►
From YouTube: Cloud Foundry Community Advisory Board [October 2017]
Description
Video from the monthly Cloud Foundry Community Advisory Board (CAB) meeting. Learn more at https://www.cloudfoundry.org/.
A
I
know
I,
don't
still
don't
have
a
use
case.
We
could
be
a
person
like
you
anyways,
especially
if
the
networking
still
does
right
and
quite
yeah.
I
am
quite
know.
If
right
now
can
I
build
a
service
instance
or
a
service
broker
that
plays
things
too
Kubo
and
doesn't?
Is
there
a
way
to
expose
that
to
its
routing
thing
that
the
apps
can
talk
to
it?
Is
that
solved
or
a
still
left,
with
a
routing
problem
that
I
have
every
time
after
your
service
broker
service
gives
that
slight
thing?
A
B
Word
three
months
would
look
even
at
ease
now.
You've
got
like
an
opinionated
way
if
you've
decided
that
you
can't
just
run
your
data
service,
like
everyone
has
been
moved
20
years
of
the
recently
by
getting
a
really
large,
bare
metal
box
and
running
vertically
scale
database
on
it
with
a
backup
you
can
now
use
kubernetes.
Do
that
for
some
reason
that
makes
you
happy
well
I'm
excited.
B
Disagree,
I
think
we'll
mostly
say
I'm
wrong
about
what.
B
A
B
B
B
A
Recover
it
isn't
the
idea
of
running
one
Postgres
akin
in
their
so
saying
there
we
shouldn't
use
cloud
package,
just
not
touch
that
box
ever
again,
I
mean
all
or
something
like
we
went
down
the
path
of
the
world
native
adapts
and
we
learnt
and
not
write
to
the
file
system
are
gonna,
be
ephemeral
right.
We
build
up
yes
strengths
around
the
best
end.
We
know
how
to
write
yes
so,
but
we
don't
know
how
run
long
lived,
how
available
donor
babies.
Look
we
refused.
B
To
talk
about
it,
I
think
I
think
it
turns
out
we
sort
of
do,
but
the
point
is
not
everyone
needs
to.
This
is
what
Google
found
with
MapReduce
right.
They
didn't
want
everyone
reinventing
how
to
process
large
amounts
of
data
right,
but
everyone
in
the
company
needed
to
process
large
loads
of
data.
So
what
did
they
do?
They
had
the
people
who
understand
how
to
build
scalable,
clustered
stuff
for
a
build
that
and
then
everyone
else
just
writes,
map
jobs
and
reduce
jobs,
and
now
has
it
run,
wins.
A
Where
does
that
make
or
break
down
in
the
cloud
boundary
world
as
an
app
developer?
I?
Don't
care
about
post
graves
yeah
that
doesn't
mean
that
I
want
my
post,
Chris
I,
don't
I,
don't
care
fundamentally,
I,
don't
care
whether
my
Postgres
database
design,
shared
machine
or
not
right
until
there
is
a
thing
which
I.
B
Do
care,
let
me
put
it
this
way:
RDS
exists
right
over
time.
I
think
the
direction
of
travel
is
for
things
that
look
like
IDs
right
now.
There
were
some
reasons
not
to
use
things
like
that,
because
they're
still
expensive,
but
they're
going
to
get
less
and
less
expensive
as
we
get
better
and
better
at
running
them.
I
can't
imagine
why
you'd
want
to
run
your
own
on.
Yes,
there
is
such
a
hilariously
solved
problem
having.
B
That
that
to
me
the
platform
provides
and
the
platform.
Maybe
someone
in
your
company
there
may
be
a
team
who
is
providing
your
highly
available
eight
days,
but
the
way
we
generally
run
those
things
like
Liam
fans,
peace,
you
get
from
containerization
aren't
really
advantages
that
help
you
much
with
this
databases.
Don't
get
modified,
often
you're
doing
rapid
development.
A
Assessment,
yes,
my
entire
story:
I
I
appreciate
we
sang
so
I'd
like
a
fight.
