►
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Introduction to the CNCF App Delivery SIG - Alois, Dynatrace & Lei Zhang, Alibaba
The CNCF App Delivery SIG focusses on all aspects of delivery and operations of cloud-native applications from a developer and operations perspective. Current working topics include application definition, packaging, and distribution, as well as the application delivery workflow. Key deliverables of the working group are guidelines and reference material for these topics. In this intro session, we will walk through the detailed goals of the SIG and present an overview of current work items and intermediary results. This will be followed by a discussion of the current roadmap and potential future contributions. If you are developing or managing cloud-native applications or defining company best practices you should get engaged with the application delivery SIG.
https://sched.co/WeWJ
A
So
before
we
get
into
the
topic
here,
who
is
familiar
with
what
the
Stig
does
in
the
CDF
right?
Who
is
familiar
with
what
the
app
delivery
sick
to
us
in
this
means?
Yes,
some
friends
over
there,
okay,
so
that
might
actually
be
the
good
start
in
here.
So
so
what
yeah
and
there's
one
more
thing
here.
So
as
we
know,
this
is
just
above
after
the
lunch
break.
A
We
want
this
to
be
kind
of
like
an
interactive
exercise
and,
as
some
of
you
not
familiar
with,
was
thick
thick
steel
part
of
what
sticks
to
especially
and
use
a
focusing
six,
they
collect
feedback
from
you,
the
community,
where
you
need
help,
collect
or
input,
and
what
to
do.
If
you
use
this
bar
code
over
here,
it
will
actually
take
you
to
a
small
service
that
we
believe
well.
A
This
talk
to
collect
some
feedback
on
some
topics
that
are
related
to
the
sake
work.
So
after
that
session
we
will
obviously
post
the
results.
Sent
emali
also
use
your
feedback
as
a
little
blog
post.
So
just
summed
up
if
it
works,
it
should
actually
work
excited
before
I
get
to
great.
What
particles
can
do
right?
A
Yes,
it's
not
by
where
the
barcode
is
not
part
of
the
work
of
that
the
day
after
the
visit,
so
I
think
the
most
important
yeah
again
briefly
mentioned
it,
there's
from
the
people
that
we
have
Brian
Harry
and
myself
for
the
chairs,
and
the
day
from
that,
you
see,
we
have
Alexis
and
Michelle
who
are
Delia
software,
it
he
is
busy,
and
it's
also
gets
us
to
the
point.
What
does
the
statistics
actually
do?
The
six
support
that
you
see
on
specific
topics?
A
That's
where
you
might
be
familiar
with
some
texting
platform,
Signac
working
was
from
mainstage
and
I.
Think
after
stomach
was
around
September,
where
we
formed
sick,
app
delivery.
The
goal
of
the
app
delivery
sick
is
really
help.
People,
as
the
name
suggests,
build
and
deliver
applications
in
a
cloud
native
way.
Interim
importantly,
we
focused
on
that
thermos
in
a
cloud
native
way.
While
a
lot
of
the
world
is
focusing
on
kubernetes,
it's
not
necessarily
required
that
the
world
was
eventually
will
burn
and
Benitez
and
there's
obviously
a
lot
of
other
people
who
helped
us
boots.
A
A
So
why
does
this
stick
even
exist?
So
why
are
we
even
around
here-
and
why
are
we
even
talking
about
this?
So
let's
get
a
bit
into
the
target
audience
for
the
app
delivery
state
during
that
whole
process
of
forming
the
cigarette
delivery,
we
identified
sweetheart
get
all
the
answers
that
we
talked
to
in
a
cloud
native
space.
The
first
one
is
really
people
who
want
to
get
started,
building
cloud
native
application
so
once
they
start
to
move
the
first
project
into
production
and
move
their
first
world
goes
into
production.
A
The
goal
was
to
help
them
to
get
a
quick
understanding
of
practices
tools
and
how
to
build,
manage
and
run
cloud
native
applications.
That
was
one
of
the
the
drivers
here.
So
if
you
want
to
get
started,
if
you
look
at
the
CN
CF
landscape,
for
example,
you
will
see
a
my
read
of
tools.
What
should
you
use?
How
do
these
tools
work
together?
What
should
you
care
about?
How
should
you
train
your
team?
A
What
are
best
practices
up
then
for
people
who
move
to
a
large-scale
cloud
native
environments,
the
goal
was,
and
especially
obviously
into
production.
Some
questions
suddenly
start
to
come
up,
so
everything
looks
simpler
when
you
deploy
your
first
service
and
kubernetes,
and
you
run
your
first
couple
of
deployments,
then
you
engage
in
a
service
mesh
and
then
you
start
to
really
build
an
application.
