 from Seattle, Washington. It's theCUBE, covering KubeCon and CloudNativeCon North America 2018. Brought to you by Red Hat, the CloudNative Computing Foundation and its ecosystem partners. Over and welcome back to theCUBE's live coverage, day three here, theCUBE covering KubeCon and CloudNativeCon 2018 in Seattle. I'm John Furrier, Stu Miniman and Justin Warren. Here to break down the action, Justin Warren, as you know, is guest analyst for us at many events, chief analyst at pivot nine, coming all the back over here again to break it down. So we're going to dissect what's going on here at KubeCon, CloudNativeCon. This is, some say me, the last stand to stop Amazon. Justin, good to see you. Good to see you as well, man. Stu, our first question is, as the show winds down day three, a lot of people had left all the big execs are gone. This is kind of last day, people coming together, the party was last night. So we kind of see all the action. We kind of fish this pond dry in the Kube here the last couple of days. The themes are starting to emerge. What are you seeing? What's your thoughts? Yeah, I mean, first of all, John, 8,000 people. This is, you know, geeks that are really excited. And I mean that in the best of ways, of course. There's actually, there were people here before the show started doing lightning talks in full day session. Tomorrow there's an operating session that another 250 or 300 people will be doing Friday. So, you know, and people want to just, you know, suck the marrow out of the, you know, the bone that is everything going on here. Just get every ounce of knowledge here. And they are deep into the session. So this is a great community. Question I want to ask you guys is, you were at Amazon re-invent two weeks ago. Yeah, you know, we've watched that show. I want the comparing contrast of this ecosystem and show, not just compared to like, say, OpenStack, which we've been teasing apart all week. And I think there are some things we need to worry about, but a lot of good differences. But compare against the big one in the room, which is Amazon. And a big difference is Amazon is here, and they have a seat at the table because they have to and customers will force them there. But, you know, should this worry Amazon and how does this ecosystem compare with the Amazon ecosystem? The big thing for me is, I understand how people make money in the ecosystem of Amazon. I'm still trying to figure that out here. Yeah. Ah, it is a different ecosystem. It does have a bit of a vibe of, it could be the new re-invent. We've had conversations over the last couple of days about- Or is this the independent cloud, you know, open ecosystem? It is the independent show that we've been waiting for, that we've wanted since, you know, Comdex and Interop kind of went away. And it's all been vendor shows. And now we have an independent show where all the vendors can come and have kind of a neutral meeting place, and we can all gather together and have some common ground, which is like, that's what Kubernetes is. I've been saying over the last couple of days, Kubernetes is like the ethernet of cloud. So it's something which is an agreed standard we can all collaborate on, and then, you know, you never bet against ethernet. So now you can build all of these other things on top of that platform. Yeah, just a quick note on that, right. That's Interop. Like networking, was that the core of that? Yeah. It was basically everybody, oh, yeah. It's the chance if we give true interoperability, maybe we can do multi-vendor and it won't all be Cisco, who dominated that market, Amazon's the same. Stu, this is to me, ethernet's a great example. I say TCPIP as well. Both are enabling technologies that are standardized or actually started as de facto standards. They weren't necessarily, you know, bonafide standards. They emerged and people rallied around them. Those de facto standards emerged and become a catalyst point for people to build on top of and around. Remember, it was still a lower level below the stack. On ethernet, Stu, you had, you know, physical data link layer in the OSI model. The grandfather of all stacks. That really changed, I think, 20 years of growth and innovation. I think Kubernetes is exactly right, Justin. This is exactly your point. I see that as well, that it's not so much Kubernetes is going to be all end-all. It's what it enables. And I think the innovations on top of Kubernetes and underneath Kubernetes take the holy trinity. I've been saying this in theCUBE now over the past year, the holy trinity of infrastructure in IT is storage, computing, networking. And those things are now being repurposed in a way that is highly scalable, dynamic, and resourceful for a lot of things. AI is a great example, everyone talks about AI, but storage, policy, the knobs in Kubernetes can be managed and Google saying the guys are Kubernetes, that's one of the most underutilized aspects of Kubernetes is the networking guys managing the knobs from below and then app guys with service message maybe on the top. This is just an absolute growth engine and the comparison to Amazon is similar because Andy Jassy talks about builders, the right tool for the job. This is essentially the same mantra. I mean, this is tools, platforms. But with one very important difference in around the money side of things. You don't have this massive behemoth which is going to come in and one year you're on the keynote and the next year we just announced a product which completely killed your business. It's open source. That's not really going to happen. So you've got that common core of things where there's no real competitive advantage on this stuff. So that's, you know, Linux, where's the competitive advantage on a kernel? There isn't one. So open source makes great sense for that kind of core of things that you then build upon and then all the money is in all the innovation, all the value add that goes on