►
From YouTube: DOTS WG Interim Meeting, 2020-06-25
Description
DOTS WG Interim Meeting, 2020-06-25
B
A
Yes,
we
can
hear
you,
okay,
thank
you.
Yeah,
okay,
I!
Think
it's
done.
It's
two
o'clock
at
my
time,
a
different
time
for
everyone.
I
come
from
different
area.
Okay,
but
anyway,
thanks
for
everyone
attending
our
ITF,
the
doors
open
will
open.
The
register
signal
ring
working
group
entry
meeting
this
time:
okay,
I'm
I'm,
Leon
Shah
and
my
my
my
culture
is
virus.
Smith
Smith.
Sloth
worry.
Are
you
idea.
B
A
A
Everyone
is
here
in
our
WebEx
and
okay
and
if
you
want
to
download
the
meeting
slides
you,
we
have
provided
our
link
here,
meeting
slides
and
we
have
our
interpreter
for
them
for
the
note-taking
and
you
can
go,
you
can
go
to
that
link
and
to
see
what
our
note
taker
has
write
down
and
the
later
anytime.
You
want
to
see
the
result,
something
you
can
see
that
okay
and
we
have
our
Trevor
Trevor
drop
link,
but
I,
don't
think
it's
a
it's
very
necessary
because
everyone
now
is
seeing
the
we
are
online.
B
A
Yeah
yeah:
yes,
yes,
because
now
it's
a
relatively
new
way
for
us
in
this.
In
this
new
age,
for
the
epidemic
for
the
core,
Nevada
pandemic,
I,
think
more
and
more
meeting
will
be
held
online.
So
we
we
do
not
have
paper.
We
will
write
down
on
the
online
on
the
hour.
Is
a
pattern
pages?
So
please,
please,
please
write
down
your
name
on
the
disappeared.
A
Okay
and
okay.
Here
is
some
tips
for
these
for
our
new
way
of
the
injury
meeting,
okay
and
so
let
so,
please
please,
please
aware
that
our
current
WebEx
inter
meeting
is
being
recorded
so
for
people
who
want
to
see
who
won't
know
our
discussion
later.
So
we
have
recorded
all
the
all
the
other
discussion
or
the
meeting
and
some
questions
please
make
sure
your
video
is
off.
Unless
you
are
you
want
to
present,
you
want
to
discuss
okay
and
mute
your
microphone
and
to
avoid
the
some.
A
If
you
are
not
very
familiar,
please
read
it
and
becoming
more
familiar
with
it,
because
that's
all
the
policies
that,
if
you
are
the
one
of
the
attendee
of
the
ITF,
meet
no
ITF
this
organization.
You
should
be
familiar
with
this
policy,
okay
and
okay,
and
then
this
we
need
some
help.
Firstly,
proof
sheets:
okay,
everyone
can
write
down
your
name
on
the
on
our
etherpad
and
subscribe
thanks
for
watering.
He
is
on
that
he's
taking
care
of
yeah
yeah
and
no
taker
and
the
sense
of
a
pad.
Here
we
are
P
our
note-taker
yeah.
A
B
Just
recently
published
two
policies:
87
82
and
87
83.
Is
it
a
core
of
dots
protocol
state
for
signal
channel
and
data
channel,
and
it
was
a
very
hard
job
and
thanks
to
everybody
involved-
and
it
was
particularly
a
lot
of
discussion
with
signal
channel-
arrived,
see
it.
Finally,
finally,
they
they
had
been
published.
So
congratulations
and
thanks
to
everybody
involved.
So,
yes,.
B
There
is
currently
a
long
delay
in
publishing
overseas,
because
I
receive
editor
switched
from
xmn
to
Russia
version
to
take
cement
to
Russia
version
3,
and
it
was
just
a
lot
of
burden
on
the
editors
for
publication
process
currently
beat
slow
down
and
fulfill
its
for
quite
a
long
time
in
the
edited,
but
I
hope
it
will
publish
soon
so
control
and
filtering
rules
draft
is
currently
in
is
G
evaluation
and
as
far
as
I
remember
I
just
looked
yesterday.
It
was
no
discusses
from
ASG
members.
Yet
so
we
we,
we
hope
it
will.
