►
From YouTube: IETF113-IABOPEN-20220324-1200
Description
IABOPEN meeting session at IETF113
2022/03/24 1200
https://datatracker.ietf.org/meeting/113/proceedings/
A
B
A
I
you
can
do
it
dude,
it
doesn't
change
like
you
know
exactly
what's
going
on.
C
Okay,
hello,
everybody.
I
think
remote
people
heard
already
half
of
our
conversation
here
at
the
front
desk,
because
the
mic
was
on
the
whole
time.
But
hopefully
there
was
some
entertainment
for
you
welcome
everybody
to
the
iab
open
meeting.
We
have
one
hour
today
and
we
and
my
name
is
mia
coolevent.
C
I'm
the
ib
chair
in
case
you
don't
know
yet
let's
go
to
the
next
slide.
This
is
a
completely
expected
slide.
It's
a
note.
Well,
I
hope
you've
seen
this
already.
I
hope
you
know
what's
on
there,
if
not,
please
make
yourself
familiar
with
that.
C
This
is
the
rules,
how
we
work
together
and
contribute
here
in
the
ietf,
there's
also
a
link
to
the
code
of
conduct,
so
this
is
also
something
that
I
would
like
to
have
everybody
on
their
mind
that
we
work
together
in
a
constructive
and
friendly
way
in
this
meeting
and
in
all
other
meetings.
C
Okay,
what
is
ibo?
Maybe
you
have
an
idea
about
this,
because
you
are
already
in
the
room,
but
our
idea
about
this
is
that
we,
as
the
ib,
really
want
to
have
some
interchange
with
the
community.
We
would
like
to
give
you
a
good
impression
about
what
we're
doing,
especially
from
the
technical
point
of
view,
and
I
just
lost
connectivity
but
you're
from
a
from
a
technical
point
of
view.
C
You
can
stick
with
the
previous
slide,
but
we're
also
looking
for
actually
getting
feedback
from
you
and
having
an
exchange
in
communication
with
the
community.
So
thank
you
for
being
here
next
slide.
This
is
also
mainly
kind
of
for
the
proceedings
just
to
make
sure
beyond
this
session,
how
you
that
you
know
how
to
interact
with
us.
We
have
a
couple
of
mailing
lists.
C
The
easiest
way
to
reach
the
iap
is
just
send
an
email
to
the
iab,
be
aware
that
this
mailing
list
has
a
few
more
people
than
just
the
core
ib.
So
that's
why
this
is
listed
on
here.
You
can
also
send
an
email
to
me
or
any
ib
member
individually,
if
you
want
like
an
open
mailing
list
with
community
discussion,
it's
best
to
go
to
architecture
just
discuss.
This
is
also
hosted
by
the
iab
and
we
use
it
for
various
cases.
C
There's
a
bunch
of
mailing
lists
for
not
a
bunch,
because
we
only
have
two
programs
there's
our
programs
related
meetings
for
most
specific
topics
and
if
you
have
any
kind
of
questions,
feedback
or
other
things
related
to
liaisons,
we
have
an
own
mailing
list
there.
That's
the
easiest
way
to
reach
the
right
people
in
the
iab.
C
Okay,
now
we're
starting
a
little
bit
of
just
like
an
official
reporting
about
what
we're
doing
in
terms
of
documents,
workshops
and
programs,
and
I
will
start
with
this
slide
on
the
the
documents
we
are
working
on,
what
we
have
published
recently.
So
we
did
publish
one
rfc
since
the
last
meeting.
C
That
was
a
lot
of
discussion
about
this
was
in
the
edm
program,
but
we
also
it
was
also
presented
in
this
ib
open
meeting.
At
some
point,
I
think
so,
that's
out
and
published,
we
adopted
a
new
document
in
the
iab
since
the
last
meeting.
That's
the
ib
pass
signals
collaboration,
draft
and
there's
some
discussion
ongoing
on
the
architecture
discussion
list.
So
if
you're
interested
in
that
work,
that's
the
right
mailing
list
and
we're
happy
to
get
your
feedback
on
this.
C
There
is
the
output
of
the
rrc
editor
program,
which
is
ready
to
be
published,
and
I
will
say
more
about
this
in
one
of
the
next
slides
and
there
is
one
document
which
is
under
discussion
by
the
edm
program.
This
is
already
an
iap
document
which
has
been
around
for
more
than
two
years.
I
think,
but
it
hasn't
made
a
lot
of
progress,
but
there's
some
work
on
this
going
on
again.
So
hopefully
we
can
publish
that
at
some
point.
C
There's
one
new
workshop
report-
and
we
will
hear
about
this
next
and
there's
an
update
on
the
workshop
report
from
our
workshop
last
summer
about
network
measurements
or
network
quality
measurements.
I
guess
yeah
program
updates,
so
we
have
two
to
three
technical
programs.
I
would
say
sure:
let's
try
it
this
way.
We
have
two
to
three
technical
programs.
One
of
them
is
edm.
C
C
So,
as
I
just
mentioned,
rfc
9170
is
published,
and
that
was
discussed
and
worked
in
this
program
and
there
was
an
interim
meeting
of
this
program
beginning
of
february
where
they
talked
about
this
related
implementation
text.
So
that's
something
that
is
already
in
the
data
tracker
and
you
can
use
it
in
order
to
make
it
easier
to
find
related
implementations
code
for
indirect
testing
and
these
kind
of
things.
C
E
I
am
thank
you,
so
this
is
a
program
looking
at
the
changes
on
the
treadmills
in
the
the
internet,
we
had
an
entry
meeting
after
the
previous
itf
we've
had
actually
three
documents
being
updated.
Since
last
time,
there's
been
some
discussion
on
mailing
list,
not
a
huge
amount,
but
some,
and
I
think,
we're
converging
on
trying
to
publish
a
couple
documents
on.
We
call
principles
or
documents
describing
principles
around
data
minimization
and
involvement
of
intermediaries.
E
The
work
remains
on
that,
and
also
we
have
a
bigger
question
of.
Do
we
actually
want
to
update
the
official
itf
threat
model,
which
you
know
the
iab
to
begin
with,
cannot
do
and
and
exactly
what
lengths.
So
we
go
to
do
that
and
so
on
that
that
still
remains
to
be
decided.
