►
From YouTube: IETF101-IPWAVE-20180319-0930
Description
IPWAVE meeting session at IETF101
2018/03/19 0930
https://datatracker.ietf.org/meeting/101/proceedings/
A
A
A
F
G
A
A
A
B
B
So
this
is
the
agenda
for
today
so
well.
The
first
part
of
the
meeting
will
be
focused
on
the
working
group
documents,
so
we
have
to
at
this
moment,
which
are
the
only
two
charter
items
that
we
have
one
on
the
ipv6
over
Ohio
2.11,
MCB,
that
we
we
hope
that
we
will
be
able
to
to
close
the
daemon
issues
during
the
session
to
date.
B
As
a
goal,
then
we
have
a
another
slot
for
the
other
document,
the
use
cases,
survey
and
problem
statement,
and
with
this
document
we
also
go
through
that
will
serve
a
starting
point
for
the
last
part
of
the
meeting,
which
is
about
red
chartering.
So
at
this
point
we
are
just
exploring
what
topics
could
be
in
the
in
a
potential
rich
artery
working
group
and
and
the
ABS
believe
that
this
is
a
good
move
to
do
so.
B
B
B
We
have
a
lot
of
discussion
after
the
first
working
group.
Last
call
under
been
a
lot
of
revisions
is
then
and
lively,
discussion
and
a
mainly
especially
during
the
last
few
days.
We
also
started
the
second
work
to
move
last
call
presently
before
ITF,
because
we
believe
that
the
only
pending
maintained
in
issue
was
on
the
Q
s,
data
header
thing
that
Alex
will
will
discuss
later
on.
B
So
the
aim
of
that
working
group
last
call
was
just
to
check
if
people
were
happy
with
with
the
rest
of
the
document,
knowing
that
this
is
an
issue
that
needs
to
be
resolved,
and
the
goal
is
that
we
close
that
issue
today
and
if
there
isn't
any
other
comment,
big
issue,
we
will
start
the
final
process
or
the
publication,
meaning
that
we
need
to
state
it
in
our
charter.
We
need
to
send
the
document
with
six
months
for
review
and
then,
after
that,
we
can
submit
to
the
AEC.
So
that's
that's
a
plan.
B
Let's
see
we
if
we
close
the
the
issue
during
the
meeting
today.
So
that's
on
the
on
the
first
document
on
the
second
document,
this
document
requests
quite
a
lot
of
discussion.
We
haven't
seen
almost
no
any
activity
during
on
the
mini
list
since
or
between
meetings.
So
this
is
something
that
we
that
has
to
change.
If
we
want
to
really
publish
the
document
without
any
review
activity,
that
will
be
not
possible.
B
So
this
is
a
clear
message
for
the
working
group
that
if
there
is
no
activity
on
this
document,
we
will
have
to
drop
it
because
I
mean
doesn't
make
sense
that
we
publish
a
document
that
nobody
has
removed
or
contributed.
So
please
do
look
at
the
document
contribute
provide
reviews.
We
will
talk
about
that
later,
one
when
the
document
is
present
and
it's
not
important
this
document,
because
since
it's
about
the
plan,
a
statement,
this
should
be
the
basis
for
potential
return.
B
B
B
A
A
B
Okay
thanks:
the
last
kind
of
update
is
on
the
IPS
communication
standards.
There
was
a
meeting
of
that
in
in
March
99th,
so
like
one
week
ago,
a
bit
more
and
we
were
invited
to
present
on
that
meeting,
worries
what
we
are
doing
an
ITF
in
AP
way.
So
we
remotely
presented
that
there
were
different
presentation
from
different
SDOs
working
on
maker
communications
on
on
that.
But
basically
you
want
to
take
a
look
at
the
agenda.
What
was
presented
there
and
what
were
the
materials
that
were
presented?
B
You
have
the
link
letting
they're
on
the
end
of
the
slide
and
then
just
before
going
into
the
actual
presentation.
Yes,
brief,
summary
of
next
steps.
So
again
the
main
starter
item
of
the
group
is
on
the
app-6
over
OCB.
So
the
goal
is
to
clear
pending
issues
by
this
meeting
and
then,
of
course,
with
the
confirmation
by
the
money
list
and
after
that
request,
the
six-month
review
of
the
document
and
proceed
with
the
last
steps
of
publication.
Then
again
a
call
for
action
on
the
order
on
the
second
document.
We
need
activity
on
that.
B
Otherwise
is
not
possible
to
really
move
forward
and
then
we
will
have
some
retarder
in
discussion
and
one
one
message
is
that
during
the
meeting
we
will
not
take
any
decision
of
course
on
that
is
yes
about
discussing
on
and
see
what
is
what
the
working
group
wants
to
do
on
that
topic.
So
if
there
is
any
no
no
comment,
we
can
then
go
to
the
first
presentation.
Alex.
C
C
C
First
I
will
overview
the
changes
that
happened
in
the
draft
since
the
last
meeting
and
second
I
will
illustrate
the
proposed
changes
that
happened
after
the
deadline
of
graph
submission
because
yeah
we
have,
we
can
have
many
versions
of
the
draft,
but
at
some
point
there
is
a
deadline
and
the
deadline
was
on
March
5th
and
that's
where
we
stopped
with
since
then.
Some
changes
were
proposed
that
I
will
list
here.
So.
C
C
C
C
C
H
C
We
removed
also
the
paragraph
about
addressing
modern
subnet
structure
and
the
easiness
of
using
link
local
addresses
and
right
before
the
submission
deadline,
because
we
had
no
agreement
or
there
was
no
concordance
between
the
opinions
expressed
on
the
email
list.
We
wrote
eight
or
2.11
header
instead
of
fatal
2.11,
Q
s,
data
eiders
or
eight
or
2.11
data
headers,
that's
what
the
draft
has
in
it
today,
just
a
toe
2.11
headers
in
order
to
be
a
little
bit
resilient
to
what
may
happen
after
that
means
in
the
last
week
or
two.
H
C
Add
when
the
submission
opens
again,
the
text
is
this:
I
am
reading.
It
start
at
the
beginning.
The
ipv6
packet
transmitted
on
802
dot,
11
OC
B,
must
be
immediately
preceded
by
a
LLC
header
and
an
eight-oh
2.11
header
in
the
LLC
header
and
in
accordance
with
the
epd
heather
type
protocol
discrimination.
C
The
value
of
the
type
field
must
be
set
to
86
dd
hexadecimal
in
the
eighth
or
2.11
header.
The
value
of
the
subtype
subfield
in
the
frame
control
field
must
be
set
to
eight.
That
means
curious
data.
The
value
of
the
traffic
identifier,
subfield
of
the
qsr
control
field,
in
the
eighth
or
2.11
header,
must
be
set
to
binary
0
0
1.
That
means
user
priority
background
and
qsr
access.
Category
a/c
underscore
BK
in
the
ethernet
2
header.
C
The
value
of
the
type
field
must
be
set
to
86
DD
means
ipv6
and
remove
this
text
down
here,
which
means
the
other
alternative
use
of
learning
the
EPD
other
alternative
views
of
layering
are
epd
described
in
Appendix,
II
and
snap
okay.
So,
basically,
because
if
it
is
not
a
return,
ative
view,
it
is
the
view
and
snap
is
because
we
do
not
use
snap.
C
A
I
Hello,
this
is
a
port
from
Callao
in
Korea,
so
actually
I
didn't
Apollo
the
melon
disk
in
detail,
but
I
have
some
concern
about
a
class
of
Q
s,
I'm
sure
killed
actually
background
and
best-effort
IP
has
a
minimum.
Our
window
slot
and
maximize
related
to
same
PG
means
in
terms
of
our
service.
Our
posts
classes
are
equal.
I
My
concern
is
I
feel
way.
The
track
may
be
important,
such
as
deliver
some
collagen
avoidance
or
adaptive
cruise
control
message.
That
message
should
be
are
delivered
quicker
than
other
background
traffic.
So
if
we
stick
to
our
background,
my
concern
is
the
kind
application
use
cases
doesn't
work,
so
I
have
I
needed
to
get
some
feedback.
Why
you
decide
the
background
apart,
I
mean:
could
you
describe
previous
email
discussed?
Why
did
you
decide
the
user
priority
background?
Yes,.
C
C
To
be
short,
this
use
of
background
is
to
compare
with
other
traffic
in
the
network
that
is
not
IP
traffic
and
that
exists
today
in
some
places,
and
that
may
want
to
be
more
priority
than
the
IP
traffic.
But
at
the
same
time,
I
also
agree
with
you
that
in
some
circumstances,
like
cooperative
adaptive,
cruise
control
or
platooning
IP
may
be
used,
and
some
of
the
IP
messages
for
these
platooning
scenarios
must
have
a
better
priority
than
just
background.
C
But
I
do
not
worry
very
much
about
this.
I
do
not
see
it
as
a
problem,
because
I
think
it
is
possible
to
have
background
very
background
value
for
IP
today
in
general
and
later
have
maybe
another
document
that
has
some
special
IP
over
other
priority
other
priority,
like
maybe
a
priority
that
is
not
even
yet
defined
right,
because
right
now,
only
four
priority
are
defined,
but
they
they
don't
have
names
that
are
appropriate
for
a
vehicle
network.
C
I
I
totally,
are
they,
as
you
know,
wait.
Our
protocol
stack
has
two
natural
layer:
r1
is
both
IP
version
6
and
TCP
UDP
application
layer,
the
other
one
is
l1.
Sixteen,
oh
nine
point
two
three,
that
of
layer
case.
They
specify
some
other
wave,
a
short
message,
e
or
basic
safety
message,
something
like
that.
They
think
that
is
good
enough
for
some
applications,
such
as
adaptive,
control
control
or
the
pro
tuning.
I
Something
like
that,
but
I
believe
IP
based
some
adaptive
cruise
control
and
a
pro
tuning
also
will
be
a
popular
because
many,
the
programmer
in
Internet
they're
familiar
with
IP.
So
perhaps
in
the
near
future,
we
discussed
this
issue
or
with
I
Triple
E.
Do
we
need
some
another
category?
Probably
can
ask
that?
Okay,
thank
you.
Probably.
C
B
H
H
I
Frankly,
speaking
internally,
there
is
a
lot
of
some
suggestion
from
especially
Charlie
for
kids.
He
gave
a
lot
of
comments,
ok,
a
part
of
his
command.
We
reflect
either
so
still
be
working
on
it,
and
then
this
version
we
try
to
address
Charlie's
other
comment
and
then,
after
our
meeting
over
during
meeting,
you
suggest
the
comment.
This
document
very
important
that
we
are
made
a
lot
of
effort
to
make
a
survey
and
problem
statement
use
cases
for
the
basis
for
our
next
steps.
So
please
read
this
document
and
they
give
a
give
us
comment.
I
So
those
are
the
author's
next
slide,
so
the
updates
from
the
previous
version.
So
this
is
the
major
changes.
The
first
one
is
the
Left
runs
of
all
the
TSR
see
is
added.
The
second
one
is
that
the
finish
on
our
seu
is
updated,
so
RS
you
can
be
connected
to
internet
or
not
so,
and
also
the
definition
of
a
DMM
is
used
for
the
problem
statement
and
the
survey
and
the
first
one
is
the
communication
of
radio
access
network
for
emergency
network
with
4G
LTE
for
ITV
we
I
we
could
infrastructure
the
currently.
I
I
So
the
sixteen
oh
nine
point
three
describes
a
network
layer
or
wave,
so
they
has
their
own
IP
configuration
mechanism
to
reduce
the
delay
caused
by
enable
discovery
of
original
IP
version.
6
stack
so
the
previously
I
mentioned
the
starless
address,
auto-configuration
I
change
the
wave
or
the
rest
of
the
computation.
Something
like
that.
So
the
first
ok,
let
me
I'll
show
you
the
major
updates
one
by
one.
The
first
one
is
we
added
the
DSRC
reference
for
FCC
channel
assignment
for
wave
DSRC.
I
I
The
third
one
is
we
added
the
DMM
an
acronym,
so
this
will
be
T.
The
mobility
management
in
an
terminology
section
first
one
is
I
mentioned
the
lady
access
network.
Our
land
is
mainly
constructed
by
the
body
LT,
but
I
mentioned
in
near
future.
Dsl
see
the
especially
l
to
the
11
OC
b
will
be
used.
So
lastly,
so
I
clarified
the
comparation
of
IP,
so
you
can
see
long
side.
I
16:09
3
describes
the
network
layer
in
the
wave
of
protocol
stack,
so
they
providing
their
own.
