►
From YouTube: IETF101-ALTO-20180319-1550
Description
ALTO meeting session at IETF101
2018/03/19 1550
https://datatracker.ietf.org/meeting/101/proceedings/
A
C
E
B
This
is
the
aisle
to
working
group
if
you're
here
for
also
that's
great,
if
you're
not
you're
in
the
wrong
room,
but
you
are
welcome
to
stay
and
my
coche
BJ
will
go.
Bonnie
could
unfortunately
make
it
to
this
meeting.
He's
online
I
see
him
right
there
so
he's
on
meet
echo,
so
he
will
be
able
to
speak
up
with
her
something
on
me
to
echo
and
he's
definitely
listening
and
my
name
is
and
zero.
B
B
We
have
the
blue
cheese
going
around,
so
please
fill
them
out
or
I
mean
fill
your
name
and
if
you're
here
and
the
other
thing
is,
we
have
to
know
Tyga's,
Sabine
and
Richard
F
candy,
you
greet,
can
somebody
do
a
java,
scrap
anybody,
I
see
Lyle
in
the
last
row
because
he
wants
to
hide
from
trouble
scribing
not
on
that
laptop
yeah.
That's
the
same
here!
Any
anybody
do
we
do.
We
definitely
need
a
Java
scrap
I
mean
we
do
have
me
to
echo
on
fine.
You
know
we
do
we
do
do.
B
B
B
Right
so
I
think
we
should
we
should
have
somebody
can.
Can
this
I
mean
we
have
already
no
checkers,
which
I'm
very
grateful
for,
but
so
with
that
we
can
get
started.
Can
somebody
log
in
to
meet
echo
and
speak
up
if
somebody
says
something
there
on
the
chat
room,
anybody,
nobody,
you
can
do
it
period.
Thank
you.
That's
that
would
be
great.
Okay
with
that
I
think
we
have
the
logistics
ready,
I'm
presenting
here
from
a
Macbook
that
the
secretary
gave
to
the
Secretariat
gave
to
me
so
I'm
very
thankful
for
that.
B
Good,
thank
you
Brian,
so
either
working
group,
let's
get
started
and
I
know.
Most
of
you
know
this,
but
just
for
the
record.
This
is
the
note
well
of
the
ITF.
This
basically
means
that
we
have
certain
rules
here
at
the
ITF,
so
please
I,
hope
you
all
aware
of
this.
If
you're
not,
please
check
the
corresponding
rfcs
as
you
as
you
see
here.
Most
importantly,
if
there's
any
IPR
issue
you're
aware
of
you
should
mention
it
here
and
all
the
other
stuff.
B
So
please
check
this
out
if
you're,
not
I,
assume
that
everybody
is
aware
of
this
policy.
So
what's
the
status
of
the
how-to
working
group?
Well,
the
status
is
very
simple
and
with
the
chairs
who
is
here-
and
we
all
agree
on
one
thing-
we
want
to
first
focus
on
finishing
the
Charter
items
and
that's
our
core
focus
right
now
and
the
agenda
has
been
set
up.
B
Accordingly,
we
have
about
half
a
presentation
on
core
charter
items
and
another
half
your
presentations,
the
letter
M,
the
second
half
of
the
presentations
on
other
interesting
work,
which
we
are
happy
to
discuss
here.
But
again
we
want
to
really
focus
on
finishing
the
Charter
atoms
first
and
so
I'm.
Actually
thank
you
live
for
the
blue
cheese,
blue
sheets,
anyone,
okay!
Maybe
we
okay,
there's
one
more
there
and
one
more
here:
okay,
so
we
we
give
it
a
second
try,
and
so
the
goal
is
definitely
do
to
finish.
B
The
milestones
first
and
I
actually
had
some
discussion
also
today,
with
Richard
on
the
mailing
list,
I
think
where
you
proposed
some
extensions
and
Richard.
Maybe
I
can
say
this
right
here.
We
are
not
against
extending
more
the
current
work,
but
only
if
it's
really
needed
right
and
I
think
you
responded
accordingly.
Maybe
on
some
occasions
we
can
say:
okay,
let's
stop
the
extensions
extensions
of
our
drafts
right
now
and
finish
them
first,
so
here's
the
agenda
and,
like
I
said
we
have
about
half
of
presentations
regarding
working
with
items
and
finishing
the
milestones.
B
The
focus
from
my
perspective
as
chair
and
I've
discussed
them
with
Vijay
is
and
that
I
will
be
asking
on
each
of
these
presentations.
How
close
are
we
to
working
with
last
call
because
we
as
chairs
think
they
are
very,
very
close
and
I
think
there
are
some
dependencies
in
certain
issues,
but
I
think
this.
That
should
be
the
focus
of
our
discussion
and
I
should
also
say
that
we
have
a
very,
very
tight
agenda.
B
So
if
we
I
would
ask
and
I've
said
this
on
the
mailing,
this
also
and
I
would
ask
our
presenters
please
to
stick
to
their
time
slot,
which
is
very,
very
short,
as
you
can
see
in
most
cases
here.
But
most
importantly,
if
we
have
any
discussions
with
respect
to
finishing
working
with
items,
I
would
finish
those
discussions
at
the
expense
of
the
other
documents
at
the
end
of
the
chair
of
this
agenda.
B
Any
okay
I
see
none
so
working
with
progress
and
since
the
last
meeting
last
meeting
was
IGF
track,
we
did
have
a
virtual
interim
in
December
last
year
and
we
have
one
more
RFC
published
Sabinas
year,
any
other
co-authors
of
this
RFC
who
are
here
right
now,
anyway,
congratulations
and
you
are
see
from
the
alto
working
group,
eight
one,
eight,
nine
and
yeah.
What's
the
status
of
our
working
group
items
like
I
said
we
had
an
extensive
phone
call
with
Mara
and
Vijay
and
I
think
a
couple
of
weeks
ago.
B
So
and
one
of
the
documents
I
have
a
slide
on
this
later
on
that
we
think,
is
ready
for
working
plus
called
as
cross
room
and
server
discovery,
and
fortunately,
neither
margin
or
and
Sebastian
could
make
it
and
that
we
had
some
discussions
with
them.
A
new
version
is
ready
and
we
all
think
it's
ready
for
working
last
call
I
talked
about
this,
a
bit
later
cost
calendar
and
we
had
a
working
group
last
call.
B
There
were
no
objections,
so
Vijay
actually
sent
this
on
the
mailing
list,
I
think
two
or
two
or
three
days
ago.
We
will
move
this
ahead
and
because
it
has
been
around
a
long
time
and
working
blast
call
has
raced
no
more
further
objections
so
and
we
will
move
this
ahead.
Performance
metrics
is
a
document
that
is
a
working
with
item,
but
we
have
some
extensions,
extensive
discussions
in
the
past
regarding
certain
things,
I
put
here
a
very
short
note
on
this.
B
We
had
some
about
semantics
feedback,
loops
privacy,
so
what
Mary
I
actually
suggested
and
which
I
thought
this
was
a
very
good
suggestion.
We
requested
a
review
from
the
transport
area.
Expert
team
and
I
have
informed
the
authors
of
this
document
that
we
will
right
now
for
the
feedback
and
I
think
this
is
a
good
thing.
We
get
a
kind
of
a
third
party
view
outside
of
the
other
working
group.
What
is
reasonable
or
not,
because
you
know
sometimes
we
are
too
much
into
our
own
world
and
we
think
everything
anything
is
great.
B
We
have
a
presentation
from
Shan
remotely
today
and
so
we'll
see
there
and
we're
moving
in
this
ahead
and
we're
making
good
progress,
but
I,
don't
person
I,
don't
think,
is
close
to
working
class
call,
but
maybe
Sean
can
present
something
on
this
ssee
and
Richard
you're
presenting
today,
and
so
let's
see
I
will
not
say
much
further,
but
the
question,
of
course,
is
how
close
are
we
to
working
last
call
I?
Think
we're
very
close
and
unified
property
is
an
path
vector.
