The Cranky Admin
What VMworld 2017 Says About VMware's Future
The company's successfully pivoted; but its new direction must be navigated skillfully, and potholes are everywhere.
After a year of mixed announcements, it's worth taking a moment to analyze VMware. The brief pause between VMworld U.S. and VMworld Europe feels like a great place for such introspection, if for no other reason than that it orders our thoughts for questions to ask at VMworld Europe. So just what exactly is VMware up to?
VMware has come a long way in just a few years. VMworlds 2015 and 2016 were disappointments, as was the lack of any concrete progress towards something resembling a functional vision of the future. Competitors were stealing marches on VMware left and right, and things were looking a little grim.
This all started to turn around with updates to VSAN earlier in the year. These accompanied VMware's "proactive" HCL.next approach to hardware compatibility lists that signaled a shift from VMware's traditional role as a passive infrastructure provider to a Microsoft-inspired service provider.
VMware, it seemed, wants us all to start relying on them. Rather than merely giving us some of the tools with which to build our own datacenters, it started to become clear that VMware wants us to think of them as helping us run our datacentres, day to day. This message would only be reinforced as the year progressed.
By the time VMworld 2017 had ended, the year had been flooded with numerous announcements, including critical updates to vROps, the incorporation of Wavefront and Pulse.
For the past several years, VMware has been simply a virtualization vendor with benefits. VSAN gave it a storage platform, but one that was only relevant in the context of virtualization, and eventually containers. And containerization is really just virtualization without benefits.
NSX, Photon, the various End User Computing (EUC) products, the analytics products... all of it is nothing more than ancillary support infrastructure to prop up the core virtualization vision. Old VMware would sell you software for you to run as many workloads on as few servers in your datacenter(s) as possible. VMware would help make storing, moving and launching those workloads as easy as possible, but its interest ended where the operating system began.
This conceptual barrier is important. VMware was nearly dogmatic about it. It invested heavily in making everything below the operating system as easy and efficient as current technology would allow, but it emphatically did not want to get wrapped up in deploying, managing or maintaining operating systems or applications. That was someone else's job.
Unfortunately for VMware, though, Amazon didn't go away. Microsoft, for its part, was succeeding in bullying its customers onto Azure. It was even having some success convincing its own channel to commit suicide. Just plain old infrastructure wasn't working, and by the time of VMworld 2015 it had already become clear that many of VMware's competitors were well ahead of it at things cloudy.
VMware failed miserably at its own public cloud endeavors. It couldn't make an on-premises cloud that was easy to deploy or use. Their initial attempt at hyperconverged appliances, EVO:RAIL, was something of a disaster, though VXRail has allowed VMware to at least play catch-up.
Most critical among the announcements to date has been VMware's unveiling of the long-rumored AppDefense and VMware Cloud on AWS.
VMware Cloud on AWS represents VMware's outright capitulation on clouds. It could not beat the competition, so it elected to join them. In this case that meant cozying up to the 800-lb. gorilla of the public cloud marketplace, and nearly giving away their software; while details of the cut VMware itself is getting from VMware Cloud on AWS deployments aren't public, it can't be much.
The prices as marketed by VMware are roughly at parity with the cost of native AWS instances. VMware is also pitching VMware Cloud on AWS as being less expensive than running those same instances in a datacenter you designed and control. If true (which I sincerely doubt), this would eat into VMware's on-premises licensing sales in a big way.
If VMware Cloud on AWS actually proves cheaper than designing and maintaining one's own datacenter, we'll see American organizations adopt it wholesale and in a right hurry. Consider that the bulk of the large volume customers VMware has are American enterprises and government customers. Americans don't have the same concerns about privacy that, for example, Canadians or Europeans have. They don't have to worry about putting their data in the public cloud because their laws don't give them (or anyone else) any privacy anyways.
VMware needs to have a cloud solution if it's to survive. There's no question about that, and marrying AWS is probably the smartest long-term strategy. If, however, I am correct, and the AWS-backed solution results in smaller margins for VMware than the on-premises version, VMware is going to need to make that up elsewhere.
Security Cash Cow
AppDefense represents one means by which VMware can do this. The virtualization market and all related markets (storage, networking, ECU, cloud, analytics and so on) are experiencing strong commoditization pressures. By comparison, the security market is exploding, and probably will be a cash cow until the end of time.
Virtualization, storage and the other infrastructure-related technology markets are all about building a better mousetrap. Eventually, you see convergent evolution; all the solutions from all vendors more or less look the same because function dictates form. The most efficient and pragmatic way to do x means y needs to look and behave like z. Once we've reached that point, every subsequent development is incremental. There's no real money to be made there.
