 Live from Barcelona, Spain, it's theCUBE covering KubeCon CloudNativeCon Europe 2019. Brought to you by Red Hat, the CloudNative Computing Foundation and ecosystem partners. Olah Barcelona, I'm Stu Miniman and my guest host for this week is the one and only Corey Quinn and you're watching theCUBE, the leader in live tech coverage. Actually the fourth year we've been doing the KubeCon and CloudNativeCon, this is KubeCon CloudNativeCon Barcelona 2019. We've got two days of wall-to-wall life coverage. Last year we were in Copenhagen, it was outside, a little bit windy and we had this lovely silk above us. This time we are inside at the Fira. We've got some lovely Kube branding, the store with all the t-shirts and the little plushies of Fippy and all the animals are right down the road for us and there's 7,700 people here. So I have been, I did the Austin Show in 2017 at the Seattle Show last year in 2018. We had done the Portland Show in 2016. So it's my third time doing one of these but Corey, it is your first time at one of these shows. Wait, this is an AWS show, so what are you doing here? I'm still trying to figure that out myself. When people invite me to go somewhere, do you know anything about insert topic here? Absolutely, smile and bluff your way through. Eventually someone might call you on it but that's tomorrow's problem, not quite today's. Yeah, I have this general rule of thumb, the less I know about something, the more I overdress to overcompensate it. Oh, so here's the guy in the three piece suit. My primary skill is wearing a suit, everything else is just edging details. All right, so let's set the stage for our audience here, Corey. As I said, we've got the foundation, we've got a lot of the big members, we've got some of the project people. What I'm really excited, we actually have some excellent users here because it is five years now since Kubernetes came onto the scene, of course, built off of Borg from Google and as Dan Kahn said in the opening keynote, he actually gave like a nice historical lesson and the term he used is simultaneous invention and basically those things that, there are times where we argue who created the light bulb first or who did this and this because there were multiple times out there and he said, look, there were more than a dozen projects out there, many of them open-sourcer, a little bit open as to these things like container orchestration but it is Kubernetes that is the de facto standard today and it's why so many people show up for this show and there's such a large ecosystem around it. So you live in the cloud world, what's your general view on cloud native and Kubernetes and this whole kind of space? Well, going back to something you said a minute or two ago, I think that there's something very strong to be said about this being defined by its users. I've never yet seen a successful paradigm take off in the world of technology that was vendor defined. It's at some point you wind up with these companies doing the digital equivalent of here. We've crafted you this amazingly precise wrench and you hand it to a user and the first thing they say is, wow, it's kind of a crappy hammer but it's at least good for a first attempt and tools are going to be used as users want to use them and they define what the patterns look like. Yeah, so I'll give you the counterpoint there because we understand if we ask users what they wanted, they wanted better buggy whips, right? So we could go faster. So the compare contrast we had done a few years ago is this is OpenStack was user driven and it came out of NASA and I mean, if it was good enough for the rocket scientists, it should be something that we could learn on and rack space had done good and gave it to the open source community and step back and let people use it and first of all, OpenStack, it's not dead, it's being used in the telco world, it's being used outside of North America quite a bit but we saw kind of the boom and bust of that but it very much- We're a long way past the head end. The vendor ecosystem of OpenStack was, oh, it's an alternative to AWS and maybe some way to get off of the VMware licensing and I've actually said it's funny if you listen to what happens in this ecosystem, well, giving people the flexibility not to be totally locked into AWS and oh, it's built on Linux and therefore I might not want to have licensing from certain vendors, still echoes from previously but it is very different. Very much so and I will say the world has changed. I was very involved in Eucalyptus which was a bit of a different take on the idea or the promise of what OpenStack was going to be. What if you had cloud APIs in your own data center? In 2012, that seemed like a viable concern. There were, the world we live in today of public cloud first for a lot of shops was by no means assured. Yeah, Martin Meekos, a cubalum by the way, fantastic leader still heavily involved in open source and one of those things I think he was a little bit ahead of his time on these. So Corey, one of the reasons, why are you here? You're here because I told you here and we do pay you to be here as a host. You're not here for goodwill and that but you, your customers are all users and tend to be decent size users. As I say, Corey, you can help people with their Amazon bills and no, no, that's the AWS bills, not the I have a pile of boxes with smiley faces on there. Oh my God, what did I do around Christmas time? Exactly. So the discussion at this show is this all hybrid and multi-cloud world. When I talk to users, they don't use those words. Cloud strategy? Sure, my pile of applications and how I'm updating some of them and keeping some of them running and working with that application portfolio and my data, all hugely important but what do you hear from users and where does the things like cloud and multi-cloud fit into their world? There are two basic archetypes of user that I tend to deal with. Because