Design, deliver, and run enterprise blockchain workloads quickly and easily.
All servers and systems
The Internet of Things (IoT) is a hot topic, and getting hotter with every IoT rollout. However, a more accurate description is the "Network of Things," according to Dana Gardner, principal analyst at Interarbor Solutions, and Dr. Tom Bradicich, vice president and general manager, Servers and IoT Systems, at Hewlett Packard Enterprise.
IoT didn't just happen. There has been a perfect storm of data center improvements—from speed to data storage, sensors shrinking, artificial intelligence, and improved network speeds—to create the right environment for IoT to succeed. As Gardner points out, it's not just what we can do with the IoT now, but that the payoff makes it a great business decision. And it introduces a new way of thinking about our lives.
Dana Gardner and Dr. Tom join Paul Muller, vice president of strategic marketing at HPE, to discuss what the IoT is, explain some practical applications for it, and debunk common myths.
Get an overview of the benefits, challenges, and future of IoT in 10 minutes!
Paul Muller: Internet of what? Hello, and welcome to another episode where we're going to look at the issues, the topics, and, I guess, try to debunk some of the myths and legends behind the Internet of Things, especially in an industrial context. To do that, I'm joined by two fantastic guests, one who you might recognize from previous shows who is, of course, Dana Gardner, principal analyst at Interarbor Solutions. Hello, Dana.
Dana Gardner: Hello, Paul. Good to be with you.
Muller: And a new face to us, Dr. Tom Bradicich, who is the vice president and general manager of Servers and Internet of Things Systems at Hewlett Packard Enterprise. Thanks so much for joining us.
Tom Bradicich: Hello, Paul. Great to be here.
Muller: OK, so I'm going to jump right into it. I opened up with Internet of what, because I guess that's, from my mind, as I talk to our viewers around the planet, a lot of them are struggling with what is the Internet of Things.
Bradicich: Well, let me address that. First of all, as you say, the Internet of Things has achieved celebrity status today. Everybody thinks about a different dimension of it, and there's some ambiguity. However, I'd like to start out by saying I think the word "Internet" is also a little bit of a misnomer. It probably should be "Network of Things" because things like the smart power grid are not going to be hooked up to a public Internet, or an accessible or hackable system like that.
Think of it as Network of Things where devices of any sort, like you said, from a dog food bowl tweeting someone when it's empty, all the way to a Fitbit so you can know how many steps you've taken, more than someone in Brazil, for example, to very, very real airtight business cases in the industrial sector, such as monitoring and sensing the phenomenon of jet engines to see if they're going to fail or when they'll fail, as well as turbines that generate electricity and the smart factory comes to mind. A lot of marketectures under this Internet of Things, such as the smart city, the smart factory, the Internet of Everything, big analog data, for example.
Muller: Let's just take the kiosk example, because I guess a lot of us interact with this sort of technology. Is a kiosk or a smart vending machine an IoT device? What's the difference?
Bradicich: I think it would be because it's network-connected and, it has sensors. These would be two definitions of the Internet of Things—network-connected and has a sensor technology that senses, so if it's going to scan a credit card, if it's going to scan a barcode, or some other type of a scan, even a fingerprint, or if there's physical action with a keyboard type, there needs to be a sensor that takes that, what we call analog phenomenon, converts it to a digital, ones and zeroes bits, so it can be processed.
Muller: I guess the big differentiator or discriminator, then, is if it's doing A to D processes, analog to digital, and it's transmitting it, what, back to a central site, that is considered to be an IoT device?
Bradicich: Well, in some way network-connected and many times, as we were talking earlier, the transmission of the data is not necessarily required and that is this notion of being able to do the computing and generate the insight at the edge of the IoT solution. It doesn’t have to necessarily go back anywhere because sending data back has several issues. First of all, there's a security issue, there's a bandwidth issue, etc., so computing at the edge and accelerating the time to insight from what you like to get from the data is also a very real definition in the Internet of Things.
Gardner: I like the idea of not referring to it as the Internet of Things because it doesn’t have that hop, skip, and jump back to a central server. It's a different architecture. You want to do as much as you can close to the edge, take advantage of new compute and storage technologies, so that the constant flow of the data and information from the sensors is processed, maybe even using machine intelligence, and that can then be put into a meta stage where that intelligence or analysis can be shared. So the sharing needs to be at a higher level, but you still want to get all the bits, all the data.
