VDI versus Reality

I have a long history in the remote desktop and remote application market, so for obvious reasons I like the notion of virtual desktops.  But when creating slideware for a series of one-day TechTarget seminars on VDI I started thinking about the “reality” of VDI.  A couple of VDI facts I collected are confusing but also interesting when looking at them side-by-side.  Both Citrix and VMware claim that they are the market leader and currently have a VDI user base of roughly 1.5 million each – in a market of 600 million desktops in total.  This means that VDI has a 1% market share at most, but only if all VDI licenses delivered by all VDI vendors are used in production.  That’s not too impressive, but okay, it’s an emerging market. Even if it was 5%, traditional desktops would still be dominating the market. And this is exactly what I hear when talking to customers: VDI is still a niche. Migrating their physical PCs to Windows 7 is far more important for most customers.

On the other side, in some of their public slide decks Microsoft states that VDI will reach a market share of more than 30% in 2010 and 60% in 2012 according to Gartner, IDC and Microsoft internal projections.  Okay, if this comes true, the next years will be extremely successful for ALL VDI vendors, such as Microsoft, Citrix, VMware, Quest, Red Hat, Oracle, Ericom, Propalms, Virtual Bridges, MokaFive, PanoLogic, NComputing, Kaviza, Leostream and Synchron.  Assuming that none of the big vendors and market analysts is lying, the reality may be that VDI is used in parallel to traditional desktops, adding to the total number of existing desktops rather than just replacing them.  It may still get tough for some of the smaller vendors as Gartner’s virtualization hype cycle published in July 2010 shows that hosted virtual desktops are on their way down from the peak of inflated expectations to the trough of disillusionment.  But there’s light at the end of the tunnel as Gartner also predicts that it will take hosted virtual desktops less than two years to mainstream adoption.  This means high risks for small vendors, but big opportunities for investors.

Here’s another VDI thing:  In their commercials and banner ads, Citrix claims that their products are all about simplicity.  What?  Citrix is either ignoring the facts or has some interesting interpretation of the word simplicity.  Today’s VDI solutions are all about solving provisioning, brokerage, networking, storage, scalability, user experience and personalization issues.  Without any doubt these are all very complex things, which not only influence VDI environments based on Citrix products.  VMware and Microsoft are not better here.  Installing a production-ready VMware VDI environment is also far from being easy, no matter how experienced you are.  Ever tried to setup a Microsoft-only VDI environment with all the different server roles and role services?  If you did you know what I’m talking about.  That ain’t simple!  It takes me a full day to introduce the core VDI principles, products and challenges to seasoned IT professionals – and I can only scratch the surface.  I interpret simplicity as something similar to the iPad user interface (even though I don’t like Apple stuff too much, but that’s another story).  Everything that takes longer than ten or twenty minutes to introduce and everything that requires an expert to make it work is not simple.  Telling an IT professional or his boss that introducing and maintaining VDI is simple just sucks, it’s misleading.  Citrix, can you hear me?

But wait a minute.  Maybe Citrix was looking at VDI from a user’s perspective.  Provided that a troop of house-elves installed and perfectly configured a full-blown Citrix XenApp and XenDesktop environment (including some add-on products), things may indeed look simple for a user.  He or she can use any endpoint device and get intuitive access to any application – independent if installed locally or remotely – while enjoying superior performance and full personalization.  The nature of the endpoint device doesn’t matter; Windows is only the mandatory backend runtime environment required for a subset of the applications requested by the users.  The endpoint device can be whatever the user prefers to use.  If this is what Citrix had in mind when creating their marketing campaign around simplicity, I’m all in.  If only it was not so hard to find good house-elves and get the endless budget approved.

But it gets worse.  In two weeks I will be speaking at BASTA, a major developer conference in Europe.  I’m on the Azure track, talking about “Windows applications in the cloud” (Tue, Sept 21, 5:15pm).  My goal is to educate developers about the impacts it has when running their Windows applications from one of the major VDI and remoting platforms.  I’m really looking forward to meet with the Microsoft developer community.  But when browsing the agenda and seeing what other speakers have to share with the audience I get the feeling that I come from a different planet.  They are talking about Software as a Service, SOA and Silverlight while Microsoft confirms their confidence that this will be the future.  Wait a minute, that’s the same Microsoft saying that VDI will be the dominating desktop delivery model of the future.  I mean, let’s face it, as of today Azure doesn’t even have an IaaS offering.  The Azure platform cannot host virtual desktops you can connect to through RDP from the outside, even though Prashant Ketkar, marketing director for Azure, promised such functionality in February.  If it was really important to Microsoft, they would be offering this service by now.  Is Microsoft cheating on us VDI guys?  Well, I wouldn’t go that far, but Microsoft seems to have multiple agendas.

