►
From YouTube: IETF106-DMM-20191118-1550
Description
DMM meeting session at IETF106
2019/11/18 1550
https://datatracker.ietf.org/meeting/106/proceedings/
A
B
B
Okay
welcome
this
is
the
DMM
working
group
meeting
I'm
shrinking
overly.
C
B
Please
look
at
the
IPR
policy.
There
are
RFC
references,
3,
3,
9,
7,
8
and
other
documents.
Please
look
at,
please
review
them
and
there
is
note
well.
This
meeting
is
dominant
of
you,
IET
of
rules
and
regulations.
You
need
to
be
familiar
with
these.
Please
be,
please
review
this
material
and
a
few
other
things.
We
need
a
note-taker
place.
B
D
Yes,
just
a
Hummer
from
entity
I'm
a
cosa
of
the
5g
user,
prey
on
Isis
draft.
So,
as
you
may
know,
3gpp
over
they
started
released
16
in
say
to
the
city
city
groups.
We
will
finish
the
release
released
16
in
London
next
much
so
we,
the
authors,
would
like
to
decide
the
range
of
the
goal
of
this
work.
So
we
continue
to
update
the
draft
within
ladies
16
and
after
that,
we'd
like
to
proceed
it
to
the
next
step.
We
believe
that
this
draft
will
be
help
for
many
idea.
E
B
Think
solutions
also
they're
all
straight
my
opinion,
so
yeah
I
understand
the
status
of
that
work
in
3gpp
for
release
16
right,
so
I
think
we
need
to
look
beyond
I
think
the
way
I
look
at
it
is
like
you
know,
so
what
we
can
do
in
the
release?
17
timeframe-
that's
that's
my
person
I'm.
Not
we
have
the
working
group,
but
Suresh
is
also
there.
Let's
hear
what
he
thinks
with
ya.
G
Subscription
responsibility
right
so
I
think,
like
you
know,
it
requires
some
kind
of
review
from
the
working
group
right
and
one
thing
we
can
do
to
speed
up
the
process
is
we
can
send
an
LS
right
like
if
the
working
group
decides
to
send
earlier
since
statement.
We
can
write
one
and
send
it
right,
but
I
cannot
see
anyway.
This
will
be
done
by
March
right
of
2020
right
like
so,
because
there's
like
reviews
from
the
working
group
complete
the
working
group
last
call
okay,
then
it
needs
to
come
to
me.
G
Get
the
Directorate
reviews
go
to
ITF
last
call
and
the
holidays.
Christmas
holidays
come
in
between
right,
so
it's
very
unlikely
that
this
is
gonna,
be
done
before
that.
Okay,
we
can
do
as
fast
as
we
can,
but
I
think.
If
you
want
to
inform
the
3gpp
officially
I
think
the
best
way
is
a
liaison
statement,
but.
B
H
I
think
this
case
on
our
intention
is
a
kind
of
the
year
the
others
work
based
on
the
end
of
the
state
of
the
end
of
the
March
not
to
complete
during
this
procedure
or
the
ITF
some
procedure,
so
I
the
Kokua
this
time
they'll
be
so.
We
won't
also
want
to
put
the
milestone
two
to
finish
the
walk
based
on
the
busy
16.
If
we,
they
need
to
think
the
force
analysis
based
on
the
17
that'll
be
another
chance
to
work.
B
B
B
Son,
this
is
a
presentation
start.
We
will
review
that
it
in
your
presentations
pocket.
Then
the
FPC
document
I
think
this
is
a
really
a
bizarre
situation.
It's
a
credible
amount
of
work
went
into
this
FPC
document,
but
overnight
it's
like
a
turning
out
the
lights.
Write,
complete
silence
on
this
FTC
stuff
right,
so
I
think
the
documents
are
expired
and
you
know
the
FPC
work
was
on
Charlie's
request
and
essentially
we
split
that
up
into
the
young.
The
young
portion
was
separated
out
right,
but
I
think
we
need
a
waveform
for
this.
H
B
G
So,
for
both
of
these
actually
the
first
to
like
the
the
shepherd
right
up
is
very
bad.
It
didn't
summarize
pretty
much
anything
about
the
working
group
process
and
I
really
cannot
put
this
in
front
of
the
iesg
in
that
state.
So
I'm,
whenever
you
one
of
you,
get
a
chance
to
update
the
write-up
I'll,
try
to
scale
a
little
shack
and
and
carlos
has
been
really
good,
catching
up
with
the
Directorate
reviews
and
addressing
them,
so
so
I
think
it's
good
and
the
last
one
it's
RC!
So
it's
like
you
know
it's
it's
a!
B
G
B
J
J
Derivation
of
seven,
we
beat
a
lot
of
editorial
updates
changes
to
the
terminology,
some
changes
to
improve
readability
and
unabated
referenced,
and
then
the
last
important
point
is
that
we
have
removed
the
support
for
an
instructor
to
use.
So,
as
you
know,
in
3gpp
there
are
depending
PD
users
346,
either
veteran
instructor
due
to
some
changes
that
we
have
having
as
hobby
six,
particularly
the
conservation
of
Internet
videos.
Now
we
can
no
longer
support
a
Star
Trek
videos.
This
could
be
done.
