What I've Learned About Desktop Virtualization

With applications running smoothly, the days of the fat old PC are gone

By , ITworld |  Virtualization, desktop virtualization, lessons learned

Two years ago, Kane Edupuganti came to Saint Vincent Catholic Medical Centers, a healthcare system that sprawls across New York City's five boroughs, to fix some woeful technology problems. As director of IT operations and communications, he's since revamped the system's wide area network infrastructure, virtualized the data center and is now rolling out virtual desktop infrastructure (VDI). He's not one for pussyfooting around a project -- he's pushed server virtualization to the limit, for example, running roughly 50 to 60 virtual machines on one physical server. "Otherwise, you're right back where you started from," he says. With that aggressive approach to virtualization, he brought the number of physical boxes down from about 300 to five, and did so in 90 days. Now he's zeroed in on the desktop. In this interview with contributing writer Beth Schultz, he reflects on ...

WHY THE DESKTOP: Everywhere we turned, we saw a old six-year-old IBM desktop running Windows 2000 Professional with only 256MGs of RAM. It literally took those machines five to 10 minutes to boot up. It was a very painful experience for users. We had already virtualized the servers, so we figured the next logical step would be the endpoints.

WHY VIRTUAL DESKTOPS: Initially people were like, 'Yeah, whatever, virtual desktop, we can do this in Citrix or we can do this on terminal servers.' And for most part, I think you can. But with the price coming down and the back-end farms that serve these virtual machines being able to give high-availability scenarios even to desktops and being able to give a full-blown desktop experience to the users, the industry is turning around a bit. There are still shops that can get away with Citrix farms, but that can get fairly expensive with licensing.

ESTABLISHING A VDI USE CASE: We were running our WAN on Verizon's MPLS network and were having some latency issues, especially between the main hospital and data center. One of biggest problem areas was the Emergency Department. Because it has such a high number of patients coming in and out and the application was running so slow because of network latency, doctors and nurses weren't able to complete their patient charts. People had gotten so fed up that they even had stopped using the software and were manually filling out charts. If charts weren't done, we couldn't bill, which affected the revenue flow. We knew we had to address this, with its domino effect.

CITRIX TO THE RESCUE, OR NOT: We have a Citrix farm in our data center that users can access when they go home. So we started letting Emergency Department folks launch Citrix sessions for their primary chart application, and we found the performance improvement was like day and night.

Join us:
Facebook

Twitter

Pinterest

Tumblr

LinkedIn

Google+

Answers - Powered by ITworld

ITworld Answers helps you solve problems and share expertise. Ask a question or take a crack at answering the new questions below.

Join us:
Facebook

Twitter

Pinterest

Tumblr

LinkedIn

Google+

Ask a Question
randomness