CIOs: VDI? Maybe, just maybe, it really will be different this time... Sometimes technology visions bubble along under the radar, quite often scorned then, following a change of name or emphasis, ubiquitous. Utility computing became the cloud, and we know how that’s taken off. Groupware became collaboration. Thin clients have morphed into VDI, and… Well, seems like a good time to get a bunch of …
I explored VDI about 6 months ago. We're a CAD house, and the guy who sells VDI units for a living literally told me 'do not bother'. It's not a good solution, and even if the CIO is feeling all happy-happy joy-joy about it, the user base is not. You don't really save much in the way of money from it (we calculated we'd need a $40k host for every 10 users or so, plus a complete upgrade of the network infrastructure to cope with bandwidth demand), the user experience is pretty awful, IT staff hate them... it's a fool's errand, just like it was ten years ago. And twenty years ago. And thirty years ago.
As a VDI user they have pros and cons like most things. Mostly I like the portability: it's easy to move desks and work remotely, without having to lug a laptop around.
I think that most of the downside for me is a result of the company doubling down on "virtual": not content with setting up VDIs, we have to run as much as possible in Citrix. IE in Citrix in a VDI, anyone? This also leads to usability frustrations such as not being able to drag a file from Windows Explorer (in the VDI) into an Outlook message (in Citrix).
I was looking at the feasibiilty of a reasonable size (5000 user) VDI rollout a few years ago, and video performance was the killer then too. Either you send the compressed video over the network and decompress it on the client, or decompress it on the VDI server and sent it over the network uncompressed. The latter kills the network and the former requires such powerful thin clients that they are no cheaper than a full blown desktop, and are often Windows based so there is no saving on licences or support either.
It was enough of a problem that even though the client did not have much of a need for video (mainly training presentations), that they ended up going with full desktops anyway,
That may have been true, but it seems hard to believe it is still true. You used to need a fairly beefy CPU to do this, and discrete graphics, but no more. A simple ARM SoC that includes h.264/h.265 decoding capability and GPU that easily handles getting it to the screen costs what $25?
If a $150 Android tablet can handle full screen HD video, why can't a thin VDI client - assuming it was designed with the appropriate hardware? I'm not familiar with the market, but if a solution doesn't exist there's a billion dollar opportunity waiting to be exploited.
Aren't we trying to solve deployment and management problems mostly?
This could be solved very well with desktops. Just generate an image every morning, pump it to the desktop, possibly into a virtual shell layerd on top of the desktop changing hardware, voilà the user is set up for the day. In the evening pull out any configuration, centralize saved files, clean up etc, then blow the desktop away.
If necessary, user calls up Mission Control, which generates a new image from a menu as directed by user requests.
It would be nice. Can't be done with Windows Endpoint Messmaker though.
It has its uses though. A few years ago I had someone very senior complain that the scientists working for him couldn't process their data at home. It turns out that they were trying to open and save the files (up to about a Gig) over domestic ADSL; I put the software on a server and they logged on from home and were very happy and some of them even found it was quicker in the office too.
"If they decide that now you really can’t tell the difference..."
It's like asking anything that starts with, "If wishes were horses..." It doesn't matter what the question is if it starts with an impossible premise. Users. Can. Tell.
Heck, I haven't even seen a company switch from PBX to VoIP without bringing their network to its knees and requiring a massive expensive overhaul to defeat the purpose of their "cost-saving" effort. And that's just audio, highly compressed. Now multiply that by EVERYTHING. It has never gone smoothly, and it will never go smoothly, and anyone who tells you otherwise is probably trying to sell you something. Do not buy.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
