Monday, March 18, 2024

Enterprise IoT: Experts Discuss Strategies

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Listen to the podcast and read the full transcript below.

The Internet of Things is an emerging technology with a stunningly rapid growth curve. It’s predicted that nearly $6 trillion will be spent on IoT over the next five years. The lion’s share will spent by enterprise customers, who look to IoT to boost productivity, lower business costs, and help expand into additional markets. In recent QuinStreet research, more than half of survey respondents considered IoT to be critical to the growth and success of their organization.

In this video roundtable discussion, four IoT thought leaders will discuss issues like security, analytics, deployment strategies, and the future of the IoT sector.

Download the QuinStreet research: Business and Tech Decision Makers Expect Big Impact from IoT

Nigel Upton, Worldwide Director & General Manager IoT/GCP, Hewlett Packard Enterprise

Jason Shepherd, Director of IoT Strategy and Partnerships, Dell

Mike Matchett, IT analyst, solution architect, Taneja Group

Chris Preimesberger, Editor, Features & Analysis, eWeek

Internet of Things: video transcript

J. Maguire:

Hi, I’m James Maguire, Editor of Datamation, and our topic today is the Internet of Things. We’ll talk about the IoT market overall, as well as development, security and analytics. To talk about that we’ve got four leading experts, including Nigel Upton, worldwide director and General Manager of IoT at Hewlett/Packard Enterprise. Hello to you, Nigel.

N. Upton:

Hi, thank you.

J. Maguire:

And you are there in your home office in Scottsdale today, correct?

N. Upton:

I am. It’s nice and warm here.

J. Maguire:

I bet. 120 degrees, they way, on the way.

N. Upton:

Something to look forward to.

J. Maguire:

Also with us is Jason Shepherd, Director of IoT Strategy and Partnerships at Dell. Hello to you, Jason.

J. Shepherd:

Good afternoon.

J. Maguire:

You are there in sunny Austin, correct.

J. Shepherd:

Sunny and warm Austin.

J. Maguire:

Mike Matchett, IoT Analyst and Solutions Architect at Taneja Group. Hello to you, Mike.

M. Matchett:

How you guys doing?

J. Maguire:

I know today is your birthday, so we have you working on your birthday.

M. Matchett:

I’m looking forward to the cake.

J. Maguire:

There’s cake around here somewhere. I definitely have cake. And Chris Preimesberger, Editor of Features and Analysis at eWeek, and my colleague here are Quinstreet. Hello to you, Chris.

C. Preimesberger:

Hello everybody.

J. Maguire:

I know you are very close to the IoT beat, so that’s good.

C. Preimesberger:

The IoT beat seems to be everywhere.

J. Maguire:

Right. IoT is definitely growing, which brings us to our first question. I want to get the take from the four of you on where the Internet of Things market is right now. Of course, there’s a lot of rosy predictions about where it’s headed in the next several years. I think of IoT as more of an emerging tech and not a really truly established tech. Agree or disagree. So, where is the IoT market? Nigel, what is your sense right now and what do you see getting a lot of traction in the IoT market?

N. Upton:

So, we judge where the market’s going based on how much revenue we drive out of it. Funny, my boss insists on revenue out of this for some strange reason.

J. Maguire:

They’re picky that way, aren’ t they?

N. Upton:

I know, I know. So, we are driving a lot of money today out of connected car, out of smart cities and out of smart energy. But it tends to be quite specific. You know, connected car is such a broad spectrum and if you talk to a telco they claim they own connected car, and if you talk to the head unit manufacturers they claim they own it. So, there’s so much going on it that space and we’re seeing real traction there. But we also see it happening, you know, that’s the vertical… but on the horizontal layer we see a lot of interest in edge computing. We see a lot of interest in saying, look, I’ve got all of this heavy equipment stuff, particularly from the OT guys; The GEs and Siemens and Schneiders and ABBs. They’ve got these heavy machines out there. Tons of sensors, tons of data, and their very reluctant to send all of that data up into the cloud so they’re doing as much pre-processing as they can. They’re getting more and more data so they need more and more horsepower at the edge. So, we’re seeing a real uptick in saying, ‘give me more horsepower at the edge so that I can do more pre-processing, pre-analytics and use the cloud for the exception side rather than try to send everything over expensive lines.’

