►
From YouTube: IETF-MPLS-20230525-1400
Description
MPLS meeting session at IETF
2023/05/25 1400
https://datatracker.ietf.org/meeting//proceedings/
C
C
C
Okay,
sending
a
keep
alive
so
again,
thanks
for
whoever
joined
and
we
were
waiting
a
little
bit
more
to
see
more
participants,
the
usual
culprits
or
the
people
that
help
usually
in
the
discussions.
C
C
Okay,
I'm
gonna
go
to
sleep
for
one
more
minute
and
then
come
back
on
the
mic,
hopefully
by
then
Lowa
is
back
and
I
didn't
say
hi
to
Nick.
So
if
you
can
hear
me
hello
Nick,
you
can
acknowledge
that
everything's
fine
on
your
side,
I.
E
E
C
Hello,
yes,
now
you're
back,
we
were
I
was
trying
to
reach
out
to
you
and
you
probably
was,
and
now
you
went
offline.
C
C
Okay,
Nick
I
guess
you
still
can
hear
me.
C
Okay,
I
was
gonna,
ask
you
if
you
know
we
can
get
started
any
objections.
Nick
and
it's
got
maybe
through
way
lower,
can
I
get
his
technical
problem
solved
and
then
join
in
okay,
so
you
you're
you're,
okay
with
that
I
guess.
So
let
me
reiterate
my
gratitude.
Thank
you
for
joining
the
first
M
A
interim
meeting.
C
This
is
one
of
a
series
of
meetings.
Loa
feel
free
to
interrupt.
If,
when,
if
you
want
to
test
your
audio
again
so
I
don't
mind,
so
this
is
a
a
first
of
a
series
of
interims
that
we
scheduled.
This
is
a
work
and
collaboration
between
three
working
groups:
mpls
files
and
net.
C
We
have
a
number
of
chairs
I
think
today
and
the
number
of
secretaries
as
well
Mac
and
David
are
here
as
well.
Well,
actually
he's
yeah
he's
listed,
yeah,
I,
guess
yeah
Secretary
of
both
thousands
that
the
usual
usually
we're
in
in
our
previous
M
A
meetings.
We,
the
attendees,
where
you
know
the
numbers
were
hovering
around
15
participants.
C
Today
we're
a
little
bit
lower
than
this
I'm,
hoping
it's
not
a
confusion
caused
by
our
switching
into
using
meat
Echo.
So
this
is
more
of
a
of
a
formal
way
of
doing.
The
meetings
that
we
were
carrying
out
meet
Echo
will
will
offer
you
know
the
recordings
they've
done
automatically
and
I'm
gonna
go
through
the
tools
that
we
usually
use
in
ietf.
C
C
Some
useful
pointers
we
are
using
meet
Echo
today
and
we
will
continue
to
use
meet
echo
in
series
if
we
continue
to
hold
this,
we
are
taking
minutes
using
the
usual
tool
at
the
address.
Url
I,
left
and
I
will
try
to
post
this
as
well
in
the
chat.
So
people
have
the
opportunity
to
contribute
in
taking
minutes.
C
Right
now
it's
in
the
chat,
so
please,
if
you
can
log
in
and
contribute
to
taking
the
minutes.
C
Okay,
today's
agenda:
we
have
a
number
of
things
that
we
wanted
to
talk
about:
I'm
going
to
be
walking
with
with
Loa
that
the
plan
was
to
have
lower
and
myself
go
over
things.
C
I
can
hear
you
good,
not
good,
so
we
have
the
the
agenda
is
on
two
slides.
So
it's
a
number
of
items
there,
but
I'll
leave
a
chance,
for
you
know
I
will.
If
anybody
wants
to
bash
the
first
slide,
it's
it's
your
chance.
C
We
have
an
all
other.
You
know
any
other
business
at
the
end.
So
if
any
one
wants
to
add
an
item,
that's
not
listed
on
these
two
slides.
It's
your
chance
also
to
suggest
that.
C
I
know
people
started
to
question:
do
we
need
the
series
or
so
feel
free
to
speak
up
or
when
we're
going
when
we
move
on
to
the
second
item-
and
we
can
definitely
have
a
fruitful
discussion
about
that,
we
do
keep
a
track
of
action
items
so
I'll
I'll
have
a
you
know
more
pointers
on
that
in
one
slide
and
there's
a
discussion
on
the
first
nibble
and
some
polls
that
we
might
be
preparing
and
coming
up,
that's
number
five
number
six.
