►
From YouTube: IETF99-CCAMP-20170720-1550
Description
CCAMP meeting session at IETF99
2017/07/20 1550
https://datatracker.ietf.org/meeting/99/proceedings/
A
A
Okay,
let's
start
from
there,
you
know
not
where
so
we
can
say
that
the
text
of
the
load,
where
this
time
it
is
not
different
from
the
previous
one-
the
major
changes
it's
about.
Ip
are
things
which
is
defined
enough,
say
81
79,
you
know.
Actually,
yesterday
there
were
lots
of
discussion
on
type
er
topics
and
so
I
think
if
people
and
join
the
discussion,
we
can
get
more.
A
B
A
Stand
the
inside
of
the
pink
box
so
for
the
minute
takers,
and
we
ask
how
many
and
so
I
think
how
me
and
Oscar
they
we
hope
has
to
take
some
minutes
and
also,
if
anyone
can
take
some
minutes
slow,
they
use
a
pad.
That
will
be
appreciated.
Thank
you,
yeah.
So
for
the
chamber.
We
also
will
monitor
the
chamber
loom
to
see
if
there
are
some
comments
from
the
chamber.
Loop.
A
C
A
The
second
one,
the
right
one,
which
is
they
can
specific
you
can
get
some.
You
know
pure
information
from
the
link,
so
this
time
actually
for
this
session
we
have
two
hour
sessions.
So
even
though
we
have
two
hours
actually
you
know
you
can
see.
The
agenda
is
still
a
little
bit
of
pillage,
so
we
really
hope
that
presenters
could
use
their
time
slots
efficiently
and
leave
more
time
for
discussion.
Yeah.
A
D
Let's
proceed
with
the
stages
of
the
working
group:
we
don't
have
any
new
RFC
since
the
meeting
in
in
Chicago,
but
we
have
a
new
draft
in
the
editor
queue
which
is
the
flexicord
OSPF.
We
decided
to
put
it
on
all
them
and
to
let
me
say,
define
new
values
against
the
gmpls
as
SCSI,
which
is
now
which
pasta
della
Scala
and
was
approved
at
publication
if
a
color
to
remember-
and
so
we
decided
who
to
use,
is
use
it
as
the
basis
for
for
the
OSPF
extensions
that
was
practically
it.
D
What
else
we
have
a
lot
of
working
group
document
discussed
today,
the
first
of
which
is
the
DWDM
interface
management
and
control
framework.
We
just
wanted
to
wanted
to
remind
make
you
notice
that
we
asked
them
for
an
earlier
you
of
the
document.
It
was
what
we
asked
them
for
a
for
a
review
from
at
the
Rockland
electorate,
just
to
be
sure
that
it
reaches
there
a
good
quality
before
the
della
Scala.
There
are
four
more
drafts
in
the
type
we
did
our
working
group
documents
that
will
not
be
discussed
today
today.
D
D
E
E
The
new
version
was
submitted
in
July,
so
two
weeks
ago,
document
was
updated
twice
since
March
we
addressed
all
the
comments
from
the
list
did
a
lot
of
rephrasing
again
and
error
correction
and
we
hope
that
we
approve
the
entire
eight
ability
readability
of
the
of
the
document.
So
let's
have
a
short
look
on
the
details,
so
I
updated
the
introduction
section,
so
we
were
written
that
yeah.
E
We
have
you
did
this
completely
new
and
then
we
updated
all
the
figures
explaining
the
the
the
text
and
harmonizing
the
text
and
the
figures
a
little
bit
more.
We
updated
as
well
the
control
plane
consideration
section
of
the
document
that
the
use
case
section
especially
didi
first
use
case
service
set
up
and
yeah.
E
E
One
thing
is
still
open:
we
have
to
update
the
the
reference
section,
so
I
run
the
unit's
thing.
On
the
only
tooth
page.
There
were
some
some
minor
things
regarding
the
references,
so
we
have
to
do
here
a
little
bit
more
work
and
yeah.
Then
we
think
the
document
is
ready
for
working
group.
Last
call.
F
My
name
is
Jana
congrat
I'm
from
society.
I'm
a
newcomer
is
my
first
idea
ever
so.
First
of
all,
really
thanks
for
this
document,
because
it's
for
me
it
was
quite
nicely
readable
and
so
on.
I've
got
one
comment,
something
more
on
the
editorial
point:
I
notice
that
the
text
refers
to
a
yellow
triangle
when
it
comes
to
diagrams
and
measuring
points,
so
just
maybe
removed
over
here
all
too
okay.
C
G
A
I
So
Joanna
Martinelli
presenting
you
the
update
on
the
W
son
with
an
impairment
validation,
information
model.
As
daniel
said,
the
the
draft
was
a
big
door
want
and
I
mean
I
was
off
for
the
last
year
from
the
ATF.
But
now
we
try
to
let's
say
finalize
this
work
that
is
around
since
a
while.
You
know
on
the
my
co-editor.
This
is
no
longer
active
in
ATF.
I
The
other
beautiful
information
just
their
mind
that
is
independent
or
from
control,
plane
options
at
the
time
when
all
this
work
started,
PC
was
on
I,
so
everyone
was
looking
at
PC
as
a
final
solution.
Now
there
are
moral,
it's
a
reference.
Tor
was
solution,
likes
that
goes
as
the
end
or
or
based
on
young.
I
So
a
remind
of
the
context
on
the
content
of
the
draft.
There
are
two
section
that
basically
provide
them
from
the
definitions
and
some
introductory
test
section
3
was
the
one
that
was
disgusted
with
itu-t
and
include
all
the
content
from
ITT.
Then
the
remaining
sections
are
relates
more
to
IETF
and
control
plane.
B
J
Can
I
ask
my
question
now?
Yes,
sure
right,
okay,
yeah
I
was
actually
wondering
what
the
added
value
is
of
defining
all
these
optical
parameters
without
having
a
model
for
the
optical
visibility.
Calculation
I
mean
I
have
made
that
comment
many
times
before.
I
think
this
is
still
an
unresolved
issue.
In
my
opinion,.
I
J
K
Gabriel
Francisco
I
think
deter
this.
This
document,
let's
say,
allow
us
or
allow
who
wants
to
use
those
parameter
to
exchange.
The
parameter
from
the
network
to
the
controller,
of
course,
is
not
I.
Think
is
not
the
objective
of
this
working
group
defined
in
the
algorithm
and
how
to
use
those
parameters.
H
B
A
So
I
think
a
deters
comment
means
that
actually
unite
UT,
there's
no
articulation
impairment,
calculation
or
model,
and
then
even
we
have
computed
capability
to
exchange
there.
You
know
our
impairment
information
from
the
can
Chaplin
perspective,
but
it
does
not
make
sense
because
there's
no
foundation
to
how
to
use
it
right.
L
M
Jake
I
think
that
everyone
of
knowledge
is
acknowledge
that
this
is
not
the
full
solution
to
the
big
picture,
but
it's
chicken
and
Iraq,
seeing
you.
We
need
to
start
with
something
standardized
to
start
the
next
move,
and
this
is
one
of
the
first
items.
I,
don't
think
the
document
is
claiming
to
solve
the
full
problem,
but
it
it
remains
useful
on
required.
J
G
I
J
I
I
So
actually
is
a
in
this
case
we
have
only
draft
refresh.
There
was
a
some
major
update
presented
in
IETF
while
ago,
and
here
we
just
update
the
draft.
The
draft
take
some
interest
with
a
new
new
modeling,
with
new
young
draft
coming
on
the
working
group,
so
a
quick
refresh
on
the
content.
