►
Description
Moderated by Diane Mueller, Red Hat
Panelist:
Ken Owens, Cisco
Brandon Philips, CoreOS
Aaron Williams, Mesosphere
Doug Davis, IBM
Joseph Jacks, Apprenda
Chris Wright, Red Hat
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
But
one
of
the
things
that
that
really
we
want
to
make
sure
we
get
through
today
in
this
is
that
you
all
get
a
chance
to
ask
questions
of
this
board
sort
of
and
ask
me
anything
as
well.
I
have
a
few
questions
to
set
it
up
and
to
get
the
conversation
going.
But
please
raise
your
hand
we'll
get
you
a
microphone
and
ask
you
two
to
come
up
and
ask
your
questions
as
well.
A
So
without
further
ado,
we'll
see
who's
ready
back
there
and
bring
out
our
contestants
contestant
number
one,
two,
three
four
five
and
six.
Hopefully
we've
got
ken
Owens
from
Cisco.
We've
got
Aaron
coming
from
mesosphere
whoops
whoops,
whoops,
whoops
and
Doug
from
IBM
Brandon
from
core
OS
there's
two
others
and
I
solved
as
long
as
I
could.
If
you
hear
a
photo.
B
A
Only
missing
one
person
so
that
that's
my
right
if
he
runs
up
on
stage
late,
that's
ok,
cuz!
He
works
with
red
hat
and
I'll
just
put
in
the
Red
Hat
comments.
So
the
CMC
f's
mission
is
been
helping
to
get
this
new
computing,
wonderful
computing
paradigm
off
the
ground
and
getting
the
messaging
and
the
enthusiasm
behind
the
adoption
of
cloud
native.
A
But
we
all
have
different
opinions
on
what
cloud
native
is,
and
all
of
that
so
I'm
going
to
get
these
guys
to
introduce
themselves
and
maybe
talk
a
little
bit
first
about
what
their
idea
is
of
what
cloud
native
means
to
them
as
they
introduce
themselves
and
then
I'm
going
to
hit
them
with
some
questions.
So
Joseph.
D
Sure
so
my
name
is
Joseph
Jax
and
my
entry
into
this
space
was
through
a
company
called
cos,
Maddock
that
I
helped
start
and
we
we
were
sort
of
the
the
first
pure-play
company
behind
communities.
So
it's
been
a
lot
of
time
with
a
coup.
Brandis
project
joined
company
called
aprenda,
which
is
an
exciting
platform
as
a
service
vendor
based
in
New,
York
and
I'm,
also
fairly
involved
in
cnc
F
across
a
few
different
initiatives.
There.
D
D
So
but
I
think
Craig
McKee
codified
it
pretty
well.
With
you
know,
container
package
dynamic
we
scheduled
and
micro
services,
oriented
I
think
that
that
definitely
embodies
some
things
about
about
cloud
native
I
mean
I.
Think
cloud
native
is
really
trying
to
sort
of
get
people
behind
this
idea
of
distributed
systems
as
a
first-class
set
of
principles
when
you,
when
you
start
building
applications
and
deploying
services
in
a
modern
way
and
I,
think
we're
as
an
industry
starting
to
get
much
better
around
defining
what
that
means.
Yeah,
that's
kind
of
my
general
take
that's.
E
E
Today
when
I
was
working
at
Rackspace
and
I'm
sure
every
other
cloud
provider
infrastructure
team
can
corroborate,
is
that
you
know
machines
are
super
bored
they're,
just
blowing
out
compute
out
the
back
end
as
heat
and
so
being
able
to
use
these
patterns
to
better
densely-packed
compute
workloads,
I
think,
is
a
pretty
exciting
out.
Come
out
of
this
style
of
structure,
perfect.
B
Doug
Davis,
an
SCS
m
at
IBM,
been
working
on
open
source
of
standards.
Now
for
about
16
years,
I've
worked
on
various
open-source
projects,
I
laid
back
in
Apache
openstack
cloud,
foundry,
docker,
arnette,
caruso,
doing
this
open
source
stuff
for
a
while.
Now
to
me
cloud
native,
the
very
very
short
answer
is
containers,
but
the
broader
answer
is
onion
black.
The
broader
answer
is
sort
of
the
things
that
brings
to
us
because
for
years
we've
been
talking
about
things
like
you
know,
treating
VMs
as
cattle,
not
pets
and
all
the
you
know,
12
factor
applications.
B
We've
been
talking
about
things
like
that
for
many
many
years
having
repeatable
build
process,
repeatable
test
stuff
and
we've
talked
about
it,
but
it's
never
really
become
a
reality
and
for
some
reason
everything
started
kind
of
converging
between
containers,
cloud
and
everybody
to
sort
of
jumping
on
board.
To
a
point
where
it's
like
you
know,
web
services
is
SOA
on.
Steroids
is
a
ways
a
new,
but
web
services
is
like
the
cool
way
to
say
it.
B
Forgive
me
but,
and
it's
wadad
sneeze
would
cloud
native-
is
all
about
it's
taking
this
new
not
even
knew,
but
it's
the
the
adoption
of
containers
and
wrapping
it
around
all
the
stuff
they've
been
talking
about
for
years,
but
actually
making
it
a
reality.
You
know
the
cic
d
pipelines,
because
everybody
today,
you
would
even
think
about
doing
an
open
source
project
without
the
Travis
setups
are
routed
all
that
stuff.
That's
actually
relatively
new
from
my
experience,
my
first
artem
source.
B
C
C
Are
we
coming
together
to
create
a
market
that
we
can
all
figure
out
how
we
live
in
and
being
able
to
define
ourselves
based
on
that
market
I
think
is
so
helpful,
not
just
for
us
as
platform
developers,
not
just
to
be
able
to
say
you
know
this
is
how
DCOs
fits
into
that
space,
but
also
for
end-users
and
for
all
these
different
constituents
that
have
a
perspective
on
that
market.