C
We
want
to
have
a
good
discussion
or
a
decision
on
whether
we
need
a
post
stack
data
or
not.
And
finally,
let's
see
what
we
can
record
for
for
next
meeting
in
terms
of
action
items
or
topics
that
we
want
to
talk
about
and
the
last
one
is
any
other
business.
That's
how
we
track
it
and
I
don't
see
anyone
in
the
queue
as
of
now.
C
A
The
first
thing
and
I
would
like
people
to
actually
use
the
raised
hand,
so
we
can
actually
keep
a
queue
in
the
discussion.
Please
use
that
sometimes
it's
hard
for
some
people
to
actually
get
to
the
top
of
your
queue
and
actually
say
what
they
want.
A
We've
done
the
agenda
bashing
and
that's
a
thing.
It
comes
with
the
formal
ipf
meeting
we
talking
about
note-taking.
There
is
a
tool
and
we
should
use
that
tool.
A
It's
also
on
the
top
of
the
meet
Echo
page
with
the
pen
on
a
white
on
on
a
paper
go
in
there
and
read:
what's
this
written
and
correct
things
that
you
think
are
wrong
and
actually
add
things
you
need
I,
think
I'm,
missing,
Andrew
or
A.D
said
that
the
notes
are
more
important
now
than
there
was
before
and
I
think
it
means
from
a
formal
point
of
view.
A
A
Nope,
okay,
so
this
is
still
a
joint
meeting
with
mpls
paths
and
that
that
yes,
so
I
have
that
ticked
off.
I
made
a
mistake
on
the
first
slide.
David
is
not
the
secretary
for
that
net.
It's
Ethan
and
we
will
be
changed
that
on
the
slide,
the
working
Group
shares
will
continue
to
meet
I
guess
on
Tuesdays,
and
we
will
have
the
regular
meetings
on
of
the
interim
myths
and
thoughts
that
they
are
booked
all
the
way
up
to.
As
far
as
we
can
before
iitf
fund
17.
A
I
guess
I
jumped
the
action
item.
Terry
will
talk
about
that
in
a
minute
and
we're
going
to
be
the
week.
Keeper
still
be
used
for
document
things
and
it
will
be
updated.
I
I
want
I
will
check
what
I
need
to
do
about
it,
but
I
haven't
done
that
yet
and
that's.
A
I
think
one
more
thing:
I
actually
booked
two
hours
for
each
meeting
and
I,
don't
think
we
normally
will
use
all
both
hours,
but
we
go
as
long
as
we
have
things
talk
about,
and
then
we
just
stop
and
have
the
next
meeting
in
the
week.
A
C
Yeah
yeah,
so
we've
been
talking
about
action
items,
so
this
is
something
we've
been
doing
in
the
past
meetings
and
we
were
keeping
track
of
those
on
a
on
a
Wiki,
and
we
will
continue
to
do
that,
so
the
wiki
will
be
updated
to
have
a
new
page
for
these
meetings.
You
know,
if
anything
comes
out
of
them,
we
don't
have.
C
At
the
moment
an
active
I
was
checking
the
the
open
I
action
items
and,
except
for
yeah
I
I,
don't
see
any
active
action
item,
but
you
know
when
what
we
discussed
today
will
be
recorded
on
a
new
page.
C
We
want
to
remind
owners
of
action
Islands
to
come
prepared
with
a
report
or
status
on
anything,
that's
assigned
to
them
and
we
will
discuss
these
action
items.
Obviously
in
the
interim
meetings.
So
that's
not
much,
you
know,
but
a
reminder.
C
Okay,
I'll
move
on
the
first
technical
discussion
we
will
have
is
about
first
nibble
and
again
Loa.
Please
go
ahead
and
talk
about
it.
A
A
The
the
point
is
that
we
haven't
really
converged
so
the
results
are
not
conclusive
and
we
have
had
a
handful
of
people
cannot
stating
their
opinions,
but
we
were
actually
we
wouldn't
need
to
have
more
and
we
will
come
back
to
this
one
way
or
another
that
I
don't
know
how.