My
story
of
this:
it's
not
necessary
a
story
of
the
app
development
stuff
to
care
more
about
this
people
sitting
at
Katanga.
So
if
the
app
developers
still
doesn't
care,
which
is
why
I
could
confuse
modules,
fascination
communities
and
same
trainings
are
interesting.
Yes,
all
things
are
not
interest
exactly
right,
I
mean
so
just
verb.
A
Anyone
watching
historically
for
the
last
two
years,
I
was
doing
a
thing
called
Ingo
PostgreSQL,
which
was
like
on-demand
provision,
two
containers
of
Postgres
inside
those
containers.
It
did
the
whole
lifecycle
of
forming
a
cluster.
A
TED
CD
for
coordination
of
leadership
had
a
thing
called
Patroni,
which
did
the
the
state
machine
of
what
role
to
play.
I
must.
A
B
Exactly,
but
perhaps
something
does
what
I
would
argue
is
that
what's
happened
this
week
and
I
think
it's
the
right
thing
if
I'm
honest
is
that
we've
made
a
decision
that
CF
isn't
what's
going
to
solve
that
problem
by
the
problem
of
running
that
cluster
post
race
we
used
to
be
using.
Is
it
our
CF
include
like
different
things
that
pertain
to
Orkut
rate
since
their
attractions
they're?
Not
there
is
container
n
times
and
coud
mainly
not
solve
your
problem
and,
frankly,
I
think
most
problems
are
better
sold
by
not
doing
that
stuff.
B
In
that
stateful
environment
CF
is
going
to
focus
on
we're
gonna
make
it
really
easy
push
your
app
and
bind
to
someone
else,
a
solution
back
right
so
see
it's
really
important
thing
about.
That
is
that
the
number
of
people
who
are
pushing
apps
is
large
and
they're
getting
pushed
repeatedly
and
waiting
repeatedly.
You
do
rapid
development
on
them
that
you
really
want
to
speed
up,
and
there
is
value
me
out
to
a
really
fast
they're,
getting
new
vertical
apps
out.
B
So,
on
the
other
hand,
you
can
solve
the
problem
of
being
able
to
deploy
a
database
for
someone
or
frankly,
just
get
highly
scalable
database
and
pull
out
an
account
once
in
one
team
and
it's
solved.
I
don't
need
that
to
be
made
substantially
faster
in
the
same
way,
I
need
app
developer
to
be
a
subset
of
the
cluster.
B
Provide
so
that
was
another
thing,
I
was
gonna,
say
I
do
think
we
are
now
starting
to
get
now
that
we
realize
that
there's
two
separate
things:
there's
kind
of
foundry
which
isn't
going
to
try
and
solve
the
problem
running
those
things
there's
actually
the
problem
of
kubernetes
is
trying
to
solve
too
many
problems.
I
think
clarify
to
me
is
going
to
be
a
focus
thing.
There's
also
called
more
famous
apps
now
that
you've
made
that
decision,
I
would
say
includes
upside
that
wears
off
the
site.
B
Now
we
can
try
and
solve
this,
and
what
I
think
is
the
real
way,
which
is
now
have
this
end
of
Orion
s.
Ischia
work
that
we're
talking
about
this
contains
from
saying
that
working
all
that
stuff.
It's
not
about
running
those
things,
look
forward.
It's
about
actually
saying
these
things
amongst
going
on
outside
platform.
How
can
you
make
it
really?
Well,
okay,.
A
So
anyone
that's
not
here
to
miss
the
whole.
The
other
exciting
thing
to
me
is
probably
not
keeping
any
so
much
as
their
networking
sites
SEO
in
boy
stuff.
So
what's
exciting
about
that
is,
as
I
understand,
Google
themselves
put
local
proxies
next
to
apps
to
say,
and
that's
your
app
talks
or
local
proxy
as
a
representation.
Sidecar
is
the
only
word
2017
for
this
proxy.
You
talk
to
the
sidecar
and
the
sidecar
is
entirely
responsible
for
that.
