►
From YouTube: Network Service Mesh Meeting - 2019-04-30
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
B
C
A
C
D
Hey
guys
yeah
looks
like
Lee
has
been
talking
with
pram
and
edy,
so
you
know
so
when
you
have
a
chance,
like
you
know,
let
me
know
and
I
can
actually
introduce
the
project.
It's
a
it's
a
you
know,
it's
of
course,
like
you
know
it's
very
early
stages,
but
but
definitely
like
you
know
we're
trying
to
actually
take
it
through,
not
sure
like
if
you
guys
have
some
background,
but
you
know
so
when
I
mean,
as
you
guys
go
through
the
agenda
and
I,
can
you
guys
give
me
a
spot?
I
can.
E
B
A
A
It's
possible
we'll
get
to
this,
though,
so
we
have
cube
Connie
you
showing
up
with
in
Barcelona,
and
so
that
is
from
May
21st
to
23rd.
We
have
an
intro
in
deep
dive
and
there
is
a
potential
elephan
demo
booth
that
we
may
end
up
doing
in
conjunction
with
ODL
I,
believe
we
have
on
May
20th
the
day
before
we
have
a
few
co-located
events
at
cue,
Connie
as
well.
So
we
have
the.
We
have
talks
in
the
phyto
many
summit
and
we
have
talks
and
the
cloud
native
network
service
day.
A
It's
same
at
the
exact
same
time,
we
also
have
cube
con
North
America
in
San
Diego,
where
the
call
for
papers
opens
up
in
a
few
days
and
in
May
6.
So
if
you
have
anything
that
you
would
that
you
would
like
added
to
this
to
this
events
list
it
definitely
definitely
let
us
know
and
we'll
we'll
add
it
in
there
and
with
that.
E
As
folks
actually
get
things
lined
up
for
events,
please
feel
free
to
push
a
few
are
to
the
site,
particularly
on
the
upcoming
cube
kind
of
you.
So
I
know
quite
a
few
looking
to
place.
You
know
booth
demos,
talks,
we've
got
the
general
talks,
we've
got
toxic
of
located
events
and
and
so
just
feel
free
to
push
your
PR
and
add
whatever
you're
doing
related
to
network
service
mesh
there
to
the
site.
A
Yeah,
so
one
of
the
nice
things
about
our
site
is
that,
if
you
add
the
metadata
with
the
time
in
in
the
date,
it'll
it'll
order
it
properly
in
the
events
page
and
it'll
migrate
it
to
past
events
once
it's
a
past
event.
So
it's
super
easy
to
add
something
in
it's
just
copy
over
one
of
the
templates
modify
it
then
you're
good
to
go.
B
B
Then
I
mean
it's
a
really
big
thing
now,
so
we
had
kind
of
discussion
with
Fred
and
that
last
last
week
and
we
agreed
together
that
we
would
like
to
hear
a
person
from
the
community
promoted
to
join
us
with
within
the
core
contributors,
like
the
guys
that
have
write
access
to
the
to
the
rapist
to
the
main
report
for
service
mission,
the
site
and
this
person's
name
is
under
a
Soloff.
So
he
contributed
more
than
40
patches.
Lately
he's
a
great
supporter
of
the
stability
of
the
CI.
B
E
In
principle,
exceptions
like
places
with
their
credit
card
numbers
I
am
delighted.
I
want
to
make
sure
that
we
all
leave
the
committers
have
equal
access
to
the
resources
for
a
management
point
of
view.
It's
just
generally
good
policy,
but
I'm
not
sure
that
I
remember
what
all
those
things
are.
So
if
you
come
across
one
that
I
have
an
attitude
to
you.
Just
let
me
know
yeah.
A
A
F
You,
since
last
week,
we've
gained
42
followers
we're
now
at
157
and
also
since
last
week.
We
are
following
254
and
more
accounts
than
last
week,
so
we
have
623
we're
following
623
and
I'm,
keeping
an
eye
on
the
events
to
post
about
Kubik
on
you,
as
well
as
FBI
mini
summit
and
the
elephant
cloud
native
services
day,
I'll
post
about
those
to
let
everyone
know
where
we'll
be
at
that
up
the
events
and
once
the
release
notes
are
ready
for
the
Andromeda
release.
