News

Microsoft Releases Azure Data Box Edge and Gateway

• By Kurt Mackie
• March 27, 2019

Azure Data Box Edge and Azure Data Box Gateway are now generally available, Microsoft announced this week.

In preview since last September, Azure Data Box Edge is a "1U rack-mountable appliance" made by Microsoft that can be used to help with compute and storage tasks before transferring data to Azure datacenters.

Back then, Microsoft also had announced the general availability of Azure Data Box, which is its rentable storage device for physically transferring up to 100TB of data to Microsoft for installation on Azure infrastructure. Microsoft also offers a 1PB Azure Data Box Heavy storage device, which is still at the preview stage. There's also an 8TB Azure Data Box Disk storage device that's currently available for lighter Azure data transfers.

Azure Data Box Edge is a more specialized appliance that can be "racked alongside your existing enterprise hardware or live in non-traditional environments from factory floors to retail aisles," according to Microsoft's announcement. It's designed to support a few edge computing and storage tasks.

For instance, Azure Data Box Edge supports running applications in containers. It supports local caching of frequently used data, while automatically transferring other data to the Azure Storage service. Azure Data Box Edge also lets users conduct machine learning operations locally via an included Intel Arria 10 field programmable gate array, although this machine learning capability is still at the preview stage.

Organizations can get the Azure Data Box Edge appliance by ordering it from Microsoft via the Azure Portal, a browser-based management solution. The costs to use Azure Data Box Edge are described at this page.

Machine Learning
On the machine learning front, Microsoft explained in another announcement that most of the data used in applications by organizations is "used at the edge," such as collecting images, which might be done by retail stores, factories or hospitals. Consequently, the ability to conduct machine learning processes on the Azure Data Box Edge appliance by these organizations can result in "lower latency and savings on bandwidth costs," Microsoft contended.

Organizations can use Azure Data Box Edge to do things like "train a TensorFlow model for image classification scenarios, containerize the model in a Docker container, and then deploy the container to a Data Box Edge device with Azure IoT Hub," according to Microsoft.

Microsoft also offers an Azure Machine Learning service. It's a separate service, though, that's different from the machine learning capabilities of the Azure Data Box Edge appliance.

Storage
The network storage capability of the Azure Data Box Edge appliance gets carried out using the Azure Data Box Gateway. The Azure Data Box Gateway is a software "virtual appliance" component in the Azure Data Box Edge appliance or it's available separately. Azure Data Box Gateway is also now generally available, according to Microsoft's announcement.

The data that goes to the Azure Data Box Gateway "will automatically upload to your Azure Storage account, supporting Block Blob, Page Blob or Azure Files," Microsoft clarified. The network bandwidth used for these data transfers gets optimized by Microsoft.

In another announcement on Tuesday, Microsoft announced that Blob storage on the Azure Data Box product has reached general availability status. The Blob storage support lets users copy data using REST-based APIs, making Azure Data Box "appear like an Azure storage account."

Blob storage for Azure Data Box can be used to handle Big Data workloads, such as moving "large HDFS stores to Azure as part of an Apache Hadoop migration," Microsoft indicated. The Blob storage capability of Azure Data Box is also being tapped by Microsoft's partners, such as "Veeam, Rubrik and DefendX," to help customers move their data to Azure, according to Microsoft.