You
can
participate
in
the
discussion
on
the
main
list.
E
The
link
is
here,
and
we
have
a
poll
for
the
next
meeting
time
up
on
the
on
the
mailing
list,
right
now
so
planning
to
do
that
in
the
next
couple
of
weeks.
That's
it.
C
Thanks,
yari,
okay
and
then
our
third
program-
and
I
cannot
see
any
screen
this
one
yeah
now
I
can
see
it
so
our
third
program.
This
is
a
little
bit
a
different
program,
that's
an
ib
program,
but
it
was
really
run
like
an
open
working
group
for
everybody
to
participate
participate
and
we
left
the
process
on
their
own.
C
We're
just
committed
to
host
that
and
and
take
their
output,
and
the
good
news
is-
and
I
already
announced
in
the
planner
yesterday-
but
just
to
spread
the
news
a
little
bit
more
and
the
program
basically
concluded
their
work.
So
they
finished
the
draft
and
they
headed
back
to
the
iab
for
iab
approval.
The
iab
made
a
four
weeks
community
last
call
in
order
to
reach
out
to
more
of
the
community,
get
different
reviews
in,
and
we
triggered
a
lot
of
reviews
and
and
got
some
final
feedback
made
mostly
editorial
small
changes.
C
So
I'm
very
confident
that
this
is
in
a
good
shape
now
and
we
are
ready
to
publish
that.
Basically,
there
is
a
dependency
from
the
model
itself
to
three
other
drafts
which
are
on
the
ietf
stream
and
which
are
updates
to
existing
process
documents.
C
These
have
have
already
been
on
the
iesg
agenda,
but
then
we
discovered,
like
a
small
update,
that
is
still
needed
to
the
iap
charter.
So
this
update
to
the
ib
charter
goes
through
another
itf
working
group,
not
working
group
last
call
itf,
last
call,
and
as
soon
as
and
and
we'll
also
get
additional
approval
by
isoc
by
the
iso
board,
because
they
approved
the
ib
charter
in
the
first
place.
C
So
we
asked
them
to
take
a
look
at
this
again
just
to
to
check
and
with
everybody
who
could
be
related
here,
but
as
soon
as
that
document
is
ready.
Basically,
the
whole
set
of
documents
is
ready.
C
The
whole
set
of.
Why
is
it
working
yeah?
Okay,
something's
just
changed.
I
don't
know
what
maybe
they
did?
Okay,
so
the
whole
set
of
documents
should
be
ready
and
go
out
very
soon
in
a
few
weeks.
So
at
this
point
it's
merely
a
thank
you
for
ever
to
everybody
who
was
involved,
especially
the
chairs,
the
editor,
but
also
everybody
who
provided
feedback
in
the
mailing
list
was
involved
in
the
discussion
and
provided
reviews.
Thank
you
very
much.
F
I
am
here,
and
I
agreed
to
do
it,
so
I
should
do
it
so,
as
everybody
may
remember,
the
ieb
is
responsible
for
the
liaison
outreach
or
at
least
overseeing
the
laser
on
outreaches
to
a
lot
of
other
standards
organizations.
F
We
created
a
liaison
coordinator
role
in
the
last
year
and
tommy,
and
I
have
been
fulfilling
that
role.
I
will
step
down
and
deborah
will
actually
take
over,
but
this
is
essentially
the
liaison
updates
for
what
has
happened
at
least
recently
since
the
last
ietf.
It
is
worth
noting
that
in
the
future,
we
actually
hope
to
have
sort
of
more
mini
presentations
in
the
iab
open
meeting
or
something
we've
been
discussing
lately.
So,
if
you're
interested
in
hearing
more
about
you
know
a
particular
standards
organization
or
other,
we
may
start
dedicating
some
time.
F
We'd
love
to
love
your
feedback.
On
that
the
multi-stakeholder
platform
on
ict,
standardization
is
up
for
renewal.
We
need.
We
need
to
fill
that
role.
The
the
icann
sec
has
published
113,
which
calls
for
the
icann
board
to
ensure
a
dns
label
is
reserved
for
private
use.
This
has
been
heavily
discussed
in
the
past
in
dns
op
and
in
other
places
as
well.
F
The
the
ietf
has
the
the
oversight
of
special
uses
names
and
there's
a
lack
of
clarity
on
exactly
who
should
be
creating
a
name
that
might
be
used
for
private
label
space
in
the
same
way
that
we
have
private
address
space.
There
exists
nothing
at
the
moment
and
a
lot
of
companies
and
organizations
squat
on
names.
So
look
forward
to
discussions
on
that.
Both
within
I
can
end
the
atf,
the
stuart
did.
You
have
a
question
related
to
liaisons.
G
G
F
Great,
why
don't
you
write
the
ieb
with
a
note
of
of
what
you
think
you
need
and
things
like
that,
and
then
we
will
certainly
reach
out
to
you
so
but
go
ahead.
F
You
know,
information
on
the
their
plans
in
order
to
publish
the
zone
md
record,
which
is
a
record
designed
within
the
ietf
for
providing
a
checksum
on
the
entire
root
zone
content,
so
that
you
can
be
aware
that
you
got
the
entire
file
as
well
as
there's
no
missing
records
in
the
entire
thing
for
large
scale.
Data
transfers
used,
especially
in
things
like
rfc
8806,
which
is
a
local
copy
of
the
root
zone,
for
example,
to
ensure
that
you
have
a
complete
file.
F
It's
used
well
in
conjunction
with
dnsec,
it's
not
a
replacement
for
it
and
then
our
second
58
is
success.
Criteria
for
the
root
server
system,
governance
model,
which,
a
while
ago,
icann
took
on
the
role
of
right
now.
The
root
server
system
is
consists
of
12
organizations
that
don't
have
an
overarching
governance
model
over
them,
and
then
icann
has
been
thinking
for
a
number
of
years
now
about
how
to
create
such
a
thing
and
they
created
the
gwg
ted.
Hardy
is
and
jeff
houston
have
been.
F
Our
iab
liaisons
to
the
gwg
ted,
unfortunately
has
to
step
down.
He
actually
has
been
the
chair
of
the
gwd
and
has
done
an
excellent
job
and
we
certainly
thank
ted
for
his
service.
We
are
looking
for
a
replacement.
