Ed Warnicke, Cisco | KubeCon + CloudNativeCon EU 2018


>>Announcer: Live from
Copenhagen, Denmark, it’s theCUBE! Covering KubeCon and
CloudNativeCon Europe 2018. Brought to you by the Cloud
Native Computing Foundation and it’s ecosystem partners.>>Welcome back, everyone. This is theCUBE’s exclusive
live coverage here in Copenhagen, Denmark for KubeCon 2018. I’m John Furrier with my cohost
this week, Lauren Cooney, and our next guest, Ed Warnicke, distinguished consulting engineer with Sysco Systems, CUBE alum. Great to see you, welcome back to theCUBE.>>Good to be back.>>So great developer
action, end of day one. We’re going to be here all day tomorrow. So day one’s kind of
coming into the books. Your thoughts on what’s happening here. Different crowd but active.>>No extremely active. Actually, one of the things I’ve noticed, and this is sort of a subtle point when you’ve been around a lot
of open source projects is you have a lot of people who are new to the Kubernetes community are coming in. And one of the things I found
extremely heartening is, they’ve got a really
organized approach to it. When they did their developer summit, they had an entire track for
bringing new contributors on. They’ve just revamped their documentation to help people that are here, and they’re finding better and better ways to articulate the things
that people need to hear to help them make the
leap to Cloud Native. ‘Cause one of the underappreciated things about Cloud Native is that it’s different from the move to Cloud 1.0
that we made a few years ago, is that Cloud Native is not
a lift and shift behavior. You have to change the way you
think about doing your job.>>And that’s the global platform. This is not just a transformation process. It’s a lifetime transformation.>>Absolutely.
>>Huge personnel issue. People process technology,
technology last one.>>Ed: Have you accepted
Cloud Native into your heart?>>I have come to terms with my– (Ed laughing) lift and shift problem that I have, and I’m now aware,
self-aware of Cloud Native.>>The first step is to
admit you have a problem. (John chuckling) The making amends to your
infrastructure takes longer.>>I mean, look if–
>>And you would know, so. (laughing) Well, anyways.
>>We’re all working on it.>>So, I have a question for you here. As you were talking about
how you’re seeing a lot of new developers coming on in
and things along those lines. I’m also running into
a lot of new developers at the hotel, at dinner, just walking around
and having discussions. Where do you see these guys coming from? I see them coming from banks, from large technology companies
that are based in Europe. Where are you seeing these folks?>>So that ends up matching very closely with what I’m seeing as well. From all over the place. From people who finance
large energy projects, right? From all areas of finance. Basically, all the sorts of people who have big compute problems
are starting to turn up at the Cloud Native world because this is literally
where you solve those problems. And I think that’s part of
what’s driving the ecosystem is, the folks in Kubernetes made a number of incredibly intelligent
decisions early on about how their architecture was built in terms of the modularity
and expandability of it. And the result is that
you get lots of people with lots of energy coming in saying, “I have a problem like this.” There’s an obvious well-worn
path to try and put together a proposed solution for
solving problems like this. And they engage with the community. One of the things that
you’re seeing just in terms of how the community grows itself is, they’ve got special interest groups, SIGs for various areas in Kubernetes. They’ve now had to spawn working groups that come under them. You’re just seeing things
like Kubernetes proposals for how you’re going to
do things coming to far. So there’s a lot of the
maturity process that you expect to deal with the scale of people who want to solve their problems this way.>>So you’re actually not seeing sprawl. You’re seeing highly organized
groups coming together in a way that can make the
platform more positive.>>Yeah, absolutely. Not only am I not seeing sprawl, but I’m starting to see highly
intelligent things being said by the people who work at what we think of as the core of Kubernetes. So I’ve heard a number of
people make the comment that they expect the Kubernetes
core to actually shrink in terms of what it offers because the broader
ecosystem is picking up so much of the slack. So this sort of core APIs of,
this is what is Kubernetes without having picked out some
options that meet your needs, is keeping itself very tight
while having architected it in a way where you could
have this broad ecosystem without the kinds of
