• Home
• Technical Resources
  • Windows Registry Information
• Blogs
  • Gareth Kitson
  • Jon Wallace
  • Eric S. Perkins
  • Dave Kawula
  • Michael Mills
  • Guy Leech
  • All Authors

The Beginning of the End for VMware’s VDI Push?

Jon Wallace

I am the creator of insidetheregistry.com but that's as far as it goes - the site is strictly independent and vendor agnostic and anyone is invited to contribute and post views...

I was awarded the Microsoft MVP (Windows Desktop Experience) in 2009 and have presented at various trade shows including Citrix Synergy and VMWorld. I am a frequent contributor to www.dabcc.com.

In my spare time I develop this little baby (the website), and spend time with my lovely wife Vicki and daughter Mia.

Originally for Warrington in the United Kingdom I now live in Coral Springs in Florida (near Fort Lauderdale) and although gone through Katrina and Wilma have never looked back.

As a side-note I do have a background as a professional performing magician, am a member of the Magic Circle (London) and the Magic Castle (Los Angeles). I find IT very similar, smoke and mirrors.

I hope you enjoy what I've got to say and if not, let me know.

In my opinion there have been some rather interesting developments over the past few weeks with relation to the VDI space, none more than VMware’s announcements that View is to ship late and more so, without its profile management technology.
 
I always like to start my posts with a bit of digressing so here goes…
 
I used to do Karate and in fact was quite successful in the sport.  I took part in many competitions, some which I won and some which I lost but there was one thing that was always constant, the way I reacted when behind.
 
During a “kumite” which is the word used for fight, you typically have 2 minutes in which to score 6 points.  At the end of the match, if neither competitor has scored 6 points, the one with the most points scored wins.  Obviously as you can imagine, when the 30-second remaining bell sounds and you are the competitor that is currently 3 points behind, things start to move into overdrive.
 
What I know from experience is that when I was behind in a match and that 30-second remaining bell went off I calculated the match totally different.  Emotion begins to take over which forces uncalculated risks resulting in mistakes and ultimately defeat.  I believe I’m safe in saying that it was rare for me to win a match where I was 3 points behind at 30-seconds to go and in fact, I typically lost by 4 or 5 points.
 
So what has this got to do with VMware?
 
Within the Desktop Virtualization space, I believe there are 2 players, Citrix and VMware.  Sure Microsoft and a few others like to think that they are in the space but the truth is that none are investing the efforts that these organizations are.  It is also my opinion that VMware are substantially behind Citrix with their technology and don’t have the grasp on the desktop that their competitor does.
 
Citrix and VMware originate from totally different areas of technology; where VMware comes from the pure virtualization space, Citrix have not only substantial experience in the user desktop space, but more importantly, the "remote" user desktop space.  One can argue that Citrix’s “special sauce” was always the ICA protocol but I believe those in the business will all agree that Citrix’s offerings are much more than just communications now.
 
Like myself with Karate, Citrix has been “practicing” delivering desktops to users for a long time and pretty much have it right.  Whether it be a desktop delivered by XenApp from a shared resource, or a desktop delivered by XenDesktop the experience is pretty good and has evolved from many years of “trying”.
 
The alternate solutions in my option are still trying to get things right, pulling things together and trying to iron out the kinks in order to deliver the expected experience.  Where as Citrix has all of the various pieces of the puzzle in house already, others are trying to fix up and change what they can to make things work.
 
To me there is no question that when it comes to pure virtualization VMware knows its stuff.  Sure Hyper-V and Xen are excellent pieces of technology but VMware ESX has always been ahead of its class and in most parts, VMware can claim that they created the virtualization space.  But the hypervisor is a commodity and the money is in the user hence why everyone is keen to get in the space.
 
What I think we are seeing here is an organization trying to play from behind but more so trying to pull back a substantial amount of ground and emotion is kicking in.  Products are being rushed or released incomplete and over-ambitious plans to try and mirror the functionality of others is causing problems.
 