A
So
the
goal
is
to
really
form
the
ciggies
to
collect
these
questions
and
try
to
work
on
white
papers
and
best
practices
and
how
to
answer
them,
and
last
but
not
least,
there's
obviously
people
out
there
who
already
deployed
a
large-scale
workloads
on
kubernetes
already
and
they
want
to
share
their
experience
and
best
practices.
So,
in
the
initial
meetings
we
have
people
come
up
with.
Okay,
we
this
model
on
how
we
deliver
in
ship
applications,
which
we
want
to
share
with
the
broader
community,
gain
feedback
and
also
learn
how
to
improve
an
optimize
you.
A
So
these
are
the
sweet
buckets
which
brings
us
actually
to
a
first
question.
Our
first
two
questions
is
at
which
state
of
your
journey.
Are
you
that's
exactly
one
of
the
questions
in
the
forum
here,
so
where
would
you
put
yourself
just
getting
started
deploying
made
your
first
workloads
are
already
being
that's
data
that
you're
running
major
workloads
already
and
then
the
next
one?
What
are
you
looking
for
the
most
so
the
good
thing
for
the
end
user
communities?
A
You
can
actively
shape
the
direction
of
where
this
thing
is
going
actually
in
two
ways,
obviously
can
tell
the
sick
what
you
want
us
to
work
on
what
will
build
on
put
on
the
roadmap
and
obviously
you're
more
than
welcome
of
you
to
come
to
contribute
in
our
meetings,
and
we
will
talk
a
bit
more
about
this
as
we
go
so
link
work
for
everybody.
Wait
we
look
at
the
results
then
then
later
on
it.
A
So
what's
the
goal?
What
are
we
looking
at
for
the
app
delivery
six,
so
this
was
kind
of
a
discussion,
and
so
what
how?
How
do
we
define
the
scope?
So
when
does
application
delivery
really
start,
and
how
do
we
have
like
the
demarcation
lines
you
just
design
an
implementation?
So
what
we
eventually
came
up
with
this
is
really
everything
like
from
building
deploying
managing
and
operating
out
native
applications.
A
There
were
some
conversations
in
the
beginning.
Should
you
also
talk
how
you
move
an
existing
application
over
to
Cala
native?
Should
we
have
best
practices
in
how
to
architect
the
applications
in
the
first
place,
I
think
down
the
road.
This
could
be
topics
that
we're
working
on
and
technically
we
have
them
in
the
Charter
that
we
can
discuss
them,
but,
like
with
every
working
group,
there
always
has
to
be
a
critical
mass
within
the
working
group
to
get
these
things
started.
A
So
a
lot
of
things
are
they
also
influenced
by
the
people
who
participate
in
there,
but
right
now
the
system,
the
main
focus
for
us
to
help
people
doing
this.
So
we
assume
you
have
figured
out
how
to
transform
the
monolithic
application.
You
have
a
basic
understanding
and
how
to
do
things.
Obviously,
if
we're
terribly
wrong
with
our
assumptions,
just
let
us
know
and
we'll
figure
out
the
way
how
it
work.
A
So
the
idea
is
really
to
support
people
on
a
design
of
cell
native
applications
and
now
getting
a
bit
more
specific
and
what
it
means:
building
and
packaging
cal
native
applications.
It's
a
question.
Okay,
so
I
have
a
couple
of
container
images
here:
I
have
a
couple
of
piano
files
here:
I
have
some
stuff
in
a
secret
store.
I
have
some
things
lying
around
somewhere
else.
So
how
do
I
efficiently
package
this
up
and
ship
this
in
a
way?
A
If
everybody
says
well,
I
don't
actually
know
how
to
properly
structure
my
helm,
sure
that
might
be,
and
actually
about
to
come
off
the
at
the
livery
sake
to
come
up
with
something
like
this
yeah
how
to
deliver
and
configure
the
applications
for
development
to
productions
to
what
we
have
to
keep
in
mind
how
we
would
handle
this
and
eventually
also
operate
cloud
native
applications
in
production.
So
this
goes
all
the
way
down.
A
A
Which
again
brings
us
to
our
next
interactive
harder?
This
is
going
to
be
really
interactive,
so
well,
really
figuring
out
from
the
group.
What
are
your
main
responsibilities
and
you
will
see
in
the
questionnaire?
Where
would
you
see
yourself
to
my
previous
question?
Where
should
we
focus
on
so
what
are
people
mostly
responsible
for?
Are
you
more
and
that
is
sign
and
development
side
more
on
the
operations
side
or
more
on
the
delivery
side?
A
And
you
have
right
now
this
luxury,
you,
obviously
that
you
can
provide
feedback,
that
they
will
make
it
in
there
and
then
it's
also
half
that
other
question,
which
are
your
area:
do
you
find
I
guess
the
least
defined
one
for
you
or
just
the
one?
That's
the
most
unclear
earlier
for
you,
if
you
think
of
the
whole
process
from
again,
as
I
mentioned
before,
I
have
an
app
that
I
wanted
to
put
on
a
cloud
native
platform.
