►
From YouTube: IETF-DETNET-20230822-1200
Description
DETNET meeting session at IETF
2023/08/22 1200
https://datatracker.ietf.org/meeting//proceedings/
A
A
Okay:
let's
go
ahead
and
get
started.
This
is
an
open
working
meeting
of
the
debt
net
working
group.
Looking
at
scheduling
and
queuing.
A
B
A
C
B
A
Is
great,
thank
you
all
right.
So
this
is
the
agenda
of
this
meeting
is
a
little
bit
different
from
the
past
ones.
We've
we
get
to
bash
the
agenda
in
a
moment,
we're
going
to
talk
about
what
we're
doing
to
make
progress
and
how
we're
going
to
get
things
done
from
here
and
then
on.
The
want
to
start
me
start
to
draw
try
to
draw
some
conclusions
from
the
look
we've
been
taking
at
the
TSN
mechanisms
with
respect
to
the
requirements.
A
Draft
there's
a
couple
of
things
here.
First,
is
we
have
this
new
ecqf
going
on
in
I
in
802
we're
going
to
talk
about
a
little
bit
and
then
was
help
from
psf
and
ESU.
We
have
a
summary
of
TSN
mechanism.
Evaluations
to
take
a
look
at
the
goal
here
is
to
understand
where
things
sit
with
respect
to
the
requirements
draft
based
on
what
already
exists
and
it
with
luck.
It'll
it'll
shed
some
light
on
relative
importance
requirements
for
new
mechanisms.
A
A
A
A
All
right,
oh
I,
know
we
should
do.
Can
I
get
a
few
folks
to
volunteer
to
head
over
to
the
Hedge
Doc
Page,
where
we're
sort
of
taking
notes
in
the
meeting
and
generally
take
notes
on
the
discussion.
A
Okay
for
progress
from
here,
this
meeting
is
I
think
mostly,
but
planned
for
this
meeting
and
happy
to
take
suggestions
for
doing
something.
Different
is
to
take
a
look
at
the
summary
of
the
TSN
queuing
mechanisms
with
respect
to
requirements
draft
and
along
the
way.
We
should
talk
about
ecqf
and
I'm
I've,
seen
a
bunch
of
folks
working
on
new
proposals,
who've
prepared,
one-slide
summaries
of
how
their
mechanisms
stack
up
against
requirements
in
the
requirements.
Draft.
A
A
A
Okay
show
food:
do
you
want
to
talk
about
what's
in
ecqf
and
the
if
the
initial
evaluation
against
our
requirements,
draft.
D
Who
for
easy
curve
so
for
the
first
item,
Delete
in
time
or
synchronously?
Yes,.
D
And
frequency
look,
that
means
water
nodes
must
use
the
same
second
time.
D
I
just
send
it
to
you:
I
prepare
Iceland.
A
A
D
Okay,
so
this
is
the
the
the
slide
it
based
on
our
recently
and
fully
represented
the
entire
content
of
the
standard,
so
I'm,
sorry
for
the
household
mistaken.
So
it
is
only
for
your
reference
and
for
discussion.
Next
Step,
please.
D
So
this
is
overview.
The
difference
between
these
account
for
based
on
the
SQL
in
the
next
side,
that
is
the
Seeker
for,
is
how
the
following
purchase.
D
So
on
the
right
side,
it
is
the
easy
curve
it
has
the
following
features.
The
first
is
no
time
singularization
for
water
nodes
of
for
for
formatification.
We
can
take
a
look
at
the
to
maybe
largely
negative
delay
each
load.
D
The
Cyber
begin,
isn't
all
the
same,
but
you
need
a
frequency
worker.
That
means
volunteers,
use
the
same
cycle
time
and
the
the
node
must
know
the
phase
of
the
between
other
stream.
We
know
the
under
this
node.
D
Continue
to
send
in
the
mark
of
the
cycle
and
the
park
is
born
in
the
single
cycle
wall
with
the
center
in
the
same
recycled.
This
is
the
same
version
Seeker
and
determine
the
input
the
circle
ID
back
around.
This
is
also
same
as
the
secret,
so
this
is
this
is
the
overview
for
digital
mechanism.
D
So
please
can
do
next.
D
Here
is
important
function
of
this
cup
with
C
part
for
protocol
that
is
used
to
detect
the
same
offset
between
absolute
node
and
this
node.
It
uses
two
percent
at
the
first
capacity
is
the
cpaper
timer
marker
frame,
and
the
second
message
is
to
see
purple
face
offset
mean
in
in
ctmf.
It
is
generated
by
the
CPU
and
send
the
reason
the
any
position.
D
Thumbs
down
and
then
send
it
to
the
blue
GB.
One
pollutp
received
this
message:
it
just
recorded
the
received
the
timestamp
and
the
donation
then
please
send.
The
second
number
is
the
current.
D
D
B
I
I
want
to
emphasize
Okay,
add
a
little
bit
more
information
to
this
slide.
So
there
is
a
pre-requisite
or
there
is
assumption
for
the
CPAP
to
work
that
is
the
node
a
and
the
node
B.
Actually
they
need
to.
They
are
required
to
support
the
hardware
Hardware
based
time,
stamping.