F
The
deadline
for
volunteering
for
such
a
role
is
monday,
the
march
sorry
april,
4th
2022.
Obviously,
so
please
do
consider
volunteering
for
that
role.
If
you
have
interest
in
overseeing
and
helping
create
sort
of
the
the
next
root
server
system,
governance
model,
any
questions
barry.
I
This
is
barry
lieber,
I'm
on
sac
and
was
in
the
work
party
that
developed
sac
113..
I
just
wanted
to
clarify
something
because
it's
my
understanding
from
what
we
wrote
that
this
is
not
in
the
ietf's
territory
anymore,
that
the
ietf
clearly
said
you
didn't
want
to
go
there.
So
this
is
just
a
request
to
the
icann
board
to
take
care
of
this.
So
I'm
sure
there'll
still
be
discussions
within
the
ietf,
but
I
don't
think
anybody's
asking
the
ietf
to
take
any
action
at
this
point.
Is
that
your
understanding
as
well.
F
That
is
my
understanding.
That
being
said,
the
icann
board
has
actually
talked
to
the
iab
about
getting
advice
on
how
to
do
this
sort
of
thing.
So
there
still
is
sort
of
some
confusion
in
terms
of
what
the
icann
board
can
do
without
the
ietf's
help.
Even.
J
K
K
The
root
service
system
advisory
committee
in
icann
asked
the
icon
board
to
create
this
working
group
and
asked
it
specifically
to
be
very
inclusive,
because
the
root
server
operators
per
se
did
not
want
to
create
their
own
future.
They
didn't
want
to
be
king
makers
of
themselves.
K
So
therefore,
we
reached
out
to
a
wider
community
and
there
is
a
lot
of
representation
inside
the
icon
community
itself,
but
there
are
also
certain
aspects,
especially
the
technical
ones,
where
we
would
value
the
input
from
the
technical
community
if,
if
even
if,
it's
not
more
than
sitting
there
with
the
your
foot
on
the
brake
pedal
when
we're
heading
in
the
totally
wrong
directions
that
that
would
be
valuable
in
itself.
So
please
consider
applying
for
this
position
and
please
help
us
to
to
maintain
a
relationship
with
the
technical
community.
Thank
you.
F
One
other
point
of
comment,
which
is
that
the
whatever
comes
out
of
the
gwg
will
be
a
public
document
published
by
icann,
when
that
will
mean
that
it
will
go
through
a
comment
period
and,
as
leemon
just
indicated,
the
wider
review
of
such
a
document,
the
better.
So
with
that,
I
believe
I
am
done
with
this
slide.
C
Thank
you.
So
that
is
the
reporting
part
here
and
we
can
start
with
the
fun
part.
We
have
one
more
report,
sorry,
but
maybe
more
detailed
one
about
our
last
workshop
when
analyzing
ietf
data-
and
these
will
take
this
and
then
the
rest
of
the
session.
We
have
some
time
to
discuss
about
centralization.
That
is
a
topic
that
comes
up
very
often
and
there's
a
lot
of
community
interests.
C
L
Yeah
but
jared
is
controlling
them
still.
Oh,
I
can
do
now
too,
never
mind
hello,
all
and
hello
from
sunny
amsterdam,
not
too
far.
I'm
here
to
report
back
on
the
analyzing
ietf
data
workshop
that
we
held
already
in
november
december
last
year.
It
was
hosted
at
the
university
of
amsterdam
with
a
program
committee
of
colin
perkins,
coryn
kath,
mirja,
kulavin,
sinbin
lee
and
wes
hartaker.
L
Our
question
was,
there
is
a
wealth
of
ietf
data,
but
how
can
this
be
used
to
improve
the
ietf
processes,
create
better
standards
that
see
more
implementation,
eight
ietf
leadership,
authors
and
community
members
and
increase
ietf
legitimacy
and
as
a
subsidiary
objective,
how
to
create
a
community
around
the
people
that
use
this
data
create
tooling
and
make
it
accessible
for
more
people,
and
we
did
not
come
up
on
this
on
the
fly
because
also
in
the
ietf
administrative
strategic
plan,
it
says
that
increasingly
data
should
inform
policymaking,
and
so
what
we
did
is
that
we
organized
a
hackathon
which,
unfortunately,
due
to
cove
it
was
not
on
site
but
was
online.
L
We
had
a
four-day
workshop
out,
of
which
two
were
a
hackathon,
also
inspired
on
the
experiences
we
had
at
the
ietf
with
the
hackathon,
but
thus
far
that
wasn't
done
during
an
iab
workshop.
So
on
the
first
day
and
the
last
day
we
had
sessions,
but
on
the
middle
two
days
we
worked
in
distributed
groups
on
different
issues,
of
which
I
will
give
a
short
overview.
L
Very
short,
so,
first
we
gave
an
overview
of
all
the
itf
data
sources
that
there
are,
such
as
the
mailing
list
archives,
the
rfc
index,
the
the
data
tracker
and
robert
sparks
was
there
as
well
also
provided
an
excellent
overview
and
shared
some
notes
that
are
definitely
very
useful
to
review.
If
you
want
to
have
a
look
at
this,
the
ietf
org
web
page
visitor
stats,
github
and
interrupt
testings
and
survey
data,
but
this
could
also
be
combined
with
data
of
other
sdos
to
provide
more
insights
and
then
the
tools
to
analyze
this.
L
There
are
some
really
interesting
tools
that
are
linked
in
the
slides,
the
ietf
data
library,
as
it
has
been
developed
by
colin
perkis,
his
team
and
his
projects.
The
big
bang
project
that's
been
on
this
for
a
while
rfc
prolog
database
and,
of
course,
yadi
arcos,
author
stats
and
ietf
document
statistics.
There
are
a
lot
there's
a
lot
of
data
and
a
lot
of
tools
to
work
with,
but
also
a
lot
of
things
that
we
don't
use
and
don't
know
yet.
So
that
was
also
a
part
of
trying
to
understand
what
do
we
know?
