Will VMware still exist in five years?

After Alex Barrett blogged about a conversation I had with her last week on this very topic, my inbox has been flooded with emails asking "Did Alex really quote you properly?"

This wasn't an article I intended to write right now. But after Alex Barrett blogged about a conversation I had with her last week on this very topic, my inbox has been flooded with emails asking "Did Alex really quote you properly?" and "Are you crazy?"

The answers to those two questions are "yes" and "quite possibly."

(For the record, Alex asked me point-blank whether that conversation was "on the record," and I said "yes," so she and I are cool.)

VMware deserves a lot of credit. Even though hardware virtualization has been around for decades in one form or another, we wouldn't have it in the x86 space without VMware. The hardware and OS vendors would have been happy to keep selling hardware that was only 20% utilized. VMware turned this industry on its head. They deserve credit not only for the move towards virtual hardware, but also for the whole VDI concept. (Even though VMware did not initially embrace VDI, the early adopters / creators of the concept couldn't have done it without VMware.)

So kudos to VMware for doing some awesome stuff.

But VMware will face some tough times ahead:

  • Hardware virtualization is becoming a commodity, and when this happens, you end up with a lot of competitors, feature parity across vendors, and a price race to the bottom.
  • The "easy" virtualization sales have been made already. What's left is the more complex stuff, with longer sales cycles and more complex deals.
  • Now that VMware has "proven" the concept of hardware virtualization, and now that analysts have predicted this market will be <insert some random 11-digit dollar amount> by <insert some year>, many companies are entering the space.

VMware has awoken the slumbering giant that is Microsoft. Sure, there have been isolated cases of smaller companies successfully competing against Microsoft, or smaller companies partnering with them (e.g. Citrix), but in general, if you're a software vendor and Microsoft puts you in their crosshairs, you days are numbered. Microsoft will add many of VMware's core features into the base OS over the next several years without really increasing the price. And in addition to Microsoft, many other companies are entering the hypervisor space, including Citrix, Novell, Sun, Oracle, and even Phoenix Technologies (the BIOS makers).

Of course VMware has the first-mover advantage in the virtualization market, and conventional business wisdom suggests this can help a company win long-term. But history is full of wildly successful businesses who only entered a market after another "first mover" blazed the trail. Look at Amazon, Walmart, eBay, Starbucks, Microsoft--none of these was the first company in their sapce, and each of them dominates today.

I don't want to minimize the impact that VMware has had on the industry. But I think in five or ten years, VMware will be more significant for what they did in the 2000s, not what they're doing at that time.

I'll write another article tomorrow about what VMware could do to maintain their dominant position. Feel free to share your thoughts in the comments today. Here are some points to kick-start the conversation:

VMware is doomed

  • They only make one type of product, and it's a market that everyone is entering.
  • They're competing against Microsoft.
  • People will want to buy a more complete integrated solution from one vendor, and that includes things that VMware doesn't own
  • VMware is too dependent on other vendors that they're also competing against. For Citrix servers, why not use Citrix's virtualization? For VDI, why not use Microsoft's RDP or Calista? For packaging offline Windows VMs, why not use Microsoft's Kidaro to package Microsoft Windows? For software distribution, why not use the packager of your distribution vendor (Citrix or MS) instead of Vmware's Thinstall?

VMware will continue, no problem

  • First mover advantage. ("No one ever got fired for buying IBM." We have that now in the virtualization space with VMware. VMware = Virtualization. Period.)
  • Even though their stock has lost more than 60% of its value, VMware's market cap is still $18B (3x Citrix)
  • VMware is owned by EMC. (Remember only 10% of VMware's shares are public.) EMC's market cap is $31B
  • The virtualization market will be HUGE in five years. VMware only needs to grab a small slice.
  • Only the actual hypervisor will become a commodity. VMware and others will compete on the value-adds, which VMware leads

What does VMware need to do to survive?

  • ?? Let's discuss ??

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

VMware needs to be aggressive and go after Microsoft. I think it's BS that Microsoft is allowed to push everyone around as it stiffles competition and slows down progress in the software industry. Brian, I believe that Vmware will cease to exist or will become insignificant if they do not aggressively go after Microsoft. They have a great 2-year lead over any other Virtualization vendor and they should use it NOW. VMware should come up with a bold strategy to take a big piece out of Microsoft's OS segment ownership - it is possible - by acquiring or developing on their own some new technology that allows them to wrestle the OS control from Microsoft. I don;t know what it is but there must be something they can do.

If they "just" some up with a better version of ESX - gthats not going to do it. Like Brian said, hypervisor will be a commodity and a year from now ESX just cannot be all that much  better than HyperV. Vmware needs a BOLD AGGRESSIVE MOVE. 

There are companies that Microsoft is scared of - i.e. Google. Vmware can be that too, if their smart and cunning. We don't need another Citrix who sheepishly follows Microsoft's lead. 


VMWare needs to put some effort in marketing ....until now they have been the only good solution but in the future there is competition...even at same level techically.

This is something MS does well...marketing I mean...for some reason they know how to push out products and make it sound like they are the first to do this :)


We need players like VMWare for sure...with good products and things that suits the customers.

Note this comments come from the northern part of Europe called Finland...maybe the marketing is better elsewhere!?!?!?




Even though I haven't had a lot of opportunity to work with ESX (our
one server just chugs along),  but I've been reading as much as I can.
I think VMware is overpriced, particularly Virtual Center, something
has to happen there that's better for SMBs. Nevertheless, VMware's
presence and 'first mover' advantage makes a big difference if one is
in a situation where one doesn't have a lot of choices of consultants
etc. I think VMware will have a role for a long time in that it can run
on 'old' x86 hardware, and most of the new virtualization products
can't do this, they require 64-bit processors, VT, etc. SMBs with a
smaller amount of servers don't have so many management concerns as do
bigger companies with more servers.

I think most of these companies have to stop the PR/marketing sniping at each other, too. 


the current trend is to place the hypervisor into the hardware chipset. This is what Intenl and AMD are dreaming about and this is what I've understod from Xen (SimonC) vision of hypervision... Another reason to not place a futur into the software hypervisor...