problems you sometimes get with sprawl in other communities.>>So whilst you want to get bigger, but you’ve got to get
smaller to get bigger.>>In some sense, yeah. You have to decide what’s really important to get right in the
core and really nail it.>>What are they getting
right, in your opinion? What’s right about it that’s going on? You mentioned some of the smart decisions that they’re making.>>So, a couple of the things that they’ve gotten
really, really right are our relentless focus on developer needs. So I see this particularly in networking, and I think we’ve talked
about this before. Developers don’t want
to know about subnets. They don’t want to know about L2 segments. They don’t even want to know
about IP addresses, frankly. What they really care about is two things. Reachability and isolation. Everybody can talk to everybody unless they decide you should be isolated. And service discovery and service running. Those are the only two
things they care about, and wouldn’t you know it! In Kubernetes, you have network policies that control the
reachability and isolation and services that do services discovery and service routing for you. So they’ve absolutely nailed the fundamental developer needs.>>Made you pain point.>>Yeah.>>So what’s your take
on just the ecosystem. Obviously, we’ve commented, and this is always a dangerous
game with communities, is logo farm, everyone’s here, right?>>Yeah, I mean, they took
the CNCF logos and probably, I think they broke them
into three categories now. I’m not exactly sure what that means.>>John: A whole new
sponsorship level for–>>Architecture? I’m not sure.
(John and Ed laughing) But, Ed, maybe you could
provide some clarity here.>>Well, I mean, there is a certain risk in being loved to death, right? Kubernetes is full blown into what I will sometimes
call crises of success, which is, you are succeeding so wildly that it’s beginning to be a problem. And that’s good to see. But I think you’re starting to see certain categories of
things that are emerging. And there was a good set of
readouts from the various SIGs to Kubernetes yesterday
in the developer summit. So you’ve got a bunch of
stuff around networking. You have a bunch of things around storage. These are sort of fundamental
infrastructure issues. But you have a bunch of things, literally, about how so we expand
the Kubernetes platform. How does that work? How do we produce the
constructs we need to solve the various problems that are arising, and those things are all sort of progressively moving forward. And we’re getting to sort
of the interesting point where the people who did the
original turn of the APIs are being really blunt
and honest saying, “Look! “These are the things we got right, “and these are the things we got wrong.” And there’s a lot to be
said for having that level of honesty with yourself
on stage in public, right? When you’re the guy who wrote the code, it’s unequivocally your mistake. And being able to stand up and say, “Look, “we got this one wrong.”>>But that’s the community
trust that you have, and that’s what makes the community.>>And that trust goes both ways. It’s the trust of the community in that leader standing on the stage, but it’s also the trust of that leader that we’re going to move fast,
we’re going to do things right, but there’s always a turn of
the crank to do things better. And we got to be
straightforward about that.>>And their self-awareness
around the iteration is key. They’re putting their egos at the door, checking it at the door,
focusing on the advancement. I got to get your thoughts,
from both of you guys, I want to ask you guys both a question. I know that you’re doing a lot
a work with some start-ups, and you with Cisco, the big company. What’s interesting
about this ecosystem is, the balance between the big players and the enablement for the small
start-ups to be successful. We had a variety of start-ups
here with news on theCUBE. This is the give get
between sharing in projects where there’s a balance and everyone can thrive and
survive and grow together. Thoughts on that balance. Start-ups have needs, but they’re not as big as the big guys. So what’s your thoughts on–>>Why don’t you start, Ed.>>Well, to begin with,
we can’t do everything much as we would like to. Back to the self-honesty, you have to be honest
with yourself about that. And nobody has a monopoly
on the good ideas. And so you really have to
engage with the ecosystem and figure out how different aspects of the problem knit together. I’ve had a lot of
interesting conversations. I, personally, have some interest in what I sort of call unified IO. So converged networking storage. So I’m talking to a lot of folks who are doing storage stuff, lot a little start-ups that are doing really