J. Maguire:

Do you agree that IoT is more emerging tech than established tech at this point? Are we well on our way or almost there? Where are we at?

N. Upton:

I was doing M2M six or seven years ago, so IoT for me is an evolution of where we’ve been before, and connecting devices we’ve done since RFID days. So, it’s continued to evolve, but now it has a sexy new label which means that it’s seen as brand new and every board wants to know what their IoT initiative is. But in reality, the ability to connect devices and gather data from them and take inference from that has been around for a long time. It’s now accelerating to everything.

J. Maguire:

Jason, what’s your sense of where the IoT market is these days? Are we really emerging? Are we mostly there? What’s your take?

J. Shepherd:

Yeah, we’re emerging. I totally agree with Nigel. I mean, IoT is a buzzword to apply to things that have been around for a long time. Now, the things that are different: it’s the mashup of data with backend systems, so actually connecting formally isolated things to the backend. M2M is a little bit different on that front, but from an industrial standpoint it’s connecting these formally isolated systems to backend data to make things better. That’s kind of a new trend. There’s people becoming more comfortable with that. Or even just kind of on-premise systems, just mashing up data, IT and OT types of data. And the other big thing that’s happening is that the cost points are coming down where things that have been done for a long time are now becoming accessible to the masses. So, IoT to use really represents the ability to have these things hit scale and it be changing the lives of people much broader than just the folks that have been investing lots of dollars in these types of technologies previously.

So that’s, I think, the big inflection point. We’re coming out of that hype cycle, for sure. It’s super noisy, of course, right now. Obviously, we’ll talk about that today. But, it’s that inflection point. And definitely this market goes vertical before it goes horizontal. It’s extremely important to go on use cases and really focus on that ROI.

If someone comes to us and says, ‘Hey, I want to buy some IoT,’ it’s not really the right customer. It’s people you have a real problem to solve or have to address, they totally agree with the edge compute side. I know we’re going to talk a little bit about analytics, but we’ve been driving in that since the beginning of our strategy, so really in agreement there.

J. Maguire:

Mike, what is your take about the current IoT market? Are we there, or what’s holding us back, if anything is?

M. Matchett:

We looked at a lot of different factors. We’ve seen big data. We’ve seen faster networks. We’ve seen devices roll out to the edge. We’ve seen $50 Raspberry Pi’s. Anybody can make an IoT thing that is actually internet connected. You’ve got your Amazon order ‘More Tide’ button you can stick on the wall by your washing machine. Google Nest thermometers. All this stuff is really commoditizing, as Jason just mentioned.

You know, but I think there’s still some big concerns we still have on… you talk about edge computing and you talk about data volume, but I don’t think we’ve even seen the start of the data volumes that are going to get created. I think we’re going to see people start to get really concerned about whether they have access up and down their supply chain to these data streams. How do you make that available if you’re in the middle of a supply chain. How do you negotiate that access? Security is going to be a huge problem when your devices aren’t just reporting on everything, but they might actually be programmable, as well. You’ve got a lot of surface area to attack on the Internet of Things when you get down to that level.

So, lots of great opportunity, but also a lot of concerns as we interconnect everything.

J. Maguire:

You know, the data question never fails to amaze me. It’s like everything we do creates more and more data. Especially IoT. And you think about IoT as a one-way street. IoT also has the possibility to be a two-way street, at which point the data stream increases exponentially. Where are we going to store it all? That’s a discussion for a different day.

Chris, what’s your take? The current state of the IoT market, where are we right now?

C. Preimesberger:

Well, that’s a general question. I like that, though. You know what, I like what Jason said a minute ago, too. It just struck me. Somebody going into a company like Dell or HP or anybody else and saying, ‘Hey, I want to buy some IoT. And I want a little of this and a little of that and a little of this. Put it all together so my business can work better.’  I like that idea.

