►
From YouTube: IETF112-DNSSD-20211112-1600
Description
DNSSD meeting session at IETF112
2021/11/12 1600
https://datatracker.ietf.org/meeting/112/proceedings/
A
All
right
folks,
remember
standards.
Work
is
a
contributive
process
that
we
all
do
together
and
we
need
volunteers
to
keep
things
moving.
So
it's
not
a
whole
lot
of
work.
It's
just
for
one
hour.
You
kind
of
keep
track
of
the
conversation
and
write
it
down
on
a
shared
document
and
barbara
will
help
and
fill
in
bits
if
you
missed
where.
But
we
need
someone
to
do
this.
B
Hi
everyone-
this
is
dnssd,
as
hopefully,
you've
gathered.
So
we
are
currently
sending
out
a
call
for
volunteers
to
write
up
some
decisions
in
the
notes
and
we
can't
proceed
until
we
got
one.
So
please
consider
volunteering.
We
promise
it
won't
be
too
difficult
and
we'll
be
eternally
grateful.
A
All
right,
I'm
gonna,
start
the
slides
and
we
have
one
for
the
minutes
and
then
we'll
pause
there
for
the
rest
of
the
session
until
we
find
someone
okay
good
morning
good
afternoon,
good
evening,
good
middle
of
the
night
everyone
and
welcome
to
dns
service
discovery.
A
A
Well,
this
is
the
last
session
of
the
week,
so
you've
probably
seen
it
before,
but
if
you
haven't
actually
read
it,
you
should
because
contributing
here
input
has
various
various
sets
of
rules
in
particular
in
terms
of
like,
if
you're
aware
of
patents,
you
are
like,
you
have
to
disclose
them,
and
also
this
meeting
is
covered
by
the
itf
code
of
conduct.
A
If
you
haven't
read
that
I
highly
encourage
you
to
in
a
nutshell,
it
says,
be
nice,
be
kind,
be
professional
and
anyone
who
doesn't
follow
these.
You
know
whether
in
the
chat
or
on
the
microphone
or
on
the
list
will
be
asked
to
leave.
So
we
have
never
had
a
problem
in
this
working
group,
but
just
want
to
always
highlight
this
okay.
So
as
as
we
mentioned,
we
are
looking
for
a
volunteer
from
the
audience
to
take
minutes
and
we
can't
start
the
session
without
one.
A
Thank
you
so
much
that
is
extremely
appreciated
and
barbara
will
be
there
to
assist.
I
can
take
care
of
jabber
scribing
because,
like
with
the
meet
echo
chat,
that
is
a
very
small
responsibility
anymore,
and
so,
but
if
you
want
to
say
something
and
you
don't
like
want
or
don't
have
a
microphone,
you
can
say
mick
mic
colon
in
the
chat
and
I
will
repeat
what
you
said
for
you
all
right
here
are
some
useful
links.
If
you're
looking
at
the
pdf
later
that
you
can
click
on.
A
As
a
reminder,
we
have
a
working
group,
github
organization
in
other
working
groups.
The
switch
to
github
has
been
incredibly
useful
for
tracking
comments
on
documents.
I
personally
really
enjoy
it,
because
it's
very
easy,
with
email
to
like
forget
one
of
the
12
comments
that
were
sent
to
one
email,
whereas
if
they're
issues,
it's
pretty
clear
when
the
issue
is
still
open,
that
you
need
to
address
it,
this
we're
not
forcing
anyone
to
use
github.
It
is
the
responsibility
of
the
editors
of
a
given
document
to
choose
if
they
want
to
use
it
or
not.
A
A
We
have
some
news
today,
so
I'd
like
to
introduce
our
new
co-chair
chris
box,
so
thanks
chris
for
stepping
up
to
join
dnssd,
looking
forward
to
all
working
together,
and
the
sad
news,
unfortunately,
is
that
barbara
is
stepping
down
at
the
end
of
this
meeting.
So
I
would,
if
you're
so
inclined,
please
unmute
and
join
me
in
a
round
of
applause
for
barbara's
three
years
of
service
thanks
so
much
barbara.
It's
been
great
co-chairing
with
you.
B
And
hello
everyone,
so
yes,
I'm
I'm
new,
but
yeah
hoping
to
do
what
I
can
to
serve.
The
working
group
awesome.
A
In
terms
of
working
group
adopted
document
status,
we
currently
have
three
documents,
so
the
one
that
we've
had
for
the
longest
time
is
srp.
It's.
We
started
working
group
last
call
in
july,
so
this
one's
almost
done.
We
had
a
recent
round
and
recent
revision,
which
addressed
those
comments
a
few
weeks
ago
and
some
more
comments
since
then,
so
that
one
seems
to
be
pretty
close.
We're
mainly
waiting
on
you
know,
authors
addressing
comments
to
make
progress
on
them
here.
A
We
did,
however,
notice
a
bit
late
in
the
game
that
that
draft
had
a
normative
reference
on
an
individual
draft
which
would
block
publication.
So
we
adopted
that
individual
draft,
which
is
the
second
item
here,
the
update
lease
in
september-
and
there
were
so
we're
right
now
we're
waiting
on
the
on
the
editors
to
take
into
account
some
comments
that
were
made
during
the
adoption
call
and
then
the
third
one
is
the
advertising
proxy
document
which
we
adopted.
A
A
The
first
two
at
the
top
on
replication
and
discovery
of
zones
were
first
published,
like
at
the
last
meeting
in
a
little
bit
discussed
there,
and
then
we
have
a
new
one
that
ted
just
published
a
few
weeks
ago,
and
our
agenda
really
flows
from
this
list
of
documents.
We're
going
to
start
off
with
the
three
that
are
that
are
adopted
by
the
working
group
and
then,
with
the
time
remaining,
the
three
that
are
individual
drafts.
A
A
Actually,
don't
do
the
share
the
screen
option
go
to
the
just
to
the
right
of
the
join
queue.
You
should
be
asked
to
share
slides.
A
D
D
Yeah,
so
basically,
I
think
I
think
dave
has
already
pretty
much
summed
up
the
current
status,
so
we
can
move
right
on
to
the
sleep
proxy
question.
D
The
issue
with
the
sleep
proxy
is
this
is
something
that's
it's
a
it's
an
mdns
feature
that
that
I
think
at
least
exists
in
in
apple's,
mdns
implementation,
possibly
and
others,
and
currently
relies
on
mdns
updates,
and
so
it
seemed
like
it
would
make
sense
to
use
srp
since
srp
is
a
dns
update
and
all
that
the
problem
is
that,
as
opposed
to
all
of
the
other
work
that's
been
going
on
in
srp,
we
really
haven't
spent
much
time
on
the
the
sleep
proxy
section
of
the
document,
and
so
it's
not
in
very
good
shape,
and
that
put
us
in
the
position
where
you
know.