So
the
assumption
is
the
node
a
can
record
the
first
bit
at
the
physical
layer
to
be
sent
out
and
in
node
B
can
record
the
first
bit
of
a
packet
received
at
a
particular
Port.
B
So
this
is
called
the
hardware
timestamping.
So
the
prerequisite
for
this
is
to
support
this
feature
so
that,
if
you
look
at
this
picture,
you
can
see
the
bottom
part,
it
says
a
receiving
Bridge.
B
D
D
B
B
Oh
so
when
when
did
I
disable
or
lose
my
connection.
B
That's
this
will
help
CPAP
this
new
protocol
to
be
run
in
order
to
get
quite
a
very
accurate
link,
propagation
delay.
So
that's
the
purpose
of
that
and
another
thing
is
the
the
the
picture.
The
bottom
part
of
the
picture
says
is
a
receiving
bridge,
but
at
the
same
time
we
need
to
remember
for
the
receiving
Bridge,
their
incoming
port
and
the
sorry,
the
receiving
port
and
the
transmission
Port.
B
So
here
it
is
try
to
explain
how
the
receiving
Port
of
node
B
try
to
do
the
face
alignment
with
the
Upstream
transmission
Port
on
node
a
so
it
hasn't
shown
the
transmission
Port
of
node
B.
So
this
I
want
to
clarify
that's
all
okay,.
D
D
Also
understand
the
the
sample
remaining
issues
or
say
pop
the
first
one.
Is
you
which
fees
via
to
take
it
through
to
to
you
for
this
tool,
message
the
ones
that
Define
alternative
type
or
some
nuclear
waste
in
the
existing
lgbas
and
1588
and
lldp
another
easy,
because
there
are
two
messages.
So
the
second
message
issued
bond
to
the
first
University
to
maybe
use
to
use
a
serial
number
for
these
couples.
D
We
have
another
discussion
portal
for
C
program.
We
think
it
may
be
using
one
message
in
method
for
for
the
simple
purpose:
to
detect
the
sagofits
that.
D
Free
cycle,
how
would
these
comments
I
have
just
ascended
to
in
to
Yaris?
Maybe
we
can
discuss
the
HP.
D
So
this
is
the
cool
organization
of
ethical,
different
from
the
traditional
sequel
for
the
they
may
be
one
two
four
one
chapter
class.
D
And
so
on,
so
when
the
community
received
Target
according
to
the
arrived
the
time
wall,
according
to
the
count
of
the
byte.com,
we
assigned
to
the
specific
bin
of
the
specific
Circle,
for
instance,
and
attend,
the
transmission
selection
is
based
on
the
spirituality.
D
For
example,
will
attack
IP
DHCP
to
indicate
the
Target
and
second
node.
The
cellular
here
is
a
signal.
For
instance,
each
circle
will
be
allocated
a
partial
or
for
the
bundle
values,
but
the
total
allocation
must
remove
the
over
subscribe
each
second
level.
Master
contains
the
integer
multiple
Circles
of
next
to
the
first
term
load.
The
following.
These
are
examples
that
have
four
secret
students.
D
D
A
D
A
B
A
D
The
food
is
just,
for
example,
the
wheel
and
tag
or
ibdcb.
The
following
example
is
not
afraid.
The
following
example
example
is
The
Shield
the
relationship
between
multiple
Seeker,
for
instance,
for
example.
The
the
priorities
Suite
in
the
top
level
is
the
probability
it
is
a
independent
figure,
for
instance,
under
the
second
priority.
Four
is
another
single
of
instance.
This
example
is
loaded
as
rainbow
encoding.
A
Okay,
so
it
it's
I,
think
what
I
heard
you
say
is
that
there
would
be
a
field
at
each
at
each
priority
level
to
indicate
the
cycle
at
that
level.
D
D
Okay
in
this
page
is
to
illustrate
the
timing
of
Transmissions
foreign.
D
The
fastest
level
is
priority.
Six
in
each
segment
it
has
a
group
of
rainbow
I
is,
is
something
how
there
are
some
blank
of
is
a
six.
D
D
D
So
is
there
any
comments
for
this
page.
A
D
Okay,
this
page
is
discussed
in
why
Unity
multiple
for
easy.
These
are
commonly
example.
In
this
example,
tc5
allocate
the
activity
percentage
or
for
the
bundle
virus,
while
she's
a
full
one.
Okay,
40,
positive
tc4
cycle
does
not
contain
anything
multiple
Circles
of
next
faster
level,
65.
So
in
the
yellow
sheet.
In
the
circles,
three
patches
of
Blue
Buster
normal
three,
which
is
a
five
Cycles,
maybe
squeeze
but
forget
our
vision,
the
course
of
one
which
is
a
full
circle.
D
So
this
page
is
about
the
time
of
business
curve,
the
the
key
feature
of
autonomic,
basically,
that
the
receive
the
frame
is
assigned
the
input
port
to
Output
bin.
According
to
the
arrival
time,
the
number
of
beans
required-
maybe
two
three
or
more
determined
by
how
many
cycles
I
improved
circles,
religious
span
and
output,
port,
and
one
knows
about
how
frequency
work
with
the
single
circular.
