►
From YouTube: Upstream this! Panel Brendan Burns, Clayton Coleman, Kris Nova, Aparna Sinha, Paul Morie
Description
OpenShift Commons Gathering December 5th 2017 Austin, Texas
Upstream this! Panel: Red Hat, Google, Microsoft, and Heptio
Panelists: Paul Morie (Red Hat), Brendan Burns (Microsoft), Kris Nova (Heptio), Clayton Coleman (Red Hat), Aparna Sinha (Google)
Stormy Peters, Red Hat - Moderator
A
So,
for
me
this
is
probably
one
of
the
most
fun
parts
of
the
day
and
we
will
ask
you
all
to
bring
your
questions
and
we'll
have
some
Q&A
at
the
end,
but
I'm
really
happy
to
introduce
stormy
Peters.
Who
is
the
head
of
our
open
standards
and
software?
A
group
at
red,
Red,
Hat,
leading
a
lot
of
the
community
events
and
projects
and
we're
going
to
bring
in
our
panelists
and
I'll.
Let
stormy
introduce
them
as
they
come
on.
I'm.
B
Gonna,
let
them
walk
up
and
get
seated
and
then
we're
gonna
get
started
and
they'll
introduce
themselves
while
they're
getting
seated.
This
is
the
upstream
panel
upstream
this.
How
many
people
here
know
what
upstream
means
awesome?
Let's
hear
some
answers,
chat
him
out,
shout
out
the
answers.
What
does
upstream
mean
where
all
the
fun
happens,
any
other
answers
for
upstream
community,
one
more
collaboration,
so
the
upstream
is
really
open
source
software
collaboration,
community
and
fun
happens,
and
so
this
is
a
panel.
B
We're
not
going
to
be
able
to
give
you
all
the
information
about
upstream
in
half
an
hour.
This
is
a
panel
to
introduce
you
to
some
of
our
upstream
contributors,
they're.
Also
others
of
you
sitting
in
the
audience,
and
this
is
to
start
conversations
that
you
can
have
over
lunch.
You
can
have
later
today.
You
can
have
it
the
get-together
you
can
have
during
this
week
at
cube
cotton.
So
this
is
a
conversation
starter.
So
with
that
I
was
gonna.
B
Ask
our
panelists
to
introduce
themselves
say
their
name,
what
project
they
work
on
and
then
to
you
know
how
like,
when
you're,
trying
to
describe
what
you
work
on
to
somebody
who
has
no
clue.
Maybe
it's
your
neighbor
who's,
a
surgeon.
Maybe
it's
your
kids
teacher
and
you
try
to
use
analogies
and
you
try
to
use
metaphors
and
you
end
up
in
this
big
complicated
mess.
B
B
D
Hi
everyone
I
am
Chris
Nova
I
work.
Thank
you,
everybody
who
waved
back.
That
means
a
lot
to
me.
I
work
at
hefty
Oh,
which
is
freaking
great
and
then
I
guess.
My
analogy
like
this
is
gonna,
be
really
like.
Everybody
here,
probably
already
knows
what
I'm
gonna
pick
but
I'm
gonna
compare
it
to
mountains,
because
that's
I
do
a
lot
of
mountaineering
and
I.
Think
like
there's
just
a
lot
to
be
said
about
like
going
one
step
at
a
time
to
take
over
like
a
really
monolithic
task.
D
C
Was
really
good
analogy,
so
I'm
Clayton,
Coleman,
architected,
open
open
shipped
in
kubernetes
at
Red,
Hat
I
also
work
on
the
communities
community
a
lot.
So
my
analogy
is
I,
usually
start
with
you,
don't
know
what
I'd
do,
but
when
it
stops
working,
you
will
all
regret
it
and
that's
a
nice
icebreaker
and
then
I
usually
spend
a
bunch
of
time,
which
is
like
all
of
the
things
that
we
all
build
are
incredibly
complicated
and
nobody
understands
it
anymore,
and
so,
by
analogy
we're
sitting
on
a
giant
mountain
of
stuff.
E
A
A
E
You
look
at
the
Sears
catalog,
which
is
approximately
two
feet
high
and
you
can
you
can
order
the
thing
out
of
the
catalog
and
it
comes
to
you.