B
They
are
this
serial
together,
because
I
have
a
dependency,
so
we
want
to
move.
They
had
is
a
banner,
that's,
basically
what
Vijay
and
I
decided
together.
We
have
new
versions
of
the
new
zero
or
three
versions
on
both
of
these
documents.
We
had
some
interesting
discussion
in
the
mailing
list
if
we
want
to
include
Sabine's
work
on
cellular
addresses
and
the
inclusion.
The
conclusion
has
been,
and
I
confirmed
this
on
the
marrying.
B
If
you
haven't
seen
this,
that
we've
got
to
move
ahead,
unified
properties
and
path,
factor
for
now
and
exclude
cell
your
addresses
for
now
just
to
make
progress
and
as
chair
I,
really
support
this
decision,
like
I
said
we
want
to
finish
some
milestones.
We
can
always
then
add
cellular
addresses
as
an
extension
later
on
and
Sabine
is
going
to
present
on
as
Sara
at
resistor
day
and
Don
is
presenting
this
one,
and
so
here
again,
I
think
we're
very
close
to
working
with
us
cause.
B
I
will
focus
the
discussion
on
how
close
are
we
to
working
less
calling
these
documents?
This
is
basically
a
slide
because
margin
and
Sebastian
cannot
present
you
today
on
cross
domain
server
discovery
I've
mentioned
everything
here
in
this
slide
already
we
have
version
from
our
perspective
as
chairs
I'm
speaking
for
the
chairs.
Now
and
from
the
authors,
all
comments
have
been
addressed
and
we
as
chairs
think
this
is
ready
for
working
dusk.
Also,
we
were
started
working
last
call
in
this
document.
B
Soon
after
the
IETF
meeting
any
other
business
any
more
business,
we
have
lots
of
business,
no
I
see
no
objections.
Any
I
asked
about
the
agenda
and
nobody
bashed
it
and
so
I
think
we're
good
to
go.
First
presentation
is
from
Brian
who
is
chairing
the
penology
and
yeah
without
any
further
ado.
What
is
the
panna
G?
What's
the
relation
to
I'll?
Do
Brian
I
think
you
will
talk
about
this,
so
Irish
shut
up
and
let's
see,
if
I
can
find
your
slides.
F
F
So
real
quick,
what
is
path
aware,
networking
why
we
have
a
research
group?
Why
do
we
care
about
this?
So
an
internet
working
architecture
is
path
aware
if
it
explicitly
exposed
information
about
the
pounds
to
the
end
points
that
are
sending
the
traffic
but
our
weekly
connections,
and
it
allows
any
points
to
have
control
over
pounds
over
which
their
traffic
will
be
sent.
You
can't
really
say
that
the
Carn
internet
architecture
is
path
aware,
which
is
why
we
have
a
research
group
on
this.
F
It's
not
you
know
something
that
exists
yet
we're
trying
to
see
if
it
makes
sense
to
bring
into
existence,
but
there
are
a
lot
of
IETF
technologies
that
you
could
be
used.
They
could
be
used
to
sort
of
take
what
we
have
right
now
and
add
path,
awareness
to
it.
That's
why
I
think
this
might
be
interesting
for
Altima
next
slide,
please.
F
F
Expanding
that
back
out
gives
you
other
ways
to
route
traffic
we'd
like
to
experiment
as
well,
with
cooperative
signaling
as
a
question
of
sort
of
path
properties
in
past
elections.
So
how
do
you
tell
you
know
what
you
want
to
the
network
and
how
do
you
get
information
about
what
the
network
can
provide?
And
this
is
where
the
overlap
with
Aalto
comes.
We
have
a
few
questions
that
we're
looking
at
next
slide.
F
Almost
all
of
the
ones
that
are
Aalto
that
are
interesting
to
Aalto
are
on
this.
First
bit,
so
how
do
we
define
and
represent
the
properties
of
paths
right?
So
a
lot
of
sort
of
the
information
that
is
available
from
Aalto
is
about
the
properties
of
something
that
you
can
connect
over.
We
could
we
don't
see
these
as
sort
of
cost
maps.
We
see
these
more
as
properties
of
singular
paths
for
the
network.
How
do
you
get
endpoints
access
to
trustworthy
information
about
the
pad
properties?
F
So
Aalto
has
done
this
in
certain
cases
for
servers
getting
information
about
the
cost
map
out
to
the
server
and
then
how
do
you
have
control
over
path?
Select
that
this
gets
a
little
bit
more
into
things
that
are
sort
of
outside,
so
the
altar
area,
I
think
also
has
done
a
lot
of
work
that
can
help
us
answer
these
top
two
questions.
F
F
F
Goals
yeah
do
the
afternoon
right
after
lunch,
correct
so
next
slide.
This
is
where
we
get
into
the
science
fiction.
So
here
these
are
questions
of
once.
You
have
sort
of
the
primitives
for
getting
information
about
pass
to
the
end
points
and
having
you
points,
Express
the
paths
that
should
be
used
for
certain
traffic.
How
does
that
you
know
how
does
that
interact
with
the
whole
architecture?
F
How
does
this
differ
when
you
talk
about
applying
it
to
tunnel
and
overlay
endpoints,
and
this
is
very
interesting
when
you
get
into
questions
of
how
you
would
actually
build
a
network
like
this,
like
the
way
that
we
build
anything
new?
Is
we
build
within
tunnels
first
and
then
we
hope
that
we're
going
to
replace
the
tunnels
was
something
else
and
then
never
do.
F
H
F
Like
some
larger,
larger
scale,
measurements
would
be
people
who
already
had
larger
scale
deployment.
So
that's
an
ask
Apple
question
I
think,
but
yeah
they're
famously
talkative
when
it
comes
to
real
real
data.
So
you
know
I
think
it's
all.
It's
all
been
largely
sort
of
three
or
four
groups
that
have
done
a
lot
of
work
on
multi
Kostov.
So
far,
they'd
have
done
a
measurement
there,
but.
F
So
yeah
this
is
basically
kind
of
an
ad
I
figured.
This
would
be
useful
to
some
of
the
people
in
this
room
or
interesting
to
some
of
the
people
in
this
room,
especially
with
respect
to
the
top
two
questions
on
the
previous
slide.
I
think
there's
been
a
lot
of
work
in
Aalto
that
could
sort
of
inform
that.
So
please
come
by
on
Tuesday
afternoon,
1:30.
I
F
So
it's
actually
it's
another
controller
right,
so
we
don't
assume
the
fix
this
at
the
pass
right
like
so,
and
that
that's
actually
a
really
good
question
in
terms
of
because
anytime,
you
start
building
a
an
architecture
like
this
or
protocols
that
operate
like
this.
You
now
have
the
endpoints
trying
to
figure
out
what
they're
going
to
route
their
traffic,
and
you
then
have
the
network.
That's
doing
its
own
traffic
and
trying
to
drop
is
competing
controller
problem.
F
F
So
that's
that
is,
we
know,
that's
an
engineering
problem
that
we'll
have
to
address
but
yeah
as
of
right
now
there's
this
the
assumption
is
not,
but
the
pass
will
always
remain
static
because
we
know
it
won't
happen,
but
it's
also
not
that
you're
going
to
put
a
couple
of
magic
bits
on
your
packet
and
it's
going
to
cause
a
new
bgp
advertised
it
to
happen
somewhere.
Just
because
that's
not
realistic.
J
J
B
F
B
G
Sharing
ideas
from
the
penalty,
so
next
we'll
continue
from
the
episode
problem
apt.
So
the
main
updates
from
day
100
interim
meeting
Tuesdays
101
meeting
include
the
following
and
my
chance.
Maybe
they
so
we
type
gender
document
title
from
the
extensible
unified
property
map
to
the
unified
properties
and
most
changes
are
made
made
following
that.
We
remove
the
section-
and
you
know
may,
and
it's
related
on
my
registry
and
totally
from
the
document
more
details
and
discussions
were
being
not
were
beings
in
next
slides.
G
Actually
this
problem
is
raised
and
continually
discussed
in
the
manganese.