I
will
also
post
an
announcement
for
that.
E
Two
hundred
dollars,
two
things
number
one:
that's
a
hell
of
a
lot
of
work
to
make
that
much
progress
in
one
week.
So
thank
you
so
much,
but
number
two
you've
also
been
posting
like
all
kinds
of
NSM
related
things
like
I
think
there
was
a
bit
that
the
pre-med
done
on
Telecom
TV,
the
tush
run
and
SM,
and
you
posted
that,
and
so
you
just
had
an
exceptionally
good
job
of
tracking
all
the
things
I
really
appreciate
it.
Thanks.
B
G
B
Tweet,
it's
about
okay!
Let
me
just
click
here
and
I,
guess
that
we
open
so
yeah
I
mean
we
live
in
a
free
world.
Everyone
can
have
his
or
her
opinion
open
shared.
So
my
question
is
how,
in
general,
do
we
handle
all
the
positive
and
you
know,
I'm
kind
of
not
really
positive
messages
that
are
out
there,
especially
with
Twitter,
so
I
see
not
fret
answer
here.
So
there
is
a
kind
of
small
discussion
going
on,
but
do
we
want
to
hear
for
more
structured
approach
to
this
or.
H
B
My
question
was
more
about
because
I
helped
I
I,
guess
that
I
mean
you
probably
mean
if
you
need
some
help
or
you
think
that
there
is
a
some
messaging
that
we
can.
We
can,
you
know,
send
even
not
directly
as
a
reply
to
someone
else,
but
maybe
as
a
kind
of
final
statement
from
from
us
as
a
community.
Yes,
it
has
only
indirect
apply
to
something.
I,
don't
know.
Maybe
maybe
you
can
just
reach
out
to
us
and
we
can
try
to
figure
out
that.
F
B
A
You
know,
I
and,
and
so
I
I
think
like
one
of
the
worst
things
we
could
do
is
is
to
not
be
kind,
because
then
that
not
only
does
that
not
help
the
other
person
and
or
help
us,
but
actually
destroys
our
reputation
as
well,
and
so
so
I
think
for
me
that
that's
the
best,
the
primary
like,
regardless
to
what
form
or
structure
that
that
it
has.
You
know,
that's
that's
the
for
me
that
that's
paramount
that
we
did.
A
We
act
like
that
when
it
comes
to
something
like
this
I
think
that
once
the
once,
the
website
is
a
little
bit
more
structured,
then
something
that
Lucina
could
help
us
with
is
trying
to
identify
the
list
of
very
frequently
asked
questions
you
know
and
when
we
can
point
people
at
them
just
to
seat
the
seat.
The
conversation
and
it's
not
enough
for
them,
then
we
can
then
we
and
then
we
can
further
join
in
on
the
conversation.
A
E
A
C
And
I
just
follow
up
one
other
thing
that
I
always
get
is
somehow
people
assume
that
I
mean
it's
writen.
They
assumed
that
this
is
also
service.
Mesh
simulit
with
you
so
I
tell
them
that
head
is
not
the
regular
service
mesh
and
then
explained
them.
Then
people
there
wall
effect
goes
further.
That's
the
feedback.
I
got
from
people
hey.
This
looks
interesting
in
that.
C
B
A
D
Hey
guys
so
I
first
have
to
apologize
for
the
fact
that,
like
I
haven't
been
much
time
with
NSM,
you
know
so
I
just
like
what
frame
was
mentioning.
You
know,
I
pretty
much
like
you
know,
it
would
be
pointing
people
at
the
same
set
of
questions
like
what
you
guys
just
made,
not
of
but
definitely
like
you
know.
We
would
be
very
interested
to
learn
about
it
now.
D
The
next
question
keeps
coming
up
as
to
why
this
service,
much
as
to
that
and
like
you
know,
what's
the
overhead
that
it
keeps
bringing
up
like
you
know.
What
are
the
performance
you
know
I
mean
like
I
know.
