►
From YouTube: Network Service Mesh Meeting - 2019-05-08
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
C
C
Nick,
why
am
I
are
going
to
work
on
documentation
for
I?
Guess
you
say,
updating
and
adding
links
to
the
QuickStart
guide
and
see
we
can
we're
trying
to
come
up
with
a
guide
for
basically
idea
for
where
I
see
enough
to
fill
up
or
for
our
talk,
and
we
want
to
try
to
work
out
how
how
to
tie
that
into
a
into
a
presentation
and
also
into
into
a
future
into
a
future
QuickStart
cool.
C
Cool
and
Watson
also
has
either
volunteered
or
been
volunteered
to
work
on
videos,
and
so
the
videos
are
basically
like
this
is
he
says:
Watson
would
like
to
go
through
and
install
and
write
a
concise
that
have
been
salt
steps
that
would
be
covered
in
the
script,
which
are
four
steps
operator:
persona
building
with
the
creepy
old
containers
deployed
with
helm.
The
second
one
would
be
client
connecting
to
an
SM
basic
version,
basically
like
kernel,
using
kernel
connectivity
to
start
off
with
in
the
future
another
one
where
they
use
a
shared
memory.
Client.
C
Third,
one
is
a
cmf
developer,
building
something
simple
for
NSM
and
the
fourth
one
is
something
for,
and
it
some
developers
believe
those
are,
the
four
personas
and
a
draft
script
for
what
MSM
is
and
how
do
you
use
it
for
different
personas
and
and
creating
a
video
based
upon
script
and
the
week
before
we
end
up
talking
them
about
release,
notes
the
what
is
MSM
on
the
on
the
website.
So
what
does
MSM
needs
to
be
completely
changed
or
or
modified?
There's
actually
two
things.
One
of
them
is
the.
What
is
MSM
document?
C
C
A
C
Luke
Perkins,
okay,
cool,
so
Luke
Perkins
is
he's
doing
an
initial
draft
and
then
he's
gonna.
He's
gonna.
Tell
me
when
he's
ready
to
to
show
off
a
a
draft
and
I've
had
some
detailed
conversations
with
him
about
of
what
we're
trying
to
do
with
MSM
and
how
we're
trying
to
onboard
people
and
personas
and
so
on.
So
he
should
be.
B
D
C
So
Luke
should
have
access
to
the
new
logo
so
once
it
becomes
available,
he's
gonna
slurp
it
up
and
then
he'll
design
the
website.
Based
on
that.
But.
C
B
D
A
Was
talking
into
Miu,
so
yeah,
we
just
need
to
make
sure
we're
tracking
this
stuff.
Fredrik
I'll
help
you
with
some
of
the
what
is
an
SM
stuff
like
just
put
some
words
on
paper
for
you
that
we
can
then
start
to
attack.
Do
you
or
Nikolai
want
to
share?
Now,
though,
and
we
can
start
looking
at
them
release
notes
and
going
through
those.
C
C
All
right,
I've
shared
the
document
as
well.
Can
it
can
everyone
see
yeah,
yep
cool,
so
first
from
the
from
the
table
of
contents?
Does
this
look
like
a
reasonable
order
like
does
this
convey
the
information
that
we
that
we
want,
or
is
there
any
other
information
that
we
want
to
that?
We
want
to
add
in
I.
A
C
A
C
A
C
B
D
B
B
E
C
B
C
Some
context
I
skipped
ahead
as
well,
because
I
feel
like
this
is
the
area
highest
risk,
but
for
context.
So
I
was
saying
what's
new
and
rather
than
just
this,
a
bunch
of
old
requests,
I
thought!
Well,
we'll
do
it
specially,
for
this
release
will
do
it
more
structure,
more
structured
and
say:
network
service
mesh
is
brand
new,
explaining
what
it
is.
The
API
itself
saying
that
and
just
giving
a
quick
description
what
it
is
and
then
calling
the
stuff
that
we've
built
at
the
moment.
E
Api
database
as
its
database,
it
might
be
better
to
focus
on
the
api,
is
particularly
the
EP
is
that
are
broadly
applicable
right.
So,
for
example,
the
api
the
network
service
manager
uses
to
talk
to
the
network
service
pasadena'
playing
that
api
is
sort
of
specific
to
the
fact
we're
running
in
kubernetes.
E
You
know,
you've
been
certainly
really
different
in
every
service.
Most
data
play,
but
it's
probably
not
what
you
would
choose
to
do
if
you
were
trying
to
drive
a
physical
network,
but
some
of
the
api
is
like
the
remote
network
service,
api
and
the
registry
api.
Those
things
I
think
are
generally
applicable,
and
so
maybe
worth
calling
those
out,
and
some
of
those
like
the
registry
API,
does
return
information
that
is
effectively
equivalent
to
some
of
these
resources.