D
I
got
some
comments
on
the
sleep
proxy
part
of
the
document,
and
I
basically
just
couldn't
really
address
them
easily,
and
my
conclusion
was:
let's
take
this
out
of
the
document
and
do
it
in
a
different
document
so
that
we
don't
hold
up
the
srp
work
based
on
the
sleep
proxy
work.
D
Conveniently
there
is
already
another
document,
so
so
just
a
few
things
about
this,
so
reasons
for
removing
it,
it's
not
ready
to
go.
It's
not
really
the
same
thing
as
srp,
although
it's
related,
we
have
sorry,
we
have.
We
have
another
document.
We
could
do
it
in
really
easily
the
eds
zero
owner
option
is
basically
the
thing
that
was
originally
done
to
document
the
mdns
sleep
proxy.
D
So
we
could
just
shovel
the
you
know
section
whatever
it
is
into
that
document
and
advance
that
separately,
which
I
think
would
make
sense,
and
then,
let's
see
yeah
so
and
smaller
chunks
of
work
is
good
and
yeah
eds
zero
owner
like
we
might
actually
want
to
to
be
able
to
put
that
into
into
regular
dns,
not
just
for
advertising
proxies.
So
when
I
hang
up,
that
means
I'm
not
ready
to
answer
the
phone,
not
that
I'm
bitter.
D
So
sorry,
and
then
you
know
the
the
argument
against
it
is
that
there
really
isn't
a
whole
lot
going
on
on
the
owner
option
right
now-
and
I
don't
know
if
anybody's
gonna
do
this
work.
So
if
we
don't
do
it
in
the
srp
document,
it
may
just
not
happen.
D
A
D
When
I
say
we
I
mean
a
document
has
been
has
been
worked
on
and
uploaded
to
the
tracker.
It's
not
a
working
group
document.
That
document
is
draft
cheshire,
edness,
zero
owner
option
and
oh.
A
I
see
it
and
it's
yeah
it's
expired,
but
it
could
be
revived
right
right.
D
D
It's
you
know,
we
would
need
to
revive
it
anyway
if
we
were
going
to
advance
this
in
srp.
So
it's
not
like
you
know
we
we
might
as
well
we
might
as
well.
I
think
we
might
as
well
shovel
the
work
in
there
as
well
and
and
finish
that
as
a
separate
as
a
separate
item,
because
then
it
just
decouples
these
two
pieces
of
work.
F
We
might
want
to
keep
owner
options
separate
as
a
small
self-contained
thing
in
its
own
right,
originally
ted,
and
I
thought
that
srp
and
sleep
proxy
had
sufficient
overlap,
that
it
made
sense
to
discuss
them
in
one
document,
because
they're
both
using
dns
updates
with
a
lifetime
to
put
data
into
some
server,
then
does
something
with
it.
F
But
I
agree
with
what
ted
is
proposing
now,
which
is
the
sleep
proxy
work
is
kind
of
moribund
and
it's
still
potentially
useful
in
the
future,
but
tying
it
to
the
srp
document
no
longer
makes
sense.
So
if
we
want
to
pick
that
up
later,
we
can
make
a
new
document
for
that.
It
shouldn't
hold
up
srp.
A
So
it
sounds
like
what
everyone's
or
the
proposal
is
like
leave
it
out
of
this
and
then
bring
it
in
another
document
either.
You
know
revive
that
expired
one
or
or
a
separate
one
at
some
later
date.
If,
if
someone
wants
to,
is
there
any
objection
to
this
in
the
working
group
or
any
other
stops
by
the
way.
D
Okay
and
now
the
question
is
here:
we
go
so
our
next
topic
is
the
edns
zero
owner
option
and
I'm
getting
a
big,
oh
here
here
we
go.
Okay,
sorry,
the
eds
zero
update
lease
option.
So
this
is
this
is
yeah
that
there's
our
adrik
in
the
queue.
G
A
Sorry
so
we
are
actually
we
already
did
that
so
that
sorry,
the
sleep
proxy
section
was
removed
in
version
11
or
12
during
working
hoop
last
call,
and
when
that
version
was
submitted,
we
extended
the
working
group
last
call
by
two
weeks
and
we
are
in
the
middle
of
those
two
weeks
right
now.
So
I
should.
A
D
Kind
of
the
whole
point
of
of
bringing
this
up
here
is
just
to
make
sure
that
we
don't.
You
know
somebody
doesn't
miss
the
fact
that
we're
talking
about
removing
this
and
then
later
say
well
so
anyway,
so
the
eds
zero
update
lease
option.
This
is
basically
just
like
a
piece
of
of
plumbing
that
we
need
in
order
to
get
srp
to
work.
The
problem
is
that,
and
it's
sort
of
a
general
problem:
it's
not
just
srp
when
you
do
a
dns
update.
D
This
adds
an
owner
name
and
a
bunch
of
ours.
You
know
one
or
more
rr's
to
to
a
dns
zone
and
then
those
just
kind
of
stay
there.
Unless
somebody
does
something
either.
You
know
the
client
removes
it.
The
administrator
removes
it
or
it
just
sits
there.
D
Srp
is
a
specific
refinement
of
dns
updates,
specifically
for
doing
sort
of
permissionless
advertising
using
dns
updates,
and
so
it's
not
the
only
dns
update
type
that
exists
and-
and
we
might
want
this
for
others
as
well.
So
we
have
this
this
new
option.
It's
got
a
lease
time
in
it.
It's
actually
got
two
lease
times
got.
D
One
is
the
lease
time
for
the
resource
record
that
you're
adding
and
the
other
is
at
least
time
for
a
key
record,
which
you
might
also
provide-
and
this
is
generally
specific
to
srp,
because
srp
uses
the
key
record
plus
60
to
authenticate
the
update
and
the
idea.
There
is
just
that
the
key
record
is
going
to
allow
us
to
keep
ownership
of
the
name.
Even
if
the
service
being
advertised
is
not
hasn't,
been
renewed
and
therefore
is
assumed
not
to
be
present.
D
So
it's
just
a
way
to
avoid
names
being
sort
of
snarfed,
just
as
the
service
that
was
advertising.
The
name
is
you
know
away
from
the
network
briefly,
so
so
that's
that's
the
basic
idea
and
then
the
document
is
in
pretty
good
shape.
I
think
david
mentioned
there
were
some
comments.