B
It
wasn't
developed
very
intensely
recently,
so
I
think
it's
it's
a
bit
of
stuck
and
we
we
need
your
energy
to
push
it
out
and
to
discuss
it.
I
see
no
discussion
on
the
list
and
about
multicom
and
raft
for
the
last
half
a
year
about
so
I
think
we
we
we
think
we
should
get
more
energy
to
walk
in
the
truck.
A
Okay,
yeah:
it's
no
problem.
Okay,
so
we
all
mostly
finish
you
to
our
our
our
session.
So
next
one
is.
We
have
a
relatively
long
discussion
about
mostly
important
that
draft
our
tell'em
Entry
Draft
and
then
we
have,
we
will
go
through
the
multihoming
draft
and
the
use
case
for
the
telemetry
draft
and
then
the
draft
of
from
China
Mobile
about
the
costs
of
deployment.
A
That's
all
for
our
today's
draft
presentation
and
if
we
have
more
time
we
came,
we
can
have
a
more
broader
discussion
about
our
car
in
the
charter
or
the
future
work,
and
then
we
we
can
wrap
up
today's
meeting.
So
so
so
remember
that
we
have
one
hour
but
I
think
it's
enough
is
enough
for
us.
So,
okay,
okay,
is
anyone
who
has
any
you
know
any?
A
A
C
C
You
so
next,
yes,
so
the
in
this
lot,
so
as
I
mentioned
so
I
will
provide
a
status,
and
then
we
will
zoom
on
some
key
I
would
say
design
points
that
we
have
made
already
some
decision
all
the
list
that
we
need
to
double-check.
If
everyone
is
okay
with
that,
and
then
some
next
steps
were
photographed
next
slide,
please
so
this
is.
C
This
is
the
a
we
said
yeah,
so
the
draft
have
was
adopted
last
last
year
and
assess
the
draft
was
adapted,
so
we
went
into
a
we
say,
extensive
working
on
this
on
the
on
the
on
this
on
this
on
this
document,
so
that
we
can
progress
it
as
fast
as
possible.
So
we
went
to
into
nine
revisions
since
then
we
doubled
I,
would
say
the
number
of
pages
we
we
constantly,
we
I
would
say
we
have
a
lot
of
details
there
for
the
specification,
so
that
tanks
are
really
really
clear.
C
Instead
of
one
yang
model,
we
have
now
two
tunic
models.
I
will
explain
later
why
we
have
that
in
term
of
attributes
that
we
define
we
are
current,
we
have
I
would
say
more
than
100
attribute
that
you
are
using
for
for
telemetry.
We
are
including
a
lot
a
lot.
A
lot
of
example.
Also
that
really
I
would
say
implementers
and
also
the
the
readers
can
understand.
C
The
I
would
say
the
intent
of
mini
feature
that
we
are
describing
in
that
in
a
in
a
draft
and,
of
course,
and
as
visual
as
we
have
doing
in
in
depth,
we
are
working
together
and
to
integrate
any
feedback
of
security
from
from
them.
So
and
I
want
I
want
the
list,
all
of
the
change
that
we
have.
We
have
made
sense
since
last
last
year
and
that
will
focus
only
on
few
of
them.
C
C
C
Zero
or
John,
please,
if
we
don't
have
any
of
these
issues,
please
feel
free
to
jump
in
and
continue
the
presentation.
Please
sorry
about
that,
so
yeah
yeah,
so
it
what
it
was
about
saying
is
that
we
do
some
local
information
which
is
available
where
they
look
all
DMS.
The
DOS
client
can
share
it
information
with
the
server
and
we
are
not
doing
that
in
the
mitigation
request,
because
we
won't
win
favor
with
that
and
it
we,
and
also
because
of
the
the
size
of
the
TV
to
death
as
we
are
chained
to
the
server.
C
C
C
We
are
define
a
new
type
message
which
we
call
the
telemetry
in
which
you
see.
We
have
a
lot,
a
new
operation
path,
which
is
the
telemetry,
and
then
we
there's
an
existing
list
of
parameters
and
attributes
that
we
are.
We
are
convening
in
this
in
this.
In
this
messages,
in
this
kind
of
I
would
say
in
this
type
of
message,
we
are
really
not
aggregating
the
data
we
have.