VMWare have to move one step, they know it and they have started to... What will be the futur : they own the market space, they have money (cash) and additional money available (90% of there shares are still private)... 


What about a Cocal-Cola/Pepsi-Cola world ? (or MacDonald/KFC)... aournd 50/50% worldwide, some local advantage for some of the player... 



I think you missed Google both as an example of a company that has been successful vs. Microsoft and also as an example of a company that managed to be successful without being first (there were many search companies before Google). To succeed VMware needs to emulate much of Google's behavior:

  1. As you stated currently VMware = virtualization in much the same way that Google = web search. VMware must do all it can to cultivate its brand

  2. They need to innovate, innovate, innovate. Never sit still to rest on their laurels. Otherwise you become a sitting duck for Microsoft. And then you become a dead duck

  3. If they can't innovate quickly enough then they should buy. VMware should be doing more M&A - they certainly have the cash for it. Don't worry about integration - re-branding should be good enough to begin with (Thinstall is a good example of this)

  4. VMware needs to become more of a platform to developers. They are on the right track with stuff like VMsafe but they need to be much more aggressive about it. Developing directly on top of the VMware platforms should be easy and cheap. They should hold developer conferences like Microsoft currently do (this is more of an Adobe strategy than Google)

  5. Exploit Microsoft weaknesses, e.g. Microsoft does not have a coherent VDI strategy

  6. Don't be afraid to take risks

Well, enough tips for VMware ;-)


Next week ask the same question about Citrix. Why use Citrix if I can have more or less the same from Microsoft?

I think the discussion is not about the hypervisor - that's only a small part of the code (and btw. it's still a long way to virtualize everything in hardware - CPU, RAM, NIC, disk). It's the framework around it, the services, the management tools. See Citrix.


Brian, how much money do you have to bet on this? VMware is a good company - how did they managed to piss you off? Do you have briforum sponsors?




Do I sound pissed off in this article? I'm just laying out my opinion. If you think VMware will be a leader in five years, please tell me why! I'm always open to changing my mind. In fact I'll be visiting VMware in May for a day or so with an opportunity to talk to their product folks. I'll be visiting MS in April and Citrix in May too. So I'm looking forward to hearing about everyone's vision and then writing more on this. 

Getting onto the chip would be a great move but could this not be taken one step further and sell a VMware branded box? They're backed by EMC who are storage specialists, how hard would it be to shift a branded box with ESX on a chip and a fat EMC storage solution inside as well.

VMware's lack of marketing is pretty baffling. I wish they would let
people know what they've got. The word from a lot of IT writers seems
to be that virtualization is a commodity -- but people deploying VMware
realize this is not the case. Take a few examples:

Xen can't provide anything near the stability or consolidation ratio's that VMware can.

Microsoft can't even do VMotion. 

No one else provides the ability to let your data center do its own resource management like VMware's DRS stuff. 

one else has the rich set of API's to build on, between the virtual
center and CIM API's they offer,  building scripts to allow your to
customize your deployment is a breeze.

Being able to buy mature
integrated solutions for backup, disaster recovery, and workflow
management makes rolling out a complete solution much easier.

does have a better platform than the competition today, but for some
reason they don't seem to be educating anyone outside of fortune 1000
IT organizations -- they should let the rest of the world know  this
stuff. Perhaps in 3 years Citrix or Microsoft will have a solution that
provides comperable value to what VMware offers today -- that doesn't
help me when I need to roll out a new deployment next month.


I don't think VMware has a lot to fear, as long as they continue to innovate, and lead the market. It goes without saying that they will lose market share (that's inevitable when you have been the only show in town and then face competition), but if the market keeps growing the way it has been, a smaller piece of a much bigger pie will still result in major growth for them.

As silly as it sounds, Microsoft entering the market could even help VMware, as Microsoft's marketing dollars will raise general awareness around virtualization technology and grow the market. As awareness grows and people start looking at what products are around, most will see that VMware have the better feature-set and more stable, mature offering - not to mention a better consolidation ratio than any of their competitors at the moment. (despite what MS and Citrix have been saying lately, VMware's memory page sharing mechanism is a compelling feature - That's why both MS and Citrix have similar features on their product roadmaps.)

If anyone needs to worry it should be Citrix. Their Presentation Server product has stagnated and their desperate need to jump on the VMware's coat tails has seen them spend a ridiculous amount of money on a sub-par product that is nowhere near enterprise-ready. With the new incarnation of Terminal Services from Microsoft attacking their Presentation Server (sorry - XenApp.....) market and VMware killing them in the virtualiztion space, I can't see many people flocking to the doors of Citrix.


The "system" will not allow it...too much money to made....Oil companies...Detriot.....etc..controlling how and what we buy and drive

 Why is the software/OS business like it is today? M$...good and bad..they made it easy to use and consume...but their business model, software stack is a dinosaur waiting for the comet...just happens M$ can shoot down the comet and any other THING/COMPANY that gets in their way....


dedicated to A Clarke..... still waiting for RAMA...





I'm not a teckie like the rest of you guys but I have worked in the IT industry for long enough to know that there are vendors that play a good game and vendors that talk a good game.

Vmware is in the former camp, Citrix the latter. Vmware's products work, they do what they say they are going to do. Presentation Server is riddled with niggly problems and its years old (that's not my view, that's the view of Brian's mate that's gone to work for Provision Networks). Why on God's green earth would anyone trust their mission critical servers to Citrix's hypervisor when it can't get its core product right?

In addition, Brian tells us how wonderful Citrix XenDesktop is going to be. How does he know, its not even released yet. Citrix Workflow, not released yet. Sure its all going to be wonderful, when it finally gets on the market. WANscaler was going to be the next big thing. But it doesn't accelerate ICA traffic. No, but it will do in its next release. The future's always rosy with Citrix but for now you have to make do with products which don't quite work how you hoped they might.