top of that. And that makes a huge amount of sense to have an open source show for that. And I think Stu, one of the things that we always kind of talk about networking in cloud, I think the concept of cloud is going to be old hat. You heard it here first on theCUBE because cloud is Amazon and cloud is a set of resources. When you start thinking about IoT at the edge, when you talk about moving compute to the edge, you're going to start to see mesh networks peer to peer, not a new kind of platform configurations that not isn't necessarily cloud. It's a new thing. It's a platform, open platform. And there's going to be some incentives that are going to be designed for startups that's economically beneficial to the new kinds of things versus the economic incentives that Amazon might not have. So I think we're going to see emergence of new stuff. I would still say that cloud is a state of mind. It's not a location. So we're here, it's cloud native con. It's not just cube con. It's about doing things in a cloud native way. And that, like you say, it doesn't matter where it is or how it can be communicated together, but it's the way you operate it. It's the way it actually works in practice. It's not so much about, oh, we're going to build it here and we're going to put it in that cloud or that cloud. And I think we've had some real clarity as to what that future of multi-cloud looks like, because it's not one massive cloud everywhere. It's not, oh, my application's spanning all over the place. It is we are working to solve that really tough problem of distributed architectures and giving us ways that I shouldn't have to think about, you know, where I am spinning that up or if I need to change vendor, not necessarily portability, you still do have some lock-in because Kubernetes is not the full stack. It's a piece of the overall platform and while there's like 75 different versions here that are all compliant, I should be able to move between them, but the devil's in the details and there's lots of stuff we have on top. Let's talk about multi-cloud. I want to talk about multi-cloud for a second because you mentioned Comdex, we're talking about Ethernet. At that time, during those big revolutions, the word multi-vendor was a big buzzword. Multi-vendor was like on the basis of Comdex. We all got to play together. Multi-vendor meant choice. Today, multi-cloud is just a modern version of multi-vendor. Exactly, it's multi-vendor. And that's what enterprises want. Enterprises are a bit wary now. We hear lots of conversation about lock-in and that comes up a lot and it's a real thing. Enterprises are concerned that they don't want to bet on one company and then find out that actually it's technology, it changes. Things need to be moved around. We don't want to wake up in five, six years and then suddenly find, oh my God, I can't change anything because I'm locked into this one vendor. So Justin, they say they want multi-vendor. When it came to networking, I spent years working on interoperability and plug tests and all these things. And at the end of the day, it was way better to get my standards plus with a single vendor than it was to try to root them together. And then, oh, when I change something, so hopefully the different here is actually we have loosely coupled services, we have APIs. So can we actually do multi-vendor, multi-cloud that doesn't stress out my team and have every time I want to make a change or they make a change, it moves. The new cloud world should be things change. It changes upstream and downstream I get to use them. So once again, we talk about this shiny nirvana of, oh, it's serverless and the old trinity of computing storage. I don't even need to worry about that because it'll just work. But wait, if something goes wrong, I've been talking to a bunch of vendors here that actually how do I get observability and manageability to be able to drill down because things could still go wrong. Oh, you heard Bloomberg, we had an end user come on. It's a very interesting point. And Dan Kahn from the executive director, well, Bloomberg is kind of a different case, but look at what Bloomberg does. The guy said to us, I actually don't want to buy these products and services. I just want to pay them money to be available to support me when I'm in need support because Bloomberg has fully integrated all their support internally. I think that's a trend that we're going to see in the enterprise where CIOs start building teams, real software chops. It might not be as big as Bloomberg, but the notion that we're going to run our own stuff. We'll use managed services where appropriate, we'll have a core software build strategy and I can't wait in SLA of four hour response time. I need like minutes. And that's, I think, we don't have the answers yet. There's still a lot of questions that enterprises are trying to work out about. How do I actually do that? So you mentioned Bloomberg and I interviewed them a few months ago, wrote something in Forbes about them. They are a special case in that they have chosen that we're going to invest in this technology so that we have people on staff in our company who understand Kubernetes. Now that's not a choice that every enterprise is going to make, but they decided that actually this technology, this software is so important to our business to where we get all the value for our business that we need to invest in that technology. And I think a lot of enterprises are realizing that actually outsourcing everything to one vendor and then giving all of your innovation engine to someone else. And they're realizing that was a mistake. Now they're trying to figure out, okay, what do we bring in house? What do we do ourselves? What do we get vendors to do? Which technologies do we use for what particular value creation? And that complexity, that decision-making process, that's what we haven't quite worked out yet. And that's where I think there's a