I
haven't
actually
had
a
chance
to
to
look
at
the
comments,
but
I
don't
get
the
sense
that
there's
anything
that's
likely
to
be
a
show
stopper.
D
So
the
question
really
is
you
know
right
now:
we've
got
some
text
that
talks
about
dns
updates
and,
and
so
we've
got
it's
there's
some
protocol,
along
with
the
along
with
the
the
sort
of
nuts
and
bolts
of
like
what
the
option
looks
like
and
what
the
pieces
of
it
do,
and
that
has
been
somewhat
moribund.
D
It's
kind
of
it's
kind
of
in
stewart's,
ballpark
and
stewart's
stewart's
up
it's
it's
up
to
stewart
to
update
that
in
a
sense,
but
stuart
doesn't
have
a
lot
of
time
for
doing
updates,
and
so
it
hasn't
been
updated.
So
the
question
is:
can
we
do
that
and
then
finish
the
document.
D
So
I
think
the
document's
ready
to
read
my
personal
feeling
is
I'd
like
to
make
the
document
as
lightweight
as
possible,
but
I'm
curious
to
know
what
the
working
group
thinks
about
that
and
whether
we
can
make
progress
on
this
quickly.
If
the
working
group
wants
to
you
know,
depending
on
what
the
working
group
decides,
so
that's
kind
of
that's
kind
of
where
it's
at
I.
I
guess
I
didn't
actually
put
this
in
the
slides
because
it
didn't
occur
to
me.
But
what
does
that
I
do?
D
The
question
of
the
working
group
is
kind
of
do
we
do
we
want
to
just
like
cut
out
all
the
fat
and
progress
this
quickly
or
should
we
really
get
the
you
know,
dns
update
part
of
this
solid
so
that
we
can
publish
that?
I
don't
think
it's
a
big
problem
to
do
the
the
second.
So
I'm
not
I'm
not
really
advocating
against
that.
I'm
just
saying
you
know
there
is
some
additional
work.
If
we
go
that
route.
A
Thanks
ted,
so
if
anyone
has
opinions
on
any
of
these
questions,
please
come
to
the
microphone.
Otherwise,
I'm
thinking
of
doing
a
poll
to
see
like
where
we
are
in
terms
of
having
read
it
peter
go
ahead.
E
Hello
on
this
option,
do
I
understand
correctly
that
it
would
also
apply
outside
of
dns
sd
situations
like
dhcp
leases
in
a
normal
bind
server
or
whatever.
D
B
D
A
Absolutely
our
our
goal
as
chairs
is
to
sorry
our.
E
A
Is
chairs
is
to,
and
we
we
had
a
discussion
when
we
first
adopted
this
document
between
all
the
chairs
of
the
various
dns
working
groups
and
your
corresponding
ads,
and
the
conclusion
was
that
we
would
adopt
and
progress
this
document
indian
ssd,
but
that
the
working
group
last
call
would
be
sent
to
the
nsop.
Absolutely
so
that's
that's
definitely
going
to
happen
perfect.
Thank
you.
F
I
can
just
add
a
bit
more
background
on
this.
The
the
lease
time
for
dns
updates
is
something
that
is,
I
believe,
generally
useful.
When
dns
update
was
originally
defined,
it
was.
It
was
sort
of
an
alternative
to
making
the
edits
to
a
zone
file.
F
So,
instead
of
editing
his
own
file
with
vi
and
reloading
the
name
server,
you
could
use
the
ns
update
command,
but
all
the
edits
were
permanent.
There
was
sort
of
an
unstated
assumption.
It
was
being
done
by
a
human
and
what
the
update
lease
does
is.
It
makes
it
more
suitable
for
machine-to-machine
style
communication,
just
like
a
dxep
lease
has
to
be
renewed
or
it
goes
away
automatically.
F
A
Quick
background,
stuart,
that's
helpful,
so
it
sounds.
Let's
see
so
I
have
a
poll
open
if
you
like
either
it
popped
up
somewhere
on
your
screen.
Otherwise,
if
you
click
on
the
show
of
hands
tool
on
the
top
right
on
whether
or
not
you've
people
have
read
this
draft
and
I'm
going
to
end
it.
It
looks
like
we
have
nine
people
who
have
read
the
deaf,
which
is
good,
so
I
would
say
in
terms
of
next
steps.
A
I
think
the
responsibility
right
now
is
on
the
authors
to
address
the
comments
that
were
raised
during
adoption
published
a
new
revision,
mention
it
to
the
working
group
and
then
see
we
can
have
a
discussion
on
there
to
see
if
folks
like
it.
If
there's
not
too
much
contention,
we
can
then
progress
the
document
pretty
quickly
move
on
to
work
group
last
call
which
will
then
trigger
review
from
various
other
working
groups.
So
at
that
point
I
ideally
perhaps
we
have
that
timed.
A
So
we
can
do
a
few
rounds
of
edits
and
then
have
working
with
last
call
like
during
the
next
atf.
So
that
way,
maybe
it
can
be
discussed
at
the
end
of
the
dns
op
meeting
there.
If,
if
the
chairs
over
there
are
interested
all
right,
any
other
thoughts
from
folks
on
the
update
lease
document.
D
Yeah,
so
this
document
hasn't
actually
gotten
a
lot
of
love
in
the
last.
You
know,
since
the
last
ietf
there's
been
a
lot
of
implementation,
work
on
it,
but
but
we
haven't
had
a
chance
to
do
much
to
update
the
document.
D
So
the
idea
is
for
those
who
aren't
aware
the
advertising
proxy
is
a
way
that
we
can
do
srp
updates
and
then
advertise
the
result
of
those
srp
updates
using
mdns
rather
than
using
dns,
and
the
reason
for
doing
that
was
primarily
to
enable
stub
networks
to
advertise
services
on
infrastructure
networks.
So
this
stub
network
being
like
an
iot
network.
That's
just
one
hop
away
and
doesn't
have
any
transit,
but
just
provides
reachability
to
clients
that
have
different.
D
You
know
802.15.4
radios
or
something
like
that,
and
then
another
nice
feature
of
this
that
that
comes
up
is
that
it
makes
it
possible
to
to
sort
of
eliminate
using
mdns
but
still
allow,
that
is
to
say,
eliminate,
allow
a
client
to
stop
using
mdns
to
advertise
stuff,
but
allow
clients
that
do
use
mdns
to
still
discover
the
stuff.
So
those
are
kind
of
the
two
things
the
document
does.
D
It's
been
adopted,
I
haven't
uploaded
it
yet,
as
I
said
it's
in
pretty
good
shape,
but
there
are
some
issues
with
name
conflicts
that
we
haven't
really
addressed,
and
so
in
fact,
you'll
notice.
