►
Description
Red Hat’s senior leadership is having to execute at an ever-increasing pace. That means that today's technology decisions have to balance short-term risk with long-term gains. This unique series provides host Chris Short inviting thoughtful and candid discussions with each guest.
An hour with the one and only Joe Fernandes, VP & GM Core Cloud Platforms
A
Hello
good
morning,
good
afternoon,
good
evening,
sorry
for
the
brief
technical
difficulties
there
displaying
our
zoom
schedule
is
always
fun:
hello,
everyone,
I'm
chris
shorts,
principal
technical
marketing
manager
and
executive
producer
of
this
wonderful
thing.
We
call
open
shift
tv.
Thank
you
for
tuning
in
wherever
you
are.
I
am
joined
today
by
the
vp
and
gm
of
core
cloud
product
core
cloud
platform.
Sorry,
it's
a
mouthful
joe.
It
is
go.
C
C
C
I
was
a
the
product
manager
for
our
initial
launch
of
openshift,
so
openshift
1.0,
which
we
launched
in
2012..
I
built
out
the
the
openshift
product
management
team
drove
things
like
our
pivot
to
kubernetes
around.
You
know,
openshift
3
in
2015.,
and
then
most
recently
took
on
our
openstack
and
red
hat
virtualization
solutions.
C
So
I've
been
doing
a
lot
around
those,
as
well
as
how
we're
converging
vms
and
containers
together
through
openshift
virtualization
and
the
cube
project,
and
then
most
recently,
I
I
added
sort
of
product
marketing
and
our
partner
teams
and
so
kind
of
managing
that
all
together.
But
it's
been
a
great
ride.
I've
been
here
now
it's
like
10
years,
great
company
and
just
really
amazing
people
that
we
get
to
work
with
here
every
day
and
amazing
technology
and
customers
as
well.
So.
A
Yeah,
like
that,
that
is
the
one
thing
I
think
brings
me
back
every
day
is
the
the
the
the
customers,
the
people
and
the
technology
right
like,
and
you
know
I
write
a
newsletter
as
people
process
tools
is
how
it's
organized
and
like
redhead
nails,
all
those
in
my
opinion,
so
you
know
and
you've
been
here
since
open
shift
one.
So
just
to
give
you
some
context
back
in
those
days
I
was
using
openshift
to
run.
You
know,
phishing
campaigns,
you
know
to.
A
Test,
like
our
you
know,
cyber
security
like
assurances
and
you
know,
training
capabilities
and
how
good
they
were
yeah.
So
I've
I've
used
openshift
since
the
1.0
days,
but
I
didn't
join
until
I
didn't
join
the
actual
team
until
like
right
before
the
4.0
launch
and
we
launched
that
and
we
launched
openshift
4.0
and
it's
kind
of
changed
the
kubernetes
landscape.
I
feel
like
we
did
that
at
summit.
What
were
some
of
the
goals
for
this
major
release
and
where
do
we
stand
today
with
those
goals.
C
Yeah,
so
so
we
we
pivoted
like
to
to
kubernetes
first
back
in
2015
when
we
launched
the
openshift
three
right
and
had
a
really
great
run
there,
but
there
were
some
some
key
things
that
we
wanted
to
accomplish
with
with
the
openshift
4
launch,
which
was
just
over
a
year
ago
now
at
summit.
So
first,
you
know
we
want
to
reaffirm
our
commitment
to
that
kubernetes
core
right
to
providing
the
best
kubernetes,
most
stable,
most
secure,
q-ready
solution
on
the
market,
but
also
bring
that
into
new
workloads
and
you've.
C
Seen
us
do
that
in
areas
like
ai
and
machine
learning,
analytics
data
services
vms
with
openshift
virtualization,
we'll
be
coming
out
with
our
windows
container
solution
here
really
soon,
as
well
as
use
cases
like
edge
right,
but
all
built
around
this
kubernetes
core
that
that
you
know
we
first
sort
of
brought
to
market.
You
know
back
in
2015.
C
The
second
thing
is,
you
know
one
of
the
ways
we've
distinguished
ourselves
is
being
able
to
support
that
across
a
hybrid
cloud
environment,
but
we
need
to
get
better
in
terms
of
how
we
manage
openshift
and
kubernetes
in
these
different
environments.
So
we
acquired
core
os
and
we
set
about
rebuilding
how
we,
you
know,
install
and
manage
the
platform
with
technologies
like
operators
with
technologies
like
container
linux,
which
became
rel
core
os
and
then
extending
sort
of
the
kubernetes
declarative
apis
into
managing
the
infrastructure
beneath
us
right.
So
machine
controllers
managing
machines.
C
A
So
so
you
hit
me
right
in
the
heart
right
with
the
word
devops
and
get
ops
right
so,
like
I
foresee,
or
I
see
now
get
ops
as
kind
of
being
the
way
to
implement
devops
in
a
kubernetes
environment.
That
will
also
get
you
that
culture
gain
of
devops
without
actually
having
to
go
through
a
lot
of
like
the
realization
of
culture
changes.
You
know,
and
we
recently
announced
our
teaming
up
with
argo,
but
you
can
do
get
ops
any
way
you
want
on
kubernetes
right,
like
our.
B
C
Yeah,
so
ultimately,
git
ops
is,
is
really
around
starting
with
git
as
the
single
source
of
truth
for
both
provisioning,
the
infrastructure.
So
you
know
we
use
git
as
a
single
source
of
truth
for
how
we
manage
the
configuration
of
the
openshift
and
kubernetes
clusters,
but
then
also
for
the
managing
the
configuration
of
the
applications
that
you
deploy
on
that
cluster.
So
so
you
start
with
get,
you
know,
manage
your
configuration
of.
C
You
know
your
clusters
or
your
applications
in
a
controlled
manner
and
then
leverage,
in
our
case,
argo
cd,
to
make
sure
that
you
know
that
that
state
is
always
maintained
or
if
you
need
to
make
changes,
you
don't
make
it
by.
You
know
going
directly
to
the
application
or
you
know
going
to
the
cluster
itself.
You
change
that
by
going
to
git
updating
that
configuration
and
then
having
that
drive
the
changes,
and
so
so
for
a
lot
of
customers.
