Startup Crowsnest is creating an API abstraction layer and ecosystem to solve one of the biggest pain points of the Internet of Things: how to connect a proliferation of devices. If there are expected to be more than 20 billion devices connected in an IoT-enabled future, then each of those devices will need an API, creating a lot of device integration work for developers before they can start coding applications and products. The Crowsnest founding team spoke to ProgrammableWeb about their vision for device integration, some early use cases, and how they are solving one of IoT’s other biggest pains: finding a viable business model.
The Internet of Things: Not So Fast
In a closing keynote at APIcon, MuleSoft (the parent company of ProgrammableWeb) CEO Ross Mason cautioned against too much hype around the Internet of Things, saying:
There is quite a lot of work that needs to happen to enable the IoT. Everyone is building their own proprietary networks, but to build the best applications, you have to get data out of devices.
For application enablement and applications themselves, this is what I am really interested in, as they add value for the end user. Every device is building their own application-enablement platforms. What will be really important will be creating APIs that enable devices to be captured.
Ross Mason speaks about IoT enablement platforms at APIcon.
It is this challenge that Crowsnest is seeking to solve by creating an abstracted API that can enable connections between any devices.
An Open Source Ecosystem of Device Plug-Ins and an Abstracted API
Crowsnest — a startup from an ex-venture capital policy analyst and two engineers with backgrounds in software development, systems admin and audio engineering — aims to build itself into an open source IoT ecosystem of plug-ins for connected devices. These plug-ins can then be routed through Crowsnest’s IoT API to let developers focus on creating applications rather than on managing devices. Crowsnest is working on building device plug-ins for digital cameras at present, to populate the ecosystem, but will be crowdsourcing from its developer community so that anyone can add a device plug-in.
“Cool things will start to happen when developers don’t have to deal with device integration and can use APIs instead,” says Crowsnest co-founder Michael Kruk. “A new wave of innovation will happen when developers can focus on the layer that sits above devices, when devices can communicate with each other independent of human input. Applications that could let your car talk to your coffee maker and to your alarm, so that if you are running low on gas, they can set themselves to get you up five minutes earlier.
“Developers can focus on building the applications that can make those decisions. Once you can get IoT into the hands of developers, there is an incredible amount of business potential,” he says.
Crowsnest has started by developing plug-ins for the Foscam series of digital cameras and already has some customers using webhooks (in lieu of the abstracted API to be made available soon) to ease deployment and connectivity of the cameras in retail environments.
The Crowsnest pricing model provides two free device integrations and then charges $5 per additional device integration, with additional charges for data usage above 500 MB a month.
IoT Growth Needs Use Cases
While tech analyst Gartner forecasts a $1.9 trillion Internet of Things market within the next six years (as referenced in Janet Wagner’s recent SigFox story on ProgrammableWeb), the reality is much more nuanced for any business or startup seeking to enter the market.
A report from the International Data Corp. recommends that market entrants should take a use-case research approach to understanding where the opportunities lie in the emerging market. Scott Tiazkun, senior research analyst, said:
A bottom-up research approach is crucial to understand the specific applications of IoT for each vertical market and educate vendors on the overall IT opportunity available to them today. The use cases validate where the rubber meets the road when it comes to IoT, and this report will serve as a useful tool in evaluating this burgeoning market.
The IDC study identifies healthcare and retail as the two verticals most prepared for leveraging IoT technologies, with spending in these two sectors growing at between 7% and 10% annually. At least with retail, it is a trend Crowsnest is seeing among its early adopters.
“Our camera API (which is our minimum viable product) has been used for consumer analytics in a shopping complex. There are also a couple of other customers who want to use the cameras as part of their digital displays: using the cameras to identify what sales promotions to display in real time that are relevant to the customer,” Kruk says, explaining some of the initial use cases.
A Viable Business Model?
So in this sense, the initial business model might make sense. If customers can use the Crowsnest API to reduce the burden of integrating the cameras into their in-store use and instead focus on how this camera data can flow into their retail business operations and marketing, then there could be a return on investment worthy of the initial Crowsnest pricing structure.
But like the IoT API discover platform Connect2.me, Crowsnest’s initial pricing policy seems a big ask, and is focused on the end customers (developers) paying, not the device manufacturers.
Crowsnest admits that in the longer run, it has plans to alter the business model. As part of a lean startup approach with an MVP, having some sort of pricing structure is useful as it can teach the company more about the market than a purely freemium model. But it does seem particularly problematic, given that Crowsnest’s ecosystem will require leveraging a loyal community of developers in order to expand at pace.
“One of the main things we are trying to accomplish is to build a community behind the integrations,” confirms Kruk.
“We are open sourcing the plug-in system. The open framework is free as a way to try and engage with the community,” adds Ian Wilson, co-founder and engineer at Crowsnest.
Still Some Way to Go
Crowsnest’s overall idea may be well-placed. Mason, in his talk at APIcon, showed the struggles he had had when trying to build a connected home network for his IoT devices. He explained:
At the moment for the connected home, you have to build different connectors for Hue, Nest, etc. Networking has gotten amazingly complicated. When you get into IoT, getting low-power Bluetooth to your phone and having it reliable is super-difficult. And then your phone is an unreliable network. At the edge, you have many more connectivity problems than you ever had.
To add a thermostat to my home network, I had to do a lot of heavy lifting. I quickly needed an ontology: a vocabulary that my home network could understand. Every time I add a device, I have to add it to my vocabulary and I need to explain how to turn things on and off.
So Crowsnest’s plan is to be that vocabulary. Creating a vocabulary framework that can grow to handle all types of IoT devices may take some time. The team is working on specifications for messaging protocols and hopes to provide its developer API soon, as well as launch the open source plug-in ecosystem. At present, the digital camera interactions are managed by webhooks.
“We have the confidence around the whole platform, and we are confident we have the team to make it happen,” says Kruk. “Our next big push is to open up the plug-in framework, put that more in the hands of the community.”
“This will also gives businesses the chance to inspect the code that is running on their network,” adds Joshua Sorenson, co-founder and engineer.
Crowsnest’s biggest struggle in growing its business — apart from the technical complexity of device integrations — may well be recruiting a developer audience that wants to contribute to the ecosystem while also being willing to pay for device integrations.
There may have been alternative business models that could have been considered even at the MVP stage: Again, device manufacturers may be willing to pay for the integration work to be done for them, or for the leverage they gain from device discoverability with an active and interested developer community.
Going forward, Crowsnest may have plans of providing additional analytics on top of streaming the device data via its plug-in ecosystem, but these business model discussions are very much “a young conversation,” Kruk says.
Everyone is hoping for a piece of the IoT trillion-dollar pie, and infrastructure services are a promising place to start, as it is estimated that only 1% of connected devices are actually connected to the Internet or each other:
However, the business models that will enable infrastructure growth are still far from sorted out.