I
will
be
to
allocate
a
new
Internet
Protocol
number
this
content
year.
J
Then
the
Yoast
of
how
the
wheels
participated
in
the
hackathon
this
year
we
had
a
lot
of
progress.
So,
finally,
after
several
hackathons,
we
have
finished
the
DVDs
lamentations,
so
we
have
a
coupie
implementation
of
the
code
base
in
this
truck
and
we
have
submitted
a
disco
tutorial
sucks
when
the
next
weekly
release
that
will
be
generating
all
this
good
would
be
available
to
everyone,
and
that
said,
we
really
want
anyone
to
do
more
comments.
I
would
be
the
draft,
even
though
the
progress
being
relatively
stable.
We
believe
that
it's
ready.
B
B
So
this
is
a
a
new
talk,
your
topic,
this
is
it's
still
a
I
individual
ID
right
now
use
the
plain
message:
encoding,
zero,
zero
version.
L
Yeah,
yes,
the
a
motivation
so
sweetie,
sweetie
PB,
is
a
trainee
to
suppose
here,
user
premises
associated
with
you
at
the
TTP
youthful,
so
in
K
in
the
case
of
GSL
switch
to
user
day.
So
this
message
are
also
required
with
the
user
frame
interlock
with
the
HTTP
you
so
instead
it
has
a
community
definitions.
So
author
field
has
a
capable
to
indicate
the
a
different
profile
you
it
is
SAT
there.
Also
oh,
isn't
it
here.
L
We
have
the
capability
to
not
see
a
user
community
has
a
capability
to
provide
the
a
metadata
to
see
a
polygon,
so
this
capability,
LTS
rate,
should
be
leveraged
to
not
hear
you.
The
pregnancy
include
it's
a
bit,
no,
so
so,
and
also
in
this
approach.
So
there
is
no
internal
header
or
extension
headers
totality
any
user
privacy.
L
L
L
L
And
also
some
of
user
Craig
message
recoil
is
here
some
additional
information
arrogance.
So
all
this
information
element
can
be
encoded
with
the
SLS
Tier
II,
so
we
are
proposing
to
at
IPS
container
key
away
as
a
little
XD
away,
and
also
this
theory
can
can
see
a
multiple
of
V
DP,
acidity
I
eat
here
with.
L
L
So
user
per
message
can
be
carried
with
the
insulate
aside.
What
is
you
know
the
in
order
to
carry
a
user
pay
message,
although
it
is
a
bit
narrow,
so
instead
it
s
on
Heather
must
is
sustained
here,
although
inside
of
TSA,
which
never
so,
however,
so
if
enabling
tear
PSD
so
or
in
select
header
is
ophtho,
as
he
apparently
signals.
L
So
in
order
to
avoid
here
this
problem,
so
one
solution,
eg
a
PSP
must
not
be
used
in
das
a
now
in
order
to
keratea
use
a
pregnancy
or
if
in
identity,
a
PSP,
so
another
internet
should
be
added
to
Cali
ta
use
of
a
message
on
top
of
Gaea
author
ID,
which
in
a
war
a
celestial
yep
so
OT.
Our
next
step
will
like
to
adapt
here
this
dots
with
your
walking
the
dog.
L
N
Ma
hi,
my
name
is
Omar
from
future
USA
I
read
the
draft.
What
is
the
main
use
case?
I
didn't
understand
for
this.
Like
you
know,
you
won't
remove
gtp
completely
with
us
our
v6,
so
some
parts
are
removed
with
the
current
last
call
document.
Right,
I
think
some
part
of
the
gtp
with
that's
our
basics.
This
is
a
complete
elimination
of
gtp.
You
is
it
character.
What
is
it.
N
L
N
Also,
my
was
just
like
it's
not
written
what
I
was
thinking.
Is
there
any
use
case?
You
are
thinking
to
send
the
message
with
the
gtp
encoded
message
right
from
the
UE
itself.
Is
there
any
use
case
kind
of
that
kind
of
thing
you
are
thinking,
or
this
is
not
from
the
you
a
possibility
right
with
this.
You
have
the
real
possibility.
You
extend
all
the
way
till.
B
G
Suresh
suresh
krisshnan
Just,
Answer
Roma.
The
reason
why
you
don't
want
to
do
that
is
like
the
spectrum
is
very
expensive
right,
so
you
don't
want
to
put
in
all
the
headers
from
where
the
air
interface
right.
That's
why
the
gtp
is
starting
from
the
base
station
towards
the
core
network,
not
from
the
UE.
N
B
O
B
B
I
think
this
is
a
logical
extension
to
the
earlier
work.
I,
don't
see
any
problem,
I
think
this
addresses
rather
completes
the
whole
world,
but
we
need
more
feedback
for
the
working
group
and
also
you
know
from
the
ad-
and
you
know
to
make
this
right.
I
think
my
other
suggestion
is
engage
without
keeping
the
world
can
like.
Essentially,
like
you
know,
we
can
in
our
trade
off
the
document.
Okay.
Thank
you.
A
N
Okay,
just
a
quick
recap
of
this
wrap.
This
has
been
presented
first
in
2018,
I
give
honor
to
0-0
version
because
later
mean
problem
and
strong
and
solution.
