>> Announcer: Live from
San Diego, California, it's theCUBE, covering
KubeCon and CloudNativeCon, brought to you by Red Hat, the Cloud Native Computing Foundation and its ecosystem partners. >> Welcome to theCUBE here at KubeCon CloudNativeCon, 2019. Second day of three days,
wall to wall coverage. I am Stu Miniman, John Troyer is my cohost
for the three days, and we've had a great schedule, but this one will be
super dope, of course, 'cause it is the one, the only >> That's the right phrase to use >> Kelsey Hightower
>> to bring me out. >> who is now a principal
developer advocate at Google Cloud. Kelsey, thanks so much for joining us. >> Well, thanks for having me. >> All right, let's start. You did a keynote yesterday
and I actually heard, not only did it rain in San Diego, people were talking about allergies. They were grabbing their tissues,
eyes seemed to be tearing. You had stepped back for a little bit. When I first came into this show, we've been doing it for four years, it was, you know, Kelsey
Hightower and Kubernetes almost seem to get top
billing of the show. You specifically stepped
back for a little bit, and you're here this week. So, talk a little bit about that piece. >> Yeah, so I stepped back to do some serverless stuff, right? So I worked on some cloud
function stuff at Google, launching the ghost support
for cloud functions, and really trying to
understand the serverless base by being in it, and that means stepping back from Kubernetes quite a bit. So the keynote, I wanted
people to have emotion. So no live demos, no
slides, no speaker notes, and then just telling stories
from the last six years of being a part of the
Kubernetes community, and making people feel something. And I think it resonated with folks, and, of course, people
got a little teary-eyed. I gave people a cover,
so we just kept saying the allergies are starting
to flare up in the room, and we really connected with people. >> Awesome. So you came back, which means serverless not completely taking over and obviating what we've been doing here for years. >> Yeah, I think serverless is just another tool in the toolbox, and I didn't want to miss it. So before I put it in its category, I wanted to make sure that
I got super deep with it, used it myself, gave it a fair shot, and it definitely deserves a place. But I think the idea of serverless is the thing that's going to stick. This idea of eliminating as
much infrastructure as possible and then putting that everywhere we can. >> I want to bring that idea
of a tool in the toolbox to what we're talking about at this show. >> Kelsey: Okay. >> So, you know, Kubernetes is one of the most
hottest topic at the show. The CNCF, now I mean, there's dozens and
dozens of projects here. Dan Kohn, when he kicked it
off, talked about Minecraft. And it's like there's that
board there with all the tools, and, oh boy, which one do I
pick, and how do I use it? >> How do you look at
where Kubernetes fits in the overall landscape? Obviously, 12,000 people,
it's really exciting. Why is there so much excitement around something that I think is really, it becomes another tool in the tool shed and baked into the platform? >> I think Kubernetes represents a problem that most people have. If you went down the Linux
and then virtualization path, then you ended up with a
bunch of virtual machines that you need to glue together somehow. So if you look inside
of what Kubernetes has, like the scheduler, how it takes in the pain
of running a workload. If you're running VMs in Linux, this is a problem you already have, so Kubernetes just resonates with almost everyone that
is using virtualization. This is why it's so popular. So it fits. Now every tool in the landscape may not resonate the same way because everyone doesn't
have the same set of problems around the edges, but Kubernetes is a very
obvious thing to anyone that's managing more than
a handful of machines. >> Well, I think that brings
up an interesting question of, as companies and people
assemble the stacks, right, assemble the engines
out of the components, do you have any thoughts on,
well, I guess we could take it from a couple of different ways. But maybe as a person coming
here for the first time, representing their team, getting started, maybe not involved online
with upstream Kubernetes but trying to make sense
of the landscape here and all the different, the
zoo of different projects. >> Lots of new people here. You talk to people, I think, what, 50% or more of the people are brand-new. People have been ignoring, rightfully so, Kubernetes for four or five years. "Maybe I don't need it,
I'm good where I am." But we're at a point now
where you can't ignore it. VMware's offering Kubernetes,
every conference you go, where it's KubeCon or not, this is the thing they're talking about. It's just like Linux
was years prior, right? It's just the thing that people are doing. So now, you're coming to
see for yourself first-hand. You're coming to ask
people how's it going, now that we're five years in? There's a sense of maturity,
things are slowing down, the ecosystem's getting a
lot more mature around it. So you almost have no
choice but to be here because now it's in your world. >> All right, so, there's some people that I've been seeing online that are still looking at
this a little bit skeptically, and said, "You know, we've
been down this path before." You know, "Oh, everybody's
involved in Kubernetes." you know, "There's my Kubernetes "versus some of the other environments." How should we think about that? 'Cause as you said, it's
going to be baked into VMware when they do project-specific, and they've got a couple of
