►
From YouTube: Meshery Community Meeting - Sept 27th 2019
Description
Discussion of @smi-spec and sleek logo designs this week. Raul lands his first-ever open source contribution!
A
A
B
A
A
B
A
A
A
There's
dungeon,
okay
and
you
know
great
and
then
hopefully
let
me
paste
the
link
to
the
meeting
minutes
again,
so
folks
can
go
in
and
oh
and
help
me
type
while
I
talk
and
listen
to
you
guys,
okay,
good,
we
are
about
five
six
minutes
after
so,
let's,
let's
get
started
so
some
general
announcements.
This
is
this
first
one's,
actually
a
carryover
I
think
from
last
week,
where
there
will
be
a
force
to
report
coming
out
shortly.
Dad
it's
gonna,
highlight
measure
e
and
really
highlight
the
concept
of
multi
meth,
our
multi
mesh
manager.
A
Rather
that
was
kind
of
the
language
that
they
it
used
to
describe
measure
e,
a
multi
mesh
manager.
That's
in
my
mind,
good
language
we've
been
referring
to
my
areas
of
Management
plane
this.
Actually,
since
we
have
a
bit
of
a
smaller
audience
today,
it
might
be
worth
taking
a
quick
look
at
what
I
mean
by
a
management
plan.
A
A
We
say
that
hey
in
general,
across
all
service
meshes
or
what
makes
up
a
service
mash
is
the
data
playing
a
control
plane
and
the
control
plane
is
that
the
implementation
of
the
control
plane
is
specific
to
the
type
of
service
mesh
that's
running,
but
beyond.
That
is
a
management
plane.
This
concept
is
very
familiar
to
network
engineers
and
as
you've,
seen
as
each
of
you
have
gotten
into
service,
meshes
that
they're
very
networking
centric.
So
this
management
plane
is
where
a
tool
like
measuring
resides
and
so.
A
And
so
yeah,
so
that's
what
he's
telling
folks
about
will
since
since
we're
on
the
subject,
maybe
we'd
said
that
there'd
been
some
enhancements
made
to
make
sure
that
people
can
walk
through
the
workshop
using
massery.
So
Raul
did
you
in
guadalajara?
Did
you
did
you
end
up
running
mastery
at
the
time
or
what
did
you
we
I
guess
I
know
now:
I
recollected,
the
workshop
that
we
went
through
at
waddle
ahora
is
actually
different
than
the
one
that
that
we
normally
give
here
mm-hmm,
so
it
might.
A
B
A
A
A
Just
a
recent
enhancement
in
the
mesh
reissue
yeah
I'm,
sorry
yeah,
the
massery
sto
adapter,
where
you're
able
to
apply
basically
the
sample
configurations
for
a
book
info,
and
we
use
this
as
a
tool
to
drive
their
workshops
is
that
this
is
pretty
inflexible
to
Nuno's
point
and
needs
to
show
the
user,
the
actual
yam,
while
that's
being
applied,
so
they
can
be
comfortable
with
it
and
right
now
it
doesn't,
but
there
are
features
set
up
to
do
what
you're
describing
Nuno,
which
is
for
those
that
want
to
directly
put
in
their
own
your
paste
in
their
own
gamal
that
they
they
could
do
so
to
just
in
you
know,
in
its
Diaz
case,
to
describe
a
new
virtual
service,
can't
paste
it
in
here
and
have
it
applied
in
this
case.
A
This
is
gonna
is
gonna
fail,
but
it
need
the
UI
needs
to
evolve.
Well
beyond
that.
To
make
it
such
that,
like
I,
think
if
anyone
is
used,
if
you've
ever
configured
a
firewall
rule-
or
you
know
within
maybe
in
within
ec2
or
really
anywhere,
where
you're
using
a
visual
interface
to
configure
a
firewall
rule.
But
this
that's
probably
what
the
visual
interface
for
ma
cherie
to
configure
these
different
types
of
rate
limits
and
other
things
might
end
up.
Looking
like
and
it
gets
a
little.