There's
a
new
document
that
I'm
proposing
the
tsr
document,
which
addresses
some
of
the
name
name
conflict
issues.
The
problem
is
that
our
model
for
for
dealing
with
name
conflicts
in
mdns
is
not
really
oriented
towards
proxies
and
I'll
talk
more
about
that
in
the
tsr
document.
D
So
so
I
think
it
needs
some
more
work
and
you
know
I
I
would
not
encourage
people
to
read
the
document
yet,
because
I
think
that
we
need
to
to
do
an
update
and
then
once
that
update
is
done.
I
really
appreciate
some
reads
and
reviews.
A
Thank
you,
ted.
I'm
gonna.
Similarly,
do
a
poll
just
to
get
a
feel
for
who's.
Read
it
oh
well!
I
guess
I
forgot
to
put
a
title
sorry,
so
this
one
is:
have
you
read
this
document?
I
don't
know
where
that
window
went,
but
you
can
raise
hand
to
say
you've,
read
this
one
or
or
not
raise
well,
while
we're
doing
that.
If
anyone
has
any
questions
comments
on
this
document,
please
join
the
mic
line.
A
All
right
cool,
so
we
have
seven
folks,
who've
read
it
nice
esco
go
ahead.
D
D
H
Yeah,
I
think
the
main
point
was
that
I
think
one
section
really
needs
to
be
simplified
in
some
way
and
maybe
some
other
people
can
also
have
a
look
at
it
see
you
yeah
see
if
that
can
be
done,
because
it
just
looks
to
me
like,
like
it's
almost
too
complex
to
just
parse
when
you
casually
read
it.
H
A
Thank
you
for
that
esko.
So
I
think,
in
terms
of
next
steps
here
ted
you
have
some
edits
that
you
want
to
make
to
the
doc
and
then
we
have
some
comments
from
esco
that
need
to
be
addressed.
A
So
one
actually
item
is
on
you
to
have
those
submit
a
revision
of
the
draft
and
then
let
the
mailing
list
know
so
we
can
have
people
review
that
that
revised
version
yep
okay,
if
unless
we
have
some
something
else
on
this
document,
I
think
we
can
move
on
to
the
next
one.
A
So
these
yeah
we're
transitioning
from
the
working
group
documents.
You
know
current
work
to
individual
drafts
future
work.
The.
A
What
what
I'm
noticing
here,
though,
is
that
a
lot
like
all
the
action
items
appear
to
be
on
uted,
so
maybe
from
a
chair's
perspective,
I
mean
let's
discuss
this
new
work,
that's
good,
but
I
think
we
should
probably
focus
on
the
existing
work
and
getting
that
out
the
door
before
we
add
more
things
to
our
plate.
D
Fair
enough
yeah,
I
will
say
that
that
I
have
gotten
some
offers
so
they're,
one
of
the
other
implementers
of
this
protocol,
who
is
here
in
the
room
today.
I'm
not
gonna
mention
his
name,
but
he
can.
D
He
can
mention
his
name
if
he
wants
has
authored
to
has
offered
to
to
help
out-
and
you
know
with
the
replication
protocol
document,
and
so
I'm
thinking
that
it
would
be
nice
to
add
him
as
a
co-author
and
that
would
potentially
reduce
the
workload
on
me
and
that
would
be
good,
whether
we,
whether
we
adopted
immediately
or
not,
but
but
basically
I
just
want
people
to
be
aware
of
this,
and
and
and
thinking
about
it
and-
and
you
know,
maybe
even
trying
an
implementation
or
or
criticizing
it,
because
you
know
it
well
I'll
I'll,
get
to
the
end
of
the
document
and
then
talk
about
this.
A
I
wanted
to
just
sounds
good
just
just
to
answer
to
that
quickly.
I
think
that's
a
great
idea,
because
I
think
you're
pretty
and
both
steward
and
you
are
bandwidth
limited
these
days,
which
I
totally
understand.
We
all
have
day
jobs,
and
so
adding
a
co-author
before
asking
for
adoption
sounds
like
a
great
idea.
It'll
definitely
give
us
more
confidence
that
the
document
will
be
worked
on.
So
that
sounds
absolutely
great.
D
Thanks
all
right
go
ahead
sure,
so
I
just
wanted
to
give
a
quick
overview
of
what
what's
going
on
here,
because
you
know
there
there's
a
lot
of
stuff
in
fly.
You
know.
Obviously,
dnssd
has
been
working
on
dnssd
generally
for
quite
a
while.
We
have
mdns
we've
had
that
since
before
the
working
group
started
dnssd
on
dns,
we
had.
These
are
all
in
the
base
document,
6762
and
6763..
D
D
So
then,
there's
srp
replication,
srp
replication
addresses
the
problem
that
when
I
started
doing
the
the
srp
work,
I
was
just
assuming
that
we
were
updating
like
by
nine
or
something
like
that.
I
wasn't
really
thinking
about
it.
In
terms
of
of
you
know,
home
networks
and
so
srp
assumes
an
authoritative
server,
but
that
turns
out
not
to
be
a
very
realistic
assumption
on
a
home
network
on
a
home
network,
the
home
router
is
usually
some
little
like.
D
You
know
box
that,
like
they
were
trying
to,
they
were
trying
to
reduce
the
the
the
materials
cost
to
the
lowest
possible
amount,
so
they
put
in
less
memory
less
flash
everything
so
having
that
act
as
a
as
a
resource
for
doing
service
registration
turns
out
not
to
be
it
just
doesn't
work
right
now
that
could
change
we'd
like
that
to
change,
but
that's
the
way
it
is
right
now,
and
so
the
existing
implementations
of
srp
that
that
are
out
there
in
the
world
are
actually
devices
that
are
that
are
srp
plus
advertising
proxy,
and
wouldn't
it
be
nice
to
be
able
to
have
sort
of
the
the
reliability
of
a
managed
dns
server
without
without
a
manager,
and
so
srp
replication
allows
us
allows
a
bunch
of
devices
a
bunch
of
srp
slash
advertising
proxy
servers
to
to
collaborate
in
producing
that
reasonably
stable
namespace
without
there
being
an
operator
so
like
you
could
have,
for
example,
like
a
couple
homepod
minis
each
one
of
those
has
an
srp
server.
D
Each
one
of
them
is
doing
srp
replication
if
any
one
of
them
gets
an
update.
All
of
the
other
ones
get
that
update.
It's
really
nice.