VMware will no doubt be criticized for the lack of inclusion of the RTO functionality, but why?  Because it was promised!  If RTO were never acquired, would customers have really missed it?  Sure Citrix would have had the one extra tick in the box but I’m sure VMware could have found other places to pull back points.
 
I guess what I’m trying to say is that sometimes, you have to face fact that there is a winner and that there simply isn’t the time to catch them.  For me, once I’d gone to 4 points behind in my match, I was pretty much done and I went into reserve so I didn’t get hurt and put out of my next matches.
 
One does have to wonder if VMware will switch gears and move to something  that they are clear leaders in.  The hypervisor may be a commodity but there is still many many things missing from it just waiting to be created.

Penny for your thoughts?

Comments

There sure is a big difference in the work VMware is putting in vSphere and View. Just got the vSphere 4.1 details, and they will stay ahead - where as View - well - they are even behind Microsoft ;-)
I think that one problem is that there is not that big a market - with our customers we have found out that the is no business case in most cases.
RDS/TS, web-apps and well managed rich Windows clients does a fine job covering most!
Best regards
Michael

I think there's another way of looking at this. There's two ways of driving a product release and development. You can either works on a Time Release Schedule or Feature Release Schedule. With the feature release schedule you promise that a new product will be available with X, Y, Z feature which you hope to deliver at N time. Of course, this leads to features not being ready, and the release date being pushed back.

A time release schedule - tells customers that a product will arrive at N time, and the company plus out all the stops to add the features in it roadmap. If the features aren't ready by the date. They get pulled, or in the case of VMware sometimes marked as "experimental".

For most customers a time release schedule is best. In fact being able to plan the process based on a reliable release time is MORE important than a specific sub-feature which could be added at some later date with in the product release in the form an update OR the removal of the "experimental" flag once a feature has been properly QA'd.

The curve ball (to take your sporting reference here) is when a couple of serious bugs are discovered very close to the original RC or GA date. In this case the vendor has two choices - release a product with known serious bugs, and take a serious negative hit from the market place (or people like me and you) or take the time release schedule approach - and push back the release date, and pull a feature. I think you'll find this is what happened with View 4.5. That serious bugs associated with Windows 7 we discovered, and VMware had to take this option. Whilst this isn't an ideal situation its better than rolling out a product on time crammed full of features - which don't work...

So what are to make of this. I think the problems that VMware have with View have little to do with Citrix or the history of the company, but everything to do with the nitty-gritty of software development which most customers and analyst are unaware of. In the long-term this delay will be insignificant - and the current lack of the profile feature may not be the show-stoppper that some folks imagine.

Mike Laverick
RTFM Education

I don't think it matters. VDI adoption will not be driven by VMware or Citrix because neither one of them has figured out how to reduce the cost of storage.

2 players? I think Quest would put a decent punt in against that - their vWorkspace product gives XD a better run for its money than View


I'd agree with Mike - the issues that VMware have aren't that important and are part and parcel of s/w development: its not the first time a product's been delayed or features missed from a release.


Ruben Spruijt had a good quote in his VDI smackdown "VDI fits for every customer but not for every desktop".

I think VMWare's issue is that it doesn't look to offer a desktop strategy: the play seems to be 'we'll virtualise your desktops'. When you dig into it its 'we'll virtualise your desktops on the LAN - and you'll need a good amount of tools and management to roll out an enterprise structure'

I an issue VMWare has to address is can they continue to put development time into a solution that only delivers VDI - and even then requires a number of 3rd party tools to manage it effectively? Because, as Michael says - the VDI market is (I'd suggest) not, or going to be, as huge as perhaps they thought it may be.

Andy Wood
Gilwood CS

I believe the VDI market will be larger than anyone is predicting .... eventually. However the form factor that it eventually takes, and the designs of the future state are not even on white boards at present. (i put a lot of technologies into my definition of VDI)

As i have said, the desktop is not, and maybe never was a NOUN we Buy ... but rather a mythical VERB we have spent money crafting so users can DO things.

To call the match now is kind of like saying that one team is going to win because their uniforms are cooler than the other team. A bit silly.