lot of value in the ecosystem with service providers who can provide advice on here is how you should do it based on what you need to do. That's a great point, Stu. I want you to comment on that, let's riff on this for a second because the people actually spend the money or the people reimagining IT infrastructure, IT applications. CIO, I've interviewed the VP of Advanced Technology at Proctor and Gamble, and he told me when he came in, he came from Coca-Cola, he's been an old IT guy, he says, look, we've outsourced everything to the point where we're anemic. We've got a couple storage guides, they're pushing buttons, they're jumping on, all in the vendors, they outsource everything. He says they had no ability to create a competitive advantage for the business. And what they moved quickly to was to bring talent in to be builders, to be in house. So now you have that trend happening in the modern CIO, CXO kind of roles. Now you have to say, okay, I got teams here. How do I get the investments deployed? How do I go to this ecosystem here with all these tools, all these capabilities? How do I invest? How do I build out? Look, I think Kelsey Tauer had a great point when we interviewed him this week. It is a huge opportunity for managed services because like we talk about the Amazon or even the ecosystem, how do I keep up with all of this? And the answer is, you don't. You need to be able to have people, whether it's systems integrators or partners that are going to be to help that. Look, Amazon gets criticized for not being deeper in open source. Well, they use a lot of open source and they deliver those as services and they make it easy. Frictionless is something we talked about for many years as being in that thing. The enterprise wants to be able to spend money and just go do it because they don't have the team of PhDs. Even somebody like Bloomberg or some of these really big companies. I love talking, you've got Apple and Nordstrom and some really interesting- They're here at the show. Oh, by the way, and they're all hiring. Whether or not they're actually using Kubernetes, they cannot confirm or deny, but we know how that goes. Let's unpack the end user piece here. Amazon is pushing 5,000 referenceable customers. So about the Amazon question. We end users here. How many referenceable customers are here? What are they actually Uber's here? They're hiring. They might have some Kubernetes stuff in the background. Sure, they probably do. But actually what is the end user adoption really look like? It's still early, but again, a difference between this show and Amazon re-invent. How many end customers have a booth at re-invent? Compared to here where we have people, end customers who are here mostly to hire talent, they have booths. Kudos to the CNCF. They've got 80 end users participating. There are a lot of users here. This is not the vendor fest that we see at some shows when they get big. I hear they're not soaking the vendors. The vendors that I talk to are happy because there are the users here and they're excited. Before we go, John, there's a couple of kinks in the armors and things we need to worry about. The two, if I look at service meshes and I look at serverless as a huge threat. One of the things I wanted to look at coming in was I'd heard a lot of talk about Knative. I think Knative is great, but it is not. Lambda is the de facto standard just like S3 was before Lambda is this and Knative has absolutely nothing to do with Lambda and does not connect with it. It is the difference between serverless and functions. And so all the AWS functions and all the Azure functions have nothing to do with Knative. For the people that looked at OpenWisk and all these other options, Knative seems a good way to pull. They've done a re-spin of what's happening there and it's moving things down the line. Once again, as Kelsey said, if we look at serverless as a spectrum, which many of the hardcore serverless people will debate and argue and be like, that's not real. Serverless, well, just like we said, there was only one real cloud and it was Amazon. We know that's not the case. It will be a spectrum. We want to meet customers where they are. So Knative, goodness, but the elephant in the room is that AWS and Azure are where all of the serverless really happens and therefore, you know, there's a big air gap between them. Justin, service mesh is something I know you've been looking at and it gives the good, the bad and the ugly. Service mesh is really, really early. So we're at that part where there's a diversity of innovation going on. So we do it where there's about 12 or at least 12 different companies here at the show who are all doing something with service mesh. They're all trying to sell you a different solution. This is what happens with technology. We have a new technology gets created and we have this flurry of all these startups who are all trying these different things. And they said, this is the, what is it, the destructive force of capitalism. It's like, not all of them are going to succeed, but we have to have them all out in there in the market because at the moment, it's too early to figure out, okay, well, it's definitely going to be that one. If we knew that one, then I'd be putting all of my money behind that one company today. Last year, Justin, all the talk was about Istio. I've heard a lot of talk about Istio, but it hasn't all been good. No, that's the thing. So we've had a year now and last year was definitely, hey, Istio is like the service mesh. Like, not so much. Envoy seems to be the common ground that people are actively using. That's what most people are building on top of. So it looks like Envoy is going to be that underlay of everything else. But in terms of how you actually use service mesh, it's still very early and people are trying to figure out how do I use this quite complex technology in practice. And as people use it more, as we get more