►
From YouTube: IETF103-CCAMP-20181107-0900
Description
CCAMP meeting session at IETF103
2018/11/07 0900
https://datatracker.ietf.org/meeting/103/proceedings/
A
A
Okay
for
the
load,
where
I
think
you
all
should
be
family
with
this
information,
but
here
I
would
like
to
mention
a
few
key
points
here.
It
says
that
by
participating
in
the
IDF
you'll
agree
to
follow
idea,
lures
and
I
give
process
and
policies,
especially
you
know,
our
patented
policies.
So
if
you
aware
that
any
IDF
contributions
is
covered
by
any
contribution
pattern
by
any
patterns,
so
you
must
disclose
otherwise,
you
know
you
should
not
participate
in
the
discussion
in
there.
You
know
what
include
so.
A
A
So,
as
usual,
we
have
their.
You
know
audio,
while
recording
so
please
be
in
front
of
the
mic
and
also
please
state
your
name
before
speaking
so
I.
In
this
way
we
can
know
who's
there,
something
bad
you
know.
Okay
meeting
echoes
so
I
think
the
chairs
are
in
the
meeting
echo
and
with
your
monitor,
you
know
our.
If
there
are
some
comments
from
the
meeting
echo
and
you
know
arm
yeah,
so
for
the
minutes
yeah,
you
know
our.
You
know
handsome
psychology
Oscar.
It's
not
here
he's.
You
know
always
quite
well
easy.
A
So
if
anyone
could
take
some
minutes,
you
know
as
Lewis
pet.
That
would
be
much
appreciated.
So
I
would
like
ask
how
men
that
may
be
in
tallow
are.
Could
you
help
us?
Okay?
Thank
you
yeah,
so
IPR
closes.
This
is
quite
very
important
right.
Yes,
so
you
delay
on
the
church
will
send
out
IP
appalling
before
moving
to
the
next
step,
for
example,
before
individual
job
becomes
a
booking
loop
document
or
you
know
tap
to
last
call
we
were
sent
out
to
you
know:
I
can
pull
into
the
list.
A
So
if
you
are
a
contributor-
or
you
know
our
authors,
so
please
you
know
the
sponsor
IP,
upholding
it
soon
as
possible.
I'd
want.
You
know,
it's
great,
lay
your
work
men
in
list.
So
it's
always.
We
encourage
people
to
use
the
list
actually
actively.
You
know
we
can
discuss
any
technical
stuff
on
there.
You
know
on
the
list.
For
example,
we
can
have
some
discussion
on
top
and
issues
on
some
jobs
and
also
we
could
introduce
some
new
jobs
on
your
topics
and
you
know
on
the
list.
A
So
another
important
thing
that
you
know
walking
loop.
You
know
consensus
is
determined
on
the
many
lists
you
know
rather
than
in
the
face-to-face
meeting.
So
even
though
we
we
could
get
some
consensus
here,
face-to-face
meeting,
we
still
need
to.
You
know
print
to
the
list.
You
know
for
more
discussion
so,
for
example,
some
people-
maybe
there
are
very
active
means
they
can.
But
you
know
they,
maybe
they
do
not
have
chance
to
be
here
to
the
you
know,
Mossad
place
so
I.
We
have
to
take
it
to
the
list.
A
B
So
update
on
the
status
of
the
working
group
since
the
last
meeting
that
we
have
a
new
RFC,
we
have
our
first
Eric
C
on
a
microwave.
So
congratulations
to
all
the
authors
and
all
the
people
that
contributed
the
to
this
achievement.
We
have
now
two
drafts
in
the
ASG
processing
them.
He
doubled
vm
interface
management
and
control
framework.
Actually
we
just
received
a
liaison
from
the
itu-t
on
that.
So
authors,
please
make
sure
you
address
all
the
comments
so
that
we
can
are.
B
We
can
proceed
and
the
other
one
is
our
first
young
document
in
particular
the
microwave
one.
We
requested
the
publication,
for
there
is
PPT
bandwidth
availability
document.
We
have
quite
a
good
number
of
working
groups.
All
of
the
working
group
drafts
being
discussed
that
day
and
some
of
them
are
not
them.
So
the
first
one
is
the
alarma
madula.
This
draft
is
probably
the
one
that
has
been
discussed
most
on
on
the
list.
B
It
was
two
revisions
in
been
include
uploaded
a
lot
of
work,
a
lot
of
improvements,
so
there
are
still
some
clarifications
to
be
at
vadhana,
but
we
believe
these
the
draftees
in
a
good
shape-
and
we
are-
this-
is
probably
the
next
the
draft
we
will
issue.
The
last
call
for
OTN
topology
is
obviously.
C
Youtube
cynical
Ericsson,
but
speaking,
is
PBF
liaison
manager.
My
understanding
of
the
alarm
draft
is
that
it's
up
to
version
5,
which
was
just
posted
and
that
the
comments
have
been
addressed
and
my
other
other
understanding
is
that
the
authors
are
looking
to
send
this
to
working
replies.
Call:
ok,
ok,
so
I,
don't
know
how
you
cherish
one
here.
B
Actually
references
have
been
fitted
and
alignment
with
the
requirements
in
RFC
sixty
eighty
seven
bees
have
been
a
big
done.
The
first
one
is
almost
Israeli
or
the
young
doctor
review,
while
the
second
one
is
as
more
dependencies
on
the
IETF
t
draft,
which
is
not
yet
fully
stabilized,
as
we
said
that
the
RSVP
T
bandwidth
availability
is
the
publication
for
that
draft
has
been
as
big
requested,
no
big
changes,
as
at
least
we
received
an
homage
or
not
dates.
B
Regarding
the
ID
w.zahn
IV
info
and
an
encoding
reasons,
we
received
a
lot
of
reagents
right
yesterday
and
and
the
day
before,
both
from
ITU
who
studied
15
on
the
WDM
of
framework
drafted,
so
the
one
that
is
in
the
SD
processing
and
on
the
l1,
a
CSM.
So
please
they
AB.
They
are
available
on
the
on
the
mini
list.
D
Status,
update
of
the
worker,
ok,
so
a
quick
recap
about
the
scope
of
the
design
team,
which
is
to
develop
use
case,
a
gap
analysis
of
how
the
idea
Frank
models
can
be
used
to
address
a
multi
domain
optical
transport
network
and
to
check
and
provide
guidance
about
how
to
use
them.
We
were
beyond
Middle
East,
a
conference
call
and
we
have
a
gift
of
area
where
we
track
progress
and
open
issues.
We
have
updated
the
applicability
statement
document
to
the
version
0
tree,
what
we
have
done
in
this
new
version.
D
You
erisa
it's
an
under
discussion
in
the
design
team
and
we
have
I
decided
to
change
a
little
bit
the
way
we
we
assume
the
PNC
abstracts
the
topologies,
for
mainly
reason.
Simplification
of
in
damages
own
code,
my
net,
my
answer,
and
we
have
used
the
new
tool
developed
by
Nanette
mod
working
group,
which
is
recently
adopted
as
a
working
up
your
document
to
folded
his
own
code,
because
his
own
code
lines
are
well
but
too
long.
D
They
didn't
fit
into
the
RFC
format
and
we
had
problems
in
the
past
versions
to
fold
to
put
this
text
into
the
our
document.
Thanks
to
this
tool.
Now
we
are
able
to
do
that
and
we
have
just
updated
uploaded
a
zero
for
version,
but
just
to
fix
a
sum
and
it
needs
and
formatting
issue.
We
discover
after
submission
of
the
see
activation,
so
no
technical
change
the
zero
for
versus
a
zero
A
zero
tree.
D
Okay,
a
quick
recap
of
what
we
are
addressing
is
a
network
which
is
a
multi
domain
OTN
network.
Every
node
here
is
assumed
to
be
capable
to
switch
in
the
other
layer
and
we
want
to
support,
set
up
an
end-to-end,
OD
and
audio
tunnel
and
to
create
a
services
over
this
tunnel,
and
we
have.
