Azure Stack Edge Pro with GPU is an AI-enabled edge computing device with network data transfer capabilities. This article provides you an overview of the Azure Stack Edge Pro solution, benefits, key capabilities, and scenarios where you can deploy this device. The article also explains the pricing model for your device.
Azure Stack Edge Pro with GPU is a Hardware-as-a-Service solution. Microsoft ships you a cloud-managed device that acts as a network storage gateway. A built-in Graphical Processing Unit (GPU) enables accelerated AI-inferencing.
Use cases
Here are the various scenarios where Azure Stack Edge Pro GPU can be used for rapid Machine Learning (ML) inferencing at the edge and preprocessing data before sending it to Azure.
Inference with Azure Machine Learning - With Azure Stack Edge Pro GPU, you can run ML models to get quick results that can be acted on before the data is sent to the cloud. The full data set can optionally be transferred to continue to retrain and improve your ML models. For more information, see how to use
Deploy Azure Machine Learning hardware accelerated models on Azure Stack Edge Pro GPU.
Preprocess data - Transform data before sending it to Azure via compute options such as containerized workloads and Virtual Machines to create a more actionable dataset. Preprocessing can be used to:
Aggregate data.
Modify data, for example to remove personal data.
Subset data to optimize storage and bandwidth, or for further analysis.
Analyze and react to IoT Events.
Transfer data over network to Azure - Use Azure Stack Edge Pro GPU to easily and quickly transfer data to Azure to enable further compute and analytics or for archival purposes.
Key capabilities
Azure Stack Edge Pro GPU has the following capabilities:
Direct data access from Azure Storage Blobs and Azure Files using cloud APIs for additional data processing in the cloud. Local cache on the device is used for fast access of most recently used files.
Cloud-managed
Device and service are managed via the Azure portal.
Support for standard Server Message Block (SMB), Network File System (NFS), and Representational State Transfer (REST) protocols for data ingestion. For more information on supported versions, see Azure Stack Edge Pro GPU system requirements.
Use the Marketplace experience from Azure Network Function Manager to rapidly deploy network functions such as mobile packet core, SD-WAN edge, and VPN services to an Azure Stack Edge device running in your on-premises environment. For more information, see What is Azure Network Function Manager? (Preview).
Components
The Azure Stack Edge Pro GPU solution includes the Azure Stack Edge resource, Azure Stack Edge Pro GPU physical device, and a local web UI.
Azure Stack Edge Pro GPU physical device - A 1U rack-mounted server supplied by Microsoft that can be configured to send data to Azure.
To procure a device, go to the Azure Edge Hardware Center and place an order. Azure Edge Hardware Center service lets you choose from a variety of Azure Stack Edge SKUs as per your business need. You can order multiple units of a device type, ship multiple devices to different locations, save addresses for future orders, and also track the status of your orders.
Once the order is delivered, you can configure your device and create an Azure Stack Edge resource to manage the device.
Azure Stack Edge resource – A resource in the Azure portal that lets you manage an Azure Stack Edge Pro GPU device from a web interface that you can access from different geographical locations. Use the Azure Stack Edge resource to create and manage resources, view, and manage devices and alerts, and manage shares.
Azure Stack Edge Pro GPU local web UI - A browser-based local user interface on your Azure Stack Edge Pro GPU device primarily intended for the initial configuration of the device. Use the local web UI also to run diagnostics, shut down and restart the Azure Stack Edge Pro GPU device, view copy logs, and contact Microsoft Support to file a service request.
The local web UI on the device currently supports the following languages with their corresponding language codes:
Azure Stack Edge Pro GPU physical device, Azure resource, and target storage account to which you transfer data don’t all have to be in the same region.
Resource availability - For this release, the resource is available in East US, West EU, and South East Asia regions.
Device availability - For a list of all the countries/regions where the Azure Stack Edge Pro GPU device is available, go to Availability section in the Azure Stack Edge Pro tab for Azure Stack Edge Pro GPU pricing.
Destination Storage accounts - The storage accounts that store the data are available in all Azure regions. For best performance, the regions where the storage accounts store Azure Stack Edge Pro GPU data should be close to the device location. A storage account located far from the device results in long latencies and slower performance.
Azure Stack Edge service is a non-regional service. For more information, see Regions and Availability Zones in Azure. Azure Stack Edge service doesn’t have dependency on a specific Azure region, making it resilient to zone-wide outages and region-wide outages.
For a discussion of considerations for choosing a region for the Azure Stack Edge service, device, and data storage, see Choosing a region for Azure Stack Edge.
Device availability for customers
Azure Stack Edge devices are only available to customers with the following use cases:
Validated partner workloads.
Large scale deployments.
Validated partner workloads
Devices are available for customers who utilize Azure Stack Edge as an infrastructure platform to validate their solution. Partner workloads may not be supported on all devices. For details, consult your Azure Stack Edge partner.
Approved, validated partner workloads:
Azure Government – We continue to facilitate and support Azure Government customers and the Defense Industrial Base.
Large scale deployments
Azure Stack Edge is optimized for edge scenarios that typically involve hundreds or thousands of devices across a high number of locations. Our hardware as a service solution is only available for deployments of at least 100 nodes.
Next steps for existing customers with qualified workloads
Submit an Azure Stack Edge sign-up form via Azure portal.
Next steps for alternate edge solutions
Use the following links to get information about alternate infrastructure solutions:
The users are charged a monthly, recurring subscription fee for an Azure Stack Edge device. In addition, there’s a onetime fee for shipping. There’s no on-premises software license for the device although guest virtual machine (VMs) may require their own licenses under Bring Your Own License (BYOL).
Currency conversion and discounts are handled centrally by the Azure Commerce billing platform, and you get one unified, itemized bill at the end of each month.
Standard storage rates and transaction fees are charged separately as applicable. Monthly subscription fee billing starts after delivery whether the appliance is activated or not.
The billing happens against the order resource. If you activate the device against a different resource, the order and billing details move to the new resource.
Azure HPC is a purpose-built cloud capability for HPC & AI workload, using leading-edge processors and HPC-class InfiniBand interconnect, to deliver the best application performance, scalability, and value. Azure HPC enables users to unlock innovation, productivity, and business agility, through a highly available range of HPC & AI technologies that can be dynamically allocated as your business and technical needs change. This learning path is a series of modules that help you get started on Azure HPC - you