That's
actually
a
terrible
analogy
also
because
the
houses
that
you
ordered
from
Sears
catalog,
which
is
actually
a
thing,
did
not
come
assemble
and
part
of
service
catalog
in
the
concept
of
a
broker's
that
you
don't
you
don't
care
at
all,
so
very
bad
analogy.
F
F
H
H
I
I'm,
a
pronoun,
Senna
and
I'm
a
product
manager
on
the
kubernetes
project.
My
analogy
is
that
of
a
I
guess
of
rocket
ship
I
feel
like
we
not
the
interesting
aspect
of
flying
a
rocket
ship,
but
of
putting
it
together,
so
I
think
of
the
work
that
I
do
as
being
in
a
environment
where
there
are
lots
of
moving
pieces
and
we're
trying
to
build
this
thing.
That's
gonna,
that's
gonna,
take
off
and.
C
H
F
H
D
H
D
B
B
C
B
Or
is
there
another
question
start
us
out
all
right,
so
so
this
is,
this.
Technology
is
a
pretty
kubernetes
and
their
surrounding
technologies
are
our
all
open-source
software.
They
have
a
lot
of
the
aspects
of
traditional
open-source
software
and
it's
also
a
pretty
unique
place
that
really
fosters
innovation
on
top
of
it
and
what
anybody
would
what?
What
do
you
think
is
unique
about
it
that
fosters
that
innovation.
D
So,
for
me,
the
kubernetes
community
and
the
way
that
we
sort
of
all
come
together
to
engineer
the
process
of
engineering
is
super
unique
coming
from
other
communities
and
like
the
open
source
in
Linux.
Ii
Apache
space
to
kubernetes
was
a
really
big
jump
for
me
and
I.
Think
as
a
software
engineer,
like
I,
really
really
thrive
in
this
environment,
where
I
get
to
speak,
my
mind
and
collaborate
with
other
people
and
look
at
our
software.
D
C
Was
gonna
say
one
of
the
things
that
I
really
enjoy
is
using
I.
Think
it's
somewhat
unique.
It's
using
what
you
build,
so
a
lot
of
open
source
projects-
you
don't
necessarily
use
we
build
like
Linux,
is
a
great
example
like
most
of
the
people
who
hack
on
Linux
actually
use
Linux
and
in
kubernetes
like
a
lot
of
the
things
that
are
components.
C
F
F
Really
came
out
of
a
desire
to
help
people
build
software,
better
I
think
that's
relatively
unique
in
the
open-source
space
and
I
think
that's
part
of
why
it's
developed
such
a
great
ecosystem,
I
think
also,
we've
been
blessed
with
really
fantastic
people
throughout
the
the
community
and
it
really
helped
get
the
DNA
right.
Speaking
of
the
rocket
ship
at
some
level
like
little
tiny
bits
of
angle
in
the
early
trajectory,
make
really
big
differences
and
I
think
that
we
got
really
lucky
in
the
first
three
to
six
months.
I
We've
got
some
rocket
scientists
I.
Think
what's
unique
is
that
you
can
join
it's
kind
of
rare
to
be
able
to
join
in
building
something
like
this,
but
this
is
a
community
which
thrives
on
contributors
and
a
lot
of
users
join
so
to
Clayton's
point.
It's
also,
that's
also
actually
quite
unique,
so
I
would.
F
I
I
B
D
So
I
spend
a
lot
of
time
working
like
on
the
infrastructure
side
of
things
in
kubernetes,
so
I'm
a
member
of
Sigma's
life
cycle,
and
we
have
a
working
group
for
the
cluster
API
that
we're
working
on
now,
robert
bailey
and
I
are
giving
a
talk
on
it
on
the
thursday
if
you
guys
want
to
come
check
it
out
but
like
getting
involved,
it's
it's
really
like
joining
the
sig
calls
like
comment
on
things
on
github.
This
is
a
huge
one.
D
Like
the
first
time,
I
started,
commenting
on
kubernetes,
stuff,
I
was
kind
of
like
am
I
allowed
to
do
this.
Does
it
you
know?
Is
this
okay
and
like
now,
I'm
a
maintainer
of
a
couple
of
repos
and
they
kubernetes
org?
So
what
kind
of
comment
did
you
leave?
I,
just
like
my
two
cents
right,
like
I
agree
with
this
or
I.