We
are
honored
at
this
moment
in
time
we
are
analyzing
the
what's
happening
at
NPI
number
one,
so
we
have
described
in
a
JSON
how
audio
topology
is
built
and
how
do
you
can
set
up
an
audio
service?
D
Okay,
what
are
the
open
issue?
We
need
to
complete
the
example,
so
we
have
a
two
items
to
address
in
particular
how
the
Plaga
D
is
used
to
teach
the
inter
domain
links,
and
then
how
do
you
you?
You,
you
understand
from
the
TDP
and
configure
the
consolidation
between
the
client
layer
and
the
Audion
layer.
D
Oh,
you
have
few
verification
questions
we
have
to
discuss
with
ET
topology
at
all
autos
and
we
will
do
other
offline
this
meeting
or
after
this
meeting,
and
we
complete
almost
the
existing
examples
so
for
the
next
step.
So
what
we
want
to
do
is
a
new
update
to
address
the
technical
issue
to
align
the
text
ability
tutorial.
We
have
to
make
sure
that
there
is
no
overlapping
and
if
there
is
overlap,
it
will
either
move
or
potential
relative
other
move
text
from
our
document
to
the
T
tutorial
and
we
make
references.
D
We
want
to
complete
the
examples-
and
we
add
the
other
example
based
on
priority,
and
we
are
planning
a
face-to-face
meeting
tomorrow
evening
at
6
p.m.
ok,
and
so
everybody
is
welcome
to
join
the
face-to-face
meeting.
We
will
announce
them
the
room,
we
have
a
rumor,
so
we
will
announce
that
on
the
mini
list,
them,
let's
complete
the
presentation.
Thank
you.
B
E
B
D
D
B
G
The
first
thing
is
that
we
created
github
to
put
all
the
layer
0
models
across
w.zahn
and
flux
grid
into
one
github,
including
impairment
all
actually
so
that
more
people
can
participate.
Actually,
indeed,
are
more
people
participated
because
of
this
github
and
one
major
changes
across
WS
are
in
flux.
Grade
we
created
a
common
layer,
0
types,
the
groupings
were
or
layer,
0
models
can
input,
and
so
that
we
don't
have
to
overlap.
G
Like
no
types,
application
code
and
well
length
of
Simon
policy
and
client
types
etc,
and
the
complete
documentation
of
theater
policy
model
is
now
available
and
in
particular
or
modules
they
require
tablets
and
specific
label
like
a
label
restriction
and
pendency
types
or
so
Tia.
Now
they
created
ot
entire
for
client
type,
so
we
actually
import
that.
So,
as
OTN
draft
is
stabilized
and
rotate
apology,
now
we
have
updated
or
modules
to
be
complete.
G
So
I
think
our
topology
model
is,
you
know,
probably
20
pages,
so
I
cannot
put
it
all
together,
but
it's
in
the
key
table
so
take
a
look
or
in
the
draft.
If
you
have
any
comment,
please
leave
comment
on
the
github,
so
I
think
now
is
on
par
with
the
OTN
topology
model
in
terms
of
maturity
and
completeness,
and
we
hope
that
we
can
move
for
the
working
group
last
code,
but
before
we
do
that,
probably
a
young
doctor
review
would
be
required.
H
G
B
G
G
G
G
Yeah,
this
is
a
tara
model
for
flexgrid
media
channels,
so
I
think
same
information
applied
layer,
zero
flexibilities
in
the
same
key
top
and
layer.
Zero
type
module
share
across
w
sense
looks
great
and
it's
a
complete
augmentation
of
th
tunnel
model.
Now
in
terms
of
specific
labels
and
bandwidth,
I
think
I'll
get
to
details
questions
shortly.
G
So
is
just
some
example
of
that.
Another
Tony
page
young
model,
okay,
so
this
one
is
a
little
more
unstable
because
t
tunnel
model
is
not
as
stable
as
t
topology
model.
So
maybe
hopefully
we
can
move
once
t
to
know.
Move
tourists
secure
and
now,
pending
issue
is
why
we
get
away
from
m
and
n
integer
using
frequency,
just
a
style
issue
and
style
and
also
table
stone.
We
use
frequency
and
CDMA
ability
emits
or
frequency,
and
also,
if
you
know,
label
restriction
stats,
it's
all
using
frequency
unit.
G
G
But
I
think
theater
and
also
have
really
said
you
know,
for
human
to
configure
and
a
man
would
be
easier.
I
have
no
preference.
Maybe
Iowa
may
have
some
idea
from
implementation
standpoint.
We
can
change
and
a
man
for
flexgrid
integer
or
live
it
as
a
frequency
level.
It
really
doesn't
matter
for
me,
but
you
can
comment
on
that.
I
Gabriela
Francisco
yeah,
you're,
right
Hank
and
my
suggestion
is
to
apply
the
6205
that
is
working
either
for
DWDM
and
also
for
sson,
so
NMM
can
be
applied
also
for
little
yam.
You
just
have
to
specify
the
central
frequency
the
channel
spacing.
Thus,
it
is
are
all
integer
numbers
and
then
you
get
rid
from
very.
E
I
J
E
E
J
Simple
S&W,
sound
stuff
opt
frequency
that
we
use
our
integer
number
instead
of
instead
of
floating
point,
because
this
is
according
to
the
ITU
spec
right.
This
is
the
one
and
second
is
if
I
use,
Internet
and
and
M
I
think
from
the
implementation
inside.
We
have
to
convert
it
into
frequency
anyway,.
E
E
H
I
J
G
K
K
K
G
G
So
just
related
modules
are
the
main
module,
ITF,
l1c
or
CSM
is
a
type
of
CSM
and
we
create
a
separate
service
type
for
layer
1
and
we
also
import
ITF
young
types
again.
Github
is
created
for
layer,
1
CSM
model
as
well
under
a
CT
and
CMI.
So
if
you
have
any
comment,
please
visit
this
site
and
leave
your
comment.
There
changes
since
last,
ITF
is
basically
pre
working
group
last
called
young
doctor
review
is
completed
and
thanks
to
rob,
I
believe
Colton,
right
and
changes
are
incorporated
and
I
just
met
him
today.
G
A
H
B
B
B
G
G
Actually,
when
Herman
and
I
published
in
secret,
many
people
already
jumped
on
comment,
but
that
didn't
mean
to
be
completed
and
we
even
changed
the
name
of
the
draft
from
w.zahn
to
optical,
to
convey
that
this
is
not
just
fixed
but
also
flux
grid.
So
we
had
a
lot
of
offline
meeting.
Even
before
we
came,
and
yesterday
we
had
a
great
meeting
with
all
those
authors
and
contributors
and
and
working
group
chairs.
G
So
this
there
are
other
information
in
payment
validation
model
that
proposed
by
curtain,
cut
really
and
other
people
I
think
they
represent
later.
The
difference
is
this
is
a
network
model
and
we
want
to
base
a
theater
policy
and
Douglas
areflexia
topology,
the
installment
that
ii
has
secured
mentees
have
been
building
so
that
it's
not
just
interface,
but
it's
a
te
model.
G
So
that's
and
also
it
user
layer,
0
types,
common
groupings
that
we
just
defined
in
w7
flats
grid
so
that
it
can
be
consistent
with
in
of
them
from
you
know,
modeling
perspective
is
just
another
augmentation
of
that
so
model
I
think
t
link
attribute.
Is
we
argument
that
to
describe
fiber
don't
bear
with
this
parameters
yet
because
this
is
work
in
progress
we
just
show
as
a
simple-
and
we
also
put
some
when
light
passes
are
created.
G
We
put
some
which
label
there
corresponds
to
and
what
are
the
parameters,
but
this
probably
not
a
topology
model,
so
this
might
be
tunnel
models.
So
when
you
have
tunnel
model
for
environment,
maybe
this
part
where
moved
there?
Oh,
not
I'm
not
sure
at
this
moment.
What
do
we
need
impairment
tunnel
model
as
a
separate
model?
I,
don't
know
yeah
it
we
can.