This
is
how
they
want
to
manage
their
infrastructure
and
their
applications.
C
We're
lucky.
You
know
to
have
a
great
partner
into
it
that
launched
the
argo
cd
community
and
we
like,
we
always
do
join
them
upstream
and
we'll
eventually
bring
that
as
a
fully
supported
capability
in
the
product.
But,
as
you
mentioned,
it's
not
the
only
solution
right.
So
if
customers
want
to
use
other
get
ops
get
ops
tools,
they
can
certainly
use
that
with
openshift,
and
that's
also
been
a
model
of
ours.
A
A
All
the
stuff
on
the
chalkboard
blinds
and
everything,
and
it's
just
the
cncf
landscape
instead
behind
him,
instead
of
the
chalkboard,
with
all
the
lines
and
drawings
and
cards
and
everything
and
yes,
red
hat-
does
provide
an
opinion
with
some
of
the
tooling
that
we
provide
with
openshift,
but
we
don't
prevent
you
from
having
your
own
opinion
of
what
you
should
do
with
kubernetes
right.
I
think
that's
the
coolest
piece
about
openshift
is
that,
yes,
we
have
an
opinion.
A
A
C
So,
look,
you
know
we
first,
we
start
with
sort
of
our
core
right
and
our
core
is,
you
know
our
linux
distribution,
you
know
rel
and
now
rel
core
os
and
our
sort
of
curated
kubernetes
that
runs
on
top
right.
So
that's
the
core
engine
that
sort
of
is
the
one.
Is
the
constant
that's
what
you're
certifying
and
you
know
managing
applications
to,
but
anybody
that's
tried
to
build
their
own
qra's
platform
or
work
with
third-party
solution.
Knows
you
need
more
than
kubernetes
and
linux
to
get
anything
to
work
right.
C
You
probably
want
a
network
network
to
manage
the
connection
between
those
containers.
You
probably
need
to
get
traffic
into
the
cluster,
so
you
need
some
form
of
ingress.
You
probably
want
to
monitor.
What's
going
on
with
the
cluster
and
the
application,
so
you
get
into
prometheus
and
then
you
probably
want
to
manage
the
logs
and
then,
as
you
move
up
the
stack,
as
you
mentioned.
C
What's
this
platform
for
right
it's
for
building
and
deploying
and
managing
applications.
So
you
want
to
have
build
tools
and
ci
and
cd,
and
now
you
know,
git
ops
tools
and
other
application
services
to
make
that
easier
service
measures
and
storage
and
so
forth.
So
so
all
of
those
things
you
know
we
try
to
sort
of
bring
together
in
this.
You
know
solution
that
we
call
the
openshift
container
platform.
We
have
made
some
decisions
right.
C
We
have
an
sdm
that
we
fully
support
and
include,
but
it
hasn't
precluded
us
from
supporting
some
of
our
great
partners,
whether
that's
cisco
or
juniper,
or
nuage
or
others,
who
also
have
sdn
solutions
that
customers
can
use
even
vmware
with
nsx
and
so
forth.
It's
the
same
thing
you
know
with
with
our
storage
solutions.
It's
the
same
thing
with
our
log
management.
C
A
lot
of
our
customers
will
use
splunk
with
openshift
will
use
their
own
elk
or
efk
stack
with
openshift,
and
then
certainly
you
get
even
more
as
you
move
further
up
right.
You
get
into
more
diversity
because
you
know
not
everybody's
going
to
agree
on.
You
know
what
their
proper
cicd
solution,
exactly
not
everybody's,
going
to
agree
on
with
the
development,
and
so
without
that
flexibility,
you
really
you.
You
can't
really
serve
the
broad
market
that
we
need
to
serve,
because
customers
have
really
good
reasons
for
the
choices
that
they've
made.
A
Yes,
so
that
partner
ecosystem
is
expanding
rapidly
right
and
we've
we've
built
out
operator
hub
and
like
this
bevy
of
operators
with
our
partners
that
have
just
absolutely
made
like
running
your
services
on
top
of
kubernetes,
a
click
button
experience
for
in
a
lot
of
cases
right
if
they
are
a
full,
you
know
full
fully
capable
fully
mature
operator.
Then
they
have
all
the
capabilities
of
you
know
a
service
right
for
lack
of
a
better
term
running
in
some
cloud
provider.
A
You
know
right
there
at
their
fingertips
and
you
know
what
was
it
like
when
we
said:
okay,
we're
going
to
build
operator
sdk
we're
going
to
build.
You
know
operator
hub
we're
going
to
have
this
thing
built
into
kubernetes
or
built
into
openshift
directly.
I
always
conflate
kubernetes
and
openshift
yeah.
That's
my
life
now
yeah
exactly
right,
yeah!
So
the.
How
did
that
operator
hub
and
operator?
A
Sdk
decision
come
about
because
now
it's
in
cncf
landscape
or
in
the
cncf
like
ecosystem
right,
like
we've,
donated
it
to
cncf,
and
we
expect
it
to
grow
and
you
know
bring
in
q
builder
and
other
folks
right
like
and
you
know
be,
you
know,
part
of
a
larger
ecosystem
of
operators.
What
was
that
decision?
Making
process
like.
C
Yeah,
so
so
so
really,
you
know
the
operator
model
right
really
originated
at
corollas,
with
some
great
work
that
you
know,
grant
phillips
and
others.
There
did
right
to
sort
of
define
this
model,
for
how
do
you
bring
day
two
management
to
the
services
that
run
on
top
of
kubernetes
using
in
a
kubernetes
native
way?
Right,
because
you
know
kubernetes
does
a
lot
for
the
applications
and
services
that
run
the
platform
right
it
can.
C
It
can
sort
of
you
know,
deploy
those
scale,
those
up
manage
them
in
a
declarative
way,
restart
instances
when
they
fail,
but
then
it
sort
of
drops
off
right
as
you're
dealing
getting
into
more
complex
services.
You
need
to
do
more
than
scale
up
and
down
and
restart
on
fail
right
if
you're
running
a
database,
yeah
message
queue,
there's
a
lot
of
other,
whether
it's
you
know,
data
replication,
whether
it's
you
know
rebalancing
whatever,
there's
a
lot
of
other
things
that
need
to
happen
that
normally
are
done
by
human
operators.