So
my
hope
is-
and
maybe
we're
not
quite
there
yet.
C
F
For
the
last
three
and
a
half
four
years
to
me,
cloud
native
is,
is
really
more
of
a
software
architecture
and
it's
kind
of
looking
at
at
hardware
and
software
components
together
and
kind
of
how
you
would
knowing
that.
What's
happened
in
the
past
and
I
think
we've
talked
about
it
on
the
stage
so
far,
true
that
this
is
kind
of
a
natural
evolution,
we're
not
going
to
be
done
after
this
right.
We
can't
say
this
is
now
cloud
native
we're
done.
F
We
can
sit
back
and
our
job
to
finish
right,
there's
always
going
to
be
something
different
and
so
being
extensible
and
and
making
sure
that
your
software
practices
encompass
change
as
part
of
the
norm.
I
think
is
a
big
part
of
cloud
native.
It's
not
just
a
static,
develop
it
once
deployed
and
let
it
just
sit
then
run
for
the
next
ten
years.
It's
kind
of
change
often,
and
how
do
you
take
into
account
those
change
than
the
impact
of
those
changes?
That's
what
cloud
native
is
a
lot
about.
A
G
Chris
right
and
I'm,
the
VPN
chief
technologist
at
Red
Hat
and
for
me
cloud
native,
is
a
lifestyle.
I
live
in
Portland
Oregon,
so
I'm
not
actually
kidding.
It
is
a
lifestyle,
but
what
I
really
mean
is
cloud
native
is
a
set
of
technologies.
It's
a
set
of
practices.
I
think
you've
heard
a
lot
of
like
pieces
of
a
common
theme
here,
we're
solving
problems
that
businesses
have
today.
G
The
world
is
really
changing
rapidly
and
part
of
what
we're
doing
with
cloud
native
is
focusing
tools
and
processes
so
that
humans
can
do
the
interesting
part,
and
we
really
are
evolving
the
state
of
the
art
of
our
lives.
So
that's
it's
kind
of
a
bold
statement,
but
you
know
what
I
mean
by
that
is.
New
technologies
can
really
change
how
how
we
interoperate
with
the
world
with
each
other.
G
So
you
know
wasn't
long
ago
that
we
had
no
mobile
phones
at
all
and
then
phones
become
a
pretty
much
be
big
good
as
part
of
our
lives,
as
you
imagine
kind
of
that
same
evolution
of
Technology
having
a
similar
kind
of
impact
on
our
daily
lives.
I
think
cloud
native
is:
is
the
foundation
that's
helping
us
except
continue
to
accelerate,
is
innovation?
Business
model
changes,
new
technologies,
all
kind
of
coming
together,
and
so
it's
not
just
living
in
Portland
having
a
cloudy
lifestyle,
but.
A
A
People
and
I
want
to
get
a
little
bit
more
into
what
the
CNC
f
means
to
you
and
I'm
gonna
pick
on
Aaron
here
at
mesosphere,
just
a
little
bit
because
I,
you
know
we
have
Cooper
Nettie's
and
Prometheus
under
the
umbrella
right
now
at
CN
CF,
and
you
know
what
what
does
it
mean
that
you
mezzo
fears
belongs
the
CNC,
F
and
then
separately?
What
does
it
mean
for
DCOs
the
CNC
f
and
isn't
there
just
a
little
bit
of
competition
going
on
here?
This.
C
I've
never
heard
this
question
before
us
right,
whatever
yeah,
of
course
yeah.
So
look
and
I
think
it's
it's
a
good
way
to
think
about
sort
of
when
you
think
about
why
Mesa
sphere
participates
in
a
group
like
cnc
f.
That
may
help
to
define
kind
of
what
it
is
as
well.
So
you
know
I
think
I
feel
like
we
should
all
get
sort
of
the
mission
statement
tattooed
on
us
somewhere
and
Joseph
kind
of
mentioned,
the
core
pieces
of
that.
You
know
it's.
C
It's
microservices
its
applications
running
in
containers,
it's
orchestration
and
automation
of
those
applications,
so
those
core
fundamental
pieces
of
what
this
new
computing
paradigm
is
are
very
relevant
to
us.
If
you
look
at
what
we
do
with
dcos
as
an
open
source
project,
and
if
you
look
at
what
we
do
with
our
enterprise
products,
those
things
are
the
embodiment
of
those
principles,
and
so
you
know
I
think
for
us
to
participate
in
this.
C
It's
really
about
coming
together
with
these
leaders
in
the
industry
to
create
this
market
and
to
be
able
to
help
reduce
some
of
the
confusion.
I
think
around
what
it
means
to
be
cloud
native.
We
all
have
sort
of
this
goal
to
get
everyone
on
board
with
that
single
market
and
that
sing
inside
that
single
market
and
with
that
single
definition.
But
but
that
might
be
the
easy
answer
actually
I
think
the
more
interesting
answer,
for
me
at
least,
is
looking
at.
Who
else
is
participating?
C
And
you
know,
one
of
the
reasons
we
participate
is
because
this
group
of
people
this
this,
this
CNC
f,
is
so
diverse.
It's
not
just
about
you,
know
cloud
providers
or
platform
providers
or
ISVs
or
end
users
or
blah
blah
blah.
It's
about
all
of
us
coming
together
to
provide
opinions
and
perspectives
on
this
topic,
and
so
you
know
for
us
for
Mesa
sphere.
That's
one
of
the
things
that
that
makes
it
so
exciting
for
us
makes
it
why
we
are
so
active
and
excited
about
the
future
of
ciencia
yeah.