We
may
be
able
to
use
W
Sentinel
model
or
you
know,
flex
return
or
mother
I,
don't
know,
but
we
discussed
a
later
yeah
this
one
install
what
people
have
different
views.
G
So
we
have
available
operational
modes,
which
is
I,
think
least,
and
then
which
one
to
choose,
and
one
of
them
is
I
to
T,
is
six
98.2
code
and
there's
some
other
information,
but
this
one.
It's
unsettled
people
debate
whether
we
have
to
have
explicit
you
know,
attribute
defined
or
application
code,
or
modification
of
that
will
be
sufficient
or
not.
So
so,
just
to
give
you
a
background,
this
is
application
code,
G
de
six
982
and,
as
you
see
here,
P
means
or
bi-directional
inference
one
and
these
TW
depreciation
and
so
on.
G
G
G
G
Yeah,
so
this
code
is
now
at
least
it
address
hundred
k,
which
is
better
than
previous
version.
So
I
think
the
issue
is
how
to
model
itt
application
code
versus
vendor-specific.
If
you
recall
the
recent
draft,
we
actually
allows
that
in
the
future,
when
the
specific
code
can
be
added.
So
if
this
code,
plus
some
additional
parameter,
need
to
be
added,
I
think
we
have
freedom
to
do
so,
and
another
issue
is:
is
a
vendor
specific?
G
M
M
In
my
opinion,
basically,
don't
you
have
to
solve
all
these
disintegration
problems
that
you
need
to
solve
when
you
disintegrate
optical
termination
from
optical
line,
when
you
also
do,
for
example,
disintegration
of
optical
line
itself
and
connecting
same
optical
segments
provision
in
different
domains
by
different
vendors?
It's
not
only
about
application
calls
and
modulation
codes
right,
CX
and
stuff
like
that.
You'll
have
to
consider
all
the
parameters.
It
needs
to
be
considered
right
when
you,
when
you
try
to
interconnect
alien
wavelengths.
G
Yeah,
actually,
current
scobie's
is
oil-less
segment,
a
list
within
one
all
red
segment.
How
we
use
our
application
code
to
set
or
configure
transponders
from
different
vendors
I
think
that's
actually
narrow,
focus
and
I
think
you're
addressing
a
big
or
a
problem
that
optical
network
since
day
one
have
has
been
facing
so
that
one
I
don't
know
I
may
be.
Court
may
have
some
answer
you're
about
to
write
yeah.
M
M
The
chairs,
okay,
so
do
you
do
think
it
is
reasonable
to
consider
say
a
CT
and
multi-domain
scenario:
an
MVC
controls
obstacle
tunnel,
okay,
the
ghosts,
the
multiple
domains
stacked
in
in
one
demand.
One
termination
point
goes
through
one
demand
and
another
demand
third
domain
and
then
on
another
termination
point.
So
in
a
similar
way,
as
we
consider,
for
example,
og
internals
well,.
B
E
M
B
This
doesn't
mean
that
we
don't
need
to
work
on
on
this,
in
the
sense
that
you
might
have
multiple
optical
domains,
control
divided
and
the
same
mdac,
but
controlled
individually,
in
the
sense
that
you
don't
necessarily
need
to
have
an
optical
part
that
spans
multiple
domains.
But
you
might
have
multiple
optical
domains
with
the
trails
initiated
and
terminated
inside
in.
B
M
B
M
M
Because,
like
I
said
in
before,
we
can
consider
such
a
concert,
we
need
to
solve
the
disintegration
issues
like
open
can
fix
all
this
integration
between,
say
termination
and
optical
line.
Before
we
can
start
even
considering
multi-domain
obstacle,
whines
no
yeah
anyway
get.
This
is
a
bigger,
bigger.
F
F
So
what
are
we
doing
really
so,
in
theory,
I
believe
we
should
work
here
under
the
assumption,
as
if
there's
a
some
technology
with
independent
of
any
vendor
or
producer,
and
that
we
need
to
set
up
a
let's
say
a
trifle
that
now
how
that
is
mapped
to
a
real
implication
where
you
have
different
vendors
and
there
are
different
controllers
and
into
different
things.
It's
certainly
something
that
need
to
be
discussed,
but
probably
not
with
that.
One
I
think
we
should
really
work
here
as
we
work.
J
On
slightly
different
topic,
I
was
asking
about
the
augmentation
and
because
the
current
model
suggests
to
augment
directly
from
the
T
topology
right.
Well,
I
think
that
for
optical
impairment
it
it
applies
only
to
do
it
fixed
or
flex
grid,
and
it's
optional
parameter.
So
do
we
consider
argument
instead
of
from
t
topology
augmenting,
w
sound
and
flex
create
models
actually.
G
That's
a
fair
statement
when
this
model
was
walked
out
between
home
Ian
and
me
and
my
colleague
in
HQ
a
flex
grid
model
was
not
stable
so
and
we
saw
this
one
addressed
not
only
fixed
or
flex.
That's
why,
with
it,
but
I
think
your
point
is
well-taken.
We
can
augment
both
w
on
flex
grid
on
a
separate
augmentation
and
we
can
duplicate
the
work.
That's
what
you
mean
right,
yeah.
J
N
N
So
there
are
several
use
cases
where
we
can
use
colored
interfaces
or
this
kind
of
technology
in
the
network
so
having
this
kind
of
interfaces
in
a
simple
peer-to-peer
connection
in
the
aggregation
or
peg
hole
network,
as
well
as
Igor,
addressed
in
a
multi
optical
domain
scenario,
and
there
are
not
all
points
or
issues
covered
yet
and
in
that
case,
but
for
certain
other
scenarios
that
we
can
run
in
the
network
today,
this
work
is
really
appreciated.
I.
D
Agree
with
Igor
that
you
cannot
today
with
the
technology,
you
cannot
span
an
optical
domain,
an
optical
path,
to
multiple
domains,
but
I
also
agree
with
an
area
you
can
create
an
end-to-end
ODU
and
then
trigger
the
optical
path
set
up
by
the
PNC
by
truly
MVC,
so
I
see
in
a
multi
domain
optical
network
still
available
to
a
real
city
architecture.
Of
course
you
have
to
go
through
three
re
generation
between
one
domain
to
the
next
and
this
the
job
of
the
ndse.
Okay,.
G
Okay:
let's
continue
on
terrorist
yeah
I
think
the
Tsun
is
the
one
we
need
to
augment
w
some
flux,
great
model.
I
think
that's
what
I
was
comment.
I
will
work
on
that
once
this
is
a
good
idea
to
go
on
and
another
one
is
I
think
this
is
Joan
Jonah's
medicine,
a
creo.
He
commented
order,
we're
gonna
model,
not
Ouellette
itself
but
or
well.
I
spend
because
one
fiber
has
multiple
span
and
to
catch
a
power
and
OS
NL
on
each
segment
might
be
need
to
be
modeled.
I,
don't
know
yet.
G
But
if
that's
enough
interest,
we
can
do
that,
but
today's
model
doesn't
support
that
and
also
he
also
mentioned
that
there
are
events
back
modulation,
type
and
open
or
the
MSA.
They
define
hundred
DP
QPSK
with
staircase
fare
and
upcoming
YF.
They
also
defined
400
DP
16
calm,
but
Hamming
as
defy+
staircase,
HD
effort.
I,
don't
know
what
that
means,
but
there
are
different.
Proprietary
versions
are
coming
down
the
pipe.
G
So
just
gonna
give
you
some
heads
up
what's
going
on
in
industry,
so
next
step,
I
think
if
this
is
a
good
idea,
we
can
refine
the
model
and
also
we
can
work
with
other
draft,
and
then
this
one
is
actually
exerting
them.
This
is
outcome
of
yesterday's
offline
discussion
and
I
noticed
that
there
are
two
camp
in
depending
on
where
the
device
configuration
it's
located,
whether
it's
a
part
of
controller
or
part
of
a
device
management
would
like
ENS
like
fashion.
G
So
if
there
are
two
mode
operational
mode
by
domain
controller,
that
means
device
configuration
is
outside
of
domain
control
so
that
those
yellow
line
is
kind
of
simple
operational
mode.