C
And
that's
really
what
the
operator
model
enabled
the
underpinnings
of
operators
are
crds
right,
kubernetes,
which
has
been
one
of
the
biggest
innovations
and
by
the
way,
one
of
the
innovations
that
red
hat
really
helped
to
drive
in
terms
of
our
upstream
contributions,
because
we
needed
better
ways
to
extend
kubernetes
itself
right
and
guru,
maze,
you
know
hasn't
historically
been.
You
know,
going
all
the
way
back
to
1.0
that
easy
to
extend
with
crds.
C
We
now
have.
You
know,
I
think,
a
workable
extensibility
model,
and
so
you
know
crds
and
the
work
that
red
hat
and
others
have
done.
There
came
together
with
this
sort
of
operator
model,
and
so
when
we
acquired
coreos,
you
know
we
knew
that
this
is
something
that
we
needed
to
sort
of
bring
to
the
community.
C
So
we
did
donate
the
operator
framework,
which
includes
you
know
the
operator
sdk,
making
it
easier
to
build
operators
operator,
lifecycle,
managers
demanding
the
lifecycle
of
those
operators
and
then
work
around
sort
of
telemetry,
so
metering
and
monitoring
of
those
operators.
So
you
know
just
judging
by
the
most
recent
couple
of
kubecons:
it's
not
just
red
hat
talking
about
operators
anymore
right,
like
every
vendor
customer
you're,
talking
about
how
they're
making
use
of
this
technology.
It's
just
really
really
exciting.
So
yeah.
B
A
C
Yeah-
and
I
think
you
mentioned
a
couple
things
that
are
enabling
in
that
ecosystem
right,
we
want,
we
want
make.
We
want
the
community
to
have
a
place
where
they
can.
You
know,
you
know,
find
operators
as
well
as
contribute
their
own
operators,
and
that's
really
what
operator
hub
is
all
about
and
operating
io
is
that
destination
which
we've
also
sort
of
now
contributed
to
cncf.
As
part
of
that,
you
know
the
completion
of
the
operator
framework
and
then,
for
you,
know,
enterprise
customers.
They
want
a
place
where
they
can
buy.
C
You
know
or
sort
of
procure
and
find
certified
operators,
operators
that
are
supported
either
by
red
hat
or
some
other
vendor.
They
make
it
easy
to
find
easy
to
procure
and
so
forth,
and
so,
just
this
week
we
launched
the
red
hat
mark
marketplace
and
for
the
red
hat,
marketplace
is
sort
of
the
commercial
side
the
you
know
for
folks
who
want
to
find
operators,
and
you
know
and
more
that
they
can
sort
of
get.
C
A
B
A
C
C
Yeah
with
the
global
banking
system,
if
it's
a
it's
in
a
stock
exchange,
if
it's
a
major
retail
asset,
you
know
all
these
different
customers
that
red
hat
has
worked
on
with
for
now,
20
plus
years.
That's
the
trust
that
they
place
in
us
and
they
place
in
the
partners
that
we
work
with
so
yeah.
So
it's
great.
A
You
mention
our
customers
right
and
every
time
you
know
we
have
an
internal
message
that
goes
out
about
our
customers
every
so
often,
and
every
time
I
see
it.
I
I
take
great
pride
in
the
fact
that,
like
so
many
of
the
services
and
things
that,
like
we
use
in
our
home,
are
powered
by
openshift
and
like
our
banking,
our
cell
phones,
all
this
stuff
has
openshift
as
one
of
its
like.
C
Yeah
well
so,
first
of
all
I'll
start
with
you
know,
kudos
to
our
customers
and
and
like
yeah.
If
you
haven't
seen
our
customer
service,
you
can
go
to
openshift.com
customers,
there's
just
some
amazing
customer
stories
about
how
those
customers
are
using
kubernetes.
You
know
powered
by
openshift,
using
containers
in
in
some
of
the
most
interesting
applications
and
use
cases
out
there.
You
know
and
companies
that
you've
heard
of
right
companies,
like
you
know,
like
bmw,
custom
things
like
deutsche
bank
and
barclays.
C
You
know
not
only
financial
services
but
in
retail,
in
healthcare
and
telecommunications
and
so
forth,
all
different
use
cases,
but
all
with
a
common
theme
of
wanting
to
bring
sort
of
greater
agility
greater.
You
know
flexibility,
yeah
so
forth
into
the
applications
and
workflows
that
they
care
about
the
most
so
in
terms
of
and
and
by
the
way,
that's
you
know
not
just
powered
by
openshift,
but
this
amazing
kubernetes
community
right
right.
I've
been
around
long
enough.
C
So
so
these
these
evolutions
have
happened.
It's
just
that.
You
know
with
qra
as
a
container,
I
feel
like
it's
happening
much
faster
and
we've
been
right
right
in
the
middle
of
that.
But
to
your
point
you
know:
what's
next,
it's
just
continuing
to
evolve
right
and
in
the
areas
I
mentioned
right.
We
we
definitely
see
an
evolution
in
the
types
of
workloads
right.
So
a
lot
of
the
use
cases
and
customer
stories
you
see
recently
are
about
things
like.
C
How
do
you
run
ai
and
machine
learning
in
this
sort
of
you
know,
elastic
container
cloud
native
environment
and
so
forth?
How
do
you
bring
other
those
workloads
to
new
footprints
like
edge
right?
So
we're
doing
a
lot
of
work
now
around
extending
the
hybrid
cloud
out
to
the
edge
and
dealing
with
the
realities
of
what
what
you
can
run
in
those
constrained
environments.
When
you
don't
actually
have
humans
there
to
install
and
upgrade
and
manage
those.
C
You
have
hundreds
of
potentially
thousands
of
locations
to
manage,
so
multiple
management
work
we're
doing
with
with
acm
our
multicultural
management
solution
is
key,
so
that's
important,
but
ultimately
too
it's
about
how
do
we
make
those
developers
more
productive
and
that's
where,
like
just
you
know
and
again,
you
saw
this
in
linux.
You
saw
this
in
virtualization,
but
this
ecosystem
of
solutions,
whether
it's
service
meshes
or
serverless,
and
function
as
a
service
pipelines.
