►
From YouTube: CNCF CNF WG Meeting - 2022-10-03
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Hello
Akash
good
morning,
Lucina.
B
C
A
Hello,
hello:
let's
wait
just
a
one
minute
more
and
we're
going
to
start
in
the
meantime.
Can
you
put
your
name
in
the
minutes?
Notes.
C
A
Cns
working
group
today
it
seems
like
we
don't
have
too
much
items
in
the
agenda,
so
I'm
going
to
use
review,
but
we
have
it
so
the
first
one
is
like
you
already
know
like
we
have
a
couple
of
events
coming
the
first
one
is,
obviously
they
keep
going
and
we
have
a
collocated
then
one
day
before
cubecon,
so.
A
And
the
rest
is
this:
like
foreign.
A
After
that,
in
Seattle
there
is
like
one
Summit,
just
like
two
weeks
after
the
cube
turn,
so
it's
pretty
close
and
and
in
the
same
place
during
the
same
time,
it's
going
to
be
there
lfm
developer
and
this
this
this
forum.
A
So
you
have
the
opportunity,
it's
a
nice
place
to
do
my
work
in
and
catch
all
the
Giants
and
and
interesting
things
that
other
colleagues
are
doing
so
I
don't
know
Lucina.
Do
you
want
to
mention
something
about
an
execute
corner
like
any
any
thinker
that
I,
usually
with
those
things
or
like,
or
it's
anything
that
I
have
missed
and
the
next
coming
events
that
you
want
to
highlight.
C
So
we
may
have
a
another
session
about
certification
and
then
for
kubecon
EU
that'll,
be
in
Amsterdam
in
April
and
see
if
fees
are
open
for
another
month
and
if
anyone
is
interested
in
sharing
with
their
companies,
the
sponsorship
form
there's
an
interest
form
available
as
Google
form
and
then
they'll
share
the
actual
perspective
for
coupon
Europe
and
any
Telco
days
this
winter.
After
a
coupon.
A
I
was
just
asking
you
know
the
the
clown
regarding
the
schedule.
The
scope
has
been
polish
some
place
or
like
I,
didn't
that
I
saw
Jeff
sharing
like
the
schedule
of
the
topics
that
is
going
to
be
present
during
the
12-minute
talk
away
in
in
Detroit.
C
Oh
half
day
event
it'll
be
in
person
only
so
if
we
want
to
have
a
live
stream
component
for
Europe,
please
share
that
sponsorship
perspective
so
that
we
can
have
a
sponsor
for
the
live
stream.
Next.
A
Okay,
well,
let's
move
forward
with
the
next:
let's
review
the
PRS,
if
you'll
have
any
UPR,
which
are
really
that's
good
news
or
bad
news,
but
please
feel
free
to
add
new
things
or
propose
new
changes.
If
you
discover
something,
I
think
they
should
have
been
pretty
much
the
same.
You
know
last
last
week
going
was
reviewing
them,
so
I
don't
want
us
back
too
much
changes
on
that.
A
Yeah,
we're
still
have
anything
new
on
this
issues.
Probably
the
only
the
only
thing
that
we
have
something
I
had
in
proposed
is
he's
just
practice.
I
don't
know
if
that
is
a
best
practice,
but
I
will
start
the
discussion
about
this
particular
topic.
So
basically,
what
I'm
proposing
is
to
support
multiple
annotations
for
different
cnis
or
cni
multiplexers,
those
dnis
which
provide
multiple
network
interfaces
per
bot.
A
So
my
proposal
is
to
have
all
of
them
defined
in
the
annotations
and
in
that
way,
maybe
we
can
provide
better
portability
and
support
many
platforms.
Things
like
that.
This
is
an
initial
draft
where
I
start
like
having
a
kind
of
discussion
like
review.
A
No
things
I
have
provided
a
link,
so
if
you
feel
like
a
would
be
a
good
idea
or
like
Community
can
be
beneficial
of
this
particular
topic
or
like
if
you
consulate
it's
not
a
good
idea
and
also
if
you
think
that
there
will
be
a
another
best
way
to
define,
there's
a
I
guess
this
is
a
place
to
start
the
discussion.