So
there
is
the
optical
parameter.
I
How
is
encoded
then
we
call
it
impairment,
vector
actually
list
of
optical
parameters
and
the
extension
to
the
impairment
we
call
it
impairment
matrix
actually
is
an
extension
of
the
connectivity
matrix,
define
it
in
another
in
the
arab
season.
Seventy
five,
seventy
nine
and
the
here
probably
providing
an
extension
to
the
resource
block
information
as
well
with
the
previous
information.
I
So
here
our
authors
believe
that
is
a
ready
for
a
working
group
adoption.
If
there
is
interest
in
the
working
group
in
the
Phi,
this
kind
of
information
for
sure
our
working
group
adoption
doesn't
mean
that
that
coding
is
a
let's
say,
complete
ease.
We
consider
fairly
stable.
There
are
some
few
open
points,
but
that's
up
to
discussion
for
the
working
group.
I
One
of
the
clean
opera
and
the
wanna
remove
also
their
effort.
There
is
still
one
refers
to
the
encoding
draft
that,
to
me
has
to
be
remove
it.
This
is
a
minor
section
and
there
was
only
let's
say
some
control
plan.
Consideration
doesn't
add.
Let's
say
specific
content
to
the
blue
stuff.
I
agree
with
this
I
realized
that
has
to
be
cleaner.
Doctor.
O
O
O
The
status
is
that
the
draft
was
adopted
by
C
camp
in
December
last
year
and
in
since
then,
we
haven't
really
received
any
comments
or
feedback
on
on
the
framework
draft,
but
we
submitted
a
new
version
of
the
framework
draft
in
June,
with
a
purpose,
to
take
care
of
some
comments
that
we
received
when
the
data
model
was
adopted
by
the
working
group,
and
those
comments
were
related
to
wish
to
ensure
that
these
activities
are
aligned
in
a
good
way
with
other
similar
activities
within
the
industry,
mainly
within
the
onf
I
would
say.
So.
O
O
It,
of
course
follows
the
framework
draft
and
supports
the
use
cases
and
requirements
defined
in
that
draft.
One
key
characteristic
is
that
it
augments
RFC
7223
and
thereby
it
has
a
structure
which
is
aligned
with
how
other
interfaces
in
a
microwave
node
is
expected
to
be
managed.
It
uses
established
microwave
equipment
and
radio
standards
as
a
basis
for
further
definition,
and
it
also
uses
the
draft
all
bag
and
the
onf
microwave
information
models
as
the
basis
for
the
scope
to
be
addressed
by
this
young
data
model.
O
We've
also
added
a
subsection
explaining
that
some
of
the
attributes
in
the
Augmented
7223
may
not
be
applicable
or
relevant
for
the
radio
link
model.
We've
done
a
couple
of
model
changes,
I
have
slides
for
that,
so
I
will
touch
upon
that
later
on,
and
we've
also
made
some
improvements
in
some
of
the
descriptions
of
of
the
data
notes.
O
There
are
two
new
co-authors
in
this
evasion
of
the
document,
and
that
actually
means
that
all
major
vendors
within
the
microwave
industry
is
now
participating
in
supporting
the
work
we
are
doing
here,
which
is
very
good.
I
would
also
like
to
mention
that
we
have
started
to
validate
this
modeling
practice
by
participating
at
the
ITF
hackathon
earlier
this
week.
O
One
change
that
we
are
done
is
actually
removed,
one
data
node
and
that's
the
data,
no,
the
aurochs
frequency
config,
which
was
used
to
determine
if
oryx
frequency
should
be
based
on
or
calculated
from
duplex
distance.
Instead,
we
have
clarified
in
the
description
of
the
two
data
nodes
that
oryx
frequency
will
be
calculated
based
on
duplex
distance.
If
duplicate
instance
is
the
only
parameter
provided
in
the
configuration
statement,
otherwise
its
will
be
taken
from
the
aurochs
frequency
value
provided
in
the
configuration
statement.
O
The
other
change
we've
done
is
that
we've
added
a
new
identity
for
external
commands
in
the
protection
area.
We've
added
the
identity
for
forced
switch,
which
means
that
the
traffic
is
switched
to
a
protecting
carrier
and
is
forced
to
remain
there.
Even
if
the
protection
right
here
indicates
that
it's
ok
to
switch
back
again
and.
O
The
working
group
or
sorry
the
design
team
considers
the
young-dal
a
data
model
work
to
be
completed.
We
are
currently
working
on
a
restructuring.
According
to
the
nmda
architecture,
we
have
a
proposal
which
we
have
already
received
some
feedback
on,
and
we
see
that
there
is
a
possibility
to
make
it
independent
of
the
changes
that
will
take
place
in
RFC,
72
23,
and
we
will
be
able
to
publish
such
an
update
fairly
soon
and
we
are
asking
for
a
working
group
loss
column
on
this
young
data
model
as
well.
Thank
you.
D
J
D
To
try
to
make
another
effort
to
see
what
can
be
generalized
in
the
sense
that
the
microwave
worker
is
all
fork
is
mic.
Microwave
technology
is
as
a
variable
bitrate
by
definition,
there
are
other
technologies
in
which
we
can
have
a
variable
bitrate
like,
for
example,
elastic
optical.
So
this
could
be
an
opportunity
to
try
to
make
this
work
a
little
bit
more
general.
This
doesn't
mean
I
mean
it's
too
late
to
start
with
a
brand
new
work
possibilities,
and
then
I
do
technology
specific
extensions
here,
but
there
are
some.
D
Minor
changes
that
could
help
to
the
focus
on
this
one
of
these.
It
would
be,
for
example,
my
personal
opinion
in
the
work
in
the
frame
worker.
Just
add,
I,
don't
know
few
lines
saying
that
these
most
of
the
considerations
done
in
in
the
document
might
also
be
applicable
to
other
other
tech
variable
bitrate
technologies.
D
This
is
just,
in
my
opinion,
just
a
minor
updater
to
the
framework.
Nothing
nothing
else
in
CDT
on
the
other
side
had
updated
to
the
model,
could
be
a
little
bit
more
significant
I.
Don't
expect
super
major
changes.
I
see
two
options.
One
could
be
to
have
a
new
document
which
is
which
includes
the
generic
part
and
then
leave
the
technology.
Microwave
specific
extensions
here
or
within
the
same
document
just
to
split
the
the
model
and
say
this
part
is,
can
be
generalized
can
be
applicable
to
other
technology.
This
part
is
a
purely
purely
immaculately.
P
L
Q
B
O
Can
just
mention
that
we
started
doing
that
based
on
this
Etsy
302
to
170
and
I,
think
there
are
seventeen
sections
or
seventeen
different
version
or
chapters
or
paragraphs
or
standard
documents
in
that
series,
and
it
turned
out
at
least
when
I
try
to
resolve
this,
that
they
there's
a
sort
of
a
circular
reference
between
all
these
seventeen.
So
it
was
very
difficult
to
be
very
strict
on
exactly
which
document
out
of
those
seventeen
but
a
specific
parameter.
What
had
was
defined?
Q
Least,
if
some
of
them
are
come
so
here
in
our
case,
in
idea
it
happens
the
same.
There
are
many
cases
where
the
same
is
having
several
RFC's,
but
then,
in
the
end,
you
put
two
one
RFC.
What
is
a?