L
Try
to
understand
what
are
the
relations
between
corporate
interest
and
individual
contributions
and
competitions
between
leadership,
and
that
really
flowed
up
super
nicely
in
the
hackathon
on
work
done
on
initial
analysis
on
affiliation
trends
in
mailing
list
participation,
a
lot
of
work
done
on
a
stakeholder
analysis
of
stakeholder
categories
and
groups
and
initial
comparison
between
ietf
and
3gpp
data,
so
that
that
resulted
in
some
new
data
models
in
some
hand-picking
of
of
top
contributors
to
several
mailing
lists
to
to
set
a
bit
of
a
gold
standard,
then
work
by
nick
doty
to
see
if
that
could
also
be
done
odd.
L
In
an
automated
fashion
and
compare
the
hand
coded
one
with
the
automated
coded
one
and
then
some
work
by
sebastian
benthal
to
see
if
we're
from
mailing
list
data,
we
could
actually
retrieve
affiliation
and
then
compare
that
also
with
data
tracker
data
and
the
hand-picked
data.
So
there
was
a
lot
of
a
lot
of
collaboration
going
on
and
working
there,
and
then
we
could
actually
use
that
data
on
3gpp
data.
So
that
was
really-
and
this
was
the
work
done
by
christoph
becker.
L
So
it's
really
interesting
how
this
opened
this
work
on
ietf
data
also
opened
insight
into
other
standards
bodies.
Then
another
session
was
on
community
and
diversity
asking
who
is
in
the
ietf?
Who
is
not
there?
What
is
the?
What
is
the
existing
diversity,
can
trends
and
reasons
for
that
be
found
in
data
and
what
could
be
different
if
diversity
had
been
stronger,
so
any
discussions
and
hacking
we
looked
at
what
could
be
measured
and
what
cannot
be
in
existing
tools
but
also
in
in
current
nlp
approaches.
L
The
studying
of
general
identification
country,
region
participation
and
tenure
and
considerations
of
privacy
and
ethics
and
initial
results
show
different
kinds
of
participation
in
different
groups,
confirmed
something
that
we
of
do.
We
already
know
about
a
lack
of
diversity
in
during
several
meetings
and
then
another
session
looked
at
the
publication
process
and
decision
making,
especially
how
decisions
are
made
at
the
ietf.
How
can
these
decisions
be
studied
and
how
can
these
be
used
to
improve
decision
making
and
also
help
authors
and
leadership,
make
better
choices
and
to
lead
to?
L
How
can
we
make
better
input
and
output
documents
to
understand
who
writes
the
rfcs
and
how
does
this
contribute
to
them?
Passing:
try
to
understand
doing
data
driven
mining
of
homophilic
users,
requirements
of
natural
language
processing
to
understand
the
decision
making
process
a
deployment
measurement
of
course.
So
how
does
the
relationship
between
a
successful
rfc
and
its
deployment
and
measuring
efficient
efficience
of
rfc
development
process?
So
what
are
common
delays,
and
also,
when
does
working
group
last
call
take
place?
L
Also,
there
was
some
some
really
interesting
work
being
done
there
and
what
I
think
are
the
most,
oh
sorry
and
then
some
really
interesting
work
on
environmental
sustainability,
on
which
not
a
lot
of
hacking
has
been
done.
But
some
really
interesting
papers
were
presented
on
actually
the
lack
of
discussion
of
sustainability
in
ietf
drafts,
but
also
the
co2
emissions
of
ietf
meetings
and
how
these
could
or
should
be
taking
into
consideration
with
the
environmental
crisis
right
ahead
of
us.
L
So
some
observations
is
that
there
was
really
well
attended.
The
papers
were
really
good
were
really
a
pleasure
to
read
and
the
participation
was
extremely
diverse:
iab
members,
ietf
leadership,
social
scientists,
data
scientists,
computer
scientists
and
the
mix
of
sessions
and
hackathon
work
really
well,
so
there
were
new
collaborations
code
and
tool
chains
were
tried
and
tested,
and
people
have
are
now
using
each
other's
codes
to
write
papers
and
do
things
and
no
research
questions
have
been
developed
and
I
think
that's
really
nice.
L
So
this
also
provided
for
ourselves
for
a
basis
of
our
research
agenda
for
ietf
data
to
inform
future
research
and
development.
There
are
people
working
in
different
new
collaborations.
L
So
if
you
want
to
have
a
thorough
look
at
who
did
what?
During
the
workshop,
please
have
a
look
at
the
papers,
but
also
at
the
slides
that
were
presented
during
the
sessions,
but
also
the
outcomes
of
the
hackathon
I'll
also
have
slides,
and
these
can
be
found
on
github.
We
have
a
draft
report
there,
it's
a
zero
zero,
but
here
is
the
link
to
the
github,
so
you
can
have
prs
and
comments
etc,
and
there
is
also
a
video
recording
of
the
whole
workshop.
C
M
Good
afternoon,
good
morning
and
good
evening
to
you,
colleagues,
niels
thanks
for
the
update
on
this
workshop,
I
saw
a
lot
of
slides
go
by
that
explained
what
transpired
at
the
workshop
and
what
the
next
steps
were.
I
wonder
if
you
or
other
people
who
participated
could
take
just
a
moment
and
and
provide
maybe
your
top
two
or
three
insights
in
terms
of
what
you
learned
about
how
we
participate
or
what
what
was
confirmed
or
what
what
misconceptions
people
had
that
the
data
wore
out
not
to
be
true.
L
Yes,
but,
unfortunately,
not
with
a
full
answer,
because
people
came
in
with
different
research
interests
and
different
questions,
and
I
think
where
we,
where
we
focused
on,
was
trying
to
understand
the
data
and
the
issues.
People
are
working
on
to
have
an
understanding
of
the
problem
and
understand
how
the
things
could
be
approached.
So
we
were
by
no
means
in
these
days
to
a
full
understanding
of
of
particular
issues,
and
I
wouldn't
want
to
talk
for
other
participants
to
say
what
are
the
conclusions
of
that
workshop.
L
That
was
definitely
not
the
nature
of
the
workshop.
It
was
the
first
time
we
we
did
this,
so
I
think
like
establishing
a
community
establishing
a
research
agenda,
establishing
research
questions,
and
it
would
be
way
too
early
to
say
we
would
have
answers
to
them.
But
what
could
be
said
is
that
some
of
the
code
is
now
being
used
to
try
if
we
could
generate
dashboard
kind
of
features
similar
such
to
the
ones.