Oh
I
noticed
that
Taylor
joins
the
conversation
hi
sailor,
so
yeah.
Basically,
this
is
the
the
initial
state
of
this
proposal.
A
It's
not
too
much,
but
at
least
I
expect
to
to
have
that
initial
conversation
where
you
can
I
noticed
that
I
Ian
has
been
very
important.
Some
ideas
in
this
live
Channel,
that's
great,
so.
A
I
don't
know:
do
we
have
something
else,
a
tailor
to
to
discuss
how
besides
particular
things
in
the
agenda
like
I've
just
checked,
the
PRS
will
appears,
which
is
pretty
or
the
same
I
will
they
haven't,
had
any
particular
change
on
that
and
have
just
start
talking
about
the
the
the
draft
that
I
have
very
liked
last
week.
B
A
Well,
we
have
these,
which
was
great
white
Daniel
four
months
ago,
yeah
right
and
in
that
we
have
kind
of
a
gigs
of
discussion
where
I
look,
I
guess
is
his
name.
A
She
he
started
the
initial
conversation
of
this
in
this
time
he
makes
a
few
things
about
supporting
multiple
protocols
or
like
using
in
this
case
he
was
expressing
the
usage
of
maltus
and
also
the
requirement
that
not
to
have
which
is
a
predefined,
the
networks.
So
obviously
he
was
combined
those
two
topics,
preparation
and
and
the
usage
of
motors
with
I-
don't
think
that
we
concluding
something
particularly.
It
was
a
lot
of
good
things
to
discuss
here.
A
That's
why
Daniel
Curry,
like
it's
kind
of
new
issue,
which
is
some
right,
something
like
her
related
with
the
discussion,
but
it's
not
directly
related
so
and
yeah.
Definitely,
the
the
the
best
practice
is
just
my
proposal.
Something
I
have
that
I
work
in
in
the
past,
so
I
did
some
experiments
where
I
conclude.
Basically
what
I'm
proposing
say
like,
for
example,
in
this
case,
I
have
a
CNF
that
I
tried
to
Define
with
multiple
annotations.
A
That's
that's
was
a
way
to
provide
that
opportunity,
so
I
just
wanted
to
put
it
like
a
proposal
or
somewhere
like
when
someone
can
review
it
and
maybe
started
like
the
conversation
or
like
concluding
something
like
we
don't
have
to
do
this.
Maybe
it's
not
a
best
practice.
It's
like
anti-partner
or
something
like
that.
A
B
So
Ian
was
asking
the
question:
is
this
trying
to
codify
what
people
are
currently
doing?
Are
we
trying
to
communicate
suggest
an
idea
to
improve
the
state
of
the
art.
A
Yeah,
that's
really
it's
pretty
hard
to
find
at
least
an
open
source,
CNF
ample,
so
real
examples,
I
mean
we
have
a
couple
like,
but
most
of
them
or
they're
very
complex.
For
example.
I
know
that
magma
has
Implement
something,
but
it's
pretty
hard
to
adjust
all
what
they
have
been
doing
or
you
know
they
have
a
couple
of
them,
but
it
is
just
a
very
dumb
examples
where,
like
I,
don't
think
that
as
much
positive
to
20
any
real
example.
A
So
what
I
have
seen
is
like
they
just
pick
one
solution:
they,
for
example,
just
pick
motors,
which
is
the
most
common
solution
and
they
try
to
design
their
own
tnf
models
that
I
don't
know.
If
this
is
a
good
approach
or.
A
A
B
Network
information
and
code,
like
not
putting
IP
addresses
directly
in
code,
would
be
related.
That's
not
specifying
your
network
interfaces
and
your
code
instead
you're
dynamically
allocating
and
binding
to
the
interfaces
based
on
what's
available
so
that
all
sounds
good
I'm
just
trying
to
see
where
it
fits
together
and
the
example
that
you're
showing
would
be.
A
Yeah
exactly
I
mean,
maybe
we
can
break
them
down
in
do
best
practices
like
I
mean
the
first
one
is
Maybe
I
just
said
like
pointing
just
to
not
specify
any
any
particular
Network
details
like
just
limit
and
limit
the
definition
to
the
name
of
and
that's
it
which
is
in
this
case
I
guess.
A
The
only
thing
is
like,
but
this
part
is
doing
is
basically
just
specifying
which
networks
are
going
to
be
connected,
but
it
is
not
specifying
the
IP
address,
or
you
know,
maybe
there
I
think
that
maybe
their
families
a
little
bit
more
picky,
which
I
have
to
provide
a
little
bit
more
details,
but
probably
we
can
reach
that
commit
and
say
well,
maybe
what?
What
can
we
do
for
reducing
the
amount
of
information.
A
And
in
modus
yeah
you
have
to
provide
all
like
you
have
to
provide
the
the
name
of
the
name
of
the
interface,
but
you
don't
you're,
not
trained
anything
related
with
the
IP
address
or
the
the
thumbnet
that
you're
using.
So
this
is
really
more
specific.
You
have
to
say:
well,
this
network
is
going
to
be
attached
to
this
interface
with
diamond
is
just
the
name
and
with
MFM
is,
is
given
his
nature.
Is
you
have
to
consider
it?
A
You
are
consuming
the
endpoint
or
if
you
are
offering
an
endpoint
so
which
depend
on
the
case.
You
have
to
define
the
things
differently,
but
but
I
guess
I
like
a
rich
certain
level
of
abstraction,
and
this
is
not
the
standard
way
that
they
are
saying
because
they
have
like
a
they
need
to,
and
you
need
containers
and
those
unit
containers
has
to
specify
all
this
information.
A
So
so
they
haven't
provided
that
level
of
abstraction.
I
just
reached
that
level
of
attraction
in
FM,
because
I
create
multiple
architects.
But
but
my
point
was
you:
you
can
reach
that
level
of
abstraction.
If,
once
you
have
all
these
things,
it's
easier
to
get
a
similar
thing
and
and
offer
this
network
definition
independama
patients,
but
yeah
I'm
saying
is
this
is
the
right
way
to
do
it.
I
don't
know.
B
Your
example
is
us,
you
are
limited
in
trying
to
implement
this
based
on
the
software
that
you
were
working
with,
which
is
emulating
yeah,
and
the
specific
interfaces
are
required
for
the
application
to
work
and
they're
they're
also
I
think
relevant.
As
far
as
what
we're
talking
about
in
the
group.
A
And
also
another
thing
that
I
noticed
Taylor
is
regarding
the
order,
because
most
of
the
cnns
that
I
haven't
seen
in
in
yeah
around
it's
they
have.
They
provide
like
a
manual
steps
to
deploying
things.
For
example,
they
consider
like
once
a
particular
container
is
a
poet
they
keep
it
and
they
are
not
considered
like
that
case,
where,
like
you,
try
to
boil
them,
and
eventually
you
know
reaching
that
ready
state.
So
that's
why
I
have
created
like
this
other
script,
like
I
mean.
A
Usually,
this
is
what
we
can
expect
in
in
places
like
kubernetes
you,
you
have
a
particular
order.
Some
of
the
containers
are
going
to
start
before
others
and
and
that
eventual
Readiness
is
something
that
dissolves
important
service
is
I.
A
Don't
know
if
that's
something
that
we
have
also
put
like
another
best
practice
or
another
point
to
this
cost,
but
yeah
it's
it's
important
to
highlight
that
the
order
is
not
something
that
we
can
control
and,
and
eventually
we
can
expect
some
failures
like
if
one
of
the
dependencies
is
is
working
at
the
beginning,
but
eventually
jackpot
or
container
is,
is
failing
so
and
and
then
next
time
that
it
is
running
again
and
the
trips
and
other
IP
address.
But
can
we
do
with
the
rest
of
the
the
dependencies?
B
Some
of
that
will
be
covered
with
existing
tests
that
are
happening
that
they
list
on
the
CNF
certification
around
the
resiliency
and
other
things
where
something
comes
up
goes
down.
A
component
goes
down,
but
writing
up
a
best
practice
related.
That
would
probably
be
a
good
idea,
so
it's
decoupling,
those
parts
you
were
talking
about
how
the
timing
not
not
having
dependency
on
another
component.
