►
From YouTube: IETF111-ALTO-20210729-2200
Description
ALTO meeting session at IETF111
2021/07/29 2200
https://datatracker.ietf.org/meeting/111/proceedings/
A
Yeah
good
to
hear
you
and
let's
get
started
so
welcome
to
itf
111
virtual
meeting,
and
this
is
the
auto
working
group
session.
My
name
is
ching
and
sit
online.
Yeah.
B
So
this
is
the
notebook.
B
A
You
want
no
yeah.
Let
me
take
care
of
this,
so
another
way
apply.
Actually
this
auto
session
will
automatically
record
it,
and
that
means
all
what
you
said
and
presented
will
be
deemed
as
a
contribution
to
the
ietf.
If
you
have
any
ipr,
or
you
have
been
already
aware
of
any
api
related
to
the
auto
work,
please
disclose
it
in
accordance
with
ietf
rules.
A
A
A
A
Actually,
yeah
for
blue
shield.
Actually
we
have
a
virtual
blue
shield,
so
your
attendance
will
be
automatically
recorded
and
the
notes
take.
Actually
we
we
actually
send
a
request
to
the
list
and
we
have
two
volunteer.
One
is
richard
young
and
another
player
is
tanya
king,
and
so
thanks
for
that
to
have
her
take
a
notes
and
we
have
online
agenda
and
and
also
slides.
A
A
A
So
this
is
the
agenda
for
today's
discussion.
First,
we
will
discuss
working
with
status.
I
will
ask
vijay
to
give
a
quick
update
of
document
status
and
for
auto
we
almost
finished
all
the
existing
worker.
We
are
in
the
recharging
stage.
So
the
second
topic
I
will
ask
martin
to
provide
some
guidance
for
the
next
child
and
the
future,
the
group
and
and
then
actually
you
know,
see
auto
and
pid
share
a
lot
of
commonality
and
also
focus
on
network
aware
about
application
and
pig
focus
on
you
know
past
aware,
application
and
networking.
A
A
Hopefully
we
can
find
some
common
interest
and
then
we'll
you
know
dive
into
the
discussion
as
the
chair
will
lead.
The
discussion
introduce
the
charter
and
then
we'll
reserve
10
20
minutes
to
open
to
open
discussion
to
the
floor
last
will
wrap
up
and
any
agenda
bash.
B
Great,
thank
you
all
all
right,
so
I
think
from
the
viewpoint
of
the
current
chartered
items,
all
of
the
work
has
been
moved
towards
iesg.
We
were
basically
looking
at
four
drafts.
The
performance
metric
draft
cdna
request,
routing
path,
vector
and
unified
props.
B
The
last
three
of
the
last
three
two
depended
on
unified
props,
cdna,
request,
routing
and
path
vector,
so
they're
moving
ahead
as
a
cluster
and
performance
metric
is,
is
also
moving
ahead.
So
from
the
viewpoint
of
existing
work
charter
items,
I
think
once
they
pass
iesg
and
get
in
with
the
rfc
editor,
you
know
those
will
be
done.
B
D
Okay,
I
was
just
about
to
write
on
the
chat,
so,
first
of
all,
thank
you,
martin
for
your
review
and
guidance.
We
have
internally
addressed
most
of
the
comments
and
some
questions
actually
interesting.
Questions
regarding
the
design
and
possibilities
to
to
improve
the
design
to
access
better
so
to
some
information
are
being
discussed
among
the
authors
and
the
plan
is
to
send
an
answer
to
martin
and
all
recipients
next
week
it
will.
The
response
will
be
ready
next
week
with
a
given
number
of
points
of
discussion.
B
Okay,
great
from
the
other
drafts,
I
noticed
that
you
know
martin
had
some
comments
on
them,
but
I
did
not
see
anything
that
was
blocking
or
needed
any
explanation.
So
unless
the
authors
of
those
documents
want
to
say
something
we
can
move
ahead.
C
E
Was
just
in
queue,
so
I
I
like
to
say
something
about
the
past
vector
document
so
so
martin
have
have
provided
the
any
review
on
tonight
twitter,
but
I
just
thought
because
the
email
goes
into
my
spam.
So
I
just
replied
the
email
yesterday.
So
I
think
maybe
we
can
take
a
few
days
and
see
how
martin
and
others
would
respond
to
the
to
the
comments.
B
Okay,
martin,
you
are
also
in
the
queue.
Do
you
want
to
say
a
few
words.
F
Just
to
clarify
right
so
I
mean
I've
not
had
an
opportunity
to
to
look
at
kai's
reply,
but
yes,
so
so
performance
cost
metrics
just
to
run
through
these
very
quickly
performance.
Metrics
had
I
mean
most,
the
important
comments
were
addressed,
and
so
I
sent
it
to
last
call
the
minor
comments
which
might
be
covered
in
the
latest
revision,
but
either
way
they're
just
knits
pathfector
and
unified
props.
F
Yes,
I
was
hoping
for
some
responses
from
the
authors
and
probably
a
new
new
id
in
each
case
before
we
went
to
last
call
and
cd-
and
I
request
routing
is
the
one
I
haven't
gotten
to
yet.
I
think
I
got
all
three
at
once,
so
it's
taken
me
a
while
to
get
through
them,
but,
but
I
anticipate
certainly
no
later
the
end
of
this
month
having
an
ad
review.
A
A
F
Right,
well,
I
don't
have
any
slides,
so
this
should
be
easy.
I
I
I
wanted
to
take.
I
wanted
to
take
about
five
minutes
just
to
talk
about
the
state
of
the
working
group
and
its
future.
I
wanted
to
be
really
just
open
and
transparent
with
everyone
about
where
things
are
when
I
took
over
as
the
ad
and
we
were
wrapping
up
these
documents
we
started
discussing
recharter
and
I
tried
to
understand.
F
The
status
of
this
protocol
was
in
the
internet
and
at
the
time
there
were
a
number
of
reports
and
kind
of
second
hand,
comments
that
you
know
this
was
deployed
in
a
number
of
different
places
and
the
people
were
actually
out
there
using
alto,
and
we
got
a
number
of
and
got
a
lot
of
proposed
work
items
that
spanned.