Also,
it
started
updated
in
after
updated
and
present
in
the
night
economy,
and
a
major
update
happened
in
last
night
here
with
the
original
solution
and
two
approaches
that
we
presented
to
put
a
framework.
So
they
were
multiple
options
proposed
to
canopy
MCRD
proposed
like
a
header
and
a
tent
in
CID.
N
One
is
more
satisfying,
but
the
problem
is
the
transport
networks.
There's
no
transport
Venice
for
some
of
let's
take
some
of
the
spheres,
so
this
is
gives
a
default,
mirror
our
voice
better.
Like
LTE,
it's
not
an
issue,
but
different
traffic
characteristics
are
needed
from
the
transport
network.
For
example,
low
and
deterministic
latency
are
real
dimensions
with
a
lot.
Dpr
characteristics
that
need
to
be
need
to
be
fulfilled.
For
some
of
the
traffic
capacity
types
transport
network
need
to
be
awareness
will
be
there.
N
So
the
proposed
solution
in
the
last
one,
the
last
IDF
to
here
this
is
the
last
idea
solution.
So
we
deduce
the
concept
of
John
novelist,
because
the
concept
of
available
context
identified,
which
carries
the
to
reflect
the
slice
characteristics
like
isolation
man.
You
know
that
dealing
at
the
sticks
and
slice
characteristics
mapped
in
the
PLC
identifier,
and
this
MPN
see
added
permittivity
in
is
expected
hi
packet.
It
could
be
a
solution
that
ended
SRB
fix.
N
If
it
is
the
overall
resolution
GD
you
can
be
used
and
there
are
also
other
so
other
possible
solutions
are
listed.
We
also
comment
from
DMM
what
you
wrote
that
you
know
we're
going
to
use
the
lower
64
bits
of
ipv6
address
and
some
of
the
comments
from
peace.
What
you
hope
is
why
don't
you
use
the
fast
because
kind
of
solution
that
is
which
is
this
big
by
giving
six?
N
B
N
N
P
So,
as
I
mentioned,
there
were
two
different
solutions
and
one
of
the
suggestions
was
to
come
up
with
one.
So
we
took
the
discrete
and
the
integrated
once
in
the
earlier
one
and
have
one
single
approach,
a
simple
and
that's
all
in
Chapter
two
and
the
second
part
of
it
was
what
required
more
work
section
two
here
carrying
the
transport
context,
invisible
packet
header.
So
we
had
a
number
of
options
are
here
on
gue
and.
Q
P
P
So
here
we
have
the
four
steps:
I
think
the
first
step
is
partly
3gpp
work
and
we
we
have
specified
anymore.
They
made
clear
that
we
mentioned
that.
It's
not
idea
to
come
up
with
this
mapping,
so
somewhere
I
mean
in
3gpp
that
there
are
these
functions
like
MSS
and
networks
like
that
is
defined
in
the
kv-85
30
and
they
identified
the
3gpp
and
that's
whenever
functions
and
a
transport
network
in
between.
So
that's
roughly
what
we've
done.
P
P
So
the
problem,
in
essence,
are
going
back
to
series
question:
is
that
the
3gpp
slice
aspects,
the
NSS
ai,
the
networks
like
selection,
resistance
information
is
related
to
what
a
user
is
supplying,
and
so
the
customer
for
the
3gpp
network
is
the
UE
and
the
customer
for
the
transport
network
is
the
3gpp
provider,
so
somehow
has
not
been
mapping
between
was
offered
to
the
customer
in
the
treaty.
P
become
a
plus
is
what
the
transport
network
has
to
couldn't
we,
the
those
are
not
the
same.
There
is
a
it's,
not
a
one-to-one
mapping.
P
P
Yes,
we
assume
that
actually
that
that's
a
good
question.
We
assume
that
it
is
not
made
third
session,
so
the
estimation
of
whatever
there
are
various
functions
which
we
typically,
that
estimate
over
longer
periods
of
Chinese
and
that's
programmed
in
the
transport
network.
So
when
a
session
comes,
it
doesn't
try
to
go
and
figure
out
other
resources
and
it's
been
allocated
and
in
only
the
mapping,
information
is
conveyed
and.
B
So
maybe
on
elementary
question
right
now,
this
is
I'm
going
to
understand.
You
know
not
a
name
so
today,
if
you
like
DHCP,
and
all
of
that
you
do
mapping
at
a
session
layer
alright.
So
now
we
are
dealing
at
a
slice
level.
What
are
the
pros
and
cons
of?
What
does
that
burn
days
of
being
this
way
versus
on
the
session
layer
session?.
P
So
that's
one
answer,
so
we
look
through
a
number
of
alternate
is
just
to
see
if
we
can
use
what
exists
you
know,
and
even
so
that's
how
we're
coming
to
do.
So
if
you
go
through
it,
this
the
tcp
is
more
beautiful,
VLAN
I.
Suppose
it
works
in
some
scenarios
where,
if
you
have
a
customer
and
the
provider
in
the
same
network,
it's
okay,
it's
not
they.
Three
PN's
use
that
I
mean
also
works
in
some
scenarios,
but
not
in
many
others.
Where
you,
you
know,
it's
very
rigid
provision
else.
