Microsoft expands Windows IoT at Ceatec
October 23, 2019
In a keynote speech at last week’s Ceatec show in Japan, Ian LeGrow, Microsoft partner director, announced general availability of Windows IoT board support packages for NXP’s i.MX family of processors along with other Windows IoT additions.
“The world of IoT is more exciting than ever, transforming every business on the planet, and that transformation is accelerating,” said LeGrow. “Advances in AI and ML are helping companies across industries gain efficiencies; reduce downtime, waste, and errors; and improve customer satisfaction by harnessing the power of the intelligent edge and intelligent cloud.”
During the keynote, LeGrow shared Microsoft’s strategy for how Windows for IoT could help device builders of all sizes quickly create IoT services that would bring digital transformation to life for customers with smart and secure IoT devices.
“With a special focus at Ceatec on the evolution of smart cities, smart buildings and smart infrastructure under the Society 5.0 framework, we’re excited that Windows for IoT delivers a scalable family of operating systems as the foundation for building intelligent edge devices,” he said.
This family includes:
- Windows 10 IoT Core: This is built for small-footprint smart devices, while still offering manageability and security and, along with Windows 10 IoT Core Services, ensures long-term OS support and services for managing over-the-air device updates and device health. An example of how Windows 10 IoT Core is being used is Hourfleet, a ride-sharing platform addressing urban mobility problems faced by many cities, towns and communities. The Hourfleet platform is a globally deployable, robust and cost-effective product underpinned by Windows 10 IoT Core and Azure IoT.
- Windows 10 IoT Enterprise: This full-featured edition of Windows 10 IoT delivers enterprise manageability and locked-down capabilities for fixed-function smart devices. Companies such as City Beacon are building next-generation intelligent infrastructure based on Windows 10 IoT Enterprise. City Beacon has built a smart hub that can be located on city streets and venues; these multifunctional kiosks house a variety of IoT devices, connected to fast networks providing citizens with access to hyperlocal apps and services.
- Windows Server IoT 2019: Windows Server IoT is for demanding edge-computing workloads, for connecting, storing and analysing data from large databases of connected applications, networks and web services.
“Bringing NXP i.MX and Windows 10 IoT Core together provides device builders with additional choices for building secure, commercial-grade, connected devices,” said LeGrow.
Windows 10 IoT Core enables secure, power-optimised devices for the intelligent edge. Developers can quickly build new intelligent devices using the familiar Windows development environment and commercialise them using the Azure IoT platform.
“We are pleased to have reached the general availability milestone and continue to build our partnership with Microsoft,” says Jeff Miles, vice president at NXP. “With this addition, NXP customers now have access to industry-leading edge-computing, from i.MX 6 to i.MX 7 and i.MX 8M across Windows 10 IoT Core for advanced industrial and IoT applications, harnessing the full power of Azure in their intelligent edge products.”
More than 15 boards are available or in development from the partner ecosystem. As an example, Diagramics has built an NXP i.MX8-based product with Windows 10 IoT Core to build client devices, certified by Azure IoT Device Catalog, and enabling IoT plug-and-play registration of these devices with the Azure IoT Hub Device provisioning service.
During the keynote, LeGrow also shared details of an offering from Smart Tower based on the combination of Windows 10 IoT Core, NXP and Azure IoT. This is designed to improve reliability and reduce downtime for critical public infrastructure, such as cell towers or electricity distribution networks. Smart Tower designed it to take advantage of edge processing, creating a sensor module attached to the tower and a ground control unit that analyses the health of the tower in real time, supporting high latency or disconnected scenarios.
Additionally, it leverages Azure IoT Hub to manage and connect the devices and enables historical analysis across the whole network of towers being monitored.
“Smart Tower chose the combination of NXP, i.MX and Windows 10 IoT Core because of the extreme operating conditions the device would endure, and the familiar Windows IoT platform reduced the development costs and time to market,” said LeGrow.
Earlier this year, Microsoft enabled support for Azure IoT Edge on Windows 10 IoT Enterprise and Windows Server IoT 2019. This made it easier and seamless to integrate the IoT Edge runtime, making it possible to move machine learning algorithms and other similarly complex computing functions from the cloud to edge devices. Windows for IoT also supports Windows Machine Learning and Windows Vision Skills, which allows users to run AI algorithms developed in Azure on any Windows for IoT device.
The Windows ML container, available now as an insider preview, is designed for delivering AI-inferencing workloads built using the Windows ML API.
“Windows ML container gives businesses a fast and agile platform to build enterprise-grade IoT,” said LeGrow. “It combines the advanced features of Windows with the security of the Windows 10 IoT platform and the manageability of the Azure IoT Edge service.”
Windows ML containers provide hosted AI workloads with hardware acceleration and direct access to the host’s peripheral devices. Hosted workloads can use sensors or cameras connected using USB, I2C, SPI or GPIO. Hardware acceleration is available using any Direct X 12-capable GPU or Microsoft Compute Driver Model-compliant ASICs attached to the host.
By optimising the container specifically for AI workloads, the on-disc size of the container is approximately 350Mbyte, said to be significantly smaller than any other GPU-enabled inferencing container in the market. With this smaller size, users can manage Windows ML containers more easily and at lower cost.
“SQL Server has been available for embedded usage for many years, and we are excited to bring the latest SQL Server 2019 capabilities to embedded device builders,” said LeGrow. “We are introducing SQL Server IoT 2019, which is the binary equivalent to SQL Server 2019 and will be licensed through the OEM channel and is specifically designed for dedicated-use, server-class edge appliances running application software.”
Coupled with Windows Server IoT 2019, Microsoft SQL Server IoT 2019 provides a database platform for mission-critical data analytics and scalability, performance and security, and intelligence over all the user’s data. SQL Server IoT brings real-time analytics on operational data with in-memory technologies with concurrency and scale through persistent memory. It will be available before the end of the year.
Microsoft is also making investments in cloud connectivity and using the power of Windows to accelerate its customers’ journeys to the cloud and the intelligent edge, said LeGrow.
“The innovations we announced at Ceatec and the real-world IoT solutions being built by our customers and partners are just the start of the exciting path ahead for Windows for IoT,” he said.