But, it is an evolving thing, and we have had these devices and this broadband and we’ve had this connectivity for a long time. I think that it’s got, like one of you say, there’s a buzzword that’s put on it. The media is as guilty as anybody because we use it all the time. But, once something has a clever label and a memorable and easy to use label, then all of a sudden it seems fresher and it seems newer. It’s really not that new.

However, it is being democratized. Slowly but surely. The larger players have been using these for a long time, and as time goes on and these devices become cheaper and more easy to get and use, and as the software and analytics are able to be used in the cloud and anybody can do a DIY type of thing, do it yourself. As it becomes simpler to use then it becomes more democratized, smaller and smaller companies will be using these things in their businesses. They’ll find use for them. That’s when the market is really going to start taking off. We’re not anywhere near that yet.

But I think a lot of the key foundation blocks are in place for that. Companies like HP and Dell have their platforms now, and others. Developers are learning how to put these things together. We’re going to be seeing a lot more of this as time goes on.

J. Maguire:

You know, when I think about things that are holding us back from IoT adoption, Quinstreet did a research study, and security was one of the key concerns that businesses have. Are we really secure with our IoT deployment? So, Jason, what would you say to a business that’s concerned about the security aspect of the Internet of Things? Is that real? Not so real? Could you reassure them on that regard?

J. Shepherd:

No, it’s definitely real. Security is clearly paramount to these types of deployments. You really want to start out being careful, I mean, what we talked about already, when you start pushing down actuation or control from, certainly the cloud…

J. Maguire:

And what do you mean by actuation?

J. Shepherd:

So, basically actuating a pump or turning something on that spins that people might be nearby or, literally, like being able to control an autonomous vehicle from a remote site. When you don’t have some sort of data diode to prevent that type of control coming in from a hacker it’s a problem. So, that’s why sending data north, like to the backend or to keeping it on-premise, sort of like that starting point, it’s still kind of doing that new thing which is adding less expensive sensors, you know, focusing on really targeted problems, bringing in ERP, CRM, social, but just being careful about how you approach it. So, that’s one kind of way to ask the security question.

The other way to address it is to work with a credible player that works in IT class security, and that’s why companies like us are getting involved. We’re here to go bridge those operations. Companies have been doing this for a long time. With new tools, new infrastructure, and certainly that security manageability to bridge. Everyone is seeing the ven diagram now of IT and OT kind of coming together in IoT. Our whole goal here is to lift up operations focused partners, giving them credible IT class infrastructure to go solve problems in more accessible ways. That’s the goal here.

I’d also say on security it’s all about finding the right size of security so you don’t have to buy the use case. Again, it depends on what you’re doing and the implications if there’s a problem.

J. Maguire:

The right size, did you say?

J. Shepherd:

The right size. Like, you’re not going to throw the kitchen sink at every use case because what happens is if you make it too difficult to implement because it’s so many layers of security, people won’t adopt it.

Now, on the consumer side it’s going the opposite way where they make it so easy to hook up stuff or try to that basically people go return, return, password, password, and then they get on the network and all of a sudden there’s no security. So, that’s where you see a lot of those breeches. You want to find that kind of happy medium between securing it appropriately and then working with companies that have that pedigree of securing these types of solutions and applying the right level.

So, it’s totally warranted. It’s very important. But we actually see data integration as one of the bigger challenges, and actually, use case focus is one of the bigger challenges up front. First, you need to figure out what your goal is then in very lock step you go secure it. But that’s been one of the biggest things, is fragmentation with data, use cases. Too many companies confusing customers is what’s holding us back even more so than security.

J. Maguire:

Mike, what do you think about security? What would you tell companies that are concerned about security in IoT?

M. Matchett:

Well, you obviously have to have a multi-layered approach to thinking about security. Where the devices are. The data in transmission. The data in storage. Where you’re pushing out analytics, and as you mentioned, actuating. I’m going to start using that word.

J. Maguire:

I love that word, don’t you? I’m going to start using that as well.