C
We
are
low
in
a
wide
quality
of
information
that
can
be
shared
because
the
purpose
here
is
to
to
share
as
much
so
that
we
can
have
a
website
to
be
in
sync
and
so
that
we
can
foster
the
the
mitigation
that
we
can
help
to
the,
for
instance,
the
DOS
client
or
the
server
to
to
to
assess
whether
the
mitigation
operations
are
really
really
on
track
and
on
progressing
next
slide.
Please-
and
you
can
also
so.
This
is
another
and
another
context
in
which
we
can
use
it.
C
C
This
kind
of
I
will
say
filters,
so
that's
why
it
is
important
for
the
dots
client-server
if
it
wants
to,
for
instance,
to
detect
abnormal
traffic
or
in
the
attack
traffic
to
to
have
an
idea
about
the
baseline
or
the
traffic
based
on
of
a
given
Doc's
client
domain.
So
in
other
city
that
we
are
defining
means
in
the
telemetry
specifications
to
a
lower
this
client
to
share
as
much
as
possible
the
information
about
the
the
traffic
normal
baseline.
C
C
So
Patti
can
speak
the
traffic
and
clean
the
traffic
before
forwarding
that
for
the
clean
traffic
into
the
dots
client
domain.
But
if
the
dot
server,
that
is
not
aware
about
the
capacity
of
the
the
links
of
the
dis
client
domain,
the
direction
of
the
clean
traffic
into
the
ducts
client
domain
may
be
another
source
of
attack.
C
So
that's
why
we
do
does
client
share
the
what
we
call
the
pipe
the
the
pipes
of
capacity,
that
for
the
ingress
with
the
server,
so
that
the
server
will
take
this
information
in
order
to
to
adjust
the
traffic
that
will
be
sent
or
distributed
and
among
existing
adverse
links
of
the
dots
client
domain.
Next
slide.
C
Base
would
say
matrix
in
order
to
to
to
to
make
the
the
computation
and
to
make
the
analysis
on.
For
instance,
the
DOS
client
can
disable
some
some
of
the
person
tiles,
for
instance,
if
it
is
only
interested
on
the
hyper,
some
tiles
444,
the
traffic
and
so
on.
It
can
set
that
with
with
that
server,
and
only
that
kind
of
information
will
be
sent
to
the
server
and,
of
course,
we
are
also
indicating
to
the
server
whether
the
declined,
for
instance,
is
interested
to
receive
the
server
originated
telemetry.
C
This
is
interested
more
I
would
say
relevant
in
case
of
the
server
update
census,
telemetry
information
in
the
update
messages.
It's
ended
to
the
client,
and
this
is
to
avoid
any
I
would
say
errors
or
in
the
I
would
say
exacerbation
of
the
size
of
the
messages.
While
this
is
not
useful
for
the
client,
we
are
also
introducing
this
notion
of
the
duty
limit
free
notification
interval.
C
So
this
is
also
again
another
guard
to
avoid
all
links
with
with
the
notifications
and
to
avoid
creating
new
conditions,
and
while
we
are
already
under
server
several
attacks,
we
are
also
in
negotiating
a
lot
of
types
there
and
cute
apps
that
we
that
are
used
to
filter
out
to
death
as
we
can.
We
can.
We
we
are
interested
to
receiving
from
the
the
server
next
slide.
Please
yeah.
So
once
we
have
I
would
say
when
we
have
defined
all
this
I
would
say
the
attributes
for
that
can
be
used
for
the
telemetry.
C
C
The
name
of
the
attributes,
indigenous,
if
occasion
for
a
notation
for
answers,
will
be
different.
Whether
we
are
including
this
is
in
a
pure
telemetry,
whether
we
include
it
in
into
an
existing
signal,
miss
channel
messages.
So
we
we
have
discussed
whether
we
will
use
only
one
key
value
for
for
both
of
them,
so
that
the
I
would
say
it's
absolutely
implementation.
She
decides
whether
this
is
a
telemetry
or
an
efficacious.
A
bit
or
a
normal
I
would
say,
signal,
channel,
signaling
message
and
then
map
it
to
the
appropriate
attributes,
but.
A
C
C
C
C
Okay,
so
next
set
please
so
the
decision
to
you
that
we
we
have
we
have
taken
is
to
to
simplify
the
implementation
and
that
this
is
why
we
we
went
to
with
two
key
values
for
even
if
this
can
be
assimilated
as
a
same
attributes.