O
P
P
H
Just
a
question:
that's
question,
auto
farm,
so
it's
interesting
to
to
when
we
see
the
P
out
to
be
able
to
look
up
the
UDP
source
code
source
port
to
handle
which
tries
pocket,
bringing
bring
doubt
so
on
so
which
working
group
will
stand.
Let
somebody
walk
on
the
the
dot
P
router
to
be
able
to
handle
source
port.
P
H
My
question
is:
ID
is
a
place
to
just
on
walk,
walk
on
you,
yes,
on
dot
for
the
mobility
monument
in
distribution,
mana
and
act
on
work
on
the
truck
engineering,
and
you
now
propose
the
UDP
source
port
other
yeah,
the
ID,
to
mapping
the
mobility
session
to
the
transporter
resource.
So
so,
which
idea
who
are
working
group
to
define
that
behavior.
N
There
is
no
what
Cooper
say
here:
it's
a
local
mapping
function
and
the
PE
nodes.
How
do
you
map
the
incoming
session
to
this
local
policy
based
routing
right
today
to
falsely
based
routing
based
on
various
packet
fields?
You
map
this
information
incoming
traffic
to
map
to
the
tunnels.
So
what
we
are
suggesting
is
this:
it's
not
a
new
thing
putting
in
the
UDP
source
port
that
is
used
for
today.
N
For
the,
for
example,
what
Dave
is
asking
for
load
balancing
at
both
English
side
and
that
ran
side
are
also
on
the
UPF
side,
both
sides.
So
we
are
latching
on
to
this
mechanism.
We
are
extending
one
more
indirection,
but
some
of
this
extra
indirection
we
created
to
encode
this
empty
anxiety
indirectly.
So
there
is
no
working
group,
it's
already
working
on
working
fields.
So
that's
why
this
approach.
We
felt
that
it's
backward
compatible
even
with
existing
4G,
slicing
solutions.
G
N
Encoding
happens
at
the
G
note
P
or
the
UPF,
so
we
are
not
doing
any
additional
stuff
which,
if
you
see
that
what
is
happening
is
this
operator
control.
This
particular
range
of
the
UDP
port
map
it
to
this
particular
empty
and
CID.
So
there's
no
rewriting
the
first
time
and
the
GT
encapsulation
happens.
This
mapping
happens
so
what
happens
in
the
pierrot
trees
only
lookup
and
map
it
to
the
particulate
eternal.
So
there's
no
okay.
P
E
This
is
Yasuo
and
I.
Think
the
v/line
eye
team
should
not
be
excluded
in
this
solution,
because
in
the
first
hub
from
the
run
to
the
P
note,
the
villain
can
represent
the
network
slice
information
and
in
the
IP
domain
we
can
use
other
NATO
slice
encapsulation
to
respect
the
information,
so
I
think
is
a
reasonable
solution
to.
E
E
R
Mahoney
purchases
for
correct
understanding,
so
the
since
UDP
is
a
be
directional
thing.
So
both
end
points
of
the
GDP
tunnel
have
to
use
the
same
transport
network
identifier
in
the
source
port
or
is
it
can
they
be
different?
Can
you
have
multiple
of
these
identifiers
to
unambitious
li
identify
the
slice
that
the
traffic
is
associated
with.
P
P
T
P
P
B
P
B
B
E
One
comment:
I
think
the
the
last
part
of
the
work,
the
how
to
map
in
the
user
plan
from
the
run
to
the
IP
backhaul
I
think
this
is
whether
this
part
of
work
can
be
separated
from
the
current
draft.
I
think
this
is
some
isolated
work,
and
this
is
very
important
for
the
network.
Slicing
mapping
relationship.
You
know,
data
plan.
This
is
just
a
suggestion,
so.
E
P
P
K
B
U
F
U
I'd
like
to
address
a
candle
with
a
level
4
systematics
I
need
to
be
looking
at
the
holistically
and
across
layers
when
drastic
architectural
transition
is
required,
for
example,
in
the
5c
era
we
are,
various
decisions
are
to
be
supported
and
we
are
completely
new.
Data.
Intensive
on
services,
like
an
elastic
narrative,
is
to
get
it
within
our
soon.
This
document
discusses
the
architectural
implication
of
our
prices.
Are
the
six
memory
user
plain
not.
U
Six
button,
especially
regarding
the
possible
optimization
work
with
existing
conventional
players.
Yes,
we
hope
it
suggests
that
Excel
basics,
mobile
user
claim
is
a
right
afflicted.
Choice
for
E
and
Sierra,
so
mobile
network
and
the
transport
network
are
separated
by
layers
convenient
Food,
Network,
it's
under
a
layer
to
the
mobile
network
and
GDP.
You
is
an
over
a
tonal
in
the
movie
team
network.
U
So
let's
take
an
example
of
network
slicing
into
a
network
sizing.
So
in
typical
TTP
you
over
IP
MPLS
SSL
configuration
are
3.
Gpp
data-planet
entity
such
as
UPF
is
a
c
e
to
the
transport
networks,
are
PE
or
even
needed
extra
see
each
lecture
scale
on
which
of
these
are
paper
to
analysis
easier
in
this
nicely
explained
the.
U
U
Extra
is
needed
for
secure
getting
traffic
and
nothing
also
artistic
sliced
in
the
same
week,
and
also
TLP,
is
in
ze
connection.