But
if,
if
device
configuration
is
residing
within
each
team
controller
than
the
information
that
you
need
to
configure
it's
much
more
detail,
it's
a
colored
as
a
blue
so
depending
on
which
one
is
your
implementation.
G
We
have
to
support
both
model
look
like
and
then
this
is
a
chairs
rendition
of
those
information
into
this
good
summary
look
like
we
can
do
both
and
I
think
the
draft
tyranny
did
only
interface
my
feet
in
device
configuration
to
the
transponder
block,
that
is
outside
of
yeah,
and
the
the
scope
of
this
document
that
I
presenting
is
actually
initially
for
TNC
to
MDS
C.
So
this
is
more
operational
mode
MDS.
M
So
my
question:
another
question
is
that,
according
to
eat,
ola,
the
the
reasonable
scenario
is
that
you
are
setting
up,
say,
odeon
tunnel
going
from
multiple
domain
and
just
to
support
their
segments
of
this
OTN
a
tunnel,
you
might
trigger,
say
optical
trails
within
each
domain,
which
fully
contained
within
the
domain
right.
So
if
this
is
correct,
then
I
also
heard
that,
at
the
beginning
of
your
presentation
that,
unlike
all
other
presentations,
this
one
is
exciting
because
you
are
collecting
optical
environment
information
from
different
domains.
M
O
J
I
Really
speaking,
just
to
clarify
the
question
from
from
Igor,
you
might
have
different
kind
of
let's
say
accuracy
in
the
optical
image
calculation,
either
in
MDS,
C
and
in
the
optical
controller
in
the
optical
domain
controller.
But
that's
why
we
have
two
different
kind
of
parameters
that
are
shared
and,
of
course
nothing
prevent
MDS
c2
ever
let's
say
a
very
high
level.
P
M
This
still
doesn't
make
sense
because
say
I'm
doing
whatever
I
need
to
do
inside
my
domain
to
support
my
audience.
Edmund,
okay,
so
and
I
could
even
assume,
quite
simply
that
whatever
calculations
are
happening
because
it's
done
by
TNC.
It
could
be
very,
very
well
proprietary
to
to
this
domain
Gaelic,
for
example,
we
in
other
we
did
computation,
which
are
totally
different
from
commonly
accepted
computation
of
the
optical
domains.
For
example,
we
never
used
optical
called
application,
calls
at
all.
M
We,
we
simply
computed
the
optical
terms
based
on
the
engineering
rules
provided
by
our
optical
engineers,
okay,
and
so
so.
What
I'm
trying
to
say
is
that
what
I'm?
This
has
to
say?
Why?
Why
why
I'm
this,
you
would
care
about
the
optical
impairments
inside
the
domain,
so
what
kind
of
preferences
can
actually
specify.
D
D
D
G
Think
basic
idea
is
that
we
need
to
encompass
these
two
mode
of
operations
and
depending
on
which
one
you
choose.
The
details
are
determined
by
that
and
I
don't
want
to
talk
about
multi
domain
at
this
moment.
No,
it's
a
single
domain,
whether
this
works
or
not,
and
then
I
think
egos.
Question
is
valid
when
we
have
multi
domain.
Yes,
that's
a
different.
D
F
So
get
gamma
and
try
to
come
back
to
my
earlier
statement.
I
think
one
of
the
the
issues
I
see
in
particular
here
is
that
we
associate
a
domain
with
a
vendor
thing
which
well
it
might
be
the
existing
case,
but
I
think
we
should
go
away
from
the
notion
of
vendor
X
test
this
or
whether
X
there.
Why
does
something
else?
F
A
Yeah,
thanks
for
the
comment
from
GERD,
so
I
think
from
there
you
know
our
country
blend
perspective
domain.
Here
is
many
about.
You
know
administrative
domain,
for
example,
on
this
kind
of
how
to
define
the
domain.
It's
more
like
a
kind
of
you
know,
operators
policy-
it
could
be.
You
know,
vendor
specific.
It
also
could
be
technologies.
The
best
things
right,
yeah.
B
B
E
H
This
graph
is
all
about
apology,
so
we're
we're
currently
discussing
the
scope.
I
think
it's
all
about.
That
means
we
have
to
work
on
on
what
the
what
the
link
attributes
the
te
debt
and
includes
our
four
layer
zero
and
identify
those
specify
those
we
have
started
some
discussion,
how
we
actually
should
model
the
capabilities
of
the
termination
points
which
are
the
optical
transponders,
but
there's
a
lot
of
work
ahead.
Right
and
I
would
strongly
suggest
that
we
consider
it
a
simple
case
first
and
see
where
we
can
get
before.
M
If
you
think
about
this
I
P
network,
they
really
consider
this
multi
domain
right.
It's
usually
it's
a
single
dimensional
control
and
many
devices
happen
to
work
in
the
same
domain
and
in
a
interoperable
profession.
So
this
is
not
what
happens
in
the
transport
world.
Okay.
The
reason
why
we
have
so
many
domains
in
my
opinion
is
predominantly
because
each
vendor
has
its
own
solution.
Okay,
so
so
I
cannot
see
like
you
have
like
two
domains
there
in
and
in
each
of
this
domain,
say
hallway:
our
works
was
no
care
in.
M
M
This
is
exactly
the
reason
that
we
can
actually
handle
the
network
as
a
multi
the
way
multi
way,
and
in
my
opinion
it
almost
always
the
case
it
came
out
to
be
internal
to
the
mains,
so
you
have
one
domain
which
is
hallway
domain.
Another
domain
is
look
at
the
main
and
we
have
a
interoperability
solution.
Safe,
say,
throw
northbound
interface
how
to
deal
for
this
cases.
Okay,
it
is
not
like.
We
have
a
administrative,
the
Fae
and
I
get
to
choose
those
devices.
B
M
My
question
is:
do
we
really
need
to
come
here
and
discuss
a
single
domains
in
the
winter
solution?
Okay,
so
how
long?
We
should
keep
dancing
about
the
issue
that
if
we
are
talking
about
like
a
single
domains
in
those
vendor
solution,
it
is
good
to
know
about
half
things
happening,
but
it
has
nothing
to
do
with
standardization.
However,.
G
G
M
A
Also
have
one
comment
on
egos
question
so
actually
I
think
right
now
more
and
more
people,
especially
as
operators,
our
customers,
they
are
looking
for
some.
You
know
solutions
for
the
multi-domain
mat
vendors,
you
know.
So
this
is
why
we
are
defending
a
city
and
all
your
models.
For
that
kind
of
thing.
We
are
trying
to
find
a
kind
of
open
interface
for
that.
So,
if
we
are
just
trying
to
solve
some,
you
know
think
domains
think
vendor
in
solution.
I.
Think,
of
course,
we
are
not
going
up.
L
A
L
L
Adrienne
Farrell
so
I'm
not
going
to
say,
let's
concentrate
on
a
single
vendor
single
domain
solution,
but
I
want
to
say
that
there
is
some
benefit
in
having
a
core
standard
approach
that
other
that
single
vendors
can
pick
up
and
use,
and
that
benefit
lies
with
the
operator
having
a
common
understanding
of
how
vender
domains
work
and
the
the
ability
to
transfer
knowledge
and
the
ability
to
have
a
a
common
diagnostic
approach.
And
you
know
much
much
though.
L
We
all
hope
that
gmpls
would
be
a
multi
vendor
multi
domain
end-to-end
solution
and
everybody
would
be
plug
and
play
interoperable
I.
Don't
think
we
got
anywhere
close
to
that.
But
Jim
PRS
has
seemed
quite
a
lot
of
deployment
in
single
vendor
domains
and
the
benefit
there
is
definitely
visible
in
in
it
being
the
standard.
J
Q
It's
about
a
GM
GM
B&O
I
had
a
big
sign,
applicability
draft,
it's
because,
but
this
is
not
the
first
time
away
for
analysis,
because,
according
to
comments
from
working
world,
I
think
it's
a
London
meeting
that
had
a
way
saying
that
there
is
no
political
extension
existed.