So
we
don't
have
a
session
decision
on
the
first
nibble
Paul
as
of
now,
and
we
need
to
fix
it
some
way,
questions.
A
A
A
C
Yes,
okay,
I
I
went
in
before
Greg,
but
you
know,
let
me
know
if
Loa
ready.
C
Yeah
yeah
well
I
just
want
to
say
that
discussions
on
the
poll
hovered
between
reap
reusing,
the
zero
value,
zero
for
the
first
nibble
and
a
in
a
new
value
that
gets
assigned
specifically
for
m
a
I'm,
not
adopting
any
one
of
them.
But
just
you
know,
sharing
the
results
of
what
I've
seen.
So
maybe
it
opens
up
the
more
discussion,
but
the
these
were
the
two
pro
two
two
kind
of
feedbacks
we
got
from.
A
F
So
it
appears
to
me
that
the
first
nibble
is
not
important
for
m
a
in
general,
but
for
PSD
m
a
so
would.
First,
we
need
to
decide
whether
we
have
any
viable
solution
for
PSD
and
then
decision
on
what
the
value
would
be
for.
First
nibble
PSD,
M
A
to
use
will
be
a
a
consequence
or
part
of
this
decision.
So
as
I
understand,
ISD
m
a
which
we
adopted
solution
for
as
a
mpls
working
group
does.
D
E
A
Actually,
we
are
going
to
talk
about
the
PSD
decision
a
little
bit
later
in
this
meeting,
so
I
think
Greg.
Can
you
come
back
with
kind
of
adding
that
this
actually
is
relevant
for
the
first
nibble
discussion,
when
we
do
that
yeah
of.
F
C
Okay,
I
will
move
on
and
it's
a
different
topic
now
and
you're
still
on
on
this
go
ahead.
A
You
can
hear
me:
yes,
the
so
label
14
has
a
long
history.
The
RC
was
written
in
202
I.
Think
it's
an
informational
Roc.
It
doesn't
really
follow
the
procedures
around
assigning
a
newbie
bsp
else,
but
at
that
time
those
procedures
were
not
really
set
in
stone.
So
we
got
it
to
say
a
draft
coming
in
pretty
much
from
iqt
asking
for
an
oem
alert
label
and
there
was
an
RC
written
from
informational
one
and
we
assigned
the
alert
label.
A
At
that
time
there
were
about
five
or
six.
It's
a
little
bit
unclear
how
many
it
actually
were
because
I
think
the
two
of
them
had
responded.
Two
of
the
six
that
responded
actually
worked
for
the
same
company
so
and
they
said
they
had
products,
and
then
they
were
marketing
products
with
the
label
14
using
the
label
14..
A
So
what
I
did
was
preparing
a
poll
to
be
sent
to
the
working
group
suggesting
that
we
retrieve
label
14
and
actually
deprecate
the
ERC
and
when
the
shares
reviewed
it,
we
got
a
comment
that
we
actually
need
to
send
a
liaison
to
the
study
15.
It
was
started
with
13
at
that
time.
It's
now,
starting
with
15
that
owns,
owes
that
part
of
the
technology.
A
C
You,
okay,
I,
have
a
question
lower
and
or
maybe
clarification
rather
than
question
so
the
retrieval
of
label
14
is
not
directly
tied
to
the
m,
a
work
or
even
the
PSD
in
specific
in
m
a
we're
just
trying
to
bring
it
or
retrieve
it
and
make
it
available
for
the
next
next
application.
A
That's
I
think
it's
true,
but
there
is
one
aspect
that
it
actually
makes
it.
If
we
get
one
label
back,
it
makes
it
easier
to
actually
assign
a
new
one,
a
new
label
so
and
since
we
actually
need
to
assign
it
looks
like
we
have
to
assign
an
m
a11.
So
it's
eases
up
a
bit
if
we
can
get
label
fork
and
back
and
that's
the
connection
we
have
now.
E
C
So
this
is
a
slide
on
the
PSD
discussion
and
it's
the
chance
to
talk
more
about
it.
So
Laura
go
ahead
again
still.
A
Where
we
said
that
we
need
both
ISD
and
PSD,
based
on
that,
we
actually
started
to
push
two
drafts
through
the
working
group.