D
In
the
figure
we
can
take
a
look
at
that,
for
example,
in
Bridge
e.
They
may
have
some
tools
in
the
circle.
D
Firstly,
in
the
field
is
caused
by
a
maximum
frame
with
low
priority
and
the
maybe
some
more
delay,
evaluation
caused
by
broker
accuracy,
and
it
was
so
during
the
third
time
the
that
studied
imposed
by
the
downstream
of
bgb
so
available
to
that
is.
The
ta
is
less
than
the
circle.
D
Yes,
but
so
we
can
look
at
the
blue.
D
The
blue
block,
but
yes
this
is
the
the
cycle
place
of
bridge
a,
but
when
you
ultimately
arrived
at
the
output
Port,
it
may
be
a
cycle
plus
14
delay
evaluation.
So
if
you
you
may
spawn
two
or
three
or
more
Cycles
in
the
output
port.
Okay,
is
there
any
commands
for
this
page.
D
So
this
is
a
condition
about
the.
If
this
I
inconsistent,
that
is
a
different
modes,
we
can
see
that
if
there
is
inconsistent
inconsistent,
that
is
a
the
circle
phase
offset
is
not
constant.
D
It
may
be
different
tab
so
for
each
tip,
wtp
can
imagine
a
differently
inputed
cycle
circles,
but
that
is
another
altitude
and
you
put
inside
a
service
I
guess
so
they
imagine
they
did.
The
Circle
sequence
has
the
same
cycle.
Duration
is
output
second
service.
Thus
it
is
impossible
to
assign
output
being
only
based
on
the
arrival
time
at
the
runtime.
D
So
this
is
the
limitation
of
thermal
basic
being
assigned.
So
we
can't
considering
that
if
we
don't
useless
at
the
runtime
to
determine
to
assign
adequate
bin
based
on
the
router,
but
if
we
can't
take
some
operation
that
the
degree
or
suspicion.
D
So
this
is
the
alternate
message
of
the
user
curve,
for
that
is
a
counter-based
festival.
The
key
feature
of
convincible
for
it
and
receive
the
frame
is
assigned
the
import
port
to
Output
bin.
According
to
the
bytes
account
plus
view,
each
stream
store
no
more
than
it
is
constructed.
Amount
of
data
per
cycle
into
any
given
sake
of
being
fluids
are
power
that
they
needed
are
stored
in
the
sub
cell
convenience
and
the
access
data
that
you
see
the
maximum
amount
of
purposes
allow.
D
The
data
stream
is
discarded,
so
this
is
similar
to
vcp
with
maximum
the
interval
of
cycle.
So
this
is
our
course.
Other
features
is
that
the
volanos
use
nearly
single
cycle
nowhere.
So
it
is
not
required
protagonist
semicellular
because
it
is
by
the
car
in
the
history.
The
second
phase,
corresponding.
D
A
One
one
question
on
this
page:
it
says:
TC
Reserve
by
talkers,
reservation
who's.
The
talker
here.
D
D
There
are
two
keys
for
the
flow
aggregation
first
case
is
a
typical
case
that
from
Earth
to
Pogba,
the
new
bandwidth
is
the
changes
tonight,
so
the
few
aggregation
nation
and
in
the
earth
University
May
configure
the
cycle
with
largest
other
in
the
backbone.
The
second
may
be
configured
with
symbol
sets.
D
This
document
mainly
discusses
this
function
under
the
complex,
the
First,
Signal
resistance
and
there's
no
signal
instance
of
the
same
allocated,
bundle
Riders.
So
the
case
one
is
not
the
typical
case
for
the
screen,
Mark
location.
D
The
aggregate
employees
should
use
contact
with
being
assignment
messages,
communities
that
the
time-up
is
being
assignment
requires
the
input
cycle,
duration,
the
output
circuit
duration
to
be
equal
and
the
time,
obviously
being
assignment,
is
harder
to
process
concluding
the
arrival
films
of
component
streams.
If
you
output
faster
secret
instance,
as
a
node
allocated
the
mobile
awareness,
our
recent
aggregated
segment
time
basically
applied.
D
D
We
can
take
a
look
at
the
node,
a
for
easy
stream.
This
is
the
component
industry
for
each
stream,
one
free
more
per
cycle,
for
example.
Here
the
cycle
is
one
milliseconds,
the
required
for
the
sales
for
for
the
screen,
because
somebody
that
is
one
from
the
site
for
one
milliseconds,
so
the
total
required
abundant
values
for
water
streams.
It's
way
more
volumized
divided
by
one
milliseconds,
so
the
secret
resistance
here
is
the
one
millisecond
distance.
D
Component
screens
are
regarded
as
the
same
availability,
the
stream
we
terminate
with
us
Force
streamer
as
one
flimer
per
cycle.
He
the
cycle
is
faster.
It
is
just
for
using
733
micro
signals
required
by
this
one
from
the
sites
approach.
She
misses
me
local
settings,
so
the
sake
of
submissive
is
sleep.
