The x86 virtualization market is an important part of the modern data center, and while growth may no longer be "explosive," it is still more than respectable. For the better part of a decade, VMware owned this market outright and expanded its products to build an entire VMware ecosystem. It would have been foolish of anyone to expect VMware to retain market dominance; however, Microsoft has been able to make astonishing gains in a very short period. As server consolidation reaches the point of satiation and cloud becomes the new driver, everyone in the x86 virtualization market must adapt.
Cloudy with a chance of hybrids
Microsoft has done a particularly good job of selling the cloud with its Cloud OS marketing campaign. The concept behind Cloud OS is straightforward: Create virtual machines (VMs) on your own infrastructure, which can be moved to the infrastructure of a third-party service provider, or to Microsoft's Azure cloud.
Of course, more than just raw VMs can move from A to B. Microsoft prefers to focus on services, or packaging the operating system, application and virtual configuration. A lot of Microsoft's effort is going into tightly integrating the hypervisor, OS and app. This removes the need for many kinds of emulated hardware. Microsoft is focusing on total management: the ability to manage infrastructure, OS and application in one place. They're not there yet, but the direction is clear. Microsoft sees management of an application or a workflow as the logical atomic element of administration.
In contrast, VMware is very much an infrastructure company. VMware's new vCloud Hybrid Service remains an infrastructure play -- and a limited one at that. Service providers do form part of VMware's strategy; however, VMware has not gotten the pricing right. VMware is far too expensive for service providers, especially when one considers that these providers are now competing against VMware.
In terms of OS and software support, VMware's approach is, at best, inconsistent. VMware offers some integration -- tools such as automated customization during deployment are useful -- yet others seem painfully overlooked. If VMware is to be the single pane of glass to manage infrastructures, it will need to tie into update mechanisms for OSes and line-of-business applications so that infrastructure operators can manage fleets of VMs without flipping back and forth between tools from different vendors.
Which tool is more important?
For Microsoft to win this war, it needs to commoditize the hypervisor and infrastructure management tools. This makes the application and OS the focus and the element of import. For VMware to win, it needs to make handling OS and application maintenance so simple that the focus becomes increased infrastructure automation as a driver for business efficiency.
Between the two of them, Microsoft is currently doing a better job. None of its management tools are particularly good, but they are growing by leaps and bounds with every generation. Microsoft is making the infrastructure part of the equation just simple enough that it makes good sense to turn to OSes and applications inside VMs to accomplish the real automation and magic that the cloud promises. VMware, on the other hand, continues to provide little more than lip service to caring what's inside the VM, while charging full steam ahead with an enormous pile of infrastructure automation tools.
In the Microsoft world of five years from now, I can clearly see being able to move services from my local cloud to a service provider and on to Azure without fretting about the infrastructure underneath. My job as a systems administrator will be to make sure that the OS and applications are properly maintained and automated, nothing more. When I look at the VMware world of five years from now, I see myself having to maintain and automate the OS, the application and the infrastructure.
VMware's two biggest advantages
VMware has two things on its side: It tends to make better management tools than Microsoft for any given task, and corporate inertia is practically a force of nature. Companies around the world will buy VMware because companies around the world already buy VMware. There is something to be said for the first-mover advantage. It might be worth asking a Novell executive whether that particular advantage lasts forever. If VMware is going to beat Microsoft over the next decade, it must do more than milk its existing customer base. VMware will need to evolve beyond an infrastructure company and into a full-service data center management outfit.
Neither of these companies is particularly good at listening to its customers. They both have a nasty tendency to look at the middle of the curve and assume one size will fit all. Microsoft -- and its outgoing CEO -- learned this lesson rather publicly and rather painfully. The possibility exists that the incoming CEO will learn from his predecessor's mistakes. Can VMware adapt in time to shake off its infrastructure-only roots? Of this, I am unsure.
Let us consider a practical example of change management in action. VMware administrators are facing a user interface (UI) change. First introduced with vSphere 5.0, the third major iteration of the Web client -- introduced with vSphere 5.5 -- isn't exactly winning over sysadmins. A poll run by vExpert Gabriel Chapman reinforces the general chatter on social media and amongst sysadmins regarding this change.
As of early February 2014, 187 votes had come in and 153 votes -- 81.8% -- were still using the Windows desktop client, either exclusively or in concert with the Web client. This means that 2.5 years after the younger, hipper Web client was introduced, less than 20% of respondents have abandoned the now-deprecated desktop client. While not exactly a Windows 8 level of underwhelming, that's still alarmingly slow uptake.
Features new to vSphere 5.5 are only available in the Web client. Rumors have already been doing the rounds that the next version of vSphere won't support the Windows client at all. Even if the execution is stayed for the 2014 release, the death sentence on the VMware desktop client will eventually be carried out.
Overcoming administrator antipathy will be important for VMware. Microsoft is hot on VMware's heels, and its technology is as good as, if not better than VMware's in almost every aspect. VMware's two biggest advantages right now are corporate inertia and the fact that it has the far-less-frustrating UI.
The hypervisor is a commodity
While VMware says the hypervisor is not a commodity, its marketing documents about why you should buy its stuff instead of Microsoft's rely on total-cost-of-ownership arguments that talk about decreased operating expenses if you choose VMware.
A goodly chunk of those arguments boil down to "VMware is easier to use," but does that argument hold up if any new features are isolated in a Web UI that no one wants to use? If administrators end up disliking the administrative tools from Microsoft and VMware equally, that leaves only corporate inertia as VMware's selling point.
Microsoft's corporate hubris meant it was unable or unwilling to listen to administrators, beta testers and its own customers. It stayed the course with Windows 8 against the advice of virtually everyone and nearly obliterated the PC market in the process, letting tablets take over and earning the enmity of end users who are sick of Microsoft delivering products they don't want.
VMware cannot afford to have a Windows 8 moment with the Web client. It needs to swallow corporate pride, admit there's a problem and set about resolving it. It needs to talk to administrators from all walks of life and find out why they aren't embracing VMware's vision of the future, and what needs to change. This is only one example, but it is an important one.
Tactical versus strategic thinking
Right now, VMware is not only doing well, it's doing amazingly well. To question its strategy would seem to be madness. This is the IT industry, however, and empires can be built, fall and be built again all in the span of just a few years. VMware's short-term successes are providing it with a sizable war chest, a vast customer base and a great deal of respect.
As a company, Microsoft plays the long game. It hires people who think strategically and makes plans that stretch out over years. These plans don't always work, but when they do, they tend to completely obliterate its opponents.
Sysadmins have long memories and hold grudges. Being less frustrating than Microsoft shouldn't be a particularly high hurdle for VMware to jump. Whether or not VMware has even realized that this needs to be a perpetual long-term goal remains an unknown.
The future of the virtualization industry is now as much a battle for hearts and minds as it is one of price or features. The war of the clouds is VMware's to lose.
This was first published in February 2014