We
are.
We
are
using
two
key
values
here
to
simplify
the
implementations
and
there
is
no
local
mapping
tables
that
will
be
used
as
a
function
of
the
type
of
the
message
in
which
we
have.
We
are
inserting
the
telemetry
data
next
slide,
please.
C
So
this
question
whether
we
need
to
we
have
to
define
the
attributes
as
a
comprehensive,
the
comprehension
required
or
a
comprehension,
optional
attributes,
and
we
went
with
the
as
you
can.
You
can
see
in
the
next
slide,
please
that
yeah.
So,
as
you
know,
the
telemetry
is
all
is
only
are
really
optional
and
the
television
rata
is
only
used
as
a
hint
by
the
beta
server.
So
the
server
is
really
free
to
to
to
to
take
into
into
account
this
information
for
it
for
it
for
the
mitigation,
but
it
can
it
can.
C
It
can
use
it,
of
course,
but
it
can
ignore
it.
So
this
is
really
I
would
say
optional
and
we
don't
want
to
tweak
the
server
by
the
the
person
failure
so
each
time,
for
instance,
if
you
have
defined
that
as
a
comprehensive
require
each
time
you
will
send
an
attribute
which
is
not
understood
by
the
pier
there
will
be
our
narrow
and
then
we
have
to
fall
back
into
a
normal
process
in
which
you
don't
insert
that
information.
C
So
that's
why
we
are
not
really
consistent
and
coherent
what
we
have
done
in
DF
or
the
other
specification.
So
this
is
really
a
comprehensive,
optional
attribute,
but
then
we
have
this
issue,
given
the
ranges
that
we
have
defined
for
the
comprehension
of
optional
attributes,
this
kind
of
keys
that
will
be
used
will
consume
three
bytes
so
which
will
be
I,
would
say,
and
given
the
amount
of
the
data
it
will
be
sent
in
the
telemetry,
this
may
be
problematic.
C
C
B
B
D
Mad
this
is
still
here.
I
have
a
comment
here:
I
mean,
as
you
know,
right
I
mean
telemetry.
Data
is
supposed
to
be
optional
and
rot.
Server
can
ignore
the
telemetry
messages
right
I
mean
so.
The
whole
idea
was
telemetry
is
an
optional
feature
and
whether
the
dot
server
uses
some
of
these
features
for
identifying
its
medication
strategy
or
not,
is
optional
so
making
it
comprehension
required
C
to
save
by.
It
seems
like
an
overkill
for
the
dots
over
to
understand
all
these
messages.
Otherwise
it
has
to
start
rejecting
these
messages
right.
D
D
A
E
A
B
F
So
the
telemetry
attributes
the
basically
what
red
is
asking
here
is
to
originally
the
comprehensive
required.
It
was
one
all
the
way
through
to
63
what
63
83
and
it's
just
to
make
a
sub
set
of
that
dis,
comprehension
optional,
so
that
we
can
have
the
telemetry
attributes
sitting
down
in
a
single
byte.
That's
what
that
is,
but
yeah
what
a
terminus!
Okay,
so
good!
Next
slide.
F
Okay,
so
here
is
currently
we're
sharing
the
information,
in
particular
things
like
the
different
attack
IDs.
What's
going
on
and
the
after
vision,
he
just
had
attack
name
included
in
there,
which
is
a
string
which
is
quite
long,
which
then
takes
up
quite
a
lot
of
space
sitting
there
with
a
telemetry
data.
F
In
particular,
it
could
be
covering
there's
a
new
attack
name
just
come
out,
and
yet
both
ends
have
been
synchronized
on
that
next
slide.
Okay,
so
a
lot
of
the
telemetry
stuff
made
us
realize
that
we
have
severe
challenges
with
large
data
packets
and
there
is
within
the
coop
underlying
mechanism.
Is
the
ability
to
be
able
to
do
I
want
to
block
one
a
block
to
where
data
can
be
reduced
down
into
blocks
and
then
sent
over
the
crime
channel?
We
assembled
at
the
other
end
and
things
continue.
F
We
looked
at
trying
to
break
the
data
with
and
down
into
individual
chunks
so
that
everything
sits
a
packet.
