Post to Twitter |
Comments (3) |
Views (2336) |
In case you have not been following some of the recent news on VDI and desktop virtualization, here is a recap.
Two contradictory stories
It all started about 2 weeks ago when Citrix and VMware shared the status of VDI business. There were two contradictory stories - Citrix announced some large purchases and implementations for Citrix XenDesktop and VMware announced that larger projects are being pushed out to 2010.
VDI is doomed for failure
Then came some reports on how VDI projects have recently failed. I like calling these not so successful VDI projects as 'first generation' architectures that are destined for failure. As the author notes that these projects start 'innocently' - with the goal of putting a desktop on top of the existing server virtualization platofrm. In other words 'We have done server virtualization - we can now do the same for the desktop'. Unfortunately, that is where the problem starts. Recently, Brian Madden published a story highlighting the problem. Two weeks ago, I heard customers sharing their frustrations and concerns about why desktop virtualization is different from server virtualization.
The Truth
I spoke with our CTO, Harry Labana, and Harry walked me through the top 5 reasons why in his experience VDI projects will fail if they are done with a presumption that desktop virtualization is similar to server virtualization.
The fact is that implementing Desktop virtualization or VDI is mre like implementing any other server based computing technology, such as Terminal Services. The VDI infrastructure that may be resident in the data center; however, it goes beyond the data center to the network, the end points and the user. So, unless the implementations are planned with network, device and end users in mind they are destined to fail. Good design, on the other hand, will lead to successful implementations. But that means doing the right due-diligence of products up front - including network, end points and end users. Or else you will fall into the similar trap as others are falling into.
Comments (3)
Aug 12
Stephen Greenberg says:
We have not seen any VDI projects fail, but we are seeing some functional issue...We have not seen any VDI projects fail, but we are seeing some functional issues when you combine the entire XenServer/PVS/XenDesktop stack. If projects are failing it is a lack of understanding/expectations, poor design and implementation.
The technnology is still a bit early stage and scaling may be an issue at this point, but most functionality works as advertised, it very analagous to the early days of Xen App.
One important point you raised- VDI is rehashing most of the same issues we dealt with in XenApp and TS. If you have the experience and battle scars from that genre you can be very successful with VDI
Steve Greenberg, CTP
Aug 12
Jeremy Saunders says:
Nice summary Sumit. I agree with Steve's comments. There are a lot of VDI POCs, ...Nice summary Sumit. I agree with Steve's comments. There are a lot of VDI POCs, but for the most part that's as far as they go, because more often than not they are being driven by server virtualisation guys, who push desktop virtualisation into a business where is doesn't fit, or the feature sets of the products cannot be aligned with use cases. Guys like Steve and I have been driving the "user experience" for 10+ years in the server based computing world. Server vitualisation guys don't understand these challenges and how to manage these expectations. They will learn, but sadly may damage the image of desktop virtualisation along the way simply because they are trying to flog a product. Yes, Desktop Virtualisation will succeed, but the products are still undercooked and have some way to go before they become mainstream and acceptable.
Cheers,
Jeremy.
Aug 12
Claudio Rodrigues says:
That was exactly the case with TS/XenApp years ago. I have seen so many projects...That was exactly the case with TS/XenApp years ago. I have seen so many projects that failed and guess what, TS/Citrix were always blamed for 'not working', 'being slow', 'not scaling' and so on. The main reason why these failed, like with pretty much every other technology out there, was the lack of understanding/knowledge about the underlying technologies being used.
Just because Terminal Services became part of the 2000 Server OS as a service that could be add/removed in a couple clicks that did not mean anyone with mouse/clicking skills could deploy it properly. In a way that is to be expected when you wrap several things under an ease-to-use wizard. People tend to think anyone can do it and do it properly. Then when it does not work they blame the technology/product.
Proper understanding on the technology is key to guarantee any project will succeed. Understanding clearly what needs to be achieved and how a particular technology can help, with its pros and cons, is mandatory.
And now that VDI is the buzz word pretty much anyone is now claiming to be a 'VDI expert/guru' just because they deployed Hyper-V/XenServer/ESXi at his buddies office. All this without ever installing WinNT 4.0 TSE and having to deal with profiles, printing, bandwidth constraints, RDP and so on...
I can already see what will happen down the road now that VDI is part of 2008 R2 and can be deployed following 'click here' Wizards...
Add Comment