A
The
security,
the
load,
balancing
the
relationship
that
failover
reached
on
your
app
can
be
stupid
about
it
and
that's
why
it's
big
fan
of
console
when
it
came
out
because
it's
like
the
only
thing
hey
now,
you
don't
talk
to
the
constant
servers,
talk
to
console
agent,
we'll
deal
with
that.
So
I
love
that
pet,
the
psyche
I've
had
so
that
was
exciting
on
day
one
when
they
first
saw
that
their
presentation
right
and
then
and
then
you're
watching
out,
I'm
going.
Oh,
my
god,
you
are
never
gonna
say
it.
A
Are
you
you're
not
going
to
say
services?
You
only
used
the
word
services
was
to
talk
about
other
apps
and
that
whole
talk.
It
was
all
about
apps
talking
to
other
apps.
They
could
not
get
out
of
their
head.
Cloud
Foundry
has
a
responsibility
to
talk
to
external
things
and
for
external
things
to
manage
their
relationship.
B
D
Mostly
a
to
listen
to
understand,
especially
if
you're,
for
example,
of
something
that
I'm
not
familiar
with
yet
more
about
it
at
this
conference.
Oh
that's,
interesting!
I,
like
a
concept,
that's,
let's
see,
let's
see
what
comes
from
that
I've
had
more
work
to
run
it
is
there
some
more
opinions,
but
also
there
I
think
it's
really.
We
have
to
listen
to
what
people
actually
do.
C
C
D
C
D
D
C
D
Right
so
at
runtime,
then
you
don't
use
push
anymore.
So
for
some
of
you
as
infrastructure
and
that's
the
experience
that
has
built
up
the
knowledge
to
operate.
There
are
things
with
Bosch
them.
It's
probably
the
right
choice.
So
if
you
don't
do
that,
but
you're
coming
from
a
community's
perspective,
then
you
might
not
want
to.
A
D
A
A
Run
clash.
Why
concerned
you
guys
is
that
you've
such
a
long
history
by
the
long,
little
limb
staff,
and
you
start
of
wanting
to
do
things
the
way
of
doing
them,
that
with
is
ever
the
chance
that
you'll
actually
go
back
to
doing
it
with
Bosch
I
know,
Bosch,
wasn't
very
good
five
years
ago,
when
you
decided
not
to
use
by
stuff
like
them
understand
and.
A
A
D
But
I
would
argue
that
all
these
technical
differences-
that's
probably
something
where
you
can
find
optimizations
and
solutions-
I
mean
you
know,
I,
think
it's
reading
from
I
don't
adhere.
That's
more.
My
personal
view.
Here's
what
kind
of
ecosystem!
What
are
you
familiar?
Music?
Did
you
do
after
that?
But
where
do
you
go
because.
B
Special
ed
to
me
is
how
you
deploy
your
cube,
Annette
right,
because
I'm
really
interested
now
that
there's
this
way
of
the
point
connected
with
Bosch,
that's
that
seems
like
it
gives
you
its
opportunity
to
manage
everything,
and
with
with
that,
you
have
to
deploy
your
cube.
Somehow
right,
you've
got
your
coop
in
the
first
place,
whether
that's
with
chef
or
hands
or
whatever.
B
But
that
to
me
is
the
question
of
is
Bosch
actually
a
potentially
use
case
that,
because,
as
soon
as
your
the
pull
of
cube
with
Bosch,
then
I
think
you
should
justify
yourself
watching
now.
You
really
do
have
just
one
thing:
to
learn
right
right.
The
question
is:
is
it
practical
for
you
to
before
the
coop
with
Bosch
and
if
not,
why
not?
What's
that
like?
What
does
that
look
like
in
your
organization
or
in.
D
B
A
C
A
C
A
People
desperately
wanted
to
see
if
deployment
see
if
release
and
fall
footprint
massively
smaller,
because
it
can
shop.
It's
a
defensive,
not
the
right
word,
but
there's
scratching
a
that's
such
a
long
live
pine
of
mine.
Then
it's
not
fair
to
say.
Oh,
we
have
two
days.
That's
so
problems
like
this.
You
know
we.