So, in a nutshell, it looks like all major vendors and market analysts are walking in the dark.  The Citrix versus VMware fights may become obsolete in less than two years when Silverlight finally dominates the market and Windows is just an application hosting platform within Azure - just kidding, but I scared you for a moment, didn’t I!  But there’s still some truth in my statement.  What if the Microsoft developers are right?  What if they are about to start creating the next-generation applications that don’t require a Windows desktop anymore?  Did you ever see how smooth Silverlight applications run on a Mac?  Is it possible that in five or ten years we will be looking back at virtual desktops with a smile of wisdom and knowing that they were only good for a transition phase?  On the other side, who knows what IT users will be asking for in some years?  Maybe the next-generation VDI will be based on an über-cool Windows 8 or 9 desktop that rocks the world by then.  Just keep that in mind when listening to all the VDI marketing at VMworld, Synergy and Tech-Ed.

Some Nordic Thoughts on the Future of App Remoting and User Interaction

Last week I attended the Norwegian Citrix User Group Spring Event in Lillestrom (which is close to Oslo). There were roughly 100 attendees and this crowd is really inspiring. Despite the fact that they were never cheering and screeming while I was presenting my sessions, I had some really good conversations with individual memers of this group. In addition, we had some round table discussions which made me think about our industry.

One question I was asked was “What do you think was among the most influencial technologies developed over the last years?”. My answer was touch screens and gesture-oriented computer interaction. I truely believe that these technologies will massively influence the way how people interact with computers. This doesn’t mean that they will replace keyboard and mouse in all use cases. However, touching a screen and using gestures is a lot more intuitive for many users, so they will be adopted rapidly. Unfortunately, most existing applications cannot really profit from a touch screen yet. I expect many application designers and developers to come up with new solutions for that, revelutionizing “natural” user interaction within the next five years.

What does that mean for virtual desktops and and remote applications? There are so many existing Windows applications - in particular in enterprise environments - that will not disappear shortly. I believe that Windows will slowly turn into an application hosting platform providing such Windows applications for dedicated touch and gesture-oriented frontends. These frontend devices will go beyond the iPad form factor as we it know it today. The application runtime logic including the graphical UI will be clearly separated from the device used for user interaction. As a consequence the difference between web applications and traditional Windows applications will disappear step by step. This will not only be the case at work but also at home. In other words, desktop virtualization and application remoting will be a commodity. Microsoft and Citrix (and to some extend VMware and other vendors) will provide the necessary base technology regarding application runtime platform and remoting protocol. Multiple vendors will provide the forntend devices, satisfying the requirements of a heterogeneous target audience. Do you agree? What are your thoughts?

Isn’t it funny, I didn’t even mention the hype phrase “cloud”, as cloud is only one scenario among so many that can benefit from mature desktop and application remoting technologies.

Are Pure Virtual Desktops only Successful in Virtual Markets?

Since almost two weeks, Shawn Bass and I teach Advanced Citrix Training Classes and present at Geek Speak Events in Norway and France. We had an audience of almost 30 training class students and 60+ event attendees. We always asked about the experience with virtual desktop infrastructures in production environments. The amazing outcome was that there was literally none. Sure, there are a number of proof-of-concept environments, but no real production environment with VDI - at least when talking to our peers in Norway and France. Only Shawn knows a US customer with several thousand users working on virtual desktops. But these desktops are dedicated desktops with private images on an expensive storage system.

I was never a fan of pure VDI. And again I’m asking myself if there is a real market for pure VDI? Or is it only a “virtual” market? According to articles published by several VDI experts and marketing material coming from serious software vendors, VDI seems to be the only future for them. But if this VDI market really exists somewhere, it better materializes fast, or investors may pull out there money and invest it into new stuff. And the media may fall in love with the next buzz word.

Another question I’m asking myself is what will happen to all the VDI POCs if this market never really takes off as expected? Will there be technical experts for de-virtualizing virtual desktops and anti-virtualization ghost writers cleaning up enterprise strategies by removing all references to VDI? This all may happen when people find out about some limitations with VDI that cannot be removed on short term, such as properly integrating TV overlay cards, video cameras, DVD burners, external harddisks, music players, modern cell phones or pen devices.

The virtualization war has just begun. Betting all your money on only a weapon called VDI seems not to be a smart move. So I would agree with Microsoft that VDI is good for all enterprises but not for all desktops. In particular for the users it’s the right technology mixture combined with usability that matters, not a vendor-driven one-for-all philosophy. The feedback I received from experts and peers over the last days and weeks reconfirms this statement and indicates that the world of desktops is not VDI flat. The surprising fact is that many virtualization expert out there seem to have silently accepted this already. Now they are waiting for the moment the “pure VDI flu” is over and they can go back to work and implement the full spectrum of virtualization technologies as needed.