So
we
can
personalize
everything
up
into
individual
packets,
but
the
yang
any
data
requires
a
fully-fledged
Jason
definition
sitting.
Then
we
found
that
that
would
look
there.
The
use
of
block
1
block
2.
The
challenge
is
that
they
are
lockstep
in
that
you
send
a
block.
F
The
other
end
says:
yep
I
got
it.
You
can
then
send
the
next
block.
The
other
one
says:
yep
I
got
it.
Please
send
the
next
one
and
there's
a
lockstep
process,
but
if
we
are
running
in
a
pipe
full
scenario
because
we're
under
attack,
we
can't
have
that
lockstep.
We
can't
necessarily
have
bi-directional
traffic.
We
can
have
unidirectional
traffic
and
so
a
solution
that
were
a
potential
solution
to
that
is.
F
We
came
out
with
a
block
3
block
4,
equivalent
to
the
block
1
block
2,
which
doesn't
require
packets
to
be
ignored
so
with
a
block
3,
which
is
a
push
type
environment,
is
that
all
the
data
can
be
just
pushed
at
once
and
then
the
server
can
receive
it
and
acknowledge
it
at
any
one
point
in
time
at
the
end,
likewise,
block
4
is
the
server
sending
stuff
back.
He
can
just
send
all
the
data
in
the
series
of
blocks
and
there
can
be
a
single
acknowledgment
at
the
end
embedded
in
there.
F
F
Okay,
so
this
just
really
talks
about
this
particular
draft
that
we
have
out
there
in
there.
We
have
things
to
do
with
things
like
probing
rate,
so
we
can
make
sure
that
the
actual
block
three
block
four
don't
overload
a
pipe
in
their
own
rights
by
sending
too
much
data.
So
we
have
controls
in
place
what's
taking
place
there
and
we've
put
in
something
for
max
payloads,
which
basically
says
after
ten
packets
take
a
breather,
maybe
exchange
something
and
carry
on,
and
we
have
got
within
the
block.
Three
block
option
offer
options.
F
We
have
using
the
type
II
tag
and
the
quest
tags
to
identify
that
this
particular
set
of
blocks
is
associated
with
this
particular
body
of
data
and
there's
a
new,
so
response
code.
That
comes
back,
says
well.
Yeah
I
got
everything,
but
missing
blocks,
one
five
and
ten
or
whatever,
so
that
there
can
be
recovery
sitting
in
there
next
slide.
F
F
F
Saying
give
me
these
missing
blocks
number
one
and
two
so
retransmits
blocks
one
or
two
to
still
doesn't
get
through,
and
then
it
requests
it
again
and
so
on.
So
we
can
just
see
the
recovery
mechanism
in
place
there
next
slide
and
the
same
for
block
4
is
just
transmission
from
the
server
the
whole
of
telemetry
data
coming
back
to
the
client
and
some
of
the
traffic
gets
lost,
so
the
client
can
request
that
he
wants
a
missing
blocks
or
optionally.
The
client
can
decide
well,
actually,
life
has
moved
on
I.
F
F
So,
within
the
the
current
telemetry
Dorf,
we
make
reference
to
the
fact
that
there
is
this
ongoing
block
three
block
for
work,
but
we
haven't
said
it
a
requirement.
This
is
taking
place
because
all
the
different
parts
are
not
yet
come
together.
Currently
in
the
draft
would
have
to
use
block
1
block
2,
which
does
have
its
limitations
because
of
the
lockstep
requirements
sitting
there,
and
we
do
recognize
that
a
full
inbound
pipe,
for
example,
will
causes
problems.
G
F
F
As
far
as
my
persuading
is
concerned
is
this
is
where
I'm
at
with
the
NCC
implementation,
everything
is
up-to-date
apart
from
the
telemetry,
which
is
version
7
and
I'm,
working
on
bringing
it
up
to
version
9
the
main
fundamental
difference
there
is
the
attack
ID
to
attack,
name
mapping
that
we
mentioned
before,
and
this
is
currently
using
block
I
haven't
got
the
block
for
making
stuff
in
there.
So
that's
situation
where
I'm
at
the
moment,
so
it's
a
fairly
mature
there.