So
that's
kind
of
the
goal
of
srp
replication
is
just
to
like
fill
that
that
that
box,
in
the
block
diagram
with
something
that
doesn't
have
to
be
managed
by
a
network
administrator.
D
And
then
the
other
work
that
I'm
going
to
be
talking
about
permissionless
advertising
of
dns
zones
is
basically
once
you
have
that
stable
dns
server
or
reasonably
stable,
dns
server.
Wouldn't
it
be
nice
to
be
able
to
just
query
it
using
dns
or
dns
push
rather
than
multicast
dns.
You
can
eliminate
some
of
the
dns,
so
so,
basically,
what
this
allows
us
to
do
is
what
the
the
permissionless
advertising
work
is
going
to.
Allow
us
to
do
is
to
allow
dnssd
clients
to
query
srp.
D
D
And
similarly,
if,
if
your
infrastructure
supports,
you
know
delegating
dns
zones,
then
we
have
a
way
to
to
advertise
a
discovery
proxy
or
to
advertise
an
srp
server
into
the
infrastructure
in
a
way
that
that
just
cleanly,
you
know,
works,
and
then
we
can
actually
eliminate
mdns
entirely
and
still
get
all
of
the
functionality
that
we
currently
get
with
mdns
we
get
the
you
know
the
support,
for
you
know
service
discovery
we
get
immediate,
like
we
can
have
an
ongoing
query
and
get
immediate
results,
because
we
have
dns
push,
so
it
basically
gives
us
all
of
the
stuff
we're
used
to
with
mdns
without
any
multicast,
so
back
to
srp
replication.
D
D
D
So
that
means
that
when
it
goes
away
we
have
to,
we
have
to
pick
a
different
primary
and
we
have
to
figure
out
if
the
data
that
that
server
has
is
correct-
and
it
just
becomes
like
this
big
challenging
problem
and
the
biggest
problem
is,
we
don't
actually
have
a
way
to
establish
trust.
D
So
the
only
real
trust
that
we
have
is
that
when
we
get
an
srp
update,
that
update
was
signed
by
the
client,
and
so
we
know
if
we
get
another
update
for
the
same
name,
we
can
tell
if
it
came
from
the
same
client,
and
so
I
kind
of
wanted
to
leverage
that
source
of
truth
for
srp
replication,
and
that
meant
that
I
couldn't
just
we
couldn't
just
have
a
zone
because
the
zone
doesn't
have
the
information
that
we
need
to
to
actually
validate
that
that
a
new
change
to
the
to
the
database
actually
came
from
an
srp
client.
D
It
didn't
just
come
from
a
rogue
srp
replication
server.
So
that's
kind
of
why
I
invented
a
new
protocol,
sorry,
so
the
new
protocol.
The
idea
is
that
there's
an
arbitrary
number
of
srp
replication
peers
which
can
come
and
go.
We
could
limit
the
number
if
we
wanted
to,
and
I
think
that
might
make
sense.
But
I
don't
know
what
the
limit
would
be.
There
is
no
authority
we're
trying
to
maintain
a
common
data
set.
D
All
the
servers
are
communicating,
they're
all
peers,
but
we
do
have
a
precedence
because
we've
got
to
we've
got
to
figure
out
some
way
to
decide
who
talks
to
who
whom
and
you
know
we
have
a
database
sync
and
reconciliation
process,
and
then,
whenever
you
know
we
just
whenever
something
changes,
we
we
replicate
those
updates
to
all
of
the
other
servers.
D
So
so
the
way
this
works
is
that
so
we're
assuming
that
the
srp
client
is
only
going
to
update
one
server.
So
there's
going
to
be
some
advertisements
for
servers,
clients
going
to
pick
a
server,
it's
going
to
advertise
to
that
server.
D
We
can
identify
the
client
using
crypto
because
it
has
a
public
key.
So
we
know
if
we
get
an
update.
We
know
that
it's
from
the
same
client
or
that
it's
not
from
the
same
client
for
any
given
host
record,
and
we
know
when
we
got
the
update.
So
so,
if
a
client
sends
an
update,
it's
either
newer
than
what
we
have
or
it's
older
than
we.
D
But
if
we
received
an
srp
update,
it's
always
going
to
be
newer
than
what
we
have
so
so
essentially,
the
the
client
winds
up
being
the
authoritative
server
for
its
data,
and
we
just
maintain
this
data
set.
We
try
to
keep
the
data
set
in
sync:
it's
not
a
catastrophe.
If
it's
not
in
sync
the
worst
case
scenario,
if
we,
if
we
totally
lose
a
client
update,
is
that
we'll
get
that
information
the
next
time
the
client
renews?
D
D
So
what
we're
replicating
is
you
know
the
the
the
data,
but
but
we
actually
replicate
the
the
srp
updates,
not
the
data.
The
srp
updates
are
then
applied
to
our
local
version
of
the
database
and
that
winds
up
replicating
the
data,
but
because
we're
replicating
the
updates,
we
can
validate
them.
D
So
so
you
guys
can
read
the
slide
again
later,
if
you,
if
you
want
to
get
more
details,
but
that's
the
basic
idea
so
so
the
flow
of
the
protocol
is
that
first,
when
an
srp
server
comes
online,
it
needs
to
figure
out
what
data
set.
It's
replicating
and
that's
somewhat
situationally
dependent
and
there's
some
language
in
the
document
that
talks
about
how
to
do
that.
But
but
it
may
be
implementation
dependent.
D
We
can't
necessarily
specify
it
in
full
detail
or
there
may
be
sort
of
documents
that
specify
like,
if
you're
trying
to
do
srp
replication
in
this
situation.
This
is
what
it
looks
like
so
in
other
words,
for
example,
the
stub
networks
document's
probably
going
to
have
some
text
about
that.
So
we
choose
a
data
set.
We
try
to
find
srp
replication
peers
that
are
replicating
that
data
set
and
if
we
find
them,
then
we
start
we
just
join
in.
D
If
we
don't,
then
we
become
the
first
and
we
start
advertising
and
and
that's
pretty
much
it
so
then
the
next
step
in
the
protocol
is
connecting
and
reconciliation.
So
we
connect
to
a
peer,
and
then
we
compare
our
databases
and
what
we're
looking
for
is
if
the.
D
If
the
peer
has
newer
data
than
we
have,
then
we
want
the
newer
data,
and
if
we
have
newer
data
than
the
peer
has,
then
then
we
want
to
give
that
newer
data
to
the
peer,
and
so
that
means
even
if
we
got
some
updates
before
reconciliation
occurred,
we
can
still
reconcile
them
and
we
we
can
feel
secure
that
it's
actually
going
to
be
correct
because
the
because
we're
we're
passing
around
the
actual
updates,
including
the
time
at
which
the
update
occurred,
which
lets
us
validate
it.