A
So
I
think
chrisann's
exceda
it
in
some
announcement
this
past
week
at
Linux
con
about
the
diversity,
the
number
of
startups
that
are
coming
out
of
the
space
that
are
built
on
Cooper
Nettie's
and
lots
of
announcements
came
out
this
past
week.
So
one
of
the
healthy
things
about
this
marketplace
and
this
group
of
projects
that
are
working
together
is
the
diversity
that's
coming
out
now.
Definitely.
A
B
Right
I
was
going
to
say
they
said
the
first
governor
board
meeting
seen
CF.
There
was
a
lot
of
discussion
about
there
being
confusion
in
the
industry
and
that
thing
that
confusion
has
doubled,
lost
different
aspects
to
it
and
I
think
yes,
cnc
f,
should
be
there
to
help
promote
cloud
native
technologies.
Clear
up
some
of
the
confusion
that's
out
there,
but
the
one
thing
that
I
would
really
love
to
see,
and
I
don't
we
think
we
talk
about
it.
B
B
If
you
get
an
agreement
on
things
like
api's
fight
over
the
invitationals
api's
as
a
goal,
I
just
want
that
level
of
harmonization
across
the
industry
from
the
CNC
I,
because
I
think
by
having
all
the
players
at
the
table
alone,
could
help
foster
that
without
the
c-in-c
fu
may
not
even
get
them
talking
to
another.
So
I'd
like
to
see
that
more
collaboration
there
and.
A
F
On
the
CNC
FM
on
the
technical,
Oversight
Committee,
and
then
one
of
our
goals
is
to
really
help
sort
of
incubate
projects
that
want
to
be
part
of
this
cloud
native
movement.
Maybe
they
they
have
pieces
of
it
that
fit.
Maybe
they
don't
so
kind
of
help
Shepherd
them
help
them
identify
how
to
be
part
of
this
overarching
cloud
native
set
of
capabilities
and
projects
that
we
weren't
putting
together.
A
So
you
mentioned
projects
and
the
CNC,
f
and
I.
Think
Dan
cone
put
a
whole
list
of
the
new
potential
projects
that
the
TOC
is
looking
at.
Everything
from
things
that
are
near
and
dear
to
my
heart,
like
xcd
and
flannel
and
open
tracing,
and
that
so
there's
there's
a
whole
list
of
things.
The
TOC
is
considering
now
and
reviewing,
and
not
all
of
them
will
will
make
the
cut
and
become
in
there.
A
But
are
there
any
things
that
you
guys
are
seeing
that
are
missing
or
is
there
one
that's
a
favorite
of
yours
that
you'd
really
think
that
we
should
yeah.
I
was
thinking.
How
do
we
get
a
patchy
foundation
to
let
go
of
my
toes
and
bring
it
over
and
get
it
in
the
CNCs,
then
we'll
all
be
happy
right.
But
aside
from
trying
to
figure
out
that
divorce
settlement
either
are
the
projects
maybe
Chris
that
you're
you're
thinking
that
we
really
should
be
looking
at
that
we're
not
yet
well.
G
I
think
a
couple
of
things
on
this
on
this
topic:
one
is
from
a
umbrella
project,
point
of
view.
We
shouldn't
set
a
bar
for
the
CNC
f
to
measure
itself
on
how
many
projects
do
we
have
it's
really
a
negative
incentive,
it's
kind
of
pushing
us
in
the
wrong
direction.
In
my
opinion,
and
then
you
know
that
it's
a
nice
list
of
projects
I
think
we've
got
some
really
promising
prospects
coming
in,
but
I'd
almost
take
a
step
back
and
say
we're
still.
G
You
know
and
we'll
probably
explore
some
of
that
on
this
panel
today,
we're
still
trying
to
define
what
what
cloud
native
is
from
an
architectural
point
of
view
and
the
better
clarity
we
can
provide.
From
that
perspective,
it
will
help
us
see
where
there's
good
matches
for
projects
that
exist
or
or
maybe
more
importantly,
white
spaces
that
are
that
are
work
to
be
done,
and
we've
had
some
experience
with
this
in
an
open
shift.
G
Where
were
you
know,
we're
absolutely
leveraging
doctor
and
crew
benetti's
under
the
hood,
but
we're
also
closing
gaps
in
the
open
shift
layer
that
don't
exist
in
other
projects,
and
so
I
know
that
everything
is
new
and
evolving
and
I.
Think
that
too,
you
know
great
list
of
projects,
I'd
love
to
ensure
that
we're
kind
of
taking
a
holistic
view
and
looking
especially
for
gaps
in
white
spaces.
G
E
E
Like
you
know,
these
two
things
actually
makes
sense
together
when
put
into
the
same
infrastructure,
and
then
the
other
piece
is
like
one
thing
that
the
scene
CF
has
a
unique
opportunity
to
do.
Is
that
post,
Apache
foundation,
style
projects
that
need
like
documentation,
hosting
and
mailing
us,
etc?
Like
there's,
google
groups
and
github
that
handle
those
things
today,
a
lot
of
these
projects
are
maybe
single
vendor
projects
that
we're
considering
or
their
projects.
That
may
not
have
the
best
structure
around
making
decisions
and
that
sort
of
thing
and
I
think
soon.
E
Cf
could
not
necessarily
dictate,
but
at
least
help
and
encourage
certain,
like
best
practices
of
how
you
actually
come
to
decisions.
How
you
like
make
processes
happen?
How
do
you
make
releases
and
that
sort
of
thing
so
I
think
there
are
like
projects
are
important,
but
I
think?
What
is
the
value
that
the
scene
CF
can
uniquely
provide
by
adopting
a
project
is
something
that
needs
to
be
sorted
so.