L
N
N
So
this
is
about
getting
people
together,
oftentimes
for
the
first
time
and
trying
to
understand
what
everybody
like
has
as
an
interest
and
has
us
data
and
has
his
motivation,
and
I
kind
of
expect
this
will
be
sort
of
an
ongoing
thing
in
some
way,
and
I
don't
know
if
it's
always
going
to
be
an
iab
workshop
or
if
there's
some
other
way
to
continue
this.
But
it's
certainly
not
done
after
after
this
one
event
right.
N
But
it's
it's
really
helpful
to
put
in
the
leadership
cap
on
to
have
actually
data
about
a
bunch
of
stuff,
especially
during
covet,
where
it's
sort
of
very
difficult
to
get
a
sense
from
the
community,
because
there's
no
physical
meeting
where
you
bump
into
people
and
it'll,
tell
you
things
right.
So
we're
trying
to
get
some
some
insight
into
what
we
should
do
in
terms
of
decisions
through
efforts
like
this.
O
Hello,
this
is
shane.
This
is
very
cool.
I
wonder
if,
during
the
work,
if
there
was
a
lot
of
attention
given
to
areas
where
it
was
difficult
to
extract
information
out
of
the
data
that
you
have
like
are
is
is
part
of
the
intention
to
make
recommendations
for
additional
metadata
or
additional
data.
H
L
Excellent
question,
so
what
what
I
think
was
the
preliminary
analysis
is
that
data
tracker
very
nice
mailing
lists
very
hard,
very
messy.
But
what
is
the
advantage
of
mailing
list
is
that
many
other
organizations
also
have
mailing
lists
and
many
other
organizations
don't
have
data
tracker,
so
that
makes
them
comparison
harder.
But
then
we
could
try
to
see
if
we
can
enhance
the
the
the
mailing
list,
data
with
the
data
tracker
data
and
then
see
if
we
can
just
improve
our
whole
thing.
L
There
was
also
talk
and
discussion
and
some
presentations
on
on
knowledge,
graphs
and
how
that
and
how
machine
learning
can
inform
this
process,
but
that
that
was
still
also
also
a
presentation
about
how
many
things
did
not
work
as
well
as
people
hope
so
so
that
continues
to
be
work.
There
is
a
thing,
of
course,
also
in
the
data
tracker
that
many
things
are
text
fields
and
not
drop
down.
So
you
have
many
many
different
ways
of
even
just
spelling
cisco,
cisco
systems.
L
So
with
capitals
you
know
so,
and
so
there
were,
there
were
challenges,
and
what
was
so
nice
is
that
that
robert
and
others
were
there
so
yeah.
It
really
felt
like
we
were
hacking
on
a
common
problem.
So
it
was.
It
didn't
feel,
like
research
community
needed
to
request
leadership,
to
tell
the
tools-
people
no.
It
was
like.
We
were
really
like
I'm
trying
to
understand
and
wrapping
our
head
out
where
we're
going,
what
we're
missing,
what
we
have.
C
Yeah,
so
then
we
actually
we
gained
some
time
but
on
that
presentation,
so
that
is
work
from
isoc.
That
did
some
market
analysis
about
centralization
and,
as
we
don't
have
the
presenters
here,
it's
probably
much
easier
for
you
to
just
go
on
the
isoc
web
page
and
look
at
all
the
data
there
firsthand
it's
very
interesting
and
it
doesn't.
It
does
contribute
to
this
discussion.
C
Okay,
then!
No!
This
is
the
wrong
presentation.
The
next
should
be
from
viet,
who
also
did
some
measurements.
So
let's
start
with
that
part.
C
P
Exactly
cool
thanks,
perfect,
so
yeah,
hello,
everyone,
my
name
is
viet
from
technology
university
of
munich,
and
today
I
have
the
honor
to
present
our
measurement
study
on
web
consolidation.
Here
at
the
iab
open
meeting,
this
paper
is
covered
by
roland
oliver
weipov
and
was
recently
published
in
toit.
P
You
can
find
a
link
to
the
paper
and
all
the
measurement
related
material
on
our
github
repository
as
well
as
contact
information
in
case
you
have
any
questions
so,
as
already
mentioned
before
in
the
recent
years,
you've
probably
also
seen
that
there
has
been
increasing
demand
for
measurement
studies
on
internet
centralization
in
particular.
Why
is
that?
Essentially,
because
there
are
differing
opinions
and
views
about
centralization
on
one
hand,
centralization
obviously
facilitates
the
deployment
of
the
protocols.
P
So
what
we
first
did
in
our
paper
is
we
looked
at
longitudinal
dns
measurements
for
all
domains
in
the
dot,
com.net
and
dot,
org
name
spaces,
which
were
more
than
160
million,
and
we
found
that
the
cdi
penetration
has
increased
from
roughly
eight
percent
to
nearly
double
with
fifteen
percent
in
twenty
twenty
fifteen
percent
in
this
case
means
that
fifteen
percent
of
the
websites
that
we
looked
at
out
of
all
the
measured
ones
were
hosted
on
a
cdi
and
note
that
the
cdi
identification
is
obviously
not
an
exhaustive
identification,
but
it
covers
most
of
the
prominent
providers
that
are
out
there.
P
So
this
can
be
seen
as
a
lower
bound
one
could
say
in
the
next
part.
We
then
looked
at
popular
landing
pages
in
based
on
the
alexa
top
1
million
top
list
and
see
that
the
overall
penetration
is
roughly
24
over
pv4
and
roughly
12
over
ipv6,
with
more
popular
domains,
obviously
having
a
higher
penetration,
since
these
websites
are
simply
facing
more
traffic
and
are
in
more
need
of
this
infrastructure.
P
So
to
speak
in
the
bottom
plot,
you
can
see
the
contribution
of
providers,
in
this
case
only
the
top
providers,
in
this
case
to
the
overall
penetration,
and
you
can
see
that
it's
only
a
handful
of
providers
that
make
up
the
most
amount
of
contribution
to
this
overall
penetration,
in
particular
of
ipv6,
where
we
see
that,
for
instance,
for
the
alexa
top
100
000
for
ipv6
cloudflare
contributes
roughly
80
of
the
pages
of
all
cdi
hosted
pages.
