►
From YouTube: IETF103-MMUSIC-20181108-1350
Description
MMUSIC meeting session at IETF103
2018/11/08 1350
https://datatracker.ietf.org/meeting/103/proceedings/
A
A
A
A
A
So
for
the
working
group
status,
it's
basically
the
same
as
it
was
the
last
time.
We
have
a
whole
lot
of
documents
and
it's
long
in
the
RFC
editor
queue,
the
famous,
maybe
even
infamous
cluster,
two
three
eight
waiting
for
one
another
and
we
had
a
minor
update
of
the
bundle
negotiation.
The
time
between
this
meeting
and
the
last
one
updating
a
reference.
Basically.
A
A
We
have
two
other
drafts
that
are
alive
in
the
working
group.
We
have
the
key
share
documents
that
need
some
rewrite
based
on
review
comments,
and
we
have
the
data
channel
MSRP
draft
that
has
died
in
the
working
group.
We
decided
not
to
pursue
that
one
will
soon
expire
and
we
didn't
have
any
expert
reviews
since
last
meeting.
C
A
A
E
Under
get
on
an
ice
co-chair,
so,
yes,
we
have
pretty
much
done
with
all
our
documents.
They
are
there's
no
new
documents
that
are
being
up
and
coming.
We've
been
waiting
for
the
infamous
cluster
to
finalize
and
make
make
sure
there's
no
new
things
that
need
to
be
very
nice,
but
at
least
at
the
moment
it
seems
that
we
are
ready
and
we
could
be
declaring
success.
A
Okay,
so
the
proposal
so
far
has
been
that
since
yeah,
you
go
first,
since
the
ice
documents
were
actually
mediated
in
a
music
and
further
developed
in
the
ice
working
group
for
proposal
is
that
a
music
should
continue
to
maintain
the
hospitai
specifications,
and
thus
the
music
charger
would
need
an
update.
In
that
case,
this.
D
handling,
new
maintenance
and
extensions
as
we
would
handle
any
new
work,
he
asked
the
question
one
question
he
asked
and,
and
I
think
know
the
answer
to
that
would
be
depending
on
what
part
of
ice
we're
doing
maintenance
on
you
know
if
we're
talking
about
core
timing
thing
that
sort
of
thing,
if
in
music
is
the
place
with
the
right
expertise
for
that
I.
Think
that
my
reaction
to
that
is
well,
it's
the
same
people
wherever
we
do
it
so
I
have
strong
feelings.
I
just
wanted
to
relay
that's.
D
F
G
As
individual
I
mean
I
think
a
few
years
ago
there
people
had
a
number
of
ideas
of
things
to
extend
ice
with
which
they
didn't
get.
You
know,
eventually
they
lost
energy
and
doing
them
I'm,
not
sure
how
much
of
that
was
because
they
were
just
you
know,
all
the
interview
went
to
just
getting
reberty's.
He
finished
instead
and
that
might
come
back
when
people
think
that's
all
good
and
stable,
or
maybe
they
just
still
decide.
B
B
It's
some
of
its
expressed
in
drafts
and
I.
Think
if
I
my
concern
is
that
it's
not
at
all
clear
how
it
how
it
works
in
all
of
the
corner
cases,
so
I
think
if
there
were
to
be
some
charter
work
on
it,
it
might
help
at
least
to
elucidate
what
how
things
are
supposed
to
work
right
now.
You
know.
Obviously
the
drafts
we
published
are
good
enough
to
interoperate
with,
but
interoperating
with.
What's
out
in
the
field
is
actually
pretty
difficult.
You
have
to
do
a
lot
of
reverse-engineer.
A
D
B
B
You
know
it
depends
how
far
you
want
to
go.
I
think
I,
think
the
energy
for
a
complete
rewrite
is
very,
is
relatively
low,
but
if
they
were
targeted
things
that
could
demonstrate
significant
performance
enhancements,
that
would
be
a
little
bit
different
or
you
know,
alternatively,
if
you
win
through
some
of
these
things
and
were
able
to
prove
that
they
didn't
provide
performance
enhancement.
That
would
also
be
useful,
but
people
seem
to
believe
they
do
it's
just
there.
B
C
C
B
E
So
yeah,
of
course,
that
is
a
one
option,
but
I'm
I'm
thinking
more
kind
of
how
do
you
get
the
work
done
point
of
view.
We
are
a
lot
of
a
small
group
when,
when
ice
work
was
originally
in
EMU's,
it
Emily
a
lot
of
people
a
lot
of
fast
but
activity,
and
it
made
a
lot
of
sense
to
split
the
ice
away.
