machines? How do we collect that data, manage it and enhance it?”
As an aside related to those WWII-era machines, Tapley, in response to a question from an audience member, said he has “yet to get the call from someone saying, ‘We’re going to build a brand new factory. We want it to be a smart factory. Can you help us with that?’ We’re definitely having to retrofit.”
Back to the sensors and data collection and management–Tapley said creating a digital twin provides a data-based picture of the business that increases in fidelity as additional sensors are added to the setup. “The digital twin starts to look more like the physical twin.” Layering in sensor data from the machines with enterprise resource planning data increases that fidelity. “Being able to layer in the analytics that show you now a time sequence–I processed this many bullets in the last minute. Is that good or bad? Well, I don’t know. How does that compare to how we were running yesterday?”
From manufacturing facilities to Class A commercial real estate with complex operating systems, digital twins are moving into the mainstream in an increasingly connected physical world. Research from Gartner indicates 13 percent of companies pursuing IoT projects already use digital twins while 62 percent are somewhere in the process or plan to start.
“The results — especially when compared with past surveys — show that digital twins are slowly entering mainstream use,” said Benoit Lheureux, research vice president at Gartner. “We predicted that by 2022, over two-thirds of companies that have implemented IoT will have deployed at least one digital twin in production. We might actually reach that number within a year.”
Following a flow of assets into data into information into intelligence into action into benefits resulted in DXC’s bullet-producing customer gaining end-to-end visibility into their production cycle; optimizing scheduling, process and task changes; task assignment and completion status insight; automated and manual task management and completion notification; and condition-based automation.
“More and more we’re living in a virtual environment,” Tapley said. “Until we’re in like Ready Player One land or the Matrix, there’s still physical stuff around us. What we’re talking about here is that bridge and how we can interact with the physical parts of life that are around us.”
So what does all that mean?
Grappling with the rapidly increasing importance of cyber-physical interactions that make sectors like manufacturing, mining, logistics and transport operate is abstractly difficult to comprehend. But, given its fundamental importance to global commerce, it has to be comprehended every day by IT and OT professionals. So is it possible to make something so confoundingly complex simple?
“One of the things about IoT is the fact that you have so many heterogeneous use cases, so many platforms that you need to bring bear, and every vertical has its own set of use cases and own set of vertical expertise,” Mandyam said. “This is where building an ecosystem of partners that can go solve this challenge is really fundamentally important. We’re thinking about how are these actually gonna get deployed and managed? Who is gonna bring together all these technologies to make it simple? It feels like we’ve gone from buzzword bingo on IoT…[and] we are actually now, in our conversations with our partners, we’ve gone from being in the PoC friend zone to actually doing a lot of pilots and now we’re seeing that we’re going from doing pilots to actual deployments.”