It's
a
sin,
organ
failure,
and
so
some
form
of
be
a
leader.
You
can
see
using
nothing.
Protocol
or
ancillary
kind
of
things
are
required,
which
complicates
the
architecture.
H
U
U
U
From
the
perspective
of
implementation,
Cameron
says
it's
unrealistic,
630
PP
SD
that
don't
care
about,
resulting
but
I
would
say
it's
realistic
and
even
it's
more
national
in
the
new
generation,
since
I
still
use
cases
in
I
own
flexibility
of
crowd,
negative
and
in
the
clouds
negative
as
hugs
mich
Mich
deployments
aside,
the
cod
proxy
on
the
same
solution,
good
or
the
second
copula
port
would
take
care
of
the
communication.
Thank
you
for
coming
for
more
such
kind
of
things,
so
that
service
can
be.
U
U
And
for
each
computer,
it's
a
v6
fictional
traffic's,
general
capabilities
and
the
networking
concept
for
all
freely
describing
instructions
and
metadata
RC
suitable
for
providing
it
distributed.
Computing.
In
addition,
since
is
a
v6,
can
be
accommodated
rain,
regardless
of
a
domain
such
as
access
by
the
area,
mobility,
or
there
is
a
datacenter.
Some
spreads
memory,
and
some
Cheney
that
used
to
be
concentrated
in
data
center
historically
can
be
expanded
over
our
by
the
area.
U
It
is
good
for
and
with
a
certain
six
system
and
the
QoS
information
can
be
exposed
in
the
IP
header,
which
does
not
affect
performance
with
thanks
to
the
wrong
estimate.
Genetic
mechanism
is
I,
figured
grunting,
so
only
the
services
or
applications
will
need
to
meet
that
information
for
granular
processing
to
look
up
that
kind
of
information,
so
this
would
are
effects
replacement
oxide
applications
and
for
you,
our
energy
into
undeniable
low-rated
see
communications
3gpp
tier
two
three
is
seventy.
U
U
Such
a
survey
included
and
the
DA
themes
are
based
on
the
static
configuration
or
static
mapping
for
steering
that
over
a
parking
garage
across
medicinal
and
such
solutions
do
not
scale,
and
he
does
about
economics
and
that
document
six,
one
six,
two
six
two
four
addresses
the
issues
on
how
to
support
redundant
data
transmission
via
a
single
uvf
and
signal
our
landlord,
but
how
to
ensure
that
this
do
it
fast
is
not
enough
in
the
government.
It
just
says
using
different
IP
address
or
different
ever
instances,
but
it's
not
assure
that
the
path
is
disjoint.
U
Hana
GDP
data
plane
entities
need
to
support
pack
identification
and
elimination.
This
actually
requires
the
modification
of
gtp
you
protocol
itself,
so
if
it
need
to
evolve
CPU
independently,
but
I
would
say
it's
not
good.
It's
too
much
the
SLV
six
mobile
user
plain
also
has
an
advantage
in
this
list.
You.
U
P
U
P
G
Suresh
yeah
suresh
krisshnan,
so
one
question
konasana
like
this
is
I
understand
why
you
are
doing
this
right,
I'm,
trying
to
understand
why
it's
a
separate
document.
So
why
is
it
not
something
that's
folded
into
the
satoru
sons
document
or
homo
sons
document
as
like
either
an
analysis
either
on
the
solution
to
say,
like.
U
G
I
I
think,
though,
and
that
I
think
the
text
itself
is
useful,
I'm,
not
sure
about
the
form
of
how
it
needs
to
go.
So
I
think
it's
probably
worth
my
discussing
how
it
actually
proceeds.
Ok
because,
like
as
you
like
know
right
a
lot
of
the
times
like
you
know
when
you
have
a
document
like
this,
which
has
like
it's
pretty
dependent
on
other
stuff
right.
G
So
the
question
is:
why
should
it
be
a
separate
document
right-
and
this
is
something
that
comes
up
in
the
iesg
I
told
straight
right,
like
we
have
an
IE,
AC
statement
right
on
what
is
called
support
documents,
so
I
would
really
like
the
working
group
to
think
about
it,
and
if
the
working
group
decides
it
needs
to
be
separate,
that's
fine,
but
I
just
want
to
put
it
up
there.
So
you
think
about
like
why
it
needs
to
be
separate
and
put
it
in
the
write-up
480
separate.
Okay,
thanks.
B
Absolutely
I
think
great
points,
I
think
I'll,
just
chime
in
there
right
I
think
you
brought
out
number
of
good
points
there.
I
think,
but
salacious
cushion
is
a
very
fundamental
to
this
in
the
sense
that
there's
already
two
documents
right
now.
If
we
can
fold
some
things
there,
it's
great,
if
you
still
have
some
good
stuff,
which
may
you
know
if
you're
given
justify
that
you
know
we
need
one
good.
One
document
which
you
know
covers
the
overall
thing.
B
N
My
name
is:
sumo
furious
a
so
I
read
this
document.
It's
she's
raising
very
interesting
overall
big
question
to
the
community.
Actually
she's
asking
I
think
now
I
understand
better
the
second
presentation,
because
it
has
this
directly
connection
right
like
like
you,
can
collapse,
gtp
you
and
SR
v6,
and
you
can
have
much
more
transport
benefits.