Now
we
are
much
more
group,
much
smaller
amount
of
documents
and
it's
same
people
and
music
ice
they're
fighting,
even
given
that
doing
it,
music
actually
would
make
sense.
E
G
Yeah
Jonathan
the
next
month,
the
nation
will
be
to
say,
do
it
through
a
music
unless
something
comes
up
where
it
looks
like
it's
big
enough
that
it's
worth
that
it
would
need
its
own
working
rope,
in
which
case
spin
up
a
new
mini
working
group
or
resurrect
ice
six
and
one
exactly.
Yes,
that's
exactly
so.
This
language
would
match
that
my
idea
so.
D
So
this
has
been
again
and
and
I'm
will
leave.
You
know,
consensus
choices
to
the
chairs,
but
what
I'm
hearing
is
that
people
think
this
it's
reasonable
to,
but
something
scoped
like
this,
which
would
be
small
work
and
a
music,
and
that
in
music
is
the
right
place
to
do
that
sort
of
thing.
But
then
we
don't
want
a
music
to
take
on
a
fundamental
rewrite
or
ice.
This
sort
of
thing
without
without
a
recharter.
D
Right
well,
the
process
of
new
work
has
a
lot
of
ways
you
can
come
in.
It
could
be
dispatcher
or
just
what
dispatch
does
is
a
recommendation?
There's
a
number
of
ways
we
can
do
those,
but
in
any
case,
I,
don't
think
we're
taking
on
something
like
Isis
biz,
without
recharging,
wherever
it
goes.
Turk.
E
B
E
B
Well,
just
that
I
think
there's
a
part
part
of
this
would
be.
There
would
be
an
empirical
aspect
where
you
know,
because
I
think
a
lot
of
it
is
about
performance,
optimization
and
I.
Don't
know
that
that's
necessarily
a
musics
forte,
but
I
think
you
basically
have
to
prove
out
some
way
of
proving
that
these
things
make
sense.
If
they're
all
about
performance,
you
know
and
they're
all
about
optimization.
You
kind
of
have
to
show
that
in
fact
you
do
optimize.
B
D
E
G
E
And,
and
also
on
the
topic
of
like
getting
the
right
people
who
have
the
stomach
to
talk
about
these
things,
I
mean
we
can
pull
people
from
different
communities
when
that's
what
it
takes.
I
mean
I
think
help
upper
80s
and
others
like
we
do
that.
For
example,
in
high
school
community
transport
expertise,
we
got
a
bunch
of
transfer
people
there
and
I
did
the
discussants
a
nice
group
I'm
sure
we
can
do
something
similar
in
music
and
that's
a
big
problem,
but
of
course,
but
I
do
I
agree
it
Bernard.
E
A
B
An
example
of
something
you
would
do
if
you
really
wanted
to
get
activity
would
be
to
have
a
hackathon.
You
know
have
several
of
these
extensions:
implement
it
do
tests.
You
know
whether
they
perform.
If
you
know,
provide
any
optimization,
it's
it
would
be
a
fairly
focused
and
intense
activity.
I'm,
not
saying
that
people
a
missus
I,
don't
know
that
people
would
sign
up
to
do
it,
but
that
would
be
the
kind
of
thing
I
think
you'd
have
to
organize.
If
you,
if
you
wanted
to
make
progress
on
it,
but.
D
B
I
think
there
needs
to
be.
Let
me
put
it
this
way.
I
think
there
needs
to
be
if
this
I'm
not
going
to
say
that
there's
enough
interest
to
make
this
happen,
but
if
the,
if
the
desire
is
there
to
make
it
happen,
I
think
you
have
to
have
some
framework
put
in
place
for
it.
Like
I
said
it's
optimization.
So
if
you
can't
prove
that
it
actually
optimizes
anything,
it's
somewhat
pointless.
B
D
A
G
A
F
I'm
Ronnie
event,
so
I'm
trying
to
finish
the
data
Channel
SDP
negotiation
version
21,
which
I
submitted
Lai
recently,
did
some
address.
Some
of
the
comment
made
and
I
think
most
of
the
comment
made
by
Ben
as
the
in
the
area
director
review.
So
we
need
to
look,
but
there
was
an
open
go
to
next
slide,
but
there
was
a
the
data.
Channel
setup
can
be
done
using
either
the
invent
beta
channel
establishment
protocol
this
app
or
using
out-of-band
known
this
a
protocol.
F
This
document
specifies
this
just
to
recap
how
the
SDP
her
answer
exchange
can
be
used
to
achieve
an
out-of-band
non
VSAT
negotiation
normative
reference
by
clue.