The
mechanism
based
on
wave
service,
advertisement
management
frame,
painting,
okay
to
a
address
IP
address,
DNS
server,
address
I'm
gonna
predict
something
like
that,
but
this
one
doesn't
use
original
IP
version,
6
neighbor
discovery,
so
we
needed
to
name
another
Lane
such
as
a
wave
address
or
the
computation
something
like
that,
instead
of
a
slack.
I
So
next
step
is
synchronized
with
our
OCB
document,
so
we
improve
the
cat
analysis
and
the
problem
statement
avoid
our
OCB
document
and
also
the
we
all
source
will
in
us,
the
user
KC's
action,
including
more
industry
activities
and
also
according
to
our
milestone.
We
try
to
finish
this
document
and
then
I
will
look
at
the
working
garage.
The
call
actually
plotting
your
comment,
this
meeting
and
after
meeting
an
earliest.
Thank
you
yeah
any
questions
or
comments.
I
I
C
Right
because,
during
the
discussion
on
the
working
group
list,
one
or
two
people
made
mention
about
deployments
where
IP
is
used
on
NATO
2.11
on
CB
and
I.
Think
it's
good
to
refer
to
these
okay
right
trials
like
there
is
a
trial
dinner
with
a
police
department
in
New,
York
City,
and
then
there
is
some
other
tribe
with
some
other
company
somewhere
in
America
forgotten.
C
I
C
K
Justin
Dean
from
the
chair
of
the
main,
a
working
group,
mm-hmm
I've,
looked
over
the
draft
and
my
comment
is
more
that
it.
It
reads
more
like
a
collection
of
solutions
and
less
like
a
problem
statement.
I
had
a
hard
time
figuring
out
what
the
problem
was
and
in
that
was
what
I
thought
the
document
was
going
to
be
expressing,
but
it
wasn't.
It
was
giving
solutions
and
and
options
to
solve
different
problems,
but
it
didn't
to
me
anyway
clearly
state
what
the
problems
were
to
solve.
K
So
my
input
would
be
to
clean
it
up
in
terms
of
what
you're
trying
to
convey
and
then
work
on,
selecting
solutions.
I
know,
I've
seen
many
working
groups
come
and
go
that
that
I've
tried
to
solve
everything
and
I.
Think
if
you
nail
down
the
specific
pieces
you're
trying
to
solve
and
just
go
with
something
in
or
gives
give
options,
you
might
get
too
deep
into
every
trying
to
solve
everything.
So
I
think
having
a
very
concise
problem
statement
document
would
be
good
for
the
working
group.
Don't
move.
H
I
Let
me
the
clarify
your
question,
the
comment
so
originally
this
document
it
designed
for
solve
a
two
sub,
a
IP
based
I
picture
that
talking
so
that
July
we
has
a
lot
of
some
domains,
including
yeah
many
stops,
and
then
we
try
to
analyze
a
problem
but
I
think
you're
yet
commented
correct.
We
need
to
focus
on
something
right,
so
we
try
to
some
address
what
kind
of
problem
we
have
to
solve
in
our
looking
or
so
definitely
we
are.
We
do.
K
J
Christian
so
like,
are
you
willing
to
help
like
fix
this,
or
do
you
have
like
specific
actions?
They
can
take
because,
like
one
of
the
things
that
we
are
trying
to
do
in
the
idea
here
is
not
to
sequence
the
problem
shaken
on
the
solution?
Okay,
so
that's.
Why,
like
we
have
this
like
grab
bag
of
things
of
like
all
the
support
documents
together?
Okay,
so
if
we
think
like
a
focus
problem
statement,
it's
useful
like
give
that
feedback,
and
we
can
certainly
like
accommodate
that.
K
Coming
coming
from
just
the
experience
of
working
groups,
I
think
that
has
to
be
driven
by
what
people
want
to
to
fix.
I
mean
what,
if
what
are
they?
What
do
you
try
to
do
and
the
problem
statement
should
be
driven
by
what
you're
trying
to
do?
Are
you
trying
to
do?
Do
you
need
multicast?
You
need
fast.
Switching
to
get
to
the
Internet.
Are
you
trying
to
do
vehicle
vehicle?
Are
you
trying
to
do
dad?
Are
you
trying
to
do
a
mobile
IP?
Are
you
trying
to
do
an
email
kind
of
stuff?
K
But
those
are
there.
Yes,
those
are
there
kind
of
mixed
up
together,
but
there
are
different
ways
of
solving
those
problems,
but
any
one
of
them
individually
are
problems,
and
you
can
have
a
suite
of
problems.
That
is
what
a
ban
a
would
look
like,
or
what
a
vehicle
network
would
look
like
and
what
the
problems
are
in
that
space,
and
then
there
are
multiple
different
solutions
and
I.
Think
going
into
the
solutions
in
this
document
gets
a
little
bit
too
muddled
in
in
terms
of
what
you're
trying
to
convey
with
it
right.
J
J
K
K
K
If
you're
gonna
go
with
the
with
the
document,
name
of
these
are
the
problems
we're
trying
to
solve.
I
think
it
has
too
much
in
terms
of
possible
solutions
in
a
survey
of
solutions
in
that
document
and
and
breaking
that
out
into
solutions
of
you
know,
problem
section
four
and
it
you
should
have
a
different
documents:
hey
this
document
work.
These
are
a
survey
of
what
we're
trying
to
solve
in
the
problem
statement,
document
search
but
right
now,
they're
kind
of
merged
together
right.
J
And
so
that's
kind
of
like
I
would
say
that
got
driven
by
me
right
because,
like
one
of
the
things
we
are
trying
to
do
in
the
IETF
as
a
whole,
it's
not
to
this
working
group
is
to
minimize
the
number
of
documents
like
you
know,
the
use
case,
documents
and
everything
so
like
so
not
to
focus
the
working
groups
on
that.
Okay.
J
So
that's
why
these
two
documents,
but
come
on
they're
like
separate
documents
at
some
point,
it's
so
like
we
got
like
you
know,
feedback
is
G
that,
like
you
know,
if
you
need
to
reduce
the
number
of
support
documents,
okay,
so
when
I
originally
charted
this,
we
had
like
three
documents
like
in
there.
So
if
you
can
throw
that
slide,
Carlos
right
there
were
like
separate
things
in
there.
So
if
the
working
group
wants
to
go
that
way
like
I'm.
J
Okay
with
that,
yes
I
just
want
to
see
more
support
in
the
working
group,
because
there's
like
quite
a
bit
of
effort
invested
in
putting
these
documents
together.
Okay,
like
I,
think
Charlie
did
like
work
with
you
on
that
much
peace
and
and
if
that's
the
right
answer,
that's
okay.
I
haven't
seen
like
the
same
comment
that
it
made
from
other
people.
So
if
that's
what
the
one
who
wants
to
do?
That's,
okay
with
me,
I'm
thinking.
K
A
van
a
would
look
like
and
I
think
there's
value
in
that
so
that
you
know
one
person
that
deploys
it
one
way
and
another
deploys
it
the
other
way
and
the
protocols
aren't
the
same.
It's
not
gonna
work
together,
so
I
think
having
a
document
that
is
explicit
in
saying
you
know
these
are
the
problems,
and
these
are
our
solutions,
and
this
is
kind
of
the
suite
of
protocols
that
you're
gonna
run
to
solve.
H
K
K
That
seems
reasonable
to
two
points.
To
that
my
main
hang-up
was
the
document.
Name
didn't
seem
to
follow
what
the
document
was
doing
and
the
second
one
was,
if
you're
going
to
do
that
as
chairs
I
would
be
weary
of
trying
to
do
everything.
So
don't
get
bogged
down
in
trying
to
solve
everything
and
getting
it
perfect
and
getting
it
right.
I
mean
if
you're
trying
to
get
something
it's
a
solution
and
there
are
solutions
out
there.
It's
there's
value
and
just
putting
you
know
down.
Okay,.
B
So
then,
I
I
think
it
would
be
good
if
we
tried
to
get
some
volunteers
in
the
room
that
commit
to
to
provide
I
mean
to
make
a
review
and
try
to
provide
suggestions
to
go
into
this
line,
because
I
otherwise
I'm
afraid
that
we
may
end
up
with
the
same
situation
in
in
the
next
IDF.
So
I
don't
know:
if
is
there
any
volunteer
in
the
room
that
come
is
to
take?
The
document
on
in,
like
two
three
weeks
timeframe
provide
good
review
to
the
main
list.
C
B
J
So
yeah
suresh
krisshnan,
so
we
had
like
a
larger
charter
like
for
this
working
group.
Thanos
approved
okay,
so
we
put
together
something
and
there's
like
quite
a
bit
of
discussion
on
the
iesg
and
like
we
kind
of
narrowed
it
down
to
like
these
items,
okay,
so
that
was
like
kind
of
related
to
how
much
would
fit
in
the
interior.
So
like
one
of
the
things
that
was
like
in
there
was
like
multi-hop
so,
like
you
know,
a
lower
like
and
I
discussed
it
and
and
realized
like.
J
That's
concerning
to
me
and
enter
the
chairs
as
well
that
this
document
has
not
seen
that
much
activity
as
the
document
and
I
would
kind
of
put
it
as
a
precondition
for
the
for
each
other
to
have
like
a
good
discussion
of
the
problem
and
I
agree
man
on
the
problem.
Okay
and
like
so
today's
session
is
like
you
know,
people
present
things
to
do,
but
I
want
to
see
some
kind
of
working
group
backing
to
solve
these
problems
before
I
put
up
a
charter
to
the
issue.
F
Three
in
condemning
sirisha,
from
my
point
of
view,
at
least
more
from
my
experience,
of
never
seeing
a
tight
correlation
between
the
following
statement
and
the
world
that
followed
so
far.
I'm
not
saying
this
document
should
not
be
done
right,
but
you
know
just
saying
that
everything
whatever
we
do
has
to
be
on
the
basis
of
this.
J
So
so,
personally,
right
like
if
the
problem
statement
doesn't
exist,
that
doesn't
bother
me
okay,
so
this
is
like
what
I've
been
saying
from
the
beginning.
The
problem
statement
is
like
something
for
the
voting.
It's
a
working
group
document
for
the
working
group
really
right
like
whether
or
not
it's
published
as
an
RFC
is
like
a
completely
different
story,
but
the
thing
is:
like
people
have
not
this
the
problems
that
need
to
be
solved.
Okay,
your
presentation
is
like
the
first
time
like
you're
talking
about
this.
J
H
J
A
lot
of
column
statements
we
thrown
away,
like
you
know
like
a
Laura
and
I,
have
abstained.
A
lot
of
documents
like
I'd
like
to,
because
that's
not
the
important
thing.
The
important
thing
is
the
working
group
agrees
on
a
set
of
problems
to
solve,
and
I
haven't
seen
that
discussion.
Okay,
maybe,
like
you
know
today's
presentations,
like
you,
know,
ignite
the
discussion.
Maybe
the
LS,
like
you
know,
Sharon
was
like
talking
about,
like
you
know,
doing
this
cellular
thing
we
don't
know
right,
but
that
needs
to
happen
in
the
working
group.
J
I
haven't
seen
that
happen
in
the
working
group,
like
the
main
discussion
has
been
about
ipv6
or
ogv,
like
that's
the
right
thing
to
do,
because
that's
just
like
the
solution
document
that
needs
to
go
out
first,
but
I
want
to
see
the
discussion
in
the
working
group
before
we've
each
other.
Just
it
doesn't
matter
if,
like
two
people
want
to
solve
a
problem,
right
like
us,
doesn't
make
a
working
group.
So
we
need
to
make
sure
that
there's
supporting
work,
new
purse
or
something
yes
and.
E
H
J
So
at
that
point
they
lost
the
value
okay.
So
that's
really
the
background
for
all
this
stuff
right.
My
concern
is
not
that
the
problem
Satan
doesn't
have
reviews.
My
my
concern
is
like
nobody's,
discuss
the
problems
that
need
to
get
solved
in
the
next
phase
of
this
thing.
Yeah.
So
that's
really
the
background.
F
So
a
quick
update
on
the
potential
work
items
that
this
working
group
can
take
up.
So
if
you
look
at
just
let's
look
at
the
context
right
right
now:
oh
I
connected
vehicle
and
this
vehicle
and
networks
is
an
important
topic.
If
you
look
at
it
from
the
DSRC
point
of
view,
all
the
regulatory
clearances
are
out
there
essentially
the
five
point:
nine
gig
essentially
allocations,
and
all
of
that
is
in
place
right
and
I.
F
Triple
E
has
already
made
progress
with
respect
to
six
to
nine
standardization
and
the
wave
standards
are
out
there
right
and
at
least
the
Mac
layer
and
other
stuff
right.
3Gpp
has
a
steady
item
on
the
axillary,
2x
and
they're
explored
various
use
cases,
and
some
work
has
already
happened
in
there
right
now.
