Alexa will be coming to low specification devices like light switches and thermostats. All that is required is a cortex-M processor and 1MB of RAM. Amazon’s Alexa has been used in many devices like microwave ovens, eye-glass, and earbuds. The requirements are not much, which means Alexa can be inserted into the dumbest of devices like toys or as basic as lightbulbs.
The basic requirement for Alexa in the past was 100 MB of RAM and an ARM Cortex-A processor. But since that has been changed Alexa is now applicable to devices that are not smart. The Alexa voice feature has been made available to companies that manufacture hardware so that they can incorporate it into their devices. This means you can expect Alexa on single-purpose devices.
But the availability of Alexa doesn’t mean it can perform complex voice recognition services. These simpler devices won’t have the same model and decision engine. These devices cannot retrieve media or decode audio, all of that is done in the cloud. The minimum requirement for Alexa is to be able to recognise the simple waking command, which is quite simple to configure into these devices.
Amazon Web Services (AWS) IoT VP Dirk Didascalou says “we now offload the vast majority of all this to the cloud. The device can be ultra dumb. The only thing the device still needs to do is wake word detection. That still needs to be covered on the device”. There was another important thing noted by Didascalou, NXP and Qualcomm have come up with new low powered processors, which simply just makes the entire ordeal so much more attractive!
This new feature gives manufacturers the potential to use in unfamiliar areas and most importantly, in the consumer space. Didascalou states “it just opens up what we call the real ambient intelligence and ambient computing space. Because now you don’t need to identify where’s my hub- you just speak to your environment and your environment can interact with you. I think that is a massive step towards this ambient intelligence via Alexa”.
AWS’s IoT services called IoT Greengrass is the company’s offers support to Docker Containers. AWS services have been extended to edge devices, and no cloud computing services can be discussed without mentioning containers. The reason for offering this service is pretty simple. Initially, the idea was to ask the developers to write the lambda functions for it. But according to Didascalou, there were many companies out there that wanted to bring this legacy and other third-party application for IoT Greengrass devices and some applications that were not supported by Greengrass. This meant that you were allowed to bring any container from the Docker Hub or any other registry related to Docker to Greengrass.
Didascalou explained “the idea of Greengrass was, you build an application once. And whether you deploy it to the cloud or at the edge or hybrid, it doesn’t matter, because it is the same programming model. But very many older applications use containers. And then of course, you saying, okay, as a company, I don’t necessarily want to rewrite something that works”.
Greengrass has another apparent feature which is new, and it is called Stream Manager. Up until now, developers had to work hard to build solutions to manage the data streams using lambda functions from the edge devices. But Stream Manager will relieve from doing this, and they don’t have to reinvent anything every time they deal with the data stream. They will not need to build a new solution for every connection management and data retention policies etc. They can simply rely on the Stream Manager to do that for them. It is pre-integrated with AWS kinesis and IoT analytics.
Greengrass also provisions fleet, which allows businesses to set up new devices more quickly and easily. It also secures tunnelling for AWS IoT device management. This makes it easier for the developers to gain remote access to a device and troubleshoot them easily. The IoT core also happens to feature configurable endpoints.