Don't
agree
with
this
or
like
I.
D
Don't
think
that
we
needed
to
put
an
interface
here,
or
you
know
just
kind
of
like
whatever,
and
even
if
it's
not
like
a
super
technical
comment,
I
still
think
user
experience
is
something
that
we
really
need
to
focus
on.
In
kubernetes,
so
even
if
you're
like
sitting
at
a
command
line
and
just
something
feels
weird
or
backwards
like
bringing
that
up
and
making
that
known
as
super
important,
there's.
I
Ways
to
get
involved
in
non
technical
functions,
one
thing
to
do
is
I
think
just
use
the
product
and
kind
of
understand
what
it
is
and
then
all
of
the
SIG's
and
all
of
the
meetings
are
open
and
everybody
is
invited.
So
you
can
join
the
community
meeting,
which
is
on
Thursdays
at
10
a.m.
Pacific
or
you
can
join
any
of
the
sig
meetings
which
are
all
published
on
github
and
just
kind
of
start
to
listen
in
and
then
see.
If
there's
a
role
that
you'd
like
to
play.
E
But
I
would
urge
anybody
in
this
room
that
wants
to
get
involved
start
attending
community
meetings
start
attending
sig
meetings
and
the
the
the
reality
of
the
non
textured
non-technical
contribution.
Type
track
is
that
those
things
tend
to
be
things
that
engineers
are
like
terrible
at
so
they're.
Incredibly
valuable,
even
just
taking
notes,
is
a
really
valuable
skill
for
my
own
little
slice
of
the
world.
E
One
thing
that
we
recently
discussed
in
a
open,
serviceworker,
API
face-to-face
in
Raleigh
last
week,
actually
is
adding
new
generic
action
support
to
that
API
so
that
you
can
prototype
things
without
having
to
go
through
the
process
of
changing
the
official
API,
which
we
hope
will
be
a
way
for
people
to
incubate
and
experiment
things
and
maybe
have
some
emergent
conventions
in
the
actual
spec
itself.
Yelling.
F
F
There's
just
so
many
interesting
cool
tools
that
people
have
built
out
in
the
ecosystem
effectively
by
themselves
or
with
a
small
group
of
people
like
you,
don't
even
have
to
be
involved
in
the
main
community
to
make
a
contribution,
you
can
build
a
tool
that
helps
manage
configs
or
manage
certificates,
or
something
that
is
useful,
generically
useful
on
top
of
these
api's
that
are
now
available
to
people
across
the
world.
That's
also
a
really
great
way
to
kind
of
get
a
people.
I
C
I
was
gonna,
Brennen
totally
stole
mine,
because
I
think
the
idea
here
is
that,
like
Cuba's
gotten
so
big
in
the
last
three
years,
like
there's
a
lot
of
stuff
that
it
does,
it
just
can't
keep
growing
forever.
So
we
need
to
get
really
really
good,
making
sure
and
Brennan's.
Let
a
lot
of
this
stuff
on
the
client
side
going
and
using
the
client
generators
to
make
clients
for
new
languages,
and
you
know
I
think
we
need
to
do
more
and
like
this
is
something
that
I
try
to
spend.
D
I
I
Good,
that's
a
visionary
statement
and
I
think
that
that's
a
that's
a
good
vision
and
others
should
certainly
comment
where
I
wear.
What
I
took
that
to
mean?
Is
you
know
a
lot
of
analytics
and
machine
learning
can
generate
output
that
is
useful
to
to
the
application
so,
but
so
potentially
I
don't
know
if
the
code
would
write
itself,
but
there
could
be
connections
of
existing
modules
that
you
know
create
an
application.
I
B
F
Think
another
thing
that
I've
been
playing
around
with
a
lot
actually
and
I'll
talk
about
actually
in
my
keynote
tomorrow
is
this
notion
of
how
we
build
abstraction
layers
on
top
of
kubernetes
and
I
think
that
it
is
still
way
too
hard
to
build
distributed
systems
and
in
communities.
We
have
maybe
created
an
assembly
language,
but
we
have
not
created
higher
level
programming,
languages
and
I.
Think
that
is
an
interesting
next
step.
F
I
The
concepts
comes
up,
obviously,
because
there
are
so
many
things
in
kubernetes
that
you
know
over
time
will
will
manage
themselves
so
clusters
that
repair
themselves
and
scale
themselves,
and
so
that
leads
to
okay.