M. Matchett:

Actuating. I think of it as programmable, but yeah. Control, remote control, right? I’m going to fly the drone and not just I’ll look at the video it’s taking and push the fire button as well. So, I think there’s a lot going on there.

I think it’s definitely a bigger problem than anybody can solve at a small level, so I’m glad to see that larger companies are starting to step into the breach and say, hey, we can help you approach this from a comprehensive perspective.

J. Maguire:

Nigel, I know you think about connectivity in security. Talk to me about connectivity and where that comes from in terms of IoT. What’s important about that?

N. Upton:

In my mind it all depends on who you’re talking to, and it’s similar to security, as well. If you look at it from the consumer side, then connectivity, there’s a certain perception on what they need on that. Then you look the enterprise, and then you look to the OT guys, and it’s all different.

So, just stepping back. The key for us is the data and how to make sure that data goes over secure connections. We have that phrase of, you know, the right connection for the right device. When you talk to a telco then, funnily enough, everything is solvable through a cellular connection. But, I guess if you have a hammer then everything looks like a nail. Companies like SigFox and the Lora lines weren’t around two or three years ago, and the sudden acceleration of being able to use unlicensed spectrum…

J. Maguire:

Explain who those vendors are, just if people don’t know.

N. Upton:

So, this is unlicensed spectrum where they can use connectivity using wavelengths that are longer, and therefore ideal for the low power wide area networks where you have devices that have a limited amount of power that perhaps have a battery in them, but they’re buried inside a bridge or a device or in a wall, but they’re not easily accessible so you can’t change the power supply. So, those devices need to be able to have a network that’s not constantly pinging them saying, ‘Are you awake? Are you awake? Send me data. Send me data.’ They’ll only do that once a day or once a week or whatever it is. So, these are very low power devices that need a connection that is a very low bandwidth on them. They’re normally sending tiny pieces of data, small small chunks, but they’re not doing it as often as, say, a cellular network that would continually ping to see if the device was still attached to the network.

So, SigFox and The Lora Alliance are two examples of companies that have gone and created this new type of network, which is very attractive if you’re looking for a low cost alternative to cellular. They’re right now building out their network city by city. Funny enough, they’re both French, and I’m English. So, a natural affinity.

J. Maguire:

You can get past that, though.

N. Upton:

I’ve gotten past it. I’ve gotten past it. The Lora Alliance and SigFox in the US is actually run by an old colleague of mine who was Ex-Hewlett Packard. So, we’re very close with these guys. We’ve just been demonstrating their stuff at our show last week in Las Vegas, being able to demonstrate how to overt these low power networks. They’re particularly suited for things like smart cities where you do have parking sensors, or whatever it happens to be, but they’re a significant cost advantage. The telcos will tell you hang on a second, narrowband IoT is just about to be here and that will solve it. And that will certainly help telcos. Our view is that wifi has it’s place, as does radio and as does cellular and as does the low power wide area networks. So, cellular for us with that motto of the right connectivity for the right device, is something that really resonates particularly with smart cities and smart energy, but not so much with connected car because they need much more bandwidth. Cellular is ideal for them.

J. Maguire:

What about the question of analytics? IoT is all about analytics. It’s a stream of data going every which way. There’s a disagreement. Do we do the computing at the end point? Do we transfer the data back to headquarters? Mike, what’s your view on that? Do you recommend one way or another for companies? I know it’s very situational, but what do you think?

M. Matchett:

Well, it really depends the algorithm what data you’re trying to get and what insight you’re gathering. I think in general what you’re going to see is kinds of streaming controls where you can push out the processing as far to the edge as we can, in every case, as far out as we can. That will be the gold standard, and it’s not always going to be able to push it out to the edge but it will be pushed out as far as possible in the future.

J. Maguire:

The advantage being what?