T.Rex

It’s an interesting analogy, and yes the "missing" RTO software has been a kick in the pants for some customer or partners that were "promised" it but I am not sure it is on the mark. Is VMware swinging fast and furious trying to make up the points it is behind? maybe.

But the analogy you use REQUIRES that there be a time limit. Having been in some of these same tournaments and fights I would say they may FEEL like they are in the last 30 seconds, but this is due to individuals under pressure to make numbers and that never goes away.

I think it is more appropriate to say this is the 2nd fight in a double elimination tournament. They just lost one because they started out behind in the desktop space and are trying to win one to keep in the tournament.

Obviously taking some losses early in the 2nd fight is not good. You would love to start out up 3-0 after a couple of genius combinations. But even if you are down 0-2 in the second fight it is time to SLOW DOWN. Make your technique right. Concentrate on your center and what you are good it. Then go back at them.

Anyway, enough with the fight analogy. CTXS has great technology and a deep understanding of desktops. I think VMware can have a decent product and be a foundation for a decent solution. They just need to focus on what they do well, target the market appropriately and look for something they can do better than CTXS in the Desktop space. Being even won’t cut it, they need an edge.

Ron O (Unidesk)

I agree with Ron O that the analogy is not quite applicable. I disagree with him about VMware catching up. VMware arguably created the virtual desktop space (Virtual Bridges claims differently), and continues to dominate the market because customers increasingly see virtual desktops as an extension of the vDC rather than as a silo. This change is being reflected in organizational struggles to redefine responsibilities between the desktop and systems teams. Indeed, the very definition of a desktop is changing as devices such as the iPad and upcoming Cisco Cius begin to alter the way in which employees work.

Guest just below me! (Ron O)

I think in the General desktop space VMware IS Catching up. Yes they were some of the first folks to run desktop OS's as VMs in the datacenter (Before it was even called VDI). But that matters very little.

In the desktop space (from a partner and buyer perspective) VMware is the "new guy" or FNG. This REQUIRES that they prove themselves. Citrix has a LONG history in desktops. Most of it good, some... not so much (like when their sales guys actively went out and slammed softricity because it could be used to deploy apps to desktops... but I digress). Anyway Citrix knows the buyers. Has a holistic story. Knows how to tell it. And has a proven desktop track record they can point it.

This means it is not enough for VMware to have a feature checklist that equals CTXS. They have to have an edge. They need to show their buyer that they should buy VMW instead of CTX for reason X!

Now ‘reason X’ can be a feature or set that Citrix doesn’t have (but it better be compelling) or it can be that using VMware instead of Citrix is Y% cheaper. But again that better be significant.

You don’t win a customer by equaling the incumbent. When all else is equal, I’m sticking with what I know. If you want to beat the incumbent you have to BEAT them at something.

That is the key here. They (VMW) have cast their net wide looking for that something. They initially thought they would coast into the desktop world and that didn’t happen. Then they scatter shot this, that and the other, trying to build a list of features and functionality... it always looks like catch-up to the market and to some a lack of focus.

I would rather be great at a couple of things (maybe things no one else can do), than be ok at a lot of things and trying to unseat entrenched technologies.

But hey, that’s just me... and what the hell do I know.

I also agree with Ron that the analogy is not quite applicapable. I am skeptical as to whether VMware WANTS to catch up with Citrix and others.

I wrote a blog entry on DABCC months ago http://www.dabcc.com/article.aspx?id=12798 that rhymes a lot with what you said. I think VMware needs to first decide whther it wants to be in the desktop irtualization market.

View is a datacenter only type virtualization that is made up of a collection of OEMs. That is a big problem, until they bought RTO, they had PCoIP, RTO, ThinPrint. They bought ThinApp etc... all very loosely integrated. They went cold on CVP.

It feels they are not sure they want this. I agree with the author of the blog, from a hypervisor perspective, VMware is king and will contine to be so. I am seeing a lot of implementations though with vSphere backend, XD front end.

If VMware is serious about desktop virtualization, the first thing they should do is buy all the OEM components that make up their solution so they can control feature releases and QA better. A good feature to support would be SSL for PCoIP. Etc...