I
see
I
believe
most
of
you
may
be
familiar
with
this
one,
and
the
last
is
that
we
and
also
updated
section
7
C
examples,
including
the
followings
that
is,
we
highlight
a
little
bit
ahead:
benefit
of
using
unified
properties
over
using
the
standard
and
point
property
service
and
to
to
present
to
the
purpose
of
the
endpoints,
and
this
section
is
defined
in
7.2,
I,
suppose
yeah.
G
So
since
we
have
already
remove
a
any
domain
from
this
document,
where
do
we
need
to
place
it?
It
is
not
a
specific
problem,
but
a
more
general
I
suppose
so,
because
it's
not
always
in
domain.
We
have
called
a
few
new
emergent
domains
like
sit
on
CDN
of
CI
footprints,
defining
CDN
I
and
the
in
situ
a
CGI
domain,
I
suppose
of
defining
the
celluloid
a
trisys.
So
two
choices.
G
Actually
we
prefer
to
put
them
in
the
in
their
own
documents
for
the
following
reasons
that
we
need
to
make
this
document
a
minimal
framework
and
the
reasons
that
we
need
to
approach
define
such
as
a
PID
domain?
Are
people
thought?
Oh
man
and
ipv6
domain
in
this
document
is
that
they
are
actually
quietly
and
elegantly
defined.
However,
others
might
not
let
us
do
in
development,
so
they
might
change
and
not
so
stable.
G
B
B
I
mean
we
have
this
and
many
working
groups,
these
discussions,
dependencies
and
and
things
and
there's
always
trade-offs,
but
I.
Think
in
this
case
you
raise
some
good
arguments,
but
also
from
the
chair
perspective
that
we
want
to
move
progress,
make
progress,
I.
Think
it's
a
better
choice
here.
To
use
this
one,
a
hat
and
the
the
documents
will,
then
the
other
proposals
were
later
on.
You
know
at
their
stuff
in
their
drafts,
so.
G
Another
discussion
problem
is
that
and
what's
the
relationship
between
the
auto
domain
registry
ends
Oh
to
address
tab
registry,
as
we
see
that
as
a
concept
of
Z
domain,
more
general,
Z
and
Z
address
types,
for
example,
and
address
tabs
can
be
up
to
be
for
ipv6,
but
we
can
have
an
ipv4
domain
and
ipv6
domain,
so
we
includes
them.
Besides,
we
have
a
domain
D
CGI
domain
PID
domain,
so
it
is
a
more
generic
idea
and
Zanjeer
or
to
address
tap
registry.
G
But
since
you
have
some
relationships
so
forth,
and
so
we
needed
to
study
and
what
attributes
are
to
be
defined
in
the
Auto
domain
registry
and
auto,
address
tap
registry
and
to
do
is
the
consistent
issue
between
these
two
registries.
We
have
two
possible
solutions
wrong
if,
let's
see
will
be
merged
automatically.
That
means
when
you
define
a
new
domain
and
if
it
applies
to
the
address
tab,
so
you
may
need
to
to
find
all
the
attributes
that
is
not
only
related
to
the
domain,
but
also
related
to
the
address
type.
G
So
when
you
put
this
domain
in
Z
domain
registry,
C
address
typewriter
story
will
be
automatically
automatically
generated.
Another
approach
is
that
we
will
need
to
do.
Is
it
manually?
That
is,
the
dependency
will
not
be
automatically
generated,
so
you
have
to
do
maybe
twice
you
have
to
one
the
first.
B
Yeah,
so
before
we
go
on,
I
mean
we
should
have
a
small
discussion
on
this.
I
have
a
question.
First,
there
technical
questions
so
auto
consistency.
If
you
choose
that
option-
and
you
have
somebody
who
has
you
know
a
domain
that
doesn't
really
match
the
IP
addresses
so
there's
a
there
is
no
consistency.
Actually,
you
could
always
create
a
new
dummy,
a
domain
for
whatever
IP
addresses.
You
have
right.
I
B
I
The
current
design,
so
the
congas
on
is
we
want
every
single
address
to
be
a
domain.
We
believe
that's
correct,
so
that
every
single
girls
type
would
be
domain.
Okay,
only
promise
the
other
way
around.
For
example,
a
PID
is
another
way
to
address
type.
It's
like,
for
example,
or
abstract
network
elements,
it's
a
noun
or
they
can
be
circled
as
you
attack
by
there
may
not
to
be
really
like
a
traditional
way
of
every
time.
I.
B
I
Also
trying
to
really
go
down
a
path
of
second
of
a
manual
consistency,
but
we're
not
well
we're
wondering
if
there's
any
a
like
a
confused
from
IETF
working
group
and
EFI.
Yes,
you
two
really
guided
the
process.
You
have
a
one
which
stops
another
one
mouth
or
Union
register
for
both
and
what's
the
process,
do
you
have
like
an
automatic
electrician
in
one
and
trigger
which
we
another
one
I?
Guess
that's
really
the
question
for
ie
and
yes.
B
We
actually
I
guess
I
can
I
also
cannot
answer
at
this
time,
and
the
question
is:
how
do
we
move
forward?
Do?
Do
we
just
choose
manual,
consistency,
move
this
ahead
and
see
of
what
kind
of
people
we
get
from
the
isg
I
mean
that's
a
good
idea.
I
mean
that
works
for
me.
If
you
are,
if
that's
what
they
also
prefer,
I,
don't
see
anybody
in
the
room,
raising
the
strong
opinions
and
in
the
interest
of
moving
things
ahead,
I'm
just
saying,
and
why
don't
we?
I
I
G
H
B
Saying
you
need
to
move
this
forward,
we
can
either
just
do
it
and
get
the
feedback
that
way
or
you
know,
seek
the
feet
where
seek
the
feet.
Feedback
ahead
of
time
and
I.
Think
Vijay
is
proposing
that
that's
fine
with
me,
I
think
in
the
interest
of
moving
ahead.
We
have
a
lot
of
presentations,
let's
move
ahead,
but
this
is
an
issue
we
need
to
clarify
with
Jana
and
it
is
how
we
are
what's
possible,
what
they
suggest.
Yes,
there.
B
H
J
G
B
I
think
you're
done.
Thank
you
for
as
chair
from
my
perspective,
you
make
good
promise.
You
updated
the
document
addressed
so
comments.
We
just
have
one
main
outstanding
issue
right
now,
the
one
we
just
discussed.
So,
let's
try
and
the
following
weeks
to
to
clarify
this
I
mean
we
should
have
mailing
lists
as
custom
of
each
a
may
be
involved
or
Samaria
and
and
Jana,
and
this
see
what
what
is
their
advice
and
then
we
can
move
as
soon
as
we
have
a
decision.
I
have
some
advice
from
guys
who
know
better.
B
I
B
B
G
Updates
are
the
following,
sir
fine
so,
and
the
first
point
is
that
we,
it
might
be.
We
made
a
lot
of
text
edits
and
something
like
clarification
of
the
use
case
because
I'm,
who
you
may
call
auras
document
and
your
way
of
finding
such
differences.
So
another
much
made
remark
is
that
and
the
structure
of
the
whole
document
is
all
mostly
really
organized.
We
made
clear
three
building
blocks
in
the
design
or
review
section.
G
G
So,
following
this
series
during
the
building
blocks,
we
reorganize
the
document
with
one
section
for
each
and
design.
We
have
introduced
a
new
co-star,
which
is
already
and
quite
mature
and
discussed
in
last
meeting,
and
we
have
defined
a
new
domain
that
is
in
domain.
As
you
know,
it
is
moved
from
the
unified
protease
and
we
have
to
talk
about
the
comm
system,
the
same
mechanisms
that
we
used
and
to
encourage
the
prospector
and
its
related
properties.
G
And
domain,
as
you
okay
answered
for
the
section
six,
it's
specified
with
in
with
a
domain
name,
what
is
a
domain
specific
addresses
and
that
does
it
have
any
inherent
hierarchy
or
inheritance
or
and
related
to
this
one?
You
can
find
in
domain
registration.