Medical
seconds
instance
are
located
bounded
by
this
one
thing
that
says:
police
solution,
three
magnets,
that
is
the
same
as
a
single
one,
millisecond
distance,
the
output
pin
set
of
a
single
for
33.
D
D
D
So
for
the
disaggregation
is
told
about
how
to
change
from
a
fast
and
save
level
to
slow
a
second
level
for
Speed
under
this
degree,
aggregated
node.
This
is
similar
to
aggregating
community.
This
creating
quality,
also
use
counter-based
beam
assignment.
The
Flames
of
the
universities.
Aggregated
strings
are
distributed
in
the
slow
out
proteins
in
the
domain
foreign.
D
And
the
second
is
the
overpollution
in
latency.
Trade-Off
stream
often
consists
of
offering
with
variable
sites,
for
example,
maximum
movement
is
13
South
in
the
bit
and
the
minimum
of
fluid
600.
D
In
here
are
located,
100
T
I'm,
a
BPS
I
wish
19
percentage
of
operation.
If
you
choose
the
second
time
is
596.
We
are
located
so
this
formula.
D
Code,
I
I,
don't
know
if
any
other
house
something
we
can
listen
about.
How
about
the
the
city
cells
in
the
minus
either
homophone?
D
And
the
third
item
is
a
tail
tailored
by
the
way
this
is
often
is
you
offered
the
customer
only
a
specific
set
of
choice
for
a
bundle
Riders
to
contract
corresponding
to
the
Oneness
of
TC
implemented
in
the
providers
to
improve
our
population,
and
the
first
item
is
about
if
cyber
time
will
report
to
the
regulated
in
the
internet.
In
this
case
one
free
motorcycle,
otherwise,
both
faster
the
cycle
level
can
be
used
under
this
bundle
Riders
to
visit
it.
D
For
example,
other
ATS,
he
must
be
shielded
in
the
worst
case,
the
syncope
can
still
empathy
is
being
recently
available.
A
Okay,
I
have
a
question
on
this
slide.
Very
brief.
What
should
be
converted
here
is
this
a
statement
that
everything
should
be
specified
in
terms
of
cycle
time.
A
E
Thank
you,
I.
Thank
you
for
the
nice
presentation.
I
have
a
quick
question.
Is
there
explicit
delay
bound
to
given.
D
From
the
standard
I
have
noticed
the
formula
about
the,
for
example,
under
the
under
latency,
but
just
have
some
section
describe
the.
However,
it
indicates
the
underground,
ladies,
must
include
the
the
many
types
are
due
to
the
cycle.
Phase
of
this
is
a
key
difference
from
the
traditional
TSN
Seeker.
A
E
Thank
you.
As
you
can
see,
the
cqf
is
deterministic,
as
we
all
can
see
very
clearly
but
easy
to
add.
There
is
a
certain
uncertainty.
Yeah
there
is.
There
is
a
factor
that
is
not
deterministic.
E
If
I
understood
correctly,
the
the
yellow
and
blue
packets
can
be
sent
in
a
best
case
in
node
C
in
cycle
two,
but
in
the
worst
case
they
are
sent
in
those
load.
The
sea,
the
online
cycle.
D
This
is
the
arrival
time
of.
Did
this
whole
flame
at
the
not
be,
but
the
node
B.
We
avoid
a
time
that
is
indicated
as
the
reader,
the
lyrics
to.
D
Cinder
these
two
films
inside
to
the
node
C.
We
received
this
framework
as
a
second
table
and
again
with
the
type
of
that
is
indicated
by
vertical
hello
and
send
the
in
the
next
cycle.
E
D
E
D
E
A
A
Okay
issue.
B
This
is
very
a
lot
of
details,
so
actually
I
have
a
question.
Third
qdv
is
just
started
with
its
initial
very
initial
draft.
So
my
question
is
for
for
the
for,
for
this
slide
deck,
can
you
elaborate
what
other?
B
What
are
the
comments
from
you
and
what
other
the
text
from
the
current
draft
I
can
tell
that.
There's
slide.
As
you
said,
the
slide
four
basically
is
a
comment
from
you
right.
The
slide
four.
Is
there
any
anything
else
that
actually
are
there
the
the
comments
for
the
draft,
but
not
the
draft
itself.
D
He
I'm
not
sure
I,
don't
understand
the
question.
I
just
sent
some
comments
about
the
goody
way
to
jealous,
but
in
decent
selection.
We
just
good
at
the
some
more
important
virtues
of
risk
to
know
some
big
difference.
Traditional
Seeker
under
under
two.
We
can
give
our
information
to
the
notification
requirements.
B
So
I
guess
my
question
is
I
can
see
that
the
slide
for
what
said
in
slide.
4
is
a
comment
to
current
udb
Drive.
The
text
right.
It
is
not
in
current
2db
draft.
So
is
there
any
else
anything
else,
because
qdb
is
is
still
in
very
early
stage,
so
so,
for
example,
the
last
slide
is:
is
it
in
the
current
qdv
text
or
it
is
something
like
you
comments
you
want
to
make
to
qdv.
D
No,
no
just
the
page
four
I,
just
only
the
page.
This
is
an
interesting
part
that
we
require
have
some
discussion.