We
did
a
working
group
adoption
poll
on
both
the
draft
Jackson
draft
song.
The
draft
Jacks
did
go
ahead
with
the
exchange
that
it
only
only
addresses
the
ISD
part.
So
that
part,
the
ISD
parts
are
kind
of
kind
of
clear,
but
the
decision
is
still
that
we
need
PSD
and
we
have
a
proposal
that
we
need
to
take
care
of,
that.
A
The
PSD
is
not
necessary
and
we
should
get
rid
of
it.
And
what
we
have
on
the
list
here
is
that
we
a
number
of
questions
where
how
how
the
chairs
want
to
actually
derive
this
discussion
and
it's
kind
of
a
waterfall.
It
starts
with
number
one,
and
only
if
you
say
yes-
and
we
have
sufficient
use
cases
to
motivate
PSD,
then
it's
worse
going
ahead.
Take
trying
to
address
the
other
questions,
and
so
we
need
to
take
them
in
order
one
two,
three
four
five
and
then
we
dig
through
the
list.
A
A
I
should
say.
Thank
you
to
David
is
probably
most.
A
Your
your
text
on
the
questions,
but
the
Tariq
and
I
scrubbed
a
little
bit
yesterday,
but
we
I
don't
think
we
made
any
any
real
changes.
It
was
clarification.
C
Yeah
the
Joel
you're
at
the
top
of
the
queue
so
go
ahead.
D
Thank
you
two
points.
First
I
found
the
characterization
of
the
earlier
discussion
on
the
mailing
list
and
its
conclusion
somewhat
odd,
because
the
question
when
you
asked
it
quite
some
time
ago
and
it-
and
it
was
a
very
reasonable
question
at
the
time-
was
not
do
we
need
only
ISD.
The
question
was
very
St
specifically:
do
we
need
only
PSD,
or
should
we
go
with
both
ISD
and
PST?
D
Doesn't
answer
this
question
and
I
at
least
in
looking
at
the
use
cases
have
not
seen
enough
use
cases
to
justify
the
complexity.
Yes,
there
are
some
use
cases
Now
Grant
there
are,
but
PSD
introduces
a
lot
of
complexity
and
I
haven't
seen
enough
to
justify
it
and
at
the
same
time
there
doesn't
even
seem
to
be
an
agreement
as
to
which
use
cases
are
actually
important.
D
C
C
So
one
use
case
was
you
know,
for
example,
and
not
not
only
it's
just
an
example
is
getting
timestamps.
You
know
that
you
get
stand
that
gets
stamped
as
the
packet
is
traversing
a
path.
So
in
this
case
you
know
the
timestamps
you
know
get
stamped
on
in
the
packet
and
things
in
the
in
the
stack
get
the
you
know
get
changed
the
stack
it
becomes
mutable.
C
There
are
other
proposals
not
to
do
the
the
time
stamping
inside
the
packet
I
understand
that,
but
in
general
my
point
is:
if
the
data
is
mutable,
putting
it
in
stock.
Might
have
concern.
F
F
So
now
we
have
an
agreement
and
adopted
the
solution
for
ISD
and
my
understanding
of
the
solution
that
it
does
not
preclude
future
development
of
ESD
I
agree
with
tarek
that
there
might
be
cases
theoretically
or
in
the
future.
That
would
benefit
from
a
PSD
metadata
being
posted.
F
F
M
A
use
cases
that
mandate
the
use
postal
data
cannot
be
addressed
and
support
it
using
the
adapted
ISD
solution
for
m
a
so
we
might
go
with
the
exercise
of
reviewing
these
use
cases
and
any
other
practical
use
cases
that
we
know
of
to
see
if
any
of
it
does
require
PSD
and
cannot
be
solved
with
ISD.
So
again,
as
I
remember,
the
discussion
of
PhD
scenarios
that
brought
significant
complexity.
D
B
F
So
with
we
adopted,
ISD
solution
and
PhD
is
being
separated.
So
again,
even
when
we
have
a
possible
PSD
solution,
then
of
course
that
can
be
and
should
be
discussed,
but
I
don't
see
that
there
is
a
dependency
between
progressing
the
ISD
Mna
solution
and
PSD
yeah
discussion.
