Whereas in last year’s re:Invent IoT post we noted that the focus was on providing more device-to-cloud integration options, this year the overarching theme is the maturation of many of AWS’s edge and embedded system software and services. This development should take us from simply having the ability to connect things to AWS to having a full suite of tools and features to develop IoT applications in an agile and scalable way. Let’s take a look.
Greengrass has seen a tremendous amount of adoption since it became available a little over three years ago: it has been used for applications ranging from machine learning at the edge to Internet of Things (IoT) device gateways. Nevertheless, it has lacked many of the features required to develop and maintain infrastructure at scale.
For example, many of our IoT customers often asked about the best way to perform canary deployments across select groups, monitor the connectivity status of a fleet of Greengrass cores, or easily navigate the Greengrass application logs. Customers often had to turn to third-party or custom-build solutions to address these needs.
The newest version of AWS’s edge solution is packed with features designed to significantly improve the development experience: a modular architecture, local development features, and previously-missing fleet management capabilities.
For example, before this announcement, building an application typically consisted of an arduous process of performing long deployments, digging through local logs, and devising mechanisms to deploy release versions to different groups of devices at scale. Now, the development lifecycle is streamlined by selectively adding components (potentially expediting deploys), allowing users to understand application performance locally through the Greengrass CLI, and leveraging the capabilities of AWS IoT Thing Groups to release final deployments in a specific order and at a specific rate–with rollback configurations also available.
Because of that, we believe that Greengrass 2.0 is the most significant IoT release from AWS this year.
Over the past couple of years, AWS has been hinting at a focus on industrial machine connectivity through the release of services like AWS IoT SiteWise and projects like the Connected Factory Solution–of which Trek10 is a launch partner. This makes sense given that the vast majority of industrial equipment still has no interaction with cloud-based data analytics or machine learning applications.
One of the biggest hurdles we have encountered with our clients who want to ingest data from their industrial machinery is protocol translation. There is a huge variety of industrial protocols available, both open and proprietary, that must be understood and translated to an application protocol like MQTT before data can be sent to and from AWS. To address this, AWS has partnered with vendors such as Inductive Automation to leverage their industrial connectivity solutions which can perform all of the necessary conversions. The data can then be accessed from the partner software, for example, through the Greengrass SiteWise Connector using the OPC-UA interface. This allows a Greengrass device to only support one industrial protocol (OPC-UA) instead of dozens, while still gaining access to data from industrial equipment. Furthermore, Greengrass devices can communicate with any equipment that supports the OPC-UA protocols.
The announcement of support for the Modbus TCP and EtherNet/IP protocols is massive because it means a Greengrass device can now communicate directly with hundreds of thousands more devices in industrial settings around the world without the prerequisite of a third party solution (though we believe that these third-party solutions still provide critical value for interoperability, operations visibility, and local data processing).
Cellular connectivity is perhaps one of the most important components of the Internet of Things. A very significant portion of IoT products use cellular connectivity to communicate with their backend resources. Cat-M is–to oversimplify it–LTE’s little sibling that has lower bandwidth but provides up to 7 times the coverage of other cellular technologies and is far more energy efficient--a perfect fit for many IoT applications.
Up until now, supporting cellular modules on an IoT device required companies to use vendor libraries, and having each customer coalesce them with Amazon FreeRTOS libraries and AWS keys and certificates within their application. I can speak from experience that this inconvenience, when paired with cellular certification requirements, can turn developing an IoT solution into the most time-consuming activity.
The announcement of Cat-M1 libraries for select modules built into the Amazon FreeRTOS package has the potential of drastically reducing the time-to-market for an IoT project while also providing support for what is often the best communications technology choice for IoT Devices.
Currently, connections in the kilobits-per-second range are sufficient for many IoT applications such as environmental monitoring or home automation. However, high-throughput wireless connectivity is the next frontier in the Internet of Things. AWS Services available at 5G Base Transceiver Stations make it possible for remote devices to become very closely knit with the AWS backbone.
Up until now, for example, many industrial facilities are still located far from high-speed Internet infrastructure, and running cable to them would prove difficult and expensive. With 5G, a single transceiver station is able to wirelessly serve multiple facilities within its radius, enabling machinery on a secluded manufacturing floor to interact with AWS services as if it were next to an AWS datacenter.
The general availability of AWS Wavelength starting in Tokyo demonstrates a leap in the availability of 5G for paradigm-shifting IoT applications. We listed it as a bonus since, unfortunately, it isn’t available at your nearest warehouse, but given the rapid deployment of the technology and AWS’s evident focus on this and all other edge technologies, we can still be ecstatic about the possibilities before us.
In the final installment of our AWS IoT journey, this post explores how to use IoT Jobs to update configuration information on a fleet of devices.