It's interesting that we're also looking at so many different kinds of devices, we really need some standardization as well. While we need to crunch at the edge, we need to store at the edge. We also need to have some sense of standardization, particularly within vertical industries, and we're just beginning that journey.
Muller: I guess we're starting to see that with dedicated stacks with use cases such as a home automation. We're starting to see APIs appear there. I think about things like, say, Z-Wave as protocol for home automation.
Bradicich: It's very much verticalized, if I can use that term ... and, therefore, we have, for example, a stack to run management and monitoring of the condition of assets in a factory or in a power plant, for example, in the industrial space, or to manage and monitor the facility of a home or a smart building, etc., so definitely it is, to answer you directly, very much verticalized.
By the way, the key in a future solution will be how these vertical stacks across the solution can blend together seamlessly from an optimizing point of view, both cost and performance and, again, back to time to insight, how fast can you get insight from the data you're collecting from the things in the Internet of Things.
Muller: I know, Tom, you've done a lot of research and work on this space in terms of both the opportunities and some of the downsides or considerations. As you think about what may be ... I think a lot of people can imagine the upside. Do you want to talk a little bit about some of the considerations you have on the downside or the risk side that people need to be thinking about, especially as you're talking about lots of data being created and, of course, some of them being potentially sensitive.
Bradicich: There are several risks. First of all, it's extremely complex, so this notion of integration risk and time to value, how fast we can get value is extremely important, so the end-to-end solution is very complicated. It's actually much more complicated than the data center solution. The second risk is security, and as data is personal, or at least it can be corrupted, it can be used improperly, etc., so I would say the time, the integration risk, and the security are two key ones that pop of mind as downsides.
Gardner: I think privacy is also key because we're talking about data about physical organs and temperature and even more intrinsic human dynamics and diagnostics. So not only provisioning to be secure or not to be secure, but provisioning different levels of privacy will be essential, and to protect what's going on on a factory floor. That can be very important information as well.
Muller: I guess to Dana, to your point, how do we need to think architecturally about processing at the edge versus processing at the core?
Bradicich: Let me comment on that. I believe that from the Internet of Things, we will get primarily the dominant majority is analog data from the physical world, such as vibration, temperature, location, video, audio—these analog phenomena in the physical world. Light is another one. Again, that is so big that if we look at, for example, not to get too technical here, but 40 terabytes a second—some of you know how fast that is—that's the biggest of all big data I've ever seen and it's big analog data. I would assert that big analog data from the Internet of Things is bigger than all the other types of big data combined, and we're going to see that as you look around the environment and everything is instrumented and sensor-connected.
Muller: I guess, to your point, Dana, you don't want to be sending that all over the network, do you?
Gardner: No, you can't, just can't, but what's interesting is that the economics are at play here, too. A few years ago, we couldn’t be talking about this type of scale, this type of intensive compute. It simply wouldn’t be affordable. We're actually getting to the point now where you could start to think about measuring every little thing, capturing that data, streaming it in real live time 24/7/365, and you could afford to do that because the payoff is going to be so substantial; the business implications of this are very important. As you say, we're looking at cross-pollinization between industries. What's happening in an airline, what is happening on the factory floor, what's happening in a home aren’t just individual separate silos. The processes which kind of cross the consumer, the business, the supply chain are all going to be starting to be measured and, then, the relationships teased out.
Muller: That's great stuff. We normally end the show with a tip of the week. We might need to end now.
Bradicich: Let me suggest people considering an Internet of Things solution would consider perpetual connectivity, what can happen if you, as a business, are perpetually connected to a product or a person with whom you're doing business and think about those possibilities.
Muller: Fantastic. It's a good tip.
Gardner: The Open Group, an industry consortium, has a workgroup going around Internet of Things standards as verticals, as well as cross-verticals. So go to the Opengroup.org, become a member of that working group if you are interested in getting some more standardization.
Muller: It's a great piece of advice. From our side, we've also got the Big Data Prediction Panel coming up on December 17th, which is hosted online. The details will be on the lower third of the show and in the show notes, of course. With that, I want to thank Tom and Dana for joining us today.
Bradicich: You're welcome.
Muller: I really think we're going to have to come back and dig into this topic some more. There's a lot to discover, and thank you all for watching. For more great content like this, make sure you check out our e-zine, our podcast, and of course, subscribe to our YouTube channel. Thanks again for watching.