VDI is often a pie-in-the-sky idea for many organizations. In certain use cases, it has vastly improved the efficiency of IT environments and drastically improved manageability. But, until recently, VDI had still been prohibitive to power users. Organizations like engineering or architecture firms that use 3D or computer-assisted design software have very performance and graphics-heavy requirements. Because of this, they typically struggled to adopt VDI without compromising user satisfaction and productivity.
Learn more about what it takes to adopt VDI in my previous blog series.
To address the video demand of these workloads, we now have NVIDIA GRID. With this technology, we are now able to include graphics processing on the VDI server, bypassing the hypervisor. GRID enables VDI for a larger audience by offloading graphics processing in virtualized environments to the GPU, making the VDI user experience comparable to using a dedicated PC.
NVIDIA GRID is scalable for simple office applications, like PowerPoint graphics, or graphic-rich design applications, like Adobe Creative Suite or AutoCAD. NVIDIA GRID does this through video profiles, which are sized 512 MB, 1 GB, 2 GB or 4 GB, and you can manage them in a single-pane-of-glass management console. Each virtual desktop has dedicated graphics memory, just like they would at their desk, but you can intelligently allocate how much. So, while users will always have the resources they need to launch and run their applications, you’re providing just what they need and not overbuying hardware.
Having graphic processing in VDI that can keep up opens all kinds of doors for organizations with disparate locations or remote employees. By centralizing the technological resources while providing the necessary performance, employees can access projects and collaborate from anywhere, any time. For example, engineers across the country can collaborate on the same AutoCAD design, which opens up a whole new world of resource sharing.
NVIDIA GRID is releasing second iteration of the product on September 15. NVIDIA GRID 2.0 will have double the user density and double the application performance. It will also support blade servers, not just rack servers.