D
So
so
and
the
other
thing
is
it's
possible
that
that
an
srp
replication
peer
might
have
gotten
an
update,
replicated
it
around
and
then
gone
offline,
and
so
during
rep
during
during
the
reconciliation
process.
Those
updates
are
still
going
to
be
part
of
the
reconciliation
process,
even
though
they
weren't
received
by
the
server.
D
That's
that's
offering
them,
because
we
want
to
make
sure
that
we
replicate
the
whole
database,
even
if
the
server
that
originally
got
the
information
is
gone
so
and
then
once
we're
up
to
date,
then,
whenever
appear
gets
an
update,
it
just
sends
it
to
the
other
peers
and
whenever
and
so
from
the
perspective
of
an
individual
peer,
it's
sending
updates
whenever
it
gets
an
update
and
occasionally
it
receives
updates
from
other
peers.
D
We
don't
expect
this
to
happen
a
lot,
but
you
know
happens
from
time
to
time,
and
during
this
period
we
don't
do
secondhand
updates,
because
at
this
point
these
are
live
updates.
So
so,
if
something
goes
wrong
in
this
process,
it's
basically
going
to
have
to
be
resolved
at
the
srp
level,
not
at
the
replication
level.
But
generally
speaking,
it's
you
know.
Usually
stuff
won't
go
wrong
in
the
situation.
D
So
that's
the
basic
protocol,
current
statuses.
We
have
two
implementations,
there's
an
implementation,
if
you,
if
you
have
tvos
15,
which
is
the
current
release
that
has
this
code
running
in
it
slightly
different
version
than
than
what's
documented,
but
you
know
that's
sort
of
par
for
the
course.
D
When
you're
inventing
new
protocols,
open
thread
has
done
an
implementation,
I
haven't
looked
at
the
implementation
yet,
but
I've
had
some
conversations
with
the
implementers
and
it
sounds
like
they're
doing
a
great
job
and
I'm
looking
forward
to
us
doing
some
interop
at
some
point.
So
so,
basically
I
don't
know
if
I
put
it.
No
that's
so.
Basically,
my
goal
is
to
ultimately
get
this
work
adopted
by
the
working
group.
This
is
probably
the
biggest
ask
that
I'm
ever
going
to
make
of
this
working
group,
because
this
is
a
fairly
complicated
spec.
D
D
You
know
david
made
the
good
point
that
we
don't
want
to
overload
the
working
group
given
how
many
people
are
participating,
but
but
this
is
to
me
it
seems
clear.
This
is
the
place
to
do
the
work.
So
basically,
this
is
kind
of
your
fair
warning
that
this
is
there's
I'm
going
to
be
asking
for
adoption
at
some
point.
You
know
if
I
thought
or
if
I
thought
david
wouldn't
laugh
at
me.
I
would
ask
for
adoption
now,
but
he'll
probably
laugh
at
me.
So
so
that's
where
we
are.
A
That
that
sounds
good
from
the
church
perspective.
I'd,
say:
yeah,
let's.
If,
if
you
could
prioritize
going
through
the
action
items
you
have
on
the
adopted
working
group
drafts
and
those
look
like
they're,
no
longer
blocked
on
you
and
making
good
progress,
then
I
think
you
know
then,
and
taking
a
co-author
then
bringing
this
back.
I
think
the
problem
there
should
be
appetite
at
that
point
as
long.
A
Know
as
long
as
my
only
concern
is,
I
want
to
make
sure
that
this
new
work
doesn't
delay
the
existing
work.
But
as
long
as
it
doesn't
do
that,
assuming
there
are
multiple
folks
from
different
affiliations
that
are
interested,
then
we
can
definitely
consider
adopting
this
I'll
double
check
the
charter.
But
I'm
I'm
not
yet
worried
I'll
double
check
that
with
our
id
and
everything.
D
Okay,
cool
all
right,
so,
let's
see,
I
think
the
next
on
the
agenda
is
the
discovering
dns
zones.
D
So
actually
it's
discovering
an
advertising
dns
zone
so
so
kind
of
a
crap
shoot
what
what
lands
in
the
title.
So
this
is
some
work
that
that
I've
been
working
on
with
another
colleague
and
you
know,
there's
actually
an
implementation
of
this
in
the
latest
version
of
tvos
as
well.
D
The
problem
is
that
you
know
so
so
our
our
goal
here
is
to
be
able
to
discover
a
lot
of
services
on
and
these
are,
these
would
be
like
iot
devices
so
like
light
bulbs
and
outlets
and
sensors
and
stuff
like
that,
you
can
have
quite
a
few
of
those
on
a
network.
D
The
sort
of
the
original
model
of
mdns
was,
you
were
going
to
discover
printers
or
something
like
that,
and
usually
you
don't
have
like
a
thousand
printers,
or
I
mean
you
probably
don't
have
a
thousand
light
bulbs,
but
usually
you
don't
have
even
50
printers
you
have
like,
maybe
two
or
three
and
with
with
iot
you
really
could
have
like
50
or
100
devices.
D
D
The
other
problem
is
that
there
are
devices
that
like
to
go
to
sleep
and
for
them
to
to
notice
you
know
to
to
to
to
have
a
complete
set
of
data
about
what's
going
on
in
in
md
s
they
have
to,
they
have
to
be
either
listening
all
the
time
or
they
have
to
like
do
a
point
query,
but
even
then,
there's
no
guarantee
they're
going
to
get
all
the
data,
and
so
so
being
able
to
to
have
a
tcp
connection
that
is
open
and
doing.
D
Dns
push
is
a
lot
cleaner
solution
to
both
of
these
to
both
of
these
issues,
and
so
what
we
propose
to
do.
The
problem
is,
of
course,
what
we'd
like
is
to
be
able
to
just
advertise
the
stuff
to
the
infrastructure.
D
Have
the
infrastructure
pick
it
up,
delegate
the
domain
and
then
have
have
all
the
discovery,
stuff
sort
of
work
and
that's
certainly
our
end
goal,
but
in
the
meantime
we
also
need
to
be
able
to
have
this
stuff
work,
even
if
the
infrastructure
doesn't
support
it,
which
at
the
moment
it
definitely
doesn't
so
so
the
way
to
do
that
is
to
advertise
the
zone
using
mdns
so
essentially,
or
you
know,
and
and
by
extension
dnssd.
D
So
so
we
we
still
using
mdns,
but
we're
only
doing
one
query
instead
of
50..
