The World Economic Forum says there will be more than 40 billion IoT-connected devices by 2025. Regardless of what type of product they make, most manufacturers believe they’ll need to remain competitive in the IoT business. And yet, for every IoT project that reaches the market, there’s more than one that doesn’t make it. Many IoT projects have an early death, or they’re put down before proving their concept.
According to a recent IoT Signals report from Microsoft, 38% of companies with IoT ambitions say that complexity and technical challenges are holding them back. Nearly as many cite a lack of staff resources or simply “lack of knowledge.” (Executives who ticked that last box should at least be commended for their honesty.)
Designing from the Cloud Down
Manufacturers aren’t stumbling on the “T” part of IoT. They know how to make things — that’s what they do. They’re stumbling because today’s IoT platforms were designed by and for software experts for whom cloud computing, databases, networking are second nature; that is to say, they’re designed from the cloud down.
The unique selling proposition of Golioth: it is an IoT platform created for engineers working from the device up.
The company has just emerged from stealth and is now opening its private beta to a wider
hardware developer group. “I cut my teeth at Nest, working on the networking and wireless technology that powers Nest products; then I did hardware at Particle,” says Golioth’s
founder, Jonathan Beri. “I realized that the companies that will actually build those billions of connected devices are all hardware companies, but the IoT part of the devices is all cloud software. Hardware companies don’t have that expertise and don’t want to become cloud companies.”
Mr. Beri, who spent seven years at Google, saw that disjunction as an opportunity; that was the genesis of Golioth, which he started in early last year to simplify the cloud and streamline IoT for hardware developers.
Those developers’ challenges begin with some of the first decisions they have to make, including picking a cloud provider. That choice has to be made while many other factors remain unknown, but once it’s been made, it’s baked into the new product. That decision imposes permanent restrictions on hardware – there are wireless chips compatible with Azure but not AWS, and vice versa.
Golioth’s Customer Avoid Pitfalls
Golioth’s customers will avoid such pitfalls because, as Mr. Beri says, “Golioth is built using cloud-native technologies such as Kubernetes and containers so that it can be deployed wherever customers will require it.
We’re starting with an option hosted by us, but we can run it on any of the big clouds from AWS to Alibaba Cloud, on-premise or even on an oil rig.” In effect, Golioth promises to be not just cloud-native but cloud-agnostic.
Those huge companies bring comprehensive suites of services to the table when selling themselves to hardware makers, but they blithely ignore the attendant complexity. When pitching Golioth to investors, Mr. Beri used AWS’s own case study of its client Pentair to illustrate this point.
A schematic of the reference architecture includes nearly 20 separate AWS cloud services. Each
one had to be enabled, configured, secured, and managed – to say nothing of paying for it.
Mr. Beri appeared on the Amp Hour podcast earlier this year. Highlighting this kind of complexity prompted the host, Chris Gammell, to quip-quote Carl Sagan: “If you want to make an apple pie from scratch, you must first invent the universe.” That was a slight exaggeration (OK, not so slight!), but smaller hardware devs simply don’t have the necessary personnel or expertise to
marry up with Azure, AWS, or Alibaba. Even in big companies, hardware and firmware are often siloed apart from cloud engineering.
Secure Connection Between a Device and Cloud
Examples of things Golioth will take off developers’ plates include: Establishing and maintaining a secure connection between a device and the cloud, including key management; collecting sensor data; monitoring device telemetry and health; pushing software updates; and fleet management.
Today, Golioth is also announcing a $2.5M seed investment from Zetta Venture Partners. Zetta partner Jocelyn Goldfein explained the VC’s confidence in the startup by saying, “Connected devices and edge computing are the future, and Golioth will be an essential part of the stack that makes it happen. Device makers want to design hardware and firmware – not manage Kubernetes clusters and data warehouses. The effort required to stand up a cloud backend is an obstacle to progress.”
A key member of that team is Golioth’s CTO Vit Prajzler, who obviously shares Beri’s vision because he too has seen hardware designers struggle to reach the cloud. Mr. Prajzler is an IoT pioneer on the original team at IBM Research that developed the LoRaWAN protocol.
“I have seen businesses struggle to transition from proof of concept to production, resulting in years of delays and millions of dollars lost.
Frequently, the problem was talent scarcity and a gap between embedded teams and cloud teams. Our mission is to change that. We are building a platform that scales easily and can be used by embedded engineers a cloud engineers alike.”
Golioth’s Long-Term Plan
The long-term plan calls for Golioth to support multiple SDKs and protocols. The initial beta offers support for the Zephyr device SDK and the CoAP protocol. Zephyr is the fastest-growing RTOS, vendor-neutral, and Mr. Beri and other team members contribute to the Zephyr open-source project. The CoAP protocol is the leading standard in cellular connectivity.
With seed capital and a small team in place, the next step for Golioth is to add to its existing small group of beta testers. The company seeks several hundred hardware developers who will be able to use the platform free of charge.
If Mr. Beri and Mr. Prajzler are right, there should be at least that many hardware developers who want to get devices to market without first becoming – or hiring – cloud experts.
Image Credit: tima miroshnichenko; pexels; thank you!