You
know
you
know,
that's
you
know
those
are
so
it's
not
one
thing.
It's
all
on
all
those
things.
C
New
workloads,
new
footprints
and
new
capabilities
to
sort
of
to
power
developers
all
on
this
sort
of
common
foundation,
which
I
think
is
going
to
be
enduring,
is
going
to
be
around
you
know
we'll
be
talking
about.
You
know
kubernetes.
You
know
five
years,
ten
years
from
now
in
the
same
way
that
we've
been
talking
about
linux
for
the
last
20
plus
years
and
virtualization
as
well.
A
B
C
Yeah
and
but
but
you're,
what
you're
really
driving
in
is
it's
not
about
containers
or
kubernetes
right.
Certainly,
that's
the
some
of
the
underpinning
technology.
It's
about
those
use
cases
right,
right,
yeah,
just
a
plug,
for
you
know
three
really
great
customer
stories:
one
exxon,
mobile,
the
second
hca
healthcare,
which
built
the
analytics
application
for
for
detecting
sepsis
and
hops.
Hospitals
and
now
is
doing
stuff
around
things
and
then
most
recently
royal
bank
of
canada,
which,
with
their
borealis
ai
platform,
you
know
they're.
C
You
know
all
three
customers
are
using
openshift
and
they're
solving
it.
You
know
some
really
important
use
cases.
One
is
on
the
front
end.
How
do
you
get
those
data
scientists
to
collaborate
as
they're,
building
and
refining
their
models
right?
How
do
they?
How
do
they
share
those
models?
And
you
know
x,
I
think
exxon
told
the
story
really
well
where
it
used
to
be
it's
it's
a
data
scientist
would
work
on
their
laptop
and
then
pick
up
their
laptop
and
bring
it
right.
C
Yeah
then
say
like
look
at
what
I'm
doing
and
what
are
you
doing
and
so
forth.
Now
the
ability
to
you
know
share
those
models
across
the
development.
You
know
host
them
on
the
platform.
Collaborate
and
you
know,
put
them
through
a
pipeline
for
ai
and
stuff.
It's
really
really
cool,
so
that
solves
a
great
problem
on
the
front
end,
for
you
know
for
model
development
and
collaboration
with
data
scientists,
then,
on
the
back
end,
it's
like
well,
you
know,
ai
is
no
longer
the
domain
of
just
this
small
sliver
of
your
organization
right.
C
A
C
Yeah
yeah,
no,
I
mean
look,
I
think
event
competing
vendors
will
say
what
they
say.
You
know.
First,
I
guess
I
start
by
saying
you
know:
red
hat's,
commitment
to
kubernetes
and
red
has
come
into
open
source
in
general
is
something
that
it
would
be
hard
for.
Anybody
to
question
right
for
20
years
we've
been
contributing
to
just
a
large
number
of
open
source
communities.
C
C
I
don't
think
anybody
can
question
what
we've
done
there
and
you
can
ask
not
just
us,
but
you
know,
other
vendors
we've
worked
with,
like
google
and
others
around
our
commitment
and
our
contributions
to
cube.
That
being
said,
we've
also
realized
that
we
need
to
build
more
than
you
know.
Customers
need
more
than
just
kubernetes
right.
Our
goal
has
always
been
to
drive
that
stuff
into
communities.
So
so
openshift
is
a
100
compliant
kubernetes
distribution
and
hosted
service
as
as
sort
of
deemed
by
cncf
right.
C
So
so
we
are
fully
compliant
everything
that
you
do
in
openshift.
Sorry,
everything
that
you
do
in
kubernetes
you
can
do
an
openshift
right.
So
if
you
want
to
work
with
openshift
directly
via
coop
ctl
at
the
command
line,
if
you
want
to
use
standard
ingress-
and
you
know
you
can
do
all
those
things
right-
we
we
fully
support
those
things
and
then
there's
things
that
just
aren't
part
of
kubernetes.
But
customers
still
need
right.
You
know,
like
security.
A
B
C
Security,
not
cubecon,
so
let's
start
there
right,
because
so
one
of
the
one
of
the
important
things
around
security
that
people
now
take
for
granted
in
cuba
rises
is
authorization,
authentication,
authorization,
permissions
and
so
forth.
People
may
not
realize
this,
but
that
didn't
exist
in
coup
c
in
kubernetes
1.0.
I
think
the
full
rbac
implementation
didn't
come
in
until
one
four
or
one
five
yeah
yeah
yeah.
We
had
it
in
openshift
three
on
cube
one
zero.
So
why
did
we
have
that
when
it
wasn't
in
cube?
Well,
the
kubernetes
community
wasn't
ready
for
it.
C
It's
like
that
was
never
going
to
happen.
So
how
do
you
get
enterprise
customers
to
adopt
openshift
three
to
adopt
kubernetes?
With
these?
You
know
very
real
constraints.
Well,
you
have
to
build
an
our
back
model
and
then
give
people
permission
so
that
you
can
distinguish
who
your
administrators
are
from
your
end,
users
right
and
red
hat
did
that
I
think
we
contributed
more
to
our
back
in
qrays
than
any
other
vendor.
It
initially
started
in
in
okd
our
upstream
price.
C
It
was,
it
was
an
extension
to
kubernetes,
but
then
over
time
we
pushed
that
into
sort
of
the
kubernetes
core.
I
think,
by
the
time
you
know,
q1516
came
around.
That
was
just
a
standard
capability
that
everybody
can
now
take
advantage
of
that
you
didn't.
You
know
it
didn't
matter
if
you
were
using
openshift
or
gke
or
anything
else,
but
sometimes
you
know,
like
those
things
take
time,
communities
are
hard,
you
have
to
prove
stuff
out.
A
C
People
need
to
get
traffic
into
the
cluster,
so
it's
like
how
useful
is
this
cluster
when
you
can't
expose
endpoints
to
take
exactly
so
we
built
we
built
a
model,
an
early
model
for
congress
called
openshift
routes.
Ultimately
that
led
to
you
know
ingress,
which
we
also
worked
on
becoming
a
standard
part
of
cube.
We
support
ingress.
We
maintain
support
for
routes
because
there's
still
some
capabilities
there
that
haven't
made
it
into
ingress
and
customers
want
right.
