►
From YouTube: Network Service Mesh Meeting - 2019-05-07
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
A
Great,
let's
go
ahead
and
get
started
so
welcome
to
the
next
network
service
mesh
meeting.
So,
as
a
reminder,
please
add
yourself
to
the
attendees
list.
If
you
have
not
done
so
already
and
let's
get
started
with
some
agenda
bashing,
is
there
anything
that
anyone
would
like
to
talk
about
that?
It's
not
on
the
agenda
list.
A
A
A
A
So
we
have
so
we
have
three
recurring
network
service
mesh
talks
that
we
have
this
one.
At
8:00
a.m.
the
NSM
document
call
at
8:00
a.m.
Wednesday.
We
have
the
used
to
pay
stocks
which
are
on
Mondays
at
8:00
a.m.
and
on
twos
on
the
second
fourth
and
fifth
Mondays,
and
we
also
participate
in
the
sense,
the
telecom
user
group
members
of
the
feather,
which
occurs
every
person
third
Monday
at
8:00
a.m.
on
Thursday
May
23rd.
A
We
will
be
giving
a
intro
deep
dive
that
the
telecom
user
group
and
found
made
in
the
network
functions
testbed
as
well.
We
have
cube
con
coming
up
in
May
21st
street
23rd,
and
this
will
be
held
at
Barcelona.
Please
book
your
hotel.
If
you
haven't
done
so
already,
we
have
an
intro
and
a
deep
dive,
maintainer
track
that
you
can
attend
and
we
should
be
offering
a
demo
at
the
elephant,
booth,
I
believe
in
conjunction
with
the
OTL
stuff.
A
We
have
two
collocated
events
at
q
connie
you
we
have.
The
Fido
mini
summit,
which
occurs
average
is
going
to
occur
on
Monday
May
20th
in
Barcelona
and
the
cloud
native
network
service
state.
We
have
talks
in
both
if
you
have
any
food
talks
or
demos
or
anything
like
that.
You
have
a
coupon.
Please
speak
up.
A
A
We
have
OH&S,
that
is
on
June
25th
through
26.
We
have
ons
Europe
coming
up.
The
call
for
paper
is
currently
open
and
closes
on
June
16th,
the
middle
of
the
next
month
that
occurs
in
Antwerp.
So
if
you're
interested
in
talking
feel
free
to
submit
a
paper
there
in
November,
we
have
two
events:
the
Emmy
of
2019
and
Q
come
the
call
for
paper
for
Q.
Khan
is
now
open
as
of
yesterday,
so
it
is
open
till
July
12.
So
we
we
have
some.
A
E
A
It's
and
the
the
site
also
reminds
me
that
we
have
to
start
ending.
These
frequently
asked
questions
to
the
site,
so
I
will
make
sure
to
get
started
on
that.
Actually
Bremen's
gonna
start
a
Google
Doc
track,
but
I,
don't
I
haven't
seen
one
passed
around
unless
I
missed
it.
So
I'll
I'll
open
one
up
and
email
it
to
the
earthbound.
A
A
A
All
the
pull,
the
all
the
pull
requests
line
by
light
and
I
realized,
there's
800
of
them,
and
so
as
much
as
I
would
love
to
to
do
that.
I
decided
to
write
a
higher
level
overviews
to
what
we've
what
we've
built.
There
may
be
things
missing
in
this,
so
if
you
feel
there
is
a
major
component
that
is
missing
in
the
reference
architecture
or
or
otherwise.
A
So
a
couple
things
that
we
need
to
work
out
as
well
is
for
downloads
I've
added
a
couple
things
that
we
could
talk
about
for
the
Downloads,
so
helm,
docker
images
to
the
source
code,
preferably
signed
Arneson.
So
are
there
any
recommendations
or
anything
like
that
that
that
people
would
like
to
see
that
I
missed.
B
A
A
Definitely
agree
my
my
one
preference
on
that
is
that
we
we
actually
make
three
accounts.
Okay,
and
the
reason
for
that
is
one
of
them
is
our
standard
test.
One
of
them
is
our
one
of
them.