You
can
get.
That's
all
good,
but
the
only
question
I
have
is
so
that
we
are
seeking,
like
you
know
the
it's
kind
of
envisioning
in
future,
only
ipv6
based
transport
right.
So
what
happens
to
the
backward
compatibility?
N
These
kind
of
things
I
said:
I
think
this
recipe
be
stuck
with
gtp
right
like
and
with
the
discussion
started.
They
were
only
using
an
S
for
new,
they
move
to
a
nine
now
and
they
were
new.
They
moved
gtp
u2f
anew,
so
they
are
expanding.
Gtp
usage.
While
we
are
discussing
you
know
how
to
give
the
much
more
transport
benefits
right.
So
there's
some
tussle
going
on
there.
So
I
think
this
is
a
good
document.
I
think
you,
you
laid
out
the
points.
N
F
U
F
U
Doesn't
know
routing
doesn't
know
it's
not
v6.
That
kind
of
argument
could
happen,
but
I
would
say
just
if
there
could
be
a
kind
of
proxy
or
sidecar,
which
feat
can
only
do
the
routing
for
getting
stuff
right
now,
rafting
on
the
console
implementation,
so
the
application
doesn't
know
the
wrong
thing,
but
another
thing
sits
in
de
host.
In
that
case,
we
can
provide
being
probably
from
the
host
okay.
F
V
Just
want
to
respond
to
the
same
things
approaching
no
Cisco,
so
I
think
it
was
actually
really
like
that
analogy
of
site
car
I
related
to
this
specific
content
in
the
drop,
because
the
site
card,
this
whole
motion
is
to
allow
for
the
proxy
to
sit
just
sank
to
the
application
themselves.
So,
if
you
think
about
these
different
network
functions,
these
different
functions
related
to
the
Ryan
or
than
the
FRG
Seaside.
These
functions
are
now
capable
of
hosting
where
they
are
sitting
which
are
hosts.
V
F
Site
cars
are
functioning
at
the
HTTP
level,
not
assert
v6,
which
is
quite
low
in
in
the
protocol
stack,
so
I
still
I'm
on
the
opinion
that
we
are
mixing
things
now
very
bad.
If
there
is
application
level,
routing,
HTTP
level
request
routing,
that's
the
business
of
sidecar
and
then
we
have
this
services,
which
is
an
IP
level
or
maybe
even
below.
G
Just
like
closing
off
on
this
one,
all
right
like
so
the
way
I
understood
it
is
right,
like
you
know,
the
the
the
cyclic
pattern
as
a
design
pattern
allows
for
the
application
to
be
unaware
of
what's
happening,
underneath
right.
So,
as
I
said,
it's
right
now
used
for
HTTP
like
certificate
like
offload
and
things
like
that
for
the
security
but
I
think
konasana
spoink
was
like
you
know.
If
your
s
r
v6
is
like
put
in
a
sidecar
pattern,
then
your
application
doesn't
need
to
know
about
it.
G
B
N
B
W
Okay,
thank
you
and
the
other
authors
of
this
draft
isn't
only
from
Hawaii
and
li
geom,
her
and
actually
I
guess
the
noona
has
pride
in
did
this
draft
last
time
and
we
did
a
lot
of
extensions
and
polish
it
a
lot
and
now
I
will
introduce
it
here.
W
And
we
all
know
that
the
ATF
and
other
organizations
has
standardized
my
new
mobility
management
protocols
and
we
can
put
him
into
two
categories:
the
network
based
solutions
and
the
host
based
solutions
and
in
the
network
based
solutions,
the
main
protocols,
other
PM
IP
based
protocols
and
the
extensions
and
similarly
for
the
host
based
protocols
and
mainly
based
on
the
movie
accurate
six
and
some
extensions.
So
a
question
here
is
the
item.
W
Okay
and
the
window
mobile
node,
access
to
the
network,
maybe
the
first
time
access
all
the
handover.
What
kind
of
mobility
management
protocol
can
be
used
and
because
of
the
network,
may
support
some
protocols
and
the
host
miss
supported
some
protocols
and
maybe
the
host
to
pray
for
something
and
different
from
the
capability
of
the
network.
So
this
charter
mainly
research
on
this
problem
and
to
choose
Portland
negotiation
of
the
mobility
capability
between
the
host
and
the
network.
W
So
we
may
have
some
basic
principles
here.
Is
that
the
first
one
is
to
during
the
negotiation
when
to
follow
the
network
ability,
and
the
second
is
follow
the
host
preference
and
third,
a
supporter
of
functional
extensions
and
the
first
one
is
to
support
the
performance
enhancements.
For
example,
some
fast
handover,
or
something
like
that
and
in
default
in
order
to
guarantee
are
the
performance
of
the
mobility
management.
The
network
based
schemes
can
be
should
be
used
if
it
can
be
supported.
W
So
after
the
initial
attachment
of
the
mobile
node,
a
profile
will
be
generated
in
the
management
store
to
recover.
The
selected
I'll
pray
for
the
protocol
of
this
host,
and
when
the
handover
happens,
the
network
network
will
check
the
selected
or
preffered
protocol,
but
the
network
also
need
to
notify
the
host.
If
the
slack
here
the
protocol
cannot
be
spotted
in
the
new
area,.