F
F
D
John
on
the
previous
slide
right
I
mean
I,
like
the
idea
of
exchanging
the
attack
idea
names
in
the
dots
data
channel,
especially
because
the
names
are
not
standard
right
I
mean
every
window,
has
its
own
way
of
interpreting
a
attack,
description
right
and
if
somebody
is
using,
let's
say,
NLP
or
deep
learning
to
basically
map
these
using
some
board
emitting
techniques
to
identify
whether
it
maps
to
some
attack.
They
already
have
I,
think
I,
think
I.
D
Think
sharing
sharing
as
much
this
detailed
description
of
the
attack
name
is
a
good
idea
and
only
can
be
done
in
a
data
Channel,
and
if
all
that
description
details
are
sent
in
signal
generation,
so
I
think.
If
somebody
wants
to
have
deep
learning
techniques
to
map
the
attack
IDs
to
whatever
mitigation
techniques
they
have
I
think
definitely
a
data
channel
seems
to
be
a
good
place
for
that.
F
D
That
can
all
be
automated
yeah
yep
and
on
the
question
of
block
three
and
block
four
right
looks
like
I
mean
I
mean
you
are
heading
the
path
of
of
the
quake
right
I
mean
quick
protocol.
That's
being
discussed
right
here,
where
you
can,
you
can
do
multiplying
multiplexing
and
you
don't
have
to
worry
or
worry
about
head-of-line
blocking
right.
D
That
seems
to
be
a
great
way
of
doing
it,
but
is
that
creating
a
normative
dependency
on
that,
because
that
seems
to
be
a
much
because
our
for
a
large
data
center
or
an
ISP
that
is
getting
attacked
all
right.
Our
large
enterprise
is
getting
it
at
the
telemetry.
Details
are
going
to
be
large,
set
of
messages
right
and
and
if
we
envision
the
telemetry
details
are
gonna,
be
large
set
of
messages
and
will
be
exchanged
during
the
extreme
network
conditions,
for
example,
when
the
network
is
congested
right.
D
F
C
This
part
Thiru,
and
is
that
actually
yes,
we
we
are,
as
mentioned
that
in
the
one
of
the
slides
that
we
are
not
using
our
purpose,
I
would
say
in
a
normative
language
to
avoid
any
dependency
and
delay
of
the
specification.
That's
the
first
point.
The
second
point
is
that
in
the
draft
you
have
mini
every
eight
weeks
in
order
to
optimize
data.
For
instance,
we
have
the
ability
to
filter
out
the
data
that
can
be
pop
it
in
the
signal
channel
so
that
you
can
control
the
amount
of
data.
C
C
This
block
three
and
block
four
to
be
implemented,
but
we
cannot
see
it
as
a
mandatory
right
now
in
the
heat
in
the
current
specification,
because
we
don't
control,
I
would
say
the
how
that
specification
will
go
through
the
edit
code,
working
group
and
and
and
so
on.
So
that's
why
we
are
proposing
this
approach
that
we,
yes,
we
will
record
the
issue.
C
Will
work
on
I
would
say
on
tweaks
on
how
we
can
do
tanks
better
without
frozen
the
specification
to
something
that
we
are
not
sure
will
get
in
the
soon
in
times
as
we
can
deliver
a
solution
to
be
to
be
to
be
there
for
telemetry,
this
solution
can
be
I
would
say,
hopefully
enhanced
with
once
we
have
this
block
three
and
unblock
for
what
we
are
not
recommending
to
have.
At
least
we
are
not
using
any
normative
language
for
that.
One.
A
F
Basically,
we
did
a
lot
of
interoperability
test
back
in
March
April,
there's
a
couple
of
things
still
outstanding
and
we
find
it
very
useful
to
pass
information
back
cell
phones
and
in
the
sort
of
limited
tests
we
were
talking
about
eight
or
nine
packets
required
for
basically
a
fairly
minimal
subset
of
telemetry
information.
So
for
body
of
information,
so
we
do
need
the
ability
to
be
handle
block
packets,
okay,
next
slide,
so
here
just
briefly
is
covering
where
we're
at,
as
we
can
see,
really
well
covered
in
terms
of
what's
there
in
a
specification.
F
Okay,
so
we
can
pick
up
this
a
bit
later.