Is
our
fully
released
version,
the
third
one
so
that
we
can
we
can
test
they
test
the
release,
yeah.
That
way,
we're
not
publishing
directly
to
the
to
the
main
repo
staging.
A
Or
the
other
alternative,
but
basically
I
want
the
CI
to
to
publish
to
do
some
staging
before
anywhere
I
could
either
be
a
repo
with
the
same
name
in
a
different
in
a
different
location.
So
we're
not
publishing
the
darker
hub
or
it's
unblocker
hub
and
it's
and
it's
under
a
different
name
space.
So
one
or
the
other
but
I
think
that's
a
fantastic
idea.
B
C
A
So
and
one
last
thing
about
as
well
as
I'm
also
sticking,
even
though
it's
not
part
of
the
main
release
itself,
I'm,
also
putting
the
fact
that
it's
C
D
and
C
at
membership,
and
then
the
new
logo
is
being
added
in
here
as
well.
So
once
that
logo
comes
up,
we're
going
to
we're
going
to
inject
it
in
and
we
get
to
go
one
thing
I'm
going
to
need
help
with
is
under
the
known
issues.
If
there
are
any
major
bugs
that
we
have
not
resolved
that,
we
want
to
call
out.
A
We
have
to
make
sure
that
we
populate
those
with
not
just
a
list,
but
the
actual
issues
that
attract
those
those
problems
and
then
moving
forward
will
call
out,
like
kubernetes,
will
actually
call
out
major
major
pull
requests
and
inspects
that
that
have
changed
network
service
mission
interesting
way.
Moving
forward
and
I'll
also
need
some
help
with
correlating
and
getting
started
as
to
what
do
we
want
to
show
off
and
I'm
making
sure
that
that's
all
tested
for
the
for
the
main
release
itself?
A
B
B
A
And
when
I
didn't
die
very
clearly,
this
is
an
alpha
release,
and
so
that
should
be
code
for
use
this
in
production.
I,
don't
use
this
in
production,
depending
on
your
your
level
of
risk
cool,
so
back
to
the
main
agenda,
so
I
saw
Alex's
on
the
alex
is
on
the
agenda,
so
work
she's
in
the
room.
So
you
have
oh.
F
F
F
B
B
C
Okay,
I
am
super
glad
that
we
moved
away
from
the
juicy-looking
spider
and
the
web
nonsense
and
went
with
the
matrix
ice
actually
sorta,
like
the
the
third
one
down
a
little
bit
better
or
even
the
orange
and
red
that's
below
that
one
I
kind
of
prefer
the
transition
a
little
bit
more,
as
opposed
to
the
darker
outer
matrix
with
the
lighter
inner
matrix.
But
that's
just
me,
I
I
can
handle
the
one.
C
That's
just
off
the
screen
now,
on
the
right
hand,
side,
but
I
prefer
the
light
to
dark
transition
personally
and
I,
definitely
like
the
reoriented,
I'm
heptagon.
That
is
now
got
the
flat
side
down
it's
not
too
too
overly
kubernetes,
but
it
still
kind
of
gives
a
nod
to
the
CN
CF
that
adopted
us.
So
that's
my
thoughts.
B
I'm
very
sensitive
to
color
palettes
because
I,
as
you
know,
couple
slides
here
there,
and
so
you
want
to
get
color
palette
to
work
with
and
I
was
super
happy
that
our
3c
to
be
orange,
a
yellow
one
that
if
you
invert
the
colors
on
that
just
mechanically
invert
them
you
get
something
that
is
a
really
lovely
blue
on
black,
but
also
works
and
I
drop.
That
out
to
the
thread
as
well.
F
C
B
D
B
F
B
B
Do
you
folks
have
any
strong
feelings
between
those
two
I
think
we've
had
some
strong
feelings
for
a
lot?
Oh,
but
I
don't
know
if
they
were
just
because
it's
a
cool
longer
in
comparison.
C
The
one
that's
on
the
purple,
brown
and
green
one
on
the
left-hand
side,
that's
also
very
clean
but
like
if
you
go
de
Miele,
it's
right
like
with
the
impliedly,
Cantonese
and
stuff
like
that.