A
A
Okay,
good,
so
maybe
maybe
last
thing
on
that
particular
topic
is
just
to
show
that
let
me
stop
sharing
and
share
that
a
couple
of
things.
One
not
only
has
joined
us
from
Guadalajara,
but
so
has
Karen
who's,
not
on
the
call
here,
but
she's
been
doing
some
work
around
our
globe,
the
design,
the
logo,
design
for
layer
5
and
for
measure
e
as
well
and
I've,
been
I.
Think
many
of
you
have
been
unfortunate
enough
to
have
your
opinions
solicited
about.
A
A
This
is
what
it
looks
like
a
little
more
up
close
if
I've
got
the
right
picture
up
yeah
there
we
go
so
anyway,
personally
excited
that
your
work
is
being
represented
today
at
the
the
workshop
measure,
e
I
think
yeah,
just
you
know,
like
I
said
before,
is
gonna
drive
the
workshop.
So
that's
great
on
this
point.
I
really
wish
Karen
was
on
cuz
cuz
she's
been
providing
a
lot
of
these
new
designs.
Let
me
share
it,
maybe
in
the
slack
channel
or
show.
A
That's
as
any
other
human
on
this
planet
I
think
particularly
on
something
that
that
I
considered
to
be
very
important,
I,
don't
know
in
the
end.
Maybe
it's
not
a
nick
I
have
to
say
Hashi
corpse,
sleekness
and
UX
is
inspiring
and
when
we
think
about
a
design
like
that,
so
there's
a
bit
of
hypnocil
sexiness.
That
I
think
is
associated
with
hashey
corpse
logos.
D
A
A
So
I
as
Nick
looks
for
that
logo,
which
is
I,
think
we
might
all
get
a
chuckle
out
of
I'm
gonna
go
back
to
sharing
the
minutes,
I'm
a
bit
disoriented
this
morning
and
I
apologize.
There
are
any
number
of
things
actually
I
think
to
share
and
discuss
so
a
couple
of
things.
We'd
been
a
couple
weeks
ago,
we've
been
relatively,
the
community
had
been
relatively
hard
at
work
on
the
build
and
release
process
for
Mestre
itself,
so
we've
switched
from
Travis
CI
over
to
github
actions.
A
Here
we
haven't
switched
to
that
I,
don't
know
if
there's
I,
don't
know
that
we're
necessarily
incented
to
do
that,
primarily
because
you
know
more
or
less
the
world's
docker
installations.
The
runtimes
the
default
registry
is
docker
hub
I
think
it's
just
a
little
bit
easier
to
go
referenced
there.
So.
A
So
the
nice
thing
about
that
is
that
Kieran,
who
isn't
on
the
call,
but
he
had
created
a
homebrew
formula
for
mesh.
We
CTL
to
just
do
a
quick
brew,
install
marry
CTL
now
that
we've
got
version,
2
releases
of
Mishra
CTL
and
those
are
happening
and
being
released
on
github.
That
formula
can
be
stable
and
can
land.
So
that's
good.
D
D
D
D
A
I
think
you're
right,
then
you
know
a
lot
of
times.
You'll
see,
logos,
start
relatively
complex
and
just
simplify,
simplify,
simplify
over
time
new.
No,
you
look
like
you.
You
got
first
dibs
on
stickers
from
what
I
tell
that's:
okay,
fair
enough,
we'll
see
we'll
see
what's
happening,
buy
it
buy,
cube
car,
okay,
actually,
speaking
of
cute
con
Nick,
just
given
your
involvement
in
service
mesh
day
or
service
mask
on,
rather,
he
I
think
I
think
talk,
acceptances
and
submissions
and
things
were
were
determined.
Those
were
sent
out,
the
deed.
A
D
F
D
D
Link
D
and
it's
it's
pretty
pretty
sort
of
heavily
weighted,
but
I
I
think
that's
kind
of
fine
in
some
ways,
because
yeah,
it's
all
about
what
you
call
I
guess
like
consoles,
are
pretty
new
to
the
market.