Nothing
comes
for
free
right.
People
understand
that
very
well,
so
there
are
some
of
the
immediate
questions
that
keep
coming
up
are
like.
So
you
know,
what's
the
overhead
of
you
know
bringing
in
sto?
D
What's
the
overhead
of
bringing
a
link
or
D
like
you
know,
and
and
they
also
get
into
some
specifics,
so
what
we
did
was
like.
You
know
me
and
Lee
started
working
on
this
project
like
not
just
like
you
know
a
few
months
back
and
we
thought
it'll
actually
be
a
nice
community
project,
so
we
actually
created
this
mesh.
We
thing,
and
now
we
also
have
weekly
meetings.
I've
posted
a
link
to
that.
D
Like
you
know,
of
course,
like
anybody's
are
actually
seeing
at
the
site
like
in
which
is
fantastic,
so
we
have
a
few
contributors
now
just
to
give
a
year
of
what
Mary
is
like,
so
so
make
sure
we
tries
to
address
or
helps
you
know
or
is
trying
to
help
address
some
of
these
overhead
concerns
like
in
which
people
have
been
bringing
up
so
at
the
moment.
Looking
you
can
see,
there's
not
much
because
again,
like
everything
is
pre-alpha.
D
Now,
one
of
the
first
things
we
did
was
we
started
off
with,
like
you
know,
a
simple
proof
of
concept
like
you
know,
which
turned
out
to
be
a
monolith
and
then,
like
you
know,
we
started
bringing
it
out
into
you,
know:
I,
wouldn't
call
it
like
in
our
pure
micro-service
architecture,
but
then,
like
you,
know,
into
its
own
confidence
and
we
thought
rather
than
having
the
adapters.
You
know
so,
I'm,
really
sorry.
So
when
we
started,
we
actually
started
it
off
as
a
project
forest
feel,
and
then
we
thought
hey.
D
You
know
what
this
is
a
sensible
question
that
needs
to
be
addressed
across
meshes.
So
what
we
did
was
like
and
we
thought
measure
e
could
be
a
thing
which
interfaces
with
this
concept
of
adaptors,
which
again,
like
you
know
this
is
you
can
say
I
you
know
I
borrowed
the
the
terminology
from
SEO,
but
essentially
what
it
is
is
that
measure
e
itself.
It
does
not
talk
to
any
meshes
any
so
specialist.
So
what
we
did
was
we
created.
This
concept
called
mystery
adapters.
D
Now
we
we
are
planning
to
have
one
adapter
for
every
type
of
mesh
and
you
can
see
it
there,
like
you
know,
there's
one
phrase:
do
there's
one
for
linker,
D
octarine
is
still
being
developed
and
there
is
like
a
skeleton
one
for
console
Connect.
So
essentially
the
plan
is
that,
like
you
know,
all
these
meshes
will
actually
have
diapers
and
measure
will
actually
interface
with
them.
Now,
why
should
measure
E,
which
is
supposedly
a
performance
tool
at
this
point
in
time?
D
Oh
and,
like
you
know,
try
out,
like
you
know,
testing
tools
like
not
to
see
if
you
know
it
performs
well,
so
you
know
as
you
you
can
see
like
you
know,
this
is
kind
of
like
you
know
where
the
project
is
it's,
it's
I
wouldn't
call
it
beta
like
I.
Would
say
it's
pre
alpha.
It's
been
like
three
months
roughly,
like
you
know,
we
started
this
project
it
and
now
we
try
to
come
up
with
this,
a
weekly
meeting,
which
is
at
2:00
p.m.
central
on
Fridays.
D
So
if
you
guys
are
available,
but
you
know,
please
feel
free
to
join
and
also
there
is
a
slack
channel.
You
know
I
have
a
link
to
that
and
you
you
will
be
in
the
spy
Channel.
All
the
code
is
open
source
like
and
everything
is
under
the
layer,
5
io
org
on
github.
So
again,
like
you
know,
right
now,
I
am
one
of
the
main
contributors
to
the
project
I'm
here
to
actually
talk
about
it.