C
So
so
we
end
up
supporting
both
of
those,
and
I
could
talk
about
deployments
and
and
other
things
right.
So
kubernetes
has
evolved
a
lot
over
the
last
years
and
and
sort
of
characterizing.
You
know
the
work
we've
done
as
a
fork,
I
think,
is.
I
think
it's
just.
It's
really
vendors
trying
to
sort
of
spread
fud,
but
you
know
I
think,
customers
know
better
right.
They
know
you
know
they've
seen
what
we've
done.
C
A
C
You're
using
gke
aks,
eks
tanzu,
you
know
formerly
heptio
right,
there's
projects
there.
You
know
there's
projects
like
contour
and
valero
and
the
other
things
that
they're
just
hard
at
corku,
because
we
want
to
keep
that
core
cube.
Really
tight,
really
small
right,
so
going
forward
is
going
to
be
defined
as
by
the
number
of
things
that
we
keep
out
of
that
core
to
make
that
secure
and
stable
as
the
things
that
we
allow
in.
C
Now
those
are
taken
out
of
the
core
through
csi,
because
they
shouldn't
be
part
of
the
core
right
and-
and
so
so
so
then,
like
at
what
point
where
this
pure
cube
thing
falls
over,
because
the
youtube
itself
is
sort
of
a
core
that
has
all
these
different
extensions
and
and
again
we
we
do
our
best
to
sort
of
to
make
the
right
choices,
but
also
enable
choice
for
our
customers
when
they
want
to
integrate
other
solutions.
So
sorry,
that's
a
long
answer,
but
it's
no.
A
C
Yeah,
so
so
it
basically
comes
down
to
you
know
again,
which
vendors
are
you
going
to
trust
right?
So
openshift
has
always
been.
You
know,
first
and
foremost,
again
going
let's,
let's
go
back
to
the
beginning
of
of
our
kubernetes
journey.
Openshift
3.
we're
trying
to
help
customers
have
choice
in
the
infrastructure
that
they
could
run
in
right.
So
so
you
know
red
hat
and
google
were
the
first
two
vendors
to
sort
of.
I
think,
launch
commercial
kubernetes
solutions.
C
Google
obviously
focused
on
puberty's
as
a
public
cloud
service
through
gke
red
hat
focused
on
our
hybrid
cloud
mod
right.
So
we
want
to
make
sure
customers
didn't
get
locked
into
one
cloud
that
they
could
use
kubernetes
on
google,
on
amazon
on
azure,
but
also
in
the
data
center
on
vsphere,
on
openstack,
on
bare
metal
on
red
hat,
virtualization,
hyper-v
and
so
forth.
C
Right,
so
so
that
I
think,
is
an
element
of
lock-in
that
customers
care
a
lot
about
because,
like
even
though
customers
may
not
be
thinking
about
a
hybrid
cloud
strategy,
hybrid
is
ultimately
where
they
end
up,
because
all
their
applications
don't
run
in
one
place.
They
don't
all
run
in
the
data
center.
They
don't
all
run
in
just
one
public
cloud
as
they
move
stuff
to
public
cloud.
They
need
to,
you
know,
expand
to
other
clouds,
and
then
you
know
increasingly
out
to
the
edge.
C
C
Is
that
portability
and,
more
importantly,
that
consistency
across
those
those
different
environments
and
ultimately
the
rest
is
up
to
customers
to
decide
which
vendors
do
they
trust
right
to
support
you
in
all
those
environments
and
also
to
support
you
when
you
move
those
apps
out
to
your
production
mission,
critical
environments
and
then
you've
done
your
job
by
the
way?
That's
red,
hat's
history
right?
Where
did
red
hat
begin?
C
Helping
you
avoid
in
your
in
your
infrastructure
right,
like
you,
know
like
having
to
buy
hard
servers
and
chips
from
your
unix
vendor
versus
you
know,
choosing
linux
and
then
being
able
to
choose
any
x80
commodity
hardware
yeah
that
saves
customers,
tons
of
money
right
in
hardware
costs
and
software
costs,
and
that's
that's
kind
of
the
history
of
red
hat-
is
enabling
choice
and
portability
and
avoiding
lock-in.
So
I
I
kind
of
laugh
when
people
try
to
turn
that
around,
but.
B
C
A
Possible
right
and
we
we
made
linux
the
kind
of
the
webs
de
facto
os
for
lack
of
a
better
term.
Not
we
but
like
the
world
did
right
like,
but
now,
when
you're
looking
at
kubernetes
and
it's
becoming
this
de
facto
compute
plane
and
we
know
kind
of
like
there's
all
these
different
directions.
We
can
go
with
serverless
open
service
mesh
and
you
know,
k
native,
like
of
all
those
technologies.
What
excites
you
the
most
right,
like
all
these
different
directions,
you
can
go
right
like
what
excites
joe
fernandez
the
most.
C
So
in
terms
of
what
excites
me,
the
most,
I'm
really
excited
about
sort
of
you
know
this
explosion
in
terms
of
you
know,
data
and
analytics
and
so
forth,
and
you
know
the
information
that
can
be
gained
from
that
right
and
and
so
how
we're
enabling
customers
to
do
that
right.
So
so
so
some
of
these
new
services
around
ai
analytics
data
services,
which
are
helping
customers,
you
know
better,
serve
their
customers.
You
know
make
scientific
breakthroughs,
you
know
better
sort
of
take
advantage
of
you
know.
C
What's
going
on
in
the
market
and
so
forth.
That's
exciting
for
me
and
pushing
that
out
to
the
edge
you
know
to
enable
that
where
the
data
is
being
generated
is
also
critically
important
when
you
think
about
autonomous
vehicles
when
you
think
about
5g
and
so
forth.
So
I'm
super
excited
about
that,
because
I
think
kubernetes
and
containers
are
a
great
platform
to
enable
those
workloads
and
those
and
those
footprints.
The
other
thing
I'm
excited
about
is
is
again.
C
Like
yeah,
so
again,
it's
not
that's
not
to
say
like
if,
if
you
like
kubernetes,
if
you
want
to
work
right
at
that
core,
you
want
to
work
at
the
coupe
ctl
layer
or
at
the
kubernetes
you're
welcome
to
do
that.