So
we
it's
only
really
been
this
summer,
where
we've
had
the
features
which
would
realistically
you
could
write,
you
want
to
run
things
in
production.
A
It
it's
still
a
single
track
right.
It's
just
a
single
track
day:
yeah,
okay,
okay,
fair
enough
yeah,
and
that
makes
a
lot
of
sense.
I
mean
the
more
the
the
more
that
there
is
to
talk
about
and
do
it
within
a
given
project.
You
would
expect
that
the
more
heavily
weighted
submissions
are
about
that
so
actually
yeah.
D
D
It
was
heavily
political
to
be
honest
with
you,
which
I
was
a
little
bit
kind
of
saddened,
the,
but
but
I
guess
it's
it's
kind
of
understandable
when
you
basically
get
a
bunch
of
vendors
to
vote
on
on
talks
by
my
personal
perspective
on
that
was
I
parked
who
I
worked
for
and
voted
on.
Talks
on
the
merit
of
what
I
felt
would
be
a
benefit
to
the
practitioner
rather
than
you
know,
just
kind
of
to
be
trying
to
promote
hope
and
I'm
like
throw
shade
on
everyone
else.
D
So
that
was
that
was
kind
of
I
was
kind
of
interesting,
but
in
some
ways
expected
I.
Think
in
general,
servicemen
Carney
is
is
something
which
probably
should
be.
They
go
it's
kind
of
weird
you
can
of
the
fringe
event,
and
then
you
have
also
service
which
talks
of
the
main
food
car
so
I,
don't
know,
I,
don't
know
what
what
what
the
orientation
on
that
will
be
will
be
next
year,
a
view
one.
Have
you
also
seen
that
the
CFPs
are
open
for
the.
D
D
I
think
there's
also
relevant
opportunity
there
for
for
mastery,
because,
ultimately,
if
you
are
running
a
service
match
being
able
to
kind
of
evaluate,
may
be
able
to
test
being
able
to
kind
of
run
a
bunch
of
debugging
and
things
like
that,
there's
I
think
there's
a
good,
a
good
fit
a
few.
You
are
thinking
of
submitting
yeah.
A
Yeah
interesting
yeah,
yeah,
okay,
and
so
this
is
at
Q
Khan
as
well
all
right,
interesting,
okay,
good
yeah,
it's
a
twist
lock,
is
part
and
parcel
to
it.
Okay,
yeah
yeah,
Nick,
maybe
wanna
yeah
I
know
you
don't
have
enough
to
do
it.
Cute
comes
to
me:
maybe
we
want
to
brainstorm
something
there.
Yeah.
D
D
Not
even
trying
to
think
of
could
con
right
now,
it's
it's
so
far
into
so
many
hurdles
ahead
of
where
I
am
right.
Now,
I'm,
just
kind
of
yeah,
yeah
you're.
A
D
So
I
think
a
bunch
of
good
stuff
at
Hoshi
comment.
So
we
we
didn't
have
any
new
features
for
console
because
well
direct
features
for
console
from
the
1.6
release
that
we
brought
out
in
the
summer.
But
we
did
go
GA,
one
six,
which
introduces
all
the
layer
7
configuration
kind
of
one
of
the
biggest
announcements
is
that,
where
we're
partnering
with
Microsoft
and
what
we're
doing
with
with
Microsoft
with
that
partnership
is
bringing
managed
console
to
Asia.
D
So
anybody
who
wants
to
run
console
I'm,
predominantly
probably
in-service
mash
mode
rather
than
just
kind
of
vanilla
service
discovery
and
can
can
now
do
so
and
actually
Corp
will
kind
of
look
after
the
management
of
those
servers.
So
you
you
don't
need
to
do
anything.
You
just
kind
of
click
a
button
and
if
you
get
any
points
that
you
can
then
use
with
your
existing
infrastructure.
So
we're
kind
of
my
take
on
like
why.
That's
really
really
important
is
heterogeneous
environments.
