Metro Health is an integrated healthcare system that offers expert, award-winning care from 15 locations in Grand Rapids, Michigan and surrounding areas. Their facilities include Metro Health Hospital, a 208-bed facility that offers general acute care and osteopathic teaching, and physician/outpatient offices located throughout West Michigan. Services offered include nationally-recognized heart & vascular services, various inpatient and outpatient specialty services, a cancer center, and a wellness and community education program. Metro Health also partners with other healthcare providers, a medical education program, and more.
At present, Metro Health is running approximately 50-75 users per server and is testing to see how much of the load NVIDIA GRID is offloading from the CPUs to determine how many users they can support on a per-server basis, especially as the Radiology, Cardiology, and other graphics-intensive departments begin migrating to GRID. The current VDI implementation has been proven through internal and external analysis to be delivering a time savings of 30 minutes per day to each doctor and 50 minutes per day to nurses and other professionals at Metro Health through fast roaming access alone.
Doctors, nurses, and other professional users routinely roam the hospital and between locations as part of carrying out their duties. These users require fast, reliable access to their applications and data from anywhere within the Metro Health network. Roaming users had to resort to sharing credentials and logging into other users’ workstations to access their applications. This led to an in-house response of implementing a Virtual Desktop Infrastructure (VDI) that would allow a roaming user to access their own desktop with their own credentials from any endpoint within 10 seconds. The in-house solution worked but did not offer a full virtual desktop because of the inherent limitations of both the browser-based interface and the Remote Desktop Protocol (RDP).
From there, the solution was to develop the first Epic View client that allows the same user to maintain their session and keep working from endpoint to endpoint. Despite these advances, none of the available solutions could offload graphics processing from the CPU. Even basic work, such as playing back videos in a browser or viewing still or moving images within Epic requires GPU calls that were being handled by the server CPUs, which slowed down under heavy usage conditions. The end goal of replacing CPU usage with a GPU when appropriate remained elusive.
OST worked closely with HP and NVIDIA to put together a new VDI solution spec to address Metro Helath’s virtual desktop performance issues, which included a requirement to improve overall graphics performance. Metro Health opted to use HP ProLiant servers and when Metro first adopted the HP platform OST helped deliver the DL380P Gen8 servers with NVIDIA GRID™ K1 cards. Thanks to the teamwork between OST, HP, & NVIDIA, Metro Health was the first customer to receive NVIDIA GRID card pre-installed in HP servers.
Metro Health is currently evaluating how well NVIDIA GRID technology satisfies their goal of offloading GPU processing from the CPU for every application that makes GPU calls, and will be conducting testing to determine both how many requests are being made and how NVIDIA GRID is responding to those requests.
So far, the 95 users participating in the NVIDIA GRID evaluation have responded very positively, with only two significant issues: First, a few high-end users are currently using three or more high-definition monitors and will not be happy at having to go down to two monitors, an issue that requires upgrading to vSphere 6.0 to solve. The other main issue is that one of the imaging applications used by the Radiology department requires RDP, which is not supported by NVIDIA GRID technology. Currently, RDP is being forcibly disabled for these users; however, the ultimate goal is to have every specialty that relies on imaging using NVIDIA GRID technology to deliver the graphics.