P
So
while
there
is
some
concentration
that
you
can
see
there,
it
definitely
helps
with,
in
this
case,
pushing
ipv6
then
also
went
beyond
landing
pages,
where
we
looked
at
page
resources
for
that
we
used
page
load
measurements
from
the
hp
archive,
which
covered
more
than
four
million
websites
and
roughly
400
million
resources,
and
we
saw
that
roughly
one
third
of
the
pages
were
hosted
on
the
cdi
and
in
contrast,
more
than
50
of
the
resources
that
were
loaded
with
together
with
the
pages
were
actually
coming
from
a
cdi
in
the
top
table.
P
Right
here
you
can
see
that
google
and
amazon
in
particular,
account
for
together
more
than
50
of
the
cdi,
hosted
resources.
So,
while
more
than
half
of
the
resources
in
total
come
from
a
cdi,
more
than
50
quality
come
from
google
and
amazon
in
particular.
So
there's
definitely
some
you
could
say
top
heavy
distribution
there
in
the
bottom
table.
You
can
also
see
the
penetration
by
resource
type,
which
is
especially
high
for
static
resources,
which
is
sort
of
expected,
since
cdns
in
particular,
are
used
for
distribution
of
static
resources.
A
lot.
P
P
This
obviously
also
means
that
there
are
some
risks
with
this,
but
on
the
other
hand,
it
also
helps
with
the
deployment
and
support
of
ipv6
and
tls
1.3,
as
we've
briefly
showed
also
as
media,
as
mentioned
before,
there
was
some
quite
some
discussion
on
the
architectural
discuss
mailing
list,
which
I
definitely
recommend
you
to
have
a
look
into
if
you're
more
interested.
P
In
that
I've
put
some
of
the,
I
guess,
main
questions
and
comments
from
the
discussion
here
on
the
slide,
and
I
guess
the
most
important
or
most
interesting
one
would
be
what
is
the
impact
of
and
on
itf
activities
around.
All
the
centralization-
and
I
believe
mark
has
some
comments
on
this
on
the
next
talk.
So
unless
there
are
any
questions,
I
would
conclude
here.
C
Yeah
we
have
somebody
in
the
cube,
but
I
would
actually
like
to
go
through
the
other
presentations
quickly
and
then
we
take
the
question.
So
christians
just
stay
in
the
queue
and
then
you
can
just
you
can
just
unmute
yourself
when
there's
a
question
coming
up
anytime
later.
Q
Okay,
can
you
can
you
hear
me?
Yes,
okay,
how
do
I
control
these
slides?
Oh,
I
see
interesting,
okay,
weird,
okay.
I
don't
think
I'm
going
to
need
10
minutes
here,
but
we'll
see.
So
this
is
just
a
quick
update
on
a
draft.
I've
been
working
on
for
a
little
while
draft
nottingham,
avoiding
internet
centralization.
Although
the
title
is
centralization
and
internet
standards,
there
we
go
so
the
purpose
and
a
bit
of
the
history
here.
Q
This
was
inspired
by
a
draft
that
brian
trammell
and
I
started
when
we
were
both
on
the
iab
about
choke
points
on
the
internet
before
this
became
quite
such
a
big
topic
and
it
never
really
got
enough
steam
to
actually
be
adopted
by
the
iab.
We
talked
about
it
once
or
twice.
H
Q
But
the
the
topic
stayed
with
me
and
and
as
we
went
on,
I
thought
I
wanted
to
write
more
about
it.
So
here
we
are,
and
I
very
much
want
to
keep
the
focus
here
on
exploring
how
centralization
relates
to
the
internet
standards
community
and
our
work.
There
is
a
lot
more
that
can
be
talked
about
here,
and
I
know
that
it's
it's
something
we
return
to
fairly
often
around
competition
and
economic
factors
and
and
so
forth
and
so
on.
Q
But
we
we
lack
expertise
and
authority
in
those
areas
and
and
in
in
a
lot
of
cases,
and
one
of
the
main
points
of
the
draft
is,
is
that
you
know.
While
we
see
centralization
happening,
we
don't
necessarily
have
the
ability
to
prevent
it
with
the
tools
that
we
have
in
our
toolbox.
But
but
having
said
that,
there
are
still
things
we
can
do,
and
so
that's
kind
of
the
the
area
that
the
giraffe
is
trying
to
stake
out
and
and
and
talk
about.
Q
So
this
is
the
current
table
of
contents.
It
starts
by
trying
to
define
centralization
with
the
american
spelling.
It
tries
to
talk
about
why
it's
not
a
great
thing
to
have
centralized
internet
protocols.
Why
we're
trying
to
avoid
it
and
and
then
it
it
offers
a
kind
of
a
taxonomy
of
centralization
of
different
kinds
that
that
we've
observed
this
is
as
far
as
I'm
aware
original,
I
haven't
seen
anyone
else.
Q
Do
this
and
and
as
I'll
I
think
the
next
slide
says
I
haven't
gotten
a
lot
of
feedback
on
this,
which
means
either
I
got
it
great.
The
first
go
or
people
haven't
really
looked
at
it,
I'm
not
sure
which.
So,
if
folks
have
thoughts
about
that,
I'd
love
to
hear
it
offline,
then
it
talks
about
kind
of
you
know.
Q
There
are
some
common
conceptions
about
how
decentralization
can
be
brought
by
using
certain
techniques,
and
it
talks
through
why
they
are
not
magic
bullets
that
can
be
used
to
solve
centralization
issues
and
then
and
then
it
concludes
with
a
discussion
of
what
we
in
this
community
you
know,
might
think
about,
doing
and
and
and
and
makes
a
couple
of
recommendations
or
at
least
explores
a
couple
of
areas
where
you
know
that
can
are
suggestions
of
how
to
guide
that
discussion.
Q
So
so
far,
all
the
interactions
that
I've
had
a
lot
of
them
have
been
private,
but
it
seems
like
there's
pretty
strong
and
broad
interest
in
the
topic,
and
I've
had
a
fair
amount
of
encouragement
to
pursue
this
work,
which
has
kept
me
revising
it
about
once
a
month
right
now,
I've
made
a
lot
of
tweaks
and
adjustments
throughout
based
upon
the
feedback
I've
received,
and
I
think,
as
I
mentioned,
I
haven't
had
a
lot
of
substantial
suggestions
on
the
cons
of
centralization,
but
a
lot
of
the
discussion
has
been
around.