The
last
page
is,
is
the
that
are
searching
of
considerations
in
a
good
way
for
the
just
give
some
telephone
reorganization.
A
A
F
F
F
A
F
Yes,
yes,
yes,
thank
you
actually
I
I,
think
I
think
the
features
leads
to
begin
in
the
right
hand
for
using
that
is
quite
quite
similar
to
what
we
want
to
solve
or
to
raise
the
SKF
or
key
City,
for
example,
there's
no
requirement
for
our
time.
Synchronization
there
is
a
mapping
relationship
maintained
each
node.
We
are
on
the
10
unit
councilor.
F
So
so,
if
you,
what
you
are
trying
to,
if
you,
what
you
are
trying
to
do,
is
to
provide
a
summary
of
the
existing
or
candidates,
I
I
recommend
to
and
reference
references
here,
especially
references
to
the
university
that
I
work
and
if
you
are
trying
to
propose
A,
New,
Concept,
I
I,
recommend
to
and
to
make
it
more
clear
whether
it
includes
some
other
kinds
of
mechanisms,
for
example,
as
I
think
you
have
also
released
another
mechanisms
about
your
house
like
for
Ctrl
I,
think
maybe
the
the
common
features
of
these
mechanisms,
and
especially
differences
of
these
mechanisms,
should
be
and,
and
so
I
just
want
to
clarify.
F
First,
what?
What
do
you
try
to
do
to
introduce
the
concept
of
musicals
here,
or
it
is
a
concept
that
already
being
disgusting
actually
or
it
just
newly
introduced?
Sorry,
because
maybe
I
missed
some
part
of
the
presentation.
D
So
if
I
could
import
it,
that
you
see
that
the
the
the
wall
footages
of
easy
care
for
is
just
similar
to
a
sales
curve
on
the
t-ser
curve,
I
I,
I
I,
think
we
have
the
same
point
about
the
Dead
so
for
for
for
another
mechanism
that
I
have
always
I
think
committee,
some
difference
for
each
other,
so
that
I
have
just
send
them
out
to
call
us
so
I
think
that
is
another
secret.
D
F
Yeah
I
think
make
it
more
clear.
It
seems
like
easy
class
is
kind
of
summary
of
says
Jeff
and
a
tcqf
NSS
player
MRI.
D
F
Okay,
so
so,
if
that
is
the
case,
I
think
what?
What
is
my
comment
is
that
maybe
we
we
should
consider
whether
we
we
are
going
to
introduce
so
many
Concepts,
because
because
they
there
are
have
already
a
lot
of
mechanisms
and
those
names.
If
we
also
try
to
introduce
new
Concepts
to
just
summarize
the
existing
mechanisms,
maybe
we
should
be
more
careful
about
the
school
of
the
concept
and
what
should
be
your
side
and
what
should
be
outside?
F
It's
from
my
point,
I
think
it's
we.
We
don't
have
to
use
a
new
term
for
that.
We
can
keep
it
separately,
because
it's
independent
mechanisms.
F
F
If
that
is
the
case,
maybe
I
think
it
is
more
important
to
to
clarify
the
actual
European
ITF
work.
As
I
have
raised
the
question
in
the
main
list:
how
to
connect
to
them
because
it
for
actually
I'm
not
sure
whether
they
are
fully
clear
about
the
existing
cash
Network.
F
And
if
we
just
Define
the
concept
independently.
I
think
that
will
make
perfusion.
A
It
is
certainly
the
case
that
we
should
use
common
terminology
with
electrically.
F
Yeah
because
I
think
csqf
and
tcaf
have
been
discussed
for
sometimes
in
IEPs.
F
So
yes,
if
there
is
a
new
new
update
in
I,
believe
that
can
you
know
give
you
give
new
mechanisms
for
similar
case,
for
example,
for
your
enhancement
of
the
Europe
I,
think
that
is
good,
but
I'm
not
sure
we
are
clear
now
what
actually
we
want
to
include
here
in
this
process,
if
that
is
the
whole
picture,
I
think
it's
really
similar,
or
we
should
have
some
official
discussion
about
how
to
you
know,
make
connections
between
these
terminologies.
D
I
have
a
short
responsible
for
your
service
quality.
You
can
take
a
look
at
the
easy
cups
we
we
will
see
a
key
feature
of
these
adults
is
that
you
just
either
just
assign
things
to
Flame
according
to
the
arrival
time,
that
is,
it
seems,
abandonment
from
the
at
one
time,
maybe
some
enhancement,
for
example,
College.
Some
big.
F
Yeah
I
think
so
that
is
my
proposal.
I
think
what
actually
you
should
do
is
that
will
if
they
want
to
introduce
New
Concept
it.
It
is
also
their
opportunity
to
see
what
has
already
been
done
on
your
idea,
just
like
what
we
will
do
before
we
try
to
make
some
extensions
to
existing
article.
You
work
so
I
think
we
should
clarify
the
relationship
together.
F
I
think
I
understand
your
point.
Shampoo
is
that
the
this
concept
may
try
to
take
the
common
features
of
these
locations,
but
I
think,
for
example,
the
reference
is
is
very
necessary
and
what's
more
inside.
This
also
should
be
clarified.