A
So
one
of
the
the
the
value
propositions
for
the
CNC
F,
like
Apache,
is
being
a
neutral
host
and
that
but
Apache
has
a
lot
of
other
things
that
they
add
on
in
governance
and
processes,
and
things
like
that
and
for
what
I've
been
listening
at
the
CN
CF
and
I'm,
not
on
the
board
of
governors
or
the
TOC.
Is
that
weird
kind
of
hands
off
on
the
projects?
We
let
the
Cooper
Nettie's
and
the
Prometheus
folks
govern
themselves
and
mediate
their
disputes
and
figure
out
who's
going
to
be
a
committer?
A
E
I
think
the
point
is
that
it's
there
are
things
that
the
scene
staff
can
uniquely
do
as
like
an
umbrella
organization,
and
we
need
to
figure
out
what
those
things
are.
It
is
it's
clearly
not
offering
like
hosting
services,
and
it's
it's
not
necessarily
about
like
choosing
the
one
solution
that
will
always
be
correct.
So
what
is
that?
What
is
that
thing?
I
think
the
Prometheus
and
Cooper
Nettie's
thing
is
a
good
example
of
what
could
emerge.
But
what
is
it?
A
So
one
of
the
new
things
that
the
CNC
f
is
done
is
offer
the
cluster
yep
with
Intel's
and
switch
has
helped
to
create
this
cluster.
For
us,
openshift
we've
taken
advantage
and
blowing
up
blown
through
1000
nodes
and
scaled
things
up
and
that's
kind
of
that's
almost
unique
to
any
any
foundation.
I'm
not
sure
if
apache
does
that
or
anybody
else
I've
ever
heard
of
that
me.
A
G
There's
some
actual
justification
behind
that
which
is
especially
when
we're
trying
to
bring
in
some
initial
well
established
projects.
If
we
create
a
hurdle
that
says
by
the
way
you
have
to
adopt
our
governance
policy,
you
know
we're
pushing
people
away,
so
I
think
11
very
likely
outcome
is,
as
we
create
some
inertia
and
momentum.
You
can
distill
out
some
best
practices
and
try
to
share
those
with
the
related
projects,
so
we're
not
necessarily
dictating
from
a
CNC
F
best.
You
know.
C
C
It's
all
those
sort
of
avenues
of
being
able
to
connect
to
a
community
that
I
think
CNC
f
is
doing
a
good
job
of
starting
to
create
and
foster
and
for
you
know,
either
a
member
of
the
community
or
for
these
projects
that
are
coming
into
the
community
having
those
resources
at
your
fingertips
being
able
to
go
out
to
a
community,
a
large
community
of
people
that
are
ready
to
hear
about
sort
of
what
you're
doing
and
how
you're
doing
it
is
extremely
valuable.
Yeah.
A
So
Joseph
you
you've
started
coupe
calm.
Thank
you
very
much.
One
of
the
most
awesome
call
conferences
ever
lots
of
fun
and
we're
going
to
do
that
again
in
Seattle
shortly
and
set
the
eighth
and
ninth
of
November
so
I
already,
please
come
to
that
as
well,
but
and
so
we've
got
a
bit
of
a
community
around
kuber
Nettie's,
going
more
than
a
bit
just
a
bit
just
a
bit.
What
do
you
see
as
some
of
the
inhibitors
to
continuing
to
this
or
inhibitors
to
the
adoption
of
cloud
native
yeah.
A
D
So
Brenda
and
I
were
actually
talking
briefly
about
this,
but
just
at
a
high
level,
I
think
it
all
comes
down
to
what
what
natively
gets
built
into
Cooper
Nettie's.
That
may
not
necessarily
need
to
get
building
Cooper
Nettie's,
and
I
think
cnc
f
could
actually
play
a
role
in
that
in
terms
of
drawing
some
of
the
lines
there.
D
So
I'm,
actually
a
really
big
fan
of
something
that
Brandon
is
proposed
recently
on
this,
which
is
you
know,
sort
of
a
lightweight
incubation
process
for
Cooper
Nettie's,
which
which
really,
if
you
unpack,
you
know
what
the
what
the
proposal
outlines,
which
I
won't
get
into
in
detail.
But
it's
a
fairly
comprehensive
proposal.
It
really
gives
a
lot
more
clarity
to
people
who
want
to
contribute
code
to
this
extremely
fast.
You
know
essentially
escape
velocity
level,
open
source
project
and
and
mitigates
a
lot
of
the
confusion.
D
I
think
around
what
Cooper
Nettie's
is
today
and
what
will
it
evolved
into
within
the
next
nine
to
18
months
and
I?
Think
that's
actually
very
related
to
what
what
the
inhibitors
could
be
for
something
something
successful
like
like
like
this,
this
open-source
project
so
get
getting
a
handle
on
a
lot
of
those
things.
I
think
a
lot
of
the
lessons
that
can
be
learned
and
that
can
get
translated
into
what
CN
CF
ends
up
ends
up
codifying
for
people.
D
You
know
I'm,
not
a
huge
fan
of
standards.
I
think
standards
are
very
limiting
to
people
but
best
practices,
and
you
know
guidance
and
some
of
the
some
of
the
the
aspects
of
community
development
with
the
open
source
projects
themselves
and
in
offering
those
insights
to
the
projects
with
that
sort
of
dictating
the
you
know
the
structure
of
the
you
know:
the
engineering
culture,
the
the
upstream
open
source,
you
know
frameworks
so.
D
I
think
so:
I'm
not
I'm,
not
I'm,
not
I'm,
not
really
clear,
on
what
some
of
the
implications
are
too
big
tent,
but
you
know
I,
think
I,
think
those
are
just
some
of
the
the
quick
ideas
around
what
what
we
should
be
careful
with
in
terms
of
consequences
and
inhibitors
to
continued
growth,
because
there
are.