With
respect
to
safety,
messaging
saja
say
he
has
specified
the
safety
message
set
right.
There
are
two
documents:
one
is
on
that
j2
735
on
the
weightless
safety.
F
All
the
messaging
elements
are
there
that
the
extensive
amount
of
work
went
in
there
as
well
right
and
also
the
onboard
unit
requirements.
Now,
with
the
storming
back
to
IDF
this
working
group,
the
initial
focus
was
on
specifying
ipv6
support
on
the
intrud
level
of
CB
media.
But
if
you
look
at
that
work,
you
know,
fundamentally,
all
it
said
was
how
to
transmit
a
packet
on
the
on
that
medium
right.
That's
about
it
all
right,
there's
absolutely
no
discretion
on
how
neighbor
discovery
will
solve
any
of
the
other
things.
No
discussion
on
the
link
model.
F
F
It's
all
you
know
fine
in
a
lab
environment,
if
you
put
two
RS
units
and
a
OB
unit,
sure
ipv6
Volvo,
but
in
reality,
if
I
put
you
know
the
OBU
in
a
vehicle
and
if
I'm
driving
at
60
miles
per
hour,
will
I
be
able
to
have
like
you
know,
essentially
you
know
access
some
cloud
application
or
not.
That
part
is
not
clear
right.
Essentially,
nobody
is,
you
know,
joy,
yeah,
okay,
Tony,.
B
L
F
I
understand
IP
works
right
essentially,
but
is
it
in
the
form
will
I
be
able
to
make?
You
know,
keep
my
you
know
will
I
be
able
to
use
it
in
a
reasonable
fashion
right
there
I
establish
a
session
right
to
some
application.
What
will
happen
will
I
where
I
keep
read,
you
know,
how
does
the
system
work
from
end
to
end
point
of
view
right?
How
does
neighbor
discovery
both?
How
do
I
identify
other
nodes?
How
do
I
disk
you
know,
there's
no
discussion
on
whose.
C
So
yes,
there
is
some
there
is
this
I
agree
with
Tony.
There
is
some
magic,
ok
just
works,
but
also
you
also
said
in
a
lab
environment
versus
on
the
road.
Also
in
the
road
just
works,
but
there
may
be
some
issues
with
his
hand.
Over
okay
I
will
ask
Tony
in
private.
How
is
his
hand
overdoing
and
this
hand
over
means
some
neighbor
discovery?
Maybe
this
is
what
I
think
may
be
some
more
so.
F
Every
thousand
meters,
you
know,
there's
a
RSU
unit.
You
know
argument
is
that
the
vehicle
is
moving
at
60
miles
per
hour
and
we
will
have
you
no,
no,
no
issue
to
the
communication
everything
will
or
grid
there
is
some
magic
I,
don't
understand,
but
but
I
leave
it
at
that.
I
think
you
know,
that's
that's
what
you
know.
Yeah.
C
B
F
Think
the
question
is
when
it
say
when
we
say
that
it
works
right.
What
does
that
mean
that
works
right,
essentially
that'sthat's?
The
point
we
need
to
clarify
I,
think
yeah.
Exactly
the
point
is
some
optimizations
are
needed.
What
are
those
optimizations
I
do
not
know,
but
I'm
reasonably
sure
that
some
optimizations
are
needed.
M
Harry
speaking
I
think
it's
really
about
semantics.
Working
fine,
it
works
sure.
Does
it
work
efficiently?
Is
another
issue:
I
have
a
high
mobility
and
that's
something
we
have
to
configure
quickly
and
then
we
have
to
maximize
the
communication
during
the
connection
time
right.
These
are
things
even
the
neighbors
around
one
vehicon
ipv6
labour
lists.
The
neighbor
table
might
change
very
rapidly.
We
have
to
adapt.
We
have
to
check
first
of
all
and
say
a
second
adapt
to
this
kind
of
environment,
so
we're
working
sure
working
efficiently.
This
is
something
we
have
to
discuss.
M
Friends,
even
though
that
we
spent
a
lot
of
time
on
the
first
document,
the
IPO
cb4,
my
prosperity
or
my
perspective.
This
was
just
kind
of
the
opening,
the
door
that
we
have
to
do,
the
the
real
job,
which
is
improving
the
efficiency
on
ipv6
over
a
highly
dynamic
and
mobile
environment,
which
may
be
not
specifically
in
tocb
anymore,
but
I
just
stopped
absolutely.
E
E
In
different
ways
where
they
do
some
poor
mobility
handling
on
the
networking
side,
similar
to
what
you
do
in
cellular
networks,
we
don't
actually
see
that
when
you're,
when
you're
moving
when
you're
doing
handoff,
right
and
I
have
no
idea
what
they're
planning
on
doing
so.
That
might
be
a
question.
I,
don't
know
if
anybody
in
the
room
knows
the
answer
to
that
question
or
whether
this
is
still
open
from
the
other.
The
other
STL
is
doing
work
in
the
space
or
whether
it's
up
to
deployment.
But
that
seems
like
a
fairly
basic
question.
F
Right,
yeah,
that's
a
key
point:
Eric
I
think
you
know
what
you
said
in
cellular
mobility.
It
works
because
there's
Mme
there's
a
huge
infrastructure
out
there,
which
can
do
with
l2
mobility.
Here
we
didn't
talk
about
that.
I,
don't
see
a
single
speck
which
stands
about
how
the
RSC
is
connected
to
the
back
or
network.
There's
no
definition.
There
is
no
definition
on
the
link
model.
There's
no
statement
on
you
know.
You
know
who
is
hosting
the
prefix
in
the
absence
of
that.
I
cannot
agree
that
this
spec
works.
This
approach
works
well.
F
J
So
just
responding
to
one
thing:
stre
right,
like
you
said
it
doesn't
work
right,
but
one
of
the
things.
So
this
is
like
kind
of
like
a
thing
that
IETF
says
like
we
don't
do
architectures.
We
do
protocols
right.
So
that's
if
you
look
at
it
from
that
point
of
view,
the
question
is
like
it's:
not
that
like
whether
it
works
or
not.
J
The
truth
is
somewhere
between
what
Tony's
saying
and
what
you're
saying:
okay
but
I'm
closer
to
Tony
on
this
than
you
that
things
don't
work
right,
but
if
you
have
some
kind
of
measurements
like
I,
don't
know
if
you
like
specific
pain
points,
we
need
a
fix,
because
if
something
is
broken
on
the
IP
side,
we
need
to
fix
it.
Okay,
so
I
remember,
like
you
know,
we
work
together
on
LTE
for
the
you
know
the
ending
spec
like
on
on
doing
the
fixes.
J
But
if
there's
like
specific
things
that
need
to
get
fixed,
we
can
fix
it.
But
saying
like
you
need
to
do
the
whole
thing:
it's
not
gonna
work
so
like,
if
you
have
say
like
okay,
maybe
my
neighbor
cache
entry,
like
you
know
it
times
out
even
before,
like
you
know,
I
can
establish
the
next
length.
Then
we
have
an
issue.
Okay.
So
but
those
are
the
kind
of
things
we
can
look
at
here:
okay,.
F
Absolutely
the
end
of
the
day,
some
optimizations
for
this
see
I
think
we
do
focus
on
protocols,
but
we
define
protocols
for
operating
environments.
Right
I
think
you
know
we
are
do.
We
cannot
lose
track
of
that
all
right,
so
yeah
with
that
I
think
yeah.
We
all
agree
that
you
know
the
optimizations
needed
yeah,
let's
go
with
that
and
moving
on
I
think
you
know
just
you
know
nothing
new
here.
Fundamentally,
you
know
from
you
know
from
this
deals
point
of
view.
I
think
you
know
so
far.
F
The
focus
has
been
on
the
air
interface
right,
I.
Think
now
you
know
we
are
going
on
the
track
that
you
know.
Essentially
you
know
you
want
ipv6
to
work.
That
means
that
you
know
from
the
use-cases
point
of
view.
You
know,
there's
you
know,
there's
an
application
in
the
vehicle
which
wants
to
access
some
other.
You
know
application
in
the
cloud
right
essentially
or
you
know,
safety
message
is
going
from
the
RSU
unit
to
the
car
or
you
know
vehicle
to
vehicle
right.
F
That
is
one
hot
communication
or
to
the
cloud
right
in
all
of
these
cases.
I
think
it's
in
our
interest
that
you
know
we
make.
You
know
ipv6
work
completely.
You
know,
but
you
know
in
an
optimized
manner
where
we
can
truly.
You
know
you
know
in
a
in
a
reasonably.
You
know
useful
manner
now
I
think
going
back.
I
think
you
know
I
think
the
question
is,
you
know,
let's
zoom
into
some
of
the
Indy
Indy
cushions
right
anything.
You
know.
Maybe
you
know
this
straight
phone.
F
You
know
you
think
series
you
said
that
maybe
some
is
do
is
already
thinking
about
it.
I
think,
from
my
point
of
view,
ITIL
please
focus
of
I
Triple
E
is
not
the
system
architecture.
They
are
mostly
on
the
air
interface
on
the
Mac
layer,
standardization
right
and
they
say
they're
looking
at
more
at
the
messaging
set
right.
So
nobody
is
looking
from
system
point.
If
you
like,
you
know
how
do
how
does
you
know
one
city
or
you
know,
entity,
you
know,
deploy
these
RS
units
right?
F
What
kind
of
infrastructure
they
need
right
and
what
kind
of
you
know?
How
can
we
make
nd
work
in
all
of
this
I
think
you
know
some
of
the
discussion
points.
The
question
is,
you
know
from
the
link
at
least
minimally.
We
should
have
a
link
model
definition
right.
What
exactly
defines
a
link
in
in
this
operating
environment?
Right
in
classic
sense,
like
you
know,
are
in
you
know,
a
bunch
of
nodes.
You
know
attached
an
access
point
yeah.
You
know
you
see
a
VLAN
or
layer,
two
link
right
now
in
a
regular
environment.
F
You
know
zigzag
vehicles
are
moving
at
16
miles
per
hour
and
for
a
short
period
of
time.
You'll
see,
like
you
know
few
in
a
few
set
of
you,
know
few
vehicles
in
path
and
they
may
be
in
proximity
for
one
of
the
RSU.
Those
three,
if
you
look
at
here
in
this
slide
right
these
three
nodes
for
at
time,
T
equal
to
one
or
form
of
potentially
one
link,
because
they
may
in
a
in
a
range
where
they
can
communicate
right
now.
The
question
is,
you
know
who
you
know?
F
How
do
you
know
what
is
the
prefix
that
is
being
hosted
right
and
you
know
how
do
these
vehicles
communicate?
How
do
they
vehicles
discover
each
other?
You
know
how
do
I
nd
works
now
when
T
called
1/2
T
equal
to
suddenly
the
scenario
completely
changed
you're
at
a
different
RSU
in
it
now,
if
they're
a
different
set
of
neighbors
right
now?
What
happened?
If
you
in
shaded
nd
here
and
suddenly
you're
completely,
is
your
entire
in
a
set
of
vehicles?
Are
your
notes
are
completely
changed?
How
you
know?
How
does
you
know?
F
How
does
this
work
right,
I?
Think
some,
you
know
explanation
or
some
you
know
text
will
is
what
you
know.
We
are
looking
for
right
now,
going
back
to
I.
Think
the
question
is,
you
know
the
from
system
wander.
We
should
specify,
like
you
know
the
the
link
model
and
you
know
how
the
sar
hosted,
how
what
kind
of
you
know
we
should
explore
if
you
know,
if
there's
some
optimizations
read
and
needed
with
respect
to
dad
or
with
respect
to
general
ND
or
you
know
other
things
and
also
some
discretion
on.
F
You
know
the
prefixes
that
are
hosted
in
the
vehicle.
What
is
the
topological
a
relation
to
the
to
the
rsu's
prefix?
Is
it
you
know
easy
to
overlay?
Do
we
assume
that
a
mobile,
IP
kind
of
model
way
there's
an
anchor
and
the
prefixes
that
are
hosted
inside
the
vehicle
are
coming
from
the
overlay
anchor
or
are
they
delegated
prefixes
from
the
RSU
I
think
this
discussions
have
to
happen?
I
think
you
know
some
explanation
is
needed.
You
know
from
my
understanding
the
current
idea
of
work.
What
we
have
done
over
the
years.
F
L
L
C
Well,
excuse
me:
I,
don't
know
whether
that
was
some
kind
of
provocation
or
something
no,
it's
not
going
to
be
not
in
the
car,
okay
at
least
not
for
ipv6,
and
there
are
several
nono
ipv6,
not
in
the
car.
That's
for
sure,
and
it's
there
are
several
ways
to
achieve
it
and
they
could
be
discussed
and
I.