I
think
I
mean
what's
more.
Is
that
besides
local
idea,
work?
What's
more
can
be
done
or
what's
new
in
this
contract.
A
B
Yeah,
so
the
the
easy
query
is
enhancement
to
secure.
If
it
is,
it
is
a
pretty
new
project
in
IEEE
802.1,
so
it
is
still
under
development.
I,
pretty
much
sure
there
are.
A
lot
of
things
has
not
been
fixed,
so
I
guess
the
purpose
of
current
stride.
Deck
is
trying
to
give
a
brief
introduction
to
ecqf,
which
is
defined
by
qdb.
By
the
way
it
is
has
not.
It
has
not
been
done
that
how
their
tcqf
or
CS,
what
sorry,
csqf
and
all
the
other
relevant.
B
B
For
example,
you
try
to
lose
the
requirement
for
the
time
synchronization
so
that
their
different
nodes
can
use
their
sorry
that
can
can
have
the
unsynchronize
the
cycle
boundaries
and
also
it
allows
both
the
time
based
buffer
opening
selection
in
the
account
based
buffer
of
being
selection
and
at
the
same
time,
and
there
it
allows
the
multiple
traffic
classes
to
be
included
in
the
c2f.
That's
why
it's
called
originally
called
multiple
secure,
multiple
means,
multiple
traffic
classes,
I
I
think
so.
So.
B
A
quick
comment
to
how
it
is
different
to
tcqf,
in
my
mind,
is
the
first
thing
is
because
the
ecuf
is
defined
by
IEEE.
So
it
has
a
quite
a
valid
assumption
that
between
two
neighbor
nodes
there
is
only
a
point-to-point.
Physical
link,
that's
interesting.
The
second
thing
is,
there
is
a
hardware
timestamp
available
so
that
it
can
use
the
time-based
reception
time
to
determine
which
one
which
output
being
should
be
put
for
a
particular
packet.
B
So
for
tcqf
we
try
to
divide
in
Internet
which,
which
has
a
mouth,
which
has
a
more
complex
topology,
possibly
so
it
may
not
be
a
point-to-point
physical
link
between
two
neighbor
nodes
and
also
there
are
some
other
unexpected
time
variations
between
two
nodes.
For
example,
there
could
be
some
large
clock,
sorry
clock
drift.
We
call
it
mtie
I
think
in
the
in
the
draft,
so
that
it
is
not
always
true
that
the
constant
propagation
delay
is
always
available.
B
So
that's
why
explicit
tag
is
required
in
that
case,
so
that's
very
short
comments
to
how
it
how
the
T
propose
a
TC
to
have
a
csqf
different
from
the
current
e-cqf
and
yeah.
That's
that's
my
understanding.
A
A
B
A
So
this
is
an
enhanced
version
of
the
slide
deck
we
used
last
time.
Maybe
the
thing
so
anything
to
do
quick
might
be
to
start
with
what
the
sheet
qfe
evaluations
were.
Xiaofu
had
done
one,
and
is
you
had
done
one
the
the
they're
pretty
close
to
each
other
I
think
the
only
difference
was.
Let
me
quick
check.
A
The
only
difference
was
whether
this
number
seven
down
here
was
a
no
or
a
or
a
partial.
So.
A
So
maybe
xiaofu
we
all
would
like
to
talk
about
the
e-seq
evaluation.
Do
you
want
to
start
by
talking
about
cqf
or
start
by
just
just
talking
about
the
the
ecqf
evaluation
directly.
D
By
understanding
is
that
because
the
hopes
company
is
not
just
so,
it
may
affect
the
underground
the
latency,
so
it
is
multiple
code
to
select
the
second
time
so
I
need
to
make
a
trade
Office
of
the
channel
the
other
to
under
delay
on
the
cellular
issue.
D
Okay,
foreign.
D
D
The
third
one
is
a
considered
the
higher
linear
speed.
This
is
what
is
the
policy
I
think
that
my.
D
D
D
D
D
So,
okay,
for
the
next
item
is
the
Apple
wonderful
flag
situation
from
disrupting
source
this
illumination
strategies
departure.
We
can
divided
to
two
aspects
why
the
type
of
business.
D
B
A
A
Okay
and
then
I
put
together
an
initial
scorecard,
and
the
purpose
of
this
was
to
take
a
look
at
what
is
already
available
or
in
case
of
ecqf,
hopefully
coming
soon
from
IEEE
against
against
our
requirements,
and
you
can
read
the
details
down
the
evaluation,
we're
just
counting
and,
in
the
right
hand,
column
on
notes
listed
the
ones
that
aren't
yes.
What
I
wanted
to
call
attention
to
is
the
two
requirements
for
which
we
have
no
Roots.
E
A
C
E
Yeah
I
have
a
I,
have
a
comment
on
the
evaluation
of
ATS.
E
I
think
the
80s
is
quite
a
flexible
solution,
although
it
doesn't
give
the
very
tight
endoting
latest
bound,
but
it
is
flexible,
so
I
think
the
for
the
large
number
of
flows
and
the
complex
topology.
E
E
I,
don't
think
the
flower
aggregation
is
necessary.