Thank
you.
C
G
Okay,
I
think
they
have
had
long
time
discussion
about
the
pros
and
cons
of
PSD
in
ISD,
and
we
have
a
lot
of
understanding
about.
They
are
not
competing
solutions
that
separate
components
of
the
complete
solution.
Some
of
the
functionality
can
be
provided
by
ISD
and
some
other
functionality
requires
PSD
and
I.
Think
Tarik
just
gave
an
example
that
something
put
in
the
ISD
will
cause
some
problem,
and
another
point
is
for
the
ISD,
since
the
limitation
of
the
size
of
the
ISD
space.
G
Regarding
the
use
cases,
I
think
we
have
discussed
some
of
the
potential
use
cases
related
to
the
PSD
like
the
iom
and
also
the
that
net
may
be
another
possible
use
case
and
regarding
the
complexity,
I
think
I
would
have
had
the
discussion
about
it.
The
complexity
in
the
meetings
and
also
on
the
list,
both
ISD
and
PSD,
will
introduce
additional
complexity
to
the
empires
technology,
and
if
we
want
to
improve,
simplify
the
solution,
we
can
we
need
to
look
at
those
not
just
to
the
PSD.
G
C
E
Can
okay,
so
we
have
documented
two
use
cases
in
the
use
cases.
I
think
there
is
one
for
the
that
net
time,
time
bound
applications
and
another
one
is
for
ioam
use
cases.
E
For
example,
iom
there
is
IPv6
options
defined.
There
are
two
IPv6
options:
the
destination
option
and
the
hubby
hop
option
to
carry
iom
there
is.
There
is
a
variant,
for
example,
of
doing
a
buy,
Hub
recording
of
information
it
so
for
IPv6.
So,
for
example,
it's
it's
in
last
call
and
it
should
become
RFC
and
there
is
equivalent
functionality
desired
in
mpls
and
in
in
case
like
iom,
where
you
are
recording
every
information.
E
How
by
half
PhD
is
the
ideal
place
to
record
where
each
hop
writes
in
some
information
and
forwards,
the
packet
just
like
IPv6?
So
this
is
one
use
case
for
PSD
that
we
are
interested
in
for
the
complexity.
Part
of
PSD
I
think
the
the
team
have
done
good
job
in
preparing
one
solution,
which
is
very
similar
to
ISD,
in
the
sense
that
all
of
the
encoding
using
seven
bit
opcode
land
and
data
associated
with
it
are
very
identical.
E
E
F
I
want
to
note
a
point
that
adding
operational
State
information,
IEM
Trace
option
incremental
or
pre-allocated
posted
data
or
actually
in.
B
F
Packet
itself,
it's
clearly
prohibited
for
for
that
net,
because
increasing
the
packet
size
of
a
debt
net
with
their.
F
Non-Uh
client
information:
it
means
that
all
their
expectations
of
on-time
delivery
and
will
packet
to
us
are
changing.
If
you
look
at
that
net
architecture
documents,
so
even
active
OEM
load
on
the
network
needs
to
be
considerate
and
resources
reserved
so
that
in
normal
situation,
OEM
synthetic
packets
does
not
affect
data
traffic.
So
it
seems
that
keeping
OEM
impact
in
that
net
to
the
minimum
to
the
required
minimum
is
beneficial
and
that's
why
I
believe
that,
for.
F
Extending
that
to
mpls
network
in
general,
using
IAM
direct
expert
option
is
a
reasonable
balance
and
compromise,
because
it
allows
collection
of
operational,
State
and
Telemetry
information
at
the
same
time,
keeping
the
overhead
of
IAM
in
under
control,
because,
with
the
direct
expert
IEM
option,
The
information
collected
can
be
exported
out
of
that
relative
to
their
data
traffic.
So
again,
what
I
would
encourage
and
ask
is
that
let's
not
look
at
what
we
would
like
to
do,
but
what
is
useful
and
necessary
to
do?
Thank
you.
C
Okay,
thanks
Greg
I'm
next
in
the
queue
so
I
want
to
highlight
and
maybe
get
more
clarification
on
when
we
say
PSD
PSD
solution.