MVP Summit 2010 – Microsoft’s Universe of Virtualization

When I attended the MVP Summit at Microsoft’s headquarters in Redmond last week, there was this moment when a simple truth struck me: Microsoft is sooooo different from VMware and Citrix and Apple.  Well, you may say that this doesn’t come as a big surprise.  Most probably you are right, but this time it’s for a different reason.  It happened when an MVP’s question to one of the Microsoft Windows product groups top managers was “Where do Windows workstation and Windows server product lines meet when decisions about the general virtualization strategy are made?” and the answer was “At Steve Balmer”! What? Are you serious?

Even if this answer may not be 100% accurate, it gave me some valuable insight into how Microsoft sees the IT universe, a simple fact that was confirmed again and again during a number of conversations I had with Microsoft folks over the last week.  It looks like Microsoft does not intend to convince “the market” – customers, partners, IT professionals, consultants and system integrators – of a specific virtualization strategy.  It looks like Microsoft avoids predicting a single virtualization scenario which they believe will be the one and only future option.  There is not this one truth spread by Microsoft evangelists.  Instead, Microsoft simply tries to have enough products on stock for all potential virtualization scenarios!

What does that mean?  Microsoft is always working on many new products, some of them brand new, some of them in stealth mode and others just new version of existing products.  All these Microsoft development activities follow a common design pattern, demanding that all resulting products must have the potential to be good for the masses.  I mean in the range of several 100 million units.  This doesn’t imply that all products will really become successful enough to sell at big scale.  But each of them could be a 100 million seller, if the market decided to go into a specific direction.  And if not, Microsoft just dumps a product that didn’t make it after a while, or gives it another try with a different developer team, waiting for a second chance.  Dumping a product doesn’t mean that it was technically bad; it just means that it was not accepted by the masses or that the timing was not good.  There are quite some Microsoft products I would not regard as technically superior, but have a big market share and thus are successful according to Microsoft’s definition of success.  It seems like Microsoft’s overall mission is “good enough beats great” when done at the right time and at the right prize, in particular when looking at the virtualization market.

This brings me to VMware.  Their strategy is convincing the market that virtual desktops are the future, now that their initial mission of virtualizing servers is almost complete.  In the VMware universe, there is no other valid future scenario, only virtual desktops, nothing else, no matter what.  And they are convinced that this scenario requires the best-of-breed products which VMware claims to provide.  If VMware was in the building and construction industry, offering sophisticated bathrooms and kitchens, they would say that it’s enough to have a bathroom and kitchen to live in.  In a way they are right.  As a matter of fact, you can happily live in an apartment that only has a nice bathroom and a big enough kitchen with a bed – an apartment I rented when I was at university was pretty much that way.

But there are situations when you want a living room, extra sleeping rooms, a basement, an attic or a patio.  Probably you don’t need them all at the same time or you need them in different combinations at different occasions.  But you want to have the freedom of choice, despite the fact that not all of these rooms can be of superior quality right from the beginning.  In the virtualization market, it’s pretty much the same.  We have server virtualization, session hosting (aka presentation virtualization or terminal services), desktop virtualization, client hypervisors, application virtualization, storage virtualization, cloud computing and more.  It’s Microsoft’s clear ambition to have good enough solutions for all of them – or have a trusted partner ecosystem to fill the gaps.  This is where Citrix comes into the game, and Quest, and Symantec, and many others – but not necessarily VMware with their one-size-fits-all vision of virtual desktop for everybody.  The Microsoft universe is so different from the VMware universe.

Unfortunately, Citrix seems to be impressed by VMware’s strategy on one side and by Apple’s success in the consumer market on the other side.  When translating this into the apartment story again, Apple wants to convince us all to live in a really cool trailer, just because it’s mobile and it’s all you need to be trendy.  This all inspires Citrix to concentrate on building trailers for business people – including big bathrooms – instead of continuing building great living rooms and sleeping rooms and patios for Microsoft houses.  Come on Citrix, give me a break.  I know you can do better than this.  Tell your customers that you have a spectrum of products that goes far beyond trailers.

All that being said, what are the implications?  I get the impression that Microsoft does not have that one solid vision of what the virtualization technology of the future will be.  And to be honest, they just don’t need to.  With all the money and developers they have, they can afford building good enough products for all possible virtualization scenarios and ship them right after they see that the mass market is ready for them – plus provide some early teasers to probe the market.  They were always extremely successful being a second mover, letting others – like VMware and Apple – find out what the real market trends are.  As soon as a market trend is confirmed to generate substantial revenue, Microsoft can use their marketing power to deliver the right products.  The trick is not to lose touch with the successful market seekers and keep on developing all sorts of solutions with the hope that some of them will become relevant in the future.  Not the fastest and not the most elegant method, but very successful so far.

Now, what about Citrix?  My personal opinion is that if Citrix was smart, they would just use the spectrum of the outstanding products they already own and continue to use them to convert good enough Microsoft products into great solutions for specific market segments.  I don’t believe that copying central aspects of Apple’s or VMware’s marketing strategy is beneficial for Citrix and their customers.  The IT universe has more to offer than just virtual desktops.