Imont: Has IoT broken the cloud business model?
Has IoT broken the cloud business model? Nigel Pugh, CEO at IoT software start-up Imont, believes there are issues that need to be addressed.
Pugh is the former head of platform engineering at what was once called AlertMe, now known as Hive, the smart thermostat arm of Centrica’s British Gas. Having spent almost three years developing the team and the technology behind Hive’s smart home product range, Pugh took a break in April 2016.
During this time, Pugh and two former AlertMe colleagues, Roman Sonkinas and Andrew Woodhouse, apparently asked themselves, ‘If we were to do it all again, what would we do differently with all the things we had learned?’
Inspired by advances in edge computing and the issues they perceive with the current cloud computing business model, Sonkinas and Woodhouse decided to found the company Imont, with Pugh joining in January of this year. Internet of Business spoke to Pugh about this new project, and how he sees the cloud computing model changing with IoT.
Software for the many, not the few
Ultimately, the reason behind setting up Imont was to “to make a generic piece of software that powers connected home systems and scales better than anything, but at a reasonable cost,” Pugh said.
“You can scale pretty reasonably with Amazon, but Amazon will charge you for the cost. So that’s our raison d’être, we want to commoditize the software. We want to make the best quality, the most scalable software available to everyone at a reasonable price.”
Pugh admits that the company is at its “1.0 stage” in terms of development, with plans to launch fully in six months, but added that the software is already useful today for anyone who wants to build a connected home system – if you can find out where to download it from the website, he jokes.
But who is Imont targeting? Pugh believes that Imont’s appeal comes from current flaws in the cloud computing business model.
“I think [our software] will appeal mostly to larger companies, who have in-house development teams who want to roll out systems of their own. I think it will appeal to device manufacturers as well, who, as they ship more devices, [find that] the cloud that they need to support these devices gets bigger and bigger,” he said.
Here Pugh reaches his defining argument: for consumers, he asserts, you buy your thermostat and pay once at the point of purchase. But for the business that provides that thermostat, be it Centrica’s Hive or a start-up like Tado, a monthly subscription is needed to pay for the cloud for as much as 25 years. That gets expensive.
Instead, Imont wants to introduce a hybrid model. “The connected home has a hub that doesn’t do an awful lot,” Pugh said. “We see it differently. Most people build a piece of hardware and then put software on it as an afterthought. For us, the hub is fundamentally a software piece and it can pretty much run anywhere you want.”
One obvious place for Imont to do just that is in the home router, so it’s currently in discussions with telcos to form partnerships, but Pugh said that this could also extend to things like IP cameras.
Devolving power to the edge
Imont is all about increasing the data processing power at the edge of the network. “You can have a thermostat on the wall talking to the hub, and that’s talking to the cloud over some kind of message bus, and then your phone is talking to the cloud, so the cloud passes messages back and forth. The other thing it does, is it’s aggregating data and starting to pull together insights, but it does this by taking in all the detail. So, every time your smart plug sends an update saying ‘the signal strength between me and the hub has changed’. That’s a lot of data.”
Imont says you don’t need to send all that data, you can do most of the analysis at the edge. “Even a Raspberry Pi has a lot of power, and it has a GPU, so it’s good for AI,” Pugh says. “It may not be as fast as an Nvidia GPU, but it has time.”
Pugh summarizes that, by embedding software at the edge, we can reduce the amount of data flowing to the cloud and only send insights about your particular connected home ecosystem. Then the cloud can aggregate that data across all of the homes in the country and update what it’s learned back to your hub.
“In the end, they’re not going to have the capacity to use the cloud. Our system just makes a direct connection from the phone to the hub, which is very immediate and all the control is there,” Pugh acknowledged.
Read more: Are we any closer to IoT Edge Computing?
Playing to the cloud’s strengths
But Pugh doesn’t sound the death knell for the cloud just yet. “There are some things that the cloud does really well. Software-as-a-service is obviously going to persist. [So will] aggregating data across entire ecosystems that you can’t do on the edge. [These are] the cloud’s strengths,” he says.
“It’s a hybrid between having a bit of cloud here and a bit of cloud there and using it for what it’s good for. Really all we’ve done is we’ve taken the tech that sits in the cloud and stuck what we can at the edge. But there will always that thing which you can’t and the cloud will always be great for that.”
Imont is currently privately funded, as it seeks to develop its technology further, but Pugh confirmed that it is already looking at how it can adapt its software for environments beyond the connected home. With the likes of Andreesen Horowitz, and more recently Aruba Networks, predicting this kind of hybrid model between the cloud and the edge, Pugh and the Imont team look set to be at the forefront of this change.