A
Obviously,
in
many
cases,
ku
vanitas
I'm
pretty
clear
about
certain
steps
but
I'm
pretty
clear
about
other
steps.
So
what
are
these
other
steps
that
you
are
not
so
familiar
with?
Okay,
great,
we
have
some
more
people
joining
it's
great
if
you're
joining
it.
Just
came
right
in
time
for
the
interactive
part
of
this.
Don't
worry
the
ho
test.
A
So,
following
up
what
is
it
would
that
be
working
on,
so
this
SIG
is
not
necessarily,
as
you
mentioned
before,
it's
not
a
project
for
saying
so
we're
not
really
building
software
in
the
first
place,
but
it's
more
about
giving
direction
and
helping
people
to
better
work
together.
First,
one
is
really
driving
this
collaboration
between
an
users.
So
if
you
have
developed
the
best
practice
or
if
you're
looking
at
best
practice
for
something,
this
thing
is
a
great
place
to
engage
as
I
mentioned.
A
If
you
along
your
journey,
come
into
a
situation
that
they
well
actually
would
need
more
help
around
this
or
we
have
developed
a
great
model
and
we
want
to
share
it
with
the
wider
community.
That's
what
we
want
to
do,
which
obviously
means,
if
there's
topics
that
you
want
to
have
an
active
discussion
on.
That's
also
super
helpful.
A
Then
developing
guidelines
help
end
users
patterns
and
practices.
This
came
up
a
lot
in
the
initial
discussion,
like
what's
the
best
practice
for
doing
or
the
good
practice
for
doing
X.
How
are
other
people
doing
action
really
having
a
place
where
you
can
easily
find
this
information?
So
a
lot
of
people,
especially
those
who
were
early
adopters
of
kubernetes,
have
already
figured
out
certain
things
or
there's
best
places
all
over
the
Internet
to
find
those
things.
A
I,
remember
a
keynote
on
Monday
when
Dan
was
asking
who's
here
for
the
first
year,
and
you
have
seen
how
many
people
are
here
for
the
first
year,
and
you
can
assume
that
they're
here
this
year,
because
they
want
to
learn
how
to
move
the
workers
into
production
and
I
liked
his
analogy
of
like
the
fourth
graders
and
the
second
graders
so
having
the
fourth
graders
share
the
best
practices
with
the
with
the
second
graders
here.
This
is
also
what's
who
here
and
especially
in
that
area,
where
it's
really
great,
to
use
again,
sharing
examples.
A
Project
guidance:
there
was
a
foster
collaboration
amongst
projects.
You
know
that
there's
a
lot
of
students
you
have
projects
today.
Many
of
them
have
like
touch
points,
but
projects
do
not
necessarily
need
to
work
together
that
much.
However,
the
goal,
especially
for
the
stick
is
to
get
projects
closer
to
work
together
like
if
you
think,
ok,
there
is
something
in
one
project
that
should
relate
to
something
else
or
I.
Think,
for
example,
just
to
come
up
with
an
idea.
A
The
whole
project
should
work
closer
together
with
the
open,
telemetry
approach
it
for
certain
reasons
or
other
project
to
collaborate
or
helm,
and
the
lincolni
should
work
closer
together
so
that
I
have
my
service
match
definition.
Those
are
two
my
actual
project
definition.
This
would
be
something
to
engage
with
and
where
we
can,
it
also
help
work
with
those
projects.
A
And,
last
but
not
least,
one
charter
is
always
so
provide
input
to
the
TOC
where
a
project
might
be
needed
in
a
certain
space
where
something
is
not
available
to
be
so,
but,
very
importantly,
for
us,
as
that
you
see
narcissistic
chairs.
We
obviously
look
approach
it
at
a
part
of
CN
CF,
but
I
don't
see
how
work
necessarily
being
limited
to
CN
CF
projects,
so
they
had
this
whole
idea
of
collaboration
should
really
go
outside
of
CN
CF.
Like
there's
a
lot
of
other
projects.
A
Who
will
talk
later
a
bit
about
the
application
definition
I,
don't
see
a
reason
why
they
shouldn't
is
participate
in
the
discussion
and
why
we
should
not
like
all
get
together
so
I'm
personally,
coming
from
a
background
in
standards,
development
and
usually
used
to
bring
people
together,
different
opinions
and
see
what
comes
out
of
it
and
I
see
the
same
thing
happening
here
as
well
so
going
forward.
We
will
also
invite
other
people
to
a
discussion
as
well
like
Harry's
wearing
his
open
application
bottle
t-shirt
today.
A
So
it's
not
a
CN
CF
brochure,
but
still
as
we
discuss
how
we
model
and
define
applications,
we
want
these
people
to
be
there
as
well.
So
there
really.
The
goal:
is
here
really
open,
sharing
again
next
question:
to
really
keep
this
interactive
here,
I'm
very
simple
one:
what's
the
biggest
challenge
right
now
so
obviously
related
to
application
definition?