M. Matchett:

Just because as you centralize the data you’ve got more to store. You’ve got more to protect. You’ve got expensive infrastructure to hold it. You really want to take advantage of that distributed computing that’s naturally going to emerge in any sort of networked environment. We expect, you know, as the cost of the computing device in this internet of things goes down and the power goes up for it to calculate something. You know, we talked about the power in your phone, and you can imagine the memory density and the computing power that’s going to come out in the next generations of chips. You’re going to want to take advantage of that and push the computing out in a distributed way. This is kind of what we see with scale out computing and big data already, when you scale out into a cluster rather than scale up a machine. So, internet of things takes that even broader and says lets try to push out what we can do, map it out, and then reduce that data as it comes to the center as much as possible.

C. Preimesberger:

Yeah, and James, storage and… building upon what he was talking about, storage and master data management become even more important in the IoT. There’s no question about it. There’s so much chaff in all of the data that has to be eliminated, and the more you eliminate at the end point the better it is for the central processing. This is just a huge challenge or anybody or any company that’s looking at an IoT strategy of any kind. You’ve got to look at the storage, you’ve got to look at your compression, you’ve got to look at your de-duplication.

And it’s not just security. Security is huge, but let’s get the data clean first. I think that is fundamental to the success of all of this because if you don’t it’s very very expensive to clean it up afterwards.

J. Maguire:

Jason, what’s your take? At the end point? At home base? Where is the analytics happening these days?

J. Shepherd:

I mean, really, it’s about having a flexible framework to distribute it. Processing and storage where it makes sense. I totally agree with what we’re saying. We’ve been really pushing on the notion of edge analytics since we began our whole strategy and what we’ve seen is that the same players that were talking about, ‘Cloud cloud cloud,’ like last year, are now coming to us saying, ‘Hey, how about that whole notion of edge analytics?’ because they got the bill. Sending all of the data into the cloud is like a bad episode of Hoarders. Keeping all this stuff. Another way I put it is let me Fedex myself a bunch of data and then I’m going to put it in my closet and just never, you know, a bunch of packages, and put it in my closet and never touch it, because I think the stat is that 99% of data that’s collected and sent to the backend isn’t analyzed. So, obviously there’s some work to be done there just in terms of making sense of that data. We see the edge as being the spam filter for your cloud. It’s all about running complex event processing in-memory stream, in the moment, at the edge, as close to things as possible. To do that kind of real-time housekeeping, you can also be supplying security measures at that time in-memory and you can also be applying the metadata models and stuff to where that wherever that data goes in the future, you got to the point earlier, of just around the data management. You’re kind of putting the housekeeping rules on it at the moment of inception at the edge, you know, doing meaningful analytics there, then meaningful data to the backend. Then there’s this symbiotic relationship between the kind of core analytics in the backend and what you’re doing at the edge. And the backend can be your sort of on-prem data center, and that’s as far as it goes. Or, it could be all the way to the cloud.

It’s not about any single model. It’s about applying the right model to the right use case.

And I also want to say one last thing. It also depends on the connectivity. You wouldn’t necessarily, when you’re running on cellular, for example, you definitely want to be doing local analytics so that you don’t just backhaul everything over cell.

But to Nigel’s point, Lora changes things a little big and in a SigFox kind of scenario, low power wireless. It all comes into play.

J. Maguire:

Nigel, I just heard your name mentioned. People are talking about your opinions. Your view on analytics? I mean, it is an end point? I think, going into the future, analytics is going to be more and more at the end point, you know, there will be less and less need to transfer it back. What’s your take on where analytics happens, ideally?

N. Upton:

There’s not too much more I can add to what Jason was saying. I mean, I absolutely agree with him. But a little earlier, just before we started we were talking about basketball, and the last basketball game, game 5, we launched a new ad. Hewlett Packard Enterprise did. It’s all about the machine, and the machine in the terms of IoT where we’ve created this new technology that is all memory-centric. You know, it’s exactly to Jason’s point. Everything will be done in memory, and particularly where you have real-time systems, it has to be done in memory just for speed. I encourage you, I think they’re going to do the advert again from Hewlett Packard Enterprise on the game this evening, game 6, where you’ll see the machine advert. It’s based on Star Trek, of course, because that would be cool and trendy. The ad is quite clever because basically there’s a new concept of how do you redesign the way the computer is built. Instead of being built around chips you build them around memory. So, have a look at the ad. It’s also fun. There’s a sense of humor in there.

