For all of the breathless hype that containers and cloud native get, the strange-but-true story is that enterprises still need to run servers and for that vSphere is arguably the best possible option. VMware probably understands this in its inner sanctum but, well... Kubernetes envy.
Not surprisingly, then, VMware at its eponymous event linked arms with Pivotal (its step-child) and Google (with a cloud business run by ex-VMware and ex-Pivotal folks) to announce… PKS, essentially a rerun of the company’s open source Kubernetes-on-BOSH (KuBo) project.
If that doesn’t excite you, you’re not alone. There’s a lot not to like about the announcement, largely because it basically strings together three failed strategies to try to make one successful strategy. But the bigger fail in all this is that VMware, while missing the container hype train, still has plenty of virtualization to sell.
The VMware reunion tour
The generous view of the PKS announcement is that the three shades of VMware (comprised of current and former VMware employees spread across three different companies) got together to make it easier to run Kubernetes on-premises, all while combining with Google’s cloud container service. Or, as The Register’s Simon Sharwood reported, PKS “make[s] all three partners more relevant”.
He wrote: "PKS can use VMware's VSAN for persistent container storage, making it more useful, as well as running it in vSphere. Which makes Virtzilla's code relevant to another important application and therefore more useful. Pivotal gets a better container story to make Cloud Foundry even more useful. Google shows that Kubernetes may be cloud-native, but is also relevant for those who aren't wholly committed to the cloud."
A less generous reading goes like this:
VMware fails to deliver a cloud native strategy, adopts Pivotal's instead. Pivotal fails to deliver a cloud native strategy, adopts Google’s instead. Google fails to deliver an enterprise strategy, adopts VMware's instead.
In that kumbaya announcement, in other words, there’s a lot of hope that the sum of the whole will be much better than the individual parts. Of these, the one that probably has the most strength to it is Pivotal’s leaning on Google, and Google’s leaning on VMware. (More on this below.)
What strikes me as completely bizarre in the announcement and subsequent chatter, is that Pivotal declared that its big desire is to take market share from Docker, which generates relatively no revenue – despite its popularity. At the same time, VMware has put a bullseye on Red Hat, declaring: “The three of us are coming for Red Hat.”
I’m sorry, but that’s completely pathetic.
VMware has twice the market cap of Red Hat and while Red Hat’s early OpenShift (which competes with PKS) momentum has been impressive, OpenShift’s overall revenue still pales in comparison to that of VMware. I’m a big believer in Red Hat’s opportunity, but this is ridiculous. (An inside source at one of the companies who was on the call with VMware’s COO says the quote was taken out of context, and I’ll believe him, as the alternative is crazy train. VMware certainly didn’t bring up Red Hat on its latest earnings call (password required), though it also didn’t mention containers.)
Containers, containers, containers... er, VMs text
It’s not an either/or, though. Even if you believe the world is going to containers, the world still buys servers and they need to manage those servers, whether or not they are running containers. Containers are developers’ darling, but IT still wants VMs.
It’s true that VMs running on public clouds are growing faster, as Gartner’s Thomas Bittman has demonstrated. Yet it’s also true that over 100 million x86 workloads still run in private data centres, and that number keeps going up each year at an 8 per cent CAGR. That virtualization is dominated by VMware, with upwards of 71 per cent of the market going to VMware by some estimates. At larger enterprises, that number jumps to 82.9 per cent.
Even if we believe the world is going gaga for cloud - and I do - there are different flavours of cloud, with SaaS comprising a two-thirds of all public cloud spending, by IDC’s estimates. As for those SaaS vendors, they’re running lots (and lots) of VMware. As one insider at a large SaaS provider shared with me, their model is to run on public cloud when they enter a new region, then switch to private data centre once they hit a revenue threshold ($50m).
Yes, you guessed it: those private data centres will spend gobs of cash on VMware. As he described it, cloud is where SaaS businesses start, but private data centres are where they become profitable. If an enterprise has big, predictable, mission-critical workloads, they’re likely going to be running them on-premises, and paying VMware to make those operations more efficient.
Which isn’t to say that VMware shouldn’t be working with Google and Pivotal to map out a future in a Kubernetes-dominated container landscape. Why not? But let’s not forget just how much of the enterprise is still ecstatic to see VMware park the bus outside and bring VMs into their data centres. ®