So, do I recommend a tried and tested technology which doesn't need vast amounts of marketing because its works and word-of-mouth is the best marketing tool or do I go for a solution that needs technology from two vendors with a reputation for overhype (Microsoft and Citrix) and still doesn't get close just because the "strategy" i.e. vapourware sounds better. My customers need something to do the job today not some hazy wishy,washy marketing spiel.

Never mind Microsoft winning this battle in 5 years, at the rate Vmware is being adopted, its going to be over in two.

Oh yeah, "Don't worry about integration - re-branding should be good enough..."  Thank you for saying what perfectly explains why so many software companies are missing the boat today.  Integration should be everything to a software company - if you forget that, your customers will go to companies who understand it.  Stop worrying so much about Microsoft - they're having their own integration issues - which in my estimation will cause a major slimming down of MS in the next 3 years.  Focus on your customers and their needs.  Listen to them and try your best to understand - that's what made all the great companies you've listed so successful.

I am a User Workspace evangelist these days.  Where is the reference to the end user in all of this?  Issues related to performance, reduction in the complexity's of management and maintenance within the network technology stack today?  

As to data, there is a twist in my minds eye this morning.  One of the great advantages of the RES Powerfuse (evangeilizing User Workspace Management) environment is consistent auditing, not only for the administrative work associated with console configuration, but also end user visibility.  From a single console, I can review every event log associated with a user, across every instance of an OS within the network, and every application available from this context.  I see global visibility to all events associated with administrative function as well as the user session.

I reference this as EMC is a leader in Information Life Cycle Management.  I have worked at large financial houses here in U.S. who are required to maintain six months of event logs from all servers in production.  These same organizations have similar requirements for a range of data sets maintained within the network, email should be deleted when it is six years and one day old as an example.  There are synergies to be explored beyond the context of OS instance delivery and access to the network.  The ability to integrate EMC technologies, the enormous SAN backend, coupled with the capacitiy to automate data collection and repository via ILCM is intriguing.  There are big paybacks on the regulatory compliance side, and these paybacks fall into the value proposition.

As to the issue of Citrix Virtualization, there is something to be said for the security component of the Hyper-V.  Look here for an elegant 40 minute conversation on the issue (the interviewer is in a tough spot on this one....which is funny), http://channel9.msdn.com/ShowPost.aspx?PostID=365911

VDI, given the cost I am just not convinced that VDI trumps SBC.  The additional OS, TSCAL, remote access device, and additional labor to manage that remote access device are big hurdles to overcome when stating value.  I am not convinced that VDI overcomes these hurdles.

I have experience utilizing encrypted workstation OS instances for access to large customer environments in the U.S.  The time to deliver a hardware device to consultants working remotely for the organization is an impediment.  It is not uncommon to spin up a VM on the workstation, install the required software, encrypt the drive, and you are onto the network.

As to Thinstall I am not an expert at this point....:-).

Where is the user?  Where is my single management console for all of this business?  How do we reduce the complexities inherent in the implementation engineering side of the house, to pass the repititive tasks off to lower skilled engineers in the enterprise.  These are all customer service questions I still find unanswered from the virtual side of the house. 



Do you want to dring something ?

Today, you have the choice. 2 large leader have dived the world. Pespi-Cola and Coca-Cola have quite the same size and fight as equal on the market. Locally, one is better than the other for a lot a valuable reason but WW speaking, there are equal.

They offer the same type of drink light/max/black/sparkling/maid... The good thing is that can's have the same size so I can buy a standard fridge. I can also you any glasses on the market and I can easilly change from Pepsi to Coca... 

You can even change for some local specific if you want a root beer, Dr pepper, Orangina, Guiness...

that's what will the market be. 2 strong leader (VMW and MSFT if they both play well), some specifric player (CTXS, WSOFT, ...) but a full layer of interoprability at the fridges and glasses level ;-) 

Once I said that... Let's have a drink... of red wine for sure ! 

Why should people use Citrix instead of VMware? The most companies I know are still using VMware as virtualization solution and Citrix Presentation Server (I hate the name XENapp) for there remote users. Maybe people will change their minds now ... using VMware VDI instead of Citrix technologie ... Maybe VMware buys some protocol company to have their own protocol, even better than ICA ... with more multimedia capabilities ... Think about it ... not everyone is a friend of Citrix or is paid by them :-)

Obviously integration is important - in an ideal world everything will be integrated, in particular products from the same vendor. In reality achieving tight integration can be a very difficult and lengthy process, certainly when bringing in products and teams of developers from the outside. You often have functional overlaps, incompatible approaches and architectures, NIH syndrome, etc.

What I'm saying is that such companies should not suspend M&A until all existing products are properly integrated. Instead, if they see a product that can benefit their portfolio they should go for it, even if they know that it will take a long time to integrate it.

As an example, Microsoft has been able to make very good use of SoftGrid without really integrating it. Let's see if VMware is able to do the same with Thinstall.


I think the biggest barrier is getting support from Hardware vendors like HP/IBM/Dell.  VMWare is acting like the 800 lb gorilla in this space and their Linux Source code is becoming more proprietary while XenServer is based on XenSource which completely open source with thousands of Penguins developing new features for it everyday.  If VMWare opened up their source code to the Linux community it would provide them with more leverage with the big Hardware shops to get their Embedded Hypervisor factory installed.  Right now XenSource is considered to be more of an industry standard @ most large Hardware companies..... 


ESX ist not Linux! The hypervisor is it's own OS. Linux is only the Service Console!

Anyway ... We are you thinking that VMware is possibly gone in 5 years?  

 My opinion, MS wins, VMWare loses. Sucks since I purchased VMware at $90. But, who is the biggest benificiary of all this, Citrix. They will own the VDI space and chip away with MS at the Server space. Citrix became a billion dollar company working with MS, here comes 10 Billion.

I would rather eat my own vomit than use MS V Server 2007, over ANY VMware product.

MS is rubbish in this field and i am sick going to sites that use it.  As for 5 years time, by then MS will finally have devloped something similar to vmware server 1.