ways to get you to Kubernetes. Yeah, Microsoft just announced an update. Is it an inter-operability issue? Is this the universal backplane? Do you have a good analogy as to how we should be thinking
about where we are today and where we need to go so that we don't repeat
the sins of the past when it was the multi-vendor mess that really didn't solve
the customer's problems. >> You're going to
always have multi-vendors because there's too many customers for one vendor to satisfy. That's always going to be the
case, there's no way around that. But the way I look at Kubernetes now is like, take the web. Click around, webpages,
link them together. And out of that, we extracted REST. People can build APIs,
we build tooling on top, cloud providers built APIs
to manage infrastructure. So the REST component comes out of the larger
picture of the web. And when we take the larger
components of Kubernetes, and we extract out that Kubernetes API, you get Istio, you get
these network control plans, you get people building 5G infrastructure using that Kubernetes model. You get all the cloud providers saying, "Now, if the world's going to have "this set of APIs that
are based on Kubernetes, "then I can actually build
a global control plane "because I can assume that
Kubernetes' API everywhere." Not just for containers,
also for networking, authorization, management systems. So it's only natural that people
start moving up the stack, and I expect even more panes, ever more fragmentation, if you will, because now it's so much
easier to explore a new idea, even if it's only for a
smaller subset of the market. So I expect it to explode. >> Yeah, one of the things
we've been looking at this year is really the simplicity of the offering. You had done Kubernetes the
hard way a couple of years back. We've been looking at things like lightweight Kubernetes, the K3s. How are we with that simplicity
of the overall solution and making sure that Kubernetes
can reach its potential to get to all of those use cases and end points that
you were talking about? >> Kubernetes' job is to
manage the complexity. If you need to run in multiple
regions across the globe, that is a set up complexity, Kubernetes has one way of addressing it by sitting on top of
all those VMs globally, and then providing a set of APIs. That Kubernetes set up end cluster is going to be way more
complex than a MicroK8s, where you have a single virtual machine where you install the
components on one machine, you don't deal with networking, you're not dealing with multiple nodes. That flow is super-easy. I think I did a tweet
for the Canonical folks. They have a tool called MicroK8s, you just run one command, you have a Kubernetes
cluster, and off you go. And that's great for a developer, but as the underlying
infrastructure gets more complex, I think the overall cluster, and the components that
you need in that cluster, matches the complexity. So I think Kubernetes
has proven to scale up, and now you can see it's scaling down. So I think it's one of these things that's adapted to complexity, versus having to jump off of the platform because it can't meet either range. >> Now, Kelsey, we've
talked a little bit about both Kubernetes as this universal API, but also being embedded, right, and being below a lot of application layer and other management-layer things, I mean, did you think about talking to our fellow technologists, right? There are some people who are going to be, we've also used the
metaphor, mechanics, right? There's some people who are
going to be the mechanics, but, like, everybody drives. So, as we get to this level of maturity here now at KubeCon 2019, any advice on how people should pick? Do I need to, and also
online we hear a lot about, "Oh, I don't need, I don't
know if I need Kubernetes. "I don't know if my
particular use case right now, "boy, I don't know if I want to go there." So, I mean, how should
people be looking at it? And also up scaling, should
every IT and technologist and developer be working
towards Kubernetes? >> Absolutely not. >> Thank you. >> If you're managing a bunch of machines, you got two choices. You could build a lot of custom tooling and build something that
looks like Kubernetes, most people don't have
the time to do that. So what we want to do is say, look, a lot of people are collaborating on that obvious thing that you
should build to manage that. Now if I give you 80% of your time back, you should go and fill in that gap between what Kubernetes
brings to the table and what your developers
want to actually do. And at the end of the day, it's
always been the same thing. You check in code, it should adopt the
company's best practice, and I should be able to get an end point and some debugging tools. That has always been the north star, even when there was virtualization,
early days of cloud. Kubernetes is no different. The thing that Kubernetes
represents, though, is that you don't have
to build as much glue between either your own VW
ware or your pre-early cloud. Kubernetes has built all that
stuff way up to this line, so maybe you actually
finish that CICD part you were supposed to do anyway. >> All right, so, Kelsey,
every year we try to figure out and distill down the theme of the event. A couple of years ago, the service matched really
extensions were going at it. Here, there's so many different pieces, it's a little tough to kind of pin down. We talked about some of the
edge simplicity use cases, security has, of course, been a discussion for a couple of years. Anything that you've distilled so far or the things that you are