It
is
registered
in
the
auto
domain
registry
and
the
specified
in
the
section
11.3.
G
The
most
important
one
is
to
do
with
the
consistency
mechanism
and,
if
you're
looking
for
the
relationship
and
maybe
some
similarity
between
the
auto
customers
and
property
maps.
What's
the
cost
of
the
author
is
the
relationship
between
the
cost
map
and
the
property
map?
You
may
refer
to
the
relationship
of
the
network
map
and
the
cost
map,
but
they
did
have
some
differences
and
which
is
a
major
one
and
actually
for
the
relationship
between
the
customer
and
the
network.
G
Map
and
customers
will
depend
on
the
network
map,
but
network
map
is
independent
of
the
customer,
so
it
is
like
unique
directional
dependency.
However,
for
path
vector
a
and
protein
maps,
it
looks
more
like
a
parrot
directional
dependencies
because
in
in
network
Emmons,
an
impasse
vector
their
proteins
will
depend
on
the
unified
properties.
However,
on
the
generation
of
the
Proteus
in
of
the
alias,
were
defined
as
unsecured
a
of
the
cosmic.
So
it
has
a
much
stronger
coupling
issue
and
we
actually
really
look
for
the
multi-part
relate
multi-part
service
is
to
do.
G
Is
this
sure,
but
also
actually
encounter
some
problems?
Yeah?
If
you
want
to
use
to
see
these
two
services
separately,
you
may
need
to
put
in
Z,
but
you
made
me
to
get
a
response
of
Z
and
1
cos
map
and
the
same
use.
This
responses.
There
is
the
returned
and
is
to
put
it
in
the
curate
input
with
a
property
maps.
So
how
to
do?
Is
it
relationship
or
how
to
deal
with
such
operations
is
a
new
issue
to
be
a
tailed
and
we
are
seeking
for
some
solutions.
I
So
the
question
is,
you
say,
you're
looking
for
solutions,
you
have
clarify,
don't
you
were
looking
for
solutions?
We
have
solutions
which
are
looking
for
decisions
right
problems
back
to
work.
Okay,
so
you
have
a
preferred
solution.
Again,
oh
I
will
send
out
email
right
before
we
want
people
to
think
a
little
bit.
I'll
give
you
hard.
I
Hear
would
have
any
strong
feeling
about
which
design
so,
basically
is
what
I
call
absorption
embedding
and
versus
independent
and
embedded,
which
one
people
prefer.
So
the
later
active
air
found
them
as
I
Central
program,
language
of
problems
as
well.
So
why
it's
the
independent
of
what
non-independent
so,
which
one
do
we
prefer,
as
the
ietf
I
think
right
now,
the
current
design
is
a.
We
go
down
the
path
of
independent
in
a
sense
that
were
you
in
the
multi
part.
I
Basically,
we
allow
objects
to
leave
alone
by
himself
we're
composing
the
multi
part
of
compose
the
second
one
problem.
The
poor
people
are
taking
logging
in
long
I'm,
young
language,
and
they
essentially
now
you
take
this
line
type.
You
take
your
type
anymore,
you
can
your
type
and
we
slide
it
down
like
that,
because
I
would
break
our
protocol,
but
we
can
go
down
at
a
pass,
so
we're
just
taking
the
advice
from
from
a
working
group.
Okay,
anybody
want.
B
I
G
B
G
B
We
others
thank
you
John
and
yeah.
We
need
to
see
if
we
can
resolve
this
issue
as
soon
as
possible,
but
of
course
you
just
raised
it
right
in
a
pretty
close
to
the
meaning.
We
should
have
some
discussion
on
your
mailing
list
yeah,
but
from
the
share
perspective
again,
I
think
this
is
a
very
close
to
working
last
call.
We
just
have
one
main
issue
that
we
need
to
resolve.
So
that's
good.
B
I
Right,
yeah,
okay,
let
me
start
some
Gretchen
King
SSI
increment
updates,
and
we
think
this
is
very
cool
document
and
right
now
it's
getting
almost
ready
and
they're.
Only
like
a
while
two
very
small
issues.
We
won't
get
feedback
and
I
think
we
can
wrap
up.
So
what
we
have
done
and
is
reviewing
the
interim
meeting
and
this
ITF
is
we
made
design
it's
very
stable.
I
When
we've
read
the
talking
very
carefully,
we
found
a
bunch
of
text
which
you
can
revise
forever,
I
guess
and
here's
what
we
did
and
we
did
a
lot
of
to
achieve
consistency
facing
with
all
text,
editing,
editing
and
I
think
the
part
is
in
terms,
so
we
could
use
a
new
term
section
and
we
think
you
can
truly.
Maybe
the
IESE
reveal
the
cost
issues.
Therefore,
we
introduce
a
term
section
which,
and
all
the
exams
also
use
a
single
example.
We
just
have
like
two
or
three
examples
in
the
document.
I
We
said
why
don't
we
have
two
example:
two
three
examples:
why
don't
switch
to
a
single
example?
So
therefore,
now
we
are
using
a
single
example
and
go
across
for
document,
so
that
I
think
that's
a
very
nice,
that's
very
clean
and
also
between
a
second
structure.
They
say
that
we
have
Jason
hack
and
we
have
more
to
patch
their
sections
structure
slightly
different
and
we
do
not
like
it
and
we
trim
them
to
be
consistent.
So,
therefore,
those
are
the
major
changes
and
then
what
we
did
was
will
clarify.
I
The
data
are
the
choices
you
can
add
a
major
new
capabilities
in
a
document.
Therefore,
we
want
to
clarify
this
one
as
much
as
we
can,
and
then
we
clarify
our
requirement.
So
we
did
have
a
few
requirement
up.
The
words
if
we
start
me
to
go
over
like
a
mast
should
and
really
clarify
how
to
start
very
nicely
and
without
you
with
some
some
examples.
Next
slide.
Please.
So
here
is
for
people
who
do
not
read
a
document
or,
for
example,
you
read
it
and
we
clarify
the
words
of
lots.
I
Never
let
me
give
you
some
like
really
high
level
overview
and
now
the
call
the
whole
thing
is.
The
increment
updatable
consists
of
set
of
added
streams,
and
if
each
upstream
would
have
set
of
these
are
big
messages
and
set
up
a
country
of
your
messages,
obviously
update
and
then
working
with
it,
which
is
a
major
change,
is
now
we
call
it
an
automatic
data
of
the
message.
It
can
be
a
full
replacement
of
your
message
or
can
be
an
incremental
change
of
the
message
we
use.
I
We
just
say
incremental,
update
message
now
with
the
update.
It
really
means
a
bigger
things,
because
a
full
replacement
is
also
update
and
now
increment,
our
training
also
obvious
sort
of
a
wild
GPS
metric
as
custom.
How
you
call
it
sort
of
or
between
everything
increment
are
between
come
in
a
change.
If
you
don't
like
it
to
us
Valhalla
where
they
sent
a
wallop,
she
became
asymmetry
severe
for
now,
Alex
much
cleaner.
You
don't
have
an
ambiguity,
so
that's
one
meter
change.
This
is
not
to
interrupt
that.
I.
B
I
Then
the
second
thing
is,
we
talked
about
how
to
make
choices
and
turns
are
color
making
choices.
You
look
document.
We
only
said
okay,
it's
the
server's
responsibility
to
decide
how
you
push
the
increment
updates
and
in
terms
of
accuracy,
most
startups.
Therefore,
we
want
to
essential
bring
up
all
the
start
with
a
little
more
in
the
documents.
Therefore,
to
help
for
the
designer
of
the
server
and,
for
example,
here
is
I,
don't
know
if
you
can
read
a
column,
that's
a
newer
version.
I
realized,
ok,
ok,
new
subjects,
for
example.
What's
not?
I
Okay
is
if
you're
using
this
on
one
catch
and
if
you
put
a
new
value
to
b
now
actually
I
use
this
attribute,
we've
removed,
that's
really
unfortunate
or
your
phone,
whatever
depend
on
your
polynomial,
but
at
least
it
means,
if
you
want
a
change,
for
example,
and
actually
better
to
be
now
instead
of
undefined,
actually
Jason.