I
just
know,
because
I'm,
not
artistic
and
I've,
tried
to
church
things
up
too
much
in
the
past
that
when
you
start
expanding
and
shrinking
things,
they
start
to
look
super
funky.
So
the
thin
letters
in
that
you
know
purple
green
and
brown,
one
or
the
one's
the
lotto,
it's
just
very
crisp
and
clean.
B
B
F
F
B
And
then
Alex
I
leap
into
your
able
hands
to
do
the
various
and
sundry
professional
machination.
They
don't
understand,
but
we
will
notice
the
difference
in
terms
of
it's
sort
of
like
I
was
that
someone
explained
to
me
that
a
really
good
suit,
unless
you're
a
professional,
the
art.
You
will
know
that
it's
a
really
good
suit,
but
you
won't
be
able
to
explain
why.
C
F
B
F
Great
we're
going
with
the
orange
one
with
that
different
font
than
the
other
font,
and
then
oh
yeah,
so
great
so
I'll
just
revise
that
and
then
create
the
fun
there
right
file
doesn't
send
that,
to
you
add
by
the
end
of
today
or
tomorrow
morning.
Does
that
work.
B
D
D
We
still
have
a
lot
of
things
on
it.
So
I
guess
that
probably
the
question
that
we
should
answer
today
is:
do
we
still
want
to
try
keep
our
plans
for
having
curly
said
cube
corn,
or
we
should
just
postpone
and
try
to
figure
out
a
proper
way
to
do
the
police
even
calling
it
alpha,
but
still
having
something
that
that
we
feel
like
the
one
to
show
to
the
public
yeah.
B
I
mean
I
particularly
like
to
hear
from
all
the
various
trips
in
the
community
who
are
in
the
trenches,
finding
and
fixing
all
these
bugs
I
think
there
may
be
more
stuff
in
the
backlog.
We
may
need
to
arrange
to
go
prune
this
board
somewhat,
because
I
think
there
may
be
some
things
in
the
backlog
that
are
in
progress
or
gone
because
I
don't
think.
B
G
G
D
B
Very
much
like
what
we
normally
do,
which
is
you
have
some
it's
something
that's
stable
and
you
you,
you
may
be
still
be
adding
tests
to
it.
You
may
be
still
finding
bugs
and
you
may
still
be
fixing
bugs.
But
the
point
is
it's
a
place
for
your
only
finding
your
increasing
testing
testing
finding
and
fixing
bugs
so
that
we
have
stable
demos
and
we're
moving
towards
a
release.
But
you
guys
I,
think
you
know
that
that
seems
more
realistic
to
me
than
trying
to
show
something
prematurely
out
the
door.
A
Yeah
I
definitely
agree
with
that,
and
my
I
would
I
I
think
is
as
much
as
it
would
be
nice
to
be
able
to
get
the
first
one
out
a
coupon.
It's
more
important
that
that
initial
release,
even
though
we
have
the
word
alpha
stuck
on
it
and
Andromeda,
is
the
first
release.
I
think
it
is
very
important
that
we
do.
A
We
hit
the
quality
bar
high
enough
that
when
we
get
people
who,
because
we
were
going
to
be
looking
at
this
very
critically,
they're
gonna
want
to
see,
does
this
work
or
not
and
they're
gonna
get
their
first
impression
based
upon
that
that
cuts,
that
branch
and
so
I
would
rather
I,
would
rather
delay
it
and
to
be
clear
this.
This
isn't
an
implication
on
on
anyone
on
the
team
or
anything
like
that
as
well,
I
mean
you
all
have
worked
basically
hard,
and
so
you
know
kudos
to
to
you
all.
A
D
Okay,
so
from
the
things
that
I'm
involved,
2:14
I
can
say
because
of
some
kind
of
fading
sea
ice
for
various
reasons,
so
we
are
pushing
forward
the
ipv6
verification.