If
the
infrastructure
takes
care
of
itself,
can
we
do
something
like
that?
Also
at
the
application
level.
J
So
one
fundamental
dilemma
that
you
guys
upstreams
have
is
creating
something
that
collaborate
across
different
companies,
but
the
people
on
top
would
actually
create
the
product.
It's
create
something
different
because
they
want
to
flavor,
create
something
different
on
market,
so
each
company
has
different
flavor.
How
you
guys
can
guarantee
the
freedom
the
user
can
jump
from
one
product
to
another.
B
F
Another
thing
is
that
for
a
lot
of
us,
I
don't
have
to
differentiate
in
terms
of
product
I
differentiate
in
terms
of
service
right,
I
differentiate
in
terms
of
kubernetes
as
a
service,
as
opposed
to
in
terms
of
on
top,
so
I
think
it's
kind
of
like
you.
Don't
really
differentiate
on
the
x86
instruction
set,
I
think
there's
a
degree
to
which
it's
just
a
baseline,
that
cloud
providers
provide
for
you
and,
and
then
the
product
is
the
reliability.
The
price
point
all
of
that
sort
of
thing,
yeah.
H
C
Would
actually
add
to
like
to
that?
The
idea
of
API
is
like
the
kubernetes
api
will
not
be
last
API
ever
created
the
history
of
mankind.
What
will
actually
happen
is
just
like
every
other
API
like
an
assembly
is
a
good
metaphor,
because
there's
x86,
but
then
there's
micro
code
and
then
there's
the
transistors
and
then
there's
the
actual
modules
and
there's
the
bite
like
I
think
we're
actually
gonna
see
a
lot
of
that.
We're
seeing
it
today,
like
everybody,
has
build
systems
in
their
company
today.
Sometimes
they
want
to
change
them.
C
Sometimes
they
actually
want
to
hook
things
in
and
so
we're
starting
to
see
like.
We
should
be
really
easy
and
like
to
Burnham's
point
to
stack
layers
really
cleanly.
Not
everybody's
gonna
choose
the
same
layers.
If
you
can
switch
layers
out
bring
pieces
together.
That's
like
the
writing
code
is
what
are
the
pieces,
the
tools
that
help
people
build
distributed,
systems,
how
people
tie
together
build
change,
roll
things
out
to
production,
because
at
the
end
of
the
day,
that's
all
that
actually
matters
right
is.
How
do
you
deliver
software?
C
So
some
of
the
I
think
some
of
like
the
opinion,
a
ones
like
open
shift
is
somewhat
opinionated,
tries
to
tie
all
these
together
and
it
hits
certain
optimized
points,
but
not
everybody
is
going
to
build
software
that
way,
and
so
I
think
it's
layering
up
multiple
layers
of
very
well
designed
independent
pieces
that
can
be
composed
is
how
like
it
is
the
only
way
that
this
will
evolve.
Like
you
can't
just
design
the
perfect
system
and
everybody
to
use
it
I
think.
I
We
would
be
remiss
not
mention
the
conformance
program
which
I
talked
about
earlier
and
which
launched
I
think
just
before
Thanksgiving,
which
is
a
set
of
tests
that
any
kubernetes
distribution
can
run
and
then
publish
results
back
on
github.
That
shows
that
it
is
conformant
and
it's
it's
not
a
complete
set
of
tests
right
now,
but
over
the
course
of
this
year
we're
gonna
be
expanding,
that
set
of
tests
and
the
goal
is
to
ensure
that
the
kubernetes
api
is
fully
exposed
in
that
applications
that
are
running
on
kubernetes
can
be
portable
across
distributions.
I
B
G
Referring
to
analogies
and
all
that,
so
we've
gone
in
the
computer
building
the
normal
compute,
there's
analogies
to
that
what's
happening
again
today
in
the
distributed
systems
right
has
anyone
mapped
that
journey?
Sort
of
you
know,
like
you
mentioned
assembly
and
obviously
things
have
evolved.
We've
been
doing
computers
for
a
lot
further
than
the
distributed
system
so
and
the
networking
has
the
OSI
layer.
So
we
were
all
talking
about
these
things.