But
just
if
someone
is
writing,
telemetry
information
to
the
server
and
the
server
is
giving
his
own
version
of
the
telemetry
information
back.
Who
is?
Who
is
so?
We
just
need
to
clarify
the
height,
so
we
can
pick
that
up
in
the
working
groups
all
the
time.
Okay,
next
slide.
F
A
Yeah
Frank,
as
a
chair,
we
think
that
we
should
make
it
faster,
makes
a
whole
process
so
I
think
for
us.
As
far
as
chairs
we
were
requested
because
current
abortion,
you
there's
a
lot
of
young
attributes
for
the
DDoS
mitigation
Elementary
so
and
it's
a
very
long
young
young
young,
your
model
draft,
so
I
think
we
should
risk
requested
the
young
doctors
review
so
to
help
improve
the
whole
quality
of
the
young
data
model.
So
that's
what
we
we
are,
and,
secondly,
I
think
all
this
tossed
telemetry
information.
A
It's
also
very
related
to
the
traffic
analysis,
to
the
potatoes
traffic
situation,
something
so
I
think
it's
also.
Where
necessary.
We
ask
the
OBS
area
Directorate
to
have
the
review
of
the
current
draft
from
the
you
know
from
the
traffic
aspects.
So
all
of
your
all
of
you
all
of
your
new
attributes.
So
that's
what
we
want
to
do
and
we
hope
that
there's
modest
caution
about
especially
from
the
operators
view
if
they
can
give
some
comments
on
main
lister
about
auto
value
of
this
way
of
these
attributes.
That's
we.
That
will
be
great.
C
A
C
A
C
Really
really
fast,
this
one
great
next
slide,
please.
So
we
have
only
one
slide
for
ya,
so
this
is
yeah
so
that
we
have
released
the
analysis
version
last
in
in
May,
so
to
fix
some
normative
language
up
to
to
fix
a
man
or
a
that's
there,
you
see
from
the
others
and
because,
because
it
would
say
to
familiar
with
the
the
documents-
and
it's
sometimes
difficult
to
I
would
say
to
see
issues
so
I
have
asked
some
of
the
co-authors
to
double
check.
C
If
there
is
any
I
would
say
issues
or
any
suggestion
to
enhance
the
DES
specification
that
began
the
feedback
received
from
weapon
particle
that
yeah,
the
Texas
is
I
would
say
the
he
don't
Sydney
in
issue
in
issues
there,
so
I
wait.
Wait
we
have
sent
to
the
list
will
really
need
I
would
say
fresh
eyes
to
look
on
the
text
and
to
to
share
reviews
and
feedback
on
on
that
one.
So,
I,
don't
think
that
the
document
is
really
for
working
up
last
call
yet.
C
A
E
E
And
this
is
the
objective
of
the
draft
so
so
objective
over
the
draftees
are
showing
how
to
use
those
to
meet
area,
network
and
diffuse,
or
this
Dutch
territory
and
contents
of
the
draft
is
some
produced
cases,
including
aim
of
the
use
case
and
what
complaints
are
deployed
in
the
network
and
how
they
cooperate.
And
what
information
is
extracted
next
piece
and.
E
F
E
F
F
E
E
E
A
C
A
F
D
Frank,
my
opinion
is
I
think
we
should
definitely
pursue
this
draft.
This
draft
actually
talks
about
a
lot
of
scenarios
right,
I'm,
not
sure
how
we're
gonna
convert
all
of
them
and
squeeze
them
into
the
telemetry
draft
right.
So
I
think
it's
better
that
the
telemetry
draft
refers
to
this
one
as
an
informational
draft,
and
we
continue
to
make
progress
on
this
one.
D
So
for
me
option
one
looks
like
in
more
preferable
option
to
make
more
progress
on
this
and
have
more
discussions
right,
I,
think
I
think
it
discusses
several
use
cases,
maybe
they're
few
more
use
cases
that
it
needs
to
add,
but
I
think
definitely
it
could
cover
the
various
reasons
why
telemetry,
both
from
the
and
the
client
and
server,
would
help
both
the
peers
to
exchange
and
then
make
use
of
it
right.
So
I
support
this
work
and
I
think
we
should
wrap
this.
A
Okay,
as
I
individual,
my
personal
opinion,
I,
think
of
that
for
the
for
simplified
or
the
process.