Q
What
should
internet
standards
do,
and
I
think
that
needs
some
more
development
and-
and
I
realized
when
I
was
putting
these
slides
together,
I
need
to
update
the
acknowledgement
section
as
well
so
next
steps,
my
gut
feeling
is
we're
something
like
75
of
the
way
there.
We
could
spend
a
lot
of
time
working
in
this
area
generally,
but
I'd
like
to
ship
something,
and
so
I
have
some
open
issues.
Q
I
I'm
getting
some
feedback,
so
I
saw
yara
gave
some
feedback,
which
was
quite
interesting
just
about
an
hour
ago,
which
I
think
I'm
gonna
enjoy
talking
through
with
him
and
incorporating
the
document
right
now.
My
plan
is
to
submit
it
to
the
independent
stream
and
I've
been
talking
to
the
ise
about
that
a
little
bit,
but
I
do
want
to
incorporate
more
review.
Q
From
the
community
and
publish
that
relatively
soon,
so
that's
the
the
first
url
is
is
the
the
repo
that
I'm
using
to
develop
this
issues
and
discussion.
There
is
more
than
welcome,
and
so
far
the
public
discussions
happened
on
architecture
discuss
if
the
ib
is
happy
to
continue
that
I'm
happy
to
continue
to
use
that
or
I
can
try
and
get
another
mailing
list
if,
if,
if
the
iab
doesn't
want
me
to
do
that,
and
that's
all
I've
got.
C
E
And
if
you
display
my
one
slides,
there's
only
one
actual
slide.
This
will
be
brief.
So
we
have
some
time
for
discussion
so
very
much
in
the
same
vein
as
mark
was
talking
about
now.
So
this
is
about
the
technical
things
that,
if
you
can
go
to
the
next
slide,
technical
things
that
we
could
do
about
some
aspects
of
centralization-
and
I
don't
want
to
go
into
the
details
of
this.
But
I
wanted
to
highlight
that.
E
E
How
do
you,
you
know,
bundle,
os
or
device
or
browser
and
some
application
system,
some
service
services
together
and
how
you
find
the
different
components
and
how
how
tight
those
bindings
are,
and
there
are
things
you
can
do
about
this
just
to
list
on
the
right
side,
a
few
things
that
the
ietf
has
recently
worked
on.
So
for
data,
for
instance,
oblivious,
dns
or
pre.
Bpm
type
of
solutions
are
one
approach.
E
Of
course
we
can
also
try
and
eliminate
some
of
the
data,
don't
send
as
much
data,
but
there's
other
tools
as
well
and
then
for
these
two
type
bindings.
You
have
discovery
that
discovery
and
modularity
that
you
can
apply,
and
ddr
is
one
good
example
of
that
coming
out
of
the
add
working
group.
So
I
think
that's
it.
For
now
you
can
look
at
the
slide
in
more
detail
and
we
can
go
to
discussion
thanks.
C
R
R
The
the
the
one
thing
I'd
like
to
be
clear
about
is
whether
we
should
measure
cdn
concentration
by
itself
or
as
the
part
of
the
web
delivery
process.
R
Now
that
kind
of
stuff
I
mean
a
big
organization
like
wikipedia,
but
the
same
is
true
for
facebook
or
google
or
microsoft
or
many
others.
R
R
And
so
I
I
don't
know
if
you
want
that,
so
that
was
one
question
and
the
other
question
is
that,
yes,
I
mean
we
have
to
to
go
into
this
publication
access.
Also,
the
the
world
of
platforms.
We
see
the
concentration
happening
in
the
underlying
platforms,
I
mean
the
the
the
place
where
the
hosts
are,
I
mean,
being
dns
providers,
virtual
network
with
virtual
machine
providers,
etc.
R
C
Okay,
thank
you.
Then
we
take
the
next
in
the
queue.
H
Hi
and
thanks
for
the
presentations
prior
to
mark's
draft
mark
mcfadden,
myself
and
elliott,
has
it
have
a
draft?
It's
still
out
there
revision
three
on
centralization
as
well,
so
I
was
thinking
one
way
forward
would
be
to
have
a
reconciliation
or
have
a
look
at
our
both
of
our
drafts
together
and
maybe
see
where
each
of
us
can
fill
in
the
holes
and
that's
just
something
to
think
about
going
forward
I'll
post
the
the
link
here
and
also
on
the
also
on
the
mailing
list.
H
The
other
thing
I
just
wanted
to
bring
up
just
more
generally,
maybe
perhaps
following
on
from
yari's
presentation,
is
that
steven
in
saad
mentioned
consolidation
around
protocols
that
are
going
from
two
to
three
different
protocols
and
that's
something
that
I
don't
think
has
been
looked
at
yet
in
terms
of
standardization
here
at
the
ietf.
So
it's
just
another
thought
for
something
we
can
put
together
on
on
all
these
drafts.
So
thanks
thanks
a
lot.
M
Hi
again,
all
right
a
couple
of
things
now,
first
of
all,
I
think
my
name
has
fallen
off.
That
draft
dominique
I
had
hoped
to
actually
contribute,
but
for
various
reasons,
mostly
involving
the
rfc
update
process
and
a
few
other
things,
I
ran
out
of
time
and
steam
at
the
time
that
that
that
it
was
going.
M
M
This
is
the
there
was
a
workshop
on
centralization
in
the
internet.
I
will
post
the
link
into
the
chat.
It's
probably
good
for
anybody.
Who's
interested
in
this
topic
take
a
look
at
the
presentations
there
and
I'll
just
end
with
while
mark-
and
I
haven't
indeed
been
talking,
regardless
of
whether
we
decide
to
go
forward
in
the
independent
stream.
If
we
do,
I'm
certainly
hoping
that
that
wouldn't
be
the
end
of
the
conversation
but
more
to
spur
additional
conversation,
both
in
the
ietf
and
elsewhere.
M
This
is
a
topic
that
extends
well
beyond
the
etf
and
other
as
ddrc
and
wise
many
other
places
where
policy
implications.
Economics
considerations.
This
is
something
where
I
think
we
should
reach
across
in
order
to
to
form
better
shared
understanding.
Thank
you.