J. Shepherd:

You might see a basketball game in there, too.

J. Maguire:

That’s a minor detail, yeah.

N. Upton:

It’s not the Super Bowl, right?

J. Maguire:

Let’s turn to our crystal ball to sort of wrap up our discussion. If we’re thinking about it, what are we going to be talking about when we talk about IoT, say 3-5 years from now? And what kind of advice would you give to businesses that are either in IoT now or are going to deploy to it? What does the future hold for the internet of things looking out a few years? Mike, what does your crystal ball say?

M. Matchett:

Well, you know, I think we might have already started talking this way. The technologies have existed for some time to connect devices to a network and talk about them. There’s nothing really magically new in that sense. But I think in a couple years or 3-5 years we won’t be calling it the internet of things. It will just be the internet again.

J. Maguire:

It’s like the cloud. The term cloud computing will eventually become obsolete.

M. Matchett:

You know, cloud still have some public/private thing. But I think internet of things just… we’re going to start to assume that much like your phone and your watch and your FitBit and your Nest thermometer and your car, that everything is just really going to be connected. You’re not really going to be thinking about that as a separate kind of paradigm for a data flow architecture. So, I think it normalizes.

But, between here and there there’s a lot of work to do on reinventing how we talk about it, how we document it, how we design it as engineers. One of the things that I think there’s a big gap in is standards. Standards for protocol. Standards for interchange. If you’re even going to take data and pass it up and down a supply chain, it would be great if there were standards among those people participating in that. So, there you go.

J. Maguire:

Chris, what do you think? What do you see in your crystal ball? What are we going to be talking about for IoT a few years ahead?

C. Preimesberger:

Yeah, I agree. the IoT is a new and exciting term now, but it’s for some old stuff and some new stuff. Three years from now we might still be talking about it. Five year from now? I don’t think so. We’re going to be use to this, these activities, these functions, the smartness within devices. We’re going to take them for granted like we take XML for granted or Java for granted right now. Or the internet itself. It will evolve and it will become part of the culture and we won’t even need to mention the name that much. I think it’s just going to be the internet, like Mike just said. So, that’s the way it is. That’s the way the cloud has evolved and that’s the way PCs evolved and watches and tablets and smartphones and all of them.

J. Maguire:

We can use our FitBit to text our refrigerator or our coffee machine and get things going that way.

C. Preimesberger:

Can I insert just a really interesting use of the IoT? There’s a new startup here in Redwood City where I live called InIt. It’s a kitchen, okay? What they do is they connect all of the devices in a kitchen. We used to talk about the toaster. The smart toaster and the smart refrigerator and all of this, but here’s one result. You’ll be able to look at your smartphone, push a button, see what’s in your refrigerator, there’s a camera in there that will tell you what’s in there. Then, the refrigerator will look at it and weigh everything and see the freshness of everything and judge it and then give you a recipe for some of the ingredients that you have in the fridge right now for dinner tonight

J. Maguire:

Wow, okay. That’s impressive.

C. Preimesberger:

Within a couple minutes.

J. Maguire:

The problem is you still have to cook it, but okay.

C. Preimesberger:

How lazy are we getting.

M. Matchett:

I can’t wait until there comes a day that I go shopping for a refrigerator and have to conduct touring test kind of interviews on the different refrigerators to find the smartest one to buy for my house.

C. Preimesberger:

We’re going to need to have an IoT IQ.

J. Maguire:

Jason, going forward in the future, what’s coming up for IoT a few years from now?

J. Shepherd:

I mean, like I said before, this market goes vertical before it goes horizontal. It’s all about use cases and all banding together on some of these lower level standards for that interoperability so where you can really get to more of the “Internet of Things”. It’s really a series of increasingly larger intranets. And so, it’s like we very much subscribe to starting small today, go solve some real problems and then you connect your smart ag solution to your cold chain logistics solution, then to your factory flow, and all of those need to be interoperable, ultimately. That’s where we’re headed. So, it’s very important to work on those core standards to make that happen. But right now, start with small real problems for customers.

