Amazon Web Services (AWS) this week unleashed a flurry of updates to its IoT platforms, and for the first time revealed how many devices are connected to its IoT core.

“We have more than half a billion unique, externally connected endpoints going through IoT Core alone,” Dirk Didascalou, VP of IoT at AWS, told SDxCentral in a phone interview. 

The most pertinent updates to AWS’ IoT feature set, according to Didascalou, are framed around three categories: software for device management and control, cloud software for managing connectivity and performance in large fleets, and analytics fused with artificial intelligence (AI) for application services.

The cloud giant completely rewrote the software stack for its IoT connected edge platform and released AWS IoT Greengrass 2.0, which now supports more programming languages and hardware. The software is fully modularized, based on Java Virtual Machine, and entirely open source, Didascalou said.

“In the past you had to deploy it from the cloud. Now you can locally develop, test, and deploy, and then automatically sync with the cloud so that you can deploy at scale,” he said. 

AWS also added long-support to FreeRTOS, its open source, real-time operating system for microcontrollers that simplifies the process of programming, deploying, securing, and managing low-power devices at the edge. The company said it will now provide security updates and critical bug fixes for at least two years to provide additional assurances to device manufacturers that use FreeRTOS.

IoT Fleet Management

In the next category, which focuses on the needs of enterprises with large fleets of IoT devices, AWS released AWS IoT Core for LoRaWAN, a managed service for low-power, long-range WAN. The release, which targets asset tracking and sensors that can last for up to a decade with low power requirements, removes the need for enterprises to develop or operate a local LoRaWAN network server.

AWS also integrated Sidewalk, its consumer focused asset tracking system, to AWS IoT Core, and introduced Fleet Hub for AWS IoT device management, which provides a web-based dashboard for customers to monitor IoT devices across their footprint.

Finally, AWS revealed AWS IoT SiteWise Edge, which is software that runs on premises at industrial sites designed to help enterprises collect, organize, and monitor equipment data before it’s sent to the cloud. SiteWise Edge can be installed on local hardware, AWS Outposts, or AWS’ Snow family of compute devices.

The industrial IoT service brings capabilities of the cloud, including modeling of different assets, key performance indicator metrics, and dashboard visualization to a local network, Didascalou explained.

“Our job is to make sure that every one of our customers know the state of everything, of every asset that they have, and that they can easily reason on top of that data knowledge,” he said.

AWS Aims to Make Sense of IoT Data

“We connect all of the dots with and for our customers now. So the building blocks have been built, now we tackle the next complication, which is how do I combine them all to make sense for my use cases,” Didascalou added.

While he argues that AWS is the leader in the consumer IoT space and that it’s gaining momentum in industrial markets, he noted that IoT is increasingly making its market at large. “It’s not about the technology anymore, but it’s useful in all of these verticals and use cases. That’s maybe the biggest difference now, that there is not one area of momentum, it’s just everywhere,” he said.

Sarah Cooper, GM of Outcome Driven Engineering at AWS, recently shared a similar view in an interview with SDxCentral in which she argued that IoT will soon disappear as a term “because fundamentally everything will be connected.” 

“There is no single one IoT," Didascalou said. "IoT is very different because it tackles every use case that you can think of. The more tailor made and specialized communication protocols you have, the more different use cases you can handle.”

IoT won’t be converging to a point where specialized services are no longer required — quite the opposite, “it’s actually divergent,” he said. 

“If at all, specialization is the trend here. What that means ... is that you have to have more and more capabilities to cover them,” Didascalou said. “The real trick here is be as specialized as you can, but make it as simple as possible for a non-specialist to apply.”