C
C
But
you
know
data
would
be
somewhere
else,
but
possibly
there
would
be
some
ISD
and
some
data
and
some
PSD
data
in
the
same
packet
and
we're
not
talking
about
a
parallel
solution
that
is
independent
of
the
ISD
work.
E
E
Is
containing
a
scope
to
say
that
IPv6
will
not
support
incremental
where
the
size
grows
as
a
packet
trousing.
So
it's
it's
kind
of
pre-allocated
and
it's
a
fixed
size
packet
that
will
be
generated,
and
so
we
we
invent
PLS.
The
idea
is
to
follow
the
same
path,
so
it
doesn't
like
grow
and
and
the
same
reason
that
what
IPv6
did
so,
we
would
have
the
the
parity
of
behavior
and
functionality
with
IPv6
in
mpls.
E
C
We
thank
you
Tony
you're
next.
B
Hi
I'd
like
to
disagree
with
both
of
the
previous
speakers,
Tariq
I
disagree
that
PSD
would
require
ISD.
We've
discussed
that
a
little
bit
on
the
mailing
list.
I
won't
reiterate
all
of
the
details
here:
new
Rakesh,
I'm,
sorry,
but
I,
don't
understand
the
relevance
to
IPv6.
B
That
is
not
a
precedent
for
what
we
do.
If
we
decide
we
need
increasing
data
space
and
PSD.
That
seems
like
a
reasonable
place
to
do
it.
Thank
you.
C
F
I'm
glad
and
thank
you
for
your
comment,
because
I
agree
and
I
don't
I
cannot
find
the
relevance
and
actually
why,
with
solution
for
mpls
data
plane
should
replicate
their
solution
applied
to
IPv6,
so
that
is
not
clear,
also
I'm,
looking
at
IAM
in
mpls
network
as
a
method
that
provides
certain
information.
So,
in
my
opinion,
IEM
direct
expert
is
absolutely
adequate
in
providing
this
information
and
keeping
their
idea
of
that.
This
information
is
useful
and
helpful,
but
not
critical.
F
So
that's
a
option
that
uses
out
of
band
collection
of
this
information
is
philosophically
most
attractive
in
my
opinion,
so
that
yes,
there,
there
is
a
original
RFC
that
defined
Trace
options
that
use
the
data
packet.
We
can
call
it
trigger.
D
F
To
collect
this
information,
but
then,
in
course,
of
discussion
that
direct
expert
was
defined
and
published
in
his
RFC.
That
provides
absolutely
same
functionality
in
terms
in
regard
to
collecting
information
so,
and
we
have
a
proposal
that
demonstrates
how
their
direct
expert
IAM
can
be
supported
as
insect
data
m
a
so,
we
don't
have
to
support
options
that
are
not
architecturally
beneficial
to
mpls,
Data
Network,
and
we
can
choose
how
to
provide
this
functionality
in
mpls
Networks.
C
Okay,
great,
thank
you
I'm
next,
in
the
queue
so
my
question
now
you
know
it's
basically
trying
to
respond
to
Tony
and
he's
free
to
respond
back
or
not.
But
are
we
saying
that
if
we
solve
the
use
case
using
ISD
when
we
propose
or
if
we
propose
a
PSD
solution,
we're
free
to
solve
the
same
use
case
in
PSD?
C
That's
number
one!
So
I
think
that
is
an
important
point
that
you
know
when
you
said:
they're,
independent,
ISD
and
PSD.
In
that
sense,
are
we
saying
that
I
could
solve
the
same
thing
in
different
places
that
that
was
one
of
the
concerns
I
had?
Thank
you.
E
B
Was
not
suggesting
that
we
duplicate
solutions?
That
was
not
the
issue
at
all.
The
point
is
that
we
may
have
situations
where
we
may
want
to
use
PSD.
It's
got
different
properties
in
ISD
and
some
use
cases
may
need
those
properties,
so
there
may
maybe
cases
where
we
make
it
makes
more
sense
just
to
use
PSD.
F
So
tarek
to
your
question,
as
we
had
a
discussion
on
ISD
draft
m,
a
header
so
when
we
found
that
the
same
functionality
can
be
realized
using
two
different
encodings
I
understand
that
we
agreed
that
it's
a
necessary
complexity
and
challenge
for
implementation,
and
we
agreed
that
to
leave
only
one.
