►
From YouTube: 2022-01-05 Monthly Multicast W3C Community Group Meeting
Description
Notes here:
https://docs.google.com/document/d/1hhtgQVjy73FTWP4AIZA_F3C2Pkqtc0Sy_NG2ebwy-fU/edit
A
Okay,
recording's
on
yeah,
so
thanks
for
coming
notes
are
pasted
in
the
chat,
and
I
guess
I
was
just
gonna-
go
through
a
little
bit
of
of
kind
of
the
status
of
where
things
are
now
and
answer
any
questions.
Anybody
has
but
just
kind
of
give
a
quick
update
to
anybody,
who's
interested
and
ask
for
feedback.
A
I
mean
major
contributions
to
the
itf.
Work
probably
should
be
done
in
an
ietf
forum,
but
I
think
there's
a
you
know,
there's
certainly
some
crossover
here
and
some
relevance
between
them.
I
I'm
hoping
that
anybody
participating
here
will
be
sort
of
somewhat
tracking
them
with,
although
maybe
a
slightly
different
focus.
A
A
Sec
dispatch
follow-up
I
need
to
I
I
think
I
I
mentioned
last
time
that
one
of
my
major
tasks
was
going
to
be
to
kick
off
the
msec
discussions
and
I'd
not
actually
managed
to
do
that.
I
was
late
and
then
there
was
the
holidays
and
I
thought
it
wasn't
going
to
be
useful
anyway,
so
I'm
I'm
still
planning
to
do
that
and
today
or
maybe
tomorrow
I
was
working
on
that
this
morning.
I
can
share
some
of
the
text.
I've
got.
C
C
A
A
C
A
I
I
think
I
have,
I
think
I
got
noticed
last
week
that
several
of
my
unborn
drafts
are
expiring
and
I
need
to
update
them.
Yeah,
there's
kind,
I
feel
like
there's
a
lot
to
juggle,
but
it's
okay
right.
So
the
the
there
are
just
a
few
topics,
so
web
transport
did
respond
to
my
request
to
consider
adding
multicast
as
a
use
case.
A
A
A
So
if
I
can
get
that
rolling,
then
that's
my
guess
is.
Since
I
mean
there,
there
was
some
w3c
support
for
this,
but
with
web
transport
dragging
their
feet,
then
it's
gonna,
be
I
don't
know
if
it's
gonna
go
anywhere
this
year,
really
or
maybe
without
getting
some
kind
of
some
kind
of
deployment
rolling
just
to
prove
it's
credible.
A
A
A
So
the
idea
there.
So
I
I
don't
know
if
I've
given
any
updates
to
you
and
my
chris
there,
mike
and
chris
here,
but
last
month
I
started
working
on
trying
to
to
stand
up
the
web
transport
tests.
There's
web
platform
tests
for
web
transport
and
to
fold
multicast
into
that.
So
it
would
be
a
sort
of
very
similar
to
so.
The
bbc
had
a
multicast
http
over
quick
draft
some
time
ago,
but
they've
kept
it
up
to
date
and
there's
an
implementation
out
there
and
it
it.
A
It
will
do
it's
designed
to
do
the
segment
transport
and
they
had
a
demo
that
was
running
in
in
london.
I
think
it
was
at
103
or
101
or
something
ietf,
103
or
101..
They
still
have
kept
the
draft
active,
but
I
have
not
updated
it
much.
They
have
the
the
source
code
published
for
that.
This
is
not
the.
This
is
not
web
transport,
but
it
is,
but
it
is
a
sort
of
multicast
based
delivery
path.
A
They
did
not
have
the
the
packet
level
integrity
that
we're
proposing
with
ambi
that
that,
I
think
is
gonna
be
important
for
the
reasons
we
we
put
in
the
multicast
security
consideration
stock
right,
it's
like
attackable
buying
because
it's
possible
to
inject
trap,
and
if
you
can
get
traffic
injected,
then
one
packet
can
eliminate
a
whole
segment's
worth
of
data
which
can
be
you
know,
a
thousand
packets
or
more
and
so
that'll.
That
would
serve
as
a
denial
of
service.
A
You
know-
and
it's
also
possible
to
like
probe
for
the
success
of
an
injection.
So
if
you
have
so,
if
you
have
unauthenticated
packet
level
data,
then
I'm
not
sure
it's
viable.
They
would
throw
away
the
entire
segment
and
and
re-fetch
it
with
unicast.