There
are
a
lot
of
risks
right
now,
with
with
Cooper
nights
as
a
projects.
Cotton,
so
much
more
successful
tons
of
people
want
to
contribute
to
it.
D
E
It's
yeah
I
think
it's
awesome
to
see
the
the
huge
amount
of
traction
and
like
the
the
assignment
around
adding
new
features
and
new
new
extension
points
to
the
platform.
I
think
it
goes
back
to
the
conceptual
consistency
and
ease
of
the
use
of
the
platform
that
people
are
able
to
say.
Oh
yeah
works
great
for
me,
but
now
I
want
to
add
these,
like
two
other
widgets
on
the
side
of
it.
So
it's
exciting
to
see
it.
E
We
also
just
need
to
make
sure
that
we
we
don't
end
up
with
like
the
API
gigantic
sprawl
that
has,
to
a
large
degree
cause
like
the
linux
kernel
would
be
so
successful
over
the
last
25
years.
Is
that
api's
were
added
very
slowly
and
with
like
a
lot
of
thought,
even
probably
less
thought
than
we
might
have
hoped
like
we're
on
my
dupe
for
or
something
in
the
linux
kernel.
But
it's
another
thing
but
like
you,
you
want
to
have
like
you
want
the
excitement.
E
They
also
need
mechanisms
like
through
incubation
or
other
things
to
say,
let's
hold
on
for
a
second
on
this
one.
Let's
feel
it
out
for
the
next
6-12
months
so
that
we
don't
have
like
yet
another
widget
that
we
have
to
put
up
on
a
slide
that
we
have
to
put
into
our
architecture
diagrams
and
that
sort
of
thing
the.
A
A
How
do
we
make
Cooper
Nettie's
the
best
place
for
high
performance
workloads,
so
the
tweaks
that
we
need
to
do
to
specific
workloads,
I
think
incubating
those
and
being
careful
about
that
is,
is
really
going
to
be
key
for
some
that
I
see
you
nodding
down
there
Ken
and
it's
not
nodding
off.
I
know
it's
been
a
long
week,
but
do
you
see
any
other
inhibitors
to
for
the
CNC
f,
any
roadblocks
ahead,
I
think.
F
My
biggest
concern
is
in
a
community.
If
you
have
bad
actors,
you
can
really.
You
know
to
split
the
community
up
across
the
community
to
have
you
know,
issues
that
you
can't
overcome
but
easily
and
so
I
think
one
of
that's
one
of
my
concerns
is
making
sure
that
we
act
as
a
community,
not
as
individuals
that
have
an
agenda
that
want
to
push
the
community
one
way
or
the
others.
F
A
C
A
Oh,
that's
all
right.
Anyone
else
having
an
inhibitor
that
they
want
to
bring
out
okay.
So
there
was
one
of
the
things.
What
are
some
of
the
challenges
that
you're
seeing
users
who
are
starting
to
adopt
cloud
native?
So
maybe
Brandon,
you
sure.
E
E
So
one
of
the
first
inhibitors
are
the
one
that
were
attacking
along
with
the
kuru
Nettie's
community
like
head
on.
Is
that
operations
team
see
this
diagram
of
software
that
they
like?
They
see
they
vanished.
They
see
the
advantage
of
being
able
scalar
thing
having
consistent
API
across
different
infrastructures,
but
then
they
see
the
diagram
of
software
that
they
have
to
manage
and
operate,
and
then
they
have
lots
of
questions
around
like.
How
does
that
monitored?
How
do
I
upgrade
it
like
who's
upgrading
it?
When
do
I
upgrade
it?
E
Is
it
safe
to
upgrade
it
when
I,
like
my
workloads,
are
going,
and
so
these
are
things
that
we're
tackling
with
a
bunch
of
other
folks
in
the
upstream
of
like
cure,
the
api's
for
upgrading
of
the
system
through
the
API
for
adding
and
removing
extension
points
with
its
networking
or
storage
or
service
discovery
and
I?
Think
that's
going
to
be
just
laying
that
foundation
from
the
media.
The
project
is
very
young,
where
one
year
out
from
andado
and
being
able
to
have
those
patterns
of
the
Opera,
not
just
the
advantages
to
the
operator.
E
The
application
teams
that
are
deploying
on
top
of
that
infrastructure
came
that
actually
has
to
manage.
It
I
think
it's
going
to
be
pretty
critical,
because
those
are
the
folks
that
are
going
to
have
to
like
sign
off
on
wearing
the
pager
eventually
and
if
we
don't
get
them
early
on
they're
going
to
be
like
pretty
pretty
well
in
there
right
pretty
well
wanting
to
avoid
those
technologies.
E
G
Starts
with
being
clear
on
what
what
the
value
is,
so
I
still
hit
questions
all
the
time
of
why
containerized
you
got.
We
deal
out
with
enterprises,
enterprises
have
a
lot
of
legacy
involved
and
so
they're,
not
even
fully
virtualized
and-
and
so
one
of
the
things
we
can
do
is
really
be
clear
about
containerization
and
virtualization
have
a
lot
of
shared
concepts
but
they're,
not
just
sort
of
redundant
or
immediate
evolution
of
concepts.
There's
there's
some
really
fundamental
differences
that
helping
people
understand.
Why,
like?
G
What's
the
value
I
hear
all
the
time
like
you
sounds
like
I
see
the
kind
of
Holy
Grail,
sprinkle,
pixie
dust
and
and
I'm
magically
going
faster
and
everybody
wants
to
go
faster,
but
I
don't
actually
have
that
problem.
So
help
me
understand
and
complexity
is
the
other
thing
that
comes
up
because
we
spend
from
an
engineering
point
of
view.