If
the
current
I
trust
your
case
drafter,
is
it
scuba
and
it's
mapping
to
the
term
entry
chapter
where
all
we
can
we
can.
We
can
make
it
good
enough
in
a
short
time,
I
prefer
the
option,
because
now
are
now
more
and
more
working,
although
80s
are
recommending
use.
A
B
Usually
it's
better
if
these
cases
at
least
briefly
explained
in
the
main
document,
so
you
can
understand,
will
okay--oh
implemented
and
what
would
take?
What
is
it
for?
But,
of
course,
if
it
makes
the
my
document
very
large
and
difficult
to
reads,
and
it's
better
to
separate
these
cases
from
the
main
document,
so
both
options
should
be
considered
properly.
We
can
discuss
it
with
the
on
the
list
and
probably
would
just
consult
our
director
would
is
what
is
things
about
it.
D
And
Frank
I
understand
where
you're
coming
up.
They
are
to
Isaac,
discouraging
use
cases
craft
right,
but
this
is
more
than
a
use
case.
Graph
right
I
mean
this
is
this
is
going
to
be
talking
about
how
this
telemetry
could
be
used
in
specific
network
environments
and
how
they
can
leverage
this
telemetry
data
to
solve
specific
problems.
D
Right
I
mean
for
deciding
the
mitigation
strategy
or
for
psych
ops
purposes
right,
so
it
seems
more
of
an
operational.
It
covers
more
than
use
cases.
It
covers
how
it
can
be
deployed
in
specific
environments
and
how
specific
environments
can
use
this
telemetry
force
for
doing
some
more
engineering,
work,
deep
learning
or
supervised
machine
learning
and
other
work
that
could
pay
for
that
right.
A
G
G
G
G
Next,
in
the
scenario
inside
SP
I
have
changed
the
net
work
structure
and
to
make
it
much
more
abstract
and
more
flexible,
so
we
can
adjust
it
according
to
the
size
of
our
network.
I
have
also
okay
in
the
figure
left
right.
Further.
Are
the
secondary
network
for
detail
analysis
and
feel
weak
inside
about
to
attack
pace?
You
can
see
turf,
rag1
and
rag2.
G
G
G
What
what
well
to
to
turn
next
version
on
my
draft
I
will
further
discuss,
is
called
deployed
manuals
in
detail
and
served
appropriate
to
know
the
scope
of
management
on
the
other
side
ways
automatic
configuration.
This
is
also
the
comments
further
from
other
people
and
for
this
problem
is
much
more
difficult
for
me.
So
I
need
more
comments
and
coercer
to
make
this
just
a
move
on
since.
A
Actually,
I
I
don't
see
too
much
progress
of
this
draft
and
to
enough
discussion
of
hit
on
the
main
list
or
other
place
and
I
personally
I
I
cannot
understand.
Clearly
what
is
the
new
things
comparing
to
the
existing
to
those
documents
and
the
solutions?
The
so
meaning
I
think
you
still
need
clarify
your
objectives
in
Middle,
East
and
with
other,
are
you
know,
other
active
people's
in
working
group
and
have
more
discussion
on
the
main
Lister
and
also
you
can
clarify
to
us
chair.
You
know
what
is
your
goal?
A
Adjuster
fear
current
content
can
be
combined
with
with
several
other
documents,
such
as
mati
homie
or
the
elementary
youth
case
or
other
documents.
I,
don't
I,
don't
see
the
reason
why
we
needed
this
individual
draft
to
describe
you
know
the
different.
Actually,
the
different
different
different
consideration
is
related
with
different
aspects
such
as
how
to
do
the
multi
homing,
how
to
do
the
how
to
do
the
architectural
designs,
so
so
I
I,
don't
see
any
new
things
here.
A
A
B
And
then
time
so,
thank
you
very
much
for
everybody
who
presented
us
in.
It
was
a
very
fruitful
meeting
and
a
lot
of
things
were
discussed,
especially
about
the
metal
rod.
It's
quite
interesting,
I
think
that
the
next
steps,
as
Frank
already
said,
that
we
will
ask
request
an
official
review
of
young
doctors
and
option
Directorate
and
encourage
people
to
more
actively
participate
in
discussion
multihoming
drug,
because
it
is
very.