So
for
we
make
it
very
clear,
so
we
miss
ok,
the
server
should
consider
the
limitations
of
your
choices.
Please,
some
hatch
motive
have
I
think
that
actually
will
be
of
interest
to
generate
EF.
I
But
that's
the
thing
we
made
a
very
clear
next
slide.
Please,
and
also
we
are
there
about
the
data
update
or
CSI.
Now
we
will
give
a
little
bit
more
guidance
about
the
human
truth
between
full
replacement
versus
increment
increment
changes
and
if
I
mean
changes
which
one
should
you
use,
which
is
a
much
pad
or
just
some
patch
we
give
all
kinda,
like
a
guidance,
I
think
will
be
very
helpful
and
I
think
we
all
clarify
a
lot
in
terms
of
document.
Great
and
one
more
thing
is
in
Clark.
I
We
start
because,
even
with
your
from
personal
experience,
when
you
go
to
is
mass
come
in
so
we
start
clarifying
all
kind
of
requirements
and,
for
example,
we
use
it.
We
add
obviously
masses
and
mass,
and
we
use
all
the
capital
and
also
one
more
thing
really
requirement
is
about
a
future
alpha
service,
we're
on
this
increment
late
to
be
reusable
or
cross
as
some
of
us
future
as
possible
future
as
possible.
But
unfortunately
there's
a
major
issue
we
realize
is
for
increment.
I
I
If
you
want
to
have
the
increment
updates
for
a
system-
and
you
only
have
to
essentially
include
your
whole
thing
at
a
single
object,
otherwise
do
some
patch
much
had
you
not
will
not
even
apply
so
I
think
that's
a
decision
we
made
and
we're
not
going
to
make
any
major
change.
It
was
okay,
let's
go
the
assumption
and
maybe
in
the
future,
people
can
come
with
a
newer
pack.
I
Algorithms,
for
example,
can
apply
to
multiple
parallel
table
level
objects,
but
we're
not
going
down
this
path
for
progress
we
decide
to
go
down
is
simpler
approach.
You
have
my
full
support.
Ok,
so
that's
all!
And
now
let
me
just
really
if
you
follow
last
final
touch
point
for
this
design
and
this
pattern,
it's
only
up
example
for
this
increment
up,
it
actually
a
generic
crime.
I
So
let
me
get
you
a
little
bit
and
let's
get
feedback
and
get
opinions
from
experts
office
working
group,
the
prongs
falling
were
using
HTTP
an
HTTP
is
sensually
a
or
p4
2.0,
it's
unit,
Direction
protocol
and
a
client
study
requests
and
a
server
will
give
you
a
response.
So
therefore,
in
our
increment
operator
stream,
and
if
you
want
send
a
request,
form
control
forum,
add
a
new
stream
and
we
have
the
other
separate
channel
because
there's
no
way
of
problems,
you
know
same
HTTP
to
send
a
request
to
the
server.
I
So,
therefore
we
send
a
request
to
the
server
interception
channel
and
now
the
problem
is:
how
can
a
server
will
tell
of
a
clan
that
this
y
is
success?
For
example,
you
won't
remove
a
street.
You
can
send
a
request
from
the
client
to
the
server
and
success.
Give
you
four
give
you
200
successful,
but
you
don't
know
if
this
request
has
been
processed
or
not
because
that's
a
separate
channel
and
often
time
the
program
even
different
threads
or
even
different
processes.
How
do
we
handle
this
issue?
I
So
therefore,
currently
a
certain
hour
to
wish
notify
success?
One
is
notified
from
a
client
from
the
control
channel
and
second,
one
is
also
in
the
email
is
a
stream.
You
know
David
chan,
two
times
multiple
channels
and
it's
essentially,
we
want
to
have
very
nice
thing.
Prom
he
electrons
foliar
protocol
and
there
are
two
designs
and
of
course
we
can
also
have.
Third,
you
know:
option
which
company
a
cargo
craft
actually
essentially
is
using
a
combination
versions.
I
I
If
you
want
go
down
only
one
and
several
proud
will
only
see
if
they're
like
typical
Amazon,
a
type
of
service
is
something
we
present
Amazon
service,
a-ok
in
progress,
and
then
you
probably
responsive
other
way
and
so
on,
and
it's
a
moment
where
you
through
three
to
notify
those
channels,
control
channel
and
centrally
the
sin
of
thing
and
you
know
their
channel.
So
let's
come
Pisa
feedback,
yeah.
K
I
B
I
B
I
assume
you
and
your
quarters
will
take
care
of
those
requests
guidance
from
from
HTTP
in
the
next
couple
days,
all
right,
yeah,
it's
a
good
idea
to.
Maybe
even
do
this
on
side
here,
just
face-to-face
right,
correct,
okay
and
so
again.
Thank
you
very
much
for
the
presentation
again.
I
think
we
as
chairs
are
correct,
that
this
is
very
close
to
working
with
last
call,
and
you
agree,
I
think
yes,
I,
agree.
I
think
this
is
a
very
well
polished
already
right,
so
we
as
chairs
agree.
B
So
actually
you,
if
I,
if
I
go
back,
you
just
presented
a
clarifications
and
updates.
So
this
is
the
only
issue
if
you,
if
you
will
so
I,
really
encourage.
You
may
be
used
to
time
here
at
the
meeting
to
refer
this
with
some
HTTP
experts,
and
then
we
can
really
move
this
ahead.
Yes,
yes,
good
good
suggestion
right
so
I'm
looking
at
the
agenda
so
now
is
Shawn
presenting
and
I
think
he
wants
to
present
remotely
I
hope
this
works.
B
L
B
L
L
M
Okay,
go
ahead:
okay,
hello,
everyone,
I'm
shanley,
and
today
I'm
going
to
talk
about
the
current
status
of
all
those
idiots.
I
met,
trust
and
then,
let's
start
from
a
summary
of
the
changes.
The
first
one
is
we
renamed
cDNA,
FCI,
auto
service
to
the
CDN
IFC
I
met
the
second
one.
Is
we
introduced
a
precise
definition,
follow
CDN
IFC
I
met
the
third
one.
Is
we
refine
the
future
on
capabilities
and
the
first
one
is
with
fine
filter
on
footprints
and
we
made
extensive
refinements
to
the
examples
and
also
we
add
a
lot.
M
We
add
the
schema
of
the
cDNA
FCI
map
response
and
we
make
the
definition
structure
of
it
consists
with
other
already
defined
maps,
and
there
are
two
nodes.
One
is
that
the
definition
structure
is
the
same
as
the
other
one,
and
we
need
to
make
clear
let
the
cDNA
FCI
map
data
definition
is
defined
by
cDNA.
Not
Auto
has
its
normal
teeth
next
slide.
Please,
and
we
clarify
it's.
The
schematic
of
the
filtered
City
AFC
I
met.
It
is
given
some
capabilities
which
foot
cream
have
at
least
one
of
least
capabilities.
So
follow
query.
M
In
the
left
hand,
side
only
capabilities
with
footprint
constraint
in
the
red
in
the
red
rectangle
will
be
returned
in
the
right
hand,
side
because
they
all
have
the
HTTP
1.1
delivery
protocol
capabilities
next
slide.
Please-
and
next
we
consider
the
arrow
handling
of
this
query.
Let
each
missing
field
arrow
will
cover
the
cases
only
capability
type
or
only
capability
value
or
the
value
of
capability
value
is
now
and
the
e
invalid
field
value
we
all
have.
Our
value
of
capability
value
is
inconsistent
with
capability
type,
and
there
are
two
special
cases.
M
Special
cases
here
means
that
these
cases
are
not
arrows.
The
specialist
one
is
that
if
cDNA
FCI
capability
is
now
all
the
value
of
it
has
no
elements
the
author
server,
we
are
returning
related
full
cDNA,
FCI
map
and
the
second
one
is
if
the
capability
type
is
not
defined
the
author
server.