A
If
it's
like
finding
a
place
to
have
dinner
tonight
without
getting
hitting
the
rain
that
doesn't
really
count,
it
does
count
like
in
general,
but
not
for
this
session,
and
what
was
the
biggest
learning
that
you
had
in
the
past
year.
So
really
take
the
time
to
you
to
provide
this
input
and
also
see
where,
rather
than
this
one.
A
So
what's
on
the
roadmap,
what
are
we
working
on
right
now
and
what
is
the
currently
equivalent
that
the
application
delivery
stick
is
working
on
the
first
one
when
we
were
tasks,
having
like
a
white
paper
on
what
cloud
native
applications
and
the
state
of
the
industry,
which
is
more
or
less
providing
a
starting
point
for
people?
So
I
want
to
build
that
cloud
native
application,
development
delivery?
What
does
this
actually
mean?
A
What
should
I
be
looking
at,
and
not
necessarily
starting
from
the
tool
question
in
the
first
step,
but
which
problems
do
they
need
to
solve,
which
things
to
a
need
to
address
and
also
which
options
are
there
so
not
to
provide
like
an
opinionated
approach
but
giving
people
a
guidance
of
what
they
could
be
using
instead
of
just
maybe
tapping
over
one
project
or
the
other,
which
eventually
didn't
help
us?
He
helps
people
to
move
into
that
space
faster,
just
think
about
it.
You
knew
and
you
want
to
solve
a
certain
problem.
A
So
what
do
you
do?
You're
using
Google
or
your
other
favorite
search
engine
and
look
for
projects
that
I
do
this,
and
then
you
run
into
a
long
process
of
trying
out
all
of
these
projects,
looking
at
them
finding
out
what
the
differences
are,
how
they
work
and
despair
a
lot
of
time
to
get
started
so,
especially
for
people
moving
in
the
space
and
leading
it,
as
we
will
learn
later
on
at
least
a
couple
of
tools.
This
is
a
very
time
consuming
and
sometimes
frustrating
process.
A
A
Very
often
nd
users,
hate
landscapes,
because
I
think
vendors
are
just
down
there
and
just
wanted
to
be
on
there
and
they're.
Not
super
helpful
I
think
it's
all
up
to
us
to
create
a
landscape.
It's
really
helpful
by
driving
individual
landscape
tools
and
solutions
that
we
have
on
their
by
their
capabilities,
the
firm.
So
for
me
this
is
a
great
or
we
just
don't
put
a
tool
somewhere,
because
somebody
claims
to
be
there
and
you
don't
want
to
upset
them.
A
No,
rather
have
a
clear
understanding
what
a
tool
in
a
certain
space
should
do
and
they
should
be
able
to
describe
how
they
do
it
and
provide
it
so
that
when
you
then
come
to
the
interactive
landscape,
you
can
say:
oh
this
tool.
Okay,
can
do
this
this
and
this,
but
it's
not
that
good
at
they
have
to
pick
something
else
too.
A
Next,
one
is
where
this
collection
of
patterns
currently
used
in
production.
That's
really
the
whole
topic
about
reinventing
the
wheel
and
run
a
build,
a
community
of
practice
and
really
key
ensuring
that
people
get
together.
There's
lots
of
end
users
of
kubernetes.
Today,
all
of
them
have
developed
their
own
best
best
practices
for
doing
things,
but
we
will
think
that
there's,
maybe
more
exchange
amongst
them
is
helpful
and
learning
from
each
other
rather
than
developing
practices
on
their
own.
A
Obviously,
we
say
with
a
grain
of
salt
and
rollout
and
delivery
models
for
applications.
So
we
all
talk
about
cloud
native
ways
of
delivering
like
cannery
releases,
Bluegreen
deployments
and
I.
Think
all
of
us
have
an
understanding
of
what
we
mean
by
them.
So
we
know
we're
like
a
Bluegreen
deployment.
A
Is
we
know
what
the
cannery
is,
but
what's
really
the
best
practice
to
do
it
or
if
you
look
at
different
tools,
their
understanding
of
how
I
can
really
works
might
be
significant
different
from
another
tool
or
how
you
really
like
do
blue
green
or
how
would
like
really
implement
this
in
the
platform
all
the
way
to
which
files?
And
how
do
you
define?
How
do
you
change
workflow
definitions
on
the
fly,
what
is
actually
domain-specific
to
a
tool
and
what
is
generally
available?
A
It
becomes
way
more
interesting
here
and
again,
defining
patterns
and
working
on
them,
maybe
eventually
than
having
a
description
that
works
even
across
tools
which
you
wouldn't
have
right
now
and
the
last
one
sounds
like
you
like:
a
really
big
goal:
defining
a
reference
architecture
for
cloud
native
application,
delivery
and
a
reference
architecture
always
sounds
kind
of
scary.
