VMware Kubernetes – Epic Fail In Flight

By Jay Valentine

VMware wakes up every couple of years, like the guy in Sleepy Hollow, and find the world is not no longer virtualization-only.  In 2013 it awoke to the fact that the cloud was a next big thing and VMware’s customers were fleeing to Amazon and Azure.

Off they went, spending billions of V-Sphere profits, building data centers while breathlessly announcing they were going to bring “the cloud” to the VMware customer base.

Not much later, VMware quietly closed down that hot new strategy and said, “….well, maybe we got that wrong.  Better to partner with the cloud leaders.”

Today, the hot new thing, now that the Software Defined Whatever appears a bit less their message, is VMware leadership in the Kubernetes market!

Wow, what’s that? Sounds pretty cool and certainly far sexier than virtualization. 

Kubernetes is an orchestration platform to build stuff in containers.   It is a big deal because it enables development of micro service based apps that are generally cleaner, easier to build and can be used as building blocks for other, bigger apps.

Kubernetes is app development.  VMware is data center plumbing.

VMware is the plumbing supply shop for the corporate IT world.  VMware builds the low level server layer enabling one server to run many apps.  Plumbing, but good stuff.  Great profits; starting to struggle.  Everyone has most of the virtualization they need.

Virtualization requires data centers.  “The Cloud” means using someone else’s data center.

Thus, the need for a new VMware story.

Kubernetes is hot.  The tech blogs are saying it so it must be so!

VMware’s problem is Kubernetes’ apps are moving to either a bare metal (thus no virtualization box) or a super light virtualization box.  Oops, no VMware!  In either case, the VMware people are pretty much screwed going forward.

Remember VMware’s core management strategy:  Wake up, read the tech blogs, see what the hot new thing is, and jump on it applying more PR than real value add technology.  Its entry into the cloud market, above, being a recent example.

Let’s take a tour away from cool sounding technology terms and apply a bit of common sense to VMware succeeding enough in the Kubernetes container world to impact its stock price.  After all, what else is their management there for?

First, the Kubernetes world is app dev and not IT hardware/software plumbing.  The VMware sales force has led a very happy, fat life for almost a tech generation as the order takers to the IT technical staff.  A VMware core rep has probably never seen an app dev executive except in a training brochure.

The VMware core sales teams need to learn a new language and entirely new way of selling to sell in the app dev space.  They are not selling alone anymore, like they did in virtualization.

VMware faces a market where it has zero market penetration (Kubernetes, app dev, containers) with a sales force that has sold for decades to a constituency where it had 97% market penetration.  Again, this is order taking.

Order taking is easier than selling complex technology when one is NOT the market leader.

Critical thinking suggests the VMware sales force just ain’t going to get it done.

VMware will spend billions on training, all kinds of sales battle cards, SPAMMING app dev types, but the needle won’t move much.  Industry analysts, not the Gartner types who reprint your press releases if your marketing budget is large enough, but the ones who lurk on those investor calls, will demand the rev from Kubernetes.

Ouch!  When VMware has to report those Kubernetes revs are not there, and they will, the story breaks down.  All tech stories eventually fail if there is no revenue.  Ask Theranos, WeWork and a host of others.

But there’s more.  VMware, as all monopolies do, has generated quarterly revenues forecast to the third decimal point accurately for years.  Since they had incredible leverage over their virtualization customer, they could deliver predictable revenue – quarterly.

App dev deals are not a quarterly revenue kind of thing.  App dev projects, particularly where VMware is one of many vendors, will happen when they happen.  The great quarterly revenue forecasts die in a forest of competition, cross claims, demands for “show me” how you have done this before.

Don’s believe me?

Well, VMware is telling us they are going to use the Pivotal technology and sales teams to lead the way.

Really?  Those same Pivotal folks, many or most of whom are former VM-ers, so missed their quarterly numbers the stock tanked over 50% from its IPO.  Things were so bad, probably getting worse, that VMware bought the place using its Dell-mandated cash/borrowing capability.

VMware is betting the Pivotal guys and gals and the VMware plumbing core sales teams can deliver manifestly growing revenues, every quarter, when they just showed they could not do it post IPO.  Critical thinking suggests this is not going to happen.

There is another issue here, and it inherent to VMware’s strategy of reading the tech blogs for its strategy.  VMware people are data center, legacy software, move to the cloud types.  Most IT people are.  Nothing wrong with that.

Except there is a growing uneasiness in the Fortune 1,000, quite pronounced in the Fortune 100, that moving to the cloud does not change anything.  Companies are finding that moving to the cloud is a massively expensive endeavor and when they get there, their apps run generally the same.

Some execs call it “moving our crappy apps to the cloud where they run just like they did before.” 

In conversation after chat with cloud vendors, there is the continual refrain that their customers are not achieving “digital transformation” via the cloud.  They are achieving app movement with a cost profile similar to having the apps in their own data center.

Why is this?

The defining characteristic for an app’s performance is the tech stack that it sits on.  OK, let’s go back to Kubernetes and VMware here.  You remember, Kubernetes, that container stuff.

Kubernetes/container apps do not run that much faster than their predecessors.  They can be written in a bit shorter time, and modified more quickly, but not appreciably much.  Pivotal tools, you remember them from 6 paragraphs ago, cost tons of money, like a million dollars is common.  And Pivotal apps are no faster than conventional apps when they are finished.

This is because the tech stack is king, or queen of the development environment.

Today’s tech stack goes from the bare metal to the O/S, to some middleware, to virtual machines, to a database, to an app and then user interface.  While not always in that order, each of these exchanges offers an I/O wait state which can be 10x.  What does that mean?

It means these things are really slow.  (Think 10 to the 7th or 8th power slower, which is really slow).  Kubernetes, micro services, and the rest are limited by the chassis, which is the current tech stack.  VMware has bought into the stack.  They are the technology stack, or a major part of it.

This is where looking beyond the tech blogs becomes important, people.

There is now an emerging market that NEEDS apps to run thousands of times faster, use far less storage, eliminate IT taxes like Oracle or VMware and deliver business models unavailable with the current stack.

Some companies do actually need to TRANSFORM, not just use someone else’s data center.

Some of these technologies are being delivered as EDGE computing.

Now our pals at VMware are all in with EDGE computing.  So are their masters at Dell. 

What they missed in the tech blogs is that EDGE computing and computing at the EDGE are quite different phenomena.  And that difference is the tech stack.

“Computing at the edge” means building mini data centers near the customer.  This has VMware all over it and it uses the current tech stack.  Unfortunately, it does NOT deliver apps that run thousands of times faster, can be built in a quarter, use 90% less storage and run on a phone or embedded device.

No VMware on a phone.

Then there is true EDGE computing.  That means delivering massively distributed apps, that run thousands of times faster, without a data center, delivered on a phone or embedded device. 

This is the world of the new tech stack, and there is no room in it for VMware.

Here, container-based apps run a thousand to a million times faster, with nor Oracle, and can be delivered on phones and tablets.  Some early adopters are getting these benefits today and that base will only grow.  Eventually even the tech blogs will take notice.

For our friends at VMware, there does not appear to be great hope they will be able to deliver Kubernetes leadership even close to what they did with virtualization.

All good things do come to an end and it appears that if one thinks critically about VMware, there is a case to be made that its best days were in virtualization.

Find The Customer Looking For You