If
this
happened,
but
I
I
am
worried
about
the
viability
of
rolling
out
something
that
does
it
that
way,
as
opposed
to
being
able
to
reject
individual
bad
packets.
A
But
there
is
an
implementation
of
it
and
what
I
was
going
to
try
and
do
is
something
similar
in
the
current
web
platform.
Tests
for
web
transport
with
datagram
support
so
basically
add
some
extension
frames
to
quick
that
that
do
a
few
things
that
we're
going
to
need,
this
is
going
to
be,
like
you
know,
a
notify.
A
A
Sorry,
well
what
you'd
be
able
to
do
with
that
is
just
take
any
existing
udp
stuff
at
this
at
the
sender
side
and
receive
the
udp,
the
udp,
packets
and
sort
of
shove
it
into
the
web
transport
data
channel,
and
then
it
would
get
unpacked
as
a
datagram
on
the
other
side.
So
you
could
do
the
same
kind
of
port
that
the
current
demo
has
where,
once
you
had
it
running
in
udp,
then
you
can
just
sort
of
ship
it
as
a
web,
app
that
that
receives
datagrams.
B
Question
and
comment,
so
I'm
I'm
just
trying
to
help
you
to
progress
this
work
this
year.
Somehow,
and
so,
as
far
as
the
itf
is
concerned,
it
sounds
like
what
you're
saying
is,
even
if
you
were
to
go
the
ball
route
or
maybe
reopen
msec,
which
would
also
include
going
the
ball
route
by
the
way.
You're
saying
that,
even
if
you
were
to
form
a
working
group,
you
still
would
need
to.
B
Since
it's
extensions
to
quick
you'd
still
need
to
get
blessings
from
them,
which
I
think
is,
I
guess,
that's
the
discussion
that
would
need
to
be
had
because
I
it
could
still
be
worth
that
process.
I
mean
that's
how
the
idf
works.
I
mean
you
create
a
community.
If
there's
enough
interest,
then
maybe
it
forces
the
hand
of
people
to
say
yeah.
Actually
we
probably
should
do
this
instead
of
jake
pushing
this.
So
it's
something
to
consider
this.
A
B
A
B
B
A
A
A
I
think
that
would
be
reasonable,
but
there
were
some
good
comments
in
the
chat.
I
think
there
are
also
a
few
reasonable
comments
in
the
in
the
sect
dispatch
mike
line.
I
think,
but
basically
I
think
the
main
one
to
me
really
was,
or
the
maybe
the
biggest
one
was.
You
should
have
a
specific
proposal,
which
is
what
I'm
trying
to
do
with
the
with
the
quick
extension
and
then
we
might
need
to
flush
out
the
security
model.
A
A
A
I'm
not
sure
what
that
means
exactly,
but
I
do
think
that
that
question
hinges
on
like
what's
the
threat
actor
here
and
what
are
they
doing
and
do
they
have
that
capability
today
and
are
they
using
it?
That's
another
question
because
you
know
the
tls
traffic
analysis
stuff,
like
I
found
a
paper
here.
That's
about
you
know,
somebody
asked
about
you
know,
can.
Can
my
service
providers
see
what
I'm
watching
on
netflix
and
the
answer
is?
A
You
know
so
I
feel
like
this
is
the
discussion
that
has
to
happen
on
the
msec
list
and
the
one
that
I'm
trying
to
kick
off
and
then
the
other
big
piece
of
it
is
like
having
a
demo
to
run
that
has
specific
security
properties
that
people
can
argue
about,
the
other
one
they
brought
up
was.
There
was
a
threat
about
the
ipsec
and
the
group
security
key
exchange
stuff.
A
And
I
I
think
it's
like
I
I'm
not
a
huge
fan
of
it
for
wanting
to
use
it,
but
I
do
think
that
that
what
it
does
is
is
maybe
something
worth
referencing
inside
the.
Instead
of
this
security,
it
does
sort
of
have
the
same
privacy,
consideration
questions,
and
it
has
this
sort
of
answer
to
to
some
of
the
issues
right.
There
is
this.
Certainly
the
you
know
the
concept
of
the
group
key
and
the
and
how
it's
shared-
and
you
know,
there's
a
bunch
of
stuff-
that's
all
documented
there.
A
A
So
yeah,
I
think
that's
the
outline
of
the
plan
is
like
have
a
demo
running
in
vienna
and
have
a
see
if
I
can
get
a
buff
put
together
for
for
philly
and
then
and
then
we'll
see
where
we
stand,
but
that
that
was
basically
the
web
transport
feedback
is
that
we
can't
pull
it
through
them.