What
is
they
considered
as
the
the
main,
the
main
source,
so
that
same
happens
to
us
in
optical
with
ideal
that
sometimes
is
hard
to
find
in
which
document
you'll
find
define
the
N
or
the
M,
or
something
like
that
back
in
the
end,
you
need
to
put
some
some
record.
Thank
you.
Thanks.
I
Giovanni
speaking,
just
a
quick
comments
regarding
related
to
the
fact
that
you
major
analyzed
some
parts,
something
that
they
see
but
I've
seen
already
somewhere
else
regarding
protection
part.
So
were
you
always
see
you
know,
mono
switch
for
switch,
and
actually
I
don't
know
if
you
have
a
controllable
priority
who
wins
against
the
two
likely
false.
But
you
know
some.
O
R
R
The
design
team
or
the
motivation
for
design
team
is
the
fact
that
in
the
IDF
we
are
developing
multiple
Yamada's,
which
are
oblique
that
could
be
applied
at
the
MBI
of
the
transport
Sdn
controller,
but
we
have
got
two
questions
we
want
to
answer
is
how
in
this
Yamada's
actually
applies
at
the
transport
networks,
and
the
second
one
is:
if
there
are
any
gaps
in
the
car
in
your
model,
IDF
is
developing
when
used
for
transport
network
er.
To
answer
to
these
two
questions,
the
design
team
is
looking
at.
R
What
are
the
use
cases
that
we
that
has
to
be
addressed
at
the
Transport
MBI,
and
we
are
analyzing
in
detail?
How
would
the
ITF
Yamada's
can
you
will
be
used
to
address
these
use
cases
and
the
working
matter
that
we
have?
Is
we
have
a
design
team
analysis,
given
the
bigger
number
of
people
in
the
design
team,
and
we
all
almost
weekly
conference
calls
on
Wednesday
at
3
p.m.
European
time
were
to
to
discuss
a
contribution
of
an
issue
and
we
have
it.
R
We
have
a
github
area
where
we've
told
all
our
minutes
and
documents
and
Franco
Kinshasa
as
sort
of
the
first
deliverable
is
a
use
case
description
in
the
internet
draft,
which
has
been
updated
since
the
last
idea
of
meeting
we,
we
have
added
the
some
description
of
the
protection
scenarios
for
use
case
number
one
for
the
single
domain
single
layer
scenario,
and
we
are
planning
to
complete
this
document
by
addressing
in
order
of
priority.
First,
we
got
feedback
that
people
wants
to
see
more
multi
domain
iographer
the
multi
layer.
R
R
How
a
transit
service
so
I
know
do
transit
and
adieu
starting
from
to
see,
can
transit
to
a
transport
network.
We
have
provided
detailed
JSON
code
examples,
so
we've
wrote
down
in
detail.
So
this
is
own
code
and
we
are
also
validating
that
the
example
that
we
are
providing
are
fully
compatible
compliant
with
a
year
model.
R
The
Yamada's
in
the
meanwhile
has
been
updated,
so
you
have
to
either
adapt
the
example
to
the
new
models
we
want
to
equate
them.
We
need
to
complete
the
use
case.
Number
one
are
the
protection
of
scenarios
example,
and
maybe
we
will
start
a
new
draft
for
the
use
case
number
three
in
our
system,
in
addition
to
the
Internet
of
the
river
boss,
we
have
also
other
deliverables,
which
is
a
feedback,
sir,
that
we
are.
We
are
provided
as
an
output
of
our
discussions
and
analysis
to
different
working
groups.
R
So
the
first
point
we
got
us
a
comment
last
the
meeting
to
qualify.
What
are
the
assumption
we
are
making
on
the
on
the
trees
working
group
activity?
So
we
are
assuming
the
llamo
decision
model
Arabic
applicable
to
the
actn.
Mpl
interfaces
are,
as
defined
in
the
Younger
applicability
draft
in
tiesm,
and
is
we
made
it
very
spicy
too,
and
it
looks
like
that.
The
assumption
is
is
good
for
our
worker.
We
are
assuming.
R
We
are
making
an
assumption
that
how
the
tea
eternal
motor
can
be
used
to
set
up
the
tunnel
segments
and,
as
result
of
this
assumption,
we
have
triggered
some
discussion
in
teaser,
which
is
still
ongoing.
So
there
are
different
opinions
of
among
these
people,
but
we
are
continuing
this
assumption
because
we
don't
have
any
counter
proposal
at
this
moment,
so
we
will
align
with
the
outcome
of
this
discussion
when
it
is
completed.
R
R
It
Soliday
MDC
is
the
only
entity
that
key
that
cannot
allocate
a
label
which
is
available
on
both
sides,
so
the
assumptions
that
the
label
is
decided
by
the
IDM
DSC,
based
on
some
information,
which
is
a
label,
set
information
that
the
two
pians
is
provided
to
the
mdac,
and
because
of
this
we
got.
We
gave
some
feedback
to
the
Titan
and
topology
model
which
have
been
updated
to
provide
the
level
set
information
and
also
we
have
an
open
issue.
Last
time
when
we
indicate
an
S.
This
is
a
route
object,
an
another
interface.
R
We
have
a
clarification
that
the
route
ID
interface
ID
are
actually
the
node
ID
and
linked
EP
ID
in
the
T
topology
information,
and
that
was
a
useful
clarification
for
people
that
needs
to
use
the
model.
Last
three
Beck's
base.
Thanks
to
the
idea
for
the
fact
that
we
are
validating
our
json
codes
with
the
Yamada's.
We
have
also
find,
during
the
validation
process,
that
there
were
some
mistakes.
R
There
are
syntax
mistakes
in
the
Yamada's
developed
by
T's,
and
we
provided
this
feedback
to
the
auto
switch,
have
fixed
them
orders
and
for
the
detail
for
the
feeder
to
the
cecum
working
group.
We
have
dedicated
slides
at
the
end
of
the
presentation.
Okay,
I
will
wait
to
be
fast.
This
is
presenting
the
use
case
number
one.
We
have
already
spent
similar
slides
in
the
ITF
98.
So
at
this
moment
in
time
we
are
focusing
on
this
interface,
the
interface
between
the
MDS
Seanna
Transpo
PFC.
R
We
will
not
cover
what
happens
on
the
interface
between
MD,
SEM
and
IP
apse,
because
it's
outside
of
our
scope-
and
we
will
we
are-
we
will
later
cover-
maybe
the
CMI
for
transport
services.
And
so
today
we
are
here-
and
we
are
considering
this
type
of
topology-
for
single
layer
single
domain
as
an
oto
topology.
R
And
so
you
can
see
the
Aster
topology
will
you
will
have
an
abstract
node
for
every
physical
node
and
we
use
the
same
naming
to
clarify
to
make
it
easy
to
understand
how
the
our
start,
node
on
the
NPI
maps,
to
the
physical
node
in
and
physical
network
elements.
We
assume
that
on
the
reuters
is
not
visible
and
the
NPI,
but
we
have
shown
here
to
make
the
example
easier
to
understand,
and
we,
as
we
have
found
that
there
is
a
some
information,
a
tree
topology
model,
the
external
domain
container.
R
If
the
MDS
scene
is
to
create
connectivity
between
route,
r1
and
Rooter
tree,
it
has
to
request
that
to
the
transport
domain
controller,
to
set
up
an
audio
Tanner
segment,
starting
from
s
3
1
and
then
into
s
TX,
6
2-
and
this
is
done,
and
the
assumption
is
that
the
English
and
the
Integris
points
of
this
Tanner
will
be
indicated
in
this
PC
root
object.