cool things with storage about things we can do to help them there from the network side, and they’re excited about that, right? And it’s that, that’s the
sort of open source spirit that makes it possible to
have all these start-ups because, I’ll be really frank, most of these start-ups, if they were having to try and
build the thing themselves, they’re simply not resourced to do it. But with so much support from
the community in the broad, on a relatively thin start-up budget, you can move mountains.>>Yeah, if you tap the formula
properly, that’s the key.>>The start-ups are getting
more and more sophisticate about tapping that formula because only… Getting a good product is only a very small part of the equation. You also have to get the
connection with the community because you have to make sure, even if you’re entirely self-interested, if you build a thing, there will be a thing in the
open source that does that. And it is a fundamental
truth in the modern era that 80% of the value or
more of all software is its connection to everything
else in the ecosystem.>>Lauren, I want to get
your thoughts on this. You’re doing this now as a new start-up, you’re a founder of and running, but you’ve built programs. Modern architectures at play here. You’re seeing microservices
growth phenomenal. Cloud Native is just
whole nother ball game, going to a whole nother level. As you’re engaging out
there, what are you seeing for this modern community
formula playbook, whatever you want to call it. There’s a way to do things
now at a whole nother level that this is going.
>>No, I–>>Your thoughts.
>>I definitely agree. I think the developer
experience is really key, making it simple, making
it just seamless, right? So folks don’t have to
wait to download something, or they don’t have to wait for, you know. They can just click a
couple buttons through a GUI and make it really, really simple, especially those on-boarding. What I see from the
start-up side is a lot of… This is interesting because
I think it’s important. A lot of start-ups coming from companies that wouldn’t allow them to do open source inside the companies. So they’re leaving these larger companies, and they’re doing start-ups. They’re raising pretty good capital for seed rounds and A rounds. And I think, this is something
that’s pretty hot right now and we want to take a look at. And the VCs are definitely looking.>>What about the big
companies that we all know, obviously Cisco, IBM, you see Amazon here. They have huge scale. Even Microsoft has had developer
programs been successful over the years, we all know that. What’s the modern tweak
that they’re making that you’re seeing work?>>Oh, I think it’s the small teams. Adrian was on here earlier
talking about microservices and micro-teams, and I
think he’s absolutely right. You have to have teams that
are building these services that are moving quite quickly and doing it in a way that’s
rapid enough to keep up or be ahead of the market.>>The micro-team point, I
think, is actually really apropos because… This is going to sound very
engineering propellor-head, but the management overhead
gets to be quite steep when you try and do anything
with big teams, right? So you got to have very loose coupling to everything else in the system, which is exactly what
Cloud Native is about. And that’s what you see
not only in the start-ups but you see these sort of
hybrid approaches emerge, where you have a start-up
that has a small team and another start-up that has
a small team that’s nearby and a large company like
Cisco that has a small team, and there’s an interaction
between all of these. And we’re sort of
operating as the growing up of this larger team
completely across boundaries. It’ll resolve actual user problems.>>I think it’s a historic time. I think you guys are right on. This is such an exciting time for, if you’re an engineer, software developer, or anyone in large-scale systems, and building applications is
going to a whole nother level. Look at blockchain
right around the corner, decentralized applications is coming soon. We won’t go there in this
interview ’cause it’s KubeCon, but I got to get your take. What’s your view so far
of what’s working here, hallway conversations you’re having? What are some of the things going on here that someone who’s not here
might want to know about?>>I tend to be very focused
on networking things, so the thing that I’m most excited about that’s happening here is, the entire world seems to
be getting meshy, right? So there’s a huge excitement
around service mesh and Istio, which I think is extremely well-placed. The fundamental thing that’s
really happening there is, they’re progressively
taking parts of the problem that you’re not good at if
you’re writing a microservice, and they’re pulling them
out into a sidecar envoy so that you don’t have to worry about service discovery