D
So,
if
you're,
using
kubernetes
and
pure
kubernetes
and
that's
awesome,
but
the
the
kind
of
certainly
larger
enterprises
and
and
I,
think
a
lot
of
people
are
still
in
this
migratory
step.
So
one
of
the
things
that
kind
of
cash
a
call
bar
are
trying
to
enable
folks
is
to
do
that.
Easy
migration.
So,
for
example,
with
console
Service,
mash
and
in
gateways
features
that
we
introduced.
You
could
run
a
single
console
service
match
which
is
broken
up
of
a
number
of
data
centers.
D
So,
for
example,
I
can
run
data
center,
one
which
is
going
to
be
my
virtual
machines
and
I
can
run
data
center,
two,
which
is
my
kubernetes,
so,
for
example,
I
have
an
application
which
is
running
in
virtual
machine
and
I
want
a
my
great
backed
to
the
cloud
into
specifically
into
kubernetes.
Then
you
know
you
potentially
have
to
do
a
big
bang
approach.
You've
got
a
lift
and
shift
all
of
the
dependencies
of
that
application
and
put
them
into
kubernetes,
and
that
is
complicated
and
it
carries
risks.
D
However,
if
you
leverage
your
service
match
and
what
you
can
actually
do
is
deploy
the
application
to
kubernetes
and
your
traffic
can
be
seamlessly
routed
between
your
virtual
machines
and
GUP
without
having
to
worry
about
flat
networks
and
that
or
or
any
sort
of
VPN
connection,
because
the
Gateway
handles
all
of
that,
but
the
other
benefit
of
using
service
matches.
You
can
all
start
doing
things
like
Canaries.
D
The
bulk
of
it
being
home
will
buy
my
virtual
machines
and
it
gives
me
the
ability
to
do
that
kind
of
gradual
deployment
or
gradual
migration
which,
having
been
only
the
rough
end
of
that
stick
and
having
to
do
a
big
bang
approach
before
I.
Really
I
really
wouldn't,
wouldn't
want
to
do
it
again,
but
I
just
was
know
tooling
at
the
time.
So
we're
we're
pretty
excited
about
that.
I
think
like
console
works,
really
good
in
in
in
Azure
and
I
would
kind
of,
without
saying
too
much
definitely
say,
watch
watch
that
space.
A
A
Or
yeah
I
guess
this
is
a
pretty
small
column,
doesn't
really
matter,
but
but
yeah
for
the
valuation
that
an
organization
like
a
Shi
Corp
has
and
of
those
that
have
got
deep.
This
isn't
necessarily
leave
enough
pockets.
Anyway,
it's
a
great
relationship.
You
can
see
why
Microsoft
was
also
very
friendly
to
Blanton,
to
link
Rd
and
kind
of
you
know
similar
motion
and
etc,
and
so
anyway,
so
so
Congrats
I
was
fantastic
to
see
you
up
on
the
stage
talking
about
mushy
things
as
you've
done,
yeah.
D
That
was,
that
was
a
kind
of
a
last-minute
last-minute
request.
A
Nick
and
I.
Think
most
of
the
rest
of
you
Haim
is
a
good
example
of
someone
whose
paycheck
comes
from
a
vendor
and
is
trying
to
do
right
by
you
know,
be
being
loyal
to
the
is
trying
to
do
right
all
around,
and
so
I
wanted
to
talk
a
little
bit
about
service
mesh
interface.
It's
a
an
ABS
API
abstraction
across
service
meshes
and
Nick,
certainly
representing
those
and
represents
well
and
even-handedly
and
equally
and
recognize
that
we're
in
a
sort
of
embracing
the
open-source
way
and
that
we're
in
a
smaller
community
and
I.
A
Think
time
has
shown
similar
behavior
as
well,
which
I
think
it's
one
of
those
things
that
sometimes
it
hurts
in
the
short
run
because
interests
weren't,
because
you
didn't
stick
your
elbows
out
and
make
sure
that
your
you
know
your
personal
interests
weren't
the
best
represented,
but
in
the
long
term,
is
the
winning
strategy.