This
is
the
assumption
that
we
have
today,
which
not
everybody
in
tis
agree
with,
but
this
what
we
are
doing
today
and
we
will
fix
when
we
have
write
instructions.
R
This
is
an
example,
this
some
code.
From
our
example
we
can.
We
are,
since
we
are
doing
a
some
validation.
We
are
putting
real
identifiers
1
and
you
will
see
also
in
the
addressing
scheme,
sir,
and
to
make
it
more
easier.
We
agreed,
we
have
put
a
sort
of
convention
to
do
the
comment
and
we
will
slip.
The
comments
will
be
sitting
down
from
the
coder
before
the
validation,
so
we
have
to
follow
some
conventions
on
the
way
we
put
the
comments
in
it
since
the
tunnel
is
not
terminated.
R
The
de
soutien
destination,
TP
ID,
should
not
be
provided
there
and
we
put
there's
a
commerce
or
the
reader
understand
that
he
doesn't
need
to
provide.
That
is
not
something
we
have
forgotten.
We
can
indicate
that
the
time
that
is
V
Direction
our
core
rooted,
though
we
have
enough
information
in
the
tea
tunnel
model,
and
we
have
also
validated
the
information
that
is
in
the
OTA
and
augmentation
for
developers.
R
I
see
Campa
for
for
configuring
the
times
not
send
to
flock
to
you
the
switching
capability
at
the
edge
of
the
domain
and
for
the
for
the
explicit
part
that
we
can
indicate
the
English
interface
on,
as
we
know
that
we
use
the
Rooter
ID
is
the
node
ID
of
a
stream,
and
we
appeal
formatted
ostlers
accordingly
and
yes,
1/3
LTP.
This
part
will
change,
because
the
tea
topology
tunnel
models
change
the
name
of
this
to
a
table,
to
make
it
more
easy
to
understand
what
they
are,
the
yarn
compared
to
the
topology
model.
R
So
what
is
the
current
state
to
serve
for
about
the
Sikh
Empire?
Okay,
we
were
expecting
some
use
case,
a
gap
analysis.
We
are
expecting
a
some
work
on
the
machine,
your
models
and
the
guidelines,
so
we
have
published
the
use
case
addressing
the
first
point
and
we
have
started
to
publish
the
analysis
for
use
case
number
one
and
we
have
given
feedback
to
tease
about
issues
on
the
exist
in
your
modules
we
have
not
yet
in
the
final
Amadas,
but
maybe
in
the
next
phase
we
will
identify
new
young
models
to
develop.
R
R
What
we
would
like
to
to
improve,
we
would
like
to
improve
the
getting
feedback
from
other
as
the
own,
especially
on
the
solutions
we
have.
We
have
started
to
circulate
the
information
about
our
what
we
would
like
to
get
more
information
about
what
they
think
about
what
we
are
doing,
and
we
will
I
caution
to
speed
up
the
work
on
the
analysis.
I
do
more
more
work
to
do
they
the
heart
start.
R
The
at
work
at
the
beginning
is
to
have
everybody
on
the
same
page,
to
understand
out
to
work,
understand
how
to
use
the
tools
that
has
been
done
now.
We
have
to
just
do
the
examples
we
have
the
process
on
developing
the
examples
has
been
started
up
and
now
we
should
be
a
little
bit
faster.
The
next
steps.
Now,
if
there
is
to
go
to
use
case
number
trim,
the
multi-domain
ask
consideration.
Sir.
R
We
got
some
feedback
from
people
that
make
the
kink
that
a
discussion
that
ended
up
with
a
need
to
have
the
available
level
sets
the
edge
interface
is
something
generic
not
only
applicable
to
young.
So
there
is
an
idea
that
maybe
we
need
to
propose
to
PC
to
do
some
extension
to
piece
up
to
provide
equivalent
information.
We
are
planning
to
do
some
discussion
about
applicability
of
transpose
API
for
the
network
licensees
we
think
ICT
and
is
applicable
for
metro.
Slicing
transported.
R
Vi
is
therefore
part
of
a
CDN,
and
we
we
don't
know
yet,
and
we
will
decide
whether
the
current
use
cases
are
already
sufficient
to
cover
time.
As
I
said,
we
need
to
develop
additional
use
cases.
So
that's
something
to
be
discussed
and
we
would
like
to
have
to
socialize
what
we
are
doing
with
the
other
as
the
o
in
particularly
or
an
F
and
any
F.
R
D
F
L
D
D
R
L
L
A
N
That's
a
good
good
point:
I
agree,
I'm,
just
referring
to
the
previous
question
about
operating,
so
our
design
team
is
dealing
with
the
NBI
aspect,
so
it
is
to
be
understood.
If
you
heard
am
that
my
view
is
device
model
is
related
to
our
work.
R
D
Q
Hi
Oscar,
nonetheless
telefónica
so
now
we
are
going
to
go
in
another
of
these
draft.
That
is
technology.
Specific
in
this
case
is
the
please
clean
the
damn
model
for
the
class
dude
optical
network,
so
you
can
move
forward.
So
we're
gonna
speak
about
these
two
draft
that
came
from
one
okay.
So
why
do
we
need
this
model?
If
we
look
at
the
topology
itself,
we
have
the
T
topology,
which
you
know
genetic.
It
is
splitted
into
the
layer,
3
topology.
Also.
Q
We
have
a
the
tablets
on
model
here,
but
we
do
not
have
yet
a
model
to
cover
the
details
of
the
place
grid
optical
network.
So
we
consider
it
is
interesting.
Also
that
you
can
request
you
can
modify.
You
can
apply
a
work
with
flexi
allocation
with
optical
network.
That
is,
please
get
ready,
which
it
differs
a
little
bit
from
the
revolution.
So
here
we
proposed
a
set
of
young
models
to
be
able
to
manage
discrete
optical
networks.
All
the
concepts
are
based
on
the
stable.
Q
It's
already
generated
a
definite
framework
document
that
we
did
some
some
years
ago.
All
the
concepts
are
there
and
basically,
what
we
aim
at
doing
is
be
able
to
represent
the
in
the
that
in
a
topology
that
when
we
ask
for
that
apology,
we
are
able
to
show
the
details
of
of
flash
grid.
We
are
able
to
show
the
details
of
the
transponders
and
we
are
able
to
establish
in
the
space
that
let
see
that
are
flexible,
LSPs
or
SS
one
less
piece
over
the
word
project.
Okay,
so
can
you
move
to
the
next
one.
Q
So
what
did
we
change
from
the
previous
version?
So
in
the
previous
version
we
had
everything
in
a
single
document,
but
now
we
have
acceded
into
one.
It
is
just
devoted
to
the
two-day
topology,
so
we
have
a
dated
to
to
keep
it
compatible
with
the
later
stages
of
the
T
topology.
Maybe
here
at
this
point,
we
still
need
some
discussions
with
you
guys
to
see,
which
is
the
best
way
to
to
go
from
the
generic
model
to
the
technology.
Q
Specific
model
I
saw
that
I
got
had
some
comments
today
for
the
Odeon
model,
so
I
guess
those
very
same
comments
would
apply
here
in
the
way
to
do
it.
Another
difference
that
from
the
previous
track
is
that
there
are
some
the
C
non-standard
parameter
for
the
transponder,
that
we
remove.
For
example,
we
refer
to
a
slice
voltage
ponder,
so
that
is
they
say
by
now,
until
that
is
define
minute
or
something
like
that.
We'd
take
that
out
and
some
some
minor
changes
as
forcibly
is
this
lot.