B
B
B
If
you
want
great
density,
tell
watched
for
more
the
focuses
on
the
same
BM
you're
now
done.
We
already
can
containerize
those
processes
you
have.
This
little
song
could
run
see
the
rapper
called
BPM,
which
most
account
engine
components
now
using,
which
means
you
can
put
all
of
the
things
on
one
cell.
If
you
like,
and
button
or
coach
make
them.
A
C
Think
I
think
you
guys
keep
missing
this
point.
You
know,
and
you
especially
do
is
that
when
I
mentioned
higher
density,
I'm
not
talking
just
about
not
foundry
I'm,
an
organization
where
I'm
learning
services-
and
let's
say
the
table
here-
is
my
sort
of
footprint
for
the
pow.
Maybe
carpentry
is
just
this
piece.
The
rest
of
the
table
is
services
for
there's
people
running
just
Kubb
and
so
on
and
having
a
unified
way
of
dealing
with.
This
is
a
great.
B
B
B
C
Lightweight
you
keep
on
saying
you
know:
I
have
my
container
clusters
that
I'm
deferring
to
cloud
to
coop
and
maybe
I'm
using
Bosh
to
do
that,
and
then
part
of
that
is
running.
Some
Cloud
Foundry
workloads,
although
that
is
running
I,
don't
know
like
my
database
is
part
of
that.
It
was
in
my
service
and
so
on
and
I
have
a
unified
way
of
deploying
all
these
things.
Beautiful.
B
C
C
Which
is
essentially
having
cloud
foundry
on
top
of
qu.
The
part
that
I
cannot,
you
know,
did
disappointed,
maybe
not
disappointed,
but
I
wish
could
happen.
Is
that
I
don't
have
to
change
the
way
I
manage
my
company,
because
that's
what
I
lose
and
second
and
I
agree?
That's
a
big
problem
but
I'm
wondering
if
there's
a
way
to
solve
it
right.
Is
there
a
way
to
you
do
that
deployment,
and
then
you
can
still
use
Bosch,
maybe
to
some
CLI
extension,
to
keep
managing
the
founder
instance.
To
do
that.
B
Something's
best,
but
it's
herb
less.
Something
is
the
best
a
lot
of
like
retainers
and
things
by
Bosch,
and
the
points
about
this
is
I
need
to
be
able
to
develop
a
distributed
system
and
use
the
primitives
that
make
sense
me
to
develop
like
this
reputed
system
right
and
the
people
who
decide
that
on
the
developers
of
that
system,
right
if
I
develop
something
that
uses
the
kubernetes
abstractions
to
keep
myself
highly
available
or
clustered
I.
Need
you
to
give
me
a
cube.
B
Let's
discuss
that
so
that
if
I,
if
I
design
it
with
surplus
I
need
a
service,
that's
back
the
car
foundry
developers
currently
have
designed
it
based
around
wash
because
Bosch
makes
their
life
simple,
keeping
a
highly
available
system
when
you
with
much
about
there
are
restrictions
in
barge.
There
are
things
that
we
have
to
design
for
Bosch
and
then
Bosh
gives
us
a
lot
for
free
right
and
that's
why
CF
uses
Bosch.
Now
you
can
kind
of
try
and
put
kubernetes
in
there
underneath
I
think
as
long
as
the
using
any
of
its
features.
A
B
We
used
to
see
a
push
to
production,
all
the
sorry
we
used
to
Bosch
complete
virtually
all
the
time.
So
when
the
states
for
the
Memphis
is
because
we
were
responsible
to
Google
model,
you
get
Rs
re.
Eventually,
we've
got
fed
up.
Eventually,
initially
we
were
responsible
to
get
to
production
and
Bosch.
Let
has
not
think
about
that
too
much.
They
solved
that
problem
for
us,
so
we
use
them.
Maybe
now
we
would
use
a
better
one
and
maybe
in
six
months.
B
Until
we
decide
who,
until
you
decide
to
use
it,
then
it's
going
to
be
limited
advantage.
