VMware View versus Citrix XenDesktop performance: who cares?

Alessandro did a great job capturing the current "Ours is better! No, ours is better" fight between VMware and Citrix.

Alessandro did a great job capturing the current "Ours is better! No, ours is better" fight between VMware and Citrix. (If you haven't read about it yet, the short version is that these two companies are fighting about whose hypervisor can support more VDI users.) They're going back-and-forth with test after test and insult after insult. At some point you have to step back and chuckle. Seriously? Arguing about which hypervisor can support more VDI users? This is like arguing whether a 1999 Ford Contour has better cupholders than a 2001 Chevy Prizm.

In other words, yeah, I get the fact that VDI is getting better and better in terms of user denisity. But let's face it. If you really cared that much about user density, you'd be using Terminal Server. :)

The only people using VDI today are people who have a tactical business need that does not allow them to use a Terminal Server-based solution.

Of course one could make the argument that "Sure, we don't choose VDI because we want good performance. But if we *must* go with VDI, then why not get the best performance possible?"

Good point. Except that the performance differences between the two products are more-or-less marginal (speaking from the viewpoint of an objective third-party who knows that anyone can make any results look great in their favor).

The bigger point is that VDI is such a huge undertaking for companies that the factors that weigh into the decisions about which platform will win out have to do with managebility, end user experience, application compatibility, etc. I don't think anyone is actually making buying decisions based on which of the two VDI platforms can support more users.

And by the time VDI gets really popular next year, we'll have new hypervisors all-around, so we'll just have to start this new game from scratch.

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

It’s still an important factor when defining a Bill or Materials for the given solution.

Of course it will become more important as the cost of VDI comes down and the take up increases.


I would absolutely agree that if I'm looking at VDI solutions today (and we're loosely investigating it), performance is pretty far down on our list, behind ease of use, ease of management, compatibility with existing vendors/apps, price, and a few other things.


Brian, Not too sure if VDI is as big an undertaking as you say.

Organisations that already have a server virtualisation infrastructure (of any flavour) may also have a TS/Citrix infrastructure and should be realising cost & efficiency savings RIGHT NOW.

I just see VDI as bringing these two worlds together to add more value where it makes sense.

Yep, my opinion is that a mix of VDI and TS/Citrix ie. Streaming/presenting apps into a VDI as being the solution that has the most potential... you get the benefits of both worlds and can harness a lot of already sunken cost! Your most recent article about admin authorities also comes into play here with the user having the freedoms to install other required apps that don't make sense to stream/present for a particular reason.

And you have everything in the data centre to reduce admin overhead.

Making the VDI project try to fit every one of your users is where it's a huge undertaking and usually where the VDI projects fail.


I agree that performance is not that critical a differentiator now. Whatever happens with the product, performance will increase next year. So, if it's a minor problem this year, it will be forgotten next year. I think the primary issue is deployment. Getting the system in place, getting the users configured correctly. And, most importantly, getting the applications working for the users - the way they need them.


I don't think user density is even on my radar because I know the top contenders are going to be relatively close. I would see my top three criteria as being...

1) User Experience (Perceived Performance)

2) Manageability. (Routine Management Functions)

3) Cost (Do I necessarily need an expensive hypervisor?)


My short list:

1) User experience (perceived performance).  If user's perceive a slowness, then eventually it bubbles up management on the business side and IT's objectives are slaughtered.  We can't do things to make our job's easier than make their jobs hellish.

2) Capability for business as usual (this means that the VDI platform can support all the business requirements for usage of multimedia, peripheral compatibility, disability-compliance, etc)

3) Security/Control of data - I spend most of my time in the financial services market and control / encryption of data is a HUGE issue).  VDI does change things slightly.

4) Be functional for WAN users (despite the fact that VMware thinks it's a niche).  I've got 3k folks in India that abhor RDP (and if I was sitting on the other side of 270ms, I would too!)  When you've got 3k people, I'd hardly call it niche, don't you think?

Repeat these ad nauseum because if they don't happen, then VDI doesn't happen.  At least not succesfully anyway.