Q
They
followed
with
generality
that
is
added
there,
okay,
so
the
the
other
version
that
we
have
is
the
extending
the
the
tunnel
model
is
the
drop
to
define
the
that
Leslie
media
channels.
So
this
one
settlement
that
the
tunnel-
and
here
there
are
some
some
parameters
or
some
new
parameter
for
somebody
like
the
link
channel
that
reduces
the
LSP
here.
Also
some
is
worth
mentioning
also
here
we
include
all
the
innocent
lead
in
detail.
The
intended
configuration
so
do
this.
Q
But
we
think
that
the
both
drafts
are
ready
to
talk
to
be
adopted
and
can
be
a
foundation
to
further
work
and
further
modification,
not
saying
that
they
are
complete
at
the
app,
thus
in
a
indeed
not
good,
shape
enough
to
be
taken
by
the
working
group
and
I.
Think
a
the
flexi
is
completely
know
who
had
the
framework
for
the
Lexus
or,
and
the
revolution
is
stable.
Now
it's
probably
nightly
so
maybe
it's
time
to
also
dedicate
some
time
to
put
this
one.
So
with
this
any
further
questions
comments.
F
B
F
Was
a
yeah
yeah,
it
was
a
discussion
about
the
M
and
n,
isn't
small
multipliers
and,
as
you
said
it,
it
brings
certain
complications,
for
example,
that
the
same
numerical
value
of
M
and
an
could
actually
refer
to
incompatible
channels.
So
my
question
is:
was
it
ever
consider
to
use
just
gigahertz
or
three
Hertz
for
both
the
center
of
a
channel
and
for
teaching
width.
K
Gabriela
speaking-
and
there
are
at
least
two
reason
because
we
choose
n
and
M
number
one
is
this:
is
the
tradition
of
ITT
number
two
using
with
this
kind
of
granularity,
terrors
or
or
lambda
or
nanometers
may
n
that
we
may
we
end
up
in
rounding
issues
so
having
five
digits
after
the
comma
after
the
dot
is
not
enough,
because
the
frequency
may
be
interpreted
in
a
different
way.
You
probably
saw
the
open
road
MA
proposal.
I
was
not
agree
with
that
proposal,
because
this
issue.
So
this
is
what
we
recommend.
F
Something
that
I
can
understand,
because
we
are
actually
designing
some
of
our
other
idioms
using
coil
modules
and
yeah
I
hate.
Quite
a
few
implementation
issues
in
some
open
source.
Libraries
when
it
comes
to
using
this
MO.
Q
And
they
say
that
in
a
name
have
they
defined
in
in
ith
row
and
they're
very
clear,
all
the
different
things
as
I
said
before.
One
thing
is
that
you
want
a
certain
in
a
name
and
because
the
devices
might
have
a
different
granularity.
You
might
get
something
wider,
but
I
think
that
the
concept
is
already
in
the
framework
and
also
this.
Q
This
is
the
this
is
the
effect
I
think
it
is
the
effective
channel
in
country
immersively,
but
it
wasn't
one
of
the
slot
with
an
is
the
effective
as
what
with,
which
is
what
you
really
get
so
here.
Also
the
beauty
of
the
jungle
is
that
you
can
say
okay.
This
is
what
I
want,
and
this
is
one.
This
is
what
I,
what
I
get,
but
also
that
the
concept
at
all
we
take
everything
from
ideal,
so
so
there
as
we
are
not
defining
the
data
plane.
J
You
yeah
I,
would
like
to
echo
with
the
previous
speakers
have
just
said:
I
concur
with
them
that
we
should
use
the
n
mm
well
dues
and
everything
is
related
to
the
N
and
M
values
is
well
specified
in
itu-t
recommendation.
G6,
94.1
and
I
would
actually
suggest
that
this
recommendation
should
be
added
as
a
normative
reference
to
the
two
documents.
F
So
if
I
can,
maybe
just
explain
how
it
felt
to
me
when
I
read
the
ITU
document
to
me,
it
felt
as
if
the
M
and
n
was
used
just
something
like
an
implementation
detail
or
a.
We
call
to
better
convey
the
idea,
but
yeah
I
understand
it.
Well,
if
everybody
in
this
room
understands
it,
maybe
I'm
wrong.
So
thank
you.
K
S
T
Q
In
the
framework
document,
and
in
the
in
that
you,
what
a
here
is
more
about,
the
media,
media
channel
and
network
media
channel
were
very,
very
similar,
let's
say
so.
This
is
why
the
question
here
may
be
for
the
mobile
model
is,
though,
we
need
to
have
different
constructs
for
media
channel
network
media
channel
the
tunnel
might
all
all
be
the
same.
So
this
is
why
we
want
to
discuss
a
little
bit
this
and
just
to
clarify
what
what
really
is
needed
in
the
in
the
model.
Okay,.
Q
So
for
the
data
plane,
of
course
we
will
align
with
with
the
ITU
and
but
there
also
something
that
that
idea
is
specific
under
the
that
uptake
control,
blinkin
strats.
So,
for
example,
the
tunnel
itself
is
an
artifact
that
you
use
to
to
represent
something.
So
this
is
maybe
this
is
what
we
need
to
align
in
in
IDF
and
take
the
those
the
the
data
plane
concept.
Is
that
the
one
that
we
are
in
here?
We
define
how
we
abstract
those
data
plane
concepts
here
in
in
IDF,
but
I
think
it's.
H
G
J
Theatre
is
a
quick
one.
Yes,
I
would
like
to
the
altars
to
reconsider
the
redefinition
of
the
connectivity
matrix
I
made
this
comment
also
during
the
t's
MPLS
pc
e
c
camp
session,
when
yang
presented
his
draft.
I
think
the
author
should
look
at
the
tea
topology
connectivity,
matrix
definition,
and
the
other
point
I
would
like
to
make
is.
They
should
also
consider
whether
this
is
a
read-only
attribute.
As
far
as
WDM
networks
are
concerned,
thanks.
H
I
think
there
are
some
late
changes
that
you
know
seeking
people
want
aware
of
it.
So
I
think
we
need
to
talk
to
e
go
offline
on
those
connectivity,
matrices
and
I
believe
four
flex
grid.
You
have
a
application
code
on
the
interface,
so
I
think
we
still
need
to
have
that
element
which
is
additional
element
for
for
this
case.
So
it's
not
the
same.
So.
Q
K
This
is
a
Gabriela,
Galimberti
and
I
will
present
5
draft
now
all
related
to.
If
you
want
to
flex
Creed,
they
start
with
the
first
one
that
is
actually
extending
the
possibility
to
signal
in
using
GPS
or
label
other
than
the
media
channel
and
the
media
channel
with
so
the
end,
and
them
also
the
possibility
to
allocate
the
subcarrier
into
this
media
channel
in
case
of
with
multiple
subcarrier.
So
basically,
gmpls
now
is
ready
and
I
would
say,
deploying
to
support
W
Sun
network.
K
We
got
two
very
good
draft,
very
good
RFC,
sorry
that
is
a
76
98,
the
frame
of
99
setting
the
labels
for
Flex
spectrum
and
the
77
92.
If
I
remember
well
that
define
the
the
RSVP,
those
draft
address
properly
the
SS
Orion
issues,
but
do
not
give
yet
any
possibility
to
identify
what
how
and
what?
What
is
the
position
of
the
subcarrier
that
are
inside
the
media
channel.
K
Of
course,
if
you
have
a
single
carrier
inside
the
media
channel,
the
subcarriers
intra
frequency
is
the
same
of
media
channel,
central
frequency,
but
issue
that
if
you
have
more
than
one
media
channel,
that
is
the
case
of,
for
example,
flexible
transponder.
