Seattle-based resin.io is a small company that thinks large. With only around 50 employees, the company is focused on the intersection of the Internet of Things (IoT) and Linux-based containers.

Both IoT and containers are expected to see significant growth over the next several years. Strategy Analytics last year said it expects the IoT market to be worth $550 billion by 2025.  And 451 Research earlier this year predicted that the container market will reach $2.7 billion by 2020.

Alison Davis, director of product marketing and strategy at resin.io, explained that the IoT space requires both security and recovery, something that a container infrastructure is ideal at providing. This is where the company's host operating system (OS) that includes a Docker container engine comes into play.

Resin.io’s efforts revolve around separating core operations of the IoT device from the application layer. This allows for a device to continue to communicate with its network host should an application fail.

The resinOS can manage two containers, with one running a supervisor agent to make sure the device is running properly and can connect to resin.io, and the other running the user application.

“The host OS interfaces with the hardware watchdog, ensuring a reboot happens if there's any issue with the low-level software,” explained Zach Walchuk, technical content lead at resin.io, in a blog post. “In the end, this makes anything above that level an application issue that can be resolved remotely.”

Davis also explained resin.io has security safeguards in place. This includes needing a specific key for the application programming interface (API) managing IoT devices; not allowing any single device to control any other device on the platform; and using the container infrastructure to be able to more quickly push out security updates.

“We have a fundamental belief that you can secure what you can update,” Davis said. “Being able to update code on the device nearly instantaneously is key.”

Davis said the company’s push into containers and IoT actually started with real containers. The company had a customer in London that was deploying smart garbage cans for the 2012 Summer Olympics. The initial tracking devices proved to be unreliable, requiring a technician to go out to every garbage bin to provide updated software.

“Docker was new on the scene, and the founders were familiar with Docker and just connected the dots,” Davis said.

Davis said the company targets a diverse set of IoT customers, but has had better success with those coming from a cloud-native background “as they are more familiar with the workflow.”

“We have seen those customers become really successful, really quickly,” Davis said. “They have been able to scale up to thousands of devices in a matter of a months."

Davis said resin.io is different from other IoT-focused cloud implementations like Amazon Web Services’ Greengrass in that resin.io is more focused on management.

“We think you can run Greengrass in a container using resin.io to manage,” Davis said. She added that Microsoft is working on a similar management feature. “I think we have a similar view on where the support is needed.”

AWS introduced Greengrass in June as a way to help edge devices process data and communicate with the AWS cloud. Greengrass allows customers to use AWS’ serverless computing Lambda platform to run code locally on connected devices, similar to how they do it using AWS Cloud.

While software-centric today, Davis did note that resin.io may eventually make its own hardware. She said that the company has found a dearth of IoT hardware options that can take full advantage of the work resin.io is putting into its software.

The company has so far been defaulting to hardware built on the Raspberry Pi architecture, but Davis said it’s not really designed for production deployment for IoT.

“It works well, but things can be improved,” Davis said. “We are really just hoping what we can put together can help our customers scale more quickly.”

The Raspberry Pi is an inexpensive single-board computer originally intended to aid computer education, but it has gone on to become the platform for a range of consumer products.