We
have
found
some
problems
around
it
which,
which
are
I
believe
our
candle
now
I
mean
which
are
not
really
inside
inside
any
same,
and
we
also
are
having
this
thing
with
the
helm.
D
G
B
B
D
D
B
But
I.
You
know,
I've
not
yet
actually
gone
through
and
trying
to
sweep
them
to
the
right
category.
So
I
suspect
that
the
backlog
is
probably
much
smaller
than
it
seems
to
be,
because
I
suspect
a
great
many
things
have
actually
been
done
or
or
in
progress,
and
we
just
need
to
actually
get
them,
swept
to
the
right
place
and
possibly
close
those
things.
D
A
Yeah
and
this
question
is-
is
independent
from
the
actual
release
itself,
so
I
want
to
make
that
back
there.
So
it's
just
a
matter
of
like
is:
are
we
cutting
a
branch
specifically
for
Cooper
cube
corner?
Are
we
cutting
a
branch
that
is
going
to
turn
into
into
the
release,
and
so
I
want
to
get
a
differentiation
on
that
in.
A
H
So
I
have
some
good
news,
so
in
principle,
I
got
approval
for
the
meter,
so
Lumina
has
agreed
to
probably
take
care
of
the
space
as
well
as
a
bit
of
sponsorship
for
the
pizzas
and
drinks,
so
I'm
still
working
on
the
budget.
So
one
other
thing
is
the
plan
is
probably
we
can
also
expand
it
to
wherever
fishes
are,
which
is
essentially
in
India
as
well
as
in
Australia.
B
H
So
that's
update
to
meet
up,
so
only
thing
is
what
we
would
need
to
do
is
we
need
to
probably
I
mean
once
I
have
the
plan?
I
will
probably
also
see
if
we
can
create
the
meetup
group
in
Mehta
org
and
add
you
are
a
bunch
of
SS,
the
local
organizers
and,
after
that,
what
you
would
need
to
do
is
we
need
to
probably
have
to
agree
on
the
frequency
as
well
as
the
list
of
speakers
we
need.
To
probably
I
mean
this
is
very
common.
H
A
H
C
Sorry
go
ahead,
you
know
you
first
well,
I'm
gonna,
say
I
would
definitely.
Obviously
the
Bay
Area
is
a
hot
spot
for
these
kind
of
things.
So
when
you
guys
kind
of
navigate
through
some
of
those
things,
I'd
like
the
lessons,
learn
because
it'd
be
nice
to
maybe
get
something
like
that
started
up
here
in
Denver
as
well.
I'd
kind
of
would
like
to
wait,
though,
and
see
kind
of
what
works
for
you
guys
in
the
Bay
Area
and
then
try
to
replicate
it
here.
Sir,.
B
A
But
I
think
the
the
important
the
most
important
thing
that
we
do
is
we
try
to
keep
the
quality
of
the
speakers
high
and
we
maintain
a
very
common
cadence,
like
we
say,
every
second
Tuesday
of
the
of
the
month
at
6
p.m.
on
the
dot,
no
exceptions,
there's
there's
a
meet-up
and
we
just
drill
it
into
people,
and
so
people
start
to
really
care
about.
A
It
start
to
associate
that
as
their
NSM
date
tonight
or
whatever
you
want
to
set
it
up,
and
so
they
end
up
marking
those
their
calendars
off
and
they're
mine,
don't
accidently
schedule
on
top
of
it,
and
so
sorry
I'm
but
I
think
hey,
maybe
had
I
tried
to
do
to
do
this,
and
even
places
like
like
Denver
or
Houston
or
so
on.
I
would
not
have
had
the
same
response
so
so.
C
I've
been
looking
into
this
for
a
while
and
Denver
actually
has
a
pretty
thriving
meetup
space,
but
I'm
kind
of
in
this
weird
spot
of
do
I.
Take
an
SM
to
an
existing
meetup
or
you
know,
I
I
think
the
comments
about
like
trying
to
stay
cloud
native
networking
well
maybe
be
seen
as
trying
to
compete
with
other
groups
here
in
Denver.