Here
we
need
to
address
this.
This
issue.
K
K
Media
channel
inside
a
composite
media
challenge
to
identify
them
the
center
frequency
eater
of
the
media
channel
and
a
single
carrier
at
this
point
containing
this
media
channel.
But
with
this
solution,
let's
say
the
the
carrier
will
go
step
at
6.25,
while
in
some
cases
for
Nyquist
shaping
we,
the
carrier
may
be
squeezed
may
be
allocated
closer
than
six
point
six
point:
twenty
five
steps.
K
So
scope
of
this
draft
is
to
extend
the
label
and
proposed
actually
to
label
one
to
identify
the
let's
say:
request
coming
from
the
client
on
the
unit
or
face
telling
telling
to
the
network.
What
kind
of
interface
should
be
used
and
also
another
level
is
to
carry
this
information
across
the
the
network
and
allow
the
control
plane
to
share
the
information
related
to
the
interface
and
share
the
information
related
to
the
subcarrier
position.
Also
not
only
at
the
two
end.
So
at
the
transponder
side,
but
also
into
the
intermediate
Rodman.
K
Okay,
the
first
draft
was
presented
in
Chicago.
We
modified
a
few
typos
and
sunder
stupid
description
test
and
also
TLD
that
was
related
to
them
to
the
label
has
been
modified.
A
removing
information
already
present
in
the
signalling
defined
by
the
6699
recommendation
RFC.
So
here
we
have
actually
the
network
configuration
you
can
see
at
the
two
end
points
flexible
transponders,
flexible
transponder
may
have
more
than
one
carrier
at
this
point.
K
K
The
transponder,
of
course,
should
signal,
as
is
telling
the
network
a
uni,
the
let's
say,
available,
available
parameters
like
available
bandwidth
or
available
lambda.
Whatever
should
also
signal
what
kind
of
interfaces
are
available
and
other
few
parameters
that
are
related
today,
let's
say
today's
signal,
or
today
circuit
that
needs
to
be
set
up.
K
The
control
plane,
the
gmpls,
take
care
of
this
information,
make
some
calculation
terms
of
routing,
but
also
in
terms
of
spectrum
allocation
and
more
in
terms
of
transparency
in
terms
of
carrier
position
inside
the
spectrum.
Of
course,
the
agony
that
are
applied
to
make
all
this
calculation
are
not
scope
of
this
document.
So
if
you
want
up
to
the
gmpls
or
SSI
on
a
control,
plane,
implementation.
K
As
I
said
in
this
example,
we
have,
for
example,
for
transceiver
or
for
interface
available,
is
up
to
the
control
plane
to
make,
for
example,
based
on
the
routing
to
choose
all
four
interfaces
and
then
carry
for
example,
the
400
gig
payload,
using
four
carriers
running
at
100
Giga,
like
QPSK
modulation,
format,
allow
or
if,
for
example,
the
name
the
channel
or
the
circuit
is
not
to
not
to
not
to
long.
Let's
say
less
than
600
700
kilometers
the
contour
pay
may
decide
also
to
use
two
carriers
and
then
apply
a
more
aggressive
modulation.
K
Okay,
this
is
the
the
first
Tilby
we
propose.
As
you
see,
the
number
of
carriers
are
exposed
from
the
client
to
the
network,
also
the
client
bandwidth,
so
the
control
plane
knows
how
much
payload
is
has
to
carry
so
in
other
word.
That
knows
how
much
payload
can
be
share
on
the
different
carriers
and
then
the
subcarrier
frequency
tunability
that
is
basically
optional.
K
Second
LV
is
add,
and
this
is
happening
in
the
signaling
and
inside
that
the
network
is
the
modulation
ID,
the
FEC
baud
rate
and
again
this
is
for
carrier
by
carrier
the
grid,
the
care
identifier
in
case
of
we
have
more
than
one
carrier.
This
Tildy
will
be,
let's
say,
use
for
each
carrier,
so
will
be
multiplied
per
carrier
and
other,
let's
say
three
parameter
that
is
J
is
the
granularity
of
the
carrier
position
inside
the
spectrum.
That
is,
can
be
zero
point,
one
gigahertz
as
a
default
value
or
even
less
0.01
gigarth.
K
So
the
granularity
is
very,
very
fine.
K
is
a
parameter
that
identifies
the
position
of
the
carrier
compared
to
the
media,
channel,
Center
frequency,
and
then
we
will
see
here
the
explanation
here.
We
have
an
example
of
media
channel,
which
is
n,
so
the
central
frequency
is
equal
for
the
same.
Then
the
width
is
m,
equal
8
and
then
the
value
of
K,
1
and
K
2
identify
the
position
of
the
sucker
compared
to
the
N
value.
K
D
K
C
K
K
K
J
J
D
D
K
D
D
M
Emrick
arranged
has
an
operator.
I
would
be
happy
to
have
other
proprietary
proposal
joining
this
initiative,
because
that
would
be
the
way
to
turn
into
solid
rock
youment,
considering
that
I
am
actually
deploying
this
kind
of
military
under
alien
wavelengths.
This
is
something
that
is
actually
required
to
stop
doing
just
manual
multiple
step
operations,
so
I
definitely
support
that
and
it'd
be
a
pity.
You
have
to
see
jitter
or
someone
else
joining
thanks.
M
K
F
B
G
Think
if
there
is
a
specific,
very
explicit
data
plane
specification
will
be
very
helpful
to
have
that
kind
of
much
wonder
staff
that
support
subcarrier.
But
currently
we
do
not
have
this
kind
of
specification
and
if
different,
when
there
comes
with
their
own
implementation,
I
don't
know.
I,
don't
think
that
that
will
solve
the
operators
problem.
K
D
K
C
S
My
comment,
inspiration
you
jogging,
and
my
comment
just
and
this
this.
This
draft
includes
the
definition
for
the
modulation
ID
inside,
but
I
need
ask
for
inverse
and
the
career
difference
and
what
kind
of
the
modulation
ID
yeah.
Maybe
this
is
the
Autobots
coping
idea,
but
at
an
organizational
bug,
yeah.
K
Actually,
I
listed
a
bunch
of
modulation
format
in
the
draft.
Now,
and
probably
this
is
not
exhaustive
list
and
probably
is
not
the
best
way
to
represent
these
parameters.
No
on,
let's
say
today
we
have,
we
have
just
numbered
each
each
modulation
format.
I
think
this
probably
could
be
a
weak
point
of
this
proposal.
He
says
that
we
should
extend
the
agreement.
Okay,
QPSK
could
be
number
one,
I
don't
know,
68
QN
could
be
number
two
or
whatever.
This
is
absolutely
open,
and
even
if
we
have
already
an
implementation,
is
not
an
issue
to
change.
R
Ether
woozy
from
away
I
think
there
is
some
standardization
working
question,
six
of
stereo
15
for
the
optical
interfaces
and
it
will
be
much
better
if
we
wait
and
we
we
develop
a
control
plane
for
the
standard
data
plane
rather
than
four
different
or
proprietary.
So
not
because
we
don't
have
a
son,
a
data
plane.
What
we
are
controlling,
that's
my
concern.
They.
N
D
I
Just
one
common
regarding
multi-vendor
the
issue
that
we
are
see
also
in
the
past
that
there
are
shortened
under
example
dated
he
doesn't
define
the
interrupts
stand
up,
but,
for
example,
we
were
making
under
gig
interface
working
actually
talking
with
a
loo
under
the
interface,
and
we
did
interrupt.