The other thing I’ll say is it’s very important to be thinking about where we’re headed with analytics long term. So, we like to say architect for analytics. So, we have a lot of people doing great things with Raspberry Pis and Arduinos and Beaglebone. You name it. It’s awesome stuff. Everyone is like, ‘Well, I want go to put these devices out there on the edge and all I need to do is collect data so I’m good. I’m going to send it to the backend.’ We’re finding already people are like, ‘Well crap, I put that stuff out there but now I’ve got to drive a truck out to that remote site because I actually need to put some intelligence down to that thing.’ So, we do caution people to not rush or put barely enough horsepower at the edge to run the job. If you can afford a little bit more horsepower for that job, you’re going to find new things that you can do with it.

Great example: Remote monitoring of an elevator. You know, we’ve got a customer we’re working with that’s doing this. So, it’s like, hey, I put this here, now all I need to do is pull data to kind of understand what’s going on with that asset so I can do predictive maintenance on that asset. Well guess what? What if you wanted to run, you know, turn the pots over to void, you know, so you can run that as a service. What is you wanted to run digital signage for all of the eyeballs going through that elevator? You might want to run multi-tenant on that gateway to do those things, or whatever computing you have up there. Don’t be too hasty with putting the compute out because you’re going to pay for it later if you don’t think about where you might head with it in that 3-5 year timeframe. It’s all about just kind of thinking ahead from where you can grow.

J. Maguire:

Build a better structure right up front.

J. Shepherd:

Invest in the right infrastructure, without going overboard, but just think about the long term. And also, it’s all about us working together on some of those standards that we can get through binding the intranets together.

J. Maguire:

Nigel, you know what the future holds, so tell us. In the year 2020, Internet of Things, what are you thinking about?

N. Upton:

So, the backend, I think, will consolidate. And the backend will consolidate down to some large vendors that are able to do the heavy lifting to be able to put in that standard space infrastructure that allows you to be able to manage it. There will be crazy innovation on the front end. Once they understand what those standards are then people will innovate like crazy on the devices. So you’ll see a ton of stuff…

J. Maguire:

Do you mean just functionality? Or, what do you mean by ‘crazy innovations’?

N. Upton:

Like the example of the fridge. I mean, for those type of devices we’re seeing all these companies popping up that are building new trackers and all sorts of clever innovations happening on the device side. There will be consolidation, in my view, on the backend and on the stuff where the heavy lifting is. This is from the edge compute platforms, through to how we treat data, though to analytics. I think that will consolidate because it will become more… there’s just more and more data and you need to be able to scale it out.

That movement to standards is absolutely critical. I mean, we chose the 1M2M standard but I’m sure more standards are going to evolve. The idea of having a common data model is critical. The example I’ll use is when talking to a car manufacturer and they told us if anybody comes to us with another killer application we’re going to shoot them. What we want is a killer platform that we can plug anything into. Because they were like most companies. They start with the device, they have an application tied to it, it’s hard wired together, and then you have your next device and your next device. And then you’ve got another application and another application and you just can’t scale it that way. And it’s particularly interesting in connected cars because cars are mobile and they will interact with smart cities, with parking, with energy as EV becomes more and more predictable they’ll have to integrate with the grid. So, connected vehicle is a great thing to think about. You’ve got to be able to cross verticals. So, that ability to have a common data model, a common way of accessing data or using it, and being to access services, it has to go horizontal in the end. But right now it’s all verticalized, and that will, over time, mature and then it will become horizontal.

J. Shepherd:

Ultimately, a vertical solution is the great horizontal scalable platform with vertical domain knowledge applied. Initially, you’ve got to go vertical to get people to adopt and to understand what they can do.

N. Upton:

Exactly.

J. Maguire:

Sounds fantastic. On that note, I’ve got to go program my coffee machine because I can’t do it with my hands. Thank you very much. I appreciate your expertise. I’ll send you the link and we can all tweet about it. Thanks very much.

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles