►
From YouTube: IETF113-DMM-20220321-0900
Description
DMM meeting session at IETF113
2022/03/21 0900
https://datatracker.ietf.org/meeting/113/proceedings/
A
B
B
Yeah,
I
just
thought
yeah,
but
he
said
that
yeah
I
was
talking
to
lisa.
That
oh
is
here.
He
said
yes,.
E
C
C
C
Please,
like
this
and.
C
C
So
first
thing:
blue
sheets.
D
If
not
there's
a
mobile
app.
There's
like
the
the
medeco
light
you
can
use
that
can
record
your
participation
as
well.
Okay,.
D
If
there's
no
bluetooth
in
the
room,
then
probably
I
would
impress
everyone
to
sign
into
the
medical
light
that
way
they
can
join
the
queue
and
observe
the
chat.
Okay,.
C
F
D
F
C
C
C
D
C
C
So
I
think,
just
to
remind
everyone.
This
document
went
through
a
lot
class
call
sometime
last
year,
but
there
have
been
two
comments
and
concerns.
So
that's
why
we
did
not.
We
asked
the
authors
and
the
folks
who
provided
the
comments
to
resolve
the
issues.
So
there
have
been.
You
know
some
progress
and
we
I
will.
I
thought
it's
now
pretty
much.
C
You
know
we're
almost
ready
to
move
to
the
next
stage,
but
we'll
confirm
that
after
this
post
idf,
so
I
want
to
give
the
opportunity
to
the
folks
who've,
provided
those
comments
to
review
and
make
sure
the
changes
that
are
in
the
document.
C
G
Okay,
thank
you
that's
right
so
good
morning
or
good
afternoon,
and
many
thanks
for
your
time.
So
this
is
an
update.
G
So
briefly,
to
wake
up
a
bit
on
the
estate,
so
we
started
the
working
group
alaska
in
april
last
year
on
revision
11.,
we
had
a
good
amount
of
participation
with.
G
Of
everyone
who
have
improved
the
document
and
well
throughout
the
last
itf-112,
there
were
a
number
of
issues
that
were
still
pending
and
we
believe
that
they
have
been
solved
nexus
like
this
attorney.
G
Where,
basically,
we
are
well,
the
document
is
defining
or
well
this
defining
how
to
you
could
use
the
srv6
endpoint
behaviors,
which
are
defining.
G
G
The
view
for
the
comment
from
joel
it
was
that
these
sections
are
not
as
standard,
and
the
discussion
that
we
had
around
them
is
that
this
text
is
providing
value
to
any
operator.
Who
is
deploying.
G
Just
mentioning
that
these
three
sections
are
merely
informative,
and
so
we
have
added
one
paragraph
about
it.
I
believe
joel
is
not.
I
don't
know
if
you
always
here,
but
anyway,
please
review
the
paragraph
and
personally,
if
you
want
to
change
the
specific
wording,
we.
E
G
Change
it,
but
for
the
rest,
it
should
be
good
and
then
the
other
pending
comment
that
we
had.
It
was
a
few
email
exchange
with
jeffrey
and
from
these
emulations
with
jeffrey.
There
were
a
few
editorial
things
to
change
in
the
same
report,
and
it
has
been
the
change
next
to
slides.
G
So,
given
that
this
draft
has
been
well,
the
working
group
last
call
started
in
april
last
year
and
given
the
the
vast
amount
of
changes
that
we
have
done
on
the
drug,
the
authors
believe
that
it
would
be
good
to
issue
a
second
working
group.
Last
call
I
don't
know
what
is
the
view
of
the
work
group
search
and
that's
really
many.
C
Thanks
now,
thank
you.
Thank
you
for
your
presentation.
So
there
have
been
few
other
comments.
Well,
I
think
you
know
I
thought
yeah
I've
discussed
with
you
and
other
authors
so
overall
perspective.
We
also
wanted
to
make
sure
you
know
from
3gpp
point
of
view
when
we
say
this
is
the
the
interfaces
that
we
refer
to
right.
C
It
does
not
alter
the
definition
of
the
pgp
right,
so
we
wanted
some
convention
some
wordings
which
says
that
this
is
I.t
of
the
version
of
the
interface
right.
That
way,
it's
not
truly,
we
are
not
altering
the
pgp
specifications,
but
has
been
one
comment.
Any
any
comments.
Popular
is
exactly
how
you
know.
You
have
resolved
that
comment
and
you
share
a
bit
more
details.
G
Yeah
sure
so,
basically,
at
the
beginning
of
the
draft,
I
believe
I
don't
know
we
call
it
section.
Four
three.
We
have
paid
particular
attention
to
the
wording,
and
so
basically,
what
we
said
is
that
we
are
going
to
use
this
reference
architecture,
the
3gpp
reference
architecture
for
5g,
with
the
corresponding
interfaces
references.
G
Okay,
so
that
that's
a
bit
how
we
solve
it.
C
Okay,
all
right,
I
think
we
can
review
the
text
and
we
can
take
it
from
there,
but
largely
the
the
issues
that
I
am
tracking.
So
I
think,
hopefully
you
know
is
addresses
all
those
calls.
C
I
think
the
next
thing
you
know
post
idea,
probably
sometime
next
week,
I'm
going
to
issue
in
a
second
last
stop
and
to
make
sure
joel.
C
You
know
and
others
right
who
are
provided,
who
have
been
quite
active
and
who
have
provided
substantial
feedback
to
review
these
the
changes-
and
you
know,
acknowledge
that
they're,
okay
with
it,
I
think
that's
that's
a
plan,
and
so
thanks
again
any
at
this
time,
any
any
questions
from
anybody
from
the
room
or
from
remote.
For
this
anything
that
you'd
like
to
comment.
C
Okay,
if
not
thanks
again,
we
move
to
the
next
topic.
C
C
C
I
It's
a
simple
update
that
I
can
talk
through.
You
know
to
give
what
the
changes
were
in
zero,
three
versions,
all
right,
okay,
so
they
were
the
main
comment
from
the
last
meeting,
and
you
know,
discussions
were
that
the
draft
text
is
all
over
the
place
and
that
it
needs
to
focus
so
and
then
there
were
some
other
comments
too,
which
you
know
there
were
a
couple
of
others
reference
to
drafts
on
slices
and
then
there's
a
few
miscellaneous
changes
too.
I
So,
focusing
on
that
first
one,
you
know
to
correct
the
fact
that
the
text
is
all
over
the
place.
We
revised
significant
portions
of
the
introduction
and
there
are
two
main
changes
which
I'll
just
talk
through.
I
I
I
think
I'm
not
sure.
C
C
I
I
Okay,
then,
I
guess
I'll
post
these
to
the
mailing
list.
The
changes
that
we've
made.
C
D
B
C
So
the
next
presentation
is
mobile,
user
plane
message
encoding.
This
is
an
individual
id.
K
Okay,
so
good
morning.
K
And
also,
we
are
simply
to
cody.
C
So
yeah
this
document
has
been
in
listening
discussions
for
a
while.
I
think
it'd
like
to
definitely
share
adoption
card,
but
before
that
we
just
want
to
make
sure
you
know.
People
have
read
and
you
know
wanted
to
see
their
views.
Probably
yeah
dean
has
joined
the
call
yeah.
B
B
C
J
I
Oh,
that's
interesting.
I
just
switched
my
browsers,
you
know
it's
a
browser
issue.
C
I
So
I
I
switched
to
chrome
from
the
windows
browser,
so
that
was
it.
C
Right,
I
think
let
me
give
an
update
on
this,
so
this
document
went
through
several
revisions.
There
have
been
some
discussions.
I
think
we'd
like
to
you
know,
issue
an
adoption
call,
probably
I
think
I'll
call
the
working
group
for
email
now,
it's
difficult
for
me
to
track.
You
know,
as
you
know,
who
read
the
document
or
not
so
I'm
going
to
do
it
online.
We
are
going
to
the
chairs
are
going
to
issue
an
adoption
call
on
this.
C
C
L
L
L
Actually,
there
one
person
has
some
idea
to
utilize
excel
policy,
but
some
of
this
questions
we
may
we
understand
that
the
authority
doesn't
need
to
be
involved
in
the
architecture,
so
we
drop
the
electrical
receive
them
binaural,
fit
as
a
network
for
session
trucks,
from
a
drought,
not
being
a
story.
L
We
have
a
bunch
of
detect
updates,
but
the
fundamental
aspects
that
doesn't
change
so
one
the
main
change
was
the
description
was
focus
on
the
segments
instead
of
the
role
of
the
node
that
helps
object.
Editors
dispose
description
and
much
simpler.
L
We
briefly
describe
what
the
session
transfer
rules
are
allowed
to
look
like
in
the
zero
one
bar
and
well
now.
This
gives
you
have
three
technical
type,
one
type
of
questions
from
the
round,
and
we
have
to
simply
that
idea
using
extended
community
to
indicate
its
direct
premium.
L
Oh
next,
one
next
step.
Maybe
we
have
some
more
clarification
in
next
division
and
also,
if
he
needed
more,
you
stay
for
some
helpful
here
situation.
C
So
this
document
is
again
an
individual
id
as
saying
we
need
to
issue
a
validation.
Adoption
call
at
some
point,
so
I'd
like
to
see
more
discussions
on
the
in
the
mailer.
You
know
there
haven't
been
more
many
on
this
topic,
so
I
think
maybe
I
think,
secretary
san
and
other
authors,
we
may
want
to
some
trigger
some
discussions
to
you
know
that
will
help
us.
You
know
set
the
ground
for
the
issuing
an
option
called.
M
No,
yes
exactly
well!
Actually
I
raised
my
hand
and
tried
to
ask
questions
since
you
did
not
see
my
hand
so,
okay,
yeah.
I
M
Yeah
yeah,
I
have
some
questions
here.
It
seems
like.
I
only
have
a
lot
of
apples
here
by
the
way
I
only
read
through
the
the
version
zero
one
now
in
zero
three,
and
it
seems
because
when
I
read
zero
one,
I
got
a
question
on
the
mup,
the
controller
itself,
instead
of
about
basically
it
force.
My
question
focus
on
the
interaction
between
the
ip
of
the
transport
domain
and
the
the
5g
mup
plan.
M
My
question
actually
is
for
your
mup,
the
controller,
and
how
are
you
going
to
you
know
basically
for
the
5gs
system,
the
architecture
itself,
and
I
really
want
to
know
how
are
you
going
to
put
your
the
mup
controller?
Are
you
going
to
consider
us
as
a
nf
network
function
or
you
consider
us
a
application
function
so,
yes,
or
something
else,
because
it's
going
to
introduce
some
complexity
for
the
5g
as
a
control
player?
So
so,
basically,
my
question
might
not
be
related
to
d03
itself.
Basically,
it's
related
to.
M
How
are
you
going
to
integrate
the
mup,
the
the
controller,
all
the
things
with
the
5gs
functions.
L
For
your
question,
this
is
saturday
I'm
not
trying
to
answer
your
question
as
we
described
in
what
well
it's
obvious
map.
Architecture
is
independent
from
any
specific
mobility
management
architecture,
not
only
for
the
5d
but
also
the
idf
architecture,
payment,
for
example,
integration,
disability
map
into
specific
mobile
satellite
cryptography.
How
to
describe
this
document?
L
M
Okay,
now
it's
clear
to
me
since
the
rest
of
the
zero
one.
I
always
had
this
question.
No,
I
know
yeah
yeah,
okay,
observing
from
your
proposal
right
now.
So,
okay,
thank
you.
M
Yeah,
you
know
it
seems
like
for
the
of
operator
this
type
of
things.
It's
it's
not
very
trivial
to
introduce.
So
that's
why
I
have
this
question
in
my
mind,
so.
C
N
How
much
time
do
we
lose?
Did
we
lose
a
lot
of
time
because
I'm,
if
I
did,
then
I
may
skip
some
slides.
It's.
C
N
Yeah.
Okay,
thank
you.
So
I'm
going
to
talk
about
two
drops
here
with
course:
skier
and
and
changi
next
slide.
Please.
N
This
is
to
set
the
stage
for
the
discussion
in
the
next
draft,
so
the
next
draft
is
about
potentially
in
the
next
generation-
let's,
for
example,
6e
to
integrate
the
genome
b
and
upf
function
together
some
thoughts
I
I
have-
I
have
exchanged
with
some
other
people
with
my
co-authors.
N
We
want
to
socialize
this
idea.
First,
the
work
would
have
to
be
done
in
3gpp,
we're
not
trying
to
do
the
3gpp
work
in
iatf,
but
we
want
to
socialize
idea
first
among
the
people
who
have
familiar
ways
and
friendly
with
active
wireline
technologies.
N
N
There
is
relay
going
on
between
the
radio
layer,
radio
broadcast
layers
and
then
and
gdpu
and
the
relay
going
on
in
intermediate
upf
and
finally,
it
reaches
the
anchoring
upf
we
decap
the
gdpu
tunnel.
We
do
the
routing,
switching
and
then
send
the
traffic
to
the
dn.
N
Here
by
routing
and
switching
I
mean
that
the
traffic
folding
is
based
on
the
ip
header
or
ethernet
header
of
the
that
is
below
the
gdp
header
next
slide.
Please.
N
So,
with
edge
computings
nowadays,
upfs
are
being
distributed
closer
to
the
general
beats
and,
as
a
result,
the
data
networks
are
also
distributed
to
to
those
sites
to
host
the
edge
resources,
and
that
means
that
the
data
networks
are
implemented
as
a
vpn
as
vpns
for
intersect
connections,
while
the
the
central
upfs
can
remain,
they
could
be,
they
could
actually
be
removed
and
with
distributed
upfs,
the
persistent
ip
addresses
can
be
provided,
even
if
the
ue
is
re-anchored
from
one
upf
to
another.
N
N
This
allows
complete
local
breakouts
of
the
traffic
to
local
to
local
data
networks
that
are
interconnected,
and
this
simplifies
and
also
optimize
the
edge
computing
broadcast
broadcast
service
and
land
type
services.
N
We
have
the
on
the
right
side
of
the
transport
network.
It
provides
vpn
services
on
p1p2,
we
have
vr
apps,
vrf1
and
vf2
for
local
data
network
networks
respectively,
and
then
there
is
a
vp3.
N
It
also
has
a
corresponding
vpns
apps
for
the
central
data
network,
which
could
be
internet
itself,
so
that
nothing,
nothing
really
changes,
except
that
the
upfs
are
distributed
next
side.
Please.
N
So
there
are
also
there
are
at
least
in
this
dmm
working
group.
We
have
been
talking
about
alternatives
to
the
upf
and
the
gdpu
implementation.
Some
operators
and
vendors
are
actually
pushing
for
for
this
alternative
implementation
deployments.
This
is
under
the
hood.
No
3gpp
architectural
signal
changes
are
done.
N
One
example
is
the
srv6
replacing
gdp
tunnel
that
the
draft
the
pablo
was
talking
about
and
the
other
one
is
the
srv6
map
architecture
that
satoru
just
talked
about,
which
is
srv6
specific,
it's
kind
of
is
really
just
a
router-based
and
partial
implementation
of
the
distributed.
Upf
uk.
N
In
particular
for
the
srv6
map
architecture-
and
this
actually
sort
of
answers,
the
question
from
from
tng
at
least
the
way
I
see
it,
the
collection
of
distributed
map,
controller
map
gateway
and
rpe
appears
as
a
as
a
single
and
central
upf
to
the
smf.
N
That's
it
nothing
else,
changes
it
is
an
alternative
to
the
distributed,
distributed
traditional
upf.
Another
observation
I
want
to
make
here
is
that
the
srv6
map,
architecture
or
solution
is
actually
a
sr
agnostic.
N
In
fact,
for
the
pgp
signaling
draft
that
that
goes
again
together
with
the
srv6
mod
map
architecture.
There
is
another
version
that
is
sr
agnostic,
based
on
that
existing
draft
with
minimum
changes.
N
I
want
to
point
out
this
is
so
far
just
for
partial
upf
functions.
If
someone
wants
to
to
once
need
needs,
a
complete
set
of
upf
functions,
then
either
that
architecture
can
be
extended
or
just
deploy
traditional
but
distributed
upfs
again.
This
is
just
the
overview
and
summarize
all
the
5g
user
plan
options
and
then
trends.
Things
like
that
next
site.
Please.
N
Now,
finally,
I'm
getting
to
the
second
drafts
I
mentioned
earlier,
what
if,
in
the
next
generation,
for
example,
in
6g,
we
completely
integrate
the
genome
b
and
the
upf
function
together
yeah.
This
is
not
for
5g
at
all,
it's
a
potential
for
6g.
If
we
can
get
enough
support,
then
we
can
bring
to
3gpp
here.
Basically,
the
geno
b
and
upfr
integrate
into
a
single
large
logical
entity.
N
N
N
The
main,
the
main
reason
for
for
for
doing
this
is
that
this
is
a
simplified,
flattened
network
architecture
with
3gpp
and
technology
for
radio
access
and
ihf
technology.
For
the
rest,
it
allows
seamless
integration
of
vana
and
wireless
services,
and
I'm
not
talking
about
www.c,
it's
really
for
that
edge
computing,
multicast,
broadcast
service
and
non-type
services.
Other
things
we
can
do
seamless
integration
for
between
the
wireless
and
wireless
services.
N
N
Currently,
we
need
a
separate
and
two
m4
signaling
to
the
genome
b
and
to
the
upf,
and
there
is
a
need
for
the
ggbt:
u
tunneling,
between
the
two
devices
when
we
combine
them
together,
we
do
not
need
a
separate
and
two
and
four
signaling,
and
we
do
not
need
that
separate
connection
between
the
a
and
upf,
no
matter
how
close
they
are.
If
the
the
large
the
two
entities
are
logically
separate,
there
will
be
a
connection
between
them.
N
It
could
be
internal
connection
also
because
the
now
the
network
functions
are
more
and
more
virtualized.
This
is
actually
very
feasible
to
do
it
on
the
servers
that
host
the
services
currently
and,
in
fact
even
the
vpn
pes
could
be
integrated
into
data
and
app
devices.
That's
as
an
implementation
choice
next
slide.
Please.
N
I
do
want
to
point
out
there
are:
there
are
still
cases
where
separate
upfs
may
still
be
needed,
for
example,
homology,
roaming
m
and
v
n
o
or
just
sometimes
you.
You
need
a
upf
for
to
connect
for
multiple
genobies
in
proximity
when,
when
there
is
that
need
that
the
a
and
upf
connection
can
can
be
the
gdpu
tunnels
or
it
could
be
ihf
pseudo-wires
everything
is
possible.
N
N
So
these
two
drafts
are
both
informational,
the
first
one
just
to
give
the
background
the
set
stage
for
the
discussion
in
the
next.
In
the
second
draft
that
talks
about
the
potential
integration
and
again
we,
the
purpose
of
this
is
to
socialize
the
thoughts.
I
I
want
to
get
your
comments
and,
if
you're
sympathetic
to
this
idea,
let's
talk
and
that's
called
collaborate,
and
if
this
makes
sense,
if
we
get
enough
support,
then
we
can
bring
this
to
to
3gpp.
C
C
This
working
group.
We
truly
don't
have
the
complete
charter
to
go
change,
3gbp
architecture.
We
are
generally
open
to
taking
a
work
that
allows
us
to
use
ietf
tools
and
improve
the
system,
but
it's
not
exactly.
We
are
not.
We
are
not
chartered
to
change
the
3gb
interfaces
right
with
that.
You
know
keeping
that
comment
in
mind.
C
C
N
So
today,
the
the
in
the
centralized
upf
the
gdp
tunnels,
gpu
tunnels,
are
used
to
bring
the
sessions
from
genome
b
to
the
to
the
central
upf
and
in
the
distributed
case,
especially
in
that
srv6
map
architecture,
we
are
sort
of
breaking
the
central
upf
into
distributed
ones,
and,
but
still
there
is
still
a
separation
between
the
genome
and
upf.
N
So
if
we
could
convince
that
3gbp
that
we
can
actually
integrate
the
genome
b
and
the
upf,
then
yes,
the
ietf
technology
will
will
will
be
used
to
connect
those
all
those
integrated
devices
together.
At
the
data
network
level,
the
pdo
sessions
terminates
at
the
end
up
and
then
from
their
point
on
it's,
it's
itf,
two
technologies,
it's.
This
is
very
much
like
you
connect
your
pc
to
a
router
using
ethernet
and
then
the
rest
is
just
the
wireline
networks
right.
N
It's
here,
you
replace
the
ethernet
cable
with
the
radio
access
and
the
rest
will
be
the
same
as
the
rest
will
be
just
a
wireline
technologies.
C
N
I
just
I
mentioned
that,
I'm
here
basically
a
socialist.
This
idea,
I
don't
expect
the
work
to
be
done
in
this
working
group.
C
Okay,
got
it
okay!
Thank
you!
So
much
yeah
any.
I
think.
Let's
open
up
the
queue.
Let's
any
comments,
questions
anybody,
oh
yeah,
yeah.
First,
let's
start
with.
C
N
Oh
actually,
the.
E
Use
google
products-
okay,
you
can
hear
me
now
yeah,
good,
yes,
okay,
thank
you
for
interesting,
thought-provoking
presentation.
I
have
a
couple
of
comments
to
it.
First,
I'm
comparing
the
proposals,
what
exists
in
5g,
not
6g,
because
it's
still
pretty
much
underworks,
so,
first
of
all
in
5g,
it's
already
possible
to
co-locate
intermittent
upf
iupf
with
the
access,
node
or
gene
or
v,
that's
already
a
possibility.
E
But
if
you
and
your
proposal
is
moving
esa
to
the
access
mode,
but
if
you
do
that,
I
I
I'm
wondering
how
how
would
you
handle
then
multi-access
scenarios,
for
instance
interworking
with
non-cheat
non-3gpp
interfaces.
This
is
not
clear
for
to
me.
Secondly,
it
seems
that
this
is
impacting
5g
quality
of
service
morale
and
related
to.
That
is
that
that
you
could
of
course,
imagine
to
build
this
5g
quality
of
service
model
into
the
t
note
b.
But
if
you
do
that,
then
you
enter
up
into
security
problems
in
comparison.
E
N
So
the
first
one,
the
co-locating
of
geno
b
and
upf
function,
and
yes,
you
can
do
that
today,
but
even
with
a
co-located
functions,
you
still
need
separate
n2
and
m4
signaling
and
you
still
need
that
connection
between
them
and
so
by
actually
combining
them
in
integrate
them
together
into
a
single
entity.
You
will
simply
optimize
both
the
signaling
and
data
plan
that
that's
the
that
that's
the
reason
I
we
want
to
have
this
option
about
the
security
of
having
the
subsequent
information
available
on
glob
the
impact
of
security.
N
I
indeed
have
not
considered
the
curve
carefully.
I
will
look
into
that
further.
E
N
I
would
think
I
would
think
that
it
is
the
same
to
me.
The
only
difference
is
whether
it's
it
comes
in
on
a
radio
link
or
on
some
other
methods.
C
Yeah,
I
think
this.
This
is
a
good
topic.
I
think
yeah.
Let's
take
these
discussions
to
the
mailing
list.
I
think
android
some
great
comments
on
not
gpp
access
and
other
things
overall
yeah.
I
think,
let's
see
the
relevance
of
this
work
to
this
this
I
dm
working
any
other
questions.
I
Yes,
I
had
a
few
comments,
but
yeah,
please
yeah
yeah,
please
go
ahead.
Okay,
thanks
somewhat
similar
observations
as
hanu,
but
the
focus
I
thought
in
the
draft
was
to
emphasize
on
the
routing
aspects
of
these
upf
and
am,
but
I
was
wondering
about
things
like,
for
example,
qos
and
the
aspects
defined
in
5g.
So,
for
example,
even
in
release
18
they're
going
to
work
on
further
extending
qos
in
xrm
and
others,
but
even
the
standard,
5g
qos
aspects
need
to
be
considered.
I
I
can
post
maybe
a
little
more
detailed
comments
in
the
interest
of
time
and
on
the
mailing
press.
Okay,.
C
Yeah,
that's
a
good
idea.
John.
Thank
you
for
your
comments.
I
think
john.
If
you
can
open
up,
you
know
this
on
the
mailer.
I
think
that
will
be
great
yeah,
we'll
you
know
plug
the
queue
here.
I
think
dean
next
and
after
that,
louise
here.
B
So
I
really.
B
Draft
and
I
and
I'm
I'm,
I
want
to
support
it
because
what
jeff
is
bringing.
P
M
J
B
J
B
B
That
we
can
then
decide
what
do
we
want
on
the
qos
part,
the
data
plane
qos.
We
know
how
to
handle
that
on
the
security
part.
We
also
know
how
to
handle.
You
know
the
the
data
plan
security.
Yes,
there
are
some
issues
that
could
be
between
the
user
plane.
That
says
in
the
sorry,
the
control
plane
that
stays
within
the
3gpp
part.
B
But
if
you
apply
a
pure
cups
architecture,
where
you
have
one
controller
with
multiple
user
plane
elements,
this
is
exactly
what
jeff
is
proposing
he's
saying
like
we
will
continue
to
have
the
upf
control
plane
centralized.
We
will
just
have
multiple
data
planes
that
will
be
exit
points
with
the
prp
traffic,
so
from
from
all
the
proposals.
Q
C
Okay,
thank
you
dean
next,
we'll
we'll
not
we'll
go
to
the
next
comment
of
lewis
is
next
in
the
queue
yeah.
Q
Thank
you,
so
probably
another
scenario
to
look
at
somehow
seeing
the
same
problem
from
a
different
angle
would
be
in
the
case
of
the
radio
functional
split
where
we
centralized
some
part
of
the
of
the
of
the
radio,
the
baseball
unit
and
so
which
is
naturally
being
centralized
so
probably
collocating.
Also
with
the
cu,
the
the
upf
could
be
the
advantages.
So
it's
another
way
of
looking
at
the
problem.
I
think.
C
Okay,
thank
you
luis.
I
think
the
key
thing
is
we
want
to
make
sure
you
know
we
are
not
seeing
you
know.
We
change
the
3gbr
picture
here.
Let's
understand
it.
If
it's
more
on
the
routing
side,
on
the
n6
side,
I
don't
know
what
what
are
the
implications,
but
we
need.
We
need
careful
analysis
and
then
based
on
this,
if
it's
relevant,
we
can
take
it
up
here.
C
How
do
I
close
the
mic,
but
you
can
quickly
if
it's
a
quick
comment,
please
respond
yeah.
I
love
you
yeah.
B
E
E
I
know
yes
again,
but
this
looks
simple
because
you
haven't
yet
hit
a
difficult
part.
For
instance,
this
multicast
demonstrates
very
well
difficult
if,
if
you
push
the
multicast
server
next
to
the
access
and-
and
you
make
a
handover,
you
are,
you
are
making
the
handover
signaling
more
complicated
same
happens
in
other
scenarios
as
well.
We
need
to
have
this
psa
anchor
higher
up
in
the
hierarchy
in
order
to
preserve
the
existing
mobility
signaling.
C
Because
we
have
to
repeat
some
few
presentations,
we'll
see
if
that
happens,
yeah.
So
thank
you,
everybody,
I
think,
there's
some
good
level
of
interest
for
this.
I
think
we'll
do
this
in
the
mail
in
the
mailing
list.
Next
topic
is
we
move
to
the
next
presentation
again,
thank
you,
presenters
who
present
us
for
the
previous
topic
and
and
all
the
feedback.
Next
one
is
5g
distributed.
Ups
for
the
5g
multicast
and
broadcast
services,
there's
a
new
document:
zero,
zero
version
from
tng
png.
It's
all
us.
M
Yes,
you
hear
me
right
now:
yes,
we
good
good,
okay,
oh
yeah,
okay,
thank
you
good
day.
Everyone.
My
topic
on
this
one
is
regarding
the
the
distributed.
Upf
is
applied
to
5g,
multicast
and
broadcast
services,
and
also
thank
you
for
jeff.
So
basically
he
has
already
presented
about
the
dupf.
So
it's
going
to
save
me
a
lot
of
time
on
that
part.
M
Next
slide,
please
l3!
Next,
please
yeah,
for
this
slide
is
about
the
introduction
or
is
the
description
of
the
dupf
in
5g
user
plan.
So
it's
the
all.
The
details
are
in
the
draft
jeff
and
kiel
and
myself
are
composed.
So
I'm
not
going
to
repeat.
I
think
that
jeff
has
already
give
a
very
good
presentation,
so
the
next
one,
three,
the
next
time-
okay,
oh
it's
all
gabo
yeah,.
C
But
I
think
maybe
the
format
yeah
just
speak
to
the
you
know
slide
down.
It's
fine,
okay,.
M
Yes,
let
me
okay,
so
here
well
actually!
Well,
I
I
need
to
read
something
here
so,
but
but
okay,
so
this
picture
is
regarding
the
5g
mbs
architecture,
and
then
I
just
put
some
simple
brief
things
here.
Without
giving
the
full
details,
I'm
not
sure
you
can
see
the
some
some
interfaces.
M
M
Okay,
how
can
I
enlarge?
Oh
okay,
oh
good
good!
Thank
you.
Thank
you
very
good.
Okay,
it's
much
better
right
now,
so
here
I
just
described
briefly
about
the
5g
mbs,
based
on
the
3gp
document.
23.247.
M
Here
you
may
pay
attention
to
three
interfaces
on
the
the
data
player.
One
is
the
n6mb
between
it's
about
the
right
side
between
mbupf
and
the
dn,
the
other
two
19
mb
between
the
upf
and
upf,
and
then
the
other
one
is
the
n3
and
b
between
mb
upf
and
the
g
node
b
here
for
5g
mbs.
There
are
two
delivery
modes:
one
is
the
individual,
the
other
one
is
the
share
so
for
both
cases.
M
So
when
the
dung
link
mbs
packet
from
mb
upf,
where
n6mb
the
mbupf
is
going
to
replicate
and
forward
and
then
based
on
the
either
individual
of
individual
delivery
or
the
shared
delivery.
So
according
to
the
23,
the
3gpp
document,
23.2.
M
From
the
mbupf
or
either
on
the
n6
side
or
a
19
or
n3
side,
either
the
unicast
underlay
transformation,
transport
or
multicast
or
underlay
transport
can
be,
can
be
applied,
so
so,
basically
you're
going
to
deal
with
either
unicast
or
multicast.
This
is
for
the
underlay,
so
so,
for
example,
even
for
the
individual
in
video
mode,
you
may
be
able
to
use
multicast
trans
transmission
to
send
from
mb
upf
to
multiple
ups,
okay,
so
yeah
next
one
place
here
is
just
a
description
of
5g
mbs.
M
Oh
three,
next
one
yeah,
so
here's
a
there
are
some
challenges.
So
the
the
first
one
is
like
for
the
for
the
5g
mbs,
because
the
new
control
plane
and
the
user
plan
functions.
There
are
never
functions
introduced.
So
it's
going
to
post
some
additional
provisioning
and
implementing
challenges
so,
like
you,
try
to
really
unders
understitching
between
upf
and
upf
and
the
dn
here.
M
Other
challenges.
That's
the
third
one.
Well,
like
the
second
one
is
like
three
new
interfaces:
it's
corresponding
to
three
different
segments
and
then
those
three
as
I
have
already
described,
the
three
segments
can
use
either
the
unicast
underlay
or
multicast
underlay.
M
This
will
make
implementation
mode
diversified
but
also
means
more
complicated
or
challenged.
So
a
next
one.
So
next
slide
yeah
sorry
about
that
yeah.
Here
I
I
use
the
the
requirement
from
rfc.
733
is
a
requirement
for
a
dmm
distributed
mobility
management.
It
say:
okay,
multiply
the
efficiency
between
non-optimal
and
optimal
rules.
M
M
M
Lc
is
kind
of
the
common
in
in
operators
network
right
now
for
the
5gs
and
then
in
in
in
the
network
in
the
field
deployment
in
some
operators,
the
the
ups
actually
is
already
distributed
close
to
the
edge
side.
That's
why
it's
here
I
emphasize
the
5gd
ups.
M
So
with
this
in
consideration,
I
can
go
to
the
next
one.
Next
slide
yeah.
So
as
I
oh,
you
can
just
click
through
and
then
to
show
the
full
contents
of
this
slide.
M
One
more
yeah,
okay,
so
here
it's
like
the
proposal
and
the
considerations
like
how
to
use
the
dupf
for
5g
mbs
the
dupf
case
and
then
we're
going
to
deploy
the
upf
close
to
access
network,
basically
the
genome
b,
so
in
that
case
for
the
shared
delivery.
M
So
if
you
still
memorize
the
architecture,
the
brief
architecture
I
showed
so
the
mb
upf
is
the
it's
like
the
the
centralized
upf
and
then
it
goes
through
the
industry
and
b
and
then
distribute
and
then
the
send
traffic
I'll,
transmit
traffic
transfer
traffic
throughout
the
genome
b.
M
So
here,
if
the
mb
ups
or
just
like
a
like
a
distributed,
ups
is
it's
deployed
close
to
g
no
b,
and
then
it's
going
to
going
to
be
made
much
simpler,
just
like
what
jeff
has
described
in
case
in
in
the
proposal
in
the
draft
to
benefit
from
the
dupf
deployment
for
the
individual
delivery
here
is.
It
involves
two
ups
by
the
regular,
the
other
one
is
1
mb
upf.
M
So
here
what
I'm
thinking
or
is
it's
like
to
distribute
and
deploy
both
upf
closer
to
g
no
b?
So,
for
this
case
the
an
a
19
mb
tunnel
and
and
the
regular
ent3
tunnel-
you
know
if
you,
if
you
look
at
the
architecture,
so
they'll
be
significantly
simplified.
M
So
it's
just
like
to
to
to
benefit
also
from
the
the
distribute
upf
architecture
and
remember
here,
as
I
described
for
all
the
three
segments
for
the
n19
mb
and
6mb,
and
the
n3mb
those
three
segments,
either
unicast
or
multitask
under
the
translation,
can
be
used.
So
here
is
if
we
deploy
the
the
ups
two
out
of
in
a
distributed
mode,
closer
to
g,
no
b
they'll.
M
E
E
M
It's
likely
to
integrate
so
sort
of
a
similar,
a
little
bit
similar
to
the
the
a
lg
node
b
integration,
the
genomic
and
the
upf
integration
part.
But
here
I'm
thinking
is
like
to
regular
upf.
M
Together
as
a
distributed
ups,
so
in
that
case,
we'll
make
this
one
even
more:
it's
a
small,
more
efficient
and
also
it's
better
for
multicast
multicast
traffic,
forwarding
it's
a
confirming
to
the
rfc
733,
the
request,
the
the
requirement
eight.
But
still
this
is
the
long
term
part
since
it's
going
to
to
put
the
the
director
upf
and
then
the
upf
together
the
distributed
ups.
M
Yeah,
I
think
they
are
oh
okay,
probably
that's
okay,
so
here
is
for
the
the
next
step.
Things
is
try
to
to
make
the
dupf
shared
and
in
the
video
delivery
mode
I
try
to
refine
more
and
also
in
epi
domain,
the
controlling
signaling
for
underlying
transport,
the
the
vpn
part,
since
you
know
the
srv6
all
this
kind
of
things,
how
to
how
to
do
the
signaling
to
consider
further
and
another
part.
Actually,
I'm
thinking
for
the
for
the
3gpp
right
now.
M
This
there's
a
project
in
2
about
the
ups,
the
the
upf
exposure
capability
to
to
5g
system.
So
in
this
in
this
project,
there
is
there's
some
discussion
regarding
to
to
use
the
the
trusted
af
to
influence
upf
traffic.
M
So
actually
that
one,
you
know
when
I
was
composing
this
draft,
I'm
thinking,
okay,
to
use
like
a
controller,
the
trust
ap.
The
trust
ef
can
be
like
a
controller
such
that
for
all
the
like,
the
relay,
the
staging
and
the
control
plan
to
integrate
between
the
ip
domain
and
the
5gs
together.
So
I
was
at
that
moment
and
also
right
now
and
later
for
the
future,
to
con
to
look
at
the
trust,
af
things
to
see
how
it's
going
to
influence
the
ups
things
so
yeah.
M
C
C
I
think
I
like
the
first
statement,
but
I
think
we
want
to
understand
exactly
you
know,
yeah
what
what
the
changes
will
be,
and
you
know
what
we
can
do
here,
but
before
that
yeah,
let's
see
you
know,
if
anybody
has
any
comments
and
questions
who
is
in
the
queue
one.
Second.
E
Yeah,
yes,
as
I
already
have
it
in
my
previous
comment,
there
is
a
reason
why
multicast
utf
is
central.
C
M
Yes
well,
but
the
thing
is
like
you
know,
it's
I
try
to
how
about
like
for
for
mb
upf.
So
the
thing
is
like
when
the
three
gp
set
up
the
the
control
plan
and
then
there
is
some
criteria
to
select
the
mbupf.
M
Yes,
I
agree
with
you.
One
amp
upf
can
be
used
to
to
talk
to
multiple
genobies.
M
Here,
it's
like
say,
even
if
you
put
this
one
close
to
some
distributed
locations,
you
still
can
do
the
similar
things,
especially
for
for
the
large
deployment
for
for
some
for
some
countries
with
large
networks,
mobile
or
5g
networks.
M
So
I
to
be
honest,
I
am
not
seeing
any
discrepancy
here.
It's
like
say
suppose
you
deploy
the
mbpf
in
the
centralized
or
in
the
distributed
location,
so
it
still
can
handle
the
multiple.
C
C
A
great
technical
discussion-
let's
you
know,
I
think
hundreds
I
know
feedback.
Why
don't
you
respond
in
the
mailer?
We
take
it
from
yeah.
I
think
we'll
also
ask
you
know,
let's
also
ask
others
to
review
the
document,
and
I
think
if
there
no,
if
there
is
interest,
we
can
see
what
we
can
do
so.
Okay,
the
next
step
is
truly.
We
want
to
see
some
discussions
on
on
the
mailer
yeah,
but
dean.
You
know
I
want
to.
I.
J
C
C
All
right
next
yeah.
Thank
you
again,
tng
for
your
presentation.
I
think
yeah
there
is
interest,
let's,
let's
stick
to
the
mailer
next
one
is
yeah
next
topic
is
on.
Next
is
the
satellite
communication.
This
is
a
again
a
new
proposal,
satellite
network
problems
and
solutions
from
l3
perspective.
C
I
was
not
sure
what
we
are
proposing
here,
but
let's
hear
from
the
authors,
can
you
hear
me?
Yes,.
P
Okay,
thank
you,
hello,
everyone,
my
name
is
linhan,
and
my
presentation
has
a
couple
of
drafts.
First
draft
was
the
problem
under
requirement
for
satellite
level,
which
was
published
in
a
111,
and
the
second
one
is
a
semantic
address
published
in
112
and
this
ietf.
We
published
two
drafts
related
to
the
solution
and
one
is
interactive
routing.
Second,
one
is
ospf,
monitor
node,
because
this
topic
is
quite
broad
and
we
don't
see
any
dedicated.
P
Wg
can
hold
it.
So
we
tried
to
present
in
different
working
group
to
seek
more
commons
and
so
far
only
ospf.
We
we
know
we
we
have
to
go
to
asr,
but
for
other
things
we
still
don't
know
where
we
have
the
best
wt
riviera
6
comments
from
different
communities.
P
So
what
is
the
objective
for
our
work?
Our
work
is
for
the
to
explore
the
open
solution,
specifically
a
layer,
three
layer
for
large
scale,
lu,
constellation,
lu,
is
a
low
earth
orbit
and
any
problem.
P
Second,
is
that
we
assume
the
interstellar
link
is
used
right
now.
The
for
example,
the
starting
first
data
service
only
uses
a
relay
technology
which
doesn't
have
iso
leak,
but
right
now
they
start
to
expand
experimented
with
isl.
The
satellite
will
be
connect
together
and
form
a
network,
and
the
third
objective
is
that
the
satellite
network
has
to
provide
global
coverage.
P
It's
just
like
a
very
local
or
very
small
distance,
and
also
we
try
to
explore
solution
which
can
be
used
for
the
regenerative
modes
proposed
by
3gpp
in
this
mode
means
that
every
satellite
is
kind
of
player,
multiple
functions
from
tinobi
to
the
network
node
and
for
that
kind
of
requirements.
Ip
is
the
infrastructure
for
the
ntn
integration.
P
P
We
won't
seek
the
feedback
from
wg
and
obviously
this
topic
is
quite
complicated.
So
far
we
only
have
a
three
draft
about
solution.
Definitely,
it's
not
enough.
I
think
maybe
more
will
come
what
we
will
not
do.
We
will
not
do
anything
regarding
the
3gpp,
it's
okay
and
first
about
problem.
That's
the
first
draft
we
had
and
the
draft
may
not
be
is
a
quite
a
long.
P
So
if
you
want
to
see
more
explanation
by
slice,
we
have
a
workshop
which
was
done
to
the
university
of
missouri
and
the
slides
answered
most
of
the
questions
collected
in
icf
and
other
discussions,
and
most
of
the
questions
are
not
really
about
the
technical
is
about
why
we
do
this
kind
of
work
and,
secondly,
that
we
have
a
little
bit
detailed
introduction
about
what
3gb
png
ram
are
published
for
the
requirement
for
satellite
level.
P
P
And
this
is
a
second
about
a
second
draft,
a
semantic
select
address
and
for
this
ietf
we
have
some
updates,
add
a
call
and
also
add
the
32-bit
semantic
address
at
this
section
and
for
because
people
may
not
be
aware
of
this.
So
I
have
a
simple
introduction
here:
satellite
network
is
kind
of
where
all
its
moving
objective
objects,
so
it
has
multiple
layers.
Layer
is
kind
of
depends
on
the
altitude
of
satellites
and
each
layer
is
a
interleaved
grid
level,
which
means
percent
satellites
moving
in
one
direction.
P
Another
fifty
percent
of
satellites
moving
another
direction,
and
by
this
organize
the
network,
we
can
easily
think
about
to
identify
the
network
by
different
index.
Right
now
we
introduce
three
index.
One
is
the
shear
index
which
will
tell
which
layer
of
the
satellite
belong
to,
and
second
is
obvious
plan
index
which
will
tell
which
object
it
will
belong
to
the
third
one
is
a
seller
index.
There
is
a
tail
which
is
a
satellite
in
the
orbit.
P
For
example,
this
one,
the
is
the
the
force
index
is
zero
means.
The
shear
index
is
zero
means
all
altitudes
belong
to
zero
shares.
Second,
one
is:
the
pain,
is
orbit,
means
that
this
orbit
belongs
to
index
of
the
tank
of
the
orbit,
and
the
last
index
is
represents
the
which
satellite
in
the
orbit
next
slide
please.
P
P
The
lattice
draft,
which
we
propose
the
rotating
solution
for
satellite,
and
why
we
need
to
think
about
new
solutions
because
of
the
currently
distributed
routing,
algorithm,
igt
or
bgp,
will
face
a
lot
of
challenges
and
we
have
a
constant
satellite
ground
station
link
flipping
and
very
frequent.
Also,
the
constant
link
matrix.
P
Changing
because
of
the
link
distances
changing
and
also
the
constant
unsteady
link,
flipping
unsteady
link,
also
defined
in
in
the
problem
statement,
draft,
which
is
the
links
between
adjacent
satellite,
moving
to
different
directions
and
also
possible
link
interaction
at
the
follow
area
which
links
are
the
links
between
adjacent
satellites
moving
to
the
same
direction,
because
those
satellites
will
cross
over
at
the
polar
area.
P
C
O
P
Earlier
again
again:
okay,
okay,
next
slice,
please,
okay,
so
this
is
reveal
how
the
how
the
satellite
network
is
moving
through.
Topology
looks
like
we
have
a
satellite
moving
at
very
high
speed
and
the
50
percent
moving
to
north
to
source
and
50
percent
moving
from
south
to
north
and
the
link
matrix
between
the
iso
is
keep
changing
because
of
distances.
P
Keep
changing
and
also
ground
station
to
cell
link
will
be
frequently
changing
and
also
the
link
matrix
will
be
changed
besides
the
satellite
moving
the
ground
station
is
also
moving
with
the
earth's
rotation.
So
that's
why
the
topology
is
every
everything
is
moving
in
this
technology.
Next
slides,
please.
P
So
what
is
the
problem
when
we
use
the
igp
for
satellite
network?
The
because
of
link
number
is
huge
and,
for
example,
starlink.
We
have
over
one
billion
ground
station
to
sell
I
loop
and
every
satellite,
a
ground
station
to
satellite
link
will
be
flipping
in
about
five
to
ten
minutes.
That
depends
on
altitude.
P
If
we
use
a
traditional
igp,
this
will
cause
huge
number
of
lsa
flooding
before
network.
So
there's
a
simple
mass
which
can
roughly
estimate
that
every
second
we
will
have
more
than
3000
times
times
link
flipping.
Definitely
this
will
dramatically
reduce
the
service
time.
Here
we
have
the
service.
Time
definition
is
a
reduce,
remove
the
reconvergence
time.
Then
the
remainder
time
will
be
service
time.
It's
estimated
that
less
than
20
percent
for
a
network
next
slice
please.
P
C
If
I
may
comment
all
right,
you
may
want
to
relate,
you
know
how
this
work
would
be
relevant
in
dmm
to
dm
working
group.
Okay,
so
actually
you
yeah
another
five
minutes
if
we
can
wrap
it
up,
because.
P
P
Okay,
so,
basically,
when
I
presently
in
other
groups,
people
think
the
dmm
may
be
more
epic
and
some
people
even
ask
me
why
the
pmm
cannot
be
used
for
free.
So
that's
why
I
come
here
to
seek
more
comments
yeah,
so
I
I
quickly
go
through.
Then
the
seller
network
is
a
transport
network
and
and
also
the
constellation,
is
where
all
it.
Even
it
is
extremely
thailand.
P
P
So
we
have
a
think
about
the
some
solution.
Principles
first
is
that
we
have
to
maximize
the
usage
of
computation,
because
I
said
the
satellite
is
the
position
is
critical,
so
we
don't
need
to
use
a
protocol
to
dynamically
populate
those
informations,
such
as
the
network
topology,
link,
matrix
and
the
ground
station
to
satellite
link,
so
we
can
borrow
the,
but
we
can
borrow
a
current
igp
for
satellite
network
topology
and
state
detection.
P
For
example,
we
we
have
this
draft,
which
will
isolate
extremely
unstable
link
and
also
probably
more
draft,
will
come
out
for
the
past
and
we
will
utilize
the
special
characteristics
for
satellite,
for
example,
the
address
part,
and
also
we
want
to
minimize
the
rotting
overhead,
because
right
now,
with
all
those
information
routine,
is
a
steel
challenge
because,
as
I
said,
the
regular
routing
the
flooding
issue,
if
we
use
the
soft
rotting
and
second
routine,
the
overhead
is
cooking.