So
it's
a
big
improvement
and
also
like
you
know,
if
you're,
if
you're,
using
a
device
that
sleeps
a
lot
your
iphone
or
your
watch
or
you
know,
I'm
sure
there
are
equivalent
products
and
other
from
other
companies.
Then
dns
push
is
a
much
it's
much
easier
to
make
that
work.
D
Even
if
you
have
to
re-establish
a
connection
every
time
you
wake
up,
you
can
get
a
complete
list
of
all
of
the
data
that
you
want
to
know
about
very
easily.
But
ideally,
the
connection
just
survives
your
sleeps,
because
you're
sleeping
momentarily
right,
you're,
like
sleeping
for
a
second
at
a
time
and
waking
up
every
second
or
something
like
that,
and
so
so
when
there's
an
update,
you'll
just
get
it
as
soon
as
you
reconnect
and
the
wi-fi
the
wi-fi
base
station
will
probably
have
remembered
it
for
you.
D
So
so
the
way
this
works,
we
have
to
choose
a
name
that
won't
appear
that
it's
never
going
to
appear
in
the
dns.
So
we
actually
need
to
create
a
new,
a
new
name
like
home.rpa,
but
something.
F
Just
add
something
which
I
don't
think
you
actually
said.
The
idea
here
is
that,
instead
of
multicasting
a
query
to
every
device
on
the
network,
the
client
that
wants
to
do
discovery
talks
to
a
server
using
unicast,
which
is
much
more
efficient
on
wi-fi,
and
the
bootstrapping
issue
is:
how
does
it
figure
out
what
entity
on
the
network
is
the
unicast
server?
It
should
be
talking
to.
D
Yeah
or
to
put
it
slightly
differently,
what
entities
on
the
network
are
available
that
can
act
as
that
unicast
server,
because
there
might
be
more
than
one
so
anyway,
so
so
so
the
the
idea
is
we
don't
want
to
create
a
security
vulnerability
here
by
allowing
arbitrary
devices
on
your
network
to
use
mdns
to
override
your
dns
configuration.
D
So
therefore,
we'll
allocate
a
zone,
that's
only
for
this
purpose,
and
so
it'll
just
be
devices
that
are
doing
the
same
thing
competing
with
each
other
and
if
one
of
them
is
malicious,
that's
not
great,
but
at
least
it's
not
going
to
be
able
to,
like
you
know,
do
a
man-in-the-middle
attack
on
your
bank
transaction
or
something
like
that.
So
and
then,
basically
the
idea
is
you
advertise
data.
D
Sets
that
you
have
so
you
might
have
an
srp
database
data
set
if
you're
doing
srp
replication,
you
might
have
one
or
more
discovery
proxy
links,
and
so,
if
you
have
any
of
those
you
just
advertise
them
with
mdns,
the
client
goes
out
and
tries
to
find
the
list
of
zones
that
that
have
data
sets.
It
cares
about,
and
and
then
so,
when
a
when
a
dnssd
query
comes
in
that,
doesn't
specify
what
domain
to
look
in
you.
You
know.
D
Sometimes
the
query
will
specify
looking.local
and
then
you
only
look
with
mdns,
but
but
if
the
query
doesn't
specify
a
particular
domain,
then
then
we're
going
to
set
up
the
browsing
domains
so
that
so
that
this,
these
zones,
that
we've
discovered
are
browsing
domains
and
so
we'll
now
do
dns
push
queries
to
get
answers
in
those
domains,
and
so
and
and
the
idea
is
that
that
any
server
that's
advertising
a
particular
data
set,
should
work,
so
data
set
could
be
an
srp
replication
data
set
where
data
set
could
be
a
discovery
proxy
link
and
in
both
of
those
cases,
srp
replication
gives
you
a
uniform
data
set,
whichever
server
you
talk
to
and
discovery
proxy
gives
you
a
uniform
data
set
because
you're
using
mdns
on
a
particular
link,
and
so
any
mdns
client
on
a
particular
link
is
going
to
see
the
same
information,
so
it
doesn't
matter
which
server
you
connect
to
in
either
of
these
cases,
so
current
status
there's
an
implementation
in
tbos.
D
The
current
implementation
is
not
doing
exactly
what
I
think
we
should
do.
It's
using
ns
records
and
advertising
legacy
browsing
domains
and
it
only
works
with
mdns,
and
my
for
you
know,
after
after
having
played
with
that
for
a
while,
I
came
to
the
conclusion
that
we
probably
wanted
to
just
advertise
this
as
a
service,
so
the
service
will
say
hi,
I'm
a
service
that
that
does
you
know
you
can
use
to
do
to
do.
D
Dns
push-
and
here
here
are
here-
are
the
databases
that
I
replicate
and
or
here
here
are
the
data
sets
that
you
can
query
and,
and
then
just
kind
of
have
the
plumbing
be
done
based
on
that
so,
and
then
also
an
additional
service
that
could
be
advertised
is
discovery
proxy
for
this
link.
D
So
now,
if
again,
if
you
have
a
sleepy
device,
it
can
just
connect
to
the
discovery
proxy
for
this
link
and
get
the
same
information
that
it
would
get
if
it
were
doing
mdns
itself,
except
without
having
to
do
the
having
to
be
awake
all
the
time.
D
So
so
that's
where
we
are
the
request
of
the
working
group
is,
you
know,
please
take
a
look
at
this
comment
on
it
and
again
at
some
point,
I'd
like
to
recommend
this
for
adoption.
So
that's
that.
A
Does
anyone
have
questions
or
comments
on
this
document?
Please
join
the
mic
line
now,
otherwise,
I
think
in
terms
of
conclusion
kind
of
similar
to
the
previous
draft.
A
A
It
sounds
like
you
know,
you're
using
this
for
your
matter
technology
is
that
correct,
or
so
sure,
okay.
Well,
if
this
is
being
used
by
like
for
for
us
who
will
want
to
adopt
this
is
the
implication
would
be
that
it's
a
general
purpose
specification
that.
D
Everyone
so-
and
that's
that's
the
intention-
my
goal
here
is
actually
to
to
so
remember
like
I
was
trying
to
do
the
home
net
naming
architecture
back
in
the
day
and
that
basically
failed
because
we
tried
to
bite
off
too
big
a
chunk
at
once.
So
what
I'm
trying
to
do
is
address
each
of
the
things
that
we
want
the
home
net
naming
architecture
to
do
piecemeal,
so
you
could
say:
well
you
know
it's
a
matter
specific
thing,
but
no
it's
not!
D
The
idea
is
to
to
fill
in
all
of
those
little
building
blocks
so
that
then
we
get
to
the
point
where
we
actually
can
have
what
we
were
trying
to
do
in
the
beginning
with
the
naming
architecture.