But
what's
behind?
This
is
really
a
way
where
we
can't
it
all
get
together
like
seeing
really
how
certain
tools
fit
together,
how
you
can
use
tools
and
use
those
tools
to
build
like
your
application
delivery
chain.
A
Ideally
by
having
a
whatever
it's
going
to
be
a
helm,
shirt
and
operator.
Whatever
you
can
click
and
choose
to
see,
it
is
a
way
how
it
could
work
and
where
eventually
might
be
able,
then
to
put
in
other
tools
in
there,
which
again
helps
obviously
also
the
standardization,
because
if
you
want
to
exchange
a
delivery
tool,
the
only
way
to
build
something
like
a
reference
architecture
only
works.
A
So
what
are
we
working
on?
Currently?
Obviously,
you
have
seen
now
the
big
goals
that
we
are
going
to
work
on
over
the
next
year,
but
what's
happening
today.
The
first
one
is
really
and
thanks
to
Harry
who
put
a
lot
of
effort
into
this.
Is
there
a
dictionary
of
Canal
native
application
in
livery?
This
is
this
white
paper,
where
we
really
discussed
core
concepts
around
application
delivery
and
like
defining
this
terminology
and
that
model
for
for
cal
native
application
delivery.
A
The
second
one,
that's
an
example:
where
did
you
see,
for
example,
reached
out
to
us
to
get
to
a
definition
of
what
an
operator
is
and
when
we
will
talk
about
like
what
operators?
You
might
think
this
is
pretty
straightforward,
it's
a
C
or
D,
and
it's
the
control
and
controller
with
a
reconciliation
loop.
A
Yes,
that's
true
and
from
an
implementation
point
of
make
sense,
but
think
about
again,
our
target
audience
and
biologists
moving
to
kubernetes
building
their
first
application
and
wondering
should
I
write
it
and
an
operator
for
my
application
or
shouldn't
I.
Should
I
write
one
for
each
individual
micro
service
and
how
should
a
structure
this
and
what
should
go
into
the
CID
for
the
definition?
What
should
be
part
of
the
the
operator
and
which
framework
should
I
be
using?
A
Is
there
a
difference
using
sacred
orally
operate,
a
framework
to
the
operator
SDK
and
which
differences
are
there,
and
why
should
they
even
care
about
them
again?
There's
a
link
in
the
slides
and
obviously
all
these
slides
are
available
to
you
of
an
ongoing
discussion
and
really
coming
up
with
a
definition,
the
last
one
that
are
they
the
current
most
current
one
out.
A
There
is
still
the
one
from
I
think,
two
or
three
years
ago,
from
the
original
definition
and
right
now
the
goal
is:
we
chose
a
pre-image,
an
end
users
day,
like
which
problem
should
you
have
as
an
end
user
with
an
operator's
resolution,
and
how
does
it
operate
the
softest
problem
for
you
not
just
talking
about
the
very
tech
definition,
what
it
means
in
kubernetes,
because
last
but
not
least,
the
goal
was
really
to
talk
about
cloud
native
applications.
So
what
if
we
told
you
about
a
purely
server?
Less
application,
it
might.
A
On
company,
so
technically,
I
can
also
implement
an
operator
with
a
file
on
s3
and
a
lambda
function
if
I
understand
the
core
concept,
what
I'm
actually
wanting
to
build
it.
So
these
are
the
two
things
that
people
right
now
can
mostly
get
engage
list
and
I
now
want
to
let
Harry
data
the
deeper
into
the
application.
B
Delivery
model
yeah,
thank
you,
so
I
always
had
to
talk
about
what
we
are
doing
right
now
in
the
community,
but
there's
something
that
we
want
to
highlight
here,
which
is
the
application
delivery
model,
because
this
is
actually
a
mechanism.
We
are
trying
to
defy
that
help.
Community
users,
every
participators,
know
what
a
project
really
about
so
I
think
you
guys
might
already
notice
that
the
application
delivery
or
application
management
things
Giggsy
clonaid
community
is
really
really
messed
up.
There
are
a
lot
of
project.
There
are
a
lot
of
tools.
B
There
are
so
many
concepts,
but
don't
know
what
do
they
are.
So
this
is
the
model
which
we
are
actually
trying
to
split
them
into
different
topics
and
each
topics
they
actually
a
common
practices
and
a
common
concepts
about
what
a
project
or
what
a
component
can
do.
Example:
helm
right,
you
know
hell,
so
what
is
helm
a
really
doing?
According
to
this
model,
we
can
really.
You
can
very
easily
to
fight
where
helmet
actually
being
example,
the
top
Huong
it's
about
application
definition.
We
know
that
in
order
to
rather
trust
fell
into
defy.
B
What
is
this?
What
is
the
application?
What
is
the
artifacts
you
want
to
run?