We
are
ignore
this
capability,
while
return
remaining
without
next
slide.
Please,
yes,
before
we.
B
On
the
next
item-
I'm
a
co-author
of
this,
so
speaking
as
individual
just
for
the
working
group
as
a
reminder-
and
you
know
we
have
introducing
something-
a
new
map-
now
we
call
a
map
now
a
new
type
of
file.
You
used
to
call
it
out
to
service
and
one
of
the
things
that
was
an
open
issue
was
the
error
handling
and
the
question
was
of
the
traditional
Al
Jarreau
RFC,
seven
to
eight
five,
you
know
has
enough
error
messages,
and
here
are
some
where
this
is
I-
think
a
key
progress
on
this
document.
M
Next,
thank
you,
and-
and
here
here
we
introduce,
we
introduced
nude
footprint
on
our
network
map
and
we
update
the
encoding
of
auto
network
map.
Footprint.
Hello
is
key,
is
below
is
an
example.
First,
we
add
the
dependent,
be
text
field
here
so
that,
in
the
right
hand,
side,
city,
NFC,
my
FC,
I'm
f-
is
depend
on
my
yo
net
maps
and
then,
from
the
right
hand,
side
the
bottom
of
the
right
hand
side
we
can.
We
can
see
that
we
used
PID
Germany
and
the
South
France
as
footprints
next
slide.
Please
again.
B
Here,
just
so
the
shorter
add-on
and
from
the
beginning
and
I
mean
I've
been
pushing
this
for
a
long
time
and
from
the
beginning,
I
thought
Alto
was
there
or
ultra
network
maps,
was
the
natural
choice
to
express
footprints.
So
here
we
have
it
used
to
be
an
example,
and
but
here
we
have
now
a
clear
way
of
doing
this,
this
doesn't
mean
that
there
couldn't
be
other
footprints,
but
obviously,
and
using
an
outer
network
map
makes
a
lot
of
sense.
B
M
So
and
we
add
and
I
consideration.
First,
we
register
a
new
city
and
I
meet
at
8
print
type,
Auto
network
map,
and
then
we
register
two
entities
domain,
one
is
E&E
and
the
other
one
is
a
ASN
and
the
other
is
the
country
code.
And
finally,
we
also
add
an
audio
CD
and
FCI
property
type,
CD
RFC
a
capabilities
next
slide.
Please-
and
here
we
want
to
clarify
two
items.
The
first
one
is
the
media
type
of
CD
AFCI
map
and
the
filtered
CD
f's
EMF.
M
M
Embedding
this
me
at
least
need
a
new
pro
approaches
such
as
the
multi
part,
which
is
tabled
for
now
and
LA,
and
the
other
clay
clarification
items
is
that
they'd
had
components,
name
of
sadena,
FC
IMF
and
the
field
healthy,
DNI,
XML
and
currently
design
is
sitting
IFC
I
slash
map
and
we
proposed
to
change
to
see
DNI
/fc
I
slash
map
so
that
it
could
consist
with
the
network
map
host
a
map
and
endpoint
cost
met
next
slide,
please!
So!
For
a
next
step,
we
work
on
more
reviews
and
I.
M
B
As
chair
I
can
say,
and
we
want
to
move
this
app
to
watch
waterglass
call
as
soon
as
possible,
of
course,
so
I
think
the
stockmen
has
made
good
progress
and
individually
speaking
as
co-author
I
think
we
are
not
quite
there
yet.
We
need
maybe
to
polish
here
and
there,
but
as
far
as
I
can
tell
a
maybe
you
can
comment,
there
are
no
current
outstanding
major
issues
like
we
had
on
the
other
documents
right.
We
don't
have
any
major
issues
that
need
feedback
from
the
working
group.
B
B
M
B
Okay,
so
next
on
the
agenda
is
now
the
first
documents
are
being
write.
That
is
not
a
working
group
ID
item
yet
or
not
does
not
address
the
core
milestones
I'm
looking
at
the
time,
so
we
are
you're
supposed
to
start
at
40
right
and
we
are
so.
We
are
about
10
minutes
and
behind
our
schedule.
So
I
would
ask
every
one
of
the
remaining
five
presentations
you.
B
O
This
proposal
was
discussed
already
are
some
time
ago,
so
I
will
focus
here
on
what
what
relates
this
work
to
actually
the
unified
property
graph.
Actually
that
the
main
discussion
there,
the
outcome
was
that
it
really
stresses
the
need
to
harmonize
and
to
specify
rules
and
in
the
auto
entity
domain
registry
and
the
relation
to
the
auto
address
type
registry.
O
So
next
slide,
please,
okay,
so
the
updates
in
the
current
version,
so
they
expose
the
suitability
of
the
proposed
so
called
ECG
I
domain
for
both
the
address
type
registry
and
the
address
entity,
domain
registry
and
so
the
placeholder
to
specify
these
consistency
features
our
is
the
unified
property
draft.
So,
as
already
discussed
today,
the
decision
is
that
a
unified
property
draft
will
stress
on
new,
auto
information
services
such
as
the
property
map,
and
how
to
propose
and
specify
a
new
address.
O
Entity,
sorry
entity
registry
entity
domain
registry
and
how
to
really
be
rigorously
making
the
difference
of
the
relation
with
a
address
type.
So
in
this
version
there
is
a
new
section
3
on
cellular
addresses,
so
it
relates
the
both
registries,
as
a
motivation,
I
already
told,
is
because
cells
can
be
either
endpoints
on
which
you
want
to
query
properties
as
if
they're
based
protocol
or
they
can
be
entities
on
which
to
query
property
maps.
Phil
talked
a
lot,
so
the
section
explains
how
the
base
protocol
and
the
unified
property
trust
both
provides.
O
A
background
to
allow
to
register
a
new
domain
type
such
as
its
each
ECG
I.
Where
is
Jerry,
stems
from
the
3gpp
terminology.
By
the
way
this
work
actually
stems
back
from
ITF
in
to
2014
in
Honolulu
I
guess
it
was
already
proposed,
but
in
the
meantime
the
use
cases
have
been
progressing
and
we
are
going
to
outside
G.
So
we
want
to
have
we
see
the
interest
into
using
that
to
expand
how
to
beyond
on
the
ipv6
or
ipv4
networks
next
slide.
O
So
the
in
this
version,
so
the
section
6
yeah
IANA
consideration-
was
field
as
well
again
account
the
working
group,
discussion
and
I
think
that
text
that
I
code
here
should
be
updated
with
respect
to
the
progress
that
is
done
in
the
Ayane
section
of
the
unified
property.
Draft
I
won't
go
into
details
next
slide,
and
so
the
further
discussions
will
be
on
harmonizing
those
two
I
have
the
Ayane
sections
in
those
two
drafts.
O
Also
I
would
like
to
have
a
discussion
on
what
is
the
placeholder
for
new
address
types
on
you
and
city
domains?
Okay,
we
know
it's
the
auto
address
tag
registry
and
auto
entity
domain
registries.
We
already
discuss
how
to
about
the
the
the
manual
harmonization
or
automatic
harmonization,
but
in
general
the
question
would
be.
How
do
we
collect
all
the
auto
registry
updates
because
reading
the
unified
property
draft,
all
the
base
protocol
gives
you
no
hint
on
the
existence
of
other?
So
do
we
really
want
to
go
through
the
whole?
O
All
the
outer
drafts
that
proposed
extensions
and
find
out
that
somewhere
there
is
a
new
domain
or
a
new
address
type.
So
the
proposal
is
to
give
a
hint
in
the
unified
property
of
the
fact
that
there
are
currently
new
domain
types
of
entity,
domains
that
are
being
proposed
and
specified
next
slide
and
yeah.
So
so,
basically,
that's
it.
The
plans
for
the
next
section
is
for
the
work
on
that
I
am
a
section
and
post
our
discussions.
Thank
you.
Okay,.
O
O
The
issue
would
be
how
to
if
we
focus
on
working
group
documents.
Their
goal
is
to
specify
how
you
define
an
information
service,
how
you
use
it.
