AWS aims to accelerate IOT adoption with new services, OS

Of all the buzzwords relevant to AWS, IoT may be "delivering the fastest," AWS CEO Andy Jassy said, but there's plenty of room for development.
Written by Stephanie Condon, Senior Writer

Video: The biggest challenges and opportunities facing Amazon Web Services

Referring to "IoT" one of the more legitimate "buzzwords" in the industry, Amazon Web Services CEO Andy Jassy on Wednesday introduced a series of new Internet of Things services, including a new operating system for devices that run on microcontroller units (MCUs).

Along with the new OS, called Amazon FreeRTOS, AWS on Wednesday announced a service called IoT one-click to easily create Lambda triggers, AWS IoT Device Management, AWS IoT Device Defender, AWS IoT Analytics, as well as AWS Greengress ML Inference.

See also: Alexa for Business likely to win in smart office | AWS not buying into the blockchain hype | Amazon to put machine learning into the hands of more users | AWS rolls out new graph database, more database functionality

"Of all the buzzwords of choice... that we've been working on at AWS, IoT might be delivering the fastest in terms of the actual number of companies doing real work there," Jassy said during the first keynote address at the AWS re:Invent conference. While there's already been significant progress on the IoT front, Jassy said, "We're now just entering a world where the growth in number of [connected] devices is going to be exponential."

The vast majority of smaller devices, however, are not connected to the cloud. Many small devices such as smoke detectors don't house a CPU -- they run on MCUs. To enable connectivity for those devices, Amazon is introducing Amazon FreeRTOS, an IoT microcontroller operating system. It extends FreeRTOS, a popular real-time OS, with libraries that enable local and cloud connectivity and security. It should soon enable over-the-air updates. It's open source and available on GitHub.

Via the Amazon FreeRTOS console, a customer can select and download trusted software. There's a Qualification Program to help ensure the microcontroller a customer chooses will run consistently across several hardware options. For customers that have these smaller devices but don't want to make trip to the cloud or don't have the connectivity to the cloud, devices can connect via AWS Greengrass.

Meanwhile, for simple IoT use cases, AWS is rolling out AWS IoT 1-Click. It allows for the simple creation of an AWS Lambda trigger to execute a specific action on any device. For instance, Jassy said, a manufacturer may want a sensor to trigger a warning any time a human walks by a particular tool in a factory.

Once a device is registered, a customer can choose from pre-built Lambda functions or build their own -- then simply click on it. It should "help a lot more companies start using IoT," Jassy said.

AWS is also launching AWS Device Management. The service's console, customers can onboard, organize, monitor and remotely manage devices at scale. The service helps a customer mantain an inventory of all the device information needed, such as serial numbers or firmware versions, and allows them to query along that information and find out where troubleshooting is needed. Devices can be managed as an entire fleet, individually, or in portions.

Meanwhile, AWS IoT Device Defender offers a few security and compliance-related services. It audits and monitors an entire fleet of devices to ensure they're in compliance with all required policies, and it alerts a customer when they're not being met. It monitors a fleet of devices for abnormal behavior.

"Effectively tell us what is your expected behavior, what ports do you want open, where do you want to send traffic?" Jassy explained. Device Defender will notify the customer if it sees something different. It will provide information like device information, device statistics and diagnostic logs to help address any problems.

To clean, store and analyze data from IoT devices, AWS announced AWS IoT Analytics. IoT data can be "hard to deal with," Jassy said, since devices often don't produce structured data, and they're often in motion with intermittent connectivity.

With this service, a customer defines the analytics channel and selects data they want stored. They can then ingest data and configure it to enrich it or transform it. For instance, if a vineyard manager collects data on soil moisture, he can enrich it with weather data to determine how much to water his crops. He could transform weather data from Celsius to Fahrenheit. There's also a built-in query engine to answer questions about a customer's data.

Lastly, with the new service Greengrass ML Inference, customers can build and train machine learning models in the cloud and then send them to the edge visa the Greengrass console.

Editorial standards