Think
the
question
is
valid.
C
What
is
the?
What
is
the
question
that
you
made
the
relationship
of
the
prefixes,
the
prefix
in
the
car
and
the
prefix
on
the
roadside
unit?
Are
they
somehow
overlaid
or
are
they
routed
together
or
is
mobile,
IP
or
proxy
mobile
ipv6,
and
that
is
these
are
things
that
could
be
should
be
discussed
and
clarified,
but
certainly.
E
A
J
I,
don't
so
that's
the
deployment
question
and,
like
limits
be
seen,
it
is
the
place
to
do
it.
Okay,
so
you
can
go
to
the
amount
right
and
right
like
how
to
run
email
v6
on
top
of
it.
We're
talking
no
here
because,
like
yeah,
make
IP
ronsis
it's
same
way
like
you
know,
we
didn't
say
like
how
to
do
mobility
on
wireless
LAN.
Okay,
we
don't
need
to
do
it
here.
Okay,
basic
thinking,
I
could.
J
And
we
can
make
things
work
right
somehow
so
either
like
by
sharing
the
same
IP
prefix
over
multiple
access
points
or
not.
So
that's
a
different
question.
I'm
not
saying
these
questions
are
not
valid,
I'm
saying
like
they
don't
have
to
be
done
here
right
because
at
some
point
this
is
like
how
some
people
deploy
things,
and
maybe
v6
ops
could
be
a
place
for
doing
here,
like
maybe
like
EMM
could
be
like
place
for
doing
it.
I'm
not
saying
it
shouldn't
be
done
here.
J
I'm
saying
like
this
is
not
the
only
place
to
do
it
because,
like
like
this,
is
like
one
layer
on
top
of
like
now,
what
we're
trying
to
yeah
we're
trying
to
do
ipv6
or
food
okay.
So
if
you
go
to
like
Bluetooth
match
like
people
are
not
like,
if
you
go
to
six
slow
people
are
not
saying
how
to
do
mobility
over
this
okay,
like
people
in
Manor-
okay,
like
for
doing
this,
but
like
so,
we
have
like
a
separation
of
responsibilities
here.
J
B
B
J
What
I'm
saying
is
like,
don't,
we
suppose,
a
link
model,
because
you
have
a
link
model
that
you
already
please
for
saying,
like
all
these
things
need
to
be
in
the
same
prefix
like
what?
If
there's
like
a
link
model
which
says
like
okay,
everything
is
point-to-point.
We
just
don't
know
like
don't
presuppose
anything
as
long
as
you
keep
it
like
Jen
Drake,
and
talk
about
what
you're
trying
to
solve
this
okay
right,
but
don't
be
suppose
a
link
model
and
try
to
come
up
with
the
solution.
F
Two
points
or
iteration
I'm,
not
arguing
that
the
work
should
be
done
here.
That's
a
problem
statement,
I'm
saying
you
know
some
somewhere.
We
need
to
do
that.
I.
Think!
That's
all
right!
I!
Think!
The
second
point
is,
you
know
we
need
to
understand
that
you
know
realize
that
a
tutorial
and
ocv
is
for
a
specific
operating
environment.
I
think
when
we
say
ipv6
works.
We
should
reasonably
make
sure
that
it
works
in
that
environment.
I,
think
that
is
the
thing
now
we
can
say
that
you
know
and
Wi-Fi
also
you
can.
F
J
J
J
F
J
E
Yeah
come
on
so
on
the
link
model
side
of
things.
I
think
that
we
did
discuss
this
a
lot
back
when
we
started
or
as
we
progressed
with
the
6lowpan
working
group,
so
I
think
I,
don't
know
how
clear
and
concise
that
description
is,
but
the
notion
that,
if
something
might
have
been
a
neighbor
a
millisecond
ago,
it
doesn't
mean
that
it's
a
neighbor.
E
Now
the
neighbors
are
the
ones
that
you
can
hear
or
can
hear
you
at
any
given
instant
in
time
over
one
radio,
and
that
is
a
very
dynamic
environment
and
it
means
that
if
you're
gonna
talk
vehicle-to-vehicle,
it's
the
same
thing
as
trying
to
set
up.
You
know
point-to-point
connectivity
with
radio
in
general.
It's
basically
just
a
reflection
about
what
the
radio
reality
is
out
there
with
it
in
terminals
etc,
and
that
one
makes
the
system
simpler
to
think
about.
There's,
not
something
that
tries
to
hide
this
from
you.
E
H
E
Case
and
if
there
are
zeros
connectivity,
okay,
we'll
like
to
get
there,
otherwise
it
won't
right
and-
and
that
way
you
can
start
building
things
up
if,
as
part
of
the
problem
and
here's
where
I
sort
of
splitting
this
stuff
apart
and
talking
about
different
things,
a
not
in
the
same
sentence,
thinking
about
okay
directly
equal
to
vehicle
communication,
what
does
that
mean?
There's
no
RSU
right,
but
I
would
not
start
that
conversation
bottoms
up
I
would
not
start
thinking
about
IEP
I.
We
started
security.
E
N
E
And
that
might
very
well
be
end
up
being
linked.
Locals
might
be
fine
for
doing
that
stuff
and
you
have
no
acquisition
delay
in
terms
of
requiring
them.
You
need
to
make
sure
you
can
actually
talk,
but
that
might
go
away
any
millisecond,
which
might
be
fine,
but
that's
a
very
different
thing
saying
I
want
my
SSH
session
back
to
my
my
server
to
stay
alive
as
I'm
driving
around
at
60
miles
an
hour
like.
Well,
maybe
that's
fine,
an
important
problem
right.
Okay,.
F
E
F
E
If
you
try
to
talk
about
all
of
these
things
in
the
same,
sometimes
things
get
very
confusing,
because
it's
not
clear
to
me
that,
from
a
security
perspective,
things
are
actually
the
same.
Okay,
if
you
want
to
say,
I
want
to
keep
long
live
sessions
to
some
faraway
place
or
more
than
to
harp
plays
away.
Okay,
you
gonna
end
up
with
some
overlay
/
mobility
solution,
no
matter
what
and
it's
a
question
bar
okay,
we
have
about
a
dozen
overlay
solutions
lying
around
in
the
ITF
and
we
have
half
a
dozen
mobility
solutions.
E
F
F
I
think,
at
least,
if
you
can
point
to
you,
this
is
the
problem
or
or
if
you
want
to
truly,
you
know
make
this
session.
You
know
continuity
or
whatever
right.
These
are
the
list
of
solutions
at
least
that
documentation
might
be
useful,
I
think
at
least
beautifully
I'm,
not
saying
we
should
go,
define
a
new
protocol
or
something
it's
not
that
right,
at
least
their
solutions,
a
B,
C
and
D,
which
ID
have
already
specified,
which
may
be
useful
for
this
operating
environment
at.
F
D
F
D
I
want
to
use
this
slide
as
a
segue
to
share
some
information,
the
the
bullet
on
I
Triple
E
protocols
and
the
dot
11000
CB
work,
so
the
work
to
date
has
been
based
on
802,
11
P,
and
that
is
the
product
and
the
standard.
That's
out
there
today
at
the
I
Triple
E
802
11
meeting
in
March.
So
two
weeks
ago,
in
dot
eleven,
we
approved
formation
of
a
study
group
to
look
at
that.
That
study
was
titled
next
generation,
V,
2,
X,
so
NGB.
D
It's
a
new
study
group
and
dot
11
forming
to
look
at
extensions
to
the
thigh,
primarily
for
enhanced
range
and
better
performance
that
would
be
completely
and
totally
backwards,
compatible
to
the
existing
11
P
Phi.
This
is
very
doable
based
on
11
AC
technology
in
a
5
big
band
that
just
would
use
existing
technologies
in
AC,
make
them
available
in
this
OC
B
mode.
So
people
are
interested
in
that
I
will
send
a
link,
perhaps
the
IP
wave
reflector
to
the
document
that
motivated
that
work.
D
It's
just
getting
started,
but
the
goal
is
to
basically
give
a
generational
upgrade
to
the
Phi.
The
underlying
fine
that
OCB
uses
okay,
the
OC
b
mode,
is
unchanged
the
all
of
the
protocols
that
work
on
top
of
that
unchanged,
it's
the
five
layer
mechanisms
that
would
be
brought
forward
to
the
modern
age.
Okay,
thank
you.
Thanks.
M
Mary
speaking,
any
thanks
for
these
updates,
as
I
was
part
of
the
c2c.
The
caucus
ocean
I
was
pushing
for
that.
I
just
want
to
give
a
disclaimer
that,
indeed
it
hasn't
started,
but
it
has
to
also
be
clear
that
all
the
first,
the
day
ones
application
for
co2
communications
can
be
done
with
the
with
the
currents
OCD
system.
We
don't
need
the
new
generations
to
be
able
to
do
the
deployments
today,
just
to
be
to
be
clear
that
it's
not
something
that
is
correcting
failures.
M
D
M
Sure
here
is
just
a
politically
aspect:
we
were
discussing
a
lot
with
some
car
manufacturers
about
these
announcements,
and
we
had
the
issues
here
is
that
you
shouldn't
need
any
wrong
statements
of
weakness
of
the
OSI.
It
must
be
very
clear
we
can
deploy
with
OCB.
Now
all
the
defined
use
cases
for
v2,
I
and
V
2
V
can
be
done
with
the
current
DSRC
technology
and
the
I-th
e5
in
europe.
So
it's
really.
It
has
to
be
clear
on
that.
So
I
fully
agree
which
enhances
this
feature.
As
now.
F
Thank
you
so
moving
on
I
think
you
know
we
talked
about.
You
know
basically
right
the
nd
and
now
I
think
the
question
is
for
supporting
I
think
we
talked
about
briefly
about
the
the
prefixes
that
are
hosted
in
the
vehicle.
You
know
what
is
the
topological
relation
to
the
the
RS
use
network
or
whether
you
know
you
should
have
a
concept
of
an
anchor
right,
I.
F
Think
again,
you
know
the
goal
is
not
you
know,
application
of
you
know
some
of
our
IV
protocols
or
something
but
fundamentally,
like
you
know
whether
you
use
you
know,
MPT
CPU
doesn't
matter,
but
the
question
is:
do
we
need
to
introduce
a
function
called
a
vehicular
mobility
anchor
for
this
operating
environments?
Right,
I,
think
I,
think
that
is
a
key
discussion
point
at
lecture.
How
I
think
you
know
I
know
we
do
protocol
work
here,
but
we
have
to
have
some
in
architecture
for
this,
for
this
work.
F
From
my
point
of
view,
at
least
you
know
when
I
see
the
other
is
do
is
none
of
the
is
deuce
focusing
on
the
front
system
point
of
view.
You
know
this
focus
on
like
you
know,
either
it's
a
messaging
said
or
on
the
air
interface,
but
from
system
point
of
view.
Not
much
work
is
happening
there.
So
the
question
is:
should
this
working
group
or
some
other
working
group
in
IETF
look
at
this
kind
of
problem
where
you
know
how
to
support?
You
know
whether
you
know
how
the
support
you
know.
F
You
know
how
to
make
sure
that
the
networks
that
are
within
the
vehicle
are
reachable
from
from
the
from
the
network
mean
from
the
internet.
You
know
in
a
in
a
continuous
fashion
without
any
impact
or
sessions
right
I
think
you
know,
how
do
we
manage
the
prefixes
and
whether
we
should
introduce
the
function
of
an
anchor?
You
know
if
I
ever
have.
F
One
argument
is:
if
there's
an
anchor,
we
can
ensure
that
you
know
all
the
traffic
enters
and
exits
can
potentially
from
one
central
point
that
way
from
security
point
of
view,
and
all
of
that
we
can
in
an
operator
or
the
either
to
the
OEM
or
a
service
provider
hosting
that
anchor
can
ensure
you
know,
apply
some
security
capabilities
and
other
things
right.
Essentially,
these
ingress
filtering
or
every
packet
that's
going
into
the
vehicle.
F
F
Other
things
like
essentially
like
you
know
now
we
haven't
talked
about
like
you
know.
If
a
city
are
potentially
you
know
in
rsu's
can
exist
a
very
400
to
1000
meters
right.
If
a
city
deploys
it
potentially
there
could
be
hundreds
or
thousands
of
these
units.
How
can
we
manage
these
RS
units
from
one
central
point
right?
What
does
it
take?
You
know?
Do
we
need
a
new
protocol
interface
right
or
is
it
will
it
be
similar
to
Kappa
right
or
completely?
F
In
you
know,
net
config,
right,
I,
think
some
discussion
on
you
know
how
we
are
going
to
manage
this.
