Tom Rebbeck, research director at Analysys Mason, talks to Mike Troiano, Vice President of IoT Solutions at AT&T, to discuss how AT&T is supporting the Internet of Things (IoT) from providing connectivity through to end-to-end solutions.
Tom Rebbeck: Can you outline AT&T’s approach to IoT?
Mike Troiano: AT&T provides solutions and a ‘menu’ of services to developers and customers across all elements of the stack. If a customer just needs connectivity, we can play that role. If they need help with application development, how to secure the data or deal with analytics, we can play that role.
If they are looking for an end-to-end solution including hardware, software and professional services, we can play that role as well.
TR: And specifically on connectivity, how do you see the market developing and what is your strategy? I’m thinking about the different connectivity types, like global contracts, satellite and low power.
MT: AT&T has been in the M2M (machine-to-machine communications) or IoT business for well over a dozen years, predominantly in the United States. We’re seeing more and more conversations about how to build and deploy international solutions.
Maersk Line is a good example: We are providing the hardware, software, and professional services for them to monitor their refrigerated containers in over 140 different locations.
TR: Why do you think that it has become more international?
MT: A few years ago, most of the conversations were with a mid-level person in an IT organisation who was tinkering with wireless technology. More recently, the conversations are not with an IT person but with the chief marketing or chief strategy officer suite. Some deals we’re working on involve board-level discussions. When the board and the C-suite are making decisions on global opportunities, they don’t want to deal with 32 different companies. They want one agreement for all geographies.
TR: What about connectivity options beyond cellular?
MT: We need solutions that are communication ‘agnostic’. Customers need to connect devices with lots of networks; cellular, Wi-Fi, Low Power, Ethernet and others. They are saying to us: “I need one management platform. I need to be able to provision and procure the communication for each one of those paths.”
That’s the direction that we are heading in with a ‘multi-network strategy’. Our customers are becoming more sophisticated with their needs, and they’re asking us to scale our systems and platforms accordingly.
A good example is what we launched with our satellite offer. We lead with cellular connectivity as a mainstay, but where there is no cellular, or when the client needs an augmented technology, we’ll provide satellite connectivity.
TR: Who do you see as using the big customers for satellite?
MT: Shipping is the biggest use case right now, but it could be anybody in the asset management space. We’re also working on some interesting deals for oil companies that want to monitor pipelines.
TR: It would be interesting to hear what you’re doing in the low power space. I understand AT&T’s quite heavily involved in LTE-M.
MT: We are a standards-based organisation, and we are extraordinarily excited about the advancements around LTE, in particular, LTE-M and NB-IoT.
If you believe that there will be 25 to 50 billion devices over the next number of years, there’ll be more likelihood of interference. Depending on the application, you want to think about using dedicated spectrum.
Customers are also concerned about the longer-term prospects of non-standards based technology. A customer I met last week told me, “I have a warehouse of proprietary devices and stacks that we’ve played with over the last 10 years.” They’re in a warehouse because nobody uses them. They didn’t gain traction.
TR: LTE-M is one low power technology and NB-IoT another. There seems to be a different role for NB-IoT versus LTE-M. Is it something you’re looking at as well?
MT: Yes, so we’re looking at both. For NB-IoT, you’re talking about even lower module costs, less mobility support, and lower throughput.
I think the important point here is that you have to think long and hard about the use case. Let’s just assume, for the sake of argument, that you install a million widgets. What if the application profile changes dramatically in year two and throughput becomes an issue? Now you’re stuck with this little box that you can’t get out of. Again, there’s a lot of buzz in the industry about the potential, but you need to think carefully about the application and its roadmap.
TR: Beyond connectivity, what else is AT&T offering to help developers build IoT solutions?
MT: We have M2X and Flow. M2X is a time data storage environment we launched in 2015. Flow Designer is an application development platform – it has tools for developers to build and deploy IoT applications. We launched Flow Designer in January 2016, at our developers’ conferences in Las Vegas.
We’ve got about 14,000 active developers using M2X, Flow and the data services today. At our ‘Shape’ Conference in mid-July in California we’re expecting over 1,000 attendees and we’ll be doing a hackathon with about 500 people.