This
is
exactly
how
many
trying
to
help
you
to
make
it
clear.
So,
hey
I
may
do
another
job
in
the
application
technician,
especially
for
the
application.
Descriptor
part,
but
helm
also
has
a
lot
of
other
functionalities.
For
example,
you
know
him
has
some
concept
of
helm
Harbor,
so
you
can
search
or
browse
the
applications
in
the
heart
right.
So
these
part
of
functionality,
they're,
actually
belonging
to
topic
1.5,
is
about
application.
B
Packaging
helm
also
give
you
these
cut
off
functionality
to
help
you
to
package
your
applications,
so
you
can
upload
the
application
to
sound
layer
and
then
other
people
can
browse
search,
application
and
download
the
application
it's
about
application,
distribution
and
packaging
and
also
has
functionality,
which
name
is
templating
right.
You
can
just
change
the
charts
demo
files
by
using
some
kind
of
templating
and
values,
and
you
can
just
deploy
these
film
cars
2
different
Academy
environment,
to
be
different
parameters.
This
is
actually
about
application,
parameter
e
and
a
configuration.
B
This
is
also
where
he'll
actually
overlap
a
little
bit
customized.
So
a
lot
of
people
ask
the
same
question:
okay
is
customize
the
application
definition
project.
I
think
the
answer
is
not
because
I'm
a
blur
because
customized
do
not
care
about
what
ending
scenario
fell.
It
will
not
help
you
to
defy
what
either
your
application
is,
not
application
definition.
B
So
we
can
see
a
difference
here,
so
this
model
trying
to
help
you
to
evaluate
what
is
exactly
a
project
doing
and
also
help
you
to
compare
a
project
visit
another
because
now
you
have
clearly
concept
of
definition,
but
the
responsibilities
of
different
project.
It
also
happens
to
topic,
2
application,
Road
and
topic
3
workload.
Instances
I
think
you
guys
might
heard
about
a
lot
of
project
machines,
for
example
the
tectum
spinnaker.
We
works
blogger.
B
They
are
all
application
world
project,
but
what's
the
difference
and
what
they
care
about,
so
all
of
these
project
they
actually
care
about
the
application
level
management,
for
example.
How
to
do
blue
green,
keep
women
how
to
do
Claro
deployment,
this
application
comes
them
and
don't
care
about
the
pods
or
containers
spinnaker
do
not
care
about
right.
B
The
only
project
or
components
who
cares
about
container
or
part
is
actually
topic.
3
project
example
spinnaker
spinnaker
route,
your
application.
They
given
strategy
are
leverage
the
deployment
controller
in
kubernetes.
The
deployment
controller
is
not
part
of
spinnaker.
It's
part
of
properties.
It's
actually
a
topic.
Three
component,
you
can
see
difference
here.
Operator
is
also
another
kind
of
or
another
type
of
public
free
project.
So
that
means
maybe
someday.
You
can
use
spinnaker
to
manage
your
operators.
It's
common
things.
B
You
should
be
able
to
do
that,
but
I
Ryan
always
can't
do
that,
because
there's
no
clear
definition:
ok,
what
a
topic
to
responsibilities
workable
and
what
it
public
three
responsibilities,
and
also,
of
course,
we
have
topic
for
projects
like
kubernetes,
which
give
you
the
fundamental
capabilities
like
container
management,
resource
management,
scheduling
the
the
the
extension
ball.
If
you
guys,
like
CRT
the
traffic
management
capability
like
East
Hill,
all
of
them
belong
to
top
four,
but
that
also
remind
us
that,
maybe
in
the
future
or
maybe
sometime,
you
will
have
other
cattle
platform
example.
B
You
are
using
function
you're
using
fast,
your
even
function
to
the
file
application.
They
also
belonging
to
same
for
native
application
to
be
processed
so
cloud
net
done
that
that
mean
you
have
to
use
ribbon.
It's
not
like
that
so
that
that's
why
we
want
to
clearly
define
what
are
these
concept
in
this
workflow
of
application
delivery?
So
we
hope
that
will
help
you
to
figure
out
okay,
what
this
project
is
really
about,
so
we
will
talk
more
about
this
model
in
detail
in
tomorrow's
deep
dive
station
of
cig
application
delivery.
A
Interestingly,
maybe
to
add
up
here
topic
number
four
is
actually
not
part
of
what
we
want
to
work
on
on
the
application
delivery
sig
per
se,
because
that's
what
our
city
company
Tostig
is
working
on
and
it
was
a
should
ideally
also
be
the
concept
that
six
that
that's
exchanged
over
here.
So,
interestingly,
we
would,
if
would
switch
an.
A
Cloud
native
platform,
you
should
be
able
things
that
you
have
to
find
on
the
higher
layers.
Ideally,
when
can
you
do
this,
the
less
they
are
tool
and
project
agnostic,
you're,
mostly
I'm,
more
likely
you
can
do
it
this
way.