But
for
a
lot
of
developers,
that's
not
their
focus
right.
So
their
focus
is
the
code
that
they're
building
the
languages
and
frameworks
that
they
use
the
services
that
they
need
to
support
their
applications
and
then
having
more
of
those
capabilities,
be
part
of
the
platform
right.
C
So
when
we
introduced
service
management,
we
introduced
serverless
pipelines.
Those
aren't
extra
things
you
paid
for
right,
those
those
become
part
of
the
platform,
and
so
the
value
that
we're
providing
is
growing
rapidly
above
and
around
that
kubernetes
core,
and
that
has
me
excited
because
you
know
like
in
git.
Ops
is
just
the
most
recent
addition
with
the
work
we're
doing
around
argo,
but
there's
more
of
that
to
come
right
so
so
really
excited
to
see
that
and
again
it's
not
just
all
work
that
we're
doing
it's
work.
C
That
is
happening
within
other
groups
at
red
hat.
It's
work,
that's
happening
at
ibm,
so
we've
been
working
with
the
ibm
middleware
teams
and
data
teams
and
the
ibm
research
group
around
services
that
they're
building
on
top
of
openshift
and
through
days,
and
then
it's
it's.
It's
all
those
partners
that
I
mentioned.
You
know
whether
you
know
different
isvs
serving
different
industry
needs
and
security
and
data
and
analytics
and
so
forth.
So
that
has
me
excited,
because
you
can
really
see
this
as
sort
of
the
foundation
for
just
a
ton
of
innovation.
A
Yeah
yeah
yeah,
I
mean
so
there's,
there's
there's
some
questions
in
chat
and
then
like
you're
the
amount
of
data
that
people
have
these
days
from
all
these
sensors
from
everything
being
connected
to
the
internet.
From
you
know,
like
we've
hit
this
we've
hit
this
critical
mass
right.
I
think,
where,
like
2020
is
gonna
be
a
year
where
we
see
okay,
there's
been
some
ipv
enough,
ipv6
adoption
and
enough
sensors
being
created
and
enough
data
being
generated
that
we
have
a
data
problem
like
companies
now
have
a.
A
C
Yeah,
so
so
some
of
it
is
enabling
the
workloads
I
already
mentioned,
so
it
won't.
I
won't
get
into
that
anywhere,
but
you
know
things
that
act
and
process
that
data.
You
know
those
the
different
ai
frameworks
and
tools
how
to
get
data
into
the
cluster
right,
like
so
supporting
things
like
kafka
on
openshift
right,
and
you
know,
support
to
support
that,
both
as
something
that
you
deploy
and
manage
yourself
or
as
a
hosted
service.
C
Red
hat's
gonna
be
doing
a
lot
more
in
terms
of
hosted
application
services,
supporting
it
at
the
network
layer
right.
So
there's
this
growing
momentum
around
kubernetes
in
the
telecommunications
industry,
with
service
providers
right
at
our
recent
red
hat
summit
we
had
verizon
was
one
of
our
keynote
speakers.
C
They
interviewed
talked
about
the
work
we're
doing
with
them,
but
they're
enabling
more
data
through
their
work
on
5g
networks
and
and
then
you
know,
through
those
5g
deployments,
adopting
cloud
native
platforms
and
containerized
network
functions
right,
but
that
stuff
doesn't
just
happen
right
right,
there's
specific
requirements
when
we
work
with
those
service
providers
like
ipv6
right
like
and
guess
what,
like
ipv6,
didn't
just
work
in
purebase.
In
fact,
you
know
there's
been
work
going
on
in
ipv6
and
kubernetes
for
a
while.
C
C
Observability
to
be
able
to
monitor
this
thing,
so
there's
quite
a
lot
of
work
that
we're
doing
around
just
the
network
around
monitoring,
to
enable
like
service
providers
to
sort
of
even
just
manage
that
data.
So
so
I'd
say
those
things
you
know
like
getting
into
those
new
verticals,
with
new
support
for
the
workloads
that
they
care
about
and
then
being
able
to
run
in
the
different
footprints.
C
Some
of
the
things
that
are
going
to
enable
this,
this
data
to
be
sort
of
managed-
and
you
know,
made
use
of
and
we're
doing
that
ourselves
so
like
we
we're
using
data.
So
one
of
the
things
the
telemetry
that
we
include
with
openshift
for
customers
to
use
customers
can
also
then
have
what's
called
a
connected
cluster
right.
So
they
can
enable
the
subset,
like
telemetry,
to
be
shared
with
red
hat,
which
then
gives
us
data
that
helps
us
better,
serve
customers
right.
C
So
so
we
can
see
what
version
you're
on
and
recommend
you
know
and
upgrade
what
patch
level
on
that
version
to
tell
you
know
to
warn
you
about
critical
cves
right
that
I
think
yeah
yeah
and
get
data
on
where
those
upgrades
are
succeeding
or
failing
and
then
and
then
getting
into
the
workloads
right,
like
you
know,
upgrading
you
know,
services
that
you
may
run
on
the
platform,
so
this
becomes
more
turnkey
right
like
so
you
don't
have
to
think
about
it
as
much
right.
I
always
wait
for
my
phone.
C
I
don't
think
a
lot
about
creating
my
phone.
I
just
you
know
the
upgrades
just
happen
and
right
working
yeah.
I
don't
have
to
think
about
when
I'm
gonna
upgrade
the
hundreds
of
apps
that
are
on
my
phone,
because
those
those
isvs
that
build
those
apps
partner
with
apple
or
with
youtube-
and
you
know
if
we
can
make
it
work
in
the
consumer
space.
Why
can't
we
make
that
work
in
the
enterprise
space
and
that's
what
we're
trying
to
do
with
telemetry
and
work
you're
doing
around
operators
and
integrating
with
those
with
those
partners?
A
Just
with
your
phone
in
your
hand,
right
and
now
we
want
to
bring
that
yeah
yeah.
Now
we
want
to
bring
that
like
simplicity
of
upgrades
and
updates,
so
that
you're
not
in
this
place
where
it's
like,
can
I
upgrade
the
os
right
now?
Do
I
have
to
have
this
big
huge
planning
session?