So
that's
the
focus
now.
That's
the
option
chosen
now,
but
on
the
other
hand,
the
I
kind
of
miss
something
that
makes
people
find
out
that
there
is
such
and
such,
for
example,
how,
if
I,
could
probably
only
be
interested
in
the
unified
properties?
Okay.
So
how
can
I
find
out
that
there
is
something
that,
like
a
and
E.
O
B
D
B
N
N
Sir
first
I
will
talk
about
one
even
be
about
the
mood
America
trade
soon
after
I
I
will
impressing
or
purposely
the
production.
Finally,
I
will
introduce
some
alt
attention
in
order
to
support
the
main
functionality
single
purpose
that
architecture
with
next
a
coroner
future
five-year
net
word
a
scenario:
corpora
call
for
a
broader
cooperation
between
a
stakeholders
in
order
to
provide
flexible
multi-terminal
to
Pareto
services.
N
This
operation
will
require
the
information
with
chain
across
multi-domain
orchestrators
and
the
O's
the
information
to
be
exchanged.
It
includes
the
entire
network
topology,
the
social
reality
Orchestrator
and
eight
points
between
difference,
administrative
domain.
However,
the
multi
my
Orchestrator
faces
different
challenges,
for
example
the
lack
of
a
structure
in
order
to
discovery
an
adequately
replacing
the
resource
and
topology
information
and
the
discovery
of
candidate
autonomous
system.
In
order
to
provide
the
requirement
for
a
mentoring
network
service
request
on
another
challenges.
N
Next,
they
see
what
these
contexts
and
challenges
or
proper,
stirrup
or
opposite
approach
is
to
create
a
federation
network.
Empowering
the
top
of
the
manager
meaning
orchestration
play
both
dominguin
is
to
discover
resource
and
topology
information
from
different
administrative
domain.
In
the
clear
ratio
and
or
design
research
on
the
ultra
services
to
offer
a
staff
network
maps
with
a
simplified
view,
get
enough
information
about
motive
America
stature
if
they
to
involve
it
in
the
parish.
N
Next,
please
here
we
can
see
a
proposal
architecture
where
the
broker
component
is
conservative
to
be
working
as
a
coordinator
between
a
set
of
multi-domain
Orchestrator
who's,
the
main
components
at
the
inter
demanded
source
in
order
to
create
the
article
database
with
the
entertainment
resource
information
in
the
domain
topology
component
to
create
a
directly
contradict
engineering
database
and
an
artisanal
use.
The
information
collected
for
the
inter
demanded
source
and
interment
ability
to
create
to
create
a
step
maps.
N
This
example
can
be
supported
with
some
extension
to
the
Alta
based
protocol
next,
in
case
of
the
property
map,
the
alto
server
must
read
two
x
rays
for
each
property
in
the
property
map.
In
this
case,
the
disposal
pacification
was
extended
for
each
property
map
name.
Defining
the
resource
capability
means
the
corresponding
property
value
must
be
encoded
as
a
JSON
array
listing
of
a
JSON
string.
Next
this
here
we
have
an
example
where
the
Alta
client
wants
to
retract
the
entry.
N
The
entry
property
map
for
pids
entities
with
the
entik
point,
CPU
memory,
storage
under
needs
of
service
access,
point
support
and
additional
network
functions
up
next
and
in
case
of
the
cosmic,
the
active
server
must
provide
connectivity
information
for
every
serving
drafting
in
a
serving
draft
path
for
an
end-to-end
requiring
this
information
is
the
a-level
topological
distance
in
the
form
of
a
path
vector,
and
this
map
is
the
set
input.
Parameter
specification
was
extended,
for
example,
they
required
filter.
N
Eight
Cosma
support
our
service
graph
with
a
formatted
body
as
a
network
function
for
working
this
name
a
thief's.
An
FFD
object
include
as
a
set
of
network
function
at
set
of
megaphone
shown
a
set
of
service
access
points,
a
set
of
logical
interconnection
between
network
function,
service,
access,
point
on
Bob
and
Alice
of
influence
network
service
requiring
me
Leslie
I
in
the
case
of
the
respondent
specification,
respond
for
each
certificate.
N
N
Only
to
say
it
is
an
example
for
definitive
code
map
where
we
have
forgot
to
client
to
request
a
perspective.
In
this
example,
we
have
to
enable
function
with
two
service
access
point,
our
four
links,
onion
200,
quite
a
minute,
and
in
the
next
response,
when
we
can
see
in
it
the
as
next
steps
radius
to
collect
some.
B
I
I
think
it's
very
cool
and
the
question
the
following:
we
do
have
relief
to
work
because
you
want
actually
another
one,
a
networking
but
also
other
resources,
for
example
CPU
right
yeah,
so
the
Mediacom
were
working
with
this
IBM
with
very
large
US,
army
and
and
and
actually
ok
project
is
for
CPUs.
They
get
confused.
So
you
have
CPU
of
a
would
have
sleep.
You
give
the
BIA
vas
a
PC.
You.
G
I
You
know
broker
and
give
you
for
me:
TP
give
me
mercy,
and
now
NP
NSA
give
me
information
2d
and
in
my
poster
exam
I
have
or
computer
storage,
and
now
do
you
might
have
fire
might
have
10,
so
I
think
you're
trying
to
solve
that
economy
using
your
past
Becker.
Is
that
the
reason
you
reasoning
that
you
must
incur
a
hope,
a
Spectre
2d
class,
essentially
to
distinguish
how
the
colors
or
labels
or
what
the
competition
is
that
really
the
reason
you
do
this
stuff?
I
B
N
B
Alright,
thank
you
very
much
in
the
interest
of
time
we
have
to
move
on.
Thank
you
and
of
course,
any
more
comments
are
on
the
mailing
list.
That's
certainly
welcome,
and
so
we
have
three
more
presentations.
I
am
and
12
more
minutes,
so
that
makes
it
very
simple
to
birth.
Divided
by
3
is
4,
so
the
remaining
three
presentations
each
get
four
in
minutes
and
first
is
our
use
case.
Research
orchestration
for
multi
domain,
geodesy
root
areas,
and
this
is
yeah-
please
grad.
P
Yeah,
okay,
thank
you!
So
how
many
time
it's
still
left,
you
have
four
minutes,
sorry,
so
yeah!
So
from
the
last
item
interim
meeting
they
just
present
have
printed
the
Unicom
data
and
information
impairments
and
odometry
earlier
type
in
the
super
come
to
in
17
and
yeah
and
actually
in
this
item,
while
one
we
have
to
meet
a
newest
version
and
it
made
the
potential
updates
in
the
previous
version
and
requests
to
better
prepare.
They
talked
to
me
for
the
reveal
from
the
working
group
and
we
propose
an
auto
extension
just
about
the
accurate
privacy-preserving.
P
P
So,
from
the
background,
we
pass
that
the
mountain
domain,
disability
patron
and
his
system
has
that
features
the
post
workload
and
resource
every
bit
is
held
a
dynamics
so
based
on
these
features,
we
conduct
several
to
Dan
requirements
in
the
next
slide,
yeah
and
from
the
user.
As
pad
we
found.
We
need
to
provide
the
performance,
predictability
and
the
efficient
resource
sharing
and
from
the
member
networks
aspect.
We
need
to
achieve
the
high
resource
utilization
and
when
implement
in
the
privacy
and
autonomy
autonomy
of
the
member
networks,
also
in
to
provide
a
capability.
P
So
next
one.
So
we
review
some
existing
design
of
the
many
systems
in
our
document
and
found
that
this
system
cannot
satisfy
these
requirements
because
they
are
lacking
our
information
model
T
spots
as
a
result
of
curry,
first
precise.
So
next
one
so
in
the
unicrone
our
system,
the
fundamentalism
decision
is
to
charge
the
out
who,
as
a
resource
information
model
for
the
several
reasons
and
cannot
talk
about
the
videos
just
next
right
yeah.