C
So
my
other
alternative
is
to
wait
until
you
know,
maybe
January
of
next
year
kind
of
see
how
you
guys
fare
after
NSM,
as
you
know,
had
another
round
through
CN
CF,
Barcelona
and
CN
CF.
You
know
here
in
the
North
America's
again
and
San
Diego,
and
hopefully
we
have
more
exposure
there,
because
the
other
thing
I
don't
wanna
do
is
you
know,
set
up
a
meet-up
that
only
two
people
come
to.
H
Sounds
good,
so
I'll
probably
create
a
document
around
it
and
we
can
probably
add
our
thoughts.
I
also
have
a
marketing
person
have
a
Miss
Stephanie.
She
is
helping
out
on
this.
I
can
invite
to
one
of
the
calls
and
then
probably
ask
her
to
give
the
update
she's
quite
connected
she's
to
work
for
telecom
council
and
she's
well
known
in
this
marketing
fraternity.
We
can
leverage
some
of
her
experience
and
then
get
it
done.
Yep.
H
I
A
C
C
Right
obviously,
I'm
biased
and
I'm
trying
to
be
sneaky
about
this,
but
I'm,
but
you
know
like
I,
think
CNS
is
something
that's
focused
enough,
but
has
a
lot
of
people's
imaginations
captured
and
then
having
that
partnered
with
something
like
an
SM
to
say
like
now.
This
could
actually
be
a
reality
and
here's
a
way
to
do
it
maybe
be
a
way
to
get
enough
people
interested,
but
also
have
it
focused
enough
to
not
just
have
something.
That's
you
know
a
bunch
of
random
people
showing
up
and
arguing
what
the
edge
looks
like.
A
H
C
Think
partnering,
when
you
talk
about
guest
topics
with
like
the
envoy
community,
the
sto
committee,
the
link
OD
community,
to
show
them
that
we're
not
trying
to
like
you
know,
put
our
hands
in
their
kool-aid
and
also
how
things
can
be
potentially
complimentary
would
be
beneficial
to,
because
I
still
always
get
the
well.
Why
are
you
working
on
this
Jeff?
Why
don't
you
just
use
this?
Do
what
like
common
at
least
once
a
week.
C
Yes
and
I
have
similar
conversations
bit
all
you
said
if
we
show
how,
like
things,
are
complementary
and
just
once
again
now
that
we're
in
the
CNCs
like
becoming,
as
you
know,
collaborative
as
possible,
will
only
help
us.
A
Yeah
a
second
thing
we
can
do
as
well
as
at
least
here
in
the
Bay
Area.
Some
of
the
meetups
are
amenable
to
partnering
with
other
groups,
so
there
may
be
cloud
native
computing
group
out
there
that
they
want
to
on
occasions
co-host
something,
and
we
can
do
like
a
joint
joint
thing
so
and
that'll
that'll
help
pick
up
our
visibility
in
those
communities
as
well.
H
A
So
it's
a
way
to
work
out
which
one
of
those
are
is
to
look
at,
which
ones
have
done
that
in
the
past,
with
other
groups
and
there's
a
few
of
them
out
there
and
basically
we
want.
We
have
to
find
one,
that's
large
enough
that
that
they
don't
feel
threatened
with
with
a
group
who
and
basically
say,
are
we
building
our
competition
and
I.
Think
we're
we're
segmented
enough
that
we
shouldn't
run
into
that
problem
too
much.
H
D
It's
just
I
mean
we
don't
have
the
time
to
go
into
the
details
here,
but
I
think
that
we
can
at
least
open,
as
someone
mentioned
a
couple
of
minutes
before
the
kind
of
wrongs
here.
But
what
I
see
is
that
we
have
this
nice
use
cases
call
where
a
lot
of
things
are
discussed,
and
then
we
have
the
CNF
test
bed
which
we
haven't
even
touched
from
what
I
know
from
the
NSM
point
of
view,
and
then
we
have
our
demos
and
our
CIA
and
I
kind
of
see
all
these
things
going.