You
know
a
lot
of
levels
of
maturity.
Some
of
them,
I
thought,
were
a
little
bit
researchy
frankly,
and
a
few
things
have
happened
since
then.
One
is
that
it
is.
F
It
has
come
to
my
attention
a
lot
of
the
things
that
were
believed
to
be
deployments
were,
in
fact
you
know
proofs
of
concept
or
experiments
and
so
on
and
that's
fine.
Like
you
know,
the
internet
research
is
important.
I
I
don't.
I
don't
want
to
discuss
that.
It
is
not.
I
don't
want
to
suggest
that
it
is
not,
but
the
at
this
moment.
F
I
am
not
convinced
that
you
know
that
there
are
actual
alto
applications
out
there
in
the
real
world,
with
consumers,
clients
going
to
alto
servers
that
isv's
deployed
to
optimize
paths
anywhere.
That
may
be
the
case,
I'm
saying
I
don't
know,
and
the
itf
is
not
in
the
business
of
of
extending
protocols
that
are
not
really
being
used.
F
F
But
really
the
idea
here
is
this
should
not
take
a
long
time,
and
I
really
want
to
give
the
working
group
an
opportunity
to
to
see
if
the
existing
leads
that
it
has
succeed
in
in
getting
this
thing
deployed,
and
then,
if
that
is
the
case,
then
we
can
talk
about.
You
know
further
extending
it,
but
I
think
if
that
isn't
the
case,
then
this
might
be
the
last
iteration
of
of
the
alter
working
group,
at
least
until
things
speed
up
now,
just
to
say
a
little
bit
more
about
this.
F
We
have
not
made
it
through
isg
review
on
the
charter.
It
is
possible.
This
charter
could
be
slimmed
down
even
more.
It
is
also
not
outside
the
realm
of
possibility
that
we
just
sort
of
freeze
the
working
groups
work
until
evidence
of
of
any
sort
of
deployment
exists,
so
we'll
see
how
that
discussion
occurs
at
the
isg,
and
I
cannot
promise
you
anything
there.
The
second
thing.
The
second
thing
I
would
like
to
address
is
that
there
is
a
lot
of
interesting
research
work
going
on.
I
mean
a
lot
of
the
proposals.
F
I
would
characterize
the
chara
when
we,
when
we
recharged,
when
we
talked
about
rechartering,
a
lot
of
proposals-
I
I
would
characterize
as
as
research
and
as
it
happens
there
and
as
much
as
the
isg
is
not
necessarily
about
ietf
is
not
about.
You
know,
research
projects,
the
irtf
is,
and
in
fact
a
lot
of
this
stuff
is
is
would
be
fit
very
nicely
into
the
pan
rg,
and
that
was
why
I
asked
the
chairs
to
invite
some
representatives
from
pnrg
to
come
here.
F
I
think
a
lot
of
the
work
that
has
been
discussed
as
potential
new
work
in
alto
would
fit
very
nicely
in
pan
rg
and
in
fact
you
would
get
much
more
robust
review.
I
think
in
pan
rg
than
you
would
here.
So
that's
why
spencer
and
jen
are
here
to
talk
about
what
is
what
the
opportunities
are
there,
and
I
would
encourage
those
of
you
who
are
who
would
like
to
move
forward
on
this
work
in
the
irtf
ietf
to
really
consider
rg
as
a
as
a
somewhere.
F
You
can
go
quickly
to
get
some
review
and
make
progress
on
that,
while
alto
kind
of
catches
up
with
you-
and
you
know
I
I
if
there
are
any
questions
or
comments
about
anything,
I've
said.
I
think
this
would
be
a
good.
F
Richard
you
have
to
unmute
yourself.
There
you
go.
G
Okay,
yeah,
I
think
that's
a
great
comment.
I
think
one
thing
really.
I
actually,
I
do
know
a
couple
of
pretty
reasonable
deployments
of
you
know
real
life,
but
everything
needs
to
be
essentially
really
clarified
and
be
really
documented
very
well.
I
think
that
really
is
the
focus
and
eventually
I
think
that
would
be
the
focus
in
this
iteration
to
essentially
get
all
these
people
publish
all
their
results
and
get
the
policy,
and
so
I
think
that's
fair
statement.
F
Yeah
I
mean
like
I,
I
honestly
thank
you
for
sharing
that,
like
I'm,
I'm
sort
of
yeah,
I'm
glad
to
hear
that
this
is
out
there
in
the
real
world,
and
I
I
don't
know,
I
don't
want
to
say
that
everyone
has
to
show
up
and
like
publish
results
necessarily
but
but
like
from
actual
practitioners
who
are
out
there
dealing
with
customers
like
for
them
to
come
here
and
help
motivate
these
use.
F
Cases
would
be
of
these
further
extensions
would
be
really
valuable
and
I
think
we
need
to
take
some
time
to
do
that,
because
my
honest
perception
of
the
set
of
proposals
we
saw
whatever
that
I
think
was
itf
107
or
108
struck
me
as
pretty
much
a
research
agenda
honestly,
quite
frankly,
and
that's
not
exactly
the
right.
This
is
not
exactly
the
right
venue
for
that
kind
of
stuff.
F
All
right
well,
thank
you
for
listening
and
I
think
we'll
hand
it
off
to
chin
or
maybe
directly
to
the
to
the
pan.
Rg
people.
H
So
yeah,
okay,
you
sharing
the
slide
deck
perfect,
so
hello.
Thank
you,
martin,
for
giving
a
very
nice
summary
of
my
talk
before
I
even
started,
so
why
we
are
here
because,
right
after
this
session,
next
one
will
be
pay
energy
session
and
you
are
all
invited.
Actually
I
don't
know,
I
suspect,
half
of
the
audience
already
knows
what
I
am
going
to
say,
because
I
see
a
lot
of
familiar
names
in
the
participant
list,
but
anyway
pass
our
network
and
research
group.