We
have
to
do
something
on
the
ietf
side
and
get
get
some
engagement.
There.
B
A
A
A
A
I
still
have
some
hope
that
we
will
do
something
real
sorry
about
that
that
we
will
do
something
real,
but
I
am
disappointed
with
with
how
far
we've
got
on
what
I
can
announce
yet,
and
I
still
don't
know-
I
guess
I
I
expect
to
be
under
what
so
it's
gonna
have
to
to.
You
know
have
scoping
that
that
keeps
it
from
getting
too
big.
A
C
A
The
universal
software
down
yeah,
all
right
and
the
other-
is
to
do
some
software
download,
so
we're
basically
and
it's
possible.
We
would
do
both,
but
I
I
think
that
these
would
end
up
being
sort
of
two
different
product
lines
essentially,
and
I'm
I'm
trying
to
put
together
a
sort
of
minimum
viable
product
to
get
something
off
the
ground
on
each
of
those
separate
question
within
each
of
those.
Is
then,
whether
it's
wild
garden
or
or
the
sort
of
external
ingestion
of
multicast,
I
think,
of
a
global
multicast
address?
A
A
A
But
the
idea
there
is,
like
you
know,
the
standards
based
external
ingest
of
global
multicast.
So
that's
one
path,
but
it's
sort
of
missing
a
big
piece
of
that,
so
the
other
path
would
be
would
be
a
wild
garden
deployment
for
for
either
of
those,
the
video
delivery
or
the
or
the
you
know,
game
delivery
file,
delivery,
kind
of
system.
A
A
I
mean
this
would
be
our
own
like
little
web
service
right,
so
it
would
be
like
a
you
know.
Checksum
database
to
decide
whether
my
ip
should
try
to
use
a
global
address
or
a
global
set
of
addresses
or
try
to
use
one.
That's
that's
from
a
walled
yard
in
the
local
isp,
so
that
doesn't
seem
undoable
by
any
stretch.
But
it
seems
like
probably
the
direction
we'll
have
to
go
to
get
some
deployment
to
get
enough
deployment
to
spur
browser
interest.
A
A
E
I
I
agree
with
the
approach
of
doing
something:
if
we're
not
going
to
be
able
to
get
browser
support,
I
mean
you've.
I've
talked
about
how
there's
there's
a
lot
of
these
chicken
and
egg
scenarios
where
people
say
yeah,
I'm
interested.
But
what
about
this
piece,
or
something
like
that,
so
you
know
just
like
with
the
little
prototype
you
have
going
right
now
with
that.
E
E
I
do
think
that
if
there
is
a
path
forward
using
that
it
may
it
may
actually
be
an
advantage
with
just
how
people
tend
to
latch
on
to
things
they
see,
as
you
know,
kind
of
quote,
unquote
the
future.
You
see
a
lot
of
work
around
quick.
You
see
a
lot
of
development
around
quick.
I
think
I
mentioned
when
I
was
talking
to
someone
internally
that
their
one
of
their
first
questions
was
something
like.
Does
it
incorporate
quick
or
could
it
use
quick
or
something
like
that?
E
So
it's
it's
almost
like
you
know,
kind
of
using
something
brand
name
or
well
known
in
order
to
you
know,
get
some
attention
to
your
product
or
whatever
so
definitely
focusing
on
any
solution
that
gets
us
forward
with
with
quick,
I
think,
might
help
even
getting
attention
from
browser
support.
Just
if
it's.
A
A
C
C
But
I
wonder
if
you
know,
if
we
let's
say
we're
successful
in
getting
this
into
quick
and
therefore
into
web
transport
five
years
from
now.
What's
that
traffic
balance
going
gonna
be
right,
multicast
datagrams,
using
quick,
may
account
for,
like
the
overwhelming
majority
of
the
traffic.
If
we're
successful
right,
existing
applications
might
then
just
might
then
just
switch
over.
A
Yeah
yeah,
okay,
certainly
it
would
be
possible
to
do
you
know
to
make
changes
to
the
sender,
also
right,
so
that
you
so
you
have
a
sort
of
cleaner
sender
side.
You
know
solution
and
yeah.
If,
if
that's
like
deployed,
then
I
would
expect
new
applications
to
probably
start
from
that
yeah.
But
you
know,
I
think
you
can
do
it
with
a
sort
of
center-side.
