In a journey to understand the typical edge AI deployment journey, Microsoft spoke to over 500 customers and partners about their biggest barriers to entry. We heard about complexities when trying to scale, no end-to-end solution platforms, even security concerns. With these findings in mind, we released Azure Percept DK to public preview in Spring 2021. It's now available to customers in 16 markets around the world.
Azure Percept is a fully integrated platform for creating edge AI solutions. It's designed to allow developers to start prototyping solutions in minutes and includes hardware accelerators that work seamlessly with Azure AI and Azure IoT services. Indeed, since we launched Azure Percept DK, developers around the world have harnessed the dev kit to build, train, and deploy machine learning models for a variety of innovative uses.
Now, to keep advancing the capabilities and adoption of edge AI, our next area of investment for Azure Percept is to bring Azure Percept to Azure Stack HCI powered by NVIDIA T4 Tensor Core GPU. Combining NVIDIA's powerful computing platform with our edge devices and services will help to further accelerate the development and deployment of world-class AI solutions. This collaboration is highlighted in the Microsoft Ignite session, Automate your Operations with Edge AI, and it is just the beginning as we look to bring Azure Percept to other platforms and endpoints in the future.
Better together: Azure Percept on Azure Stack HCI
Azure Percept DK is fully integrated with Azure Machine Learning, Azure cloud-based security, and Azure IoT connectivity and device management services. Literally from out of the box, it works with these Azure services. Azure Stack HCI, meanwhile, provides a flexible infrastructure (HCI) operating system delivered as an Azure service that provides the latest security, performance, and feature updates. It enables customers to deploy and run Windows and Linux virtual machines (VMs) in data centers or at the edge using existing tools, processes, and skillsets.
Azure Stack HCI is a subscription service for hyperconverged infrastructure from Microsoft Azure. As an integrated virtualization stack, Azure Stack HCI is built on proven technologies that have been deployed at scale including Hyper-V, Storage Spaces Direct, and Azure-inspired SDN. Each Azure Stack HCI cluster consists of between two and 16 physical, validated servers, with the clustered servers sharing common configuration and resources. With Azure Stack HCI, you can run virtual machines, containers, and select Azure services on-premises with management, billing, and support through the Azure cloud. Azure hybrid services also can easily be added to Azure Stack HCI to enhance the cluster with capabilities such as cloud-based monitoring, site recovery, and VM backups, as well as providing a central view of all your Azure Stack HCI deployments in the Azure portal. Learn more about Azure Stack HCI.
Used together, Azure Percept and Azure Stack HCI will allow customers to deploy and manage edge AI in a scalable and secure way across their solutions. For example, they will enable customers who already have numerous cameras to connect them and leverage Azure Percept without having to invest in new cameras. The below demo shows how to easily create such a pilot solution in the Azure Percept Studio and deploy it to Azure Stack HCI. The solution uses the integrated Azure Video Analyzer and Azure Spatial Analysis, and it can handle multiple cameras and training models. Additionally, with Azure Percept and Azure Stack HCI, users can set up the solution for basic insights or alerts and stream video to Power BI for visualization of real-time data. Watch to check out the new Azure Percept Studio experience.
Combined technologies to boost edge AI innovation
Microsoft and NVIDIA have been long-term strategic partners with the vision to democratize AI on the edge for our customer and partner ecosystem. NVIDIA is sponsoring Microsoft Ignite at the highest level, check out their Ignite presence for more details.
Microsoft and NVIDIA have worked jointly on advancing the adoption of AI from high-performance computing in the cloud to the intelligent edge across a wide variety of industries. Together, Azure Stack HCI with the powerful NVIDIA GPU and Azure Percept will form a better-together hybrid solution that allows enterprises to connect their existing on-premises assets with cloud capabilities. Bringing this type of cloud innovation with Azure Percept to the edge can enable an end-to-end AI platform with NVIDIA.
Customers also can use NVIDIA DeepStream, designed for efficient run-time execution of complex AI applications, in combination with Azure Video Analyzer and Azure IoT Edge to optimize video analytics applications. And with NVIDIA GPU integration, customers can build and run their own AI and machine learning solutions at the edge, or handle graphics-intense virtual desktop infrastructure (VDI) workloads. As we continue to integrate the Microsoft and NVIDIA stacks, customers will be able to run their Azure Percept solutions on NVIDIA GPUs as well.
Teaming with global hardware and solution partners
In addition to collaborating with NVIDIA on Azure edge AI technology, Microsoft is working with global partners around the world who are building high-performance hardware and solutions that work with Azure Percept. These include Intel, which provides the dev kit's Movidius Myriad X vision processing unit. The Intel OpenVINO™ toolkit also is supported by the dev kit.
Through its partnership with Microsoft and NVIDIA, Lenovo has introduced the ThinkAgile MX1020 series, an edge-specific HCI appliance built with Microsoft Azure Stack that makes Edge HCI easier to deploy and utilize.
Solution builders working with Azure Percept include Neal Analytics, a world-leading system integrator focusing upon edge AI. The company's StockView AI application, powered by Azure Percept running on Azure Stack HCI, helps retailers reduce lost sales and improve customer experience by automatically detecting out-of-stock inventory using ordinary cameras.
PCL, a global construction management company, is extending its Job Site Insights application to add worker safety AI powered by Azure Percept running on Azure Stack HCI and Azure Spatial Analysis. This solution can proactively identify and alert employees and managers to dangerous situations at busy construction sites.
DataRobot AI at the edge capabilities—including enhanced employee safety, defect detection, and occupancy management—also are powered by Azure Percept running on Azure Stack HCI.
Try Azure Percept today
We're just at the beginning of what's possible, with tremendous momentum and energy from the community as they build innovative edge AI solutions across industries and use cases. And we have plans to expand to even more platforms and endpoints in the future. If you haven'tt tried Azure Percept and experienced what'spossible now, you can order a development kit, as well as learn more about the Azure Percept Studio of pre-built AI models.
You also can try a virtualized Azure Percept appliance integrated with our first-party Spatial Analysis AI service. This solution can be deployed on N-series VMs with NVIDIA GPUs in the Azure public cloud, enabling you to upload your own videos for AI processing. Explore what this technology can do when you request access to get started.