Who
are
we
what
we
are
doing
next
slide?
Please.
A
B
Trying
to
go
ahead
in
the
next
slide,
but.
H
I
said
slate
number
two,
I
think
yeah
perfect.
So
what
is
as
our
networking
thing?
Historically,
if
you
think
about
it,
there
was
some
kind
of
disconnection
between
endpoints
and
the
networks
and
packets
and
those
packets,
magically
reappear
on
the
other
side
of
the
network,
or
maybe
not
that's
why
we
have
tcp
and
for
the
from
the
host
perspective
network
is
a
kind
of
black
box
right
break
it
thing,
packets,
sound
you
have
no
idea
what's
happening
inside
and
all
this
like
routing,
curator
atf
takes
care
of
that
right.
H
H
So
obviously
it
would
be
actually
really
nice
if
network
should
be
the
black
box
for
an
application
and
if
hosts,
could
somehow
get
some
knowledge
about
the
available
path
in
the
network
and
make
some
educated
decisions
about
which
path
could
be
used
and
how
they
could
be
used
well
to
some
degree.
It's
already
happening
right.
If
I
take
this
phone,
it
has
number
of
interfaces.
It
could
make
some
choices
which
interface
to
use
vent
right,
but
it's
all
been.
H
H
It
means
that
in
our
sessions
we
see
a
lot
of
routing
people
and
transport
people
and
other
people
as
well
actually
come
and
talk
about
what
they
think
should
be
happening
and
what's
happening
now.
So
next
slide.
Please.
H
What
is
currently
going
on
there
like
traditionally
what
we
are
trying
to
do
we
trying
to
have
two
parts
in
our
sessions?
We
try
to
do
some
research
group
items
to
be
discussed
and
we
normally
invite
some
guest
talks
from
related
areas.
H
So
we
currently
have
one
document,
which
is,
I
think,
going
through
the
rsv
poll,
which
is
open
questions
in
password
networking,
because
we
are
doing
the
research
we
try
to
do
the
right
thing
and
start
from
asking
what
not
how
so
we
decided
to
formulate
the
questions
we
need
to
do.
Research
on
first
and
microchair
brain
wrote
a
great
document
about
eight
open
questions
in
password
networking.
H
First
of
all,
obviously
we
want
to
define
what
path
is.
Surprisingly,
there
are
a
lot
of
opinions
about
this
stuff
how
those
past
properties
are
represented,
and
thanks
to
terry,
then
cereal.
We
already
have
a
draft
which
is
trying
to
answer
the
first
question,
which
is
another
working
group.
Research
group
item
pass
properties
document.
H
Secondly,
we're
trying
to
define
how
endpoint
can
discover
those
path,
information
right,
because
those
paths
might
exist
on
the
network
but
and
points
need
to
discover
that
information
and
discover
it
in
reliable
way
and
after
that,
information
is
discovered
how
the
endpoint
make
decisions
about
using
that
information,
how
to
select
or
best
path,
or
maybe,
multiple
paths
out
of
the
all
available
ones,
and
how
that
information
could
be
exchanged.
How
what
kind
of
interfaces
can
transport
an
application
layer
to
might
have
to
use
information
about
path?
Next
slide,
please.
H
H
If
you
speak
from
practical
point
of
view,
if
you
think
about
it,
we
might
probably
even
if
we
one
day,
ietf
designs,
a
perfect
path,
our
networking
architecture,
it
probably
wouldn't
be
deployed
overnight,
so
most
likely
we'll
start
seeing
kind
of
overland
isolated
islands
scenarios
which
needs
to
be
connected.
So
probably
all
these
things
which
I've
been
talking
about
relations
to
hosts
will
be
also
applicable
to
end
points
channel
endpoints,
which
interconnects
isolated
islands.
H
H
I
Is
that
better?
So
thank
you.
Thank
you.
So
this
is
the
rc
that
jen
was
mentioning
and
it
was
basically
when
we
had
the
first
research
group
meeting,
we
had
a
show
of
hands
in
the
the
in
the
participants
of
how
many
you
know
how
many
people
had
worked
on
at
least
one
pathway,
networking
technology
at
the
ietf
that
was
not
deployed,
and
we
got
so
many
hands
that
that
that
cataloging
them
became
the
first
work
item
for
the
for
the
research
group.
I
I
They
seem
to
come
up
again
and
again
in
protocol
proposals
and
they
seem
to
still
be
a
problem
the
next
time,
so
it
seemed
useful
to
write
this
down,
and
actually
I
got
comments
from
the
isg
during
isg
review
conflicts
review
that
they
really
wish
that
we'd
had
a
document
like
this
a
long
time
ago,
but
this
list
informed
the
list,
the
development
of
the
open
questions
draft.
I
That's
now
in
balloting,
in
the
irsg,
and
actually
we've
got
two
yes
votes,
so
it
should
be
approved
as
soon
as
we
revise
it,
but
you
know
so
the
key
takeaway
is,
you
know
persistent.
You
know.
Some
obstacles
are
persistent,
so
keep
your
eyes
open
next
slide,
please.
I
So
the
two
things
that
I
wanted
to
you
know
I
you
know
the
draft
or
the
rfc
is
great,
but
the
two
places
I
wanted
to
send
you
to
point
you
to
were
these
two
sections
in
the
rc.
I
One
is
a
summary
of
the
lessons
that
were
learned:
kind
of
high
level
summaries
for
each
lesson,
with
pointers
the
protocols
behind
each
lecture
and
the
second
one
is
how
to
you
know,
applying
the
lessons
that
we've
learned
especially
table
one
in
that
section,
which
is
basically
dividing
the
things
that
we've
learned
into
things
that
are
always
true
things
that
are
variable
that
you
might
be
able
to
engineer
your
way
around
and
some
that
are,
you
know
just
flat
out
research.
I
So
that's
that's
kind
of
that's
kind
of
where
that's
kind
of
where
that
the
work
that's
in
this
rfc
ended.