C
C
A
A
A
A
A
C
I
you
know,
as
we've
discussed,
there
are
a
bunch
of
good
counter
arguments
and
kind
of
countervailing
properties
that
mitigate
a
lot
of
that
right,
things
that
are
not
even
possible
with
unicast
right
that
actually
improve
the
posture
in
some
ways
right.
So
it's
kind
of
a
it's
a
trade-off
and
what
I
want
like,
I
think
the
you
know.
Ultimately,
if
we're
successful
here,
we're
it's
not
going
to
be
by
convincing
them
that
this
is
perfect,
but
that
the
trade-offs
are
reasonable
for
the
benefit
that
is
gained.
C
A
Why
is
it
a
problem
and-
and
I'm
I'm
still
looking
for
that-
I
I
don't
know
that
there's
definitely
not
a
problem,
but
especially
when
you
couple
it
with
like
the
traffic
analysis,
attacks
against
tls
that
you
know
really
do
give
the
same
kind
of
capability
and
farther
upstream
than
multicast.
Has
I
don't
like
if
we
had
information
about
what
threat
actors
are
doing
that
and
what
they're
doing
with
this
information,
then
that
would
be
helpful
too.
Maybe.
C
No,
I
mean
look
and
I
would
not
suggest
that
as
an
approach
right,
but
I'm,
but
but
that
doesn't
mean
that
it
isn't
what's
happening
right.
So
I
I
what
I
so
I
think
you
know
like.
We
need
to
recognize
the
reality
that
that,
as
an
incumbent
tls
is
going
to
have,
is
going
to
get
an
easier
ride
than
some
new
model
when
it
comes
to.
C
A
A
Mean
my
favorite
answer
here
is
like
let's
just
start
shipping
some
traffic
for
real.
That
would
be
great,
and
then
you
know
if
we
actually
do
get
threat
actors
abusing
it.
Somehow,
then
we'll
have
a
better
idea
what
they're
doing
and
we
can
talk
about
whether
we
should
stop,
but
I
feel
like
we
should
start
shipping
this
traffic.
If
we
can
I'm
trying
to
make
that
case.
C
Does
it
does
anyone
else
here,
like
greatly
disagree
with
what
I
with
what
I've
said
or
what
what
I've
sort
of
suggested
is
is
the
dynamic
here,
because
I
I
mean
I
could
be
totally
off
base
people
might
see
something
totally
different
than
I
am
right
and
I'm
not
trying
to.
I
don't
want
to
try
and,
like
you
know,
color
all
of
this
work
with
a
you
know
with
you
know,
as
like
some
crusade
against
you
know
against
recalcitrant.
A
A
B
A
E
I
do
agree
at
that
point,
even
if
it
lacks
some
amount
of
security.
I
mean
as
long
as
you're,
not
violating
some
contract
or
agreement
or
something
like
that.
It's
just
I
don't
know.
If
I
want
to
call
it
dummy
traffic
like
we
could
probably
send
something
real,
but
just
you
know
making
sure
it's
nothing
that
sensitive
or
you
know
something
that
if
it
does
somehow
get
compromised
or
watched
outside
of
how
we
intended
it-
and
it's
no
big
deal,
but
you
know
kind
of
is
like
a
higher
level
demo.
E
I
guess
you
don't
use
like
at
least
internally.
If
we
were
going
to
do
a
demo
with
something
we
don't
necessarily
use.
You
know
hbo
content
which
has
contracts
and
contracts
and
things
like
that,
but
we
can
use
some
other
types
of
video
that
are
a
little
more
friendly
and,
I
think,
there's
even
open
content.
If
that's
even
the
right
word
like
for
open
caching,.
E
A
Well,
okay,
but
what
I
mean
is
sending
real
traffic
that
people
use
mostly
right.
That's
the
and-
and
I
do
think
it's
going
to
be
important
to
have
authentication,
because
I
I
would
actually
feel
really
guilty.
If,
like
you
know,
everybody
watching
some
sports
event
got
got
to
their
computer
zoned
or
their
tv
zone
or
whatever.
E
C
All
right
yeah,
if
I
think
I
I'll
just
I'll
just
say
as
a
as
my
final
sort
of
comment
on
the
on
the
approach,
I
like
the
approach
of
actually
delivering
real
traffic
and
I
think
a
you
know.
You
can't
run
a
counter
factual,
but
my
guess
is
that
if
google
had
started
quick
by
trying
to
standardize
it,
it
probably
never
would
have
gotten
done
so
it
would
not
be
where
it
is.