So
I
would
extend
that
line
of
thought
to
your
question
and
think
that
it
would
not
be
beneficial
if
the
same
functionality
supported
using
two
different
methods.
F
So
I
would
say
that
if
we
can
solve
a
use
case
using
ISD,
then
these
use
cases.
F
So
only
unless
we
have
a
new
use
case,
then
we
can
consider
whether
it
cannot
be
solved
in
ISD
and
requires
PSD,
so
that
that
would
be
my
view.
Thank
you.
A
E
A
C
Okay,
I'm
back
so
I
want
to
agree
with
with
Tony
and
Greg,
in
the
sense
that
okay,
so
I
agree
with
Tony
on.
There
are
use
cases
where
it's
more
suitable
to
do
it
to
solve
it
in
in
PSD,
others
in
ISD
I
would
consider
in
the
in
that
sense.
The
PSD
solution
is
an
augmentation
of
the
overall
solution.
We're
proposing
and
there
might
be
cases
I
agree
with
Greg.
You
know
that
you
might
want
to
solve
I,
don't
know.
C
Why
would
you
do
that,
but
solve
it
in
two
places?
I
would
challenge
that
my
concern
was
the
PSD
solution
would
be
so
independent
of
the
proposal
or
the
solution.
We
have
for
ISD
that
it
can
be
considered
an
alternative.
In
the
sense.
The
closest,
in
my
mind,
is
our
rcpde
that
we
standardized
and
crldp,
and
that
time
there
was
a
you
know,
two
two
proposals
on
the
table
so
I.
That
was
what
I
was
trying
to.
C
You
know
avoid
in
the
discussion,
and
hopefully
we
can
agree
on
this
one
solution,
but
you
know
it
offers
the
solving
or
carrying
the
data
in
two
places,
but
it's
one
solution.
Kind
of
and
back
to
Rakesh.
E
E
Add
one
point
to
lower
that
yeah
agree
about
the
IPv6
context.
Only
one
one
point
about
that
is
that
if
often
customers
having
multiple.
D
E
Plane,
multiple
networks
and
having
different
different
solutions
going
from
one
domain
to
other
domains
and
the
complexity
that
brings
will
do
the
hindering
of
the
adoption
or
even
implementation,
going
to
one
theme
and
asking
for
one
solution.
Another
theme
for
different
solution
for
different
data
planes
in
a
vendor
environment
is
also
it's
very
cumbersome
or
costly
right.
So
this
is
just
one
motivation
but
again
yeah.
We
don't
have
to
mimic
everything
here,
but
there
is
some
advantages.
B
Regarding
Network
complexity,
if
particular
end
users
are
interested
in
minimizing
complexity,
I
strongly
suggest
they
disable
IPv6.
That's
not
useful.
Thank
you.
A
I
think
I
agree
with
what
Iraq
cash
said.
So
what
we
know
that
IPv6
has
done
something
similar
to
what
we
are
doing
for
mpls
now.
We
would
benefit
from
understanding
what
we've
done
and
what
the
problems
are,
and
then
we
go
ahead
and
try
to
do
the
optimal
thing
for
mpls.
A
F
Oh
well:
okay,
architecturally,
yes,
I
again,
I,
don't
see
why
we
are
going
into
this
direction,
discussing
IPv6
cuts
and
bruises,
but.
F
Observing
what
deployment
of
extension
headers
and
how
much
concern
operators
have
with
them
I
think
that
that
would
be
a
good
additional
input
for
us
to
replicate
this
functionality
in
mpls
data
plane
so
and.
F
F
With
mpls
data
plane,
so
in
that
big
one
of
their
areas
would
be
considering
the
M
A4
is
important.
So
I
don't
really
see
that
using.
F
Analogous
to
IPv6
extension
headers
is
a
good
proposition
for
mpls
that
that,
thank
you.
C
Okay,
thank
you.
Greg,
okay,
going
one
two
and
three
I
don't
see
anyone
in
the
queue
I
will
shift
to
the
next
slide
and
it's
a
continuation
of
the
previous
slide
really
and
it's
an
action
on
all
of
us
so
lava.
You
want
to
speak
to
that
and.