You
know
roadside
units
is
of
interest,
at
least
from
my
point
of
view
now.
I
think
I
think
summary
I
think
you
know
I
know
I
think
we
argued
a
lot.
What
does
it
mean
in
ipv6
works
right,
I
think
what
doesn't
work
but
I
think
we
can
at
least
I
agree
that
there
may
be
optimizations
in
nd.
That
may
be
useful
for
this
operating
environment.
I
think
that
is
one
point
one
point
two.
F
We
should
have
some
view
on
how
you
know
a
system.
You
know
how
R
into
n
system
can
be
deployed.
I
know
we
don't
do
n
to
end
work,
but
if
you
look
at
some
of
them,
you
know
mobility
working
groups.
We
did
specify,
like
you
know,
if
we
provide
a
reasonably
good
solution,
end
to
end
solution.
You
know
somebody
can
take
an
emo
architecture
today
and
deploy
and
complete
mobility
in
a
system
right.
That
is,
you
know
that
is
being
done
today.
F
Similarly,
for
regular
deployments,
do
we
need
to
have
an
to
end
system
like
you
know
how
RSU
is
connect
to
you
know
the
cloud
or
how
an
anchor
talks
to
all
of
these
or
how
prefixes
are
rolled
out
to
the
vehicle?
How
outing
works?
Maybe
you
know
some
one
specification,
maybe
should
be
how
in
a
focus
on
that
architecture
document?
That
is
another
point.
Third
thing
is
you
know,
ident
anything,
you
know
any
talked
about
it
before
we
talk
any
of
the
stuff.
F
First,
let's
establish
some
rules
and,
like
you
know
how
the
vehicle
identifies
to
the
network,
you
know
what
is
the
authorization
model?
What
is
the
authentication
infrastructure
I?
Think
again,
this
from
system
point
of
view.
You
know
how
do
you
authorize
authenticate
these
vehicles
right?
What
is
what
exactly
is
identity?
It
is
the
is
it
the
certificate
or
is
it
a
you
know
some
ID
derivative
of
the
win
or
something
right,
I
think
some
discussion
has
to
happen
there
right
and
finally,
you
know
mobility
support.
F
I
think
you
know
there
are
already
other
working
groups,
but
you
know
point
is,
should
be
leveraged,
you
know.
Should
we
say
that
these
are
all
the
potential
solutions.
One
deployment
can
choose
right.
Some
some
explanation
on.
You
know
what
are
the
problems
and
what
mobility
solutions
can
be
used
and,
finally,
like
I,
think
you
know,
RSA
manageability,
I
think
these
I
think
my
view,
our
potential.
You
know
our
kite
UM's
that
that
IETF
can
can
explore
any
questions.
Any
final
yeah.
M
Yes,
this
is
your
ohm
speaking
again.
I
would
like
to
maybe
start
okay,
you
have
some
items
here.
Some
I
agree,
so
my
real
s,
but
what
I
mean
is
we
should
sell
for
the
up
from
the
top
down
from
going
for
the
top
and
going
down
one
thing
which
I
don't
see
here
and
we
try
to
talk
about
it
and
I
think
we
should
talk
more
about
that.
What
are
the
types
of
application
we
have
on
top?
What
kind
of
traffic
is
being
transmitted?
M
It
was
mentioned
before.
How
long
should
the
session
start
last?
We
have
to
define
this
kind
of
things,
because
then
it
might
have
been
hot
mighty
guidances
on
that.
If
you
don't
do
that,
then
we
don't
know
what
trust
meeting
when
you're,
providing
something
that
which
works
for
everything.
Of
course,
maybe
ipv6
can
work
for
everything,
but
I
think
we
need
give
what
we
have
some
guidances
from
some
services
and
some
applications
for
there
is.
There
would
be
some
presentation
on
IOT
wonderful
and
we
can
discuss
about
IOT.
M
What
kind
of
types
of
thing
we
do
that
you
were
mentioning
this
gate,
the
central
gateway.
That's
fine,
but
this
is
for
what
or
one
type
of
service
is
this
for
security?
You
know
if
you
start
centralizing
something
you
automatically
addresses
for
one
type
of
application.
I
would
suggest
to
start
with
from
the
upper
part
to
stop
from
a
Caucasian
step
of
traffic's
that
we
are
transmitting
and
then
maybe
go
down
to
see
where
it
goes.
Thank
you.
Thank
you.
C
Alex
for
the
application.
Sorry
I
tend
to
look
at
it
as
well.
If
there
is
a
link
that
is
a
bigger
and
bigger
pipe,
then
more
and
more
applications
will
will
appear
and
that
we
have
more
and
more
constrained.
Everybody
will
want
to
run
their
YouTube
or
Skype
that
keeps
the
session
up,
but
you
cannot
run
a
YouTube
and
Skype
phone
bluetooth.
You
need
some
802
total
of
an
AC
or
this
ngv
if
I
understand
correctly.
So,
if
you
have
n
TV,
then
I'm
sure
you
have
YouTube
and
keep.
L
C
C
Each
new
protocol,
you
document
that
to
come
up
with
it's
good,
to
have
this
management
information
base
and
I
think
for
ipv6
overall
CB.
Some
of
these
masks
that
we
put
could
be
configurable
and
maybe
a
management
information
base
would
be
almost
mandatory
and
I,
don't
think
we
have
it
in
the
chart
or
maybe
in
the
future
charter.
Any
MIB
management
information
base
young
model
possible
yeah.
Thank
you
Alex.
C
L
I'm
confused
about
the
discussion
about
applications.
As
soon
as
you
turn
on
ipv6,
you
get
all
internet
applications
you
get
web
surfing,
you
get
email,
you
get
jabber
everything
is
there
SSH
sure
no
problem
streaming
falls
out.
Gaming,
it's
gonna
fall
out,
all
sorts
of
things
are
gonna
happen
and
we
can
and
exactly
anticipate
exactly
what's
going
to
be
necessary.
Hey
I
happen
to
have
played
with
a
number
of
these
and
they
all
seem
to
just
work
this
by
its
IP.
The
link
stays
up.
Ip
is
happy.
C
L
L
I
I
have
a
some
comment
about
equal:
oh
that's
architecture,
so
very
good,
a
yeah,
it's
so
this
channel.
So
but
the
one,
a
missing
part
is
a
beaker
to
beaker
communications.
You
can
see
the
highway
many
yard
research
try
to
accommodate
the
Pico
to
be
communication
to
reduce
delay
caused
by
the
little
a
by
some
RC
or
some
other
anchor
point.
So
that
is
we
need
to
consider
and
also
a
one
comment.
I
What
so
the
bigger
Network
case
that
in
highway,
our
pker
can
communicate
with
directly
sometimes
the
intersection
area
urban
rotor
network
case
the
congest
the
earlier
case,
maybe
RS.
You
can
orchestrate
the
communication,
in
the
case
the
vehicle
communication
and
sometimes
picker,
to
pick
her
up
by
other
RSU,
something
like
that.
So
that
might
be
different
from
other
mobility.
So
we
need
to
think
of
it.
Okay,
right,
yeah,.
F
J
J
J
J
J
Of
proof
is
on
you
because
the
working
group
that
something
is
required
and
so
take
an
example,
use
case
right
and
say:
okay.
This
is
why
it's
not
gonna
work
and
this
what
I'm
gonna
do
because
any
of
the
mobility
mechanisms
we
define
for
IP,
okay,
any
of
the
mechanisms
we
define
on
top
of
IP
I'd
like
to
know
by
MP,
TCP
or
like
multipath,
quick
when
it
comes
out
like
those
things,
are
all
gonna
work
on
this
okay.
J
So
that's
why
I
was
like
you
know,
saying:
MP
TCP
when
Eric
was
talking
right
so
like
they're
still
gonna
work.
So
even
if
you
have
a
constantly
changing
IP
address
right,
like
you
know,
we
can
do
MT
PCP
with,
like
you
know,
you
can
just
like
you
know.
As
long
as
you
have
like
some
locator,
we
can
just
keep
moving
on
or
we
can
have
mobile
IP
running
right.
J
F
But
that's
that's
a
fair
point.
I
think
you
know
I,
think
the
question
is
say:
I.
Think
few
data
points.
Writing
you
know
from
my
own
company
point
of
view
here.
You
know
there's
a
lot
of
work
happening.
You
know
we
work
with
many
audience
and
essentially,
at
this
point
you
know
we
have
been
doing
mobile
appear
mobility
protocols
for
a
long
long
time,
but
now
some
real
work
happening.
We
are
working
with
OEMs
and
I'm
sure.
F
If
you
look
at
all
the
industry
direction
right,
Tesla
or
any
of
that
the
amount
of
self-driving
technology
that's
coming
into
the
vehicle.
You
cannot
deploy
a
self-driving
vehicle
and
without
you
know,
having
a
connectivity
to
some
cloud
right,
there's
a
lot
of
data
that
you
know,
there's
a
lot
of
crossing
that
happens
within
the
vehicle,
but
there's
also
a
lot
of
data
that
comes
from
the
cloud.
F
So
unless
you
know
IETF
provides
some
system,
some
specification
saying
that
this
is
how
you
deploy
it
I
understand
that
the
protocols
exist
than
many
protocols,
but
at
least
you
know
document
them
in
some
form
some.
It
should
say
that
you
know
these
are
all
the
options.
These
are
the
problem.
You
know
these
are
the
issues
that
you're
going
to
hit
and
these
are
the
potential
solution.
This
is
all
you
deployed.
That
guidance
is,
you
know
what
we
need
to
provide
and
that
specifications
are
going
to
be
useful
and
some
there's
a
room
for
optimizations.
F
B
Just
cut
the
discussion
so
can
we
can
we
have
the
discussion
on
this
point
as
part
of
the
discussion?
They
promised
a
mushara
document.
So
it's
it's
basically
what
we
discussed
before
we
have
a
problem.
There
are
a
set
of
solution
that
can
be
applied,
so
we
need
to
document
what
is
there
and
we
need
to
and
if
I,
what
is
missing
if
any,
and
if
any
that
will
be
a
gap
that
then
we
can
gather
the
interest
on
the
world
working
on
that.
B
So
we
could
have
that
kind
of
discussion
as
part
of
this
document.
I
think
that
will
be
helpful
for
the
document
and
for
the
actual
potential
recharter
in
topics.
So
please,
if
you
can
do
Sol
star
thread
on
the
Middle
East
like
these
are
the
gaps
I.
These
are
the
problems.
I
think
should
be
documented.
These
are
the
solution
that
can
be
applied.
These
are
the
whatever
guidelines
best
practices
that
is
missing,
gaps
on
the
protocols
and
trillion
defy
those
I
think
that
will
be
helpful
all
right
to
continue
discussion.
M
Okay,
so
good
morning,
interim
Harry
I'm
going
to
briefly
introduce
one
kind
of
homework
that
have
been
doing
to
identify
potential
issues
about
the
mapping
of
IP
prioritizations
into
the
ipv6
and
ocb.
This
is
also
related,
so
somehow
to
also
to
the
discussion
of
the
ipv6
over
CB
and
ste
issues
about
Keywest
data.
So
of
course,
this
assumes
that
we
use
Q
s,
data
for
transmitting
IP
traffic
over
a
CB,
so
brief
introductions
of
basically
OCD
access
category.
So
if
you
take
quality
of
service
data,
then
you're
going
to
have
this
EDC
ake
classes.
M
M
This
is
a
deterministic
value
of
number
of
slots
per
different
categories,
and
then
you
have
a
contention
windows
which
is
a
which
is
a
stochastic
part.
So
then,
as
you
could
see
in
a
different
in
the
different
queues,
then
you
can
exceed
the
folk
cused.
So
one
is
called
excessive
voice,
then
access
videos,
access,
best
effort
and
access
background.
So
of
course,
just
to
avoid
any
misunderstanding.
This
absolutely
does
not
reflect
that
only
voice
can
be
transmitted
on
access.
M
Katorga
voice
is
more
or
less
more
less
like
a
voice
type
of
traffic
quality,
so
to
say,
and
you
could
see
the
values
of
each
of
the
of
the
parameters,
so
you
could
see
in
the
charts
that
you
have
two
slots
for
the
first
access
category
voice,
three
slots
for
videos
through
six
slots
for
best
effort
and
nine
slots
for
the
fourth
back.
In
the
background,
then,
the
contacting
windows
is
also
a
bit
specific.
M
You
have
four
eight
sixteen
and
sixteen
depending
on
the
owner
Casa
categories,
but
this
is
what
basically,
in
two
thousand
eleven
two
thousand
sixteen
OCB
specifies
for
EDC
s4c.
Now,
of
course,
the
sudden
says
how
do
we
match
the
different
classes
of
the
day
off
of
traffic
of
the
the
ith
traffic
classes?