Is
anyone
kind
of
tried
to
create
an
infographic
about
something
that
helps
us
map
that
journey
a
little
bit
better
I.
F
Don't
know
that
any
infographics
I
think
about
it.
A
lot
I
think
there's
a
challenge
here,
because
you
know
there's
this
really
obvious
analogy
where
you
say:
oh,
the
distributed
system
is
just
a
computer,
and
then
you
realize
that
it's
actually
not,
and
it's
kind
of
like
a
really
bad
Numa
computer
and
computers
are
also
really.
F
No
and
that's
true
and
so
like
and
then
and
then
then
you
forgive
me
because
I
can
people
cared
and
then
people
stopped
caring
and
maybe
people
will
stop
caring
again
and
I.
Think
at
a
certain
level,
people
don't
care.
I
do
think
that
I
kind
of
feel
like
with
kubernetes
we've
kind
of
created
processes
and
I've
and
I've
often
said
that
I
kind
of
felt
like
the
kubernetes
api,
or
that
some
API
layer
would
be
sort
of
the
POSIX
of
distributed
systems.
I
think
we're
headed
in
that
direction.
F
F
C
The
big
companies
have
been
doing
this
for
years,
and
so
they've
got
lots
of
these
patterns
baked
in
and
if
you
look
at
kubernetes
like
a
lot
of
it
is
folks
like
Brendan
and
Tim
and
Brian
saying
like
we
have
all
these
patterns,
let's
boil
them
down
to
their
essences
and
put
them
out
there
and
every
day
someone
comes
to
kubernetes
and
does
the
same
thing
for
networking
or
for
storage,
or
they
take
the
new
ideas.
So
I
would
actually
say.
C
The
really
thing
that
we
should
really
try
to
do
in
kubernetes
is
the
meta
right.
It's
the
mind,
the
process
of
thinking
about
how
you
build
things
so
that
they
fit
well
together
and
they
solve
problems
really
quickly
in
the
distributed
system.
Space
like
I
I
won't
cast
any
aspersions,
but
like
just
as
a
random
example,
the
ec2
API
was
not
designed
as
something
that
was
going
to
magically
make
all
of
building
applications
easier.
C
It
was
designed
to
deliver
VMs
I
think
we're
trying
to
be
a
little
bit
more
deliberate
on
what
are
the
things
that
applications
need
and
design
api's
for
that
and
think
a
little
bit
about.
What's
above
and
what's
below
and
then
try
to
put
them
together,
I
think
the
api's
will
change,
but
the
process
is
what's
important
like
can
you
plug
stuff
in
and
get
value
out
of
it
yeah.
F
What
we
need
to
do
next
is
to
figure
out
how
do
you
build
the
standard
library
and
the
compiler
that
lives
above
kubernetes
in
a
modular
way
that
doesn't
turn
into
pass
and
that
I
think
is
the
real
challenge
that
I
want
to
spend
a
lot
of
times
and
I
spend
a
lot
of
time
thinking
about
and
that
I
want
to
spend
a
lot
of
time.
Thinking
about,
because
I
think
that's
where
we
actually
get
to
what
you're
talking
about.
F
B
B
B
All
right,
so
I
asked
a
lot
of
questions
about
how
things
work.
Now,
that's
what's
a
little
bit
technology
and
each
of
you
work
on
different
aspects.
If
you
could
tell
me
one
really
cool
thing
that
you're
working
on
right
now
that
you
think
will
foster
innovation
in
your
space
or
innovation
for
those
that
are
developing.
On
top
of
it,
give
us
a
you
know
a
one
minute
update
of
what's
happening
in
your
project.
Chris.
You
want
to
lead
us
off
sure.
D
So,
like
I
mentioned
earlier,
I'm
doing
a
lot
of
work
with
folks
over
at
Google
and
in
the
sink
cluster
life
cycle
group
for
the
cluster,
API
and
I
think
the
sort
of
10,000
foot
view
here
is
we
want
to
have
a
declarative
way
of
defining
infrastructure
and
the
really
exciting
part
of
that
is
having
a
set
of
what
we
call
machines,
which
represent
a
virtual
machine
in
the
cloud
and
having
a
controller
that
sort
of
takes
this
kubernetes
declarative
first
principle
and
applies
that
to
infrastructure
is
something
that's
really
exciting
to
me
and
I.