From
my
perspective,
so
I
didn't
want
to
talk
about
SMI
service
mesh
interface,
that
much
but
but
to
say
that
Hamlet
or
the
other
somewhat
related
project
from
VMware.
A
A
A
I
think
you
might
be
able
to
characterize
it
as
sort
of
service
discovery
for
service
meshes
like
the
ability
to
in
a
uniform
way,
speak
to
a
service
mash
and
say
what
are
you
capable
of
and
to
a
different
one
where
you
capable
of
and
help
them
interoperate
a
bit?
It's
a
bit
of
a
different
initial
vector
or
perspective
angle
from
which
this
effort
is
being
drawn
up
as
as
compared
to
SMI,
which
I'll
yeah
I.
D
Think
the
kind
of
the
key
the
the
sort
of
the
key
difference.
What
SMI
does
is
it?
It
proposes
a
high-level
abstraction
interface,
whereas
the
the
the
VMware
initiative
is
all
about
low-level.
It's
it's
kind
of
a
specification
for
low-level
connectivity
between
service
measures.
So
how
do
you
kind
of
share
service
catalyze,
as
we
were
explaining
in?
How
do
you
share
service
catalogs?
How
do
you
share
and
sort
of
feathery
identity,
because
obviously
mt
elapsed
requires
to
the
gates
or
been
able
to
validate
identity
cross
cross
cross
service
mesh?
D
D
They
all
need
to
play
nice
and
yeah
and
I.
Think
it's
I
think
it's
interesting.
I
was
actually
just
as
a
side
note.
I
was
actually
just
speaking
with
Sergio
this
morning
around
SMI
and
trying
to
encourage
VMware
to
to
to
get
involved
and
kind
of
one
of
the
things
that
so
Sergio
was
saying,
and
what
I'm
saying
is
that,
like?
Maybe
it's
a
little
early
for
a
specification
and
I'm
like
well,
maybe
but
like?
D
You
know
in
the
face
whether
it
is
just
a
service
mesh
interface,
because
theoretically,
there's
no
reason
at
all.
You
couldn't
interact
directly
with
console
using
SMI
is
an
interface
configure,
the
various
different
things
without
that
being
run
on
kubernetes,
and
certainly
for
the
folks
at
VMware
with
NSX
SM.
You
know
that
that's
obviously
a
an
important.
A
Yeah
I
agree
with
you:
I
was
in
a
it
was
I
was
gonna,
try
to
pull
up
the
issue
in
github
on
the
SMI
repo,
where
Brendan
and
someone
else
were
chatting
and
it
was
like
they
were
I
had
said.
Well,
you
know
that
that
makes
sense
if
this
is
only
gonna
be
a
kubernetes
thing
and
and
then
I
went
back
to
reread
the
intro
test.
A
You
sounds
like
a
really
well
allowing
me
to
create
initial
conversation
and
that
he
had
hinted
that
maybe
there
was
more
overlap
or
he
was
considering
that
maybe
there's
a
bit
more
overlap
with
SMI
than
he
had
originally.
You
know
considered
and
I
in
some
respects.
I
guess
I
haven't
given
it
a
ton
of
thought
but
I.
A
I
am
pleased
to
see
that
VMware
spec
did
get
announced
which
finally
kicked
the
Microsoft
in
the
pants
to
actually
have
a
call.
You
know
six
months
later,
or
you
know:
okay,
fine,
four
months
later,
so
so.
I
know
that
Nick
and
I
have
been
talking
about
this.
Hopefully
folks
are
picking
some
of
this
up.
There
is
a
specific
thing
that
I
wanted
to
describe
with
respect
to
service
mesh
interface.
It
seems
to
resonate
a
fair
bit.
A
It
resonates
not
only
with
SMI
with
respect
to
measure
e,
but
it
also
resonates
with
the
VMware
spec
and
it
could
be
work
that
this
community
could
pick
up.
