Virtualization Wars, what can VMware learn from the past?

A colleague of mine referred me to an article in the Economist titled "The rise of the hypervisor" ; a story that paints the picture of the VMware David taking on the incumbent Goliath's of Microsoft, Linux, and Sun for the top spot of owning the core software that controls IT compute hardware.  The last paragraph of the article carefully avoids making any long term predictions.  This is where I'd like to venture; based on history.

As many of you know, I spent a number of years working for Novell, through its rise and fall from the grace of the IT community. My perspective from the inside is unique. But as I watch VMware grow and competitors begin to take notice and join into the fray, I find myself witnessing a déja-vu.  Lets turn back and look at what happened to Novell's NetWare product:

You must realize that Novell's wildly successful NetWare product of the late 1980's and early 1990s was filling a void in the Microsoft eco-system.  Granted, Microsoft had a LAN networking solution allowing for file and print in those days, but LAN Manager was poor compared to NetWare. LAN Manager didn't scale, and performance was so poor, it really wasn't usable for anything more than a handful of desktops.  Microsoft restarted from scratch, and built Windows NT.  While its first versions of the NT operating system were ridiculed by the market, and of course by Novell too, Microsoft persisted.  It took Microsoft just shy of a decade to unseat Novell from its throne in the x86 server operating systems market.  How did Microsoft do this?  Simple.  It focused on the whole eco-system: the whole meal from soup-to-nuts, where as Novell's NetWare only offered a fraction of the IT solution, granted a critical component, but still it was only a fraction.  By Windows NT4, Microsoft had reached the point of "good enough" to where more customers wanting the whole meal could get it from the one vendor.  With Windows 2000 Server, Microsoft added Active Directory to increase the competition.  The clincher was Microsoft leveraging Exchange 2000, its back-end email server for the Outlook mail client that was already available to nearly 98% of all client desktops; since Outlook is part of the Microsoft Office package.  Exchange required Active Directory.  This was the nail in the coffin for NetWare and Novell's eDirectory as solutions filling in the voids in the vast Microsoft eco-system.

So how is history repeating itself, you may ask?  So in late 2004, Microsoft acquired the virtualization assets of Connectix and launched Microsoft Virtual Server 2005.  By comparison to VMware, it doesn't scale and performance is so poor its not really useful to the broad market.  Hmmm - Sounds like LAN Manager doesn't it?  So Microsoft has restarted from scratch and is launching Hyper-V, its new virtualization solution.  At this point, it's a promise, and from Beta reports appears to be akin to Windows NT first releases: better than before, and good enough for a broader market, but still not as rich and mature as VMware's ESX server.  So what will the clincher be?  Microsoft is already showing that it is integrating its virtualization solution with its whole course meal; integrations with its broad menu of products.  On January 21st, Microsoft announced it's virtualization strategy. The key points are that Microsoft is integrating with its desktop monopoly and its other "virtualization" products, such as terminal server, to build a broader solution.  The "good enough" product plus Microsoft's developer community will make this a very difficult beast for VMware to deal with.  Hmmm - Sounds like what happened to NetWare when Microsoft integrated Exchange with Active Directory.

Right now it seems that VMware is invincible (and I can remember back when it seemed Novell was invincible).  The fact remains: VMware is only offering an entrée out of the complete IT meal.  It is filling a void in the market, and Microsoft is going to come back with integrated solutions that are good enough and integrated with the rest of its meal to woo customers over to its solution.  This won't happen overnight.  Microsoft didn't push out Novell overnight.  So what can VMware do?  VMware needs to build out its menu.  It needs to offer soup-to-nuts IT solutions.  And it needs to do this with disruptive technology such as SaaS, not with the traditional methods that Microsoft already has wrapped up.

VMware has it in its power to not let history repeat itself, but it's a mighty mountain to climb!

Post note:  Another colleague pointed out that NetWare ends in "ware" just like VMware does.  Hmmm.

Post note 2: Novell learned from its experiences and now has a sound strategy around Linux to offer soup-to-nuts solutions, but the question everyone asks is if it can continue to execute on that strategy for the long term?

[Post by Richard Jones]

Comments

[Tarry Singh]

Good points!

You should join our Virtualization Linkedin group as well. Some of your ex-colleagues are there as well ;-)

-Tarry

[Rey Poster]

VMware has an opportunity to capture more marketshare by working with software vendors to create virtual appliances, i.e. prepackaged VM's with OS and configured application to add directly to a customers VMware cluster.

VMware would do well to treat it's customers better by offering scaled pricing on maintenance and figure out a way to combine sales and pricing advantages with other products from mothership EMC.

My argument on the maintenance cost is that it doesn't take a linear effort on the part of VMware (or any other software vendor) to support 100 licenses vs 10 or even 1. So why is the pricing on maintenance linear??

[John Gilham]

VMWare is overrated in the long term.

As virtualization is now becoming a commodity product, VMWare's business model allows instant portability to any competing platform (VMWare to MS/Xen's VHD. That's the flaw in VMware's business...one that's beginning to look a lot bigger now that there are two other vendors with mature products in the market.

With hyper-v (and other vendor specific VS OS's) most likely to be embedded in server hardware (i.e. flash boot) in the near future...what real value does VMWare provide now that Microsoft supports dynamic clustering?

This makes the initial high cost of VMWare's solution and ten ongoing support contract costs pointless when I can convert VMWare to VHD in 10 minutes.

[Ant Stanley]

The company I work for, a VMWare and Microsoft partner (amongst others), has adopted the strategy that there is an 18 month window that VMWare will still be a viable product offering... after that window is clsoed, there is an expectation that Virtualisation technology will be a commodity product and virtually free in all forms...

thats not a major issue for us, as we find we generate more income from the accompanying hardware sale ... VMWare is invariably sold with hardware at a ratio of around 7:1 (hardware vs VMWare), and the associated services around virtualisation/consolidation assessments, installation, and migration projects...

as long as we maintain key and rare virtualisation skills (whether they be Microsoft or VMWare skills) we'll still have a competitive advantage..

PS. we're also a Novell partner :)

[Brent Hughes]

If your expectations dictate that all you need is enough to get by then that is what you will get. I have been in the IT industry for a long time. My finding is that people turn to Microsoft when they want it cheaper or are forced to do so. Microsoft is not going away and the ball is in VMWare's court. VMWare provides a great product hands down and if it were free there would not be a need for this post. However; with the market in shambles companies have to do more with less and enough to get by. ENTER MICROSOFT again...Product is cheap and provides enough to get by during times of recession. We all do what we must and keep on going.