No,
we
want
you
to
be
in
this
place.
Where
things
become,
you
know
more
ephemeral
or
not
ephemeral.
I
guess,
but
like
more
stable
and
more,
you
know
resilient.
A
C
So
that's
that's
a
measure
of
confidence
that
you
need
to
earn.
You
know
right
right,
like
yeah,
I
I
didn't
always
have
automated
updates
turned
on
for
my
iphone
or
I
said
something
that
over
time
I
became
more
comfortable
with
through
each
successive
like
upgrade,
and
you
eventually
just
get
sick
of
saying,
accept.
A
C
Seems,
like
you
know
what
you're
doing
right
we're
trying
to
get
to
that
one
because,
like
you,
said
the
iphone,
what
it
did
for
enterprise
I.t.
Is
it
changed
our
expectations
for
what
the
platform
should
take
care
of
right?
There's
a
linux
operating
system
in
your
phone
in
my
phone
and
all
this
right.
But
you
don't
see
it
right.
It's
never!
It's
embedded
in
the
platform.
C
Well,
there's
a
parallel
there
with
what
we
tried
to
do
with
openshift
4,
with
with
ralph
core
os
right
with
openshift
3
and
prior
generations
of
openshift
the
operating
system.
The
linux
operating
system
was
a
separate
platform
that
you
had
to
install
on
the
environment
of
choice
before
you
could
even
install
kubernetes
right
now.
Red
hat
was
unique
in
that
we
could
support
both
the
operating
system
as
well
as
kubernetes
and
everything
that
ran
on
top,
but
we
still
forced
you
to
manage
those
separately
with
openshift4.
We
bring
that
together.
C
So
the
operating
system
with
raw
core
os
is
just
a
feature
of
the
true
branding
platform.
It's
actually
bootstrapped
and
managed
by
the
platform,
and
it
does
it
in
an
intelligent
way.
On
the
platform
of
your
choice,
right
because
when
you're
bootstrapping
kubernetes
on
amazon,
you
have
to
make
different
choices
than
when
you
have
bring
it
on
azure
and
google,
so
so
those
types
of
things
from
a
platform
perspective
getting
to
the
level
where
it
could
be
as
seamless
as
an
ios
or
an
android
experience
is
critically
important.
C
But
then
the
other
thing
that
people
forget
is
that
there's
just
a
ton
of
stuff
on
your
phone
right.
Some
of
it
comes
from
from
apple
right
or
from
the
different
google
vendors.
So
it's
not
just
a
phone.
It's
a
you
know,
it's
your
it's
a
video
device.
It's
a
camera!
It's
a
music
player!
It's
it's
apps!
Right!
All
those
things
are
just
coordinate.
I
sort
of
equate
that
to
how
we're
adding
more
capabilities
just
into
kubernetes
and
openshift
itself
that
are
just
there.
C
B
C
Concept,
epitomizes,
why
we
invested
in
operators?
Why
we
care
about
immutable
infrastructure?
Why
we
launch
things
like
operator
hubs
and
marketplaces,
because
we
should
expect
the
same
thing
for
enterprise
software
and
in
cloud
services
that
that
we've
grown
to
expect
in
the
consumer
world.
It
just
has
to
happen.
A
I
just
want
it
to
work
right
like
if
I
have
a
master,
that's
down
blow
it
away
and
build
it
back
up
like
done
right
like
I
want
everything
to
just
happen,
and
I,
if
I
need
that
to
happen
on
a
small
footprint,
how
do
we
kind
of
make
that
work,
because
that
is
a
complaint?
Sometimes,
when
you
own
the
control
plane,
you
have
to
have
the
compute
behind
it.
Right,
like
you
know,
but
there's
a
lot
of
value
in
owning
that
control,
plane,
yeah
and
and
that's
why
we
deliver
it
to
you.
C
So
so
the
thought
of
having
hundreds
of
these
things
like
it's
just
it
was
untenable
right.
So
red
hat
drove
a
lot
of
work
around
like
multi-tenancy
capabilities
in
kubernetes
itself.
Right,
we
drove
a
lot
of
you,
know,
name
spaces
and
quotas
and
yeah
types
of
capabilities
exist
and
as
well
as
what
I
already
mentioned,
are
back
the
notion
of
the
user
and
the
role
and
so
forth.
C
We
we
put
a
lot
of
effort
working
with
google
and
other
contributors
upstream
to
make
those
things
happen,
so
that
cluster
clusters
could
be
shared
and
again
today
we
take
all
that
for
granted
the
ability
for
different
users
to
run
different
applications
on
the
same
cluster
and
have
them
you
know
fully
tenanted,
even
at
the
network
level,
with
things
like
network
policy.
So
you
know
you
don't
get
communication
or
traffic,
you
know
between
tenancies
or
what
that
all
happened
because
of
that
really
work.
C
Now
you
fast
forward
six
years
later,
and
you
know
everybody's
sort
of
much
more
comfortable
with
kubernetes
and
containers
as
a
model,
and
now
you
just
start
to
see
more
and
more
clusters
right
so
and
and
they
do
become
more
and
more
sort
of
disposable,
disposable
or
ephemeral,
as
you
will,
and
so
we've
had
to
build
new
capabilities.
So
the
reason
we
launched
acm,
which
is
our
advanced
cluster
manager,
is
because
we
see
our
customers
having
tens
or
you
know.
C
Even
you
know,
you
know
20
50,
100
plus
clusters
and
then,
as
they
go
to
the
edge
they'll
have
hundreds
and
thousands,
and
so
we
needed
a
way
to
manage
multiple
clusters
from
one
location
to
upgrade
them
to.
You
know
to
set
policies
for
how
they
should
operate
and
so
forth,
and
so
that's
what
acm
is
all
about.
C
I
think
we
need
to
kind
of
continue
going
further
right
in
some
of
the
areas
you
mentioned
around.
Like
you
know,
you
know
the
control
plane
versus
versus
the
you
know
the
compute
nodes,
the
worker
nodes
and
so
forth,
who
needs
to
own
what
right
a
lot
of
the
cloud
providers,
including
ibm
leverage
so
like
a
shared
control,
plane
type
model
where
you
know
you
just
sort
of
bring
work,
your
nose
in
the
form
of
the
vms
and
the
control
planes
are
sort
of
running
there
and
named
by
somebody
else.
