Renski talks OpenStack ambitions, Mirantis strategy, and Gartner’s OpenStack support
“Does Mirantis know how to build OpenStack clouds? Do Russians know how to drink vodka?”
Accompanied by a vodka-drinking bear, the stunt was synonymous with a man in OpenStack who is often known for his no-nonsense talking and ambitious strategies.
TechWeekEurope managed to grab Renski during the summit for a quick-fire Q&A on the state of OpenStack and where Mirantis wants to be.
Gartner’s Donna Scott took the headline spot during Monday’s keynote, describing how bimodal IT and OpenStack go hand in hand. What do you think of Gartner’s new-found support for OpenStack, something it used to describe as a “science project”?
They have a templated and proven approach with respect to what they do when it comes to new technologies. And when there is a new technology of any sorts out there that has a lot of mind share, they almost always say: “Oh be careful, it’s early, and there’s too much hype.” They take kind of this negative stance. And then over a period of some years they gradually move to. support it. So they did the same with Linux. They were like: “It’s too early, don’t touch it. Use Microsoft.”
But OpenStack has grown quicker than Linux, don’t you think?
From a standpoint of just mind share, and the number of people involved, it has. But from a standpoint of actual value-add in organisations, it’s kind of hard to gauge to be honest. I wouldn’t go ahead and make such bold statements and bash Linux and say that OpenStack is better, but I think in general the fact that Gartner is now supporting is an indication that OpenStack has graduated. It’s a stamp of approval to use in the enterprise.
Mirantis is the largest contributor to OpenStack. What’s next for Mirantis?
There’s the technology side and then there’s the business side. From the business side, which is a little bit boring but still important, there’s kind of two vectors. One, and this is a priority, is making the customers we have right now successful. We initially focused and got into the large telco accounts and Fortune 500s on the enterprise side – which is like the cream of the crop. And now we’re finally at the stage where all of the cool experiments they were doing for a while are getting the visibility at the CIO level. That’s our number one priority, not so much about getting new customers, but making the customers we already have successful, like AT&T and VW and these guys.
The second priority is actually seeding OpenStack into the Global 2000, you know, the one step down size-wise accounts. The big question there is whether there will be private cloud at all. With Global 2000s it’s unclear, so figuring out what we do there is the next thing.
On the tech front, it’s a lot about bridging Kubernetes and OpenStack together. So that’s what we’re working on together with CoreOS and Intel. There’s a number of benefits, and I think that our view is that the world needs to come to a single data centre operating system substrate, which will be able to run VMs and bare metal and containers together in a single fabric. The first step of that is using Kubernetes and containers to simplify the operations of OpenStack, by basically containerizing OpenStack services. The second wave is actually extending Kubernetes underlay so that you can actually natively run containers on Kubernetes bare metal but make it a single platform.
AWS isn’t going away. What happens if the Global 2000 opts for public cloud?
I think that Global 2000 definitely for the most part will warm to OpenStack the same way AT&T and VW did. They are standardising the technology and building all this in-house expertise. But that doesn’t necessarily mean that they’ll go all the way to the other side where they’re going to put everything in the public cloud. It’s kind of a spectrum with a lot of grey area between and what I’m trying to say is that the Global 2000 will most likely fall somewhere in the middle of the spectrum. So it’s still the private cloud, but it’s not like they built it themselves and are running it themselves. There’s going to be different delivery models.
OpenStack Summits are echo chambers of praise and support for the platform, but what are some of the negative aspects of OpenStack today? What’s not going so well?
I think that, from a technology standpoint, there are still issues, right. Like, there’s not enough functionality, it doesn’t scale so well, etc etc. But the biggest problem right now is bridging between early, pilot, pre-production stuff that OpenStack was in – that stuff was like an echo chamber of techies that were evangelising this inside the larger organisations. And the way this was viewed from the top down in the businesses was that the business does not understand the value.
The value might be there, but the people below don’t know how to quantify it or sell the business case. And this whole thing of how OpenStack has been sold historically – kind of bridging it to the value of selling in the organisation, is something that needs to happen. That’s where we personally are struggling the most.
What about complexity? VMware’s Pat Gelsinger recently claimed he’s seeing customers throw engineers at OpenStack but it’s just too complex and they’re coming back to VMware.
But this isn’t related to technology. This is related to what I was saying. What we’ve seen basically is that you can have a perfectly well working OpenStack cluster, and then the guys at the top make a decision, and the go “Oh AWS works, so, ok, AWS!” And then that starts propagating down. Or they make a choice like “Okay we’re in a downturn, let’s stop all this experimentation, let’s stick with what we know: VMware.” This is the problem. It’s not that OpenStack doesn’t work!
Quiz: The Cloud in 2016