Often, we’ll test ideas with those developers and that ecosystem first, and then decide if we want to productise it. Both M2X and Flow Designer went through that path.
We saw the developers in the hackathons using our tools, and it gave us fantastic feedback. It’s been a great, great avenue for AT&T.
TR: 14,000 is a lot of developers. Who are these developers? Can you give a sense of who they are and where they’re coming from?
MT: They are a mix. What we have seen over the last two years is a pretty balanced mix of consumer and business-centric developers.
Key sponsors for our Shape event include IBM, Cisco, Qualcomm, GE, Honda and Jaguar. Large enterprises and ecosystems are supporting our initiatives because they realise that this has become a good platform for them to tap into developers and, in some cases, get those developer ecosystems to work on some of their problems.
TR: Many organisations are competing for developer attention. How does AT&T, as a carrier, compete with all of the other developer environments?
MT: That’s a good question. If you believed you could just do this on your own, I think you’d get to one answer. If you want to maximise the opportunity, do it in cooperation with other sponsors.
I listed some of the sponsors that are assisting with our hackathon and Shape event. We do a lot of reciprocal stuff. You’ll see us at some of the other leading developer conferences and events. If you took the developer ecosystem of IBM, Microsoft, Salesforce and Amazon, and you started to look at that in the aggregate, you’re talking about millions and millions of developers that AT&T wouldn’t traditionally touch.
You’re starting to see a lot more cooperation between technology companies that have access to different parts of the developer ecosystem and trying to figure out how to maximise that.
TR: To finish, I’d like to hear more about what you’re doing in smart cities and the role you’re taking.
MT: In January 2016, we announced a smart cities framework. We can help the ecosystem and governments with connectivity, platforms to build and deploy these solutions and vertically integrated solutions.
We’ve announced six cities that we’re working with: Dallas, Atlanta, Chicago, Miami-Dade County in Florida, Montgomery County in Maryland, and Chapel Hill in North Carolina.
The ecosystem is an important piece of what we’re doing. When we announced our framework in January, we also announced that we’re not doing this alone – we’re working tightly with technology companies, including Cisco and GE, Qualcomm, Intel and Ericsson, as well as IBM and Deloitte.
More recently we announced additional partners, including Hitachi and Southern Company, which is a utility here in the United States. So, it’s among these technology companies, as well as these cities, that we’re starting to think through and implement solutions for cities.
TR: How are these smart city solutions being funded, given than most cities have limited resources? Are you doing a revenue or savings share?
MT: It’s a great question. In America, there are over 4,000 cities, and we decided not to go after every city. Many cities have a vision, but they don’t have the budget.
The cities we’re going after have dedicated the resources, the capital and are pragmatic, starting with a couple of use cases – whether it’s route optimisation for buses or to monitor the water leaks. Some of these cities also own an electric utility, so smart metering, for example, becomes part of the purview of the city as well, and we’re doing a lot in that particular arena.
TR: Are the cities looking for you to take on any of that risk?
MT: We’ve had a couple of cities ask us to think about the business model to see if we can articulate that we’re going to save them X% in productivity, or fuel, or whatever it might be, and adopt a shared risk model.
We have not entered into any these agreements yet, but it’s one of the reasons why the framework of additional technology companies at the table becomes interesting. Some of our partners are open to sharing the risk on applications and services. I believe you’ll see that trend in the future, but we haven’t used that model yet.
TR: For smart cities, you’ve got all of these different partners. How do you manage those partners?
MT: What we did very deliberately was work with a handful of companies late last year to agree on the framework. This is often one of the most difficult parts of any business venture. It’s not necessarily the technology; it’s the rules and governance with other partners.
We also spent a lot of time trying to ensure that we didn’t have too many partners in a given vertical, or industry, or part of the stack. That was done on purpose so that we didn’t have too much overlap.
Because we’re a mini-consortium, there are going to be some companies that have deeper relationships than others with the cities. You have to respect that and fine-tune the relationship model based on who has a better propensity to win.
We’re engaged with six cities and a number of trials, and the relationship model seems to be working well.