We
spend
a
lot
of
time,
building
awesome
technology
and
it's
critical
to
have
user
input
from
developer
perspective,
but
also
user
input
from
operations
teams,
because
we
do
see
this
regularly.
G
You
get
the
demo
the
demos
awesome
day.
One
day
two
is
a
train
wreck
and
you
don't
get
a
second
chance
for
quite
a
while.
So
helping
users
understand
the
value
and
see
how
they're
going
to
kind
of
live
with
this
infrastructure,
they're
building
out
that's
going
to
need
to
itself
evolve
over
time.
Probably
the
two
two
important
things
that
come
to
mind
from.
C
So
I
want
to
I
want
to
double
back
on
the
second
day
operations,
because
that
to
me
is
super
critical.
We
hear
that
all
the
time
sort
of
it's
not
just
about
how
do
I
adopt
this
new
shiny
thing
you
have,
but
how
do
I
actually
use
it
in
real
in
production
in
anger,
so
I
think
that's,
that's
such
an
important
topic
and
something
that
we
as
cnc
I
can
do
a
better
job.
C
How
can
we
put
together
the
use
cases
that
help
illuminate
how
we
solve
those
problems?
I'd
be
happy
if
we
did
a
few
of
them
really
well
and
we
were
able
to
focus
as
a
community
on
doing
a
few
of
them
really
well,
rather
than
trying
to
necessarily
boil
the
ocean
I
hear
a
lot
from
folks.
Be
careful,
don't
boil,
do
don't
try
and
do
everything
for
everybody,
let's
focus
on
a
few
things:
let's
go
off
and
solve
those
few
things
very
well
and
I.
C
B
It's
actually
good
and
a
bad
thing
in
the
sense
that
there's
almost
too
much
choice
out
there
right.
You
got
criminais
days
docker
with
the
112
orchestration
stuff.
He
got
mesosphere
and
I.
Think
customers
are
very
confused
of
which
path
to
go
for
some
of
these
technologies
right,
there's
a
lot
of
choices
out
there.
B
Can
we
figure
out
some
way
as
a
community
say
yeah,
we
may
have
different
points
of
view
in
terms
of
the
underlying
technology,
but
can
we,
from
the
customer's
point
of
view,
say,
take
your
one
application
definition
and
build
a
deployed
in
multiple
places?
No
docker
versus
communities,
you
know
and
not
ethnic
I
have
to
rock
the
entire
customers,
the
customers
entire
world.
Just
to
make
that
happen
right.
B
A
There
was
a
lot
of
conversation,
I
know
over
the
past
week
and
at
other
sea
and
CF
events
about
the
CNC
F
is
not
about
picking
winners
right
from
the
outside.
It
may
look
like
it.
We
you
know
that
we
pick
kuba,
Nettie's
and
crowned
it
the
winner
or
we're
going
to
create
the
new
stack
or
whatever,
but
in
some
senses
what
we're
doing
is
we're
creating
an
infinite
possible
ray
of
possibilities,
which
is
infinitely
confusing
when
you're
in
an
enterprise,
IT
operations,
group
and
you're
trying
to
say
well.
B
B
I'm,
not
analed
and
I,
don't
think
we
should
initially
pick
winners,
but
that
doesn't
SME.
We
have
to
expose
our
customers
to
all
those
variabilities
there.
If
we
could
we
minimize
those
into
some
extent,
I
think
that'd
be
a
little
happier.
Then
they
don't
have
to
feel
the
pressure
of
possibly
getting
locked
into
one
solution.
That
may
not
actually
be
the
right
one
for
them.
So.
A
So
so
we
have
we've
quite
a
bit
of
time
left
to
and
if
people
have
questions
in
the
audience,
please
just
spit
you
have
one
here:
I
see
a
hand.
Do
we
have
a
microphone
Oh
or
just
yes
here.
H
Excuse
me
Doug
mention
the
application
definition,
and
this
is
something
I
was
really
curious.
About
is
dude.
Do
you
see
as
a
role
of
the
CNC
f
as
defining
that
creating
that
application
specification
like
OC,
is
creating
one
for
runtime
and
44
image,
but
if
you've
got
an
application,
that's
composed
of
50
different
types
of
containers
and
have
to
be
connected
in
how
many
different
ways?
How
do
you
decide
how
to
roll
that
out?
H
G
There's
there's
six
of
us
up
here,
so
you'll
probably
get
what
seven
different
opinions,
but
we
actually
thought
about
this
in
the
beginning,
and
we
wrote
down
initially
that,
yes,
it's
it's
in
scope
and
I
think
the
challenge
is,
and
so
I
personally
do
believe
it's
in
scope
and
I
think
it's
actually
I
I'm
glad
you
brought
it
up,
because
I
think
it's
it's
that
top-down
view.
That
actually
is
what
makes
this
whole
thing
usable,
I
think
maybe
between
that
and
well-defined
services.
G
Those
are
probably
the
two
most
important
pieces
to
me
to
focus
on
the
challenge.
Is
you
risk
going
down
the
super
abstract
dsl
that
can
do
everything
and
then
doesn't
get
adopted
because
it's
too
complex?
Nobody
can
use
it
or
it's
actually
really
designed
to
be
a
mezzo,
specific
thing
or
a
coup,
burnetii
specific
thing,
and
it
hasn't
really
achieved
the
right
level
of
abstraction
and
pretty
confident.
B
I
mean
field
would
have
said
dr.
with
the
docker
image
when
thought
have
been
possible.
Yeah
they've
managed
to
standardize
that,
to
a
certain
extent,
forgive
me
for
using
the
S
word,
but
yes,
ended,
adds
to
a
certain
extent.
So
it's
not
I.