That
has
to
be
a
decision
that
we
make
as
developers
in
the
community
and
then
to
use
it,
which
we
could
do
we'd
like
we
change
it
all
the
time
we
don't.
Even
we
throw
a
con
limit
which
I,
wrote
and
maybe
garden
run,
say
and
I'm
glad
we
did
it
and
was
probably
lots
more
stars,
but
it
has
to
be
fun
within
the
community.
C
B
A
C
C
A
C
A
C
B
I
think
I
think
I
think
this
is
the
mediumship
thing
about
its
consider:
an
application
one
right.
There
was
this
confusion,
the
what
we
should
be
doing
in
that
patient
one
time
it's
dust.
This
is
container
stuff
and
it's
more
complex,
Lance
I
actually
think
the
more
we
can
like
to
forget
about
this
I'm
saying
that's
cumulates
is
they
will
stand
up?
B
B
The
question:
what
would
make
me
happy
as
a
developer
and
and
one
thing
I'll
say
about
is
one
thing
that
I
think
is
is
interesting
about
this
stuff
is
what
what
what
we
have
to
be
careful
about
this
music
aunty's
around
and
they
can
be
really
good.
But
if
you
just
add
things
away,
you
tend
to
just
add
complexity.
So
if
you
took
bunch
and
replace
the
clean
yes,
okay,
now
we
could
have
a
discussion.
B
C
B
Which
is
why
I
would
say,
no
I
think
is
the
orchestrator
one
time
versus
the
application
was
like.
Are
you
gonna
care
about
orchestration?
It's
lovely
that
you
can
push
code
and
we'll
just
run
it.
But
to
be
honest,
there
are
many
use
cases
we're
just
crazy,
an
image
and
pushing
it
is
great
I,
don't
think
you're
getting
to
all
the
extra
stuff
you
can
do
to
orchestrate
makes
for
payment
and
customize
them
for
being
a
ship.
B
Your
code
to
me
in
a
pre-compiled
way
it
seems
great,
and
one
thing
we've
been
working
on-
is
to
use
layers
whose
ozone
is
more
in
the
platform.
So
what
I'd
love
to
get
to
is
for
one
day
build
packs
are
just
another
way
of
building
an
image
and
when
you
stage
with
the
build
pack,
you
just
get
a
URL
and
it's
a
docker
hub,
silo
URL
like
an
OCR
URL,
and
it
stays
patched
over
time.
B
So
I
can
use
co-founders
to
stage
that
and
then,
because
we
use
build
that
model,
it
will
keep
being
a
patched
image
over
time.
I
can
take
that
push
it
to
coop.
So
when
you
want
their
development
efficiency
and
rapid
application
development
of
class
fountain,
you
can
either
use
build.
Tackle
you
stuff
from
FEMA
will
do
that
then,
as
soon
as
you
want
to
call
the
escape
action,
you
need
all
this
stuff.
B
A
C
A
C
A
A
Necessary
to
meet
people-
it's
so
much
different
when
you
are
face-to-face
like
super
driven
customer
customer
have
been
using
Bosch
backspace
alert.
I
would
never
have
started
any
conversation
over
the
net
about
this
whole
project.
Here
I
could
and
I
could
meet
all
the
guys,
I've
been
seeing
around
remotely.
But
it's
not
saying.
C
C
A
You're
doing
anything
with
open
source
projects
tickets
issues
using
this
is
you
know
it's
you
start
to
you
already
notice
people
buy
their
handles
or
whatever,
and
actually
you
don't
just
walk
in
as
a
stranger
like
you,
walk
in
and
you've
got
people
excited
to
see
to
me.
That's
a
good
conference
about
conferences
window
at
the
far
enough.
A
C
A
C
Talk
about
this
a
little
bit
so
I
already
know
what
we're
gonna
be
talking
about
next
next
time
there
is
the
guys
from
orange.
We
want
to
present
a
terraforming
solution
that
they
are
contribute
to
CF
extensions.
So
we
have
that
and
then
I'm
working
on
getting
more
so
too
you'll
see
more
of
the
model
of
having
presentations
getting
people
to
discuss
what
they
are
thinking.