D
Think
a
lot
of
folks
are
gonna,
get
some
value
out
of
it
when
you
start
looking
at
things
like
business
logic,
and
maybe
we
want
to
auto
scale
based
on
arbitrary
rules
that
are
encapsulated
and
in
software,
so
I
just
think,
there's
a
lot
of
modularity
and
flexibility,
and
this
is
something
I've
been
kind
of
working
on
for
like
about
a
year
now
and
to
see
it
kind
of
come
to.
Life
is
really
really
exciting.
For
me
are.
C
C
Who
is
it
that's
talking
to
me,
because
if
you
want
to
build
an
integration,
you
want
to
say:
okay
well,
I
know
that
I
should
only
give
this
secret
out
this
particular
workload
running
on
this
particular
node,
and
so
we've
been
working
with
a
bunch
of
folks
in
the
community.
The
spiffy
effort
inspire,
which
is
a
group
of
people
working
on
kind
of
a
generic
spec
that
works
across
cloud
and
metal.
C
We've
also
been
looking
at
how
we
can
break
up
kubernetes
and
make
this
easier
and
trying
to
say
like
every
time
we
break
something
out
of
cubes,
so
you
can
say
like
I
want
to
plug
Kerberos
in
or
I
want
to
plug
PKI
and
doing
it
in
a
way
that
really
makes
it
easier
for
everybody
else
to
benefit,
even
if
this
is
very
specific
to
it,
making
it
easier
to
do
extension
at
the
node
level.
So
that's
been
really
exciting.
I've
enjoyed
it.
You.
C
E
So
the
sig
multi
cluster
in
the
kubernetes
community
has
been
bootstrapping
an
effort
on
a
cluster
registry
which
is
an
API
server
that
just
has
a
cluster
resource,
which
is
where
is
this
thing
and
how
do
I
talk
to
it?
And
it's
really
exciting
to
me,
because
I
think
this
will
accelerate
and
unify
to
some
degree
different
efforts
in
the
community
around
building
tooling,
to
handle
multiple
clusters
at
the
same
time,
which
seems
like
an
inevitable
challenge
to
get
over.
So
there
is
a
cig,
multi
cluster,
deep
dive.
E
B
F
So
I'm
really
excited
about
a
bunch
of
the
work
that
that
the
azure
team
has
been
doing
an
open
source,
developer
productivity.
So
from
things
like
the
kubernetes
and
Visual
Studio
code
to
this
tool
called
draft
which
is
sort
of
a
rapid
iteration
local
rapid
iteration
experience
between
your
cluster
and
your
workspace
on
your
desktop
to
Brigade,
which
is
a
workflows
system
built
on
top
of
kubernetes,
that
people
open
sourced
a
couple
weeks
ago
to
the
work
that
I
was
was
talking
about
that
called
meta
particle
that
I'm
giving
my
keynote
about
tomorrow.
I
I'm,
a
product
manager,
so
I,
tend
to
bring
things
together
in
the
project
in
the
service
of
what
users
are
likely
to
use.
My
talk
is
on
Friday
and
I
will
be
giving
a
demo
of
a
hybrid
deployment
with
two
clusters:
one
that's
an
on-premise
cluster
and
one
that's
a
gke
cluster
I'll
be
using
the
multi
cluster
registry
and
then
demonstrating
that
I
will
also
be
using
the
open
service
broker
and
the
service
catalog
to
show
an
application
that
runs
in
a
hybrid
manner.
So
please
come
back.
That's
at
11:00
on
Friday
all.
E
So
the
question
was
whether
I
was
describing
Federation,
which
is
a
project
that
has
been
in
the
kubernetes
codebase
for
a
while
was
actually
recently
moved
out
into
kubernetes,
Federation
and
Justin.
To
answer
your
question,
the
the
answer
is
the
the
Federation
has
a
cluster
resource
that
we're
talking
about
replacing
now
with
the
cluster
resource
that
comes
from
the
cluster
registry,
good
question.
F
And
I
think
in
general
the
difference
would
be
that
Federation
thought
a
lot
about.
Maybe
what
API
objects
you
might
want
to
define
at
a
worldwide
level.
That
would
then
create
things
in
the
cluster,
whereas
I
think
this
is
merely
about
like
I.