I,
don't
need
to
update
the
framework
document
and
when
we
need
to
submit
an
a
political
ability
document
so
which
other
supplement
a
new
document
to
do
this.
Okay,
next
and
this
society
is
a
part
of
the
scope
and
aim
of
this
draft.
Q
The
reason
that
way
stopped
to
update
framework
document
is
because
we
think,
has
some
uncertainty
about
how
to
configure
using
links
to
a
list
and
this
I'm
shooting,
as
you
may
have
some
relationship
with
the
progressing
addressed
deal.
So,
just
after
that
discussing
a
model
also,
they
were
saying
I
wish
to
the
limited
us
go
through
how
to
set
up
all
the
okay
and
audio
flex
hours
he
over
configured
using
link
so
currently
wishing
that
how
to
configure
using
link
and
nothing
in
the
scope
over
the
surfaces
of
future
work.
Q
Currently,
so
the
aim
of
this
traffic
to
evaluated
whether
recurring
use
a
current
active
friend,
the
gem
has
a
mechanism
for
you,
okay,
to
apply
that
in
the
audio
saying
link
in
this
case.
So
they
say
the
aim
of
this
draft
next.
Okay,
in
this,
in
this
slide
with
there's
some
comparison
between
that
ot,
okay
and
audio
same
case
in
after
the
discussing
which
things
that
are
similar
to
audio
KR,
o
tok
healing.
So
there
is
in
the
universe,
excuse
audio
settings
can
also
be
represented.
Q
Okay,
audio,
send
healing,
has
a
automated
hierarchy
of
the
multiplexing,
so
wishing
has
no
need
to
define
new
signal
height
for
days
for
the
OSPF
he
extensive
defending
current
FC
and
as
there
is
no
new
signal
type
introduced
by
wishing,
I
say,
as
I
failed,
Chaeronea
I'll
say
ve
have
achieved,
it
should
not
be
attended.
I
shall
not
be
extending
that
we
can
cover
that
either
so
voicing
the
current
that
they
find
that
single
tackle
that
is
out.
Q
Do
the
other
1/2
circle
of
flares
can
Cavaliers
need,
so
that's
a
dependent
scope
that
the
union's
draft
and
for
audio
send
healing
missing
only
ones
rather
granulated.
They
find
us.
So
it's
a
GP.
Thank
you.
So
there
is
no
need
whether
white
ties
this
information,
as
if
amazing
can
you
have
attained
about
torrents
of
the
audio
same
ink
so
yeah.
This
is
a
robot
a
routine
practicing.
Q
Ok,
so
after
discussing
that
amount
of
the
way
we
come,
we
come
to
a
conclusion
that
way:
sync
existing
porticoes
defining
apps
a
the
current
existing,
obviously
for
single.
Only
and
writing
over
audio
K
can
be
reused
in
the
case
of
configuring,
audio
chirosecure
audio
for
exams,
G
/
audio
sync
with
us
assume
that
what
you're
sending
I
read
it
configures
are
ready
here.
Q
So
there's
no
new
protocol
attention
needed
currently
wishing
Celine
based
on
scope,
and
we
also
have
seen
wishing
that
it
would
be
useful
to
publish
an
Apple
II
capability
statement
document
at
this
query
in
details
how
these
protocols
can
be
reused
so
yeah,
okay.
Next
next
step,
we
will
request
for
working
goal
for
per
thousand
with
regard
to
the
Sun
content
in
this
second.
E
F
Q
Don't
think
we
need
to
convey
old
Hills
and
because
it's
a
one
kind
of
a
second
layer,
it
will
be
some
time
that
we
do.
It
will
be
automatically
configured.
So
we
don't
need
to
do
it.
I
doesn't
any
yeah,
but
for
oto
saying
sometimes
it
for
out
using
anaconda.
It
can
be
configured
as
a
result
of
something
like
a
server
layer.
Once
you
finish
the
configuration
of
a
sub
layer,
the
OTO
thing
where
we
are
so.
Q
K
F
D
Ii
thought
of
speaking,
the
big
difference
between
the
audio
Chien
and
the
ODU
K
is
that
the
audio
K
is
a
switching
layer
in
the
electrical
domain.
The
illusion
is
not
switched
in
electrical
domain
so
far
for
what
is
the
scope
of
this
draft?
We
assume
that
the
illusion
is
set
up
and
we
we
have
it,
and
this
is
our
underlying
infrastructure,
seven
layer
and
which
set
about
UK
over
it.
So
the
eau
du
chien
formed
from
the
front.
The
audio
case
at
our
perspective
is
like
the
OTT
UK
yeah
yeah
I.
F
Understand
foot
from
mechanism
perspective,
so
if
you
have
a
several
I
already
set
up,
you
can
pull
up
let's
everything
lines,
but
so
the
same
you
could
say
for
the
o2
you
so
once
you
set
up
the
o2
you
all
the
rest
is
already
use
it
up.
So
now,
but
the
o2.
You
is
really
what
goes
on
the
line.
The
audio
Sen
goes
in
no
too
honest
again,.
K
D
Yeah
we
have
intentionally
decided
at
that.
We
say
we
have
a
node,
you
see
N
and
what
is
below.
Do
you
see
any
is
out
of
scope,
because
then
you
have
the
OTC
and
then
you
have
all
the
optical
part
so
becomes
too
complex.
What
we
need
is
we
have
a
node,
you
see
n,
which
is
a
new
type
of
link
for
the
for
the
audio
K
setup,
and
we
want
to
say
how
we
can
set
up
all
UK
over
that
link
and
the
protocol
cities
are.
D
F
So
I
understand
that
you
want
to
limit
the
discussion
on
that
level.
However,
my
view
is
that
this
is
a
little
bit
kind
of
arbitrary
limit
to
just
say:
okay,
the
complex
thing
is
anyway
done.
So,
let's
assume
it's
there
and
then
discuss,
and
what
can
I
put
on
top
so
I
would
I
would
think
that
the
assumption
that
the
server
is
always
there
is
probably
something
that
needs
to
be
discussed
as
well,
because
it's
you
cannot
simply
say
a
let's,
my
oto
is
there
a
my?
F
Q
P
Get
in
the
in
the
draft
is
a
split
he
said
in
the
beginning.
There
are
two
type
of
scenarios
we
Shannara
wash
and,
for
example,
these,
on
the
top
on
the
right
in
which
OD
UC
n
Lincoln
to
Zn
link
is
1/1
granularity.
So
there
is
no
problem
because
the
setup
is
automatic
for
providing
good
taste.
This
is
not
the
case
below,
for
example,
and
we
decide
since
there
was
when
we
discussed
this
draft.
There
are
some
not
decision
on
SATA
2
later
from
all
the
photonic
part
that
we
still
unstable
instable.
P
So
we
decide
to
look
for
the
electrical
applicability
and
this
electric
applicability
is
provided
as
soon
as
you
have
audio
CN
link
already
provided.
So
it
is
not
in
the
content
of
this
draft
to
take
into
account
how
to
set
up
the
Odyssey
and
link
in
some
scenarios
and
Dictionary
is
describing
the
draft.
F
A
A
Q
A
A
A
S
Hello,
this
is
Amy
from
Huawei,
so
today
I
will
do.
The
presentation
was
a
young
data
model
for
microwave
technology.
S
So
this
changed
in
the
0
to
version
first,
so
we
have
a
several
changes,
stay
in
the
young
models,
but
basically
it's
kind
of
optimization.
It's
not
a
technical
change.
First,
we
change
several
types
of
change.
The
type
of
several
data
nodes,
such
as
the
link
nominal
bandwidth,
going
to
penalize
those
are
related
with
bad
ways
and
also
the
availability.
Then
the
another
changes
that
which
hilda's
a
reserved
manuals
from
read
and
write
to
it.