Don't
think
it's
inconceivable
that
we
can
move
up
the
stack
we
just
may
have
to
go
slow
as
Brendan
was
saying
you
don't
want
go
to
quickly,
but
and
his
Chris
was
saying
I
think
we
should
at
least
try
I
think.
D
You
know
docker
swarm,
there's
lots
of
conversations
in
that
community,
and
so,
if
there,
if
there
was
a
common
abstraction
at
that
layer,
think
that
would
be
a
really
good
additive
step
forward
to
then
sort
of
start
to
look
at
you
know
how
do
we?
How
do
we
solve
communication
discovery
and
secret
management
configuration
all
these
other
things
that
really
are
required
to
build
sort
of?
How
do
you
take
an
application
and
make
it
portable
across
these
different
yeah.
E
Lambda
I
think
also.
We
need
to
understand
what
are
the:
what
are
the
types
of
applications
that
are
being
deployed
on
these
infrastructures,
so
one
of
the
things
that
we're
talking
about
at
khoob
colonists
like
managing
a
complex
application
like
Etsy
d,
which
is
a
distributed
database
on
top
of
Cooper,
Nettie's,
api's
and
I-
think
it's
really
easy
to
imagine
building
an
application
definition.
That's!
Oh!
It's
super
useful,
deploying
WordPress,
but
actually
that's
called
Apple.
E
That's
like
not
the
application,
these
companies
care
about
I
talk
to
customers
where
they
have
like
multi-tiered
custom,
caching
structures
with
sharding
and
replication.
It's
like
we're,
probably
not
going
to
use
the
same
application
definition
that
model
WordPress
like
perfectly
in
order
to
make
them
successful
and
I.
Think
that's
something
that
we
probably
need
to
start
out
with
first
before
even
thinking
about
what
is
the
JSON
schema
for
this
thing
is
what
are
probably
the
for
applications
that
we
actually
feel
like.
We
can
confidently
want
to
model
like.
E
Is
it
a
distributed
database,
maybe
like
two
tier
app
and
etc
and
I?
Don't
think
we,
at
least
in
my
experience,
every
single,
every
single
customer
right
now
is
a
beautiful,
unique
flower,
but
they
really
don't
have
like
this
good
language
around
that,
and
that's
something
Joseph
and
I
talked
about
earlier
and
I.
Think
it's
an
important
thing
that
we
need
a.
We
need
a
sort.
A
I
The
traditional
vendor
is
in
need
of
some
design
patterns
at
an
application
architecture
level
to
you
know,
to
gain
maximum
out
of
cloud
and
and
and
which
is
what
you
know
I'd
imagine
cloud
native
is,
is
you
know
how
do
you
take
an
applicable?
What
are
the
design
patterns
underpinning
a
you
know
software,
which
would
make
it
run
well
on
cloud
now.
The
hyperscale
companies
have
done
it
forever
and
they
do
it
very
well,
but
not
in
the
traditional
enterprise
space,
which
is
quite
some.
I
You
know,
while
back
now
here
in
the
sessions,
I
understood
you
defined
it
as
microservices
running
in
a
containerized
ecosystem.
You
know
and
communities
and
in
all
of
that,
but
is
it
within
the
scope
and
in
perhaps
if
you
notice,
as
it
touches
on
the
question
as
previously,
but
is
it
in
the
scope
of
cnc
f
to
outline
what
those
design
principles
are?
Micro
services
themselves
have
been
a
fairly
recent.
You
know
so
in
the
last
one
or
two
two
years,
you
know
it's
begun
to
become
popular
with
netflix
cited
as
a
reference.
I
A
A
D
Bay
enterprise-
you
know
huge
huge
telco
company,
actually
voiced
something
that
we
were
talking
about
earlier,
which
is
like
how
do
you
codify
as
a
khata
fire
code?
If
I
think
both
are
about
yeah
right?
Yes,
yes,
how
do
you
codify
like
what
cloud
native
or
microservices
means
in
the
application
development
stage?
So
as
you
have
people
writing
code
and
building
out
a
set
of
services,
building
their
application
and
pushing
it
somewhere?
D
You
know
the
rise
of
the
internet,
new
protocols,
new
new
languages
and
so
on,
but
I.
I
still
think
a
lot
of
these
fundamental
problems
have
been
thought
about
previously,
and
actually
you
know
really,
you
know,
investigated
and
and
and
codified,
so
I
actually
looked
at
this
book
that
was
written
called
enterprise
integration
patterns,
I'm,
not
sure
if
you're
familiar
with
it
at
all.
D
You
know
how
all
these
different
patterns
could
be
codified,
using
intelligent
routing
mediation,
translation
and
brokering
of
messages
between
services,
and
all
of
these
things
are
very
well
codified
and
they
actually
are
the
sort
of
reference
playbook
for
how
the
technologies
have
been
implemented
against
them.
So
things
like
tibco
and
websphere,
and
on
and
on
and
on
in
the
20
billion
dollar
SOA
industry
right,
which
is
now
in
decline
because
we're
seeing
these
new
these
new
patterns
I
think
we
can
actually
take
a
lot
of
lessons
from
that
and
we're
starting
to
even
see.
D
Maybe
some
subliminal
lessons
taken
from
that.
So
there
was
actually
some
distributed
systems
patterns
papers
done
recently
by
Brendan
Byrnes
and
David
Oppenheimer
at
Google,
around
scatter
gatherer
and
the
Ambassador
patter,
and
these
other
things
that
are
actually
like
direct.
You
know
mirrors
of
what
actually
was
written
by
Gregor
Hopa
in
2003
over
a
decade
ago,
so.
D
And
apply
them
to
the
the
cloud
native
container
space,
but
there's
a
lot
of
lessons
there.