Of
course,
the
flow
aggregation
would
help
still
80s
can
manage
a
large
number
of
hopes
with
complex,
tough
ones.
As
far
as
the
the
floats
are
not
very
many,
that's
my
understanding.
Thank
you.
A
E
As
shelf,
who
has
mentioned
in
the
notes
as
the
the
hopes
increase
and
the
topology
becomes
complex,
then
obviously
the
endoting,
the
late
age,
bound
increases,
but
they
can
apply
to
any
solution
that
we
have
seen
so
far.
So
I
would
suggest.
The
answer
is
yes,
Evolution.
E
B
D
We
I
just
brief
response
to
this
items
that
why,
during
the
policy
information
I
think
it
may
be
some
more
services
that
require
low
latency.
D
Actually,
another
sub
type
of
results
that
our
last
requirements.
B
E
I
agree
basically
agree
to
you,
but,
according
to
your
argument,
the
number
of
hopes
certainly
affects
the
endocrine
latencies
of
any
solutions.
D
Multiple
instance,
the
first,
the
fastest
influence
Bay
configure,
for
example,
term
micro,
second
cycle
service.
E
Yeah
yeah
I
think
cpf
and
eckfr
has
the
entertain
the
latency
function.
That
is
linear
to
the
number
of
hopes.
D
Maybe
it
has
some
different
supported,
for
example,
ATS
I,
don't
think
it
may
provide,
for
example,
multiple
instance
for
some
type
of
class
in
May,
configure
the
pre-configure
reserved
bundle,
riders
that
detained
the
boundary
latency.
So
this
this
was
case,
abundant
latency
is
applied
to
any
flows,
belongs
to
this
travel
class.
E
Okay,
I
think
this
is
not
very
important
issue
to
spend
some
time,
so
we
can
discuss
offline.
Thank
you.
A
And
and
I
I
tried
to
capture
that
in
the
notes,
so
I
think
I
think
that
is
I.
Think
that
is
a
useful
discussion
about
whether
linear
scaling
with
number
of
hops
is
the
best
we
can
do
or
whether
better
is
possible.
A
G
Yeah
for
the
partial
evaluation
and
I
have
a
suggestion.
If
we
can
make
a
clarification
for
the
requirements
with
the
partial,
if
we
can
clarify
and
Define
our
definitely
to
figure
what
a
binary
value
or
other
clarification
for
the
repellent
with
partial
that
we
can
distinguish
the
mechanisms
with
a
death
later
low
or
does
it
yes
and
then
we
can
distinguish
their
mechanisms
with
the
the
mechanism
that
can
apply
can
be
applied
in
large
scale
and
within
a
low
the
you
have
a
low
evaluations.
G
A
G
Classification,
for
example,
the
first
part,
is
the
overall
requirements,
just
like
the
the
3.6
and
the
3.8
is
the
overall
requirements
for
the
network
capability.
I
think
it's
a
it
must
be
made.
The
the
requirements
must
be
satisfied
forward
mechanisms
under
the
other
requirements
such
like
the
3.1
3.2
and
this
3.3,
and
they
are
the
requirements
requirements
for
the.
G
Mechanisms,
it
must
be
yes
and
for
other
requirements
such
as
there's
3.4,
3.5
and
3.7.
They
are
for
the
flow
and
the
traffic
scheduling
like
even
if
they
can
be
partial
for
for
their
like
evidence,
so
it
is
okay
for
for
their
mechanisms
with
the
partial
evaluation,
I
I,
don't
know
if
I
make
my
my
clarification,
clear.
A
You
I
heard
a
bunch
of
things
in
there.
Let
me
see
if
I
can.
Let
me
see
if,
if
I
understood
correctly
so.
A
Hang
on
I
need
to
type
one
more
word
here.
Okay,
so
I
think
I
heard
you
catch
on
two
things.
The
first
is,
it
sounded
like
you
had
some
ideas
for
how
to
clarify
the
3.7
requirements
so
that
we
can
get
a
better
evaluation
of
whether
the
scaling
is
is
good.
It
with
the
scaling
is
really
good,
partial
or
not
acceptable
is
is
okay.
Could
you
that
sounds
like
its
effects
requirements
draft?
G
Yeah
I
I
think
I
suggest
to
discuss
if
we
can
to
clarify
the
requirements
more
clear
to
discontinguish
with
low
or
yes,
but
not
their
passion,
but
I'm,
not
sure.
If,
if
we
can
do
it,
if
we
can't
do
yeah.
A
Okay,
well,
it
sounds
like
we
what
I
think
I'm
asking
you
to
do
to
begin
with
is
to
focus
on
the
3.7
requirement
where
we
have
a
week.
I
hear
a
lack
of
a
an
absence
of
agreement
on
whether
linear
scaling
with
with
hop
count
is
yes,
acceptable
or,
or
or
or
only
partial
and
I,
think
the
requirements
draft
ought
to
take
a
position
on
that.
G
Yeah,
okay,
I
will
I
think
we
can
do
some
discussion
on
the
mailing
list
and
the
two
collector,
the
the
member
of
the
partial
or
no
or
yes,
or
with
the
the
the
value
of
the
level
of
the
Hops.