VMWare is the leader and it should stay that way, more hardware compatibility for ESX, simplfy ESX, including HA and the consoles.

it would be like the bush/gore election, more votes for vmware but MS VS still wins, go figure.


They understand very well that hypervisors are just a commodity these days. their focus is management now. all current products revolve around stage manager, site recovery manager, all that cool stuff. and their focus will be to exploit virtualization in the next couple of years. that means continuous availability, VMs sized bigger than their host, stuff like that. that's something the others will have a very hard time to catch up to. it's the headstart of 5 years of development that will keep them alive for at least the next five years.

some more fuel to add to the fire:

  • in 5 to 10 years, microsoft will not be making OSs anymore. linux will rule the world and it will not be lucrative for them anymore. same with hypervisors.. it'll be commodity and free. i wouldn't be suprised if they dropped their hypervisor with their OS builds. it's all about management, not hypervisors.
  • i had more but i got tired. 

I intergrate both VMware, Citrix and Microsoft products, certified and all of that nonsense.  I can tell you right now, VMware is and should be concerned about Microsoft and Citrix.  Microsoft's Hyper-V and Citrix XenServer are architected in the same manor, one has a Windows Domain0 the other Linux, I am sure you can figure out which one.  There are also many companies that are leveraging the Xen Hypervisor (Sun, Novell (I believe)).  VMware still touts VMotion as the "Ahhh" factor, and while it is cool, XenServer has the same deal.  I completely agree that ESX still reigns supreme over XenServer, but price wise, VMware will have to make cuts to compete with Citrix and Microsoft.

 From a channel perspetive, Citrix blows away VMware, of course this is my opinion.  I have to think there are more Citrix partners then VMware, which could be a negative, but for all those XenApp (PS) customers out there that aren't using VMware, the XenServer is an easy sale.  Looking forward, the new versions of XenServer coming out that will use Marathon for HA will be critical.  Citrix is dumping major money into R&D.  Add the fact that XenApp servers are supposed to perform better on a XenServer vm as compared to an ESX VM, now we are talking.

Citrix just needs to acquire a P2V company (PlateSpin) and they will be solid.  As of now they don't have any P2V tool, which should be considered a huge negative.  Let alone the enterprise features that aren't there yet, but are just around the corner.

So the question is how VMware survive???  I agree with other posters, they need to market heavily and come out with the next "Ahhh" factor.  Their latest release of products (Stage Manager, etc.) don't have that factor. 

Brian - it will be interesting to see how this pans out and I am very interested in what you have to say.

 My .02

Firstly VMWare needs to be clear that they have a beach head with Server Virtualization but the Desktop playing field is green. MS has RDP (Sucks over the WAN), Citrix has ICA, HP has RGS. VMWare has nothing. They need to address this gap first to do VDI. Perhaps extend RGS to make it better than ICA. Better bet than trying to extend RDP which Microsoft can make really hard for them. Once they have this, they have to figure out a way to bring storage costs for VDI down to make the TCO real for average people as well. Citrix has Ardence which helps, and VMWare needs to evolve their answers in this space. All this stuff has to work, so they need to innovate and keep moving up the mgmt stack which they already show signs of doing. For the hypervisor, they need to do paravirtualization to answer the others and make it compelling since the hardware can do more. Perhaps give away the hypervisor for a while to kick MS/Citrix in the balls. They should form partnerships with other vendors to make VDI happen. MS is trying to corner them with Sun announcements etc. They should beat them at their own game. Go talk to Google and conjur up ways to screw MS as well. I am sure they will be happy to help. Note all this talk about VDI, havn't even mentioned their CORE strenght the server side. Keep investing and making that better since you already have the lead there. My other piece of advice is make ACE a real product like Kidaro could have been, and MS won;t let happen. Build the controls like Kidaro had and make it work on MAC with Fusion. MS will hate you for it, but it will make VMWare really happy. Bottom line is, we need VMWare to be around to keep MS and Citrix in check. MS and Citrix will do nothing useful is VMWare is not around to kick them in the balls.

Citrix can't have Platespin.  Novell picked them up for $205 mil.

I talked to Citrix last week, they told me to focus on the provisioning instead of virtualization. Do not forget how inovating VMWARE has been of the last 5 years. No vendor at this moment can provide the same features as VMWARE. VMWARE will keep on inovating and move closer to hardware as it will eventually be integrated into the hardware itself. I think that MS should be worried, as the OS is becomming nothing more than an application.

I'll grant you Citrix (on Terminal Services) has its place in the remote application delivery space, but having used their products for 10 years, the first word that comes to my mind when the word Citrix is spoken is not reliability.  Nor is it quality.  Their Presentation Server and related infrastructure products and patches (web interface, secure gateway, client releases, printing, etc.) are a pain to deal with.  Each time a new piece of code comes from Citrix, the big question is "Oh oh, now what is going to break after we install this?".  Their quality control on software releases seems practically non-existent.  Testing for backwards compatibility is severely lacking.  Their support department's cavalier attitude of "you always need to be on the latest and greatest rollup and hotfix" to fix XYZ small issue doesn't bode well for companies with large deployments of presentation server infrastructure that must undergo significant testing to complete the release cycle.  "Always upgrade to the latest Citrix client".  Try telling that to the desktop teams who receive word that they must quickly package, test, and deploy the newest Citrix client to thousands of PCs, most of which are remote in the first place which is the fundamental reason Citrix exists in our environment to begin with.  Citrix also has developed a pattern of conveniently pretending problems don't exist.  Acknowledging the problems your customers are hollaring about in the support channels would be a step in the right direction.

Everything I just spoke about above (and much more that I just won't get into here) is why I believe Xen cannot be trusted in the datacenter to serve as the platform for production virtual machines as long as Citrix is running the show, if they are going to apply the same level of quality to Xen as they do with Presentation Server.

Brian, I'm afraid I lend more credit to your 2girls1cup reaction YouTube video that this piece of shortsightedness.

I have to admit I didn't like VMWare's position on why they don't have a true 64-bit product yet. Basically they said they don't see much benefit in performance by having a true 64-bit product as opposed to their current 32-bit product that just emulates 64-bit. So a true 64-bit product isn't a high priority for them to produce. 