and service routing. You don’t have to worry about the policies for how you’re going to figure out what things you do about
getting to the next guy in the chain of the work. You don’t have to worry about
even things as simple as making sure that you respond
to faults well, right? And there’s a whole new
set of ways that you think about problems in this
space that’s emerging there. One of the things that I’m
actually really excited about that’s also meshy is when
you get to things like people who have less
common network problems. So the operators with NFV, people who have more
sophisticated network needs. We’re starting to reimagine that stuff in the language of service mesh, right? So rather than trying to
force all the legacy thinking about networking into Cloud
Native where it’s not wanted, we try and recast the problems we have into Cloud Native ways
of thinking about them. And I think that ends up
being intensely powerful. It’s, frankly, almost overwhelming because there’s so much
conceptually going on in this space that you want to be able
to draw on for the palette for the things that you’re painting.>>Yeah, I mean, it’s
your point earlier about, and you were kind of joking but serious. This is a mind melt, you got
to buy in to the philosophy of this new era of…
(Ed laughing) Yeah, just kind of buy into, the Cloud Native is a global platform. It is a fundamental new thing. It’s not just a
methodology, it’s a new way.>>It’s a new way of
thinking about things. The C in Cloud Native does
not stand for container. Container is the smallest
possible chunk of this. If you just slap all your
applications into containers and try and do a lift and shift, you’re going to fall on
your face really hard.>>John: In what areas? Just like, what?>>Well, I’ll give you
a really simple example. Let’s say that I have an application that I’m running in vApps, right? And I’ve got my big database VM. I’ve got my big web front VM. So I pick them up, I containerize them, I drop them into Kubernetes. So I’ve got one replica of my database VM and one replica of my web front VM, and that’s going to break
sometime in the first 24 hours. Because I need to, basically,
pick them up and say, “OK, I need a bunch of replicas that are dynamically coming
up for all of these things. I need the services to
wire mesh them together.” So for whatever reason, I lose
some number of my replicas, that everything comes
back up and goes forward and we never even notice, right? In some sense, the ideal situation is, you have a major bug in your code, right? Let’s say you have a piece of code that’s leaking memory and
it dies every 24 hours. You want, if you think about
it right and you deploy it ’cause you don’t know you have this bug, you won’t even notice that
you screwed up that bug because the infrastructure
will protect you from it. But if you just try and lift and shift, you’re not going to
have a happy experience because it’s not going to
work the way you expect it to.>>And then monitoring tools
are getting better, too, and so if you’re coming in on
the other side you get that. Well, and thanks so
much for the commentary. Great, great summary of the event. Any surprises here for you? Any ah-ha moments or
revelations or epiphanies or any kind of surprises,
good or bad or ugly?>>One of the things I was
very impressed with is, I’m very impressed with what
you can do with no code. I don’t know if you saw
that keynote this morning.>>Lauren: With Kelsey.>>In response to Dan Kohn’s point about all the sort of
total attack surface area. Kelsey got on stage and
did the no code project, which has perfect security for whatever it is that you deploy it for. The fact that you can get on code, do something like that,
move an entire audience of thousands of people, that’s impressive. You don’t see speakers
who do that very often. That was, I wouldn’t say shocking, but very much a pleasant surprise. And it speaks very much to
the code of the community. The keynotes today were some
of the best I’ve ever seen. I am not a keynote person,
I seldom attend them. The keynotes today were
extremely well-done. They had good energy
and they were relevant. The walking through of the
evolution of the community in brief punctuated explanations of what’s going on and
why they’re important, I’ve never seen it done better.>>Yeah, they were
hitting their marks well. Well, great, thanks for coming on, Ed. Great to see you.
>>Yep.>>Thank you, Ed.
>>This is commentary from Copenhagen, Denmark. It’s theCUBE coverage of the CNCF, Cloud Native Compute Foundation, part of the Linux Foundation,
KubeCon 2018 in Europe. I’m John Furrier, Lauren Cooney. Thanks for watching. Be right back. (electronic musical flourish)


Please enter your comment!
Please enter your name here