A
Yes,
since
I
wrote,
it
I
better
speak.
So
this
is,
if
you
have
a
strong
opinion,
one
way
or
another
on
PSD.
A
If
you
need,
if
you
think
you
need
it
to
view
it
and
you
want
it,
what
you
have
to
do
is
actually
document
the
use
case,
and
then
you
need
to
bring
it
to
the
interim
meeting
for
for
discussion
and
as
soon
as
you
have
as
soon
as
you're
prepared
to
take
that
discussion,
you
probably
better
ask
for
a
slot,
so
we
can
put
it
on
the
agenda
and
if
this
does
not
really
happen,
then
I
don't
know
where
we
are,
then
the
I
think
the
discussion
will
will
stall.
A
But,
as
we
said
earlier,
if
we
actually
make
good
progress
and
if
we
get
to
points
where
we
can
pull
the
mailing
list
for
consensus,
we
will
do
that.
Thank.
C
You,
okay
Rakesh,
you
have
your
chance,
go
ahead.
E
A
Okay,
but
when
can
you
do
it.
A
Okay,
I
see
we
have
another
first
Naval
discussion.
Also,
we
have
to
prepare
yeah
I
need
to
say
that
before
I
say
what
I
was
about
to
say,
shares
will
meet
on
Tuesdays
and
finally
agenda.
So
you
will
only
have
two
days
if
you
request
the
slot
for
the
next
meeting
so,
and
we
are
aware
of
that.
This
is
a
little
bit
awkward,
but
we
can't
really
yeah.
Do
it
any
other
way
so.
E
A
C
C
Thank
you.
Okay,
good
I
still
have
one
last
slide
and
before
we
talk
about
that,
Dave
is
raising
his
hand,
so
I'll
give
him
a
chance.
Dave
go
ahead.
H
H
That
what
needs
to
be
documented
from
here
is
kind
of
a
you
know,
a
case
or
or
answers
to
these
questions
right
for
each
relevant
use
case.
I.
Think
getting
answers
to
these
questions
will
help
us
come
to
consensus
more
than
just
reiterating,
what's
already
been
written
in
the
use
cases,
but
I'll
leave
that
to
the
chairs.
Thanks.
C
Indeed,
I
mean
the
discussion
was
I,
mean
I,
think
Tony.
Also
a
number
of
people
pointed
as
picking
the
optimal
way
of
solving
the
use
case,
and
it
would
be
good,
I
think
Rakesh
mentioned.
There
is
a
use
case
couple
of
use
cases.
Maybe
we
have
a
to
make
a
case
on
the
optimality
now.
Why
is
this
optimal
and
and
answering
these
questions
I
think
is
very
good.
I
agree
totally
agree
and.
C
And
I
don't
see
anyone
else,
so
I'll
go
back
to
the
last
slide.
I
have
and
I
want
to
reiterate
that
the
interims
are
happening
or
they're
already
scheduled
on
a
weekly
basis,
so
we
have
an
interim
next
Thursday.
Potentially,
as
Laura
mentioned,
we
will
have
a
meeting
between
the
chairs
to
discuss.
If
there
is
enough
slots
or
there's
a
need
to
cancel
the
the
meeting
requests
as
early
as
possible
are
encouraged.
We
thank
you
for
doing
that.
We
already
have
couple
of
requests,
so
that's
a
good
thing.
C
As
of
now
yeah,
we
will
try
to
respond
as
early
as
possible
that
we
will
make
it.
You
will
make
the
slot
that
you
asked
for
for
next
meeting,
but
other
than
that.
C
What
do
we
want
to
discuss
if
it's
a
chance
to?
If
you
have
other
topics
that
we
didn't
talk
about
today,
either
you
can
speak
up
now
or
send
us
an
email
with
the
request.
C
This
was
my
last
slide
and
it's
an
action
on
all
of
us
to
you
know
if
you
have
a
topic
that
is
related
to
the
interim
meetings,
the
topic
You
Know,
M,
A
and
relevant.
So
please
send
us
a
request.
E
C
A
C
Okay,
so
the
last
is
I
want
to
thank
everyone
who
attended
today
and
for
having
a
good
and
fruitful
discussion
and
hopefully
see
you
next
time
we
meet
next
Thursday
thanks.