While it's true that there's probably not much to be gained performance wise, I think if that's defining their innovation efforts, then Brian may be right that they won't be what they are today in 5 years. We already see 3rd party companies deciding either to not integrate their solutions with VMware, or putting out VMware supported versions of their products well after their 64-bit hypervisor supported products have been released because of the limitations of VMware's 32-bit code (Egenera and Marathon Technologies to name a couple).

VMware has been the innovator in this market for some time now, but if I'm a CEO of one of these 3rd party virtualization add on companies, and my engineers tell me that the product they're engineering either won't run on the product that has 3/4 of the virtualization market, or will be a year or two before it will run on that product, they better have a damn good reason why it won't. And what it comes down to is that investing time engineering products for 32-bit code is not where the future of virtualization innovation is. But don't take my word for it. Ask the company that won product of the year at VMworld 2007 why their product won't run on VMware (Marathon Technologies Everun requires a true 64-bit hypervisor). 

So IMHO, where VMware will be in 5 years is greatly dependent on their commitment to migrating their current code from 32-bit with 64-bit emulation, to a true 64-bit product. Without that, pretty much all the innovation that VMware products will have 5 years from now will all come from in house... which is pretty much how they've operated from day one anyway. Which is my second point. Unless VMware starts to open up how they do business, and actually begin to play well with other tech companies like Microsoft and Citrix, and 3rd party add on companies like Marathon and Egenera, and especially the resellers of their own products, they're just motivating everyone else in the industry (outside of VMware customers) to rally together to combat against VMware, with the whole Microsoft/Citrix virtualization partnership being a fantastic example of that.

I don't think Brian is crazy at all for throwing this out. Obviously no one can predict the future. And this really is no different than playing the stock market, where being right and wrong in one's speculations isn't much different than throwing darts at a dart board. But those who are speculating that VMware has nothing to fear over the next five years are the one's who aren't being realistic in my book.


Agree that VMware should fear the competition.  It will keep all the players honest - good for the customer.  But I think to strongly question VMware's existence in 5 years is overboard, sensationalist, naive, and disrespectful.

Brian, I still like your stuff but seeing this article coming from you caught me off guard.



 Slightly off-topic, but could you give us your thoughts on whether Citrix will still be around in 5 years? We are mostly a Citrix shop and about to invest more. Are they still going to be around in the SBC world or are they going to turn into solely a VMware competitor?
Is it a mistake to upgrate to XenApp (CPS 4.5) now? Should we wait for Windows Server 2008? Is Citrix going to keep improving XenApp?




Uhh, who are you kidding. Do you really think x64 has no value?  Also don't believe Citrix XenDesktop has a x64 strategy. Vista is not happening, who uses or plans to use that in anger? XP x64 is a decent product and MS is just trying to force upgrades to Vista.


Citrix XenServer does have the XenAPI and Microsoft has COM APIs for Virtual Server and WMI APIs for Hyper-V. You might want to also look at System Center Virtual Machine Manager before you decide nobody other than VMware has a good management solution for virtual infrastructure.

Also, I have not tested the capacity of current release of Xen against VMware ESX. I am sure you have not either, so lets not jump to conclusions on the performance capabilities of these two platforms. ESX may only have a perceived advantage in performance. Won't know until it is proven by an independent test.


I think this is a great point because vendor affinity can play a huge role in technology decisions. Sorry to use this as an example but Hyper-V is a step in the right direction for MS and considering the price is a great option depending on requirements. Sysadmins aren't going to drop their favourite tool just for something with a better price point unless management gets the big stick out. Who knows maybe the US economic slowdown will create opportunity for more aggressively priced solutions?


I have to lean more twards the doom theory for VMware, and it is not because VMware is not trying.  I see 2 points that will make it tough to impossible for VMware to be the leader in 5 years.  First, there is way too much competition out there.  With Citrix collecting up XEN, and Microsoft creating Hyper-V, and collecting  SoftGrid, all the big players are saying "Hey you want (Server/App/Infrastructure) virtulization? I can do that!"

Second, since VMware is working building their own kernel and drivers, it forces them to work only with the big players and servers that they choose to develop for.  With Xen and Microsoft, choose any White box, and install.  Drivers are handled by other teams, and the Virtulization developers can focus on guess what virtulization.  

 I predict that with this white hot focus that all 3 companies will see great growth in the next couple of years, but microsoft is just going to price their products so that they are cheaper than VMware, and price them out of the market.  Are they the most techologically superior, no, but that does not make a winner.


If I were VMware, I'd be more concerned about Microsoft than Citrx.  What matters to me is features and today.  Comparitively, what features do VMware, Microsoft, and Citrix have to offer today?  VMware wins by a landslide.  Ok some say, but next year Microsoft will have XYZ features.  VMware, the clear leader in innovation thus far, is given no future development consideration for next year.  Assumptions are made that VMware stays static while Microsoft, Citrix, Oracle, and Novell continue to evolve, catch up, and surpass VMware.

VMware keeps the HCL on a short leash to maintain maximum uptime and minimal outages.  If they share responsibility of kernel and code development with the community, they lose control of major influences that impact uptime and stability.


It cracks me up how everyone is discussing the details around virtualization hypervisors and which is better and how you need x64 this and more features that... in the hypervisors.  Nobody will care about the hypervisor in the future.  The differences will be all based in stability more than anything else.  It will be nothing more than a selection in the bios.  Sure, there might be small nuances that make them better if you like Pepsi vs. Coke... but overall they will be pretty close.

It is all about manageability. 

Microsoft, Citrix, Sun, etc; all have a HYPERVISOR, and a bit of managabiilty around their hypervisor.  VMware has a hypervisor, and a TON of manageability around that product.  They are heavily focused on increasing that level of manageability to provide fluidity in the datacenter and beyond.  They solve problems.  Sure, SMB customers will love Hyper-V or Xen, but anyone with any real infrastructure will be running VMware because of this manageability difference.  Cost also is an interesting conversation right now, as once you buy all the products you need to be at the level VMware is at... you might as well have just purchased it to begin with.