So,
basically,
now
it's
up
to
people
with
different
protocol
proposals
to
look
at
the
lessons
we've
learned
and
see
if
you
know
see
if
they're
tripping
over
problems
that
have
come
up
consistently
as
problems
in
the
past
jen
we
can.
We
can
click
on
those
links
if
we
wanted
to.
But
let
me
let's
do
your
last
slide.
I
H
The
important
slide,
probably
is
how
we
can
help
either
close.
We
looked
at
the
proposed
charter
and,
as
martin
said
already,
there
are
a
lot
of
overlap.
For
example,
when
charter
says
that
collect
implementation
deployment
experience,
it's
exactly
what
energy
loves
to
hear
what
worked
actually
as
an
operational
person.
I
love
to
hear
about
what
did
that
work
and
why
it's
normally
much
more
interesting
to
be
honest,
well,
protocol
maintenance
for
existing,
published
protocols.
H
I
think
it's
not
really
research,
it's
more
like
atf
scope,
but
the
next
charter
item,
like
development
operation,
support
tool
for
ida
protocol.
Well,
there
are
two
things
here:
high
level
discussion
about
how
to
support
and
what
your
experiences
is
just
perfectly
in
scope.
But
if
you
say
young,
I
will
be
running
away
screaming.
H
H
I
like
we
always
find
a
slot
for
you,
but
yeah
support
for
modern
existing
protocols.
Again,
everything
which
is
already
exist
probably
support
perspective.
Not
in
this
call,
but
everything
which
is
about
what
did
we
do
wrong
or
what
we
did
right
and
what
to
do
better
next
time.
Please
come
and
talk
to
us
and
I
think
it's
all.
I
have.
A
Okay
thanks
james
and
spencer,
so
we
yeah,
we
have
someone
in
the
queue.
C
C
H
H
Well,
if
you
think
about
this
research
and
experiment,
yes
right,
I
think
it's
a
kind
of
gray
area
of
overlap,
so
we
can
probably
talk
about
that.
Well
again,
it's
you
see
like,
as
I
say,
no
young
knows
this
low-level
stuff
probably
would
be
like.
Probably
it's
not
the
best
place
here
to
come
to
us
for
this.
C
I
Yes,
exactly,
I
I
think
if
I
could
just
dive
in,
we
can
do
experimental.
You
know
drafts
targeting
experimental
rc's
in
the
irtf
stream.
So
that's,
I
think,
you're
exactly
right,
adrian
and
you'll
notice
that
I
unmuted.
When
you
said,
I
don't
think
it's
possible
to
confuse
me
more
so
challenge
accepted.
B
All
right,
quick
question,
so
richard
can
continue
a
quick
comment.
It
seems
to
me
that
there
seems
to
be
some
synergy
between
pan-rg
and
alto.
At
least
alto
seems
to
me
looking
at
what
penrgy
is
doing
appears
to
give
you
a
standardized
way
to
represent
some
of
these
abstract
network
elements
and
other
properties
of
paths
that
you
need
in
pan
rg.
B
H
B
H
G
Richard
please
go
ahead.
Okay,
so
great
I
did
read
rc
90
49.
Is
that
right,
spencer,
9049?
I
remember
right,
so
I
I
think
I
missed
the
living
correct.
I
believe
that's
nine
zero.
Four
nine
and
I
like
it,
that's
a
very
nice,
very
well
written
rpc,
so
my
question
would
be:
would
there
be
increase
from
powerg?
Let's
take
a
look
at
all
those.
I
think
lessons
will
learn
and
apply
to
the
auto
current
design.
G
Let's
say
imagine
parity
as
some
kind
of
young
doctor
and
then
you
look
at
every
single
particle
as
some
kind
of
young
model.
You
want
to
identify
the
issues.
Would
you
be
interested
and
take
a
look
at
all
lessons
you
learned
and
let's
try
to
apply
to
auto.
Let's
say
how
in
which
sense,
also
how
to
do
the
right
in
which
is
how
to
get
it
wrong
and
therefore
how
to
potentially
fix
it.
Would
that
be
a
potential
topic
which
would
be
interesting
to
penalty.
H
Well,
I
will
definitely,
I
think
I
would
invite
myself
and
talk
to
the
group
about
this.
I
cannot
answer
for
the
group
right
because
come
and
present
it
and
we'll
ask
the
group
if
there
is
an
interest
right,
so
I
can
only
say
I
would
probably
give
you
a
slot,
but
it's
all
the
chair
could
do
right,
but
yeah.
I
think,
like
listening
about
hearing
about
experience,
what
went
well
and
what
went
wrong
and
what
can
we
do
better?
I
think
it's
always
a
good
thing
for
research.
F
So
I
wanted
to
take
a
shot
at
answering
adrian's
question.
My
view
is
an
ad
when
you
look
at
the
more
interesting,
ambitious
proposals
that
we
got
for
when
we
started
talking
about
rechartering.
I
think
there
are
two
pieces
to
any
like
alto
extension
right
there
is.
F
There
is
like
what
are
you
trying
to
do?
What
date
are
you
collecting
at
what
frequency
etc?
And
then
there's
the
grammar
of
how
you
express
that
in
alto
and
in
my
view,
that
that
first
question
about
what
you're
trying
to
do
like
how
much
data
you're
collecting
that
is
like
a
perfect
pan,
rg
presentation
right
and
whether
you
choose
to
do
that
in
alto
or
in
something
else,
that's
great.
I
I
I
think
this
audience
would
mostly
want
to
do
it
ronaldo,
but
to
go
out
and
say:
look
we
tried
deploying
this.
F
F
That
is
the
time
to
take
it
into
the
alter
working
group
and
develop
the
gra
and
finalize
the
grammar
of
how
that
is
done
in
alto,
like
working
groups
generally,
are
not
solving
very
difficult
theoretical
problems.
Right,
that's,
and
so
I
don't
know
if
that's
helpful,
but
that's
how
I
think
about
that
separation.
