AWS ups IoT game by bringing Alexa to the edge

Alexa Voice Service part of broader IoT services refresh

The voice of Amazon, Alexa, is leaving the confines of bedrooms and kitchens around the globe to give the world’s ever-expanding fleet of IoT devices the gift of speech.

Alexa Voice Service, announced ahead of December’s AWS re:Invent event in Las Vegas, makes Alexa available on IoT devices with limited local processing power and storage.

Previously, devices required at least 100MB of RAM and an ARM Cortex A-class processor to support the voice assistant. Now, manufacturers will be able to integrate Alexa into devices with low-powered chips and just 1MB of RAM.

The service will effectively serve as a bridge to the cloud, which will do all of the heavy lifting (retrieving, buffering and coding etc.) once a message is captured and sent from an endpoint device. In principle, this will make any of the devices that Echo can currently control directly voice-controllable. In other words, you will soon be able to talk to your light bulb.

The announcement was one of eight IoT services updates announced by AWS, largely aimed at simplifying the deployment of IoT devices at scale.

AWS IoT Greengrass, the service that extends AWS functionality to the edge, has been injected with container support, allowing devs to deploy, run, and manage Docker containers and applications on AWS IoT Greengrass-enabled devices. The main idea is to enable applications to take advantage of Greengrass’ analytics functions even if they aren’t natively supported by the service. Developers simply package apps into containers and deploy them to the edge, rather than rewriting them.

Stream Manager for Greengrass allows organisations to develop IoT applications that can collect and transmit streams of data — data recorded continuously rather than in batches. A first stage of processing is performed at the edge before some or all of it is routed to the cloud for second-tier processing. Handling data streams on Greengrass was possible before but it required additional coding effort, what’s different now is that the functionality is built in, saving developers time and resources.

Fleet Provisioning simplifies the bulk onboarding of IoT devices, so users don’t have to uniquely configure every IoT device at the point of manufacture. Lastly, with Secure Tunneling users can access, troubleshoot and fix IoT devices located behind restrictive firewalls, as is common with medical devices.