The
the
work
is
around
the
notion
that
it's
around
conformance
and
compliance
testing,
and
it's
this
and
we
talked
about
it-
a
fair
bit
Nick
last
SMI
call.
A
And
so
most
the
time
that
you
see
a
specification
come
forth,
SNMP
HTTP,
you
know
proto
tank
protocols
and
the
you
know
these.
These
types
of
things
when
you
think
of
kubernetes
and
all
of
the
the
organizations
that
have
implemented
behind
its
api's,
whether
there's
conformance
testing
that
goes
on
as
vendors
go
to
produce
their
own
kubernetes
distributions
and
things
anytime.
You
have
a
spec,
you
have
conformance
tooling
cloud.
Events
is
another
example
that
there's
been
it
was.
It
would
seem
like
a
fair
bit
of
ask
and
agreeance
around
a
tool
like
mystery.
A
D
If
you
remember
kind
of
the
last
Essen
a
couple
of
s
of
my
calls
ago,
I
had
the
gall
to
question
whether
SMI
was
lowest
common
denominator,
run
or
not
and
received
a
little
bit
of
flack
from
wrong.
But
but
my
kind
of
concept
around
that
is
that,
like
I,
think
that
that
could
be
solved.
That
so
SMI
is
about
configuration
inbound.
But
what
about
that
outbound
feature
discovery?
Is
that
covered
by
the
the
VMware
spec,
or
is
that
covered
by
something
that
like
SMI?
Should
she
build
in
the
same
way
as
traffic
metrics?
D
Allow
you
to
kind
of
discover
metrics.
Should
there
be
a
a
formal
part
of
the
spec
somewhere
which
allows
you
to
describe
what
features
are
supported
and
the
the
reason
that
I
I
sort
of
say?
That
is
because
for
a
tool
like
ma
cherie,
if
you
didn't
have
to
maintain
an
internal
list
of
is
traffic
splitting
supported
in
glozell,
you.
D
We
can
build
a
much
nicer,
tighter
integration
for
for
from
a
practitioners
stance
and
if
surface
mesh
interface
is
always
going
to
be
lowest
common
denominator.
Then
okay,
you
don't
you
don't
need
that,
but
I
I
feel
that
I
feel
quite
strongly.
That
folks
need
comprehensive
kind
of
a
comprehensive
interface
which
allows
them
to
do
like
their
daily
tasks.
D
Regardless
I
think
even
even
from
a
application
integration
perspective,
so
maybe
that's
that's
something
worth
tabling
as
a
kind
of
a
concept
for
the
next
call
and
saying
can
we
have
a
definition
of
a
discovery
in
the
face
which
just
kind
of
says
these
features
or
those
features
are,
are
supported
by
the
various
different
meshes
and
the
various
different
versions
of
meshes
as
well
right,
yeah.
A
Yeah
I
agree:
I
mean
I,
agree
that
well
it's
before
I
go
off
and
describe
how
it
is
that
I
agree.
Let
me
say
that
yeah
I'd
like
to
have
a
win
or.
A
Do
we
have
a
relatively
small
community
of
mesh
really
moves
so
quickly
at
the
moment
and
yeah
be
nice
to
have
a
win
maybe
around
or
around
I
guess
around
anything
whether
it
was
the
compliance
tool
or
this
there,
the
the
compliance
stuff
might
be
smaller
in
nature.
It's
certainly
less
inflammatory
the
then
the
other
suggestion
around
lowest-common-denominator
yeah.
I
thought
that
it
was
a
bit
disappointing
that
so
for
those
on
the
call.
A
Quote-Unquote
proprietary
features
of
each
mesh
and
by
the
first
approach
of
lowest-common-denominator
I
think,
has
the
benefit
of
being
universally,
or
rather
you
would.
As
you
go
to
interface
with
SMI
it.
The
experience
will
be
consistent,
consistency
and
consistency,
uniformity,
people
will
readily
call
into
question
and
I
think
that
they
will
probably
call
the
question
the
value
of
a
service
mesh
behind
that
or
as
to
facilitate
load
balancing
or
some
of
these
things
be
a
traffic
splitting
because
they
would
say
well
I'm
happier
on
all
this
additional
infrastructure
run.