So
I
think
that
would
actually
help
I
think
it
I
was
talking
with
Alexis
Richardson
about
this,
while
back
I
think
it
could
be
actually
something
very
much
so
in
scope
inside
the
CNC
f.
So
to
kind
of
answer
your
question:
that's,
like
that's
my
take
on
it.
My
personal
take
but
yeah
I
think
it's
something
we
should
do
and.
F
I
can
say
on
you
know,
from
the
TOC
side
of
things
right.
We,
the
architecture,
starts
with
the
application
layer
down,
not
the
not
the
infrastructure
up,
and
so
we
definitely
see
the
value
of
focusing
at
the
application,
defying
the
primitives
that
matter
to
app
development
in
a
cloud
native
way.
We
need
to
look
at
the
translations
or
the
transformations
or
the
integration
patent
of
whatever
new
word
you
want
to
use
it
to
node
would
write
the
new
is
the
old
and
the
oldest
the
new
always
so.
F
F
Down
that
and
then
let's
see
ma'am
definitely
so
so
one
year,
I
think
we'll
have
a
very
well-defined
architecture,
hopefully
with
a
set
of
integration
patterns
and
practices
are
on
how
to
incubate
your
project
from
an
idea
to
to
an
actual
cloud
native
set
of
services
and
capabilities,
and
then
in
two
years
from
now,
I
think
well,
I'll,
be
retired
and
drinking.
You
know
my
ties
on
the
beach
yeah
Aaron.
C
You
know
I
actually
think
we
have
other
sort
of
challenges
and
problems
that
we're
trying
to
solve
that
are
actually
more
valuable
than
just
the
sheer
number
of
projects
we
have
internally.
So
maybe,
if
I
was
to
make
a
bold
bed,
I'd
say
that
you
know
we
have
maybe
fewer
projects.
Then
people
might
think
in
six
to
nine
months,
but
we've
actually
produced
things
that
are
far
more
valuable
to
the
community
and
to
see
and
CF,
as
as
the
definition
of
CNC
F,
which
will
be
better
for
us
long
term.
So
all.
C
A
B
No,
it's
funny
is
by
the
answer
for
the
36
months
was
there's,
there's
a
lot
of
choices,
a
lot
of
confusion,
a
lot
of
bickering,
and
it's
only
gonna
get
worse.
Not
that
exciting.
That
was
gonna.
My
prediction
is
gonna
get
worse,
but
now
that
you
push
it
out
to
a
year
to
two
years,
I
actually
think
in
that
time
frame.
I
think
things
will
start
to
settle
down.
B
So
it
a
little
bit
of
prediction
a
little
bit
of
hope
that
within
that
time
frame
the
community
inside
C
and
C
F
Willa
throughout
you
know,
seeing
simple
have
gotten
its
legs
by
then,
and
we
would
have
actually
started
headed
down
that
process
of
doing
the
harmonization
and
getting
more
than
your
ability
type
statement.
That
I
was
hoping
that
we'd
actually
get
to
eventually
I'm
very
impatient
so
I'm
anxious
for
the
get
started,
but
it's
a
new
organization
but
I
think
within
a
year
to
two
I
think
will
be
along
that
way.
E
People
may
not
be
able
to
articulate
what
what
it
was
or
the
name
of
it,
but
you
get
into
a
room
with
them
and
they
whiteboard
exactly
like
the
classic,
like
three
tier
application
pattern
and
I
think
in
a
similar
way,
when
people
start
to
gain
confidence
with
this
stuff
and
they
are
able
to
confidently
explain
like
this,
is
our
application
lays
out.
You
know
it's
pretty
standard
I
think
that's
well
when
we
know
that
we've
hit
success,
they
may
not
ever
say
cloud
native,
but
they
will
understood
and
internalized
the
patterns.
D
I
mean
a
lot
has
already
been
said,
so
I
would
kind
of
a
grin
and
fortunately,
I,
would
kind
of
agree
with
Doug
I
think
the
confusion
is
going
to
get.
You
know
just
sort
of
on
the
periphery
of
the
of
the
project,
probably
going
to
get
worse
in
the
next
three
to
six
months.
Hope
it's
not
the
case,
but
I
think
that's
where
things
are
sort
of
trending.
Unfortunately,
right
now
on
the
longer
term
horizon,
which
was
your
question
like
one
to
two
years,
I
mean
that
is
astronomically
long
period
of
time.
D
We're
reaching
the
event
horizon
here:
let's
see
wow
I
that
is
I,
wouldn't
even
be
presumptuous
enough
to
make
prediction
around
where
we're
gonna
be
in
one
or
two
years.
With
this.
With
this,
with
this
whole
movement,
I
would
hope
that
it's
going
to
be
very
successful
in
that
you
know,
cnc
f
is
going
to
be.
You
know,
viewed
as
an
industry
shaping
transforming.
You
know
entity
for
all
this,
but
that's
a
hope.
All.
G
Right
in
a
year,
I
do
think
we
will
have
some
architectural
clarity
I
think
we
will
have
struggled
through
a
lot
of
argument
and
come
up
with
something
that
we
can
agree
on
from
cnc
f
perspective.
What
are
we
building
and
like
I
said
the
beginning,
started
to
identify
some
of
those
key
services
and
even
most
portly
gaps?
G
I
think
around
that
same
time,
we'll
have
a
hundred
percent
utilization
of
a
thousand
node
cluster
and
I
think
we
will
have
five
anchor
projects
that
are
sitting
in
there
being
you
know,
sort
of
continuously
used
as
reference
patterns
two
years,
and
this
is
really
bold.
Two
years
I
think
we
will
have
figured
out
networking
and
kind
of
container
space
be
still
my
heart
wow.
A
A
So
I
just
want
to
thank
all
the
panelists
for
coming
today
and.