A
J
Yeah
yeah,
thank
you
just
some
thoughts
from
from
my
side
from
an
operator
perspective
looking
at
alto
and
panergy,
and
I
think
that
something
that
jane
has
committed
is
quite
important.
She
has
differentiated
between
the
how
and
the
what
no
and
I
think
that
the
alto
is
precisely
the
the
how
no
the
mechanism
for
exposing
a
capabilities
from
from
the
network
and
also
alto,
let's
say,
has
the
possibility
of
look
at
what
are
the
the
the
existing
machinery
on
idea
for
for
exposing
existing
information.
J
Let's
say,
while
panergy
probably
could
assist
on
exploring
new
capabilities
that
probably
are
not
available.
Yet
this
is
this
my
my
view
and
also
regarding
the
scope,
at
least
as
I
see
er
advantage,
I
I
see
alto
more,
let's
say
a
closing
to
the
limits
of
the
network,
so
the
infrastructure
itself,
with
the
purpose
of.
C
J
And
assisting
on
the
automation,
automation
in
the
service,
provisioning,
and
so
while
in
at
least
in
my
understanding,
energy
goes
exceeds,
let's
say
that's
boundaries,
and
also
it
looks
at
the
endpoints
and
with
these
ideas
of
selecting
protocols
and
so
the
probability
of
transport
protocols
or
evolution
of
the
transport.
I
mean
from
the
application
perspective
so
going
a
little
bit
beyond,
at
least
in
my
understanding
from
what
is
the
the
purpose
of
alto.
J
So
I
I
would
like
just
to
emphasize
this
that
probably
the
the
key
point
here
is
the:
how
versus
the
the
what
and
and
alto
would
be
the
mechanism
by
now
being
able
to
provide
what
are
the
existing
properties
or
capabilities
or
a
topological
information
that
we
have
in
the
network,
while
a
pan
energy
could
provide
extra
or
could
help
to
develop
new
new
contents,
new
capabilities
to
be
exposed
via
alto
in
in
the
future.
That's
only
my
comment.
Thank
you.
A
A
Okay,
let's
move
to
the
next
topic,
so
thanks
to
jan
and
spencer,
again,.
A
Okay,
so
we
capture
a
little
bit
about
recharge
discussion
history.
You
know
our
rechargeable
discussion
started
from
last
three
items,
especially
started
from
ie2108
and
one
zero
nine.
Actually,
we
have
a
lot
of
problems
from
operator
and
ott
and
such
as
the
tiny
bodyguard
testing
and
china
mobile,
and
we
based
on
the
discussion
we
come
up
with.
You,
know,
long-term
recharger
and
we
set
us
ago
a
little
bit
ambitious
ambitious.
Actually,
so
we
propose
a
lot
of
use
case
and
these
use
cases
actually
can
be
breakdown
into
two
categories.
A
A
Automation,
so
auto
you
know,
is
designed
as
a
you
know,
query
product
about
the
it
rely
on
some
auto
mapper
database
building,
actually
so
operation
automation,
really,
you
need
want
to
tackle
actually
how
to
get
the
data
from
different
data
sources
by
different
protocols
how
to
integrate
into
the
auto
mapper
database
and
also
for
multi-domain
setting
actually
also
original.
You
know
design
for
the
single
domain,
but
we
also
see
a
lot
of
other
user
cases.
A
For
example,
you
want
to
select
entry
and
pass
that
span
across
multiple
domain,
so
also
for
auto
deployment
we
already
have
rfc
has
been
published
and
we,
but
in
the
current
channel
we
see
a
lot
of
new
items
like
because
canada
performs
magical
pass
vector.
These
need
to
be
integrated
with
auto
deployment
rfc,
so
we
need
such
kind
of
enhancement
may
be
needed
for
new
worker
build.
On
top
of
the
auto
protocol.
We
have
some
of
a
new
proposal
like
a
movie
actually
and
a
cellular
related
worker.
A
A
Do
we
do
in
the
last
itm
meeting
when
we
have
the
recharge
discussion?
We
actually
we
have
too
many
use
cases
really
distracted.
You
know
people,
and
so
we
don't
know
those.
Many
people
actually
don't
understand.
You
know
which
use
cases
should
be
high
priority
and
which
one
should
a
research
base
based.
Actually
you
know
money
also,
you
know
confirm
for
this
so,
but
we
also
see
a
lot
of
you
know
interesting
activity
on
some
of
you
know
k
topic
for
operation
automation.
A
You
know
one
of
our
problems
actually
from
anybody
car
they
really
want
to.
You
know:
focus
on
the
cdn
service
delivery,
use
cases
they
want
to
want
to
integrate
overlay
and
on
delay
and
to
provide
the
better
the
service
delivery
for
the
city
and
service
in
the
multi-domain
setting.
We
see
a
bunch
of
the
money
domain
related
job
that
that
has
been
posed
to
the
old
working
group,
although
they
get
inspired,
but
we
see
a
lot
of
interest.
We
also
see
a
lot
of
research
paper.
A
You
know
related
to
this
money
domain
setting
in
the
movie.
We
see
a
lot
of
discussion
and
a
lot
of
energy
over
there
actually,
but
this
topic
actually
also
relates
to
what
3gb
is
doing.
Actually
we
we
see
the
problem
is
like
a
good
collaboration
with
gpp
and
so
because
it
crosses
many
areas.
So
this
kind
of
topic
may
be
broken
down
into
different
pieces
and
some
pieces.
A
A
Actually
we
come
up
with
this
term
recharger,
which
has
already
been
posted
to
the
list,
and
we
also
have
a
weekly
to
track
this
recharge
and
the
the
the
motivation
the
goal.
Actually
really.
You
know
to
build
a
solid
basis
for
the
auto
protocol
to
allow
us
to
work
on
the
immediate
issue,
such
as
protocol
maintenance,
operation
support.
But
in
addition,
actually
we
also
encourage
people.
You
know
to
look
into
some
new
emerging
services,
new
use
cases
that
we
can
discuss,
investigate
some
best
uk
that
can
fit
for
the
future
work.