E
Yeah
I
mean
particularly
the
these
sort
of
globally
applicable
api's,
because
the
thing
is
I
can't
really
see
people
confused
about
is
there
are
some
api's
like
I
mentioned
the
network
service
manager
to
SMB
they're
kind
of
locally
applicable.
You
can
totally
do
it
differently
if
you
wanted
to
and
interoperate
just
fine
with
the
broader
network
service
mash.
But
if
you
don't
implement
remote
dot
network
service
and
you
don't
implement
the
registry,
API
is
you're.
Gonna
have
a
super
hard
time
getting
along
with
the
rest
of
the
network
service
mesh.
C
Okay,
so
I
think
there's
two
things
that
we
need
to
do
in
the
scenario
then
first
is
call
a
specific
API
is
globally
applicable
and
the
second
to
do
sounds
like
we
need
to
create
a
visualization
as
well
showing
like
where
they,
where
they
fit
in
in
those
global
AP,
is
that
it
shows
only
the
AP
ice.
They
don't
have
to
show
anything
else
and
that
way
that
people,
when
they
see
this
they'll,
say
okay,
so
that's
across
connect,
or
rather
that's
the
MS
and
I
know
Sam
and
and
so
on.
I
think.
A
It's
important
to
start
with
the
global
API
like
paradigm
as
well,
and
then
specifically
calling
out
the
kubernetes
is
just
the
first
reference
architecture
we're
working
on.
So
if
someone
who's
like
an
e
CS
wizard,
you
know,
and
they
do
all
their
stuff
in
Amazon,
and
they
like
that
container.
You
know
scheduler,
they
can
come
in
here
or
you
know,
if
you're,
a
nomad
guy
like
we've,
positioned
it
as
a
tool
that
can
be
used
across
multiple.
A
E
Think
that's
definitely
a
very
important
because
a
lot
of
people,
because
a
lot
of
people
ability
if
I,
had
a
nickel
for
every
time.
It
said
somebody
said
it.
This
is
awesome.
Could
I
have
this
four
beams
and
physical
networks
only
to
say
look
that
actually
is
the
way
the
architecture
is
built.
We
focusing
on
the
invitation.
A
Exactly
so,
we
just
need
to
make
sure
that
that's
it's
apparent
that
this
is
like
a
reference
architecture
and
we
just
started
with
kubernetes
right.
Cuz,
I'm
I
get
the
same
things,
and
we've
got
like
this
small
group
of
people
who
like
to
use
nomad
for
whatever
reason
and
trying
to
like
you
know,
sell
them
on
the
idea
that
this
isn't
just
a
kubernetes
tool,
especially
since
we
put
the
heptagon
right
inside
of
our
our
logo.
We
might
want
to.
E
A
E
A
C
C
E
C
Okay,
so
so
for
so,
this
will
I'm
going
to
rewrite
portions
of
this
to
emphasize
that
the
this
is
a
reference.
Kubernetes
architecture
and
described
I'll
describe
what
it
is
and
describe
that.
That's
the
that
there's
other
architectures
that
can
fit
in
with
kubernetes
is
not
a
required
component
of
network
service
Mash.
C
C
C
C
C
A
A
C
A
Speaking
of
Twitter
and
stuff
too,
we
should
have
them
something
in
the
getting
started
guide.
That's
good
like
not
just
the
technicals,
but
since
this
is
the
very
first
release
like
a
small
little
section
of
different
community
resources,.
A
Like
getting
started,
you
know
we're.
We
have
like
the
getting
started
guide,
which
is
the
technicals,
but
there
should
also
just
be
like
as
part
of
the
getting
started
a
like.
This
is
how
you
get
involved
with
the
community.
Here's
our
Twitter,
here's,
our
this.
You
know
check
out
our
calendar
for
weekly
things,
but
yeah
like
exactly
I,
had
to
get
involved
perfect,
just
because
it's
the
very
first
release
right.
So
when
people
look
at
these
release
notes
and
like
oh,
this
is
interesting.
You
know
how
do
I
get
involved
in
this.
C
A
That's
important
because
you
know
as
different
reference
architectures
come
out.
You
know
there's
going
to
be
different
types
of
those
four
personas
you
were
talking
about
that
Watson
is
going
to
help
us
engage
as
different
personas
are
involved,
they
might
come
in
five
releases
in
and
if
we've
cut
it
out
at
that
point,
they're
gonna
be
sitting
there
trying
to
hunt
for
information
versus
it
just
being
readily
available.
