This story was originally published in my June 30, 2023 newsletter. You can sign up for the newsletter here.
First, a disclaimer. Particle is currently a sponsor of the newsletter, but this article was not paid for and represents only my editorial judgement.
The company recently held its annual Spectra conference, where it released four new software products and three new pieces of hardware. And while doing so, it also highlighted exactly how the concept of “the edge” is changing.
For lack of something cute, I’m going to call this new edge concept, the complex edge. The idea is the same one I explored in the earlier story: a complicated edge that has layers of compute performing different tasks while communicating locally between various edge nodes, then back up to the cloud. And some of Particle’s new products help demonstrate how to handle the software and processes that arise from a complex edge.
Particle is one of the first in a wave of IoT platform companies that arose in the early 2010s. It provided a hardware platform with connectivity and compute built in. It also offered a cloud and software to both manage the hardware remotely and get data from the hardware to the cloud. Its customers included consumer brands such as Jacuzzi and industrial clients such as Field Intel.
During its Spectra event last week, Particle CEO Zach Supalla focused a lot on the insights derived from connected devices and how those insights can help change businesses. The company launched an updated version of its hardware, called the Photon 2, with more processing power (so it can perform ML tasks), dual-band Wi-Fi, modern Bluetooth, and pricing that serves both developers making prototypes and businesses that want to use the Particle hardware in millions of units.
It also launched a customizable industrial automation platform called Monitor One Developer Edition, which provides customizable hardware for industrial gateways. The hardware includes a board with a processor and connectivity, but also a board that can be customized to the client’s wishes by attaching to a wide variety of external sensors and providing a home for custom hardware (such as hardware that might be required to run ML or process sensor-specific data on a device).
I view the customizable version of the Monitor One hardware as a recognition that the demands of the myriad connected edge devices out there are so complex that a company may need to design these gateways themselves. But because these complex use cases may end up performing computing in a product that could sell hundreds of thousands of units, it’s still economical to embrace that complexity.
Consider, for example, a connected lighting system designed for offices. A gateway computer for managing the system might need direct access to lux sensors or even air quality sensors, as well as specialized chips inside to run algorithms associated with building management. This is a workload or computing use case that may not become as ubiquitous as a smartphone, but it is something that will be needed in millions of offices around the world. Particle is trying to provide a more customizable, off-the-shelf option so the lighting company won’t need to hire a host of chip and connectivity engineers to build their product.
There are dozens if not hundreds of similar potential use cases that will make up the complex edge.
On the software side, Particle launched Asset OTA, a product that will manage over-the-air updates for both the Particle board inside a device as well as any sensors, secondary computers, or even displays that might also be on the device. This product acknowledges the multiplicity of connected devices that might be part of a single connected product, all of them part of the complex edge.
In an email, Supalla said, “We think of IoT as a distributed computing problem where the logic/compute can run in multiple places and the “application” is the sum of each of the separate bits of software.” He agrees with this idea of the complex edge and adds that the cloud could also have this level of complexity as data get aggregated from different machines and services.
I’m pretty excited about this, because it delivers on more of the opportunity that we have with a connected device. Now we can update not just the primary software that controls a connected product, but the sensor capabilities or even the models that might be running on the sensor edge.
Today companies have these capabilities, but they have them because they are giant tech firms or have hired the types of embedded engineers needed to handle this level of complexity. But as we shove computing into more aspects of more products, we also need to find ways of making it easier for more developers to address those computers. Especially as we embed AI further out into the edge.
Update: This story was updated July 6, 2023 to correct the name of the Asset OTA product. It is Asset OTA not OTA Asset. I regret the error.