C
So
we've
done
work
in
that
area
with
ibm
that
we're
looking
to
kind
of
expand
into
so
the
project
is
called
hypershift,
but
there's
sort
of
a
lot
there
around
around.
How
do
you
unpack
those
and
then,
like
I
mentioned
when
you're
at
those
edges
locations,
it's
just
got
to
be
plug
and
go
right
like
right.
The
cluster
is
sort
of
some
appliance
that
gets
unloaded
by
off
the
truck
plugged
into
the
wall,
and
it
just
has
to
boot
up
and
start
working
right.
There's.
A
C
Of
things
to
look
into
to
watch
out
for
so
the
work
we're
doing
around
smaller
clusters,
we
already
certified
a
three
node
cluster,
we're
working
on
distributed,
compute
nodes
so
having
a
control
plane,
that's
shared
by
individual
compute
nodes
running
in
different
locations
and
then
ultimately,
two
node
and
one
node
clusters
is
sort
of
the
next
thing
that
we're
working
on.
Now
I
mean
as
a
smile
when
you
mentioned
hard,
because
we've
done
a
lot
of
work
to
support
other
architectures
right,
so
openshift
today
isn't
just
supported
on
x86.
C
It's
supported
on
support
on
power.
It's
supported
on
z
and
soon
we'll
have
sort
of
at
the
worker
node
the
ability
to
do
windows,
machines,
yeah,
yeah,
x86
architecture-
and
you
know,
maybe
in
a
year
from
now
we'll
be
talking
a
little
bit
more
about
what
we're
doing
around
arm
and
stuff,
because
that's
that's
another
architecture,
whether
it's
for
edge
deployments
or
even
database
center
deployments
that
we're
getting
asked
a
lot
about-
and
I
think
yeah
I
think
openshift
on
on
q
a's
unharmed
is-
is
a
big
part
of.
A
C
Yeah,
I
think
it'd
be
like
everything
else.
Right,
it'll
solve
some
very
specific
problems
that
are
very
important
for
certain
customers
or
certainly
it'll,
introduce
new
problems
that
we
never
knew,
existed
and
and
ultimately
it'll
just
become
part
of
that
landscape
and-
and
you
know,
red
hat's
right
there,
like
we've
done
as
much
work
around
linux
on
arm,
as
you
know,
as
any
vendor
in
the
space
right
with
or
red
hat
or
president
nixon
has
done,
and
you
know
kubernetes
and
stuff
that
runs
on
kubernetes.
C
A
C
C
Talk
about
bare
metal
but
like
bare
metal
both
in
the
data
center
and-
and
I
mean
soon
in
the
cloud-
is
one
of
the
environments
that
customers
are
most
excited
about,
like
it's
one
of
the
fastest
growing
environments
that
customers
want
openshift
on
and
we
have
a
lot
of
great
customers
going
there.
So
so
that
came
from
the
guys
at
packet.
A
Right,
yeah,
yeah,
that's
awesome,
that's
great!
That's
so
fun!
I
love
working
with
them
and
yet,
like
I
mentioned,
they've,
been
on
the
show
a
few
times
or
the
channel
a
few
times,
so
this
might
be
too
far
out,
but
one
of
our
regular
watchers
listeners
or
whatever
you
want
to
call
them
attendees.
You
know
three
to
four
was
a
huge
change.
What
is
openshift
four
to
five?
Have
we
even
like
started
looking
at
that,
like
what.
C
Like
do
we
even
know
that
yeah
yeah
there's
nothing
imminent
on
that
front
right,
like
these
big
migrations,
and
I
know
for
existing
customers,
they're
they're
challenging,
but
like
we
hope,
customers
know
that
that
we
make
them
for
good
reasons
right.
So
so
the
two
big
transition
points
that
the
two
to
three
transition
point
was
really
around
the
moving
to
this
standardized
oci
container
model
with
kubernetes
right
and
we
had
used
containers
in
the
form
of
c
groups
and
name
spaces
and
so
forth,
going
all
the
way
back
to
openshift
1.0.
C
So
it
wasn't
that
containers
were
new,
but
we
needed
to
support
you
know.
Well,
we
actually
helped
to
drive
docker
into
becoming
this
oci
standard,
but
it
didn't
come
for
free
right
in
that
moving
to
a
new
version
of
linux,
new
version
of
rail
and
and
then
essentially
rebuilding
the
entire
platform.
So
this
is,
I
always
joke
with
my
engineers.
C
It
started
out
as
like
yeah
we're
going
to
integrate
docker
into
openshift,
and
then
it
ended
up
with
we're
going
to
throw
it
all
out
and
rebuild
openshift
from
scratch
and
well
leveraging
shootings
and
so
forth.
So
so
there
there
was
only
that
was
a
destructive
change
for
existing
customers,
but
I
think
it
was
a
necessary
change
and
obviously
led
us
to
where
we
are
today
with
openshift.
Three
to
four.
The
big
change
was
around
redefining
the
boundary
of
the
platform
right
openshift,
free
and
prior.
C
For
you,
you
know
what
are
you
going
to
choose
right
and
obviously
we
chose
the
latter
so
with
that
with
the
additional
railcore
os
rebuilding
around
operators
that
was
sort
of
a
disruption
that
could
only
be
done.
You
can't
just
in
place
upgrade
your
way
to
that
right.
Now,
there's
no
big
platform-centric
things
in
the
offing.
That
would
you
know
precipitate
in
openshift,
five
and
same
thing
in
the
kubernetes
community.
Largely
it's
just
iterative
point
releases,
so
I
don't
see
that
happening
for
the
next
few
years,
at
least
but
yeah
but
yeah.
C
C
A
I
mean
just
to
give
you
an
idea
right
like
daniel.
Oh,
like
one
of
our
you
know
fellow
red
hatters
here
he
and
I
were
both
at
devops
days
raleigh
last
year
and
he
was
doing
his
talk
on
quarkus
and
he's
done
the
same
talk
on
this
channel
before
and
you
know
in
the
audience
the
java
developers
sitting
in
raleigh
were
like
wow,
someone
behind
me
said
whoa,
as
he
like,
compiled.