One of the major themes of GTC this year is how customers can use NVIDIA’s Kepler-based GRID GPUs to virtualize desktops, enabling an “anything, anywhere, anytime, on any device” usage model. This tech was one of the big announcements at GTC12, with demonstrations of GRID for gaming and graphically intense applications.

In this session, Luke Wignall and Jared Cowart, NVIDIA senior solution architects, take us through the history of virtualization. These guys aren’t the typical product marketing types; they’ve spent a lot of time working with customers to design and implement virtualized infrastructures.

The first part of the discussion covers virtualization from a historic standpoint, looking at the factors driving server virtualization and, now, desktop virtualization. Later on they talk about the different ways to virtualize desktops, making the point that there’s not any “One True Solution” in this space. Like everything else, there are trade-offs; every organization needs to figure out who their users are and what they need from their work platforms.

The session includes a case study that illustrates what customers really need to keep in mind as they go down the VDI road. There are definitely bumps in that road – plus a few places where the guardrails might be missing or broken.

For anyone who is considering a VDI initiative or needs some ammunition to justify a VDI project, watching this presentation is well worth the time. I’d also recommend it for anyone who is new to the wide world of all things virtual or could use a refresher course. You can see the session here.