P
P
P
The
instruction
list
is
composed
of
the
function
and
arguments
the
functions
that
will
tear
fall
into
which
direction
and
until
which
satellite,
for
example,
this
incremental
satellite
id
means
that
the
incremental
direction
until
reached
a
satellite
id
equal
to
x17
and
the
secondary
instruction
is
that
the
forward
to
increment
the
direction
of
the
orbital
plane
index
until
the
orbit
plane
equal
to
c
and
third
one
is
incremental
direction,
forwarding
to
a
set
of
satellite
id
and
until
cell
id
equal
to
four
last
one
is
the
end
of
the
folding
and
the
the
package
to
the
ground
station
by
the
interface
id
next
slides.
P
P
Yeah,
this
is
how
the
user
plan
looks
like
if
we
use
the
tonerless
mode,
which
means
the
instruction
can
be
directly
embedded
into
the
some
extension
right
now
we
propose
to
use
the
rocking
header
and
then
the
packet
will
be
formed
at
the
ground
station
and
then
send
to
another
ground
station
and
for
3g
pp
point
view.
Then
of
course
we
can
insert
the
gtpu
tunnel
here.
Then
everything
is
the
same.
P
P
So
we
can
reduce
the
tcam
usage
because
the
tea
time
is
still
pretty
expensive
in
satellite,
because
satellite
environment
is
different
and
by
interactive
routing
it
will
has
much
less
size.
Overhead
and,
for
example,
overhead
will
be
number
of
segments
plus
one
times
two
octaves
normally
and
for
regular
satellites
network.
If
iso
is
not
broken,
then
very
limited
number
of
segments.
P
C
Thank
you.
Thank
you,
lynn.
I'm
not
going
to
take
any
questions
at
this
time,
but
I
think
lynn,
if
you
think
the
working
group
can
help
on
you,
know.
P
P
See
some
comments
from
the
dmm
because
the
people
trying
to
change
them
asked
me
to
come:
ask
the
dmn.
If
a
german
can
solve
the
problem,
if
not
cannot
solve,
then
this
definitely
is
a
work.
We
can
do
yeah.
C
Yeah,
let's,
let's
let's
discuss
on
the
mailer,
let's
see
you
know
how
it
relates
to
mobility
and
what
we
can
do
here
and
if
it's
in
the
chat.
So
with
that,
thank
you
so
much
and
next
topic
is
john.
Can
we
go
back
to
your
quick
quickly,
an
update,
yeah
thanks
again,
this
is
an
update
on
the
working
group
document.
John,
this
time
we
can
hear
you
please
go
ahead.
Okay,
great.
I
Okay
next
slide,
please
so
the
the
changes
in
this
I'll
try
to
be
as
brief
as
possible
in
the
interest
of
time,
thanks
for
giving
me
this
opportunity
to
present
again
so
the
the
new
draft
I
mean
updates
from
version
two
to
three:
is
that
the
comment
that
the
draft
text
is
not
focused
is
being
addressed
and
also
add
changes
to
reference?
I
The
drafts
on
slices,
for
example,
in
teas,
and
then
a
few
other
miscellaneous
changes.
So
if
I
could
go
to
the
next
slide,
please.
I
So
the
changes
to
focus
this
draft
is
mostly
in
the
introduction,
but
there
are
a
few
miscellaneous
changes
also
so
the
key
changes,
I'm
just
putting
a
few
snippets
here
to
give
you
a
feed,
a
look
into
what
has
been
done.
I
Basically,
the
slice
aspects
as
it
relates
to
3gbp
is
highlighted
here,
so
basically
that
the
slice
in
3gpp
is
a
logical
chunk
of
3gpp
resources
and
then
in
volatilized
form.
We
add,
you
know
why
we
need,
or
what
is
being
done
in
this
draft-
the
fact
that
3gbp
standards
do
not
address
the
ip
transport
network,
capabilities
or
slices,
and
so
on.
You
know
you
can
read
through
those
items
and
also
you
know
what
what
we
do
in
terms
of
the
network
interfaces
and
the
slicing
in
in
the
transport
network
and
how
it
is
mapped.
I
I
Okay,
the
actions
that
are
remaining
that
came
up
in
previously
were
to
share
this
with
other
working
groups.
That's
something
we've
got
to
do
and
then
I
think
we
would
like
to
get
more
detailed
reviews
within
the
this
working
group,
and
I
think
at
this
at
that
point
we
should
be
ready
to
move
forward
okay
and
address
everything
else.
C
Sorry
yeah.
Thank
you.
Thank
you,
john
for
your
presentation.
So,
from
the
author's
point
of
view,
you
think
the
document
is
in
a
good
shape.
You
know
ignoring
the
review
part
from
that
from
the
text
point
of
view
and
from
all
the
open
issues
that
have
come
up
so
far,
but
of
course
I
have
not
seeing
many
discussions,
but
at
least
from
your
point
of
view,
this
is
on
track.
I
Yes,
I
think
we've
addressed
everything,
that's
been
coming
up
and
has
come
up
and
focused.
It
also
yeah.
C
Okay,
so,
but
we
want
some
reviews,
some
serious
reviews,
john.
I
think
that
is
the
problem
right.
I
think
the
next
steps,
probably
we
should
focus
on
that.
I
O
C
C
K
Thank
you,
okay,
thank
you.
So
I'm
gonna
make
you
a
brief
update
about
your
mobile
user
friend
message,
including
so
please
go
to
the
next
slide.
So
here
is
a
change
from
here.
There
also
button
to
the
one
button,
so
we
define
admob
message,
totality
a
sequence
number
for
gtpu
message
and
also
added
flavor
consideration
sections
next
week.
K
This
is
a
change
from
zero
one
to
zero
two
buttons.
So,
instead
of
here
item
draft,
we
report
rfc
8754.
K
Next
speed,
so
this
is
a
change
from
territory
and
category
button.
We
added
security,
constellation
sections,
please
next
yeah.
So
this
is
a
change
from
here,
03
to
04
and
0.205
latest
version,
so
we
just
fixed
here
some
title
in
the
graph.
K
C
Questions
comments
all
right.
Thank
you,
murakami-san.
I
think
what
are
the
next
steps?
I
think
adoption.
I
think
we
need.
I
think,
let's
paul.
Let's
take
this
to
the
mailer
and
see
what
the
folks,
you
know
just
want
some
more
discussions
and
we
can
absolutely
initiate
a
working
group.
Adoption
call
in
the
next
before
the
next
eight
year
yeah,
but
if
you
can,
if
authors
can
trigger
some
discussions
and
make
sure
everybody
something
people
read
this
document,
I
think
that
will
be
great.
C
D
Just
thank
you
to
everyone
for
bearing
with
the
I
know,
the
issues
with
the
hybrid
meeting
style.
I
appreciate
everyone's
patience.