Around quality - as well as VMware's HCL.  Think about Apple.  Why do you think their products are more stable in many cases?  They don't support every peice of hardware, only select, tested, controlled hardware.  Like it or not, that model works.  People seem to be just fine with deploying within that bubble today. 

VMware's products line up with their powerpoints, Citrix has products in the powerpoint that rarely line up with reality in the field.    This is a fact.  I can't stand it.  Underpromise, and overdeliver... isn't that what people always say about expectations?  Citrix does the opposite, and it shows.

Last and /rant.... VMware sure seems to be much more open with knowledge as well as their products than Citrix will ever be.  Citrix loves to control.  VMware does too, however they understand they need to listen as well.


Brian, did you read the label before you opened this can?
Here's my two cents: I have a gut feeling that you're right.  I've watched the mounting attacks on VMWare with some interest. Perhaps it's because VMWare is arrogant, channel-unfriendly, and untogether. (ask anyone who uses PartnerCentral...) Perhaps it's just the baser aspects of us that smells blood in the water. Whatever it is, I see a common thread where people just don't LIKE VMware. Oh, the technology is wondrous, paradigm-shifting, and stabile. But the company's public face resembles a soccer mom's, if  she has four kids, all playing at the same time..."You are VERY important and I'm VERY proud of you!" (Envision a forced, tight-lipped smile, before she runs to the next field).
Yeah, I know, what does LIKING a company have to do with anything?  Perhaps nothing.  Perhaps a lot.
I also agree with the person who spoke earlier about VMWare having an advantage with so many trained partners. Someone has to architect, install, and maintain this stuff. And this stuff is all easy until something breaks...
BTW, the title has a grammatical error on purpose, as tribute to some of the folks before me.

I'm over failed Windows Server restores on raw hardware.  I laugh at  Microsofts early virtualisation trials pushed through their controlled software training environment when connectivity issues were traced to their test servers having duplicate MAC addresses.

IMO, Citrix is getting slower by the release. And a roadshow my CIO attended claimed that Citrix PS 64bit could run any application - yet I launch connections to 32 bit apps published on a 32 bit server farm member from my 64bit published desktops just to get them to run.  And even then the 64bit printer won't autocreate with a UPD in the 32bit app if I happen to patch the 32bit farm member.

I sleep at night because they're almost all on VMWare ESX. It's stable and secure, and my DR works. The thought of these clowns entering this market gives me nightmares.

I agree! VMWare needs to make an application delivery play and they need their own Protocol.  With its own virtual channels with multimedia support.  VDM is not a real protocol and I think they have to at least be looking at this.  It just makes sense.  If they are not, its a problem.  ICA and RGS are what make Citrix and HP stand out right now.

don't forget also that VMWare only have "Server Hypervisor" as a source of revenue. MSFT have a lot of cash cow and Citrix still have Xen-PS-App that is providing a lot of money. They both can put price pressure on VMWare and bring VMW down... It is possible. Remeber Aventail in the SSL VPN market Space. They were a clear technology leader, far ahead form the other and... they went down because of the competition pressure bring on price and R&D efforts.


I have to admit I didn't like VMWare's position on why they don't have a true 64-bit product yet. Basically they said they don't see much benefit in performance by having a true 64-bit product 


As this appears to be the basis for your entire argument, you must have a list of features or improvements that you think a move to 64 bit will produce. Could you explain what you think will be better with a full 64 bit release?

 as opposed to their current 32-bit product that just emulates 64-bit.


Okay. Now you are just getting silly. VMware does not emulate 64 bit. If they did, why would VT hardware be a requirement for 64 bit VMs?

Just to clear up any misinformation you might have received - The VMM (which is the bit of software that is CPU-facing) has been 64 bit since 2004, so no 'emulation' is required. The hypervisor may be 32 bit but since that it just a time slicer, and has nothing to do with interraction between the VMs and the hardware, it makes no difference whether it's 32 or 64 bit. 

So a true 64-bit product isn't a high priority for them to produce. 

Why would they make something a high priority that is going to provide no benefit?


VMWARE & SUN and their new launch of their take on VDI....

They have a secure global desktop and now this.  High end servers that eat little power, and let you max out their use...

Why the he"double hocky sticks" would you, after looking at this, subject your company to the pains of Citrix??? 

VMware is to server virtualization what BEA Systems was to application servers. Early leader and innovator, strong customer adoption of the technology, and technical advantages across the board. But in the end BEA was doomed due to non technical issues: poor management, poor channel business model, price (i.e., margin) pressure from multiple sides, and inability to market the brand beyond IT admins. In the end, you can still use WebLogic ... you just need to call a different sales rep.
I think the best example of the 64-bit issue is why Marathon Technologies Everun product won't run on VMware. Everun won product of the year at VMworld 2007, but if you ask Marathon Techonologies why they wrote such a great product and failed to produce it to run on the product that has 3/4 of the server virtualization market, they will tell you flat out it's because VMware is still 32-bit. The point is today the vast majority of VMware customers couldn't care less that it's only 32-bit because there's not a whole lot of benefit of it being 64-bit. But when we're looking at what the future looks like 5 years from now, do you really think people will still not care that VMware's product is still 32-bit? Marathon, Egenera, and some other companies innovating in the server virtualization space, are proof that 64-bit is important when it comes to the innovation of the technology and what new beneftis will exist in server virtualization 5 years from now, even if customers don't benefit from 64-bit today. Until VMware has a 64-bit product, their innovation has a bottleneck. When the topic is "Will VMware still exist in five years", that bottleneck is certainly an issue of concern even if it's meaningless to how the product works today.
I don't think anyone is saying that it's the existance of Xen that could make the industry leader cease to exist within 5 years. If VMware ceases to exist in 5 years it would take the entire industry ganging up on VMware to take them down, as opposed to the existance of one competitor. Given the way VMware operates, the entire industry has motivation to do just that. 

I completely agree!

