Is VMware the power it once was?
World's changing. Is VMware adaptable enough to cope?
Sysadmin Blog It is commonly held that if an article asks a question in the title or lede, you can safely answer no and avoid reading the body of the article itself. While not always true, the aphorism is accurate enough to be considered reliable.
With that in mind, I have been asking myself, "is VMware the focal point of power in IT that it once was?" Given the role that VMware plays in our industry, it seems worth writing about to see if I can answer it. Even if only for myself.
In the purest and most pedantic form the answer is obvious: VMware isn't what it once was because the world is constantly changing, as is everything and everyone that occupies it. We could have a lovely discussion about entropy, or talk about version numbers and feature growth, but that is getting distracted from the forest by obsessing over counting beetles.
VMware's power over our industry has traditionally been demonstrated by VMworld. VMworld has grown to become one of the largest and most important conferences in all of IT. Companies launch at VMworld. Careers, partnerships and multi-billion dollar deals are made there. Entire teams of staff are bought and sold in meetings at that event and for the better part of a decade the business of doing business in IT has occurred at VMworld.
So many partners crowd the Solutions Exchange at VMworld that it has developed communities, each with its own character. There's startup alley, which is brightly lit, overpopulated and filled with salesmen desperate for leads. There's the badly lit also-ran ghetto where the booths are miniscule and the vendors aren't seeking leads so much as reminding us all they exist.
In the middle are tier 1 booths, each a world unto themselves, flanked on either side by the middle-class vendor suburbs. Here you'll find established vendors and wealthier startups. They are more ostentatious, more well-rehearsed; they've drilled and trained, and meeting with any of the executives present requires having set up a meeting months in advance.
As the storage wars unfolded across the industry, so too did they dominate the floor of VMworld. Software defined networking, containers and DevOps all rippled throughout the event; the booths and deals, meetings and discussions echoing the broader industry framework to which we have been collectively welding our futures.
Does VMware still hold the data center together?
Over the course of the past decade, VMware has been a sort of virtual glue. Their software has held data centers together, being the practical point of intersection between servers, storage, networking and operating systems. VMware's hypervisor and associated management tools solved so many very real and very practical problems that if any vendor wanted to sell their product to either the mass market or to the enterprise, they needed to work with VMware.
While I certainly am not going to stand up and say VMware's influence has been exhausted, there are worrying signs that it is being marginalized.
The item that triggered this introspection for me is the switch of VMworld from the Moscone Center in San Francisco to the Mandalay Bay Center in Las Vegas. I'm personally not fond of the move; I loathe Las Vegas, especially in August.
I did, however, figure that I would be in the minority with that opinion. To my complete surprise, rather a lot of influential individuals (and several vendors) are openly discussing not going to VMworld 2016. The rationale behind these decisions seems to be deeper than an aversion to the heat.
Almost universally, those considering not attending have said they simply aren't sure it's worth the time and money. These same individuals and vendors are planning to go to OpenStack in August, and – perhaps more critically – they are also planning on attending Amazon's re:Invent at the end of November.
When I started asking pointed questions, I got a shockingly uniform answer: if there was only money enough to attend one conference every year, nearly everyone I've asked thus far would choose re:Invent. Two years ago, those exact same individuals would have answered VMworld.
A picture is slowly emerging of the future of IT. In the middle is Amazon, or at the very least software and services built on top of Amazon. Around that lies everything else; a web of public, private and service provider-hosted workloads radiating out like a spider's web.
This is a future where VMware could still be the glue that binds together the data center, however, that data center is but one point along the edges of tomorrow's IT design.
If Amazon's usurpation of VMware's crown as IT's focal point weren't worry enough, real threats to VMware's dominance are emerging both from Microsoft and from the open source world. Microsoft's Azure Stack, and the associated Azure-in-a-can appliances, are flat out better than anything VMware or its myriad partners can bring to the table. Significantly more functionality for less cost that's easier to use. Oops.
Despite a lot of nay-saying, KVM hasn't gone away either. It is proving increasingly popular, especially among hyperconverged appliance vendors. These vendors are building Azure-in-a-can competitors, and they're doing so with KVM and other open source options at the core of their offerings, not VMware. That does not bode well for VMware at all.
So far, my purely anecdotal evidence says that VMware may be in serious danger of becoming less than it once was. But VMware is by no means doomed. They have many of the technologies required to meet the challenges ahead of them.
What remains to be seen is if they have the vision to put them together in an easy-to-use fashion, the will to price the whole thing competitively, and the foresight to do so before it's too late. ®