Data
channel
documented
clue
signaling
to
this
document,
and
that's
why
I
would
like
to
finish
this
document
to
feel
to
finish.
Also
clue
next
slide.
F
So
the
the
open
issue
that
I
had
from
the
review,
which
it's
good,
that
it
was
reviewed
because
I
missed
it
when
I
was
doing
the
document.
Does
the
SDP
TCS
a
attribute
allows
the
negotiation
of
a
data
general
SAP
protocol
attribute
I,
give
you
the
DCC
attribute
our
SDP
attributes
that
are
specified
to
be
used
under
this
dcs,
a
level
that
is
similar
to
the
SSRC
attribute
RFC
55
76,
the
transit
on
registry.
So
this
is
the
SDP
definition.
F
The
the
name
is
DCS
a
value,
and
what's
there
is
that
the
syntax
is
stream,
ID
and
then
attribute
test
attribute,
which
is
from
RFC
4566,
which
means
can
be
any
attribute
from
RFC.
But
the
text
says
I
mean
not
a
reference.
Everything
is
viable
for
not
everything
can
be
really
used,
as
in
the
TCS
a
level,
but
most
of
them
can
be
used
and
as
an
example,
it
helps.
You
know
that
the
a
the
last
line,
I
equal
DC,
say
to
accept
type
protects
slash
plane,
for
example.
So
this
is
so.
F
F
The
the
problem
is
that
there
is
no
such
a
way
to
describe
it
because
there
are
no
registry,
doesn't
have
this
level
B,
and
that
will
go
to
the
next
line
so
for
to
address
this,
just
just
the
land
of
the
last
bullet.
So
the
proposal
is
to
define
a
new
registry
attribute
field.
B
si
si
si
si
level,
since
there
is
no
option
to
add
a
usage
level
to
the
current,
a
80
attribute
field
registry.
F
Currently
we
have
the
following
attribute
fields:
we
have
attribute
field
session
level,
attribute
field
boss
session
and
media
level,
media
level,
only
source
level
and
unknown
level.
We
have
all
these
registries
and
each
attribute
an
attribute
can
be
registered
in
more
than
one
of
them,
so
the
proposal
was
to
merge
them
into
one
registry
and
have
the
usage
level
as
a
column
inside
this,
the
the
registry
itself.
Now
this
is
not
I
mean
and
that's
what
our
area
director
said.
F
I
mean
this
is
nothing
I
mean
this
should
not
be
done
by
the
document
that
I'm
editing.
If
we
want
to
do
it,
then
it
needs
to
be
done.
It's
I
think
it's
I,
don't
know
if
it
requires
any
document
or
just
require
some
work
to
prepare
the
to
Diana
how
to
merge
them
into
one
registry.
I,
don't
know
how
to
do
this.
F
The
question
whether
we
should
do
this
and
if
we
do
that,
then
the
other
question
would
be,
should
I
wait
for
that
with
my
document
or
should
I
do
especially
specific
at
registry
and
that
will
be
merged
later
I.
Don't
have
that's
why
I'm
trying
it
as
an
open
issue,
I,
don't
know
how
to
do
to
go.
Is
it
I'm
looking
for
advice,
this.
D
Is
Ben
I'm
not
going
to
give
you
advice
in
that?
Second
part
of
the
question.
I
will
comment
on
the
first
part
of
what
she
said,
which
was
what
does
it
take
to
do?
This
kind
of
reorganization
and
I
think
that
ini
would
probably
like
to
see
a
draft
that
might
be
nothing
but
I
ana
considerations
that
probably
updates
the
RFC's
for
these
things
that
that,
but
they
created
this
stuff
it
just
to
tell
them
what
to
do.
It
could
be
really
short
and
defined.
B
H
F
I
H
D
F
F
F
F
F
F
F
F
G
B
A
G
Actually
something
that
defines
that
but
yeah
I
think
maybe
the
assumption
had
been
that
this
document
would
beat
4566
best
to
completion.
So
this
was
the
patristic
says
this
would
update
this,
but
it
didn't
so
or
maybe
it's
not-
maybe
it's
so
good
and
when
you
publish
it
like
this
and
then
varieties.
If
this
would
update
this
but
I
don't.
D
H
B
I
J
G
There's
a
bunch
of
stuff
in
WebRTC
that
was
Remini
mukhda
literally
made
the
decision.
This
is
not
going
to
be
an
SDP,
it's
only
going
to
be
in
the
JavaScript
API
and,
and
that
was
lens,
and
they
said
this
is
the
document
that
says,
if
you're,
not
using
WebRTC
you're,
just
using
SDP
offer
answer.
This
is
how
you
do
the
equivalent
things.