So
what
I
mean
what
enjoy
this
here,
but
just
what
we
did
in
Etsy
and
there
was
also
similarities
with
SAE
for
the
DSM's
so
for
the
access
Kotori
voice,
so
the
high
one?
M
There
is
the
image
you
see
them,
so
it's
like
even
driven
and
can
be
transmitted
with
a
very
high
high
urgency.
The
second
one
is
dams,
regular
dams,
then
the
third
one
which
is
best
effort,
is
the
cams
and
the
last
one
background
is
for
any
kind
of
other
traffic
if
I'm
at
it
match
it
to
SAE.
At
least.
We
know
that
the
access
Kotori
give
voice
is
for
emergency
or
even
bass,
VSM
and
axial
cattle
give
best.
M
Efforts
is
for
regular
periodic
PSM,
so
this
was
actually
the
sort
of
the
source
of
the
discussions
about
the
curious
data
in
the
data,
because
this
is
the
the
basically
number
of
slots
you
have
to
wait.
So,
if
I
have
a
specific
dam
to
transmit
with
the
access
katryca
voice,
I
will
have
to
wait
at
least
once
if
in
two
slots
and
then
randomly
between
zero
and
three
slots.
M
What
is
also
important
is
that
if
I
don't
use
this
yay,
if
I
use
DCF,
which
is
the
baseline
data
style,
then
it
is
the
equivalent
is
a
diff
which
is
also
two
slots
and
sips.
So
basically,
the
deterministic
waiting
time
for
diff
is
also
is
equivalent
to
the
access
category
voice,
of
course,
the
contention
window
a
little
bit
bigger,
but
basically
the
seat
that
was
the
source
of
potential
interferences.
M
One
thing
it
is
also
important
to
understand
is
that
so
far
most
of
the
traffic
on
OCB
is
broadcast,
so
the
contention
windows
max
is
actually
never
applied,
so
the
contention
window
we
have
is
just
to
contention
windows
min
then,
let's
move
one
step
must
have
up
so
in
total
11
800
dot,
1
D,
which
is
the
mock.
Bridging
documents
in
a
Triple
E,
provides
a
set
of
user
priorities.
So
we
have
difference.
We
have
seven
classes
or
user
priority
and
a
default
one
which
is
basically
if
nothing
is
provided.
So
you.
H
M
Here
a
chart
taking
from
the
standard
showing
that
we
have
a
seven
user
priorities
and
the
issues
it
at
the
end.
We
have
also
acronyms
for
the
four
seven
potential
traffic
class
or
not
traffic
class,
but
it's
seven
potential
priorities
from
network
control
to
background
and
spare
bed
effort.
The
issues
in
mapping
is
that
you
are
going
to
map
something
that
has
seven
options
to
something
that
has
only
four
options,
and
this
actually
and
all
another
thing
that
is
important
is
to
check.
M
What's
the
definition
of
video
and
voice,
for
example,
it's
also
related
to
some
latency
and
jitter,
so
that
was
the
original
definition
for
for
the
1d
later,
also
used
for
for
11
e
and
e
dot,
11
2016.
So
now
the
issue
is
when
you
map
it.
So
if
you
have
seven
classes,
I
mean
seven
user
priority
and
you
have
to
map
it
on
a
specific
number
of
queue.
So
again,
if
you
look
at
the
standard,
you
know
you
know
1d,
they
provide.
They
propose
a
few
options,
so
you
can
see
under
on
the
vertical
part.
M
You
have
the
user
priorities
and
the
upper
part
is
the
number
of
queues.
So
what
we
have
in
in
811
2016
is
we
have
four
cubes,
so
we
have
a
mapping
from
basically
the
first
two
user
priorities,
a
map
to
background,
then
the
second
next
two
best
efforts.
Then
we
have,
of
course
the
others
CL
and
videos,
videos
and
NC
mapped
again
to
something
that
would
be
called
videos
and
voice
and
videos.
So
this
is
a,
but
this
is
again
a
potential
suggestions.
M
There
is
also
this
if
there
is
issues
about
that
that
it's
not
part
of
it's
not
specific
part
of
ITF,
but
this
is
something
that
was
the
recommendations
come
out
of
best
practice,
but
not
specifically
hard-coded
install
now.
If
we
move
one
step
further,
we
have
the
IP
IP
traffic
classification,
so
these
are
the
deserts,
gscp
and
so
the
difference
in
service
code
points
with
a
few
names
default,
forwarding
and
pretty
sure
you're
perfectly
expert
in
that.
M
So
these
are
actually
specifying
the
prioritization
a
pitch
IP
traffic
and
of
course,
what
we
hope
is
that
if
you
have
a
specific
reciprocation
and
the
IP
traffic,
it
would
be
mapped
to
the
equivalent
in
the
Wi-Fi
layer.
So
the
IP
traffic
is
tropical
cific
ations.
You
will
have
in
ipv4
traffic
code
type
of
service
in
a
pv6
traffic
class,
both
of
them
take
the
six
bits
of
the
gscp
and
extending
with
two
two
extra
bits.
M
So
basically,
if
you
take
the
DHCP
six
most
significant
bits,
then
you
have
I
mean
if
you
take
the
six
Mocksville
in
a
mix
of
the
type
of
service
and
traffic
class,
you
get
the
DHCP
and
there
is
a
again
some
examples
again
taken
from
the
IETF
RFC
445
94.
It
was
trading
the
different
service
classes
and
the
mapping
to
different
DHCP
values.
So
you
have
telephony,
you
have
voice
and
so
and
so
forth
and
of
course,
let's
see
what
happens
afterwards.
M
So
the
issues
after
that-
that's
that's
we
have
that
was
it
force
looking
at,
is
one
with
map
and
l3
priority,
which
is
6
bits
to
nail
to
user
priority,
which
is
only
3
bits.
It's
current
best
practice
is
to
take
the
3
most
significant
bits
and
simply
Muppets.
So
if
you
have
a
dscp
that
is
one
oh
one,
one
one
zero,
then
we'll
end
up
into
the
priority
one
over.
M
This
is,
of
course,
best
practice
not
default,
but
if
you,
if
you
do
that
without
really
taking
care
of
it,
then
you
are
going
to
have
confusion.
For
example,
you
have
a
voice
type
of
traffic
expected
forwarding,
1,
1,
1
1
0
and
ends
up
being
you
p5,
which
is
actually
not
voiced,
which
is
actually
video.
So
you
have
a
you:
have
a
voice
map
in
video
and
I
live.
You
also
have
a
look
at
the
other
potential
confusion
which
is
a
bit
annoying.
M
In
that
case,
the
upper
part
is
the
same
at
the
reception
you
get
an
L
3
and
L
2
user
priority
3
bits,
and
yet
you
map
it
to
a
six
bits,
a
three
priority
and
then
what
the
best
the
the
best
practice
is
to
multiply
by
eight.
You
simply
shift
it
three
bits
and
then,
as
an
example,
we
have
the
SCP.
You
have
sorry
if
you
have
the
that's
the
opposite,
that
you
do
but
cut
and
paste.
M
So
if
you
have
a
you
p101,
then
you
have
you're
going
to
have
a
dscp
101
0
0
0.
If
you
do
that
again,
you
are
going
to
have
some
strange
mapping
or
something
that
will
actually
be
voice
or
telephony.
Then
it
will
actually
map
to
a
specific
cs8
which
actually
is
network
control.
So
that's
something
that
we
also
have
to
avoid
so
again.
Few
confusion,
examples
that
you
can
have
a
look
here.
All
of
this
actually
are
from
the
RFC.
It
will
83
25,
which
I
was
recommended
by
Roland
bless.
M
So
thanks
a
lot
for
to
him
to
have
pointed
out
these
documents.
So
that's,
basically
what
does
this
documents?
Its
mapping
deserves
to
a
2.11,
so
what
day
was
this
document
was
is
to
identify
the
potential
confusion
and
suggest
to
adapt
the
mapping
such
that
the
confusion
are
actually
relieved
and
another.
Another
aspect
they
were
suggesting
is
to
also
leave
it.
Let's
say
configurable
such
that
it's
not
sealed
in
stone
what
the
mapping
is.
M
Intermediates
systems
could
configure
it
according
to
a
specific
controller.
So
this
is
interesting
to
do
to
have
a
look,
and
we
might
actually
part
of
the
IP
wave
also
have
to
do
to
look
at
that
document.
So
my
slice
line
here
is
as
input
IP
wave
as
we
now
mention
that
we
have
to
accurate
us
data
part
of
IP
wave
for
IP
over
CB.
We
should
also
refer
most
likely
to
this
RFC
to
the
RFC.
It
is
we
25
for
the
mapping
between
diff
serves
to
8.11.
M
What
we
should
also
do
is
to
try
to
understand
what
kind
of
traffic
are
we
going
to
send
because
so
far
these
are
generic
IP
traffic.
As
far
if
there
is
no
issues
about
that,
that's
fine
if
there
is
specific
traffic
that
is,
that
is
either
doesn't
enter.
The
classical
deep
source
traffic
that
are,
for
example,
illustrated
here
sorry
here,
then
we
might
also
want
to
add
some
specifications
so,
for
example,
to
avoid
any
confusion.
M
If
you
trust
me,
the
specific
shut
up
of
traffic
that
would
be
critical
for
Thomas
driving
and
he
ends
up
being
mapped
on
the
wrong
hues.
Then
that's
something
we
might
want
to
avoid
and
last
points.
We
also
have
to
look
at
one
thing.
That
is
important.
If
we
look
at
the
current
specification,
we
have,
if
you
have
a
user
priority
zero,
which
is
the
fourth
one.
It
is
map
to
best
effort
and
not
background.
M
So
we
have
to
double
check
that
because,
as
we
explicitly
mentioned
to
the
RFC
22,
the
idea
we
discussed
before
that
it
should
be
that
we
should
map
it
to
background.
Then
we
have
a
confusion
with
a
default
mapping
zero.
So
one
potential
option
would
be
to
have
a
to
mention
that
we
should
have
a
default
mapping
to
background,
and
that
would
mean
that
we
simply
will
rather
change
the
definition
of
the
default
mapping
from
best
effort
to
background
such
that
which
we
share.
The
same.
M
The
same
language,
but
nevertheless
this
is
just
up
to
discussion
so
point
here.
Let's
see
I
mean
that
we
still
need
to
dig
in
to
be
more
about
this.
This
aspect
to
see
if
there
is
something
that
is
worth
bringing
into
this
group,
but
at
least
that
goes
also
to
follow
respect
to
my
previous
comments.
We
need
to
defy
the
type
of
traffic
we
transmit
and
based
on
the
traffic
we
transmit.
We
should
see
if
the
priorities
are
correctly
mapped
into
OCB.
Thank
you.
L
So
Tony
Lee
again
I
apologize
for
repeating
comments
that
I
made
on
the
mailing
list,
but
I
have
to
bring
it
up
again
because
I
don't
see
how
this
works
operationally.
We
have
no
security
over
dscp.
Anyone
can
say
anything
in
their
dscp
value
today
and
unless
you
happen
to
traverse
a
network,
it
actually
implements
QoS
and
not
all
do
and
the.
Unless
that
network
actually
remaps
dscp,
then
the
dscp
value
you
get
is
whatever
the
user
decided.
L
L
M
L
Is
a
broader
Internet
do
not
have
a
way
of
preventing
that
today
and,
as
a
result,
the
DSC
B
values
are
not
guaranteed
unless
your
network
has
overwritten
and
because
we
are
an
end-user
network
here,
there
seems
to
be
no
mechanism
for
doing
that.
So
I
don't
see
how
you
can
do
anything
except
to
cause
yourself
to
be
hacked
by
doing
this.
Okay,.
M
Well,
I
mean
we
can
start,
we
can
work
in
three
in
three
steps.
The
first
step
would
say:
okay,
assuming
we
don't
have
the
DHCP,
then
we
still
have
a
default
mapping
and
default
mapping,
as
it
is
mentioned
now,
is
still
access
category
best
effort.
It's
not
even
it's
not
even
that
the
SCP.
That
is
mentioning
it.
It's
actually
11.
It
is
one
D.
You
can
see
that
if
you
have
only
one
queue,
then
it
should
be
best
effort,
which
is
not
what
we
should
do.
M
If
we
basically
have
folk
huge,
then
we
have
the
background
and
the
the
default,
which
is
I,
think
the
one
that
is
the
die.
I'm.
Sorry,
all
right,
I,
don't
I,
don't
show
it
here
is
not
mapped
to
back
to
background.
So
this
is
something
that
if
we
leave
it
as
such,
if
we
take
a
default
user,
priority
would
be
Matt
the
best
effort
and
not
to
background
which
is
opposite
to
what
we
discussed
before
that's
the
first
step.