I deal as you mentioned with predominantly large customers. You have the born in the cloud types who have more or less a single application, picture a startup that hits meteoric growth and now is approaching or is in the IPO stage. They have a single application, they're generally all in on one provider and the idea of going multi-cloud is for ancillary things. If we take a step back, for example, they're saying things like, oh, pager duty is a service that's not run by one of our major public cloud providers and there are a bunch of SaaS applications like that that factor in but their infrastructure is predominantly going to be based in one environment. The other large type of customer you'll tend to see is one of those multinational, very divisional organizations where they have a long legacy of being very data center first because historically that was kind of the only option and you'll start to see a bunch of different public cloud providers inside those environments but usually they stop at the line of business boundary or very occasionally on a per workload basis. I'm not seeing people say, well, we're going to build this one application workload and we want to be able to put that on Oracle cloud and Azure and GCP and AWS and this thing that my cousin runs out in the Ozarks and it just becomes, no one wants to do that in the traditional sense because as soon as you go down that path you're constrained to whatever the lowest common denominator across all of those things are and my cousin's data center in the Ozarks doesn't have a lot of frills. So you wind up trying to be able to deploy anywhere but by doing that, you're giving up any higher level offering. You're slowing yourself down. Yeah, the thing we've always been worried about is back in the day when you talked about multi-fender, do we go by the standard and then we go to least common denominator and what has worked its way through the environment and that's what customers want. I want today, if I'm a user, agility is really one of the things that seems to be top of mind. What IT needs to do is respond to the speed of what the business needs and a cloud native environment that I look at is it has to be that lever to be able to help me deliver on the next thing or change the thing or update my thing to get that working. It was, you know, so, disclaimer, Red Hat is our headline sponsor here. We thank them for a presence but actually had some great conversations with OpenShift customers and they didn't talk about OpenShift, OpenShift, OpenShift. They talk about their digital transformation. They talk about their data. They talk about the cool new things that they're able to do and it was that platform that happens to be built on Kubernetes that was the lever to help them do this. At the Google show where you were at that was the same conversation we had. Whether it is in GCP or whether it was in my own data center, you know, yes, we can do it with containers and everything like that. It was that lever to be able to help me modernize and run new apps and do it faster than I would have done it in the past. So it's that kind of progression that is interesting for me to hear and just there is not, you know, there's this tendency now to be like, oh, look, everybody's working together in this wonderful open source ecosystem. It's like, well, look, the world today is definitely co-opetition. Yes, you need to be up on stage and if customer says, I need to work with vendors ABC and D, ABC and D, you better work with that or they will go find an alternative because there are alternatives out there. Absolutely. And when a company embarks on a digital transformation and starts moving into public cloud, there are two reasons they're doing that. The first is for cost savings, in which case, let's talk. And the other is for a capability story and you're not going to realize cost savings for a lot longer than you think you will. And in any case, you're not going to realize capability story, if all you view public cloud as being is another place to run your VMs or now your containers. Yeah, so thank you. Corey, your title, you know, in your day job, you're a cloud economist. I am, two words that no one can define so no one calls me on it. So Kubernetes, it's magical and free, right? That's what everyone tells me. It feels like right now we are sort of peak hype as far as Kubernetes goes. And increasingly, whenever you see a technology that has gotten this level of adoption, we saw it with OpenStack, we've seen it with cloud, we've seen it with a bunch of things, we're starting to see it with serverless as well, where what problem are you trying to solve? I'm not going to listen to the answer today, that answer is Kubernetes. And it seems like everyone's first project is their own resume. Great, there has to be a value proposition, there has to be a story for it. I'm not suggesting that there isn't, but I think that it is being used as sort of an upscale snake oil in some cases or serpent grease, as we like to call it, in some context. Yeah, and that's one of our jobs here is to help extract the signal from the noise. We've got some good customers, we're going into the environment. One of the things I try to do in the Open Keynote is find that theme. A couple of years, for a couple of the shows, it's been like, oh, service mesh is the new hotness. We're talking about Istio, we're talking about Helm, talking about all these environments that say, okay, how do I pull together all of the pieces of the application and manage that together? Because there's just moving up the stack and getting closer to that application. We'll talk about serverless in one of the other segments later this week, I'm sure, because there's the, okay, here, Knative can help bridge that gap, but is that what I need? When you talk a lot about Kubernetes is how much is in the public cloud versus in my data center, and there's some of the guys they talk