A
A
Next,
so
I
don't
want
to
repeat
this
actually,
this
you
know
author
in
recharter
stage.
Actually
martin
has
already
sent
this
for
isg
review
and
ib
review,
and
the
chat
has
all
been
narrowed
down
and
next.
A
So
to
to
to
discuss
our
new
new
chat
with
first,
let's
revisit
the
current
charter,
the
current
chapter,
you
know
the
foundation
is
all
auto
based
protocol
and
auto
service
discovery
and
auto
deployment
based
on
this
foundation.
Actually,
a
set
of
the
volcano
has
to
be
proposed
for
the
current
charter,
such
as
incremental
update,
which
allow
you
know
you
support
the
server
initiator
notification
and
a
partial
update,
and
also
we
have
a
auto
server
service
discovery
service
server,
discovery
max
that
has
been
developed
by
the
auto.
A
Actually,
we
have
two
offices
related
to
this
work
item
and,
in
addition,
actually
we
extend
try
to
extend
auto
protocol
to
convey
some
attributes
to
allow
application
to
know
where
to
connect
and
when
to
connect.
So
we
have
a
course,
canada.
Actually,
this
you
know
in
in
the
current
charter-
and
this
is
working
in
progress
and
we
have
the
past
vector,
also
work
in
progress.
A
A
So
take
a
look
at
the
new
charter
proposal,
so
auto
protocol
is
originally
designed
for
the
pdp
application,
cd
application.
I
think
we
should
continue
to
support
the
cdn
application,
but
we'll
also
need
to
look
into
some
new
use
cases
such
as
computing
aware
networking
as
computing,
and
then
we
we
think
to
support
some
future
deployment
of
future
use
cases.
Actually
we,
the
working
group,
actually,
should
you
know,
look
into
some
new
activity,
for
example.
A
First,
actually
we
need
to
connect
implementation,
develop
deployment
and
experience
to
help
us
understand
where
the
author
gets
deployed,
what
is
issue
and
and
where
the
auto
protocol
gets
bricked,
and
so
this
is
something
we
really
want
to.
You
know
tackle
and
we
this
can
be
documented
in
some
wikipedia
or
in
in
the
internet.
Draft
for
second
work
item
actually
protocol
maintenance
for
existing
publishing
protocol,
and
I
think
this
is
a
really
you
know-
help
you
to
identify
some
back
or
some
issue
you
can
really
see.
A
A
One
example
you
can,
you
know,
define
the
om
mechanism
or
auto
young
model
you
to
have
to
configure
the
outer
server
and
also
for
all
the
protocol
is
built
on
top
of
the
h2d1
now
actually
http,
2
and
ud3,
actually,
as
a
new
transport,
is
very
popular,
so
how
to
migrate
to
the
http
2
and
3
is
any
gap,
so
this
is
something
we
need
to
investigate
as
well.
A
In
addition,
actually
for
some
of
new
use
cases,
actually
we
think
we
still
need
more
discussion
and
get
it
solid,
so
we
allow,
you
know,
provide
such
a
venue
for
people
to
discuss
these
future
use
cases.
A
A
So
this
is
a
milestone
and
deliver
deliverables
that
we
provide.
We
actually
have
four
deliverables.
Actually
one
is
we
do
the
survey
for
the
auto
implementation
and
deployment
experience.
The
second
is
that
we
may
consider
to
develop
a
standardized
at
least
one
om
maximum
to
support,
auto
and
also.
A
third
item
is
auto
over
the
http
2
and
s3,
and
we
need
to
investigate
the
to
to
try
to
identify
the
gap
and
last
actually,
if
we,
we
also
encourage
other
participants
actually
to
come
up
with
some
new
use
cases.
A
So
this
is
a
table
actually
for
the
new
work
item
document.
Actually,
we
have
four
work
items
actually
for
the
first
implementation
report.
We
already
created
the
weekly
page
to
check
the
implementation.
I
I
see-
actually,
you
know
actually
have
decreases
wikipedia
and
solicit
some
input
from
the
auto
community
and
we
have
some
progress
actually
and
for
om
maxims
to
support
the
auto
actually
and
we
have
a
new
job
to
be
being
proposed.
A
Actually,
I
think
this
is
still
an
individual
job
and
need
more
discussion,
and
we,
we
think
you
know
this
head
towards
the
right
direction
and
for
auto
over
http,
2
and
3,
and
we
also
have
relevant
jobs
to
discuss
this,
and
so
also
can
provide
update
for
this
kind
of
work
and
use
cases.
Actually
we
have
a
bunch
of
use
cases
so
so,
for
all
these
use,
clicks
can
be
broken
down
into
two
categories:
new
use
cases
and
updated
use
cases
for
new
use
cases.
A
You
know
we
keep
on
detecting
the
money
domain
use
case
and
movie
cellular
use
cases,
so
this
money
domain
get
updated,
and
so
we
can
see
actually
where
we
are,
and
so
that's
it
and
I
pause
here
to
sort
of
see
the
feedback
and
input
for
this
recharge
contents.
A
F
F
We
may
end
up
taking
out
the
text,
but
that
would
still
remain
something
we
want
the
working
group
to
continue
to
address,
but
yeah
I
mean.
Typically,
a
charter
is
really
just
focusing
on
the
document
output,
which
would
be
the
I
guess,
the
two,
the
two
medium,
the
second
and
third
work
items
here,
but
there's
a
special
emphasis
compared
to
other
working
groups
on
this
station
deployment
experience.
F
Just
because
of,
I
think,
where
the
state
of
this
the
perceptions
of
where
alto
is
in
the
internet,
and
if
it's
actually
something
that's,
that's
it's
worth
ietf
resources
to
extend
it.
No
one
else
is
in
queue.
So
I'll
keep
talking
the
see
the
the
new
use
cases.
F
F
It
would
be
really
nice
if
we
had
if
there
were
if
there
are
designs
behind
these
use
cases
to
go,
have
some
pocs
and
have
you
know,
actual
service
providers
or
applications?