M
The
second
step
is,
we
can
already
look
at
work,
I
mean
if
we,
if
there
is
a
way
to
consider
the
SCP
first
at
not
being
a
source
of
hacking,
but
just
to
see
if
it
can
work
and
then
the
third
step
we'll
see
okay,
if
it
is,
if
it's
a
source
of
being
hacked,
then
maybe
it's
not
going.
It's
not
useful.
We
need
to
secure
it.
That's
something
beyond
my
my
expertise,
but
there
is
a.
M
C
So
I
agree
with
you
with
that
default.
Zero
value,
I
did
not
know
it
I
think
that
could
be
a
problem
and
should
be
fixed
quickly
now,
with
respect
to
DHCP
not
being
secure.
If
you
look
at
it
like
this,
yes,
anybody
on
the
internet
could
read
the
DHCP
in
traffic
class
and
then
replace
it
with
something
else
like
a
man-in-the-middle
attack
and
send
some
bad
traffic
to
the
car
instead
of
the
legitimate
some,
but
I
think
this
could
be
secured
with
the
IPSec,
because
dscp
is
part
of
ipv6
header.
C
M
But
I
mean
just
as
a
disclaimer
I
do
not
advocate
to
use
the
SCP,
because
the
only
thing
that
we
have
been
advocating
for
couple
of
weeks
from
now
is
that
we
have
QoS
data
at
OCB
because
we
just
have
to
know
quite
the
second
quad
and
the
underlying
discussion
was.
If
you
have
a
edca,
so
curious
data,
it
would
be
nice
to
be
able
to
map
it
to
the
SCP,
and
that
was
the
reason
why
I
have
this.
M
So
it's
not
that
I'm
advocating
it
I'm
just
mentioning
that
there
is
an
RFC
about
it
suggested
by
Ronan
PLAs,
and
it
would
be
nice
to
look
in
within
this
working
group
whether
this
applies
to
to
wave
or
to
a
p-wave
or
whether
it
does
not
apply
to
happy
waves
and
needs
to
be
changed.
That's
all
what
I
was
trying
to
say
here.
L
C
J
Okay
description,
so,
if
I
understand
it
correctly,
okay,
so
because,
like
it's
not
clear
from
the
presentation,
so
do
you
want
to
take
the
traffic
types
and
map
them
into
da
CPS
or
take
the
DSC
piece
and
map
them
into
like
lower
level?
Both
points,
okay,
cuz,
83
25-
does
the
latter.
It
takes
the
DCPS
and
map
them
onto
like
you
know,
whatever
the
you
know,
they
CBO
and
those
kind
of
things,
so
so
I
have
a
hard
time.
Seeing
like
what
you're
proposing
to
be
very
frank
right.
J
So
are
you
proposing
abyss
for
this
like
for
based
on
some?
If
there's
a
change
needed,
or
are
you
mapping
a
parallel
document
this
which
talks
about
what's
on
the
leftmost
column,
so
that
that
is
my
first
question?
Okay,
so
a
second
thing
is
this:
this
just
passed
through
I,
don't
I
would
say
two
months
ago,
right
and
I
have
a
hard
time
seeing
something
has
changed
radically
in
the
last
two
months.
J
So,
quite
again
like
just
like
a
toaster
II
like
we
need
with
a
justification
on
like
why
we
want
to
do
it
and
actually
also
what
you
want
to
do
so
so
explain
if
you're
gonna
change
the
mapping.
Are
you
just
gonna
change
the
labeling
of
the
kalo?
That's
the
first
thing
on.
M
Okay,
so
first
of
all,
it
was
just.
This
was
just
meant
to
basically
discuss
so
I'm,
not
proposing
anything
yet,
but
the
direction
I
would
at
least
try
to
dig
into
bit.
More
is
first
to
be
sure
that
if
I
mean
I
have
a
traffic
class,
whether
you
gscp
or
not,
is
another
question
its
own
I
will
have
a
direct
mapping
will
be
default.
Mapping
and
I
want
to
be
sure
that
this
default
mapping
goes
to
background,
not
past
effort,
so
this
is
something
we
have
to
double-check.
M
Second,
we
have
to
check
any
C's
related
to
my
my
first
quad
my
remarks.
We
have
to
define
what
kind
of
traffic
we
are
going
to
transmit
in
within
IP
wave
and
how
that
could
match
such
that
whatever
is
mapped.
So
if
I
use
it,
if
I
just
with
default
mapping,
they
will
be
automatically
background
if
I
want
to
add
a
bit
more
I
mean
that
was
also
related.
M
J
So
the
DHCP
802
11
mapping
is
done
right.
So
if
you
solve
the
first
problem
like
how
would
I
kill
a
traffic
map,
it'll
be
a
CP.
The
second
problem
is
automatically
solved.
If
you
believe
the
mappings
need
be
prefer
correct
right,
because
you
know
like
what
the
8025
DHCP
is
map
into
on
the
l2.
So
you
the
only
job
we
have
this
like
then
take
the
IP,
the
traffic
and
map
it
onto
the
right
DHCP
all
right.
So
that's
why
I'm
saying
so.
There's
like
two
different
questions
so.
H
H
J
I
So
let
me
introduce
to
some
three
and
the
topics
for
possible
work
items.
So
I
will
briefly
explain
so
so
I
would
write
to
a
suggested
three
one
item.
The
first
one
is
the
pickle
arm.
Neighbor
discovery,
especially
the
prefix
discovery
for
the
loading
and
second
one
is
a
lab.
The
lab
field,
service
discovery
and
second
one-
is
a
DNS
naming
service
in
terms
of
our
name
of
the
compression
over
in
vehicle
devices
and
also
device
and
service
discovery
and
DNS
name.
Graduation.
I
So
this
is
a
devii,
color
or
natural
architecture,
so
you
can
see
our
beakers
can
communicate
with
each
other
directly
or
it
can
communicate
with
the
infrastructure.
Also,
we
have
a
traffic
control
center,
it
is
central,
ruled,
are
maintaining
the
position
and
the
trajectory
of
vicars.
So
maybe
our
this
currently
control
center
may
be
the
mobility
our
anchor
on
the
three
I
mentioned,
the
people
so
I
assume
in
Turner
in
vicar
or
RSU.
I
I
Second,
one
or
so
similarly,
the
vicar
to
vicar
can
communicate
with
each
other.
The
scenario
is
a
reading.
Vicar,
it
accommodate
multiple
server
nets.
The
question
is
some
post
and
on
another
posted
in
another
vehicle,
such
as
the
adaptive
cruise
control
or
pro
tunings
of
this
case
they
should
communicate
with
each
other.
So,
in
order
to
that,
we
needed
to
support
the
prefix,
exchanging
I
would
say
prefix
discovery,
which
means
our
life
it'd
only
find
that
the
other
vicar
internal
network
or
prefix.
The
second
one
is
a
service
to
discovery.
I
So
internally
they
accommodate
the
some
services.
I
mentioned
the
adaptive
controller
or
routing
service
other
some
social
network
services.
So
the
question
is
how
to
advertise
their
services
quickly,
so
I
suggest
the
to
option
for
them:
Deacon,
prefix
information
and
service
information
predicts.
The
second
topic
is
our
DNS
naming
service,
so
the
first
one
is
in
vehicle
devices,
maybe
some
engine
or
some
other
on
the
hill
or
the
user
to
mobile
devices,
maybe
the
in
vehicle
devices.
So
the
question
is
how
to
accommodate
the
auto
configuration
of
their
name
without
intervention
of
the
users.
I
The
second
one
is
a
device
''tis
discovery,
so
in
vehicle
devices
can
discover
other
devices
by
getting
DNS
database
DB
such
as
a
jean
file,
so
it
can
achieve
it
by
from
the
DNS
server
so
and
service
discovery
a.
Similarly,
we
can
add
discover
service
last
one
in
the
name
resolution,
so
this
figure
shows
the
stab
our
DNS
of
the
comparation,
so
IOT
device.
We
can
say
the
in
vicar
device
can
get
loud.
I
Elavil
high
demand
having
the
Guinness
SL
tiene
socialist,
and
generated
its
name
and
double-check
the
uniqueness
using
the
ad,
because
DNS
name
I
can
generate
its
global
IP
address
and
then
launcher
can
collect
the
DNS
name
and
address
using
another
implemented
quarry
and
then
router
has
some
authentication
to
deliver
its
IT
device,
DNS
name
to
a
DNS
server
using
DNS
update
after
that
user
device.
Can
figure
out
the
in
vehicle
device
using
the
contacting
with
the
DNS
server?
I
The
last
one
is
our
Cindy's
IP
our
networking,
the
handling
of
a
MAC
address
changes
because
of
the
pseudonym
to
protect
the
privacy
of
our
address.
The
problem
is
MAC.
Address
change
emails,
the
IP
address
change,
so,
in
that
case,
ongoing
TCP
or
UDP
station
can
be
broken,
so
we
need
to
do
handle
so
on
a
possible
solution.
Is
we
can
using
some
mobile
IP
a
binding
update
so
also-
or
this
issue
case,
some
picker
of
moving
to
some
some
earlier
and
other
vicar
using
the
same
ip
address
the
addressed
conflict
right?
I
In
that
case,
we
also
need
to
resolve
with
this
IP
address
the.
Secondly,
beaker
can
approaches
another
the
RSU
coverage
e,
so
we
assuming
another
RC
you
providing
different
the
prefix
IP
address
will
be
changed
it
also.
In
that
case
also,
we
need
to
handle
the
change
of
our
IP
addresses.
So
next
step
is
our
we
can
consider
on
this
topic
for
our
work
item
for
the
chattering.
The
problem
statement,
the
case
I,
also
described
a
total
problem,
so
we
need
to
design
our
detail.
The
connector
architecture
consider
we
to
be
able
to
I
together.
C
Hello,
my
name
is
Aleksandra
Petrescu
and
I.
Have
a
very
brief
presentation
of
my
my
and
my
group
current
concerns
of
the
about
some
trials.
We
do
with
IP
for
V,
V
and
V.
I
communications
so
I
write
this
in
two
parts.
First,
I
will
talk
about
V,
V
I
have
two
slides
and
second
about
v2
I
in
our
trial
we
use
here
on
the
left
side.
C
We
use
three
vehicles,
you
see
them,
they
are
following
each
other
and
they
are
a
little
bit
a
particular
compare
to
other
vehicles
that
we've
seen
in
other
trials
is
that
they
well
first
they
are
small.
Ok,
so
you
have
self-driving
cars
that
are
small.
This
is
a
little
bit
new
and
also
they
feature
each
one
of
the
vehicles
features
actually
not
one
interface
on
the
top
of
the
roof,
but
three
interfaces,
one
on
the
top
one
in
the
front,
bumper
and
one
in
the
rear,
bumper.
C
C
H
C
16:9,
instead
of
4:3
so
but
basically
you
have
on
the
top
of
the
on
the
roof
of
the
car.
You
have
five
4G
today
in
the
future,
five
GNR
that
talks
to
an
anchoring
point
and
that
uses
DHCP
prefix
delegation.
So
you
realize
v2v
communication,
but
through
a
first
hop
router
that
is
fixed
in
the
infrastructure
that
router
is
not
called
roadside
unit.
C
It's
called
something
else:
I
forgot
PG,
now
your
sample
okay,
so
that
those
are
the
two
v2v
topologies
that
we
must
make
work
and
we
will
make
work
soon,
and
these
are
the
requirements
that
we
have.
Some
of
them
are
easy
to
satisfy.
Others
are
very
difficult
and
I'm
not
sure,
but
first
is
the
dynamic
distributed
decision
of
which
5.9
gigahertz
to
use
between
cars
such
that
to
generate
the
minimum
interference.
C
Potential
solutions
that
we
have
experimented
with
are
the
Bible
babble
routing
protocol
is
trialed,
okay
or
nato
2.11
OCB
for
v2
e.
We
also
experimented
with
prefix
exchanges
in
ipv6
router
advertisements,
trialed,
okay,
on
802
dot,
11
OC,
b
and
dhcpv6
prefix
delegation
trialed,
okay,
on
4g
deployment.
C
So
that's
about
v2
v,
and
this
is
our
our
requirements
right
now
that
we
look
at
and
maybe
some
of
them
are
compatible
with
what
other
people
would
like
to
do
at
ITF,
or
we
will
try
to
understand
this
now
going
to
V
I.
We
also
have
a
v2i
context.
The
v2i
is
something
like
this.
You
have
one
car
that
wants
to
talk
to
maybe
a
home
agent
or
IOT
platform
in
the
internet
and
that
sense
it's
its
own
data.