to, serverless is in the public cloud? We'll call it functions as a service if you put it in your own data center because, well, yes, there are servers everywhere. If you actually manage those racks and everything like that, probably doesn't make sense to call it serverless. We're trying not to get in too many semantics arguments here on the queue. You can generally tend to run arbitrary code anywhere. The premise of serverless to my mind is more about the event model, and you don't get that on-prem in the same way that you do in a large public cloud provider. And whether that's the right thing for you or not, I'm not prepared to say, but it's important that that be understood as you're going down that path. So, Corey, any themes that jumped out for you were things that you want to poke out at the show? For me, Kubernetes has really kind of crossed that chasm, and we do have large crowds. You can see the throngs of people behind us and users that have great stories to tell, and the CNTF itself has a lot of projects out there, and we're trying to make some sense of all of those pieces. There are six now that have graduated, and FluentD is the most recent, but a lot of interesting things from the sandbox through that kind of incubating phase there, and we're going to dig in with some of the pieces there. Some of them build on top of Kubernetes. Some of them are just part of this whole cloud-native ecosystem, and therefore are related, but no necessarily need it, and can play in all these various worlds. What about you? For me, I want to dig a little bit more into the idea of multi-cloud. I've been making a bit of a stink for the past year with a talk called The Myth of Multi-Cloud, where it's not something I generally advise as a best practice, and I'm holding to that fairly well, but what I want to do is I want to have conversations with people who are pursuing multi-cloud strategies and figure out first, are they in fact pursuing that, the same thing that we're defining our terms and talking on the same page, and secondly, I want to get a little more context and insight into why they're doing that and what that looks like for them. Is it they want to be able to run different workloads in different places? Great, that's fair. The same workload, run everywhere in the lowest common denominator? Well, let's scratch below the surface a bit and find out why that is. Yeah, and Corey, you're spot on, and no surprise, because you talked to users on this from our research side on our team. We really say kind of multi-cloud or hybrid cloud, because hybrid cloud means you've got your own data centers that's supposed to multi-cloud, could be any of them, and there's a little bit of a Venn diagram you could do between that. But I am prepared to be wrong as well. I'm a company of two people. I don't have a research department that's called the spare time I get when I can't sleep at night. So I don't have data, I have anecdata. I can talk with individual use cases, but then I'm telling individual company stories that I'm generally not authorized to tell. So it's more a question now of starting to speak with a broader base. So just to finish on the thought from our team is everything from I have all of these pieces and they're really not connected, and I'm just trying to get my arms around them through some of the solutions, like in the AWS world, we're looking at the VMware and AWS and Outpost type of solution that pull out, or what Azure does with Azure Stack and the like, or even companies like IBM and Oracle, where they have a stack that can be both in the public cloud and the private cloud. So those kind of fully integrated pieces versus the right, no, I'm just putting applications in certain areas, and then how do I manage data protection, how do I manage security across all of these environments. It is a heterogeneous mess that we had, and I spent a lot of my career trying to help us break down those silos and get away from the cylinders of excellence, as we call them, and we should work more towards generalists, so how much are we fighting that? I will just tell you, most of the people we're going to have on the queue probably aren't going to want to get into that. They'll be happy to talk about their piece and how they work with this broad, wonderful ecosystem, but we can drill into where Kubernetes fits. We've got the five-year anniversary of Kubernetes. We'll be talking to some of the people that help create this technology and lots of the various pieces. So with that, Corey, want to give you a final take here before we talk about the stickers and some of the rest. Oh, absolutely. I think it's a fascinating show. I think that they're the right people who are attending to give valuable perspective that quite frankly, you're not going to get almost anywhere else. It's just a fascinating blend of people from large companies, small companies, giant vendors, and of course the middleware types who are trying to effectively stand between, in many cases, customers and the raw vendor for a variety of very good reasons. Partner strategies are important. I'm very curious to see what that becomes and how that tends to unfold in the next two days. Okay, so the queue, by the way, we're not only a broadcast, but we are part of the community. We understand this network and that is why Corey and I, we come with stickers. So we've got these lovely sticker in partnership with women who go. It made this logo for us for the Seattle show and I have a few left, so if you come on by, Corey has his platypus last week in AWS, so come on by where we are, you can get some stickers and of course, you know, hit us up on Twitter if you have any questions. We're always looking for the community and the network to help us with the data and help us pull everything apart. So for Corey Quinn, I'm Stu Miniman. Two days of live wall-to-wall coverage will continue very soon and thank you as always for watching theCUBE.