P
So
we
just
use
a
kind
of
architecture
of
the
unicrone
system
which,
including
the
auto
server
and
plant,
and
some
excuse
agent,
to
take
the
tasks
from
the
application
and
discover
a
resource
from
the
different
member
networks
and
fill
the
straight
face.
We
saw
this
car
really
precise
and
we
will
talk
the
videos
in
the
next
night
yeah
and
at
the
first
date,
faith
I.
Think.
B
P
Sure
great
finalized,
since
that
in
day
Webster,
so
so
you're
encouraged
to
provide
a
you
know,
land
for
the
last
key
medical
meters,
in
her
case,
at
least
in
the
auto
working
touch,
and
we
plan
to
submit
a
full
paper
to
Melissa
super,
come
to
18
and
from
the
last
Keo
demonstration
and
the
deployments
just
by
maybe
idea
by
all
serene
yeah.
Thank
you
so
much
thank.
B
You
and
any
comments
right
now,
I
mean
I'd,
see.
Obviously
you
put
a
lot
of
effort
into
updating
this.
So
that's
great
and
obviously
it's
not
a
charter
item
right
Mouse.
We
have
to
see
how
we
can
move
this
forward,
but
thank
you
for
the
presentation.
Are
there
any
quick
comments,
otherwise
to
be
fair,
I
would
allocate
another
four
minutes
for
the
next
two
documents
sure
so
next
one
is
the.
P
So
so
we
just
highlight
the
major
update
of
the
the
latest
version
of
the
absolute
card,
the
move,
the
second
sticks
as
well.
The
flow
based
Curie
out
of
these
documents
and
which,
in
the
output
rest
have
complex
registry
to
the
auto
dressed
have
bility
registry
and
the
reason
I
will
talk
about
in
the
next
slide:
yeah,
okay.
So
so
before
we
talk
about
the
news
is
a
review
just
recall
the
our
requirements
and
proposed
all
the
FCS?
P
Probably
in
the
last
item
meeting
we
just
give
it
yeah
sure,
so
so
in
the
in
this
working
out
the
got
away
just
so.
There
were
many
issue
from
the
last
item
meeting
the
first
language.
We
want
to
show
how
to
achieve
a
unified
theory
model,
because
we
have
two
different
model
design
options
and
a
other
issues
how
to
resolve
the
flow
attribute
complex
because
a
well
definition
from
the
clan
reference
may
be
embarrassed.
So
we
want
to
allow
the
server
to
notify
this
ready
T
to
the
clan
as
early
as
possible,
so
network.
P
So
we
can,
we
have
not
found
a
unified
model
to
integrate
outdated
and
options
that
we
just
compare
the
existing
to
desire
genes
and
you
can
see
in
the
for
the
person
master
technician
here.
That
decision
can
achieve
the
battery
with
mobility
and
the
can
reduce
the
request,
ties
and
it's
extensible
enough,
and
there
is
simple
enough
I'll-
do
it's
the
flexibility
of
this
design
decision
better
than
the
another
one,
but
we
will
introduce
some
at
rest
time.
Newest
new
aggressed
had
to
make
his
to
support
some
use
cases.
P
Yes,
and
enough
so
next
day-
and
this
is
a
new,
a
grass-type.
We
just
introduced
in
this
document
or
use
each
age
age
to
suppose
a
maybe
layer
to
flows
between
the
inter
personal
networks,
and
we
use
it
to
mean
to
spoil
the
city
in
your
cases
and
the
way
you
CP
and
you
in
code
so
get
dressed,
and
it
can
support
the
floor
level,
scheduling
and
the
other
major
use
case
about.
We
want
to
saw
in
the
photo.
B
I
think
because
we're
running
out
of
time,
we
refer
any
discussed
to
the
meatiness
regarding
working
group
item,
you
know
that
you
as
chairs
want
to
finish
the
myosin,
so
not
at
this
time
and
let's
see
what
goes
on
after
we
finish
the
milestones
with
the
working
group.
I
see
there
is
a
lot
of
interest
in
new
work,
and
so
we
have
to
see,
but
not
at
this
point
in
time.
Thank
you.
Yes,
thank
you.
B
B
B
D
D
B
P
Q
Okay,
wonderful,
so
thank
you
Jason,
so
I
do
a
remote
presentation.
Thank
you.
Young
and
I
hit
a
ground
KY
and
today
I'm
going
to
prison
in
the
ER
say
draft.
Since
we
are
running
out
of
time,
I
would
give
some
of
the
pictures
that
involve
technical
details
and
key
Bonio,
quick,
summary
so
next
time,
peace.
So
here
the
summary
of
the
draft
for
those
who
may
not
be
familiar
with
learning
state
abstraction,
so
it
is
short
for
running
state
abstraction
and
instead
of
algorithms,
to
provide
information
for
a
set
up
correlated
flows.
Q
So
the
information
is
encoded,
as
in
Auto
has
spectral
response,
so
how's
our
civility
to
our
working
group
drafts.
First,
it
provides
a
commutation
of
the
path
vector
extension
as
a
stand.
Is
there
no
implementation,
and
you
can
also
be
used
to
compress
and
improve
the
privacy
of
an
existing
path,
spectral
response
without
loss
of
information.
Q
So
next
up
is
here.
We
summarize
the
changes
since
the
last
version,
since
the
overall
structure
of
the
is
pretty
stable
after
0-7,
we
are
focusing
on
improving
the
clarity,
so
in
this
revision,
which
changes,
symbols
that
might
look
complex
and
extended
examples
to
demonstrate
the
intermediate
state,
for
example,
if
I
always
omit
a
multiple
iterations,
we
demonstrate
the
results
after
each
iteration.
Well,
in
version
7,
we
only
demonstrate
the
results
after
the
first
iteration.
Q
So
next
up
is
here's
just
a
summary
of
the
contents
of
the
RC
draft.
It
contains
Tuesday's
algorithm.
The
first
dating
includes
a
three
August
and
to
tell
people
how
to
do
the
compression
and
the
same
set
includes
the
encoding
and
decoding
of
information.
Basically,
how
do
we
extract
formation
for
the
passport
or
extension
and
how
to
put
information
in
into
a
pass
vector
response?
B
Q
So
here
is
a
summary
of
the
ours
to
draft.
We
have
this
kind
of
rows
of
this
draft
and
also
the
updates
in
the
rated
version,
so
it
is
targeting
informational
track
and
we
have
not
updated
that
this.
You
know
draft
and
I
will
do
this
in
the
next
revision,
and
we
also
like
to
cover
reviews
from
the
ring
group
and
to
adopt
this
document
as
a
Wynkoop
draft.
After
the
milestones
are
set.
Q
B
You
Kai
and
regarding
the
usefulness
of
this
work,
I
as
speaking
as
Cherokee,
you
see
the
usefulness,
I
mean
compressing
path.
Factors
regarding
process
moving
is
through
working
bottom.
My
answers
to
the
same
as
for
the
last
one,
let's
finish
the
milestones
first
and
then
see
if
we
reach
out
at
the
working
group
or
not
and
so
I'm
going
to
coach
other.
B
We
had
a
discussion
with
Mara
and
I
think
she
fully
agrees
with
me
Jaime
that
we
want
to
finish
the
current
milestones
and
then
we
have
to
see
from
each
other
so
and
if
we
do
that,
then
certainly
this
is
a
possibility,
but
we
wouldn't
make
it
work
in
vitam
right
now.
Okay,
thank
you
for
presentation,
I'm,
happy
that
it
worked
out
remotely
and
we
are
running
over
time.
B
So
I
think
we
are
done,
I'm
happy
that
we
made
all
the
presentations
and
we
got
a
good
overview
on
the
status
of
all
working
group
items
and
one
last
comment:
vijay
mentioned
that
he
is
looking
forward
to
the
meeting
in
montreal
and
he's
very
hopeful
that
he
can
be
there.
So
I'm
not
sure
yet.
I
have
to
see,
but
we
are
mostly
meeting
in
montreal
and
hopefully
there
we
can
close
a
lot
of
the
or
by
then
actually
I'll
be
impressed
most
of
the
working
with
less
cause
that
we
have
on
the
agenda.