So
so
it's
more
that
I'm
using
matter
as
an
excuse
to
to
to
to
succeed
in
this
hobby.
Then
it
is
the
opposite,
but
that.
A
Makes
perfect
sense,
I'm
just
saying
like
for
us
to
you
know,
adopt
this.
The
goal
would
be
to
have
interest
for
various
folks.
If
it's
just
a
protocol
between
an
iphone
and
apple
tv,
then
you
don't
need
a
working
group
document,
obviously
yeah.
That
makes
sense
david.
F
Let
me
answer
your
question.
Actually,
I
want
to
start
off
by
thanking
ted
for
the
enormous
amount
of
work
he's
done,
both
on
coding
and
debugging,
and
writing
the
documents
too.
I
also
want
to
thank
chris
box
for
stepping
up
as
a
new
co-chair.
F
Thread
is
being
used
by
mata,
which
used
to
go
by
the
working
title
project
chip
connected
home
over
ip,
and
that
work
is
done
by
the
csa.
The
connectivity
standards
alliance,
which
also
changed
its
name,
that
used
to
be
the
zigbee
alliance,
and
it
changed
its
name
recently
to
reflect
its
new
focus
on
ip
technologies,
so
mata
will
be
using
thread
as
one
of
its
lower
layers.
F
It's
also
using
wi-fi
people
in
mata
are
looking
at
the
work
done
on
thread
using
unicast
discovery,
to
make
it
more
reliable
and
less
expensive
on
the
network,
and
there's
now
been
discussions
about,
we
should
do
the
same
for
wi-fi.
Wi-Fi
access
points
should
all
have
srp
servers
in
them,
so
I'm
not
pre-announcing
anything
there,
because
this
is
a
long
way
from
being
done,
but
there's
already
discussions
going
on
there
that
mata
would
like
to
see
wi-fi
doing.
F
Slp
just
like
thread
is
so
there's
a
ton
of
stuff,
and
this
is
what
is
keeping
ted
busy
with
with
so
many
groups
interested
in
this,
and-
and
I'm
mentioning
this,
because
all
of
that
is
largely
invisible
in
the
ietf
to
people
who
are
not
involved
with
those
groups,
and
we
are
actively
trying
to
educate
those
people
about
ietf
and
explain
to
them
how
they
can
participate.
A
Thanks
for
the
backup,
stuart
and
thanks
for
like
doing
that,
work,
I
think
having
a
fresh
set
of
new
eyes
and
new
ideas
in
the
working
group
is
always
a
good
thing.
So
that's
what
we're
looking
for
in
terms
of
bringing
new
work,
great,
all
right!
So
ted!
You
have
your
last
document.
We
have
four
minutes
left
and
I
want
to
keep
like
one
for
the
chairs.
So
if
you
could
just
zoom
through
it,
please
go
ahead.
D
Yep,
I
will
do
my
best
so
so.
The
issue
is
that
we
have
a
problem
when
we
have
srp
replication
or
when
we
have
enter
independent
advertising
proxies,
which
is
that
we
can
get
conflicts,
even
though,
what's
actually
happened
is
a
legitimate
update.
D
Mdns
assumes
that
if
something
has
registered
a
service
and
something
else
comes
along
and
tries
to
ever
register
a
service
under
the
same
name,
that
the
something
else
is
the
interloper
and
should
be
rejected
and
the
original
should
stay,
but
with
a
proxy,
the
new
update
is
we're
able
to
validate
with
srp
that
it
came
from
the
same
device.
So
it's
not
an
interloper,
it's
the
same
device,
but
we
still
get
a
conflict,
and
so
the
idea
is
in
the
case
of
proxies
to
have
a
resource
record.
D
That's
mdns
specific
that
just
says
when
we
got
the
update
so
that
if
we
get
so
when
we
send
an
mdns
probe,
we
include
that
information
and
then
an
mdns
server
that
has
that's
doing
the
same
thing
sees
that
probe
and
sees
that
there
is
a
conflict,
but
that
the
new
information
is
that
the
information
being
advertised
is
newer.
D
Then
it's
going
to
not
report
a
conflict
because
the
new
and
the
information
is
newer
and
so
it'll
ultimately
just
drop
the
information,
the
stale
information
that
it
has
and
allow
the
new
information
to
be
advertised
and
with
srp
replication.
Of
course.
Ultimately,
both
of
them
will
wind
up
advertising
the
same
information,
but
this
gives
us
this
just
prevents
all
these
messy
conflicts
that
can
occur
and
the
renamings
that
can
occur
so
so,
as
requested.
I
don't
think,
there's
a
reason
to
go
into
too
much
more
detail
about
that.
D
The
slides
are
up
on
the
on
the
server
and,
of
course,
you
can
read
the
document
which
I
think
explains
the
situation
pretty
well.
So
the
request
to
the
working
group
is
to
take
a
look
at
this.
It's
pretty
simple.
I
think
it's
pretty
necessary
for
actually
for
the
advertising
proxy
work.
So
I
think
we
we
want
to
include
this
in
the
advertising
proxy
work,
but
I
don't
think
this
is
a
really
big
task
for
the
working
group.
D
I
mean
I
wrote
this
up
in,
like
you
know
an
hour
and
I
I'm
sure
that
it
can
use
some
review,
but
I
think
it's
pretty
solid.
So
so
that's
where
we
are-
and
you
have
a
minute.
A
All
right,
thank
you,
ted
we're
gonna,
take
questions
on
this
specific
presentation
to
the
list,
boop
and
then
put
these
back.
There
done
all
right
so
and
I
think
same
kind
of
conclusion
as
the
other
ones
ted.
This
looks
interesting.
It
appears
to
be
in
scope,
especially
if,
in
the
context
of
advertising
proxy,
if
the
working
group
thinks
that
advertising
processes
should
have
a
dependency
on
this
that'll
help
motivate
bring
it
in
but
similar
to
the
previous
ones.
A
We
want
to
see
progress
on
the
other
drafts
and
we
would
love
to
see
some
co-authorship
with
diversity
of
affiliation.
All
right
thanks.
Everyone
for
joining
our
dnssd
session
and
this
great
ietf
week
now
comes
to
an
end.
We're
all
going
to
be
able
to
sleep.
It's
going
to
be
amazing
and
best
parting
words
see
you
all
on
the
list
and
thanks
so
much
barbara
for
everything
we're
going
to
miss
you
here,
and
I
hope
you
still
show
up,
even
if
you're,
not
in
the
at
the
chair
table.