E.g. CPS4.5 did't work right before R01 (and still doesn't in some areas)

But in the end, marketing always wins... See for instance the demise of Word Perfect and Novell (as an network OS). Microsoft won, because of clever marketing, not because of better technology.

So I guess I really hope VMware gets their marketing act together! 


Unfortunately, I've been consulting with all three products M$, Citrix, VMware but IMHO, VMware is the strongest player in the virtualization world due to its product robustness and maturity and customer awareness.  Fortune 500/1000 cannot afford to take VMware off their virtualization solution due to 1:20 virtualization ratios and how much $$ will saves them.  Frankly speaking, none of the products out there can do better than HA, DRS, VMotion enterprise class solution like VMware.  We all know Xen Enterprise is still far behind to match with VMware and work better with Linux/Redhat environment. Microsoft virtualization product is really crappy especially Virtual Server 2005 although M$ is going to expand the virtualization much stronger due to financial stability.  That doesn't mean VMware just sit their and watch their competitors growing.  VMware has the same kind of people and geniuses out there working for them as other companies would.  I think VMware will be stronger with other products their trying to release is the Site Recovery Manager and vmSafe but doesn't prove it's going to work as expected.  VMware should definitely openly to partnerships and marketing programs to attract more sales leads.  One thing I like about VMware VMTN forum is really good informative and sometimes quicker to post questions there than placing VMware support ticket.  It would be nice if Google has to team up with VMware on marketing and R&D efforts that will definitely knock of Citrix/Microsoft aggression.  I could see Sun and Oracle VM will be hitting the market as it mature but Oracle is a huge company they can invest and join the virtualization market easily and should adapt quickly. Key features to competitive products are performance, management, interoperability, and scaleability.  Within next 5 years, web base applications going to be huge and if VMware would implement full blown Virtual Center Server management inferface 100% with web base and securely access via internet that would be a plus.  Believe me, VMware products still have some crappy issues that system engineers like us hate about just day to day operation and tidious things paranoid us but besides that I can't see any products out there match VMware features and stability. 



Toatlly agree with you, I have not seen the Sun VDI Software 2.0 nor do I know much about their protocol. A good protocol that can be used over WAN is exactly what they need to be able to play a big role in that market space. But this tells me that VMWare is not sleeping either and is competing. I hope we'll still have VMWare around in the next 10 years with great solutions that are easy to manage.



Just to add some information where WMWare is going http://www.sun.com/datacenter/consolidation/virtualization/vmwarevdi.jsp




but if you ask Marathon Techonologies why they wrote such a great product and failed to produce it to run on the product that has 3/4 of the server virtualization market, they will tell you flat out it's because VMware is still 32-bit.

Do you have a link to any statements from Marathon to that effect?

It's got nothing to do with ESX being 32 bit and everything to do with the fact that VMware won't support ESX if you go installing 3rd party apps within the service console - which is what would be required for Everun to operate.

Aside from that, the indications from VMware are that the service console will disappear altogether, meaning that Marathon would be wasting their time developing for ESX as they would have nowhere to run the app anyway.

Nothing at all to do with 32 vs. 64 bit.

I've seen post on Expand Networks people has WAN optimization that boost 1000% faster that would be surprisingly to anyone out there, but its that good VMware might already snatch it like Thinstall to put it on their product line to compete with Citrix/Microsoft. 

Your claim is crazy talk. Do you really think Everrun isn't available for ESX only because VMware won't support ESX with Everrun on there? If VMware customers wanted to buy Everrun and run it on ESX, Marathon wouldn't care less if VMware stopped supporting the customers. It's a sale for Marathon. They would sell it to the ESX customer in a heartbeat.

The reason Everrun isn't available for ESX is because nearly the entire thing would have to be rewritten to run on VMware. And that is a task so big that it's not worth the effort for Marathon to do, even with VMware's huge market share. 


Your claim is crazy talk.

We'll see......

Do you really think Everrun isn't available for ESX only because VMware won't support ESX with Everrun on there?

It's a major factor, yes.

 If VMware customers wanted to buy Everrun and run it on ESX, Marathon wouldn't care less if VMware stopped supporting the customers. It's a sale for Marathon. They would sell it to the ESX customer in a heartbeat.

Do you really believe that enterprise customers are stupid enough to put themselves into an unsupported configuration? Marathon are well aware that customers (for the most part) know what they are doing and employ proper risk management techniques. Anyone doing even the slightest bit of research into such a product would rule it out straight away.

Why do you think ESXRanger was re-written for VI3 and taken OUT of the Service Console?

The reason Everrun isn't available for ESX is because nearly the entire thing would have to be rewritten to run on VMware. And that is a task so big that it's not worth the effort for Marathon to do, even with VMware's huge market share. 

If that was the reason (which I don't believe it is) they would have written it for ESX to start with and ignored Xen. Why target a product that has <10% of the market when you can get to >80% by developing for another platform?

IMHO, it's not available on ESX for 2 reasons.

1. It would never be supported and Marathon know that customers are not stupid enough to risk their datacenter on an unsupported platform. (you obviously think customers are stupid.....)

2. The Service Console is going the way of the dinosaurs. They would be developing for an environment that won't even exist in the near future.

It's got absolutely nothing to do with the 32 vs 64 bit garbage that's being dragged up by some of the sillier people posting here. Developing for a 32 bit platform is no more difficult than developing for 64 bit.


Yes, I'd love to hear your opinion after visiting all three.

VMware is protected by patents and is the ONLY, I repeat ONLY virtualization platform that runs on bare-metal, truley independant of a general porpose OS like, well... lets see... EVERYONE ELSE.

If you fall into the whole "Vmware is expensive" camp, you are totally missing the ball here and need to look at where the company is headed. I am a VMWare partner under NDA with and have been using betas of MS and VMware, and can tell you that the true value proposition of what VMware brings is managebitlity, and lower cost per VM, speaking of... you may be interested in taking a look at this blog:


And this doesnt even TOUCH on what they do with SRM and VDI, to forces to be reconed with in late 08/early 09. 