Just
have
a
bunch
and
I
want
to
organize
them
and
I
need
to
get
at
them
in
a
bunch
of
different
ways
and
yeah.
H
F
I
Ii,
the
multi
cluster
thinking
is
have
state
from
multiple
clusters
in
a
single
registry
and
to
be
able
to
operate
on
them
in
an
iterative
way
from
from
the
registry
and
to
have
that
be
a
highly
available.
Control.
Plane
versus
Federation
was
very
much
like
a
tree
where
you
had
sort
of
a
single
point
from
which
you
could
control
all
of
the
api's
and
that
that
single
point
was
to
look
like
any
other
cluster.
These
are
two
different
architectural
approaches
both
meant
to
manage
multiple
clusters.
The.
B
So
we
don't
have
time
for
more
questions,
but
it's
awesome
that
you
have
questions.
You
I
have
one
more
question
that
I'm
gonna
use
to
close
with
and
you
can
find
them
all
at
lunch
or
this
afternoon,
but
before
I
do
that
I
think
any
session
you
go
to
any
conference
you
go
to
to
at
the
end
of
every
session.
You
should
think
about
like
what
did
I
get
out
of
that
like.
B
Why
did
I
sit
there
for
half
an
hour,
and
it
may
be
something
you
heard
that
was
new
and
maybe
something
that
sparked
a
thought
of
your
own
that
you
want
to
follow
up
on,
so
that
you
don't
lose
that
in
this
you
know,
you're
gonna
be
in
a
lot
of
sessions
this
week
and
you've
already
been
a
lot
of
sessions.
This
morning,
I'd
like
you
to
share
with
your
neighbor.
B
All
right
so
I
gave
you
just
enough
time
to
be
a
teaser.
Sorry
about
that.
So
I
encourage
you
to
have
these
conversations
later.
If
you
had
ideas
or
thoughts
that
it
sparked
or
something
you
want
to
share,
please
tweet,
please
blog
and
I'm
gonna
have
our
panelists
close
they're
going
to
tell
you
their
name.
I
need
everybody's
attention,
so
they
can
thanks.
So
awesome
ideas.
Glad
there's
conversation
happening.
Please
talk
to
your
neighbors
at
lunch.
Please
continue
the
conversation.
Please
tweet,
please
blog!
B
Let's,
let's
get
thoughts
flowing
I'm
gonna
ask
our
panelists
to
close
by
saying
their
name
again,
so
that
if
you
walked
in
the
lock
up
to
them
at
lunch,
you
don't
have
to
like
pretend
like
you
know
who
they
are,
and
you
can't
remember
their
name
so
they'll
say
their
name
again
and
then
I
asked
them
to
share
the
cube
con
presentation.
That
is
not
their
own,
that
they
are
most
interested
in
at
any,
are
not
most
interested
by
interested
in
attending
this
week.
So
Chris
you
want
to
lead
us
off
again
sure.
D
So
again,
Chris
Nova
we're
gonna,
have
to
do
the
presentation.
I'm
most
excited
about
this
week
is
a
really
close
friend
of
mine.
Lucky
he
works
at
Microsoft,
he's
doing
a
presentation
on
sto
and
that's
like,
admittedly,
one
of
the
things
I
don't
know
very
much
about
so
I'm
gonna
get
to
go
and
sit
in
this,
like
with
a
pair
of
like
noob
eyes
and
learn
about
his
geo
for
the
first
time
so
I'm
really
pumped
about
that
I'm.
C
E
F
Burns
at
Microsoft,
Azure
I
think
I'm
gonna
cheat
and
say
that
I'm
actually
most
interested
in
talking
to
people
in
the
hallway.
Actually,
the
one-on-one
conversations
is
always
really
Mike,
my
favorite
part
at
the
booth
or
in
the
hallway.
So
please
don't
hesitate
to
come
up
and
chat.
That's
it
I'm.
I
B
Well,
thank
you
all
these.
These
are
awesome
people
in
our
panelists.
They
are
totally
approachable.
So
please,
you
know
bug
them
all
week.
Long
I'll
buy
them
a
beer.
If
you
bug
them
too
much
anyway
there
they
are
all
here
to
meet
all
of
you
and
to
have
these
conversations.
So
thank
you
very
much.
Thanks.