I do think VMware can still catch up, time is not up, the question is do they want to?

Elias Khnaser
www.eliaskhnaser.com

Eli>
Very quick response to this:
" I think VMware needs to first decide whther it wants to be in the desktop virtualization market."

I think the answer is YES. I just think in execution it is much rougher then they initially anticipated. Remember they came from several years of growth and having a product that did something no one could even compete with and have a straght face.

There was an expectation that anything they put the stamp on would just continue that track. But, if you learn nothing else, know this:

DESKTOPS

ARE

DIFFERENT.

In many ways server virtualization was easy for the end user. But users are a fickle picky bunch. Change a server... eh ok. Take my desktop!!!! WHAT!!!!

and thats just user acceptance not to mention all the tools and technology that must go into getting the user and business to think its a good idea!

Sorry. enough for me. back to work.

Ron O (Unidesk)

No, I agree with you 100% that desktop are different and 100% it is all about user acceptance. VMware has all the components, they just need to consolidate them and integrate them better.

Buy PCoIP so that when ESX U1 comes out it does not break PCoIP, they already bought RTO, great now integrate it. Better integration with ThinApp. Buy ThinPrint and integrate.

extend support for SSL when on PCoIP. These are not hard to do.

Eli

These are all good comments folks - thanks for your thoughts...

@TREX - I don't think I'm calling the match just yet, I was saying that typically, when people are behind, and emotion kicks in, it brings about the end of the match that much quicker. Is VMware out yet? No but I don't think they will ultimately win hence the problems they're having now.

Ron - your comments are very insightful and certainly give me a different perspective on things. I think I may actually write a follow up article to this one based on some of your comments...

Andy G - I think you're right on the money there, it's not a question of whether they can put the investment is, its whether they should - hopefully I understand your point correctly there?

Indeed a very good one. IMHO brilliant.
It does come to competition if you look out of the fences.
Linux, mac os X, Oracle Solaris as VDI.
Some developers desperately need an *nix based VDI's
to develop. RedHat just jumped in the game with Spice.
X11 is a hard one for VDI.

Still considered as beta - We should not take them as an underdog.
Competition drives. An evolution was started.
KVM is very interesting.
Not bad for the customer.
IMHO - as we (stil) have Moores law.
In the near future it is regardless which (type 1) hypervisior is on.
The multi technology will eat up every issues (e.g.Context switches)

Besides that:
Citrix is now on the hyped iPad, iPhone ...
They are there ...
It is there

I agree with the observation that the cost of storage needs to be eliminated. For VDI to be successful, a virtual desktop needs to be less expensive to deploy than a PC regardless of how few or how many you deploy, and it needs to offer a good user experience.

Checkout Kaviza - scales on inexpensive commodity hardware and provides high-availability with direct attached storage. Citrix recently invested in them, and they run Citrix HDX on their desktops.

How does Kaviza make it closer to desktop costs? Perhaps a cheaper vdi but vs. desktop I am skeptical.

Intresting thoughts....I'm currently a VAR engineer building out a desktop virtualization practice primarily around VMware View. I guess what I'm focused on right now is whether View has sufficient features/options to meet enough use cases for it to be worth our time (i.e. meets needs for enough customers).

Right now I'd say that while there's items I'd definitely add to View, we can meet a lot of use cases right now. Given a firm commitment from VMware (i.e. helping people be comfortable that new stuff will keep coming), there's a lot of people comfortable with a smaller feature set in exchange for simpler infrastructure underneath/lower cost.

Just what I'm seeing right now in working with customers...

How do your customers connect via PCoIP externally? A promise from VMWare? That I think is the problem with VARs out there. They do 50 user pilots internally on a LAN and never present a full use case.

What we are seeing is that customers decide they like VMware or Citrix. Then they work out the cost against a PC and decide to abandon VDI.

Robert
(Atlantis Computing)

•  There are currently 11 visitors online  
• © 2010 insidetheregistry.com | Sitemap | Privacy Policy | Contact