S
Only
because
we
agreed
that
those
parameter
cannot
be
changed
by
the
user
and
then
we
also
change
the
interface
route,
which
is
the
month
point
of
the
microwave
radio,
meaning
in
the
interface
model
to
an
optional
feature,
because
if
you
don't
want
to
France
at
acknowledge
model
to
go
through,
the
interface
model
could
just
make
knowledge
ease,
and
then
we
have
a
new
section.
This
is
according
to
the
discussion
of
masterchef,
to
expand
how
this
microwave
topology
is
related
with
the
client
technology,
especially
with
ethernet
technology.
S
So
we
conclude
that
the
ethernet
technology
is
that
no
burn
a
t-top
nourlangie
on
microwave
to
project.
Then
there's
another
comment
from
NASA
idea,
medium
to
say:
if
there's
any
possibility
that
we
can
reuse
those
models
to
add
technology,
and
actually
we
could
explore
more
we're
trying
to
add
a
new
section.
S
We
don't
have
any
concrete
tests
yet
and
then.
Finally,
of
course,
we
have
to
some
editorial
improvement,
so
this
so
shows
the
changes
of
cell
types
first,
for
those,
for
example,
nominal
bandwidth
or
current
and
wise
we're
changing
the
types
of
random
arty
types.
Band,
wise
I
should
be
afloat
in
the
32
to
unity
system
and
the
reason
is
scary,
quite
simple:
we
want
to
make
the
left
either
so
because
candle
from
attribute
float
point,
it's
not
so
human
friendly.
S
So
we
must
convert
a
certain
bandwidth
value
into
the
I
should
be
floating-point
coding
and
then,
after
the
discussion,
we
would
like
to
prefer.
You
need
assistive,
oh,
so
that
will
be
more
straightforward
and
then
another
change
is
for
the
availability,
which
changes
a
type
from
the
arty
types
percentage
to
decision
the
same
assistive,
oh
because
they
in
the
arty
types
percentage
it's
defined
in
unit
8
and
it's
not
a
load
Frankie
fraction
parts.
Well,
in
our
case,
our
availability
in
uni
halves
of
recognition
parts,
for
example,
ninety-nine
point
nine
or
something
exact.
S
So
we
would
like
to
change
that
to
dissimilar
sista
for
and,
however,
I
will
notice
that
in
the
resonant
LED
mode
are
based
discussion.
They
are
also
considering
to
rewrite
all
have
a
unbiased
and
the
arty
types,
so
percentage
a
is
also
within
the
consideration,
so
they
may
like
to
have
a
further
extended
to
allow
the
fractional
part.
S
S
Thank
you.
So
the
first
are
saying
is
that
we
want
to
be
clear
that
this
is
isn't
it
top
neurologists
over
90
technology
or
microwave
deployment.
It's
kind
of
the
clan
and
the
server
relationship
and
on
the
left,
you
will
say
a
very
single
case
that
the
Ethernet
link
is
supported
by
the
microwave
eternal
in
the
under
name
network
technology
and
while
the
microwave
kernel
is
supported
by
only
one
microwave
link.
S
I
think
this
is
without
any
doubt,
and
there
might
be
some
problems
that,
because
in
our
case,
we
normally
have
multiple
microwave
links
to
provide,
for
example,
a
higher
capacity.
Then
in
such
a
case,
if
the
isn't
ending
is
supported
by
the
microkernel,
that's
will
be
the
same
as
than
they
have
the
case.
S
So
we
are
looking
for
more
discussion
on
that
and
then
the
next
step,
as
the
first
way
is
regarding
the
how
to
make
the
model
to
begin
right.
And
so
actually
we
have
send
the
email
to
the
list.
I
think
said
in
September,
but
unfortunately
we
didn't
receive
any
reply.
So
same
charts
there's
limited
interest
to
make
that
use
for
other
technology,
because
the
current
over
we
are
just
have
the
microwave
background,
stand
where.
If
we
have
to
make
that
use
to
other
technology,
we
really
meet
some
help
or
input
with
some
more
experts.
S
So
this
is
an
open
issue
and
of
course
our
West
will
come
have
to
move
you
and
comments
and
I
said
I.
Think
I
mention
this
before
we
have
a
practice
round.
It's
a
IMT
story
mwt,
which
is
general
or
Monsieur,
and
those
top
knowledge
model,
and
several
others,
such
as
T
top
learning
model,
will
be
verified
during
the
practiced
and
there
the
code
implementation
is
starting.
S
S
B
Discuss
it
there
regarding
the
adoption.
Well,
we
have
a
teat
apology
for
all
technologies,
for
WDM
flexibility
and
I.
Don't
see
any
reason
why
we
should
have
had
one
also
microwave,
and
it's
very
good
that
yet
she
decided
that
we
use
the
ITF
models
for
very
large
character,
so
it
would
be
nice
to
have.
But
the
draft
has
a
working
group
documented
before
before
that
event,
which
is
in
January
right
yeah,.
B
B
K
Couldäôt
good
morning
this
is
a
vehicle,
Holley
and
I'm
presenting
this
work
on
the
framework
on
CP
control
for
OTN
Network.
So
CP
is
actually
a
quite
frequently
used
term.
But
now
we
just
limited
scope
to
the
OT
n
CP
and
we
try
to
have
draft
on
how
to
define
the
use
case
and
the
corresponding
requirement.
K
So
the
motivation
of
this
work
is
to
address
the
CP
problem
automatically,
so
we
have
a
list
of
functions
that
need
to
be
satisfied,
including
the
discovery,
the
configuration,
the
connections
that
helps
management,
so
on
so
forth.
So
with
the
growth
of
scalability
for
CPE
application,
there
was
a
highly
demand
on
how
to
make
this
kind
of
drops
in
an
automatic
way,
rather
than
integrate
a
lot
of
manual
labor
labor.
B
K
Great,
thank
you.
So
in
this
picture
we
show
a
reference
architecture
of
the
CP
control,
just
notice
that
the
CPU
location
is
just
a
access
to
the
OTN
network
and
the
domain
network.
So
controller
and
hierarchies
something
existence
was
very
mature
architectural
reference,
but
with
what
we
are
working
on
here
is
just
a
how
to
connect
this
one
into
the
CP
device
team.
K
It
works
that
access
to
the
OTN
and
because
the
Ural
is
the
equipment.
Cpu
box
is
physically
isolated
in
the
customers
office,
but
should
be
a
kind
of
configure
and
management
by
the
operator.
So
this
is
a
kind
of
remote
and
though
we
probably
need
the
existing
solutions
to
be
enabled
here
to
guarantee
the
access.
K
And,
moreover,
if
we
have
a
special
component
to
do
the
control
and
management
of
the
CPE,
we
will
have
corresponding
young
models,
including
both
from
this
control
component
to
this
orchestrated
system
and
also
from
the
component
to
the
CP
itself.
It
can
be
called
either.
Why
are
the
core
network,
or
directly
from
the
block
to
the
CP,
and
given
this
kind
of
different
variation,
we
can
easily
draw
some
kind
of
variant.
Architecture
example
saying
that
we
just
approve
the
CP
control
and
management.
K
The
component
in
two
different
place,
for
this
example
shows
how
it
is
embedded
in
the
orchestration
system
with
operators,
and
so
that
would
allow
people
to
configure
the
CP
directly
from
the
operators,
but
we
may
need
some
kind
of
models
here
to
do
the
configuration
and
GMP
RS
uni
is
again.
We
just
need
a
enable
a
list
and
we
are
looking
for
the
opportunity
whether
there
is
a
need
to
do
the
to
some
extension,
but
for
the
sauce
for
an
interface
for
the
multi
domain
interconnection.
This
is,
there
are
just
a
mature
tactics
here.
K
So
the
the
summary
this
work.
We
are
trying
to
provide
some
basic
background
and
the
needs
for
CP
control
and
we
have
detected
the
tube
potential
directions
that
we
may
need
to
work
on.
One
is
the
young
models
that
need
to
be
configured
on
the
CP
box
and
the
other
is
the
GM
PRS.
You
need
protocols
that
need
to
be
used
to
connect
to
the
CP
and
the
OTN
networks,
so
both
of
them
may
have
a
potential
need
for
the
extension
or
augmentation,
and
we
also
need
to
understand
more
about
the
efficiency
problem.