J
Yes,
hi,
I'm
victor
well,
I
just
wanted
to
suggest
that
the
iab
takes
a
look
at
what
I
think
is
one
of
the
key
issues
that
are
open
around
this
centralization
thing.
There
are
many
aspects,
but
the
one
that
I've
seen
coming
up
around
and
around
in
the
end
in
different
forms
is
the
relationship
between
the
applications
and
the
network.
J
If
you
look
at
the
global
space,
hundreds
of
thousands
of
corporates
that
are
providing
network
access
and
200
different
governments
scattered
around
the
world,
and
now
these
functions
have
been
centralized
into
the
hands
of
four
browser
makers.
Two
mobile
operating
system
makers-
and
I
mean
in
terms
of
messaging,
maybe
three
four
or
five
companies
so
globally.
This
is
really
one
of
the
the
reasons
why
we're
seeing
this
scenario
and
we're
getting
more
and
more
of
it
as
soon
as
we
continue
moving
functionalities
from
the
network
to
the
application
level.
J
So
there
should
be
an
itf
position,
idf
idea
of
in
the
end.
Where
do
we
see
in
the
long
term
the
relationship
between
the
applications
and
the
network
to
be
if
any
data
on
connections
and
activities
has
to
be
shared
from
the
applications
with
the
network
operator
or
not
under
which
conditions
and
part
of
this
will
be
a
policy
decision?
So
there
will
be
I'm
sure
there
will
be
regulators
involved
in
these
competition
authorities.
J
There's
so
many
other
aspects
which
are
not
technical,
but
at
least
if
the
fdf
could
come
up
with
ways
to
do
this
communication,
when
it's
necessary
when,
in
the
end,
it
will
be
decided
that
it's
necessary
to
share
information
and
discussions
on
how
to
do
it
well,
rather
than
how
to
do
it
in
a
bad
way.
I
think
this
could
be
a
contribution
to
advance
the
discussion.
Thank
you.
C
Yeah,
as
you
said,
the
the
iab
is
working
on
a
document
here.
I
I
see
this
mostly
as
a
separate
topic,
but
as
he
also
said,
there
is
a
loose
relation
by
by
this
trend
to
move
up
the
stack
and
then
having
less
players,
and
there
are
pros
and
cons
of
that.
C
Okay,
php.
S
Hi,
yes,
I
centralization
is
the
frame
that
we've
received,
but
that
doesn't
necessarily
mean
that
that
is
the
right
frame
through
which
we
should
be
looking
at
this
problem.
I
don't
worry
about
centralization
itself.
I
worry
about
walled
gardens.
S
I
worry
about
switching
costs
if
I
can
switch
from
one
vendor
to
another
without
having
pain
without
having
stickiness,
I
don't
care
about
the
centralization.
So
much.
If
I
can
talk
to
somebody
who
is
using
a
different
messaging
format,
I
don't
mind.
I
don't
worry
so
much
on
the
social
media
side,
I'm
not
that
worried
about
the
publication
side
of
things.
S
I'm
worried
about
the
curation
side
of
things,
because
it
is
the
curation
power
of
facebook
that
was
abused
to
set
an
agenda
that
promoted
a
particular
set
of
politics,
and
so
I
think
that
we've
got
to
be
rather
skeptical
of
the
frame
that
we've
given
and
address
the
real
frames
and
those
are
actually
the
ones
that
we
can
do
something
about.
We
we
don't
do
economics
here,
but
we
do
do
interoperable.
S
C
Yeah,
thank
you.
I
think
this
is
very
much
the
point
of
the
discussion
to
figure
out
what
we
in
this
community
could
be
doing
and
how
we
can
move
forward.
We
have
carl
in
the
queue
who
was
supposed
to
present
a
few
slides
from
isaac,
but
now
he's
here
and
so
go
ahead.
Carl.
T
Yes,
thank
you
very
much
and
first
of
all,
apologies
for
for
joining
late.
I
missed
the
time
for
jumping
in,
but
I
just
wanted
to
flag
that.
We've
been
working
a
little
bit
on
this
at
the
internet
society
for
a
few
years.
T
Thinking
about
consolidation
and
centralization,
and
quite
recently
we
launched
a
what
we
call
a
focus
area
kind
of
a
thematic
view
on
centralization
on
our
pulse
platform,
which
is
basically
trying
to
curate
information
and
data
on
different
thematics
and
one
of
them
being
centralization,
and
I
won't
go
into
much
details
about
more
about
that
background.
More
than
to
say
that
what
we're
doing
there
in
in
the
context
of
you,
know
trying
to
break
down
this
question
of
centralization.
T
So
the
market
share
in
this
context
is
understood
as
the
ratio
of
sites
using
a
provider
to
all
sites
using
that
technology,
and
that
is
a
bit
of
a
mouthful
but
effectively
means
if
you
have
a
sample
in
which
20
of
the
websites
uses
say
cloudflare,
then
cloudflare
is
given
a
20
market
share
and
then
we
use
those
market
shares
to
to
calculate
concentration,
metrics,
two
of
them,
one
of
them
being
the
gini
coefficient
that
you
might
be
familiar
with
from
other
contexts
like
income
inequality
and
one
of
them
being
hhi.
T
T
We
received
some
very
useful
feedback
from
many
people
in
the
ietf
community
already
that
we've
incorporated,
but
we're
definitely
looking
for
more
so
encourage
everyone
to
have
a
look
at
that
that
website
scrutinize
what
we're
doing,
and
we
very
much
welcome
feedback
to
the
pulse
at
icelock.org
address.
Thank
you.
C
Yeah
thanks
carl
glad
that
you
made
it.
I
think
this
is
a
very
valuable
input
to
the
discussion
and
I
think
all
what
we
heard
today
is
just
like
the
start
of
discussion.
We
need
to
have,
and
we
should
have
so
please
go
to
the
architecture,
this
class
mating
list
or
contact
the
ib
directly.
We
are
at
time
but
mark
and
yet
as
you're
still
here,
if
you
want
to
add
something
or
say
something,
we
have
a
few
seconds.
C
No
okay!
So
then
was
great
to
have
you
here.
Thank
you,
everybody
for
contributing
to
this
discussion
again.
This
is
the
start
of
discussion,
not
the
end.
Hopefully-
and
I
wish
you
a
good
remaining
week.