C
C
So
one
of
the
things
that
I
that's
been
bothering
middle
bits
I
actually
like
the
idea
of
listing
out
as
many
people
as
we
can
remember
to
as
part
of
the
as
part
of
the
release
and
saying
these
are
the
people
who
have
helped
us
out,
but
one
of
the
problems
I
run
into
is
that
I
know
for
certain
that,
where
there
are
so
many
people
would
have
helped
us
out
that
we're
going
to
mrs.
Borden
people
and
so
I'm,
so
I'm
trying
to
work
out
like
how
to
how
to
deal
with
this.
This.
C
E
Who's
contributed
to,
and
then
the
thing
is
no
absolutely
not
just
way
more
than
is
to
do
because
we're
super
blessed
to
have
people
who
very
early
in
the
community.
Picking
up
that
are
incredibly
important
and,
and
even
things
are
not
obviously
our
community.
So,
for
example,
the
unbelievable
amount
of
support
that
the
sky
dive
team
has
provided
for
us,
I
mean
various
points
have
literally
turned
features
overnight
in
sky
dive
in
response
to,
wouldn't
it
be
great
if
we
could-
and
so
we
just
had
just
unbelievable
amounts
to
support
for
our
number
of
foes.
E
C
Yeah
that
makes
sense
to
me,
because
I
would
love
to
just
like
be
able
to
list
every
group
every
every
person
because
like
for
me,
every
person
who's
gotten
involved
in
helping
in
some
way
whether
it's
jumping
in
and
producing
code
and
landing
committer
stuff.
It's
that
it's
all
the
way
down
to
even
just
like
asking
like
you
know
what
does
this
mean
and
it
helps
us
clarify,
you
know,
even
if
it's
it,
even
if
it's
just
a
couple
questions
but
I
mean.
B
C
E
C
E
You
know
more
more
more
personal
anecdotes
of
some
of
the
contributions
of
people.
You
know
so
one
of
the
one
that
comes
immediately
to
mind
for
me,
and
this
goes
way
way
back
was
you
know
Prem
when
he
was
basically,
you
know,
literally
the
day
after
the
first
network
service
mesh
talk
was
willing
to
stand
in
for
a
brainstorming
session
in
the
phyto
cape
and
lead
that
discussion,
because
I
couldn't
make
the
time
that
everyone
else
could
make.
E
C
C
C
C
C
E
C
E
C
A
C
C
E
C
A
C
E
E
B
C
C
C
E
Would
tie
all
ways
tend
to
put
the
getting
started
at
the
top,
and
I
would
tend
to
put
the
downloads
under
getting
start
as
a
supple
it
under
getting
started,
because
we
all
know
how
engineers
think
the
first
thing
they're
going
to
want
to
know
is:
okay,
how
do
I
run
it?
They
may
not
even
bother
to
understand
what
it
is,
so
I
would
sort
of
put
getting
started.
E
You've
got
sort
of
the
I
mean
I
tend
to
think
not
so
much
in
terms
of
download
for
getting
started,
but
I
tend
to
think
in
terms
of
a
zero
day
one
day
two
right
days,
a
zero
is:
how
do
you
an
MSM,
enable
your
cluster
day?
One
is:
how
do
you
install
a
network
service
into
your
cluster
day?
E
Two
is:
how
do
you
consume
it
and
that's
sort
of
how
my
brain
works,
and
then
we
talked
about
wanting
to
have
a
list
of
demos
that
we
could
actually
point
people
to
and
I
think
under
getting
started.
We
also
want
those
demos,
so
they
they
basically
have
the
most
immediate
sense
of
oh
okay.
This
is
how
we
do
it
and
they
can
get.
You
know
the
faster
something
works
and
they
feel
good
about
it.
The
happier
they're
gonna
be.
E
C
Who
so
I
think
something
that
I
would
really
enjoy
some
help
with
on
this?
If
people
have
time
would
be
to
help
fill
this
document
up,
make
it
more
clear,
make
sure
that
there's
no
typos,
you
know
make
making
making
it
look
really.
They
can
look
really
good
and
when
we're
ready
to
do
so,
we
will
we
will
copy
it
over
to
to
github
once
the
release
branch
has
been
well
when
the
release
branch
is
being
cut
and
we'll
start
to
to
throw
in
all
of
the
information
there
for
the
for
the
release,
notes.
C
C
It
should
not
be
a
problem,
so
Geoffrey
and
I
think
and
I
can
handle
it.
So
that
will
do
is
next
week.
Will
circle
back
up
on
the
Tuesday
and
Wednesday
meetings
and
get
more
feedback
on
this?
So
I
want
to
have
a
couple
iterations
before
we
before
we
do
our
release
to
make
sure
that
it's
that
it's
clear
that
people
are
generally
happy
with
it
than
that
were
we're
good
to
go
so
I
think
we
only
have
I
think
we
only
have
five
more
minutes
left.