F
Well,
actually,
both
say
yes,
this
is
actually
really
something
that's
promising
and
we
want,
and
we
want
the
ietf
to
to
fix
the
grammar
and
make
it
make
it.
You
know,
give
it
the
proper
review
and
then
we
will
write
write.
A
big
check
to
deploy
this
at
scale,
like
the
cdni
case,
for
example,
is,
is
an
interesting
one.
F
I
mean
this
is
the
first
case
of
that
right,
as
I
understand
it,
alto
was
originally
written
for
peer-to-peer
and
we
went
and
we
did
we
said
well,
you
know
it
didn't
really
take
off
peer
peer,
here's,
here's
the
cdn
use
case,
so
that
was
an
interesting
speculative
thing
and
if
a
bunch
of
cdns
go
and
pick
it
up,
then
we've
been
very
successful,
but
I
would
have
been
better
in
my
opinion,
if
we
had
a
bunch
of
cdn
show
up
here
and
say
yes,
we
actually
really
are
very
interested
in
this
and
would
like
to
see
this
as
a
standard,
because
we
were
going
to
deploy
it
at
scale
just
to
question.
F
B
F
F
Right
so
that's
true,
I
mean
I
don't
I
don't
mean
to
cast
dispersions
on
the
work
on
alto.
Like
I
don't
I
don't
think,
there's
anything
wrong
with
the
work,
but,
regardless
of
why
just
it
is
not,
the
work
is
not
is
whether
it's
a
fault
of
the
protocol
or
of
the
use
case
itself
and
its
failure
like
it
it.
It
has
not
been
impactful
in
that
sense,
and
that
is
not
where
we
want
to
be.
B
That
you're
casting
dispersions,
just
as
a
as
a
point
of
just
talking.
I
think
the
peer-to-peer
use
case
was
very
interesting
when
we
started
developing
it,
but
the.
F
F
Right,
yes,
thank
you
for
the
clarification
I
I
that
is
helpful.
G
Yeah,
I
do
have
a
question
to
martin
and
I
think
my
understanding
is,
let's
really
make
sure
all
the
use
cases
will
be
somehow
backed
up
by
by
deployment
and,
for
example,
one
use
case
here
is
the
multi-domain
use
case,
and
I
don't
know
if
you
read
on
the
paper
and
of
course
maybe
we
shouldn't
really
only
read
a
paper.
We
should
really
get
all
these
people
showing
up.
For
example,
there
is
a
ben
knox
deployment,
a
paper
from
connex.
I
think
that's
the
best
paper
of
2019.
G
They
also
got
the
arp
best
applied,
networking,
research,
price
and
and
when
I
was
talking
to
them-
and
they
mentioned
that
at
least
for
their
deployment
case.
Multi-Domain
was
a
very
important
use.
Of
course.
They
only
mentioned
that
one.
I
think
they,
unfortunately
they
do
not
show
up
in
ietf.
So
my
understanding
just
to
clarify
is
one
say:
okay,
that's
a
real
use
case.
That's
really
driven
by
real
case.
It's
got
all
the
people,
for
example,
from
ben
knox,
from
all
their
users
to
go
to
the
ietf.
That's
a
way
to
demonstrate.
F
I
don't
want
to
impose,
for
I
don't
want
to
make
it
so
that
people
have
to
like
fly
to
madrid
or
whatever,
but
I
do
want
to
distinguish
between,
like
the
research
arm
of
a
certain
organization
being
interested
in
poc
and
actually
like
the
real
dollars
being
ready
to
be
thrown
at
this,
and
you
know
I
don't
have
a
hard
and
fast
rule
on
when
we've
crossed
that
line,
but
like
an
example,
something
we
could
see
is
I
know
that
for
the
maui
in
particular,
there
was
talk
about
this
solving
a
3gpp
use
case,
there's
a
liaison
statement
from
dgpv.
F
Saying
yes
like
we
have
this
problem
and
we
think
alto
is
the
solution.
That
would
be
a
powerful
thing
if
there
were
actually
like
deployment.
So
if
there
were,
if
there
was
actual,
if
there
was
an
actual
deployment
that
you
know
had
some
rough
edges
based
off
an
early,
I
mean
these
individual
drafts
obviously
have
some
have
some
solutions
in
them.
If
people
have
like
run
those
solutions
and
there's
been
a
large
experiment
and
people
say
yes,
this
has
been
a
promising
experiment.
Can
you
please
standardize
this?
F
That
would
be,
that
would
be,
I
mean
it
doesn't
have
to
be
them
showing
up.
It
can
be
an
email
it
can
be,
it
can
be
a
liaison
statement.
There
are
a
lot
of
things
it
can
be,
but
yeah
like
just
something
beyond
just
research
interest
is
really
what
I'm
looking
for.
J
Oh,
thank
you
just
to
highlight
one
one
point
that
you
martin
has
said
that
which
is
the
deployment
at
the
scale.
I
think
this
is
a
very,
very
key
point,
and
I
think
that
the
the
issue
is
probably
that
we
are
now
in
the
momentum
where
all
the
these
capabilities
for
automation
also
are
taking.
Let's
say
that
the
moment
of
being
deployed
into
the
operator's
network,
with
all
the
all
the
artifacts
from
itf,
with
the
jam
models,
the
netconf
protocol,
the
sdn,
the
pc.
J
So
it's
probably
now
the
momentum
where
probably
these
deployments
at
the
scale
also
could
I
mean
alto-
could
be
also
part
of
this
all
this
deployment
of
at
the
scale
and
all
the
automation,
and
so
so
a
few
words
about
the
case
of
cdn
that
you
have
mentioned.
J
We
can
differentiate
here,
the
internal
cdns,
the
ones
that
are
owned
by
the
operator,
and
that
could
be,
for
instance,
with
the
team
mentioned
before
the
the
proof
of
concept
that
we
are
working
out
in
in
telefonica.