D
I
mean
you
mean
we
spend
out
of
time
talking
about
these
things
planning,
but
then
we
somehow
fail
to
put
the
the
attention
into
animal
implementing
at
least
some
of
the
ideas
we
start
implementing
or
having
some
some
basic
I
don't
know
code
going
in
that
direction.
I
mean
understand
timing.
This
is
the
state
of
the
project,
I'm,
not
saying
that
we
should
change
something
tomorrow
or
whatever,
but
maybe
at
some
point
we
have
to
figure
out
how
we
want
to
approach
all
these
because
they
are
essentially
different
direction.
D
B
Things
and
so
effectively,
you
know
if
we
have
people
who
are
interested
in
say
doing
putting
together
a
new
demo
or,
for
example,
it's
unit
test
bed
stuff.
If
some
of
us
would
be
willing
to
stand
up
and
sort
of
mentor
them
through
the
process,
I
think
that
that
would
probably
make
it
much
less
daunting
your
people
to
pick
up
doing
some
of
those
things.
D
B
My
experience
with
mentoring
people
is
that
it's
incredibly
low
cost,
because
women's
of
usually
happening
is
someone
will
go
and
grab
a
problem
and
they
need
a
little
guidance
on
how
to
get
started,
see
I'm
a
few
bars
and
they
go
off.
Make
you
a
thing
and
I
think
it's
stuck
someplace,
and
so
they
come
back
and
they
say:
hey
I'm
stuck
here
and
usually
the
place.
B
They're
stuck
is
really
easy
for
you
to
get
the
month's
talk,
and
so
you
save
them
a
week
of
bumping
around
in
the
in
the
dark
by
sort
of
saying,
oh,
oh,
go
try
that
and
they
you
know
you
do
a
couple
rounds
of
that
and
then
half
a
day,
they're
moving
forward
again,
and
so
that
ends
up
being
sort
of
a
really
good
way
to
help
people
get
moving.
Okay,.
D
D
D
Founding
like
it's
it's,
it
should
be
part
of
the
foundation
of
the
cloud
or
next-generation
quality.
If
you
want
to
call
it
like
this,
but
then
you
can
build
a
lot
of
things
on
top
of
that
right
and
what
we
are
building
today
is
very,
very,
very
simple
and
basic,
which
is
OK
right.
I
mean
that's
nuts.
D
What
we
do
now
the
thing
is
how
we
want
to
move
further
I
mean:
do
we
want
to
start
putting
some
effort
into
the
whatever
is
discussed
in
the
use
cases
like
telco
type
of
things,
how
we
want
to
tackle
the
problems?
If
you
want
with
the
with
a
public
cloud
item
in
because
today
we're
rolling
case
overlay,
how
we
are
going
to
integrate?
D
B
D
I
B
B
B
Well,
we
have
these
five
things
that
we
care
about,
and
if
we
cut
these
two,
we
can
pour
resources
into
those
three
open
source
right,
it's
sort
of
a
broader,
an
inclusive
thing,
but
at
the
same
time,
if
you've
been
sort
of
say,
ok
here
is
the
bogie
list
of
things
that
people
feel
are
important,
and
so
they
give
us
a
sort
of
a
thing
that
people
can
turn
to
for
stuff.
They
could
work
on.
If
it
interest
is
them
interest
and
in
that
can
be
very
productive.
I
Know
exactly
so,
maybe
perhaps
we
could
sort
of
maybe
kick
this
off
at
a
use
case
call
like
basically
start
putting
together.
I
mean
even
a
last
use
case
because
they're
trying
to
get
into
the
details
on
specifics
and
breaking
down
further
right,
the
use
cases,
and
then,
of
course
everything
is
interesting,
but
from
execution
standpoint,
where
is
the
most
community
interest?
Likewise,
we
could
factor
reason
and
then
sort
of
you
know
can
be
parts
around
it
together.
B
B
Six
support
it
should
be
fairly
easy,
but
I
suspect
that
for
many
of
the
cases
they'll
be
sort
of
a
set
of
oh
we're
gonna
meet
this,
and
then
we
have
it
on
a
road
map,
but
maybe
we
want
to
accelerate
it.