There
should
be
no
reason
why
I
use
one
container
definition
then
put
it
somewhere
else.
There
was
a
question.
We
have
a
microphone
here,
so
you
come
down.
A
A
B
A
A
If
you
have
a
disconnect
that
repository
network
yeah,
okay,
you're
kept
environments
like
world,
you
can
kept
like
store
your
repositories
and
so
forth,
so
I
think
TTIP.
The
quick
answer
to
your
question
is
obviously
having
tooling
in
there
for
repository
management.
Other
kind
of
things
would
fit
into
the
1.5
application,
packaging
yeah
area.
So
how
you?
Because
that's
exactly
the
point,
how
you
package
something
up:
yeah.
B
B
A
That's
good
any
other
question.
So
let's
stop
here
for
a
bit.
If
there's
questions
feel
free
to
ask
them,
you
don't
have
to
okay.
So
now,
let's
look
at
the
intermediate
results,
so
this
might
actually
also
drive
some
of
their
discussions
here
and
now
we
can
15
answers.
That's
okay,
a
good
starting
point.
So
let's
have
a
look
here
and
I.
Think
that's
I'm
gonna
make
this
a
bit
bigger
here
so
from
the
people
in
the
room
here.
I
think
that's
actually
not
that
bad.
A
So
what
you
can
see
here
that
we
have
over
a
lot
of
people
who
run
large-scale
workloads
on
kubernetes
so
like
in
the
spirit.
What
then
was
talking
about
in
his
keynote?
Maybe
you
raise
your
hand
because
then
the
other
people
can
all
come
to
you
to
ask
you
all
their
upper
questions
right
now,
but
the
rest
is
really
split
between
just
getting
started
and
scaling
their
deployment.
So
we
see
we
have
like
quite
a
mix
in
here.
A
A
It
would
be
great
to
have
documentation
and
visual
aids
to
help
teams
grow
in
the
right
direction.
Around
best
practices,
I
would
like
to
take
a
more
active
role.
In
the
sake
we
will
come
to
this
one.
Yes,
you
actually
can
so.
This
is
super
anonymous
fall
in
GDP,
our
compliance,
whoever
will
prove
yes,
we
will
come
to
this
employing
apps
to
air
kept
environments.
Think
it
was
request.
The
gentleman
head
over
there
image
tagging
versioning
rolling
upgrades
actually
a
great
topic.
How.
A
Upgrades
and
performance
we're
in
the
deployment
area
haven't
built,
see
ICD
pipeline
me
yet
I
think
the
question
is:
how
do
you
best
build
the
ICD
pipelines?
Yes,
packaging,
an
image,
relocation,
great
one,
yeah
customized,
okay,
does
thing
just
how
customized
fits
into
this
I
would
like
to
learn
about
best
practices
around
managing
yeah,
best
practice,
question
large-scale.
B
A
clusters
so
actually
spawn
I
think
multipolar,
Jessica
yeah.
We
actually
will
talk
about
a
little
bit
about
what,
in
the
relationship
between
application
delivery
base,
a
multi,
cluster
or
multi
cloud.
It's
actually
very
choose
interesting
topic.
We
will
cover
that
topic
in
quality
taxation,
I
believe.
A
What
are
your
main
responsibilities?
That's
an
interesting
one
as
well,
so
we
see
that
obviously
build
packaging
and
configuration
management.
Is
the
majority
in
here
and
I?
Think
it's
cuz.
It
goes
well
in
line
with
what
our
selection
was
at
the
beginning,
that
we
want
to
focus
more
in
the
lower
parts
and
lesson
development
and
design
ones,
which
is
me
we
don't
do
it
at
all.
B
A
I
mean
that's,
we
love
to
help
you.
Maybe
you
just
rethink
your
answer
here.
Let's
start
with
one
air
captain
I
think
we
say
air
is
the
air
gap.
Autocorrect
is
always
a
disconnected
environment.
I
think
is
equal
one
man
teaching
application
the
leverage
with
focused
stateful
services,
I
think
scale
stateful
services,
it's
a
good
one
specifically
focusing
on
managing
stateful
apps
need
more
than
helm
air
gap,
again
security.
How
to
keep
things
up
today?
A
How
do
I
really
identify
the
Sun
scale,
converting
apps
to
cloud
native?
Yes,
this
was
a
discussion
that
we
had
before
and
as
the
SSE
we
were
kind
of
asking.
What
exactly
what
exactly
we
should
be
working
on.
So
ever
who
wrote
this?
It
would
be
good
to
get
maybe
a
bit
more
detail
that
year
on
maybe
by
the
mailing
list.
What
kind
of
help
do
you
need?
Is
it
really?
We
like
we
architecting
from
monoliths
to
micro
services?
Is
it
best
practice?