finding most interesting and new, kind of at the edges of
this whole ecosystem? >> This whole thing is a Swiss army knife, so it depends on who's holding it. Whatever problem they have,
that's the piece of the tool that they're going to
make front and center. So that's what this is. And right now I think there's
a lot of confusion on, do I even need all the other components in this Swiss army knife? Some people are just like, "Well, this tool looks interesting. "I don't have a problem
that this tool is for." And some people are
actively creating a problem so they can use the other
tools in the Swiss army knife. I think the biggest thing that I've seen in the last two years is, make the new thing work the old way. So you're getting the more
traditional vendors showing up and adding their Kubernetes integrations, and they're making the new thing more familiar to the people
who have the existing tool. And when I look around, that's
the thing that I see arise. "Hey, that firewall you were using? "We now have Kubernetes support. "That security tool you were using? "We now have Kubernetes support." The security tool works
fundamentally the same, it's just now easier to adopt and maybe make Kubernetes
things that are deployed in it, leverage those thing. >> So you're saying that's a
good thing, not a bad thing. >> It's a good thing, but it can also be dangerous in some cases where we may get complacent a little bit, and what we end up doing
is recreating the world that we tried to run
away from a little bit. We try to create a little distance and maybe rethink a few
of these approaches, maybe eliminate some need
for some of these things. But if we get stuck in
recreating the old world on top of the new thing, it
doesn't really benefit anyone if we did that for too long. >> Yeah, it's interesting 'cause
you talk to the enterprise and only 20% of applications
are in the cloud, and if you talk about, out
of my entire portfolio, how many are really new
Cloud Native applications? Its much smaller than that 20%. So we know it's the long pole
in the tent of modernization, but you spend a lot of
time talking to customers, you're traveling the world, what are some of the best things
that you're seeing out here that are helping people
adopt those new environments and not just stake a
place in, as you said? >> Pragmatism and leadership, if I see those two things. If there is someone that
can make a decision. I see Spinnaker, I see Jenkins, I see a thousand things,
I see the options. Leadership is pick one. They roughly do the same exact thing. You get someone that
knows what they're doing, hires someone, get some
help, make it work. And then the pragmatism is just be honest about your velocity. You might only bring in the VMs, and then you go to containers. So, this all or nothing
approach never worked. You know it doesn't work. So I think when you have
those two fundamental things, then you see a lot of success. And it's not about the age
of the enterprise, either. There are hundred-year-old
companies are making it work because they have the
leadership component, and they're very skeptical, so they approach the
problem with pragmatism, so they actually get to production. Sometimes faster than the startups that are trying 7,000 things
in more of a reckless fashion, the whole thing catches fire. So, those are the positive outcomes that, there's so many tools now. You have your traditional vendors
now with skin in the game, giving you documentation. I think right now, if you've
got those two components, you're on your path to success. >> Yeah, I guess last thing,
I want to get your thoughts just on this community these days. A couple of the keynote speakers today really talked about project over company, and definitely the open-source ethos is front and center at our show here. Give us your viewpoint
how the community's doing and any highlight you want to share. >> So I have one more thing
on top of that hierarchy, is people over projects always. And then that means that the
people should be able to say, "Hey, I am not wedded
to this project forever. "There's going to be a time
when we have to jump off, "there's going to be a
time when we have to learn "from the other communities." And if you do that, then we can actually be
on the straight path. If we put the projects
too much front and center I think we start to miss the boat. Kubernetes, Kubernetes, and the rest of the world is moving on. And then we look up, we've missed it, and we actually didn't even get to contribute to the new thing. So I think the biggest
part about this community is that hopefully we keep the thing going where we keep reminding people, it's people over these projects. And I think in my keynote, I was trying to address the idea that we're just kind of pacesetters. You come in, you contribute,
all contributions are welcome, documentation, code, or leadership, and then sometimes you
got to jump back out and allow someone else to
come in and set the pace and let the ecosystem become the marathon and let it keep running. >> All right well,
Kelsey, thank you so much for sharing with our community. I tell ya, I've had countless stories of people over the years
that have talked about how they've reached out to you, you've helped them along the way, and I know everybody in this ecosystem really appreciates
everything that you've helped to move this to where we are today. >> Awesome, thanks for having me. >> All right, for John
Troyer, I'm Stu Miniman. Super dope coverage of KubeCon
CloudNativeCon continues. We'll be right back, thanks for watching theCUBE. (electronic beats)