W
And
we
have
some
examples
in
the
draft
and
I
at
least
one
example
here
is
that
a
host
and
network
post
supports
a
network
based
protocols
and,
for
example,
when
the
host
of
support
to
the
basics,
proxy
mobile
ipv6
and
the
network's
for
some
extension
extension
solves
the
based
on
the
proxxon
mobile
ipv6.
Then,
if
the
extension
protocol,
no
parity
racing's
has
no
requirement
for
the
mobile
node.
In
a
moment
it
can
be
used,
and
otherwise
the
basic
theme
of
the
ipv6
should
be
used.
W
And
the
second
case
is
is
the
item
when
the
network
and
the
host
post
is
supposed
some
extension
protocols
and
if
the
extensions
are
same,
it
can
be
used.
Otherwise,
the
basic
approximate
actives
proximity
heuristics
can
be
selected,
and
then
we
need
some
solutions.
Here
is
that
the
signaling
messages
should
be
exchanged
between
the
host
and
the
network
to
exchange
this
preference
and
the
ability
and
may
be
based
on
the
extension
of
the
icmpv6
or
some
authentication
messages,
and
another
problem
we
need
to
consider
is
who
will
initiate
this
negotiation?
W
And
we
give
out
a
possible
solution,
is
based
on
icmpv6
is
a
new
option
here
is
proposed
on
it,
namely
mobility
capability,
MC
option
with
the
corresponding
flag
in
the
eyes,
a
message
when
the
mobile
node
and
network
exchange
of
the
eyes
I
a
message.
The
private
preference
and
the
supportive
protocols
can
be
can
be,
can
be
included
in
this
option
and
then
the
protocol
get
disliked
either
final
yeah.
G
Okay,
suresh
krisshnan
ass,
like
responsibility
here
right,
so
one
of
the
things
is
that,
like
everything,
you
say,
has
been
tried
before
the
idea
yeah
multiple
times.
So
if
you
go
back
to
like
2007-2008
the
net
lmm
mnar
interface
is
exactly
this
okay,
except
that
it
never
worked
because,
like
a
network
usually
decides
if
the
network
decides
is
going
to
do
some
kind
of
mobility,
it's
gonna.
Do
it
right.
So
it's
not
going
to
negotiate
with
the
host.
Why
it's
gonna?
Do
it
that's
kind
of
something
we
found
out
like
as
a
result
right?
G
It's
not
that
we
didn't
try.
So
there's
been
multiple
iterations
of
this
I
know.
There's
one
like
Raj
and
I
did
for,
like
you
know
see
my
purses,
pmf
is
something
you
tried
like
she
did
some
prefix
coloring
stuff,
so
there's
been
like
multiple
failed
attempts
at
this
right,
mainly
because
of
not
technical
reasons.
Okay,
so
there's
like
you
know,
it
doesn't
fit
any
deployment
scenario
there,
like
so
I'm
positing
that
there's
no
deployment
scenario
where
this
would
be
useful
right,
because
if
you
take
the
architectures,
none
of
them
are
made
to
allow
this.
Okay.
G
So
if
you
can
figure
out,
like
you
know
where
you
would
use
this
like
you
know,
who
are
the
people
interested
in
this
I
think
we
can
pursue
this,
but
otherwise,
like
we
know,
there's
like
a
lot
of
past
history
of
failures
in
this
thing,
because
the
incentives
are
not
aligned
for
people
to
do
this.
Okay,
so
think
about
it,
how
you
would
apply
this
where
you
would
apply
this
I
think
that
could
be
very
useful
input
for
the
group
to
decide
if
it's
useful.
B
Thanks
I
think
that's
a
great
time.
Input,
I,
think
er
interface
document
that
was
Julien,
published
it
and
say
I.
Think
the
network
to
end
point
interface
was
one
thing
we
truly
missed.
You
know
we
don't
have
it
even
today
right
in
any
of
this
right
having
a
standardized
interface
makes
sense,
but
we
could
not
somehow
we
always
feel
we
tried,
but
we
fail
so
so
I
think
it's
the
question
is
you
know
we
need
to
see.
You
know
we
need
to
look
at
the
previous
work
and
see
what
we
can
do
now.
B
F
B
V
Just
one
quick
one
minor
one
Rajeev
Cisco,
so
if
you
think
from
the
endpoint
being
attached
to
cellular,
then
I
presume
this
message
would
go
from
potentially
the
end
point
towards
the
UPF
of
the
side
gateway
in
case
of
the
mobile
area,
happening
between
cellular
and
say,
Wi-Fi,
then
this
message
is
assumed
to
be
going
from
the
endpoint
to
where
it's
the
same
thing:
access
rotor
it.
So
there
may
be
two
distinct
right.
Excess
water
or
in
case
of
EP
digit,
might
be
the
same
thing
as
a
psychic
way
when
upstairs.
V
But
then
there
are
number
of
possibilities.
So
this
this
notion
of
having
something
in
the
RS,
an
RA
would
mean
to
allow
for
the
nobility
to
work,
have
to
be
comparable
at
not
just
within
the
cellular
realm
but
outside
as
well,
potentially
something
to
think
about,
though
I
do
like
the
fact
that
this
is
not
a
sari.