A
Okay,
I
made
note
of
that
and
then
beyond
that
I
think
you
were
suggesting
at
least
a
couple.
I
heard
at
least
a
couple
of
things.
One
is
I
think
I
heard
a
possibility
of
separating
queuing
versus
transmission
scheduling
in
the
requirements
did.
G
G
Yeah
for
the
requirements
for
the
particular
query
mechanisms
that
it
must
be
yes
and
the
other
folks
under
scheduling
the
flow
in
the
traffic
scheduling
may
be
partial.
G
So
this
is
the
the
they
have
different
importance.
A
Okay,
particularly
because,
as
of
right
now
for
the
let
me
go
down,
hang
on
too
many
windows,
just
a
minute,
I
want
that.
A
Go
particularly
because,
certainly
for
the
first
few
requirements,
we
have
no
problems
for,
let's
see
for
at
least
these
the
first
few
requirements,
we
don't
have
a
problem
with
the
existing
existing
mechanisms
that
meet
them,
whereas
for
at
least
the
two
scalability
ones.
Since
we
initially
came
up
with
no
yeses,
we
do
have
clear,
clear
issues.
There.
A
A
I
think
I
captured
your
comments.
Did
you
want
to
say
anything
else.
A
Okay,
all
right,
let
me
in
15
minutes
that
remain.
Let
me
suggest
where
what
we
might
do
for
the
neck
for
the
next
meeting,
which
is
it's
in
it's
in
three
weeks
that
the
calendar
was
carefully
picked
to
avoid
the
early
September
holiday
in
the
U.S
Monday
September
4th
is
a
major
U.S
holiday,
and
so
a
lot
of
people
will
may
not
be
around
on
on
Tuesday
the
5th.
A
So
we're
going
out
we're
going
to
go
three
weeks
in
general,
and
let
me
suggest
that
we
in
the
next
meeting
start
looking
at
this
sort
of
single
slide
evaluation
for
the
new
mechanisms,
but
I'm
a
little
bit
concerned
that
I'll
pick
on
ecqf
a
little
bit
concerned
that
just
asking
for
the
single
slide
is
going
to
be
too
short.
I
mean
we
sort
of
saw
that
in
this
meeting,
where
starting
from
this
slide
was
the
wrong
place.
To
start
the
right
place
to
start
was
the
was
the
more
detailed
explanation.
A
So
maybe
and
I
will
take
this
to
the
list.
I
will
take
this
to
the
list,
so
this
please
please
don't
take.
This
is
anything
more
than
David's
initial
suggestion.
Maybe
what
we
want
is
a
summary
slide
like
this,
for
each
of
the
new
proposals,
accompanied
by
a
small
number
of
slides
that
help
explain
what
what
these
evaluations,
why
the
evaluations
are
what
they
are
in
a
little
more
detail
than
is
possible
in
the
notes,
column.
E
Thank
you,
yeah
I
think
it
is
plausible
and
it
is
time
to
do
that
right
now.
Yeah,
as
you
have
remember,
that
I
have
presented
the
z-score,
the
the
solution
that
we
have
suggested
and
the
evolution
table
of
the
c-score
I
presented
in
the
San
Francisco.
A
I
can
yeah
and
what
I
think
I'm
gonna
do
is
I'm
going
to
put
I
will
send
something
to
the
list.
I'm
gonna
put
I'm
going
to
want
to
put
a
both
a
slide
count
and
time
restriction
on
this,
because
in
order
to
do
a
fair
job
on
the
new
mechanisms,
we
really
should
try
to
make
a
first
pass
over
all
of
the
evaluations
in
the
next
two
hour
to
our
meeting
at
30
seconds
at
30
minutes.
Excuse
me
holy
smokes.
I
need
more
coffee
this
morning
at
30
minutes
of
mechanism.
A
Enough
yeah
three
minutes
is
too
much.
Think
10,
please,
please
think
10
to
15.
A
If
I
start
with
10
to
15,
that's
I,
don't
think
we
have
eight
new
ones,
but
that
that'll
mean
if
we
run
over
slightly,
we
can
fit
them
all.
A
A
E
Planning
to
review
like
four
or
five
or
six
Solutions
in
the
next
meeting
together.
A
Yeah
but
I
would
like
to
to
have
the
next
meeting.
Is
an
evaluation
slide
like
this
one
for
every
one
of
the
new
proposals
people
would
like
to
have
considered
and
then
the
evaluation
slide
would
probably
need
a
supporting
slider
to
particularly
for
the
two
scalable
requirements.
Three
four
and
three
seven
that
are
causing
us.
The
most
difficulty.
A
Okay
and
I'll
put
this
on
the
list,
and
I
will
try
to
get
that
done
done
this
week.
So
we
can
have
some
time
for
discussion
and
agree
on
what
it
ought
to
be
before
people
start
investing
a
lot
of
time
in
preparing
slides
and
I
mean
if
we've
already
prepared
something
that
looks
like
this
evaluation
slide,
you
should
be
able
to
reuse.
It
I
think
the
new
slide
work
might
just
be
a
little
bit
more
explanation,
particularly
on
the
two
scalability
requirements.