J
This
creates
a
new
momentum
for
alto
and
even
alto
is
being
considered
for
a
couple
of
drafts
for
one
regarding
the
footprint
of
the
cdns
and
another
one
for
for
capacity.
So
this
I
mean
now
is
where
we
are
reaching
the
momentum
of
squeezing
and
taking
profit
of
all
these
capabilities
because
is
linked
with
this
overall
momentum
of
automation
and
so
that
we
are
foreseen,
and
this
will
happen
and
it's
clear
that
the
case
of
the
cloud
computing
edge
computing.
This
will
happen
in
in
the
in
the
next
future.
J
So
this
is
are
the
kind
of
of
use
cases
that
will
require
of
that
automation,
capabilities
that
probably
before,
could
not
be
exploited
as
as
much
as
we
could
display
today
with
all
these
artifacts,
the
overall
artifacts
of
an
itf
models,
netcon
the
programmability
or
programmatic
interfaces,
and
so
on
so
far.
Thank
you.
J
A
F
No
I'll
let
people
talk,
and
maybe
we
can
just
talk
about
this
on
the
list,
but
I
would
like
other
people
to
have
a
chance
to
comment.
A
K
Okay,
hello:
this
is
chancellor
from
china
mobile.
I
want
to
make
some
explanation
on
the
network.
Topology
exposed
skates
for
cdn
by
auto
is
the
same
mixes
of
noto
server.
To
choose
better
sydney
surrogate,
however,
our
city
may
not
be
able
to
passively
listen
to
rotting.
Protocol
normally
have
access
to
other
network
topology
data
so
based
on
our
analysis,
student
is
built
on
top
of
an
underlying
network,
which
runs
this
proto
routing
protocols
and
the
and
the
network
topology
can
can
be
gathered
via
bgp.
Our
airs
and
network.
K
We
are
wondering
whether
we
can
use
netcom
protocol
to
collect
this
network
data
and
translate
into
auto
automap
data.
So
in
this
case,
we
we
we,
we
want
to
integrate
your
model
natural
voltage
data
into
auto
map
database
and
to
translate
network
quality
mode
with
young
into
auto
map
data,
so
I
think
we
think
that
auto
can
help
to
to
to
to
improve
the
the
civilian
delivery
path
in
this
case.
L
As
we
proposed
the
draft
of
computing
aware
network
use
case
and
since
the
computing
related
routing
has
under
other
issues,
has
attracted
attention
and
we
think
that
autumn
may
do
something
combined
with
it
and
on
the
one
hand,
just
like
the
draft
of
lewis
about
the
choice
of
about
deployment
of
the
edge
service
and
on
the
other
hand,
it
may
also
help
to
select
the
nodes
of
service.
So
we
just
propose
the
this
use
case
to
find
out
that,
if
auto
has
the
potential
to
do
some
work,
yeah,
okay,.
M
Oh
I'm
moving
around
from
telecom
and
I
want
to
introduce
the
air
map
measurement
exposure
use
case
briefly,
because
we
are
as
network
operators.
M
We
always
want
to
better
know
what
happened
to
our
network
and
the
air
map
framework
defect
in
ietf
provide
a
good
tool
for
us
to
collect
the
network
measurement
results.
M
But
traditionally
this
measurement
results
is
a
cell
with
a
file
which
is
not
common
convenient
and
our
consideration
is
to
use
auto
protocol
to
retrieve
potential
aggregated
network
performance
measurement
results
for
a
certain
network,
and
in
this
way
the
type
of
aggregation
can
be
selected
as
needs
as
needed
to
get
the
desired
desired
measurement
results.
M
And
so
this
is
the
motivation
to
proposed
the
air
map
use
case
and
well
always
welcome
the
comments
and
input
thanks.
A
N
Yes,
I
want
to
do
some
explanations
about
the
computing
aware
use
case.
We
think
that
the
collaborative
scheduling
of
the
computing
and
network
resources
requires
the
discovery
of
the
deployed
network
resources
and
also
computing
resources.
N
So
we
hope
that
to
extend
the
auto
to
support
the
queries
of
the
computer
resources
and
network
resources
and
also,
as
mentioned
in
the
draft,
including
the
aesthetic
information
and
also
dynamic
information
and
everything
that
auto
can
assist
in
deciding
which,
which
route
to
choose
and
also
which
computing
node
to
choose.
That's
for
the
explanation.
Thanks.
A
Okay,
thank
you.
So
we
need
to
wrap
up.
Actually,
I
think
in
this
meeting.
Actually
we
here
could
talk.
You
know
from
the
pid
as
we
see
actually
good
synergy
with.
You
know
pid.
The
pig
will,
you
know
happen
after
auto
meeting
if
you,
paris,
even
actually
interested
in
the
pig
topic
of
failure
to
attend,
and
so
so
bj
yen.
Do
you
have
any
anything
to
add
here.
F
Else,
thank
you
chin
for
driving
this,
and
I
think
the
other
thing
should-
and
I
I
do
want
I
know
we're
over,
but
I
do
think
we
should
take
at
least
a
moment
to
recognize
vj,
who
has
been
a
chair
since
at
least
2013,
although
maybe
earlier,
I
I
got
confused
on
this
in
earlier
conversation
and
he
is,
he
is
stepping
down.
After
all,
his
work
against
the
the
many
specs
that
that
are
now
on
my
desk
and
you
know
they're
good
specs.
F
So
thank
you
for
that,
and
so
you
know
as
as
we
wind
down
that
work.
Vijay
is
stepping
away
chin,
of
course,
is
here
now
driving
things,
and
I
think
if,
if
we
get
the
charter
done
and
through
then
we'll
allow
to
to
break
free
as
well
and
give
chin
a
new
chair
with
with
fresh
energy,
a
new
co-chair
with
fresh
energy.
So
thank
you
vijay
for
everything
you've
done
for
alto
and
for
the
itf
and
we
hope
to
see
you
at
the
next
time.
B
Yes,
absolutely
thank
you
all
for
the
opportunity
and
it's
been
a
great
pleasure.
Thank
you
very
much.