A
It's
like
12
fact
the
replications,
because
this
is
a
very
wide
topic
that
that
has
to
be
in
there.
So,
the
more
maybe
you
can
be
a
bit
more
specific.
There
yeah
the
next
one
of
us
is
design.
Transform
model
is
2008
it
good
one
continuous
deliveries.
We
exclamation
mark
I
think
somebody's
very
passionate
about
this
yeah
good
point,
yeah
transferring
applications
to
air
gap.
Again
the
air
gap
came
up
like
this
connect.
Environments,
I
think,
is.
A
I
think
it
really
depends
on
what
do
you
require
from
your
end
users
as
a
minimum
education
definition,
and
what
can
you
hide
from
them
at
because,
as
of
today,
we
would
have
to
ship
everything.
But
what
do
you
have
to
read
is
the
minimum
requirement
here
minute
large-scale
deployments,
defining
an
EPS
metadata
named
version
convict.
There
is
actually
something
today
improving.
It
is
where
you
can
do
some
of
that
stuff,
at
least
for
services.
It's
a
new
communities,
documentation,
but
Brian
pointed
us
to
this
one.
A
A
B
A
So
many
people
have
the
same
problem.
Yes,
that's
why
we
all
get
together
here,
that's
cloud
native
anonymous!
That's
we
are
it's
all
on
video
right
now,
so
I
might
get
into
probably
how
well
developed.
Are
your
practices
I
think
this
is
now
an
interesting
one
where
people
see
themselves
and
that
might
be
interesting
to
track
really
over
time.
For
us,
this
is
really
for
application.
Definition
I
think
it's
a
wide
way
right.
The
people
seem
to
be
quite
happy
with
this
one.
A
However,
a
lot
people
with
roll-up
processors
I
think
that's
also
an
interesting
answer.
One
hand
there
were
many
requests
around
delivery.
At
the
same
time,
people
say
we
are
really
good
with
our
deployment
and
delivery
processes.
So
I
think
it
makes
up
an
interesting
discussion
as
a
follow
up
in
the
group
and
Brookfield
Automation
thing
quite
distributed
here.
A
Ok
yeah
how
to
engage
with
us.
That
was
one
question
from
somebody
you
want
to
work
or
with
us.
That's
great.
We
want
to
work
more
with
you
too,
and
how
can
you
engage
a
couple
of
things?
There's
a
github
repository
in
the
CNC
F,
which
is
for
every
livery.
If
you
want
to
present
some
of
the
work
that
you
have
done,
obviously
people
are
submitting
projects,
that's
what
they
also
do
a
lot
or
if
you
want
to
have
certain
discussions,
just
file
an
issue,
so
we
usually
track
issues
there.
A
Even
if
you
say,
hey,
I
have
like
this
specific
problem.
I
want
to
discuss
as
father
the
issue
there,
so
we
can
track
them.
In
a
cloud
native
you
seen
CF
slack.
There
is
a
dedicated
channel
too
for
the
C
gap.
Delivery,
that's
the
mailing
list,
obviously,
and
we
have
two
meetings
every
second
and
fourth
Wednesday
at
8
p.m.
Pacific
time.
So
all
of
you
should
all
of
you
do
have
access
to
the
CN
CF
calendar,
the
official
one,
the
Apple
everything
is
in
there
yeah.
A
They
also
deliberately
move
this
off
a
lot
of
the
other
meeting,
so
it's
offset
a
lot
of
other
meetings
so
that
people
can
actively
participate
in
those
meetings.
The
first
way
actually
to
engage
more
with
us
and
more
deeper
on
the
delivery
and
state
of
so
where
Harry
has
done
some
work
on
how
the
relating
currently
existing
projects
to
the
a
delivery
framework
and
getting
started.
Aaron
morphic
workshop
that
discussion.
That
will
be
the
first
working
session
tomorrow,
yeah.
B
Yeah
also,
some
people
ask:
what's
the
difference
between
the
flag
and
the
Middle
East,
so
the
personal
feelings
from
my
Saturday,
that's
likely.
If
we
want
to
ping
someone,
you
want
to
buy
me
one
final.
Always
you
want
everybody
here,
the
ping
someone,
but
if
you
want
to
you,
know
scarce
official
between
some
proposal,
you
want
to
you
know
fighter,
be
someone
go
to
the
Middle
East
and.
A
Also,
what
we've
just
started
to
do
is
we
had
like
these
discussions
on
the
mailing
list
in
order
to
have
a
bit
of
better
chance
to
track
them.
We
also
started
to
create
issues
about
very
specific
topics
that
we
wanted
to
talk
and
discuss,
because
even
after
the
mailing
list
at
some
point,
they
simply
can
get
lost
and
it's
hard
to
find
agreement
there
right
yeah
to
that
point.
If
you
really
want
to
reach
out
and
ask
questions
best,
you
stack
mailing
lists.
They
usually
follow
up
discussions.