Uh... VMware's "linux source code", that's news to me.

I've been a Vmware admin for a few years now and the only thing linux is the service console, which they used to boot-strap ESX before they went to 3i, essentially getting RID of the service console.

I support the idea of an open source community, but VMware has invested billions in R&D and have more PHD's working on virtualization right now that MS and Xen combined, or so I hear.


"Do you really believe that enterprise customers are stupid enough to put themselves into an unsupported configuration?"

If the additional fucntionality that Marathon provides is something they need, than yes. Call it stupid, or whatever you want, but to Marathon a sale, is a sale, is a sale. Obviously this wouldn't appeal to the entire VMware customer base. But there would be enough willing customers to be profitable to Marathon. And given how many more VMware customers there were than Xen customers when Marathon was developing Everrun, the potential profitability of writing it for VMware was far greater than writing it for Xen. Marathon's decision had nothing to do with VMware's willingness to continue supporting customers who bought Everrun. They would have developed it for VMware if it could have been done.

"If that was the reason (which I don't believe it is) they would have written it for ESX to start with and ignored Xen. Why target a product that has <10% of the market when you can get to >80% by developing for another platform?"

That entirely the point that is being made here within the context of what the industry will be like five years into the furture. If you were the CEO of Marathon, and the engineers were devoloping a solution so appealing that it can win "best of" awards at VMworld, but your engineers tell you it won't run on VMware, they better have a damn good reason why it won't. If the reason was VMware wouldn't support customers, the CEO would say "build it anyway" because everyone is on VMware and no one is on Xen. The bottom line is lack of VMware support is not the reason the engineers gave the CEO. The reason is VMware places very little importance on replacing the old code in their hypervisor, so developing Everrun to run on VMware is a huge challenge compared to how easy it was for them to develop it to run on new 64-bit hypervisors. If you seriously don't believe me, then ASK THEM!!! They get asked it all the time given all the VMware customer that exist who would like to run their stuff but can't. This isn't some private inside information that I'm sharing here. This is what they tell their potential customers. 



If the additional fucntionality that Marathon provides is something they need, than yes.

You have clearly never worked with enterprise customers.

developing Everrun to run on VMware is a huge challenge compared to how easy it was for them to develop it to run on new 64-bit hypervisors.

Utter garbage.
*If* that is what they are telling potential customers (which I doubt), it sounds like it would be coming from the sales/marketing people - NOT the developers. As I already stated - developing code for a 32 bit platform is no more difficult than devloping for a 64 bit platform.
It sounds to me like they are trying to come up with any excuse, however unlikely, for not having their product available for ESX. That's if they are making such claims anyway. They have made no public statements to that effect and the only place I've seen such a claim being made is here.

If they are, in fact, making these claims - the next time they make it, ask what functionality they think is too difficult to achieve running with a 32 bit kernel.

If you're only hung up on the 32-bit/64-bit attribute, then you've completely missed the entire point on how it relates to where VMware will be five years from now. When 3rd party add on companies, like Marathon and Egenera, are putting out their innovations on the new hypervisors well before VMware's hypervisor (or not at all), then that speaks volumes as to where innovation in the industry will take us 5 years from now. We can argue the specifics of the reasons that these 3rd party companies are doing this, but it's the fact that reasons even exist that's pertinant to where innovation is going to take us. If 3rd party companies have any reason at all to delay VMware supported versions of their product, or not release VMware supported versions of their product at all, then that's a liability on VMware's part regarding their future. It's pretty much a given fact that VMware has a reputation of not working well with other companies. Pretty much the only people that herald VMware are their customers. The rest of the industry could pretty much tell VMware to go to hell.

Personally I work with Citrix, VMware ESX and Microsoft technologies on a day-to-day basis, I've also installed and supported customer environments based on Xen (And XenSource) and hold the relevant bits of paper on all of the above.

VMware won't die in 5 years, 10 years maybe, in 5 I'd expect to see Citrix struggle much more. The feature creep in 2008 / Longhorn Terminal Services is the beginning of the end for Citrix, they've simply been overtaken. The last 2-3 years was the chance for Citrix to show that they could still add value in the Enterprise thin application delivery market, they blew it. We still have a WinFrame product (or MetaFrame / MetaFrameXP / Presentation Server / XenApp - whatever you want to call it it's basically the same) which is still flaky (and in many of the same ways) as when I started working with WinFrame 1.6 all those years ago.

VMware is different, they have several features that no one else is even close to being able to match. Features like memory overcommit, VMotion (obviously), Storage VMotion in ESX 3.5, DRS and HA mean they have a massive lead over all the other players, most of which won't get close to having them production ready for 2 years. VMware already have them IN production for large enterprises today.

Eventually Hyper-V (and more importantly, the management tools) will mature to catch up, and if VMware stay still now (the way Citrix have for the last 2-3 years) then they'll find their market dissappears into the Microsoft collective and they'll be left as a niche has-been (Novell anyone). I for one hope that doesn't happen, they've brought a lot of cool toys to the enterprise space way before anyone else.

Having access to some of the technical EMC and VMware guys means I've been lucky enough to see what some of the plans VMware have coming up involve, if they can deliver on them then they'll be maintaining a very strong lead for some considerable time yet - and based on experience VMware is one company I would say have a very good chance of actually delivering.


I have sat in on meetings within the last month about new products under NDA for VMware, and they're DEFINATELY going to stay FAR ahead of the rest of the vitrualization pack in their core competency of server virtualization.  They have newer features that are already working in beta, and are production ready for enterprise customers.  Add to that the promising purchase of thinstall, and I think they will be here for another 20 years.

Citrix has the problem now of consolidating all the companies it has purchase in the last 3 years into a logical combined product set, something they have yet to do at all.



Agreed on Server stance, Thinstall FORGET ABOUT. That product is so limited in the real world it's not funny.

Wow, what a blatant piece of FUD about Vmware from someone i thought was an ok blogger. very depressing i must say...

What's next headlines like "Nobody ever got fired for buying Microsoft so wait for Hyper-V"?