I
mean
Cisco
and
I
lose
picking
each
other.
So
there
is
a
need
to
let's
say
understand
if
they
do
to
interface,
even
from
different
vendors
has
to
talk,
are
able
to
talk
to
each
other.
So
this
kind
of
work
is
somewhat
needed.
I
I,
don't
I,
say
I,
don't
express
my
opinion
right
now
about
the
path,
but
something
we
need
to
move.
Even
if
standard
doesn't,
let's
say
will
come
later
on
then
the
other
Paragard
in
the
modulation
format
and
so
on.
We
did
in
the
past
the
optical
interface
class
discussion.
This
is
a
way
where
you
actually
carry
around,
let's
say
numbers
and
you
defied
the
meaning
of
the
number
somewhere
else.
So
we
may
define
an
external,
let's
say:
registry
table
were
you
cover
and,
yes,
you
don't
need
to,
let's
say:
go
through
changing
your
encoding.
C
C
So
next
slide.
Okay,
let's
first
talk
about,
but
the
emphasis
over
this
draft,
the
top
of
the
mainly
focus
on
that
comes
true
or
be
under
100,
Giga,
specific,
escape
ability
using
GM
PS
some
bullets,
some
infamous
and
just
needed
to
be
noticed
that
when
the
first
why's
that
OTO
scientist
layer
network
away
the
for
monitoring
capability,
but
no
flexibility,
it's
because
the
stir,
as
described
in
a
cute,
easy
dota
872.
C
There
is
no
concept
of
you
saying
something:
I
work
that
which
means
that
the
odious
in
that
another
support,
Tom
slaughter,
switch
and
a
sixth
one.
Second,
one
second
capability
is
the
odious
how
to
transport
without
using
audios
in
the
oto,
saying,
audio,
saying
single
or
can
be
viewed
as
a
being
found
by
Intel,
available
and
OTO
say
audio
signals
we
sell
about
want
to.
Thank
you
and
each
of
them,
maybe
transport
or
what
they
found
the
interfaces
the
the
old
using
I
was.
C
He
should
be
able
to
find
a
single
instance
together
and
the
distributors
and
over
different
interfaces.
Also
there's
a
special
of
you're
saying
signal.
We
say
the
ought
using
the
same.
This
signal
is
same.
Have
the
same
over
here.
I
thought
you
same,
but
only
am
tributary
slots
over
capacity.
Also,
the
audio
say
unless
about
5g
that
tum
start
chronologic
about
as
a
CPI
as
Yoda
of
the
TV
and
the
same
as
a
teen
I'll
be
okay,
so
that
young
has
come
true.
C
Alliance
should
be
able
to
support
for
the
general
okay
so,
and
it
was
easy
once
and
that
as
reminder
is
that
I'm
wait.
They
discussed
about
the
audio
say
as
he
what
he
taught
us
a
SP,
but
it
looks
like
a
way
forgot
to
include
this
in
you
know
draft.
Maybe
after
this
meeting
we
needed
to
make
some
minor
changes
after
two
addresses.
So
next
one
okay
about
the
history
of
this
topic,
the
first
version
of
this
topic
of
our
present
night
last
year
in
November.
C
And
the
next
one,
also
just
the
same
since
last
I
kept
Academy
team.
We
have
the
South
discussed
a
several
different
topic,
the
first
wise,
whether
flexo
it
out
is
in
the
scope
of
the
district,
and
the
answer
is
no
way
out
of
some
Peace
Corp
sandwich
after
that.
To
describe
the
transport
about
using
this
talk
him
in
the
test
viewers.
What
you
saying,
even
if
there
are
some
peas,
it
was
a
players
in
wording
in
director
Singh.
C
This
document
has
two
videos,
this
and
I
as
a
transport
mechanism
at
the
lower
layers,
Oh
tubes,
audio
say
that
define
I
accept
earlier
and
therefore
free
so
and
as
a
transport
mechanism
are
considered
outlaws
so
also
another
issue.
Another
topic
that
we
discussed
rail
as
flexi
you
and
400
Giga
use
cases
which
was
written
have
some
a
little
note
about
it.
Is
that
all
these
cases
actually
map
now
OD
incarnate,
a
single
event
or
dfrac
single
of
the
various
rates?
C
C
Okay,
about
networks
met
steps,
as
I
mentioned
before
this
stuff
that
had
been
pried
into
three
times,
so
a
request
for
walking
corporate
wheels.
Also
asking
her
about
a
topic.
Has
we
also
have
an
agreement
that
this
rather
had,
including
almost
that
they
have
abilities
and
that
we
need
to
cover
we
needed
to
use
MPs
to
come
true,
so
it
has
a
request
for
or
in
corporate
offices.
So
we
have
this
draft
that
pace
it
would
for
developing
some
sluice
and
best
document.
S
You
did
this
again.
Could
you
back
to
the
previous
writer?
He
got
into
the
use
case
for
the
flexi
and
Mohammed
G.
Does
it
mean
a
client
for
the
you
freaks
right
and
I
have
read
the
register
draft
button,
0
1
and
for
flexi
client,
and
there
are
two
use
cases:
wines
are
freaks
area
into
the
audio
effects.
Yet
that
is
a
signal
on
Anya
I
have
no
program,
but
the
other
one,
the
section
4.5
yeah
yet
another
example
and
a
flexi
client
into
the
ODU,
but
I
cannot
this?
S
L
T
L
T
So
I
do
what
I
can
here?
First
of
all
I
test
us
thanks
to
the
this,
doesn't
look
the
one
okay,
this
one
look
correct
to
the
one
I
saw
on
the
slide:
respiratory
earlier,
didn't:
okay
yeah,
so
we
have
a
draft
that
we
call
flexi
framework
signaling
framework,
it
very
often
signaling
for
configuration.
T
So
actually
both
the
document
is
just
now
in
from
a
label.
Just
informational
I
heard
that
people
want
to
make
it
experimental,
though
I
don't
really
understand
what
an
experimental
framework
is
is
kind
of
odd.
So
if
we
go
there,
someone
has
to
explain
it
to
me
and
so
look
at
what
happen
with
this
draft
since
lost
time
and
lost
time
was
Chicago
and
I
wasn't
in
Chicago.
So
partly
for
me,
it's
like
little
bit
longer.
We
have
added
two
new
contributors
and.
T
The
number
of
contributors
are
going
up
all
the
time
we
restructure
the
document
quite
a
bit
when
I
first
read
this
document.
It
just
struck
me
that
it
was
a
flexi,
tutorial
and
I
think
we
need
a
flexi
tutorial
somewhere.
It
will
probably
write
something
for
the
idea,
fewer
nil
or
something,
but
it's
not
what
we
want
to
have
in
the
working
group.
So
here
we
need
a
typical
ITF
framework
document
and
we
cannot
tweak
it
to
become
that
a
little
bit
more
than
it
was
before.
T
So
we
put
priority
what
is
interest
for
ITF
and
I,
also
simplified
use
cases.
I
hope,
I
have
not
oversimplified
them.
People
need
to
review
and
you
can
give
feedback
on
that,
but
remains
to
be
done.
We
need
to
do
a
final
clean-up
on
the
requirements
requirements.
A
chroma
is
a
pretty
good,
but
sometimes
it's
not
requirement
language.
You
can't
a
requirement
is
something
that
you
can
measure.
T
T
T
Did
you
do
that?
