►
From YouTube: Lightning Talk - Realizing the Multi-Cloud Promise of Kubernetes by Blake White, The Walt Disney Co.
Description
Lightning Talk - Realizing the Multi-Cloud Promise of Kubernetes - Blake White, The Walt Disney Co.
Disney's diverse business units and applications require running in multiple cloud environments. This talk will touch on some of the tools and techniques used to realize the cross cloud promise, as well as some of the challenges and their solutions.
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.
A
Hello,
my
name
is
Blake
white
and
I
work
at
the
Walt
Disney
Company
with
our
cloud
architecture
team,
one
of
Disney's
core
principles
is
making
magic
happen.
Sometimes
it's
easy.
Sometimes
it's
not.
We've
been
exploring
kubernetes
for
most
of
the
last
year
in
our
own
OpenStack,
environment,
Google's
cloud
and
in
AWS.
Here
are
some
of
the
things
we've
learned.
A
First,
things
that
we
thought
about
were
placement.
When
working
with
multiple
development,
QA
and
engineering
teams,
you
have
to
be
very
flexible,
be
able
to
work
within
what
whatever
confines
they
want
to
work
within.
We
need
to
think
about
connectivity
and
data.
Different
clouds
offer
different
approaches
to
connecting
back
to
a
corporate
network.
Does
the
project
you're
working
on
need
to
be
able
to
reach
code
repos
artifacts
other
services
in
your
corporate
network?
Is
there
data
that
needs
to
follow
certain
privacy
standards?
How
much
latencies
latency
is
tolerable?
A
A
Aws
has
crossed
VPC,
be
careful
with
that,
because
sometimes,
when
you've
got
something
one
VPC
connected
back
to
your
direct
connect,
it
can't
connect
to
something
else
in
a
different
VPC.
So
there's
some
some
things
to
watch
out
there
once
you
figure
out
what
your
data
depends
see
is
what
cloud
you're
gonna
land
in
and
all
of
that
you
can
figure
out
how
you're
going
to
get
your
cluster
up
and
running
in
Google.
A
You've
got
gke,
obviously,
and
it's
great
you
can
get
a
cluster
up
and
running
just
a
couple
of
minutes,
but
if
you
need
some
more
configuration
and
customization
you'll
want
to
go
with
GC
at
GCE
or
another
cloud
entirely.
If
you
need
to
use
another
cloud,
the
ways
that
you
can
bring
this
up-
automation,
you'll,
probably
want
to
use
something
based
off
of
cube
up
or
cops,
is
a
fantastic
tool
for
operationalizing
kubernetes.
A
A
The
trickiest
part
that
we
had
with
that
was
setting
up
the
dns.
We
moved
from
Sky
DNS
to
cube
DNS
that
helped
the
cluster
a
lot,
but
in
AWS
things,
just
weren't
working.
So
basically,
our
deed
CP
set
for
the
V
PC
was
skipping
the
Amazon
internal
and
pointing
just
back
to
our
corporate
network,
which
is
what
we
needed.
But
kubernetes
was
unhappy
because
it
couldn't
find
all
the
nodes
couldn't.
A
So
we
set
up
a
bind
server
pointed
that
at
the
AWS
internal
for
internal
stuff
and
back
to
our
corporate
network
for
everything
else
for
our
internal
stuff
and
normal
DNS,
everything
started
working
again.
That's
the
sort
of
thing
that,
unless
you're
really
in
there
digging
with
the
add-ons
figuring
out
what
the
VPC
set
looks
like
and
watching
where
the
dns
is
resolving
to
you,
wouldn't
really
notice
that
in
something
that
was
brought
up
through
automation,
once
you've
got
your
cluster
up
in
unfunctional.
It's
time
to
talk
about
operation
of
the
three
things
up
here.
A
A
lot
of
people
are
talking
about
deployments
and
updates,
and
monitoring
and
scaling
I'm
going
to
talk
a
little
bit
about
logging
in
an
enterprise,
especially
across
clouds.
Stackdriver
things
like
that
are
fantastic,
but
it
lives
within
your
account.
How
are
you
going
to
query
that
across
accounts?
You
don't
want
to
ship
everything
to
a
central
repository
because
then
you're
paying
for
egress
and
you're
paying
for
all
of
that
sort
of
thing.
So
one
of
the
main
things.