D
This
call,
like
you,
know,
I'm,
definitely
intrigued
to
actually
learn
more
I
mean
it
started
like
I
wasn't
intrigued
earlier
and
and
by
the
way
I'm.
Also
one
of
the
Twitter
followers
of
NSM
now
so
I'm.
Definitely
intrigued.
I'm
very
intrigued
to
actually
know
more
about
NSM,
and
so
hopefully
you
know.
D
Okay,
so
yeah
so
I
mean
eventually,
like
you
know,
we
want
to
actually
make
my
sri
in
a
way
that,
like
you
know
it
will
I
love
you
to
actually
perform
a
wide
variety
of
tests.
But
today,
like
I
mean
like
I
said,
like
you
know,
we
just
started
like
you
know,
working
on
this
mastery
project,
it's
pretty
new.
So
to
start
with,
you
know
we
actually
have
a
very
basic
load
generator,
but
I
mean
out
of
the
box.
D
Today
you
will
be
able
to
actually
see
that
we
are
I
mean,
like
you
know,
with
respect
to
the
technology
stack
like
you
know,
I'm
I'm,
a
fan
of
go.
You
know
from
Google,
so
pretty
much
like
an
all
the
backend
and
the
adapters
at
the
moment,
or
actually
all
written
in
though,
but
the
adapters
don't
have
to
be
written.
Go
like
you
know,
because
you
can
see
that
the
connection
is
over
G
RPC.
D
So
that's
the
technology
stack
and
at
the
moment,
like
you
know,
we
have
a
close
dependency
on
kubernetes,
like
which
we
will
eventually
be
moving
out
because,
for
example,
when
we
spoke
to
a
council
folks,
like
you
know,
they
actually
said
that
you
know
they
give
as
a
percentage
that,
like
hey,
you
know
what
only
30%
like
enough
people
they
know
are
on
corn
eighties.
So
the
remaining
70
are
not
so
you
know
so,
which
is
actually
a
definitely
a
good
thing
for
us
to
consider.
D
So
we
will
eventually
be
moving
away
from,
like
you
know,
being
specific
to
cover
names.
But
again
you
know
the
measure
II
app
itself
does
not
have
that
binding.
It's
mostly
the
adapter,
so
you
know
as
an
adapter
developer
like
you
know,
the
person
will
have
complete
independence
as
to
the
way
they
want
it.
Just
that,
like
you
know,
I'll
have
you
know
we'll
have
to
make
some
minor
changes
in
the
user
interface
to
actually
remove
the
tight
dependence
and
kubernetes,
but
but
yeah
other
than
that.
You
are
free
to
actually
try
my
sri.
D
E
A
So
it's
it's
very
easy
for
us
to
to
to
work
with
things
that
are
not
that
are
not
kubernetes
based.
So
something
like
this,
like
the
fact
that
you've
thought
about
it
not
being
tied
directly
to
kubernetes
and
you're
looking
at
pulling
it
out,
so
it
can
run
as
a
standalone,
orinsa
orinsa
mother
system
is
is
definitely
is
definitely
valuable
to
us
as
well.
So
awesome.
D
E
D
A
A
D
A
C
A
G
Oh,
can
you
eat,
can
you
me
okay?
Yes,
great,
that's
the
second
latest
I'll
I
sent
another
one.
Yesterday.
G
Sorry
about
that
that
was
that
was
a
second
last
one.
I
will
send
the
link.
G
E
No
people
seem
a
lot
like
we've
had
a
lot
of
people
talking
about
like
ruin,
come
row,
one
column,
I'm,
sorry,
row,
two
column,
two
people
really
like
the
gradient
there.
We
had
people
comment
on
row,
four
column,
one
row
column,
one
people
really
liking
the
sort
of
heptagon
inside
the
buckyball
and-
and
I
think
that
we
also
have
a
super
interesting
comment
there.
Where
someone
was
like.
Yes,
it
would
be
great
if
we
could
get
heptagonal
faces
on
the
buckyball
and
and
I'm
sort
of
like
thinking
through
my
platonic
solids
going.