adoption, then we'll start to see that one or two of the methods and the approaches will win out over all of the others. And that's where we can expect to see, okay, we'll have an anointed winner, that will then win out because it's useful, because it's functional, because end users want to do it that way. And Envoy, by the way, had traction, they had it sold out. Envoy Con, on the first date, 350 people, Lyft is driving that, and they're just heads down solving problems. I think that seems to be the formula for some of the successful projects. If you take away all the window dressing and the hype, it comes down to who's solving what problems. And that's the thing with open source. You can't just throw a whole bunch of marketing dollars at it to make it succeed. If end users don't like the code and they don't use it, then it won't work. Yeah, John, I want you to give us the word on the open source business model. We watched in the last year, Red Hat bought CoreOS for 250 million. Then they were acquired by IBM for 34 billion, pending, final, all that stuff and everything. And then, reading through the VMware SEC filing, $550 million for Heptio, big, big dollars. So, is open source just to get a lot of customers and they get acquired by the big guys? What's the... Well, I think it's interesting. First of all, Red Hat might not like what I'm about to say, but I'll just say it. I think there was a steal with CoreOS. If you look at what Heptio got for the valuation, CoreOS was an absolute steal. The team was phenomenal. They were doing some amazing work. At that time of the acquisition, the debate of how to make money dominated versus just getting behind the technology. And I think CoreOS was a fantastic team and they had the right track and you can see what's happening now, now part of the Red Hat. So, Red Hat got a massive lift on that. So I think Kudos to Red Hat for taking that up the table at the time, great acquisition. I think that helped them propel and that showed that to IBM that there's real value there. Now, I think open source as a business model is interesting because it's changing, right? You now have a new generation of builders and developers coming in. Open source has to evolve. And I think the CNCF I think is a cutting edge experiment or a Petri dish of how to stay true to open source principles. It's still nurture and enable a downstream impact for the commercialization. I think it's an opportunity, but it's also one of their biggest challenges because if this is Comdex, Comdex is an open source, it's Hawkenwares, right? So there's a different business model. So, this is going to be a very interesting test in the industry to see how the current open source momentum, which is looking really strong right now, how that can interplay with commercialization. Because certainly the money's there, the value's there. And if we can get these value spots identified, the white spaces for startups, and let the big guys also play as well, it's going to be a very interesting landscape. It's certainly dynamic. I don't have the answers, but my gut's telling me that a whole new level of sets of services and platforms are going to be composed around these services. And I think it's all going to be driven by open source. That's clear. How it shakes out valuations of the talent buys, of momentum, market buy. We'll be watching, I don't know. Yeah, it's exciting times. We're here at the beginnings of what I hope is going to be this massive new ecosystem and we get to watch it grow. We get to watch it change. It's a great place to be. All I can say is to, I wish I was 25 years old again right now because for young entrepreneurs and young tech folks, this is probably one of the most exciting times because you have real computer science and dormant computer science now re-energized with cloud computing and stale. Yeah, but John, they don't appreciate what they had. They don't know what it was like to have a computer that wasn't actually connected to things, let alone what we had. I used to build my own graphics libraries. I used to walk to school and bare feet in the snow. It's so hard. It's so easy now. Yeah, Craig runs in there for my token ring. Grading ones and zeros by banging ropes together. Yeah. It's so easy now. You guys got it made. You have no idea. Great stuff, Stu. This is a great analysis. And I think, again, KubeCon is the beginning with cloud native. This is just a small signal, I think. I think there's going to be a Comdex moment soon unless this thing just blows up, which I don't think is going to happen. Yeah. I mean, look, last thing, John, I want a big thank to the Linux Foundation, CNZF, for working with us. We've been there within the early days. Great partnership, this community. They've got a great media sector. All our friends over here, they're creating a lot of content, working really hard. The amount of work that goes through, and as we have the people from the CNZF talking, they've got a core team, but it's people that volunteer and work community too, and all our sponsors, John. Yeah, thanks to the community. And again, one more final point is that this market, Justin, as you know, we all cover it, is in the learning mode. There's a lot of education oriented stuff that people are interested in. You got, you know, Alex Williams over at newstackdevops.com, TFI are over there. Everyone's pumping media out there. There is a thirst for content. There's a thirst for community learning. The sessions are packed. I mean, the hallways are interesting. You see people huddling, and I overhear the conversations. They're not talking about what party to go to. They're talking about how to implement a community-based cluster. So there's really people working on and off the court here, so to speak. So it's been great, great coverage. So day three, breaking it down. I'm John Furrier, Justin Warren, Steve Miniman, back with more coverage. Day three after this short break.