C
Today
when
we
use
Wi-Fi,
if
we
do
it
with
OCB,
we
don't
have
beacons
so
how
to
implement
this
movement
detection
when
we
have
no
beacons.
So
the
requirements
are
here
the
problem.
The
problem
is
that
for
mobile
IP
to
work
in
v2i
mode,
a
movement
detection,
protti
procedure
is
needed.
Current
movement
detection
procedures
on
mobile
rotors
measure
and
compare
and
compare
signal,
strength
of
different
beacons,
but
all
it
802
dot.
Eleven
OCB
has
no
beacons
the
so
the
requirements
is
that
a
movement
detection
procedure
that
measure
the
signal.
C
Strength
of
an
IP
message,
some
IP
message
that
could
be
used
as
a
beacon
and
this
message
that
these
IP
should
contain
enough
IP
and
layer
2
information
to
assist
the
mova
router
to
decide.
It
must
change
the
subnet
and
then
decide
to
send
the
binding
of
it
arrived
at
our
beloved
mobile
IP,
ok
and
it
is
not
sufficient
to
just
put
the
MAC
address
and
the
IP
of
the
IP
RSU
and
the
prefix
of
the
link.
This
is
not
sufficient.
You
need
much
more
to
realize
it
so
yeah.
C
Maybe
a
router
odd,
maybe
an
enhanced
router
advertisement
or
some
other
IP
message
that
could
be
used
as
become
so
that's
the
end
of
the
requirement
for
movement
detection
in
OCB
mode,
and
this
could
probably
be
fit
into
the
neighbor
discovery
discussion
if
neighbor
discovery
means
router
advertisement
or
neighbor
advertisement
and
yeah.
This
is
the
things.
H
E
H
E
C
C
E
To
go
target
by
using
a
multicast
to
find
something,
and
then
it
needs
to
have
a
something
that
can
secure
that
communication,
like
the
license
plate
having
a
QR
code
with
the
public
key,
the
the
key
fingerprint
or
something
right,
but
it
has
the
bootstrap
from
something
not
saying
it's.
The
DNS
now
I
have
nothing
against
the
DNS,
but
it
seems
like
the
wrong
recreating
the
same
stack.
We
have
created
four
fixed
communication
for
this
entirely
local
and
extremely
ephemeral.
Communication
between
two
neighboring
cars
is
like
you're,
starting
with
the
wrong
premise.
Yeah.
C
C
E
Is
why
a
QR
code
with
a
key
fingerprint
on
the
license
plate
might
be
reasonable
thing,
because
now
I
can
actually
see
that
that,
if
I
send
something
over
radio,
I
actually
know
that
it's
that
thing
that
has
that
QR
code
printed
on
its
license
plate?
Yes,
whether
that's
a
good
solution.
What
the
privacy
considerations
are.
There's
lots
of
questions
there,
but
at
least
I
have
a
way
of
bootstrapping
this
in
the
physical
world,
as
opposed
to
thinking
about
the
current
internet
approach.
To
do.
C
M
To
the
second
part,
you
know
in
different
stos,
usually
what
happens
is
you
have
an
application
that
sends
a
service
announcement?
It's
now
saying
that
is
a
service
provider
and
the
other
one
is
to
use
consumer
and
that,
basically,
if
you
look
at
1609,
I
still
have
to
look
with
details
on
how
it
works.
H
M
M
H
C
Description
protocol
service
announcements:
there
is
also
we
could
reuse
this
SDP
from
from
ITF.
Why
not?
And
why
using
IP
and
not
the
link
layers
link
layer
protocols
defined
that
other
ratios
in
one
of
the
slides
I
showed
this
v2
I.
It
is
a
very
strong
requirement
to
be
able
to
connect
from
the
car
to
a
server
in
the
infrastructure.
C
Basically,
you
end
up
sending
the
same
data
the
same
IOT
data
both
to
the
server
and
to
some
nearby
car
on
kill
you
they
want
it
very
often
to
send
it
twice,
and
this
is
obviously
inefficient
and
you
just
want
to
send
what
you
just
want
to
be
connected
to
the
internet,
whether
that
internet
is
the
nook
in
the
car
next
to
you
or
the
the
cloud
or
you
see
these
are
this
is
why
only
fuse
IP.
You
are
able
to
homogenize
them.
I.
L
L
H
M
Sea
in
the
fact
that
you
exchanging
sensor
data
with
l2
protocols
only
so
there
will
be
exchange
no
matter
what
and
then
the
question
would
be
okay.
If
you
want
to
transmit
it
somewhere
else,
then
you
have
an
additional
protocol,
but
using
IP
will
not
be
place,
it
will
not
say
oh
I
can
use
by
using
ipv6
I
can
avoid
transmitting
twice.
This
is
efficiency
issues,
but
the
Alto
protocol
will
transmit
sensor
data
to
other
vehicles
without
ID.
That's
already
done
and
will
be
specified
very
soon.
M
You
have
an
Etsy,
a
specification
called
cooperative
perception
message:
NSA
is
a
standardized
exactly
the
same
types
of
messages
is
l2
only
and
therefore,
if
you
want
to
transmit
it
to
io-22
to
the
internet,
which
I've
nothing
against
I,
think
it's
a
good
point.
In
any
case,
you
will
not
be
able
to
block
that
this
is
a
standard
service
that
will
be
provided
for
a
tuning
in
car
to
car
communication.
Well,.
M
C
E
E
H
B
So
well,
there
are
many
minutes
who
are
supposed
to
be
for
additional
open,
open,
mic
discussion
on
the
red
car
train.
I,
don't
know
if
anybody
else
have
any
comment
from
from
my
point
of
view,
what
I
think
we
should
do
next
is,
as
we
have
been
discussing
to
basically
for
each
of
the
problems,
people
think
work
is
required
to
actually
present
display
the
problem
analyze.
What
is
what
are
the
solutions
out
there
that
could
be
applied
to
that
document?
B
This
as
part
of
the
the
other
document,
we
have
this
prana
statement,
a
survey
document
and
identify
potential
gaps
that
you
believe
are
not
sufficiently
covered
or
that
require
additional
team
sessions
from
the
staff
that
is
already
there.
If
we
have
that
discussion,
I
think
that
could
be
used
or
then
decide
whether
the
working
group
wants
to
do
additional
things
and
there
the
the
require
commitment
from
people
to
work
on
that
and
then
based
on
that
we
we
can
see
whether
there
is
something
that
needs
to
be
done
or
not.
C
That
I
said
on
that
we
exchanged
a
few
with
two
or
three
participants
on
the
image
about
saying
these:
what
what
they
call
LP
messages
like
in
Europe
or
DSM
in
America,
sending
them
over
IP
that
the
topic
that
could
be
also
explored
and
over
IP
instead
of
sending
come
over
yo
networking,
for
example.
C
H
G
G
B
M
This
term,
Harry
speaking,
you
know
during
the
the
when
we
wrote
the
first
draft
on
CB,
is
one
thing
we
postponed.
We
put
it
out
its
security.
If
we
don't,
if
you
use
your
CB,
we
don't
have
any
security
l
to
security,
so
the
open
question
is:
is
I,
be
providing
a
bulletproof
security?
What
what
we
do
already?
If,
yes,
wonderful,
if
not
at
least
we
should
provide.
M
Because
in
if
you
use
actively
1609
or
HC,
they
have
a
very
strong
I
mean,
maybe
not
the
truth,
kind
of
things,
of
course,
but
there
is
a
strong
security
mechanisms,
including
anonymization
and
change
of
pseudonyms,
and
the
question
would
be.
We
have
to
do
that
in
that
group
we
have
to
change
our
IP
prefix.
How
often
do
that
should
be
synchronized
to
60
or
9.3
do
that
and
how
about
security,
because
otherwise
my
perspective
is
an
issue
with
that.
I
Okay
about
our
survey
and
problem
statement,
our
document,
so
during
and
this
what
what
can
grow
bidding,
we
discussed
a
while
its
next
expansion,
so
a
current
document.
So
they
are
many
topics,
so
we
need
to
focus
on
some
of
them
right.
So
I
would
write
such
as
the
some
additional
section
to
focus
on
our
work,
every
walk
right
so
3.
What
this
meeting,
or
also
after
or
melanistic,
discussion
we
can
speak
of
and
what
what
we
should
have
done
this
working
and
then
I
will
lurch
and
then
I
add
additional
section
for
our
whatever.
I
L
Lucy
B
IP
lack
security
thought
that
I
should
be
pretty
well
known,
that
we've
got
giant
and
security
problems,
yes
by
design,
so
don't
don't
go
looking
for
it
to
the
ITF
for
help.
We
don't
have
any.
As
far
as
the
security
mechanisms
that
are
already
built
in
my
understanding
of
the
situation
is
that
it
amounts
to
rotating
MAC
addresses
and
that
is
typically
going
to
break
everything
in
IP.
The
MAC
address
is
associated
strongly
with
the
IP
address.
L
The
IP
address
is
strongly
associated
with
your
sessions,
and
so
so
you
change
the
MAC
address
you're,
going
to
basically
break
the
TCP
connections,
so
you
can't
change
the
MAC
address
when
anything
is
active.
If
you
only
change
the
MAC
address,
when
the
vehicle
is
say
powered
on
initially,
that
means
you
can't
have
TCP
connections.
While
the
thing
is
off.
That
also
means
that
you're
gonna
break
anything
when
you
key
on
and
that's
problem
so
I.
For
my
my
perspective
and
my
implementations,
you
can't
actually
change
it
and
that
actually
doesn't
work
at
all
on.
J
Okay,
so
I
can
point
you
to
that
after
the
meeting-
okay,
so
like
other
than
that,
like
I'm,
finally
exploring
all
these
things,
but
in
a
reasonable
time,
I
would
like
to
see
a
shorter
list
of
items
like
there's
like
a
lot
of
things,
you
can
do
first
thing
you
should
figure
out
whether
we
should
do
them
here
and
the
second
thing
is
like
you
know,
who's
interested
in
doing
it,
and
then
we
can
continue
this
process,
but
I
really
want
the
other
document
done.
Okay,
so
before
we
do
our
each
other.
H
M
Ip
wave
with
other
standards,
because
if,
for
example,
you
have
you
have
ipv6
I
mean
within
a
few
way
that
is
operating
on
the
vehicle
with
the
1609
operating,
then
you
will
have
to
change
your
IP.
Your
MAC
address
because
the
DSRC
will
change
is
back
dress.
So
then
you
will
have
to
do
that
not
to
break
any
security
in
an
elimination
from
the
second
I
mean
at
the
same
point
see
if
the
car
does
not
have
69
out
there,
then
maybe
big
a
speaking
too
quickly
on
that.
M
N
Hi
Sandra
Cespedes
I
just
have
a
question
of
adding
a
document
of
the
problem
statement
because
you
said
we
need
to
focus
on
specific
problems
and
then
maybe
provide
some
of
the
service
solution.
But
I
don't
see
at
the
regulation
with
a
problem
that
they
put
together
today
for
Deborah
chartering.
So
are
we
supposed
to
include
those
because
so
some
of
them
are
mentioned
in
the
in
the
document
like
the
link
models
and
IP
nobility,
but
somewhere
else,
or
not
from
those
that
we
saw
today?
N
B
My
points
that
we
need,
of
course
it
should
be
the
consensus
of
the
working
group.
What
are
the
problems
that
are
worth
being
explore?
So
we
need
to
have
that
I
mean,
and
this
requires
consensus.
We
need
to
discussion
in
the
mailing
list,
then
out
of
this
or
from
the
set
of
problems,
we
survey
solutions
that
apply
and
maybe
document
how
those
solutions
could
be
used
and
as
part
of
that,
we
identified
potential
gaps
unless
the
problem
kind
of
the
prana
statement.
B
Part
of
the
document
like
okay,
this
is
not
yet
done
or
address,
or
there
is
some
optimization
that
could
be
applied.
This
kind
of
thing,
but
I
mean
we
need
to
agree
on
the
problem
that
we
want
to
document
in
the
in
the
document,
and
that's
applies
to
the
system
solutions
they
may
be.
That
is
I
mean
we
need
reviews
from
the
working
group
to
see.
What
are
the
working
group
agrees
on?
B
What
is
in
the
document
is
fine,
something
needs
to
be
removed,
something
needs
to
be
added,
so
we
need
that
anyway,
and
the
proposal
will
be
that
I
mean
for
the
resulting
discussion
to
ensure
that
whatever
topics
are
meant
to
be
potentially
addressed.
Those
should
be
in
the
document
because
it
doesn't
make
sense
to
have
a
potential
resort
or
topic
recharging
topic,
which
is
not
in
the
in
the
prana
statement,
so
that
for
sure.