K
B
E
G
K
D
You
ammeter,
whose
I
am
presenting
on
behalf
of
Dakota
rough
for
the
Yamada
to
transport,
the
classiness
over
a
transport
network
we
have
made.
We
already
presented
this
graph
in
the
previous
IDF
and
we
made
a
few
changes
in
the
latest
version.
So
we
have
updated
the
transports
for
service
for
Internet
added
few
service
parameters
like
the
service
ID.
D
With
the
alignment
with
the
topology
default
from
T
types,
we
aligned
with
a
young
mother
from
younger
on
the
layer
one
service
type,
so
we
have
a
common
way
to
call
the
client
types
in
the
in
the
ACMA
and
in
the
NPI
and
a
DDM
in
status.
Configuration
and
the
transport
types
has
been
aligned
to
create
to
define
it
the
type
for
the
lifecycle.
D
So
a
quick
recap
on
the
transparent
client
signal.
So
we
have.
What
we
need
to
do
is
to
carry
client
like
SDH,
fibre,
channel
or
ethernet
physical
links,
so
the
coding
of
the
internet,
nor
the
frames
over
an
OTA
over
a
transport
tunnel,
the
transport
tunnel
can
be
any
technology,
can
be
OT,
n,
microwave,
for
example,
and
in
this
case
we
have
no
switching.
We
have
a
No
and
only
point-to-point
for
base
mapping.
So
what
we
need
to
say
is
which
port
is
and
which
client
is
as
attach
to
put
over
the
tunnel.
D
D
Because
we
can
have
multi
function
as
links
where
we
can
configure
different
type
of
clients,
and
then
we
can
say,
which
is
the
transport
tunnel
that
has
to
carry
the
client,
and
this
quick
snapshot
of
the
model
for
eternal
client
is
not
non-transparent,
is
the
case
where
you
want
to
set
up
internet
career
services,
for
example,
I
keep
ele,
VPL,
etc.
In
this
case,
the
stake
you
can
have
point-to-point,
or
also
multiple
in
term
and
routing
multi-point
connectivity.
D
The
access
links
are
all
Ethernet
based
the
frame
based,
but
they
can
have
different
rates
than
coney,
so
you
can
hang
gigabit
detail
on
one
side
and
then
Gigabit
Ethernet.
On
the
other
side.
There
again,
you
don't
have
a
tunnel
model
in
this
scenario,
but
you
still
need
the
topology
model
to
know
what
are
the
SS
link
and
we
have
seen
another
draft
submitted
to
address
this.
D
But
it's
outside
the
scope
of
this
document,
and
here
you
have
to
configure,
in
addition
to
the
basic
service
and
in
the
SS,
think
you
have
to
configure
all
the
classification
and
operations
on
the
villain.
You
have
to
configure
the
vendor
profile
and
again
map
into
which
transport
tunnel
you
can
they
can
carry
added
traffic.
D
So
what
we
have
a
few
comments
in
the
passage
about
generalization
and
the
map
pinger
between
the
service
and
the
transport
tunnel
is
already
generic,
because
what
we
provide
is
a
pointer
to
a
tea
tunnel
name
and
that
eternal
name
is
a
generic
attribute
in
dat
tunnel
model.
The
technology
specific
attributes
are
all
in
the
Tetons
model
as
a
technology,
specific
communications,
and
we
don't
need
to
do
any
any
any
specific
here,
depending
on
the
different
type
of
tunnels.
D
What
we
have
is
the
SS
information
is
client
specific,
that
the
reason
that
is,
that
we
there
is
no
commonalities
between
the
transparent
and
internet
classiness
found
and
more
or
less
what
we
are
mimicking,
is
the
same
separation
that
we
have
on
the
semi
between
the
air
masses
and
layer
to
SM
and
the
SM.
Even
if
we
are
not
addressing
really
services.
So
we
don't
see
there
is
any
advantage
to
have
an
almost
empty
generic
model
from
which
we
create
technology,
specific
commentation,
which
is
as
big
as
the
car
and
modest.
D
D
In
some
cases
we
need
to
map
service
to
a
specific
transport
tunnel,
as
in
this
current
raft,
for
example,
what
the
customer,
the
client
application
wants
to
specifically
say:
I,
want
this
service
to
go
over
this
tunnel
or
when
you
have
a
multi
domain,
technologist
scenario,
and
you
want
to
coordinate
the
two
ends
to
select
the
same
tunnel.
So
if
you
don't
have,
if
you
have
a
karuta
requirements
or
you
need
to
know
from
where
the
traffic
is
coming,
you
need
to
make
sure
that
the
both
handsets
take
the
same
tunnel.
D
And
then
you
really
need
to
point
to
a
specific
tunnel.
But
there
are
other
situation
where
you
want
to
say:
I
want
I,
don't
care
about
which
tunnel
is
taken
as
long
as
it
has
some
properties.
So
in
this
case,
what
you
can
say
I
want
this
service
to
go
over
this
tunnel
pool,
and
one
idea
that
we
are
starting
to
think
of,
but
we
can
refine
later,
is
to
have
a
sort
of
choice.
So
we
pour
either
point
to
a
tunnel
name
or
to
attend
that
pool.
D
So
so
we
think
the
document
is
ready
for
working
local
option
and
we
can
further
discuss
any
open
issues
and
further
improvements,
as
thank
you.
M
D
Yes,
yes,
most
of
the
services
that
we
can
support
our
TANF
services,
but
what
we
have
us
is
all
the
layer,
2
sm
services
on
the
CMI.
You
can
support,
also
them
or
using
this
model
on
the
NPI
or
FC.
No,
so
some
requirements
to
have
a
much
richer
capabilities
in
terms
of
villain,
manipulation
at
the
edge
that
can
be
done
so
map
services
are
definitely
a
subset
of
what
you
can
configure
with
these
models.
M
When
we
say
this
is
a
service
that
we
map
on
the
OT
on
tunnel,
we
can
refer
to
something
else,
or
we
are
talking
about
services
that
people
widely
understand.
So
basically,
my
my
question
is
that
in
my
mind
it
is
in
material,
for
example.
Well,
the
service
is
implemented,
say
on
IP
network
on
transport
network.
It
is
the
same
service
mm-hmm.
E
M
F
M
D
Perceived
by
the
customer
or
is
the
way
you
implement
the
service
in
the
network
and
sometimes
people
think
they're
VPN
is
what
the
customer
sees
and
then
I
agree
with
you.
There
isn't
no
difference
if
they
say,
if
you
say
the
layer,
2
VPN
is
the
way
implement
the
service
of
a
network
by
using
hello.
I
am
pls
LSP
with
maybe
BGP
and
only
stuff.
Then
it's
different
between
IP
and
transport.
D
M
M
F
Yeah,
do
it
speed
a
little
bit
that
that
kind
of
assumption,
right,
I,
don't
think
it's
it's
the
same
and
you
can
kind
of
fix
it
on
the
way
you
adapt
the
service,
which
is
layer
2
to
what
you
have
on
the
server.
So,
for
example,
if
you
just
there
was
a
what
many
years
ago,
the
GFP
mapping
in
novi
n
had
a
limit
in
the
size
of
the
packet
could
transport.
So
if
you
sent
a
jumbo
frame,
you
got
lost
it's
fixed
now,
so
just
bringing
it
up
as
a
it's
a
plastic
sample.
F
So
obviously,
if
you
run
on
a
router
that
may
have
a
specific
jumbo
frame
size,
which
is
not
very
well
defined
and
different
implementation
of
different
composites,
so
there
is
a
difference.
Definitely,
and
the
same
goes
for
for
the
mapping
itself.
So
how
much
transparency
you
get
from
you
for
Ethernet
service,
so
is
it
the
packet
switched?
Do
you
really
look
into?
For
example,
if
you
have
some
timing
information,
you
really
don't
want
to
have
a
GFP
map
you
want.