So
it's
really
low
hangings,
small
footprint
in
terms
of
protocols,
extension
and,
thankfully,
doesn't
require
anything
to
be
done
in
the
mobility
protocol.
So
that's
one
difference
thing:
I
see
from
the
past
Suresh.
C
N
G
G
The
architectures
are
done
somewhere
else
right
like
so,
unless
you
go
and
anchor
this,
like
the
other
places
where,
like
I
think
like
you
know,
I
know
is
like
she
has
the
same
scars
right,
so
we
have
to
go
to
3gpp,
go
to
3gpp,
to
like
and
everywhere
to
make
sure
that
our
stuff
is
anchored
there
right,
because
without
that
happening,
these
things
are
not
going
anywhere
right.
So
that's
kind
of
the
path.
That's
really
missing
is
really
the
deployment
support
for
this
thing.
So
Security's,
like
one
thing,
we
can
fix
it.
G
So
if
you
look
at
the
Julian
and
Pete's
traffic
I,
which
is
am
an
IR
interface
like
everything
Rajeev
said,
was
actually
looked
at.
We
spent
like
a
couple
of
years
on
it
address
everything,
except
that
it
was
not
deployed
right.
So
it's
I
think
it's
mainly
a
deployment
issue.
The
incentives
are
not
right
to
do
this
like
so.
Nobody
was
really
interested
like,
for
example,
like
Qualcomm,
was
doing
the
chip
and
they
didn't
care
about.
B
H
H
2014
I
think
so
DVM
the
working
group
adopts
that
document
as
a
working
group
document.
So
we
work
a
lot
to
to
update
the
document
spec
based
on
the
our
discussion,
how
to
configure
the
foreign
policy
for
the
user
playing
data
plan
note.
So,
as
a
result
of
the
our
discussion
we
are,
FPC
did
not
touch
any
outside
what
a
wire
protocol
so,
instead
of
that,
we
defined
API
to
configure
the
user
playing
from
the
application.
So
so
that's
the
reason
we
decide.
We
decide
to
define
the
we
hang
model
to
provide
the
API.
H
So
after
the
a
discussion
javi
around
the
2016
after
the
Bursar's
working
on
meeting,
so
we
decided
to
change
the
data
model
and
the
information
model
to
divide
to
the
topology
and
mobility
and
policy
to
those
three
model
conjunction
to
derive
the
action
forum
post
configuration
to
the
to
the
data
plane
node.
So
after
the
zero
for
revision,
we
take
that
model
and
if
you
see
document
describing
the
information
model
of
that
and
data
mode
data
model
order
by
Jung
product
information
model.
H
So
after
that
decision
we
make
bunch
of
the
division.
Maybe
yes,
yeah
have
a
revision,
though
I
think
it's
being
mature
and
live
from
Sprint
take
out
with
a
walk
to
edit
the
document
and
the
Chari
volunteer
to
to
divide
the
young
model
from
the
main
document
to
make
the
documents
short,
because
the
the
hold
of
the
document
becoming
over
the
hundred
page.
So
it
makes
the
viewer
really
difficult
to
make
write
comment
and
to
decide
whether
it'll
be
good
or
not.
H
So
that's
the
latest
status
of
the
the
FPC
work
after
the
the
last
update
of
the
young
model
by
charlie.
There's
no
update
happened.
So
then,
the
boss
of
the
information
model
document
and
yang
model
document
is
expires.
That's
the
latest,
a
total
of
the
FPC,
so
so
the
reason
why
I
shared
in
the
background
of
the
FEC
work
after
the
long
silence
of
of
the
last
update,
so
how
to
be
happy
to
discuss
whether
we
proceed
or
not
the
defi
C
work
into
the
DMM
working
group
with
with
you
guys,
I.
H
Think
many
people
put
the
much
effort
to
this
work.
At
the
beginning
of
the
discussions.
A
the
control
frame
is
a
frame
or
data
plane.
Separation
so
I
think
to
be
good
outcome
here,
but
it
depends
on
the
interest
of
the
industry
or
community
for
the
mobility
area.
I'd
like
to
hear
your
comment,
please.
R
So
I'm
co-author
and
I
know
that
the
work
that
we
put
into
this
that
would
just
been
revised
in
details
quite
a
lot
and
but
it's
pretty
mature
as
well
pretty
stable
since
a
while,
it's
a
stable
as
drafts
have
been
expired,
so
I
think
there
was
a
plan
how
to
further
progress
the
reviews
on
these
two
documents.
Now
there
are,
there
was
one
document
which
was
version
12.
Then
we
split
it
wait
and
I
think
only
the
young
part
has
been
published
so
far
and,
as
you
said,
your
opinion
is.
R
R
B
Okay,
so
I
think
we
will
confirm
the
same
in
the
mailing
list
if
I
think
this
sufficient
interest
will
move
forward
with
the
obligation
process
or
a
complete.
The
last
call
forces,
but
we
definitely
I
do
think
the
document
is
in
a
great
shape,
but
still
you
know
we
need
to
go
through
the
process.
We
need
reviews.
I.
Think
it's
not
just
authors
to
you
right,
but
in
the
working
working
group
has
to
say
that
I
think
that
that
hasn't
been
established.
If
it
that
happens,
we
move
it
forward.
Otherwise
we'll
kill
it
Suresh.