At Microsoft Build 2017 we shared our vision for Azure IoT Edge. At that time we demonstrated the idea of what Azure IoT Edge could do with Azure Stream Analytics job and Azure Machine Learning workloads by moving from Azure cloud to an IoT Edge device so that decisions could be made quickly and reliably without having to rely on the public Internet. This allowed million-dollar machines to operate reliably and safely with Azure IoT Edge, making intelligent decisions based on data from the machine. Deployment of these workloads can now be at a scale of thousands or millions of devices.
In November 2017, we shipped the public preview of Azure IoT Edge with the ability to move workloads from the Azure to Edge devices. These workloads included the AI Toolkit for Azure IoT Edge, Azure Machine Learning, Azure Stream Analytics, Azure Functions, and your own code. With these services on the edge, you could build rich and intelligent applications which operate locally, but can easily be deployed and managed from the Azure IoT Hub. We also articulated the security model for Azure IoT Edge and announced partnerships with NXP and Microchip for support of this model.
Since the public preview, we have seen many of our partners use IoT Edge in unique and innovative scenarios starting from controlling machines in the industrial IoT space, to controlling machines which may not have stable network connectivity, to using IoT Edge on machines that move, and more.
I am pleased to announce that Azure IoT Edge will be in general availability in the next couple of months. We have added many new features and services to Azure IoT Edge, all of which can be packaged from the cloud and deployed to the edge. Furthermore, we’re adding a module marketplace which lets anyone buy and sell logic for use on edge devices!
Below are some of the key innovations you will find when Azure IoT Edge is generally available.
AI on Azure IoT Edge
Azure IoT Edge will allow deep integration with AI services, which are part of Microsoft Cognitive Services, bringing true AI to the Edge. Services such as Custom Vision on an edge device can enable scenarios which previously required calls to the cloud. Custom Vision models can be trained in the cloud, and these models can then be packages and deployed on an edge device as modules. Using Azure IoT Hub to manage a fleet of edge devices, these models can be updated when new data is available and the model adapts. Soon, there will be additional AI services coming to Azure IoT Edge.
Azure IoT Edge core technology
Azure IoT Edge, currently in public preview, offers the following functionalities:
- Seamless deployment of Azure services to Azure IoT Edge, including Functions, Stream Analytics, Machine Learning, and SQL Server databases, from the cloud
- Routing of messages from and to modules, and to the cloud
- An application model based on message passing between modules, devices, and the cloud
- Monitoring and managing modules remotely
- Store and forward in case Edge is operating in unstable networking environment
- IoT Edge used as a gateway for downstream IoT devices, with transparent device management
- Scaled deployment of Edge modules from the cloud
- Tooling for developing, debugging, and deploying custom modules in multiple languages
- All of this done securely and at scale
Based on feedback from our early adopters in the public preview we are adding new and exciting features to Azure IoT Edge which will be available when Azure IoT Edge is generally available.
Azure IoT Edge will have deep integration with Device Provisioning Service for zero touch provisioning so that a device can simply be provisioned in the field with no operator intervention. Azure IoT Edge also supports custom modules which can be coded in any language. We will support five languages including C#, C, Node.js, Python, and Java.
Azure IoT Edge can be deployed on both Linux and Windows for X64 and ARM architecture. Azure IoT Edge can be deployed on a range of hardware devices from an industrial gateway class device all the way to a Raspberry Pi or equivalent device. IoT Edge offers store and forward, which enables it to be operational even in unstable network environments. Workloads on Azure IoT Edge are deployed in modules which are containerized for ease of deployment. In addition, IoT Edge can be deployed at scale using Azure IoT Hub portal user experience or via APIs which can be used to build your own deployment and management portal. Finally, in order to enable our partners and developers to fully utilize and extend the power of Azure IoT Edge, we will be open sourcing the Edge runtime on Github.
Azure IoT Edge security
Azure IoT Edge will deliver the first phase of its security framework. Specifically, it will deliver a framework that will facilitate managed security in an ecosystem, as well as adoption of the right security solution.
Azure IoT Edge will see the introduction of the Azure IoT Edge Security Manager as a well bounded security core for protecting the IoT Edge device and all its components. The IoT Edge Security Manager is the focal point for security hardening and is there to provide Original Device Manufacturers (OEM) the opportunity to harden their devices based on their choice of Hardware Secure Modules (HSM).
Azure IoT Edge will enable our partners and customers to make the right choices of security models for any given IoT Edge deployment. Choosing the right hardware for an IoT deployment involves several considerations and can be daunting without guidance.
Azure IoT Edge will avail support of a class of Hardware Security Modules (HSMs) which includes trusted platform modules and custom secure silicon.
Use Kubernetes to manage Azure IoT Edge deployments
Azure IoT Edge is built on open container technologies which allows it to integrate seamlessly with other amazing projects in the ecosystem. To demonstrate this, we have built a tool that allows users to manage IoT Edge deployments using Kubernetes concepts and vocabulary. It enables some interesting use cases that would be more difficult to achieve otherwise. For more details check out this blog post.
Edge marketplace
I am pleased to announce that partners and developers will soon be able to share their Azure IoT Edge modules and even monetize them through the Azure Marketplace. Azure IoT Edge users will be able to browse and find pre-built modules to accelerate their edge solution development. This is a first step toward building a rich Azure IoT Edge ecosystem with ready-to-use solutions for all industries. Start developing your edge modules now and get ready to onboard them!
Azure IoT Edge hardware certification
Azure Certified for IoT program is expanding to support hardware for Azure IoT Edge. The Azure Certified for IoT device catalog provides an easy and intuitive way to discover the right IoT device for intended use cases today.
We plan to highlight certified hardware for Azure IoT Edge in the device catalog to show hardware that can provide core functionalities such as AI, device management, and security promises. To start, we have partnered with Advantech, Beckhoff Automation, HPe, Moxa, NexCom, Plat’Home, and Toshiba.
If you are hardware manufacturers or customers who are working on hardware for Azure IoT Edge, please send email to iotcertdisc@microsoft.com for additional program information and requirements.
You can learn more about Azure IoT Edge through our product page. I am looking forward to announcing the general availability of Azure IoT Edge in the next couple of months, and to see how our partners, customers, and developers will use it to bring the power AI and cloud services to IoT deployments!