Do
we
have.
C
A
Well,
the
only
thing
that
concerns
me
with
the
with
the
two
line,
one
versus
the
three
line,
one
like
I-
think
the
two
line.
One
definitely
looks
better.
The
only
thing
that
concerns
me
is
when
I
talk
with
people
and
I
say
we
need
to
think
of
network
service
and
I
should
think
of
it
as
a
mesh
of
network
services.
So
you
matters,
but
like
the
network
service
in
quotes,
because
service
mesh
itself
is
a,
it
has
a
defined
meaning
and
a
lot
of
people.
A
F
A
E
G
Definitely
so
maybe
I
should
I'll
do
another
board,
obviously,
and
I
think
that
I
mentioned
in
the
email
thread
that
I
was.
You
tried
some
other
colors
too
I've
been
just
using
blue
and
green,
because
it's
just
easier
to
design
quickly
with
the
same
color.
So
maybe
I'll
try
some
other
colors
and
I'll
focus
based
just
on
the
the
ones
that
were
selected.
I.
Think
in
the
thread
so
far
were
row,
two
column,
one
that
was
one
of
them
and
then
wrote
too
long
to
write
those
two
I
think
and
then
wrote
for
column.
One.
G
So
those
three
were
the
ones
that
are
the
only
ones
I
saw
selected
yet,
but
if
obviously,
if
there's
any
others
so
so
I'll
just
focus
on
those
three
and
then
make
the
lines
thicker
make
sure
that
the
text
is
as
we
spoke
about
and
then
try
some
other
colors.
Is
that
and
then
I'll
make
them
I'll
show
them
at
small
scale
as
well.
Thank.
E
G
E
E
J
Hi
everyone,
the
first
thing
is
I,
would
like
to
introduce
myself
because
I'm
a
new
to
the
project
and
I'm
Bart
and
I'm
working
for
VMware
daily
basis
in
the
part
of
the
Cuban
net
esteem.
But
I
have
some
free
time
and
I
decided
that
I,
like
the
project
and
I,
would
like
to
be
involved
a
little
bit
more
so
I.
J
So
far,
I
tried
and
I'm
trying
to
trying
to
help
with
CI
CD
and
during
this
period
of
time,
I
saw
that
one
of
the
things
which
is
needed
to
the
CII
certification
is
the
code
style
best
practices
and
I.
Think
that,
especially
from
my
expertise
after
this
initial
period
of
time,
where
you
are
doing
everything
fast
and
a
lot
of
features,
etc.
It's
always
good
to
have
kind
of
stabilization
in
a
way
of
this
style.
Guide
of
of
the
code.
J
So
I
would
like
to
propose
to
create
a
kind
of
document,
and
I
would
like
to
be
kind
of.
I
can
be
the
owner
of
it
to
to
moderate
that
kind
of
discussion
about
the
code
style
guide
of
the
document,
which
will
help,
for
example,
to
do
proper
code
reviews
and
to
point
the
people
if
there
would
be
something,
maybe
not
easy
to
explain
to
the
exactly
place
in
the
document
where,
where
are
our
generic
rules
about
cold
style
guide?
J
B
My
how
we
see
here
plus
on
from
it
already
I,
mean
I,
don't
think
that
this
anyone
who
would
minus
one
such
such
a
proposal,
my
question,
would
be:
are
there
any
tools
that
can
help
us
in
that
like
to
check
these
things
for
us?
Because,
okay,
you
said
that
we
can
leverage
this
during
the
reviews.
We
can
point
out
to
it
and
say:
okay,
we
agree
that
this
is
the.
This
is
going
to
work
like
this
etcetera
etcetera,
but
it
would
be
always
much
more
helpful
if
there
is
some
automation
that
can
actually.
J
So
the
problem
is
I,
I
think
that
goal
is
very
opinionated
and
it
is
much
easier
to
maybe
differently.
We
don't
have
to
create
a
document
for
the
things
which
call
already
solve
via
go
formatter
or
something,
but
there
are
places
which
are
not
covered
that
much
and
I'm,
mostly
thinking
to
put
inside
this
document
the
things
which
can't
be
easily
really
automated
by
our
tools.