Security, on the other hand, has nothing to do with logically iterating designs towards some platonic engineering ideal. Security is all about monetizing the fact that humans are lazy, arrogant, cruel, vicious, prideful, willfully blind, greedy, creative, curious and miserly. We're horrible, horrible beings that never really collectively evolved from the sort of monsters that fry ants with a magnifying glass.
Except the ants are organizational networks and the magnifying glass is the entire Internet itself.
There's a lot of money to be made protecting one group of people from another. There's even more money to be made protecting us from ourselves. There's crazy, stupid, military-industrial complex money to be made if you can invent the specter of terrifying, but vague and unknown, bad guys.
If you're a trusted household name that's on the preferred vendor lists of governments and enterprises, you can add the word "cyber," ride the hysterical wave of "us vs. them" paranoia manufactured by the physical security industry over the past century, and just print money. Truth be told, security looks like a fantastic move for VMware!
While VMware Cloud on AWS and AppDefense both look like they could be the salvation of VMware, there's lots of room remaining for VMware to shoot itself
in the foot. Recent comments by CEO Pat Gelsinger are especially concerning.
Combining the selected quotes available in the tech press from VMware executives about AppDefense with my own discussions with VMware employees, I'm left with a sense that VMware expects to roll into the IT security industry and slurp up the bulk of the market share and most of the profits simply because they're VMware. Larger, more experienced companies than them have tried. All have failed miserably.
IT security is nothing like infrastructure. You can't simply tell everyone that they will now do things as you dictate, and only as you dictate, and expect to win. This is not only the basis of AppDefense in its current form: "allow nothing unexpected to occur," but it's also present in comments calling for the dramatic reduction in the number of security vendors on the market.
Those security vendors exist for a reason: there's more diversity in IT than any one person can possibly imagine, let alone truly comprehend. This is the glory -- and the terror -- of our species. Give us a new tool -- in this case computers -- and a generation or two later we'll be doing so many different things with that tool that entire industries will exist just to support individual implementations.
Readers of Frank Herbert's "Dune" might understand this more completely if I referenced The Scattering. Humans don't react well to monoculture.
The reason this is an area of concern is that VMware is now a mature tech titan
. They are a powerful force in our datacenters, and they could, if they wanted, take a page out of Microsoft's book and try to leverage a near monopoly in one area in an attempt to create one in another.
If VMware tries this, it will end badly for them. Microsoft gets away with treating its customers and partners poorly because Windows is legitimately next to impossible to get away from. Few, if any companies could afford the real-world cost of rewriting all their applications for an alternate platform.
VMware isn't that hard to get away from. VMware can be replaced relatively easily by any number of competitors. There is a certain amount of friction that makes sticking with VMware worthwhile for most organizations, but this should not be confused with burdensome economic costs or impossible timelines. VMware has market share and mindshare, yes, but it simply doesn't have the leverage that Microsoft does, and it would do well to remember that.
VMware certainly seems to be enamored of Microsoft's goal of pushing all customers towards subscription services. It also seems to have been evolving its licensing towards Microsoft's approach. Furthermore, everything about nomenclature, packaging and marketing of the vRealize suite is reminiscent of Microsoft's System Center at its mid-aughties worst.
Trust is at the very core
of the next decade of IT. The public cloud can't function without trust. Managed services can't function without trust. Subscription IT -- especially mission-critical (to say nothing of life-critical) IT -- can't function without trust.
Trust is a very real form of capital, one which will become more valuable than money amongst tech vendors in the 2020s. Trust is hard earned and erased in an instant. Once destroyed, trust can be impossible to rebuild.
Microsoft doesn't have customers, they have hostages. They get to dictate terms, and there is no choice but to accept. If VMware copies Microsoft in this -- reducing choice and taking a "vendors knows best" approach -- VMware will lose everything.
VMware has pulled the plane up before it hit the ground. The paint's a little scuffed, but if they stay the course they'll get to incomprehensible rolling-around-in-it money in short order.
The Knife's Edge
All they have to do to ensure arrival at their destination is to keep customer-hostile moves to a minimum, assiduously avoid any perception of corporate arrogance, and never, ever jeopardise customer trust.
As we close out 2017, keep this knife's edge in mind. VMware has many capable competitors, and they will be watching out for any opportunity to turn even the slightest faux pas into a crisis of confidence that could torpedo the company for good.