B
A
Yeah,
but
not
only
during
the
day
Zero
also,
they
won
them
too,
as
well
right
like
let's
head
up,
everything
is
up
and
running
and
it's
working
fine,
but
eventually
one
of
those
components
are
are
bound.
So
how
about
CNF
has
to
behave
like
a
to
need
to
redeploy
everything
or
like
just
modify
the
configuration
to
point
to
a
new
component,
which
you
said.
B
B
You're
working
on
with
this,
this
GW
tester,
this
is
related
to
the.
B
Mobile
core
components
and
they're
expecting
at
least
like
the
session
manager
or
something
to
always
be
available.
You
look
at
like
the
ink
yeah
and
this
is
only
showing
a
a
couple
of
the
components.
But,
of
course,
there's
a
whole
lot
more
here,
including
yeah.
A
B
The
the
right
now
it's
dependent
on
specific
address
of
stuff.
Once
it
comes
up
always
to
be
available,
so
it
would
think
if,
if
one
of
them
goes
down,
then
at
least
the
session
manager
is
always
available
to
go
back
and
talk
to
to
get
the
IP
address
for
the
other
components
and
connecting
via
that
specific
address,
even
if
it
was
Dynamic
at
when
it
was
bootstrapped
into
the
system.
B
B
A
service
that's
going
to
be
communicated
by
kubernetes,
here's
the
new
address
versus
whatever
it
had
hard-coded
and
I'm
I'm,
not
saying
like
the
whole
machine
goes
away,
but
maybe
it
reboots.
So
now
your
gateway
rebooted
and
it's
coming
back
up.
Well,
how
does
it
find
the
MMA?
Maybe
the
MMA
actually
moved
to
a
a
new
system
and
has
a
new
address
completely
well,
it
should
be
able
to.
A
A
A
So
so
yeah,
basically
I
tried
to
do
this
example.
Just
like
try
to
experiment
with
something
a
little
bit
more
complex,
so
in
in
terms
like
or
more
realistic.
It's
not
pretend
to
be
a
very
functional.
A
I
tried
to
exercise
most
of
the
my
learnings
in
all
these
Technologies
and
I
just
discovered,
like
that
good
case,
where
I
can
summarize
all
the
Network
definition
all
in
internal
patients
and
obviously
when
I
started
doing
all
these
things,
I
discovered
other
use
cases
and
SNL
that
I
have
to
cover
and
I.
Don't
pretend
like
to
use
this
in
production?
Definitely
it's
not,
but
my
purpose
but
I
just
try
to
you
know,
reach
those
limits
and
try
to
find
places
to
improve.
B
It's
a
Google
doc
in
comment
mode
for
anyone
that
gets
access
to
that
link.
B
If
you
make
it
where
it's
comment,
accessible,
well,
yeah,
anybody
with
link
could
comment
and
we
can
change
it
if
you
end
up
with
a
ton
of
spam,
but
probably
okay
dropping
this
for
sure
and
they
have
discussions,
but
probably
in
the
GitHub
issue,.
A
B
A
Yeah,
and
also
if
you
think
that
it
should
be
there
like
to
reduce
the
steps
before
this
one,
like
I,
don't
know
the
network
definition.
A
How
can
we
Define
all
these
things
before
proposing
having
all
these
no
patience
like
I,
don't
know
like
breaking
down
this
particular
best
practice
in
in
other
two
or
like
three
yeah.
It's
also
a
facility
as
well.
B
A
So
anyone
else
has
something
to
share
comment
or.
A
Okay,
I,
don't
know.
B
Ping
pinching
I
think
you
had
you
had
a
comment
about
extra
Villa
is
a
new
release.
B
We
forgot
to
reschedule
the
talk
on
music
group,
so
I
went
over
there
just
before
the
call
he
was
over
on
that
call
from
China
mobile.
B
Maybe
it's
not
there.
All
right
may
drop
a
link
in
there
for
the
new
release
and
get
some
info
on
XG.
Villa
release
like
to
see
some
of
the
stuff
that
they're
doing
I
mean
the
multi-interface
multi-network
stuff.
Was
it
when
I
was
looking
at
the
XG
Villa
a
while
back
I
mean
it
was
part
of
the
discussion,
so
I'm
sure
they've
looked
into
that
as
well.