A
This
additional
layer
under
understand
this
stuff,
updated
as
it
goes
rather
than
just
using
the
EEO,
beat
the
load
balancer
in
front
of
it
and
updating
its
config
and
like
this
is
just
what's
what's
the
benefit,
if
it's
so
low
that
the
power
of
these
meshes
and
the
auto
instrumentation.
These
other
features
that
you
get
yeah
I
think
that
it's
a
lot
more
interesting
and
it's
a
lot
more
helpful
to
have
a
what's
more
or
less
a
common
interface,
but
the
one
that
dynamically
says.
A
A
It's
the
same.
Not
it's
the
same.
Looking
interface,
you,
you
know
it's
the
yeah,
so
you
gain
the
benefit
of
both
worlds
and
yeah.
The
I'll
digress
for
a
second
and
say
a
big
reason
why
I
wholeheartedly
embrace
that
philosophy
is
because
I
built
a
team
of
just
less
than
a
hundred
folks
at
Cisco,
and
we
built
out
Cisco's
cloud
management
platform
and
you
know
I,
think
I
think
it
might
still
be
being
released,
it's
probably
in
a
sustaining
mode,
but
but
it
was
a
anyway.
A
On
top
of
a
many
different
types
of
infrastructure,
you,
as
a
user,
had
no
clue
as
to
what
the
actual
infrastructure
behind
the
scenes
was
and
and
for
that
particular
product.
That
was
part
of
the
point
that
was,
but
you
know
that
doesn't
have
to
be
the
point
here,
but
but
my,
but
when
you
as
a
user
or
when,
when
your
deployment
was
down
on
top
of
ESXi
or
V
Center,
there
were
quite
a
few
things
that
you
got
to
do
as
a
user.
A
If
it
was
deployed
on
top
of
OpenStack,
there
were
lesser,
but
the
experience
was
the
same.
You
just
either
gotten
more
or
less,
but
so
yeah
I
certainly
believe
in
that
I
believe
in
building
out
a
compatibility
matrix
in
the
table
in
part
I'm,
hoping
that
the
landscapes
or
the
layer
5
landscape,
to
kind
of
builds
towards
that
that
I'll
speak
for
myself
and
say
my
my
attention
is
way
too
divided
to
accomplish
some
of
these
like
really
meaningful
things.
I'm
totally
on
board.
A
I
will
be
open
with
you
and
say
that
I'm
not
sure
if
I'm
willing
to
expend
my
political
capital
on
or
like
this
or
I
was
disappointed
that
there
was
a
response.
Basically,
the
other
vendor
on
the
call
was
saying:
well,
these
things
don't
make
sense
and
mostly
because
they
didn't
have
those
capabilities
like
that.
Didn't
that
wasn't
that
it
didn't
so
the
other
way
of
encouraging
this
and
saying:
let's
do
it.
D
Like
10
years
ago,
you
say
mum
when
people
were
talking
about
configuring
firewalls
and
things
like
that,
and
the
the
space
was
just
so
kind
of
open
and
not
to
me
my
dad
and
that
that
it
needed,
like
a
couple
of
years
for
the
market
to
consolidate
around
a
sort
of
a
few,
a
sort
of
set
of
products
to
be
able
to
kind
of
get
the
conversations
but
which
I.
You
know,
I
see
his
point.
But
my
kind
of
point
is
that
we
work
in
a
slightly
different
world.
Now
we
are
a
lot
more
open.
D
Don't
disagree
about
the
consolidation
and
the
evolution
that's
happening,
but
actually
agreeing
these
things
now,
as
a
as
a
kind
of
a
group
protect
everybody,
because
if,
if
you
know
NSX
I
Sam
has
a
new
feature
and
I
look
at
that
I'm,
like
that's
an
amazing
feature
like
we
will
copy
that
one
day
it
it's
actually
nice
to
be
able
to
say
well,
hey
I'm
tabeling,
this
SMI
specification
for
my
feature
that
nobody
else
has
and
it's
my
kind
of
my
you
know.
