One of the major announcements from Amazon’s re:Invent conference in Las Vegas is AWS IoT, a new cloud-based platform - currently in beta - which will support the Internet of Things (IoT). The platform connects IoT devices to the current AWS services and makes it easier to then store, process, analyse the data from the connected devices.
The devices connect to AWS IoT’s Device Gateway, and the device manufacturers can set rules – via the rules engine - for how AWS IoT handles the data they send, and the actions they take when various conditions are met (such as sending an alert when a pressure sensor reports an unusually high reading or a motion detector is triggered).
For each device AWS IoT creates a virtual version, or “shadow” that includes all of the information about the device’s state and is always available – even if the actual device is off-line temporarily to conserve power - so that applications can check the device’s status and take actions that are automatically sent to the device once it reconnects.
To speed up adoption of IoT AWS has developed an SDK and has encouraged semiconductor manufacturers including Arrow, Broadcom, Intel, Marvell, Mediatek, Microchip, Qualcomm, Renasas, SeedStudio, and Texas Instruments to produce “Starter Kits” Powered by AWS IoT that embed the AWS IoT Device SDK and offer connectivity to AWS IoT out of the box.
Devices connect to AWS IoT via the Device Gateway using both HTTP and Message Queue Telemetry Transport (MQTT), an industry-standard, lightweight communication protocol designed for sensors and mobile devices, making them interoperable independent of the protocol they use.
AWS IoT is integrated with AWS Identity and Access Management (IAM) and also provides mutual authentication so that data is never exchanged between devices and AWS IoT without proven identity, and encrypts all data coming into and out of connected devices.
Using the rules engine developers can create rules that apply to data from a single device (such as a sensor), a group of devices (such as a sensor array), or a mix of devices and data sources (such as a sensor array and data stored in Amazon DynamoDB). Rules specify conditions that, when verified, instruct AWS IoT to take actions such as routing data to Amazon Kinesis, Amazon S3, Amazon Redshift, Amazon Machine Learning, or Amazon DynamoDB. For example, AWS IoT may receive messages from connected industrial equipment that produce vast amounts of telemetry data each hour, not all of which may be relevant to the business. With AWS IoT’s rules engine, the business can instruct AWS IoT to filter certain types of sensor data (e.g. pump pressure) as it comes in, and route only the specified data to Amazon Kinesis Firehose to be streamed into an Amazon Redshift data warehouse for later analysis. AWS IoT rules can also trigger AWS Lambda to run code that will take more complex actions, such as compressing data, or sending a push notification to an operator if an anomaly is detected.
Many businesses are already experimenting with the AWS IoT platform including NASA, and LinkNYC a smart city consortium for New York that will replace the aging network of public pay phones.
The solution is available in the US, Dublin and Tokyo areas and pricing is based on the number of messages published upstream and downstream from AWS IoT. The first quarter of a million per month are free and are charged at $5 per million messages after that.