This past Thursday, I sat down with Leostream CEO Karen Gondoly to discuss hosting graphics-rich applications in Microsoft® Azure® Clouds in our latest webinar – “Graphics-Rich Applications at Cloud Scale with Azure and Leostream.”
Ever since the announcement that Microsoft could make licensing available for Windows 10 in the public cloud, the tech world has been buzzing with talk of Azure as a viable platform for the long awaited “VDI in the cloud”. For 10 years, AWS has dominated the space, but heavy restrictions on Microsoft licensing have stuck cloud-hosted desktop users with a server OS and confusing pricing models. Instead of the cloud desktop revolution we were promised, a lackluster user experience has slowed transition to VDI in the Cloud down to a trickle.
Azure is a game changer. With the potential to license Windows 10, we finally have a cloud-hosted desktop that looks and feels just like desktops in an on-prem environment. What’s more, Azure offers the GPUs to host graphics-intense applications at cloud scale without sacrificing performance. This is exciting news for everyone, from organizations with large on-prem deployments that need to burst into the cloud during times of high demand, down to the SMB trying to save on hardware costs by setting up their first ever hosted desktop environment.
In yesterday’s webinar, Karen discussed the GPU power of the Azure N-Series VM family. Azure NC-based instances are powered by NVIDIA Tesla® K80 GPUs and provide compute power for high-performance computing jobs. Additionally, you have the option to utilize Remote Direct Memory Access over InfiniBand to scale jobs across multiple instances. InfiniBand provides nearly bare-metal performance even when scaling out across hundreds of machines.
On the flip side, Azure NV-based instances are powered by NVIDIA Tesla M60 GPUs and provide NVIDIA GRID capabilities. You can use these new instances to run hardware-accelerated workstation applications, meaning your graphics-intense applications. These instances support applications that use both DirectX and OpenGL.
As I’m sure you can imagine, these are not inexpensive instances, but the nice thing about Azure is that you pay hourly for compute. If you deallocate the instance when it’s not needed, you can lower your bottom line.
Karen also discussed how choosing the right display protocol makes a world of difference when it comes to graphics-intense applications in the Cloud. GPUs won’t matter a bit if you aren’t using a high-performance display protocol to transmit and render a pixel-perfect display on the users’ client device. In the webinar presentation, Karen discussed in detail the various options for high-performance display rendering, including Teradici Cloud Access Software and HP RGS.
All of this is well and good, but one major drawback to VDI in the Cloud remains – desktop management can be cumbersome and requires careful monitoring to avoid any unexpected compute costs. Here is where Leostream can help. With our new support for Azure, the Leostream Connection Broker can handle all the heavy lifting in your environment and provide customizable management options not found using the web interface alone. Features such as pooling, automatic provisioning/deallocation, AD authentication, scheduling and more allow you to leverage the full potential of Azure for your VDI environment. Additionally, with support for most major protocols, you’ll never sacrifice end-user experience whether you are on-prem or in the cloud. To create a true VDI environment in Azure, you need a true VDI connection broker and no one has more experience managing on-prem, cloud, and hybrid environments than Leostream.
To learn more about the Leostream Connection Broker for managing desktops in Azure, please contact [email protected]