But
but
ultimately
folks
would
look
at
that
and
go
well.
D
D
D
A
To
the
extent
that
these
two
is
as
large
as
it
is,
that
has
probably
prevented
the
the
absorption
of
surface
machining
to
kubernetes
and
that
what's
helped.
That
is
the
number
of
implementations
of
it.
Then,
unlike
a
helm
or
some
other
tool,
that's
just
has
borne
out
of
kubernetes
in
a
specificity
that
I
think
as
we
go
to
end
the
call
here
it
I
reminded
that
I'm
reminded
that
I've
got
a
little
bit
of
work
to
go.
A
A
Well,
no,
no
it'll,
open
trait,
no,
no
you're,
right
tracing
in
open
census
as
being
like
competitive
to
the
extent
that
open
tracing
was
there
for
some
time
and
then
people
that
implemented
toward
it
and
open
sense.
This
comes
out
in
this
kind
of
Auto
instrumenting,
a
number
of
other
things
bring
the
other
metrics.
It's
not
just
distribute
racing
and
we
ended
up
coming
into
open
and
telemetry
possible
for
something
like
that
to
happen.
With
these
two
yeah
think.
D
There's
hints
to
to
some
stuff,
like
this
I
mean
looking
at
like
some
of
the
new
API
v2
ingress
in
in
kubernetes.
There
are
definitions
inside
of
the
tube
network,
experimental
namespace
for
defining
things
like
traffic
splitting
so
I.
As
you
know,
though,
the
there's
therefore,
and
I
sort
of
an
element
of
crossover
and
we
got
a
bounce
but
dude
I
would
I'd
love
him
like
a
101
on
the
sig
networking.
If
you've
got
a
yeah
half
hour
next
week,
yeah
no
I
really
appreciate
that.
A
A
So
thanks
for
that,
there
are,
if
any
of
you
guys,
know
rogue
heat.
If
you
see
him
poke
him
tell
him.
We
miss
him
here.
Cuz
that
guy
was
like
I
was
moving
fast,
we'd
love
to
keep
him
engaged
and
then
time
I
think,
like
hey,
we
landed
the
that
dot.
Your
docks
I
think
we're
ready
we're
ready
to
go
from
beta
to
stable
and
have
octarine
adapter,
be
part
of
every
anytime.
Anyone
downloads
that
actually
hi.
A
Yeah
I
was
gonna
suggest
that
we
were
giving
a
workshop
in
that
we
update
the
workshop
to
include
Audrina
and
I
remembered
like
actually
we
don't.
We
don't
have
link
or
do
your
console
or
anything
else
in
there,
because
it's
an
HDL
workshop,
but
but
it
is
exciting
that,
with
the
workshop
does
use
misery
and
everyone
would've
been
sitting
right.
There
earn.
G
Actually,
you
haven't
lend
me
a
book
that
is
I
have
to
organize
the
adductors
alphabetically,
so
I
am
looking
for
the
food
that
was
NAB
dot,
html'
to
zero
sum
of
something
like
grandchild.
30
virus
and
child
avoidin
note
dirty
water,
so
I'm
not
getting
worth
of
food,
so
I
have
to
edit
blue
food
right.
A
Yes,
I'm
trying
to
remember
the
framework
that
we're
using
for
these
dogs,
so
your
goal
is
to
organize
these
alphabetically,
as
opposed
to
as
opposed
to
insinuating.
That
console
is
the
most
important
by
putting
it
first
yeah
I
understand
what
you're
asking
let
me
do
a
bit
of
homework
and
and
come
back
to
you
because
I'm
not
sure
what
Ruby
is
doing
this
on
the
back
end
or
like
what
oh
good
so
yeah.
Let
me
I'll
hit
you
up
here
a
little
bit
later
today
with
some
info
only.