Okay,
so
here
is
how
we
propose
to
split
up
the
documents.
We
do
one
framework
and
that's
the
document
we're
talking
about
today.
That's
the
normal
framework
architecture
requirement
use
case.
Information
goes
into
that
document
and
I
think
we
would
kind
of
not
stress,
use
cases
more
than
we
do
just
now,
and
then
we
need
signalling
and
routing
documents.
I
see
that
we
need
extension
to
rsvp-te.
We
need
extensions
to
the
IDPs.
T
There
is
one
document
that
Mac
is
going
to
talk
about
in
just
a
few
minutes
and
then,
since
this
is
link
technology,
it
needs
at
least
we
need
to
ask
the
question
that
we
need
to
touch
a
LMP.
It
could
be
the
case
that
you
might
must
do
that
and
then,
of
course,
where's
the
nor
we
need
a
John
model
for
for
flexi
Olson,
and
this
is
work
in
progress.
T
I
try
to
come
up
with
a
reference
model,
it's
kind
of
there,
but
I
need
comments,
feedback
help
to
refine
it.
What
it
basically
says
is
that
we
had
three
types
of
connections
for
flexi.
One
is
actually
when
you
connect
routers
back-to-back
the
bottom.
One
is
when
you
connect
it's
through
an
alkane
network,
but
where
the
area
network
is
not
aware
of
the
the
the
Flexi
and
the
last
one
is
that
you
have
an
OT
n
aware,
flexi,
aware
network
and
at
the
top
you
need.
T
We
need
to
kind
of
put
interfaces
reference
points
and
things
into
this
picture
and
I
need
help
to
do
that.
And
if
you
have
ideas,
please
contact
me
and
send
me
text
them,
so
we
looking
at
flexi
data
plane
just
quickly.
So
what
we
have
is
see
if
I
can
do
this.
This
is
animated,
and
with
this
one
it's
gonna
be
here.
So
there
we
have
a
nose,
can
I
do
a
pointer,
no,
that
one
okay.
T
So
we
have
notes
here
note
in
the
middle
note
at
the
end,
and
we
have
some
type
of
connectivity
in
between
them.
So,
first
of
all,
we
call
this.
It's
not
that
bad.
A
much
like
it
says,
flexi
link
here
and
we're
pointing
at
the
blue
big
pipe
here.
So
what
what
we
have
in
flexi
is
a
pipe
between
here
between
two
paranhos
from
here
to
there
and
from
here
to
there
and
then
the
next
thing
is
that
we
have
flexi
sub-links.
They
have
the
same
reach
ability.
T
T
U
I
was
gonna
question
the
diagram
you
have
that
middle
walk
component.
What
are
you
switching
there?
So
flexi
is
not
a
switching.
U
T
To
one
slide:
okay,
yeah,
so
this
is
why
we
have
the
what
comes
out
of
flexi.
The
interface
is
sub
interfaces,
links
and
siblings,
so
Ethernet
and
flex.
It
is
flexibly
that
are
a
link
technology.
So
here
me
getting
into
what
what
you're
talking
about.
We
have
a
we
want
to
establish
an
LSP
over
lexi
and
I
thought
and
think
we
can
do
their
so
in
this
example,
I
assume
that
the
Flexi
group
is
pre-established.
It's
not
necessary
if
we
could
also
potentially
establish
the
Flexi
group
by
bike
gmpls
configuration
though
there
are
problems.
T
So
if
we
accept,
for
example,
if
you
have
for
100
jig
files
between
two
nodes,
do
you
want
to
establish
one
two
or
what
size
would
those
two
be?
Or
would
you
want
to
establish
for
different
Lexi
groups
and
the
Shan
should
make
is
not
really
readily
available
coming
in
from
requesting
a
LSP
over
flexi?
T
So
next
thing
we
have
the
it's,
not
my
friend,
so
you
have
the
links
you
have
the
sub
link,
the
green
one,
and
then
you
have
the
flexi
LSP
and
to
end
over
the
who,
over
the
reflection
network
and
you're.
Actually,
switching
on
a
MPLS
label.
U
V
An
email,
so
there
are
actually
two
different
approaches
that
one
can
use
with
labels
and
flexi.
One
is
the
gmpls
approach,
the
others,
the
the
MPLS
approach.
The
gmpls
approach
is
that
a
label
does
not
exist
on
the
data
packets.
Rather,
a
label
signifies
calendar
entries
on
a
particular
flexi
link.
So
you
map
the
label
to
the
calendar
entries
that
the
data
is
going
to
go
into
that's
like
time
slots
right.
So
so
is
it's
just
like
map
to
time
slots
in
OTN.
T
And
you're
doing
where
I
was
going,
but
thank
you
what
I
was
going
to
say.
The
the
thing
is
that
when
you
come
to
the
green
thing
here
that
one
in
this
example,
you
establish
that
through
the
impeller
signaling
and
actually
telling
the
shim
and
the
switching
equipment
in
nodes
what
what
what
how
to
treat
them
and
where
to
place
the
MPLS
label.
Yes,
the
MPLS
label
is
requested
from
here
all
the
way
up
here
and
then
established
back
again.
So
and
you
do
both
you
can
also.
T
U
U
W
U
L
U
So
I
I,
don't
you
know
it's
just?
It
would
be
I
think
very
helpful
if
you
could
show
the
architecture
that
is
in
of'
implementation
agreement,
and
it's
also
in
the
Geo
at
7:09,
which
is
public
available
right.
The
amendment
one
it's
purely
I
mean
flexi,
is
purely
a
link
and
the
calendar
slots
are
terminated
each
side
or
an
OT
any
of
three
mappings,
but
then
you're,
switching
and
oh,
do
flex
or
a
GFP
type
container
and.
U
U
U
U
S
S
So
yeah,
maybe
at
maybe
at
it,
looks
like
Ethernet
cry
and
what
a
cry
aunt,
but
maybe
at
this
fight,
introduced
it
had
too
many
terminology.
Is
that
for
industry
clarify
on
these
points,
I.
S
W
W
W
W
T
T
L
A
Actually
I
think:
are
you
guys
on
the
picket
net,
flexing
it's
a
kind
of
link
technology
and
cannot
be
switchable,
but
actually
look
at
this
picture.
I
think
this
picture
it's
a
little
bit.
You
know
misleading.
It
sounds
like
to
give
up
people
a
kind
of
feeling
that
flux
can
be
switchable
so
because,
based
on
based
on.
U
It's
stepper
again,
I
think
lo.
We
here
is
that
you're
trying
to
control
that
flexi
link
and
and
it's
it's
not
controllable.
This
is
a
hard
adaptation
function
that
is
mapping
just
internally
that
MPLS
packets,
which
become
then
Ethernet
actually
right.
It
has
to
be
MPLS
mapped
into
Ethernet
thighs,
because
it's
mapping,
the
actual
you
know
fiying
coding
into
the
calendar
slots
and
you've
no
control
of
any
of
that.
That's
a
hard
adaptation
that
just
is
just
automatic.
U
T
W
D
D
M
I
think
I
already
commented
that
one,
the
ACS
meeting
your
document
I
agree
with
you.
Usually
we've
been
doing
a
GP
extension
within
the
cecum
working
group,
so
if
it
makes
sense,
it
does
belong
here
around
here.
Just
to
do
my
comment
again,
I
think
the
solution
on
the
problem
statement
here
don't
match,
because
it's
about
using
switching
capability
descriptor
to
do
something
which
is
not
aimed
at
being
switched.
So
clearly
it's
not
ready.
Yes,.