F
B
M
Q
It's
another
new
one,
but
it's
fun
so
the
pregnant
insane,
but
a
focus
on
young
model
child
shake
someone
discussing
or
what
can
we
do
to
free
technology
with
ietf
consume
mechanism?
The
solution
for
Policy
Council
discussing
in
this
protein
is
some
different
from
that
mean
that
the
frame
of
a
document-
okay,
next,
okay,
this
night
is
some
general
discourse
about
the
difference
between
flexi
and
traditional
internet.
It's
because,
according
to
a
discourse
and
other
finishing
in
the
fleshy
ie
or
after
the
Frankie
same,
it's
just
that
built-in
function
in
the
Usenet
appeasers
module.
Q
If
a
policy
corporation
said
have
a
congress
pounding
huge
pieces,
module
is
created
as
well,
so
doesn't
based
on
is
significant,
while
a
significant
difference
between
the
pharisee
and
a
traditional
Eastern
era
is
that
traditionally
Sunanda
pieces
had
a
one-to-one
relationship
with
a
kind
to
make
flow
while
flashy
he
had
a
more,
for
example,
structure
when
she
says
module
can
be
associated
and
be
associated
with
multiple
chronic
mega.
Follow
they
say
is
that
different
than
ever
that
I
was
saying
that's
most
a
significant
difference
and
the
just
based
on
the
previous
residence.
Q
My
analysis
at
the
way
sent
him
to
general
control
requirement
that
should
be
considered
for
the
first
one
is
what
whether
we
need
to
config
Ferretti
work.
The
second
one
is
the
whether
we
needed
to
allocate
the
for
ICD
corporate
calendar
slaughter
to
a
local
resources
to
the
colony.
The
thing
he's:
an
egg
honey.
Okay,
that's
a
base!
London!
Okay!
Next
is
the
next
aside.
It's
about
some
analysis
of
about
whether
we
needed
to
configure
for
a
seat
group.
Q
Okay,
that's
based
on
some
describe
singing
for
the
CIA
that
that,
when
I
knew
for
Isis
go
away
the
karate
in
two
surveys,
the
initial
configuration
should
be
prevent
a
football
signs
that
to
find
the
server
different
fight
together.
So
just
based
on
that
discourse
and
unify
see
a
that
was
saying
that
we
should
use
some
consume
magnesium
to
configure
the
phrase
equal
to
criticism,
they
say
and
the
next
day's
about
the
local
resources.
Q
Yet
for
the
kind
of
signal
we
visit
some
to
question
here,
the
question
is
how
to
allocate
the
dust
last
week
for
a
sequence
and
signalized
does
external
consume
needed
to
be
involved?
We
found
some
discourse.
I
mean
for
the
CIA
is
about
some
description
about
a
country
function
defining
in
the
flashy
is
the
control
function
can
be
used
to
put
the
flashy
crannies.
Had
a
salsa
in
the
into
the
philosophy
group
into
some
specific
estas
under
then
take
some
taken,
a
the
frat
he
planet
out
at
the
same
kinda.
Q
So
does
it
based
on
there's
a
way
similar
to
that
for
a
cdata
/
I
already
to
find
out
how
to
install
extract
the
flashy
kilometer
from
calendar
sauce?
So
we
don't
think
that
we
needed
to
control
for
that.
You
can't
you
want
to
do
enough
for
this
point,
I
think
it's
a
different
formula
that
this
working
in
a
frame
of
a
document.
Okay,
our
next
slide-
is
about
to
some
general
description
about
the
odor
control
anchor2,
it's
a
summary.
Some
of
them
may
not
be
included
in
current
model
job
of
addressing
the
task
Samuraizer.
Q
We
need
to
configure
that,
for
a
sequel
we
need
to
take
here
will
fly
away
a
case,
but
we
don't
want
to
attach
the
sauce
and
also
we
should
I
use
control
measure
to
trigger
the
switch
between
different.
It's
a
different
tab
place.
It
can
be,
it's
gonna,
be
a
I'll,
be
different
calendar
type
and
also
some
malama
that
we
needed
to
take
into
consideration
hot
okay-
and
this
is
a
pricing
model.
Q
It's
about
generally
stay
young,
a
valid
model
about
the
definition
this
society
is
about
that
this
curve
symbol
for
as
you
go
when
the
difference
is
throughout
the
community
and
they
for
a
5-pound,
a
that
I
can't
be
100,
I
get
100
grand
a
400,
o
gig
and
the
different
effects.
Ie
incidence
connects
later,
but
I
come
telling
the
configuration
to
be
different
type.
It
could
be
a
type
of
e
type,
also
some
acknowledgement
mode,
and
this
side
is
about
it.
Q
Some
describes
a
policy
cranny,
but
I
want
to
give
something
explanation
that
I
won't
wait.
We
don't
want
to
use
this
flashy
cranny
model
to
config
to
square
to
set
up
some
kinetic
configuration,
such
as
I
said
on
the
you
may
know
it
doesn't
want
to
report
the
for
I,
see
culinary
information
may
be
external,
maybe
controller
after
the
planetary
is
configure
data
that
you
couldn't.
So
it
doesn't
want
to
do
this.
B
Q
B
O
O
So
we
know
what
happened
when
I
T
use
that
until
the
MPLS
design,
Authority
holler,
if
you
want
they
can
we
give
you
a
very
illustrative,
a
description
of
that,
and
we
don't
want
to
make
the
same
mistake
so
I
don't
think
we
should
take
any
decision
here
or
any
kind
of
progressing
of
a
document
before
we
go
to
Deborah
and
the
working
group
shares
sort
out
what
type
of
relationship
do
we
have
the
o
AF?
Can
we
do
do
they
want
us
to
do
the
that
type
of
modeling
here?
O
Having
said
that,
I
think
it's
a
good
idea,
because
we
have
the
Jiang
expertise,
so
we
can
do
it
and
if
you're
AF
says
okay,
we
don't
care
or
they
say
well,
you
can
do
it
or
actually
we
are
very
excited
that
you
want
to
do
it
anything
like
that.
It's
fine!
We
can
go
ahead
and
do
it,
but
if
they
said
no,
we
want
to
do
the
Jiang
model.
Ourself
then
I
think
we
kind
of.
We
asked
me
to
stop
this,
but.
B
O
Q
I
Money
is
a
Gabrielle
investor,
speaking
on
behalf
of
quarters,
I
think
for
sake
of
time,
I
will
not
go
through
the
light
so
give
back
the
just.
To
summarize,
we
have
fourth
draft
here
to
relate
you
to
element
P
and
to
relate
it
to
a
modest.
The
two
related
to
the
LMP
are
for
DWDM
interfaces
and
flex
grid
interfaces.
The
one
related
to
the
DWM
interfaces
is
actually
addressing
the
application
code
I
to
the
application
code
plus
few
parameters.
I
We
have
discussed
with
the
itu-t
in
the
past
and
is
addressing
also
the
vendor
specific
application
code.
This
vendor
specific
application
code
will
follow.
Probably
the
discussion
for
sure
the
discussion
that
are
in
progress
for
the
with
the
other
draft,
like
the
younger
presentation
before
the
flexgrid
LMP,
is
actually
adding
more
parameters
that
I'm.
A
I
I
I
Last
I
think
there
is
a
excellent
discussion
in
progress
all
this
stuff
and
we,
of
course
we
need
to
coordinate
this
fault
raft
with
the
discussion
as
well.
In
my
opinion,
the
best
way
is
to
follow
the
upgrade
process
is
to
promote
the
the
draft
to
working
group
documents
and
open
to
the
other
author,
so
that
we
can
work
all
together
in
in
a
set
of
slides
that
are
all
related
together.
Thanks.
B
Is
now
in
post
working
group
last,
all
stages
of
which,
which
means
we
probably
go
to
the
RFC
editor
too
soon.
It's
probably
good
timing
to
move
forward
that
the
two
WDM
interface
drafts,
so
is
anyone
against
the
adoption
of
the
two
WDM
interface
drafts,
which
are
in
the
specific
DWDM
interface
parameter
yang
and
WDM
interface,
LMP.