J
I
agree
and
I
think
even
the
conversation
about
it
is
really
valuable
to
have
like
a
one
place,
for
example
github
issue
or
the
document
where
we
can
discuss
that
kind
of
thing,
because
I
don't
have
a
special
kind
of
automation,
tool
in
mind
right
now.
But
this
is
a
good
opportunity
to
do
a
research
for
that
and
if
there
will
be
a
for
example,
places
which
we
want
to
include
in
that
kind
of
document
to
then
find
a
proper
tools
which
can
help
with
that.
E
E
B
J
B
So
the
first
point
here
is
that
we
obviously
missed
the
deadline
for
branch
Inc
for
creating
the
branch.
Last
week
there
was
some
back
and
forth
within
the
CIA
enabling
disabling
parts
of
it
and
I
believe
that
we
are
kind
of
getting
where
we
want
to
be,
or
at
least
we
have
peers
lined
up
to
to
to
get
us
there.
E
Because
my
personal
sense
is
that
I
would
rather
get
a
good,
solid
release
out
of
the
week
after
cube
con
than
shaky
release
out
the
week
before
now
that
that
that's
sort
of
my
my
top
priority
concern
that
said,
I
would
really
like
to
get
a
really
subtle
before
cube
con.
So
that
would
be
my
secondary
priority.
There
does
that
sort
of
match
other
people's
feelings
and
opinions.
F
E
I
mean
I
said
it
would
be
nice
to
get
CIN
CIN
CIN
place
for
some
of
the
demonstrations
as
well,
just
that
they,
okay
in
the
program
I
get
abused
because
I'll
go
to
stand
up
the
demo,
because
I
was
broken,
even
though
all
the
functionality
is
working
beautifully
because
I've
got
the
integration,
just
improve
it,
and
it's
just
silly
things
like.
Oh,
the
fix
for
the
namespace
move,
which
was
a
really
good
idea,
didn't
quite
make
it
into
the
helm.
Chart
that
had
us
do
things.
Yeah.
B
So
that's
actually,
where
Bart
started
I
mean
with
adding
home
deployment
into
the
CI
to
actually
somehow
verify
this
yeah,
but
actually.
J
With
that,
it's
almost
finished,
but
it's
the
kind
of
discussion
about
style,
guide,
etc,
was
because
I
faced
few
problems
related
to,
and
especially
middle
pointer
dereferences
in,
like
at
four
or
five
places,
which
may
make
like
kind
of
race
condition.
In
so
you
had
to
specifically
have
an
order
which,
which.
J
B
Mean
I
mean
that
we
are,
we
are
on
track.
There
I
mean
like
we
are
moving
in
that
direction.
We
have
a
plan,
we
have
someone
working
on
it
so
once
this
is
kind
of
stable
and
so
in
then
yeah,
but
otherwise,
yeah
I
do
I
do
agree
that
that
maybe
even
even
just
giving
a
branch
where
we
keep
things
kind
of
stable,
--is--,
even
without
really
releasing
it
I
can
still
be
still
be
a
better
approach.
Yeah.
E
I
tend
to
personally
really
like
the
Masters.
Yo
master
is
where
the
development
happens,
and
then
you
pull
a
throttle
and
you
do
do
bug
fixing,
but
the
throttles
aren't
really
going
anywhere,
because
that
way
master
can
mostly
stay
open
most
of
the
time,
and
you
can
really
get
the
fist
a
bolt.
The
throttle
is
really
nice
and
stabilized.
B
G
E
C
C
The
second
thing
is
also
what
I
was
thinking
was
at
least
if
we
can
start
exposing
the
api's
like
an
EPA
guide,
wherein
if
someone
want
to
try
it
out
with
the
CNF
developers
or
anyone
who
wants
to
do
an
e
and
a
some
type
of
thing
I
was
thinking.
It
would
be
good
so
that,
as
part
of
the
meetup,
we
can
also
make
it
more
interactive
in
terms
of
having
hands-on
lab
similar
to
what
Fred
does.