►
From YouTube: IETF112-LPWAN-20211109-1430
Description
LPWAN meeting session at IETF112
2021/11/09 1430
https://datatracker.ietf.org/meeting/112/proceedings/
A
A
Okay,
so
we
we
have
challenged
the
network
a
little
bit
this
time,
because
there
is
some
emphasis
which
is
being
given
to
the
code
of
conduct.
So
let's
go
through
that
as
usual.
Your
reminder
that
there
are
number
of
not
next
like
previous
slide,
please
as
usual,
you
you
must
be
aware
of
the
not
well
of
the
itf
which
which
involves
involves
ipr
in
particular.
A
A
So
as
a
reminder,
this
meeting
is
being
recorded.
We
are
actually
recording
it
and
publishing
it
on
youtube.
If
you
don't
want
your
voice
to
be
published
on
youtube,
please
refrain
from
speaking.
That's
the
only
way
presence
is
loved
automatically,
so
there
is
no
blue
sheet.
Metico
takes
care
of
the
blue
sheets
for
us,
but
mitigo
does
not
take
care
of
the
minutes,
so
please
don't
hesitate
to
join
the
the
code
dnd.
A
A
A
A
A
So
the
engine,
as
I
said,
is
pretty
down,
so
I
don't
have
much
more
time
to
just
go
through
it,
we'll
be
having
a
few
words
on
the
state
of
the
architecture
and
where
we
want
to
go
with
this,
then
15
minutes
with
loho,
giving
us
the
latest
on
the
young
daika
model,
then
to
draft
which
derived
from
the
world
on
sick
fox
by
juan
carlos
next
slide.
Please.
A
I'm
hearing
nothing
so
so
we
will
stick
with
this
agenda.
So
here
is
the
status.
We
are
good
on
chic
co-op,
as
promised
the
baseline
technologies.
We
have
covered
laura
one,
and
now
we
have
six
fox
on
the
work,
so
we're
pretty
good
the
data
model.
We
appear
to
be
a
bit
late
before
we
we
decide
for
another
date.
A
A
This
one
is
the
last
one.
Oh
I'm
sorry,
so
the
next
slide
will
be.
Actually
you
one
one
quick
information
before
we
we
go
to
the
different
presentations.
The
way
we
try
to
do
it.
This
time
is
that
everybody
will
own
his
own
slides,
but
we
uploaded
already
the
slides
to
mitico.
There
is
this
new
feature
that
alexander
is
demonstrating
right
now.
A
If
you
look
at
below
your
name
on
the
top
left
of
the
screen
there
there
are,
there
are
slides
and
the
speaker
will
be
allowed
to
if
the
second
button
from
the
left,
you,
you
have
join
q
and
then
you've
got
this
this
page,
which
is
kind
of
right
now
about
with
this.
You
can
select
a
slide
deck
within
the
preloaded
slide
decks,
and
so
we'll
ask
you
to
select
your
slide
deck
within
that,
and
that
will
allow
you
to
do
the
navigation
without
sharing
your
screen.
B
So
with
this
one
and
actually
before
jumping
to
the
lpn
architecture
overview,
I
would
like
to
to
say
a
couple
of
words
for
our
progress
since
the
last
time.
So
we
didn't
have
any.
We
didn't
held
any
interim
meetings
and,
of
course,
work
progressed,
but
it
progressed
at
a
pace.
That's
let's
say
less
visible
than
when
we
had
the
interims.
B
So
this
is
why
we'll
be
having
interims
starting
the
month
of
december,
so
there
are
already
two
that
have
been
planned
and
more
will
come
right
after
so
we
are
profiting
from
this
itf
to
actually
give
up
give
the
status
of
of
give
a
little
refresher
of
the
oklahoma
architecture,
and
so
I
think
I
have
a
little
bit
too
much
slides
for
the
a
couple
of
minutes
that
are
left.
So
I'm
just
going
to
go
through,
like
you
know
the
the
general
architecture
that
we
have.
B
We
have
the
devices
they
get
the
radiogate
with
the
network
gateway
and
the
application
server.
So
this
is
the
architecture
that
we
have
seen
the
start
and
just
as
a
slow
as
a
refresher,
so
we
have
different
technologies
and
in
each
of
the
technologies
you
have
like
the
specific
terminology
per
technology
and
the
the
technology,
the
the
terminology
that
we
decided
to
use
within
in
in
the
itf
documentation.
B
B
As
of
today's,
the
the
point
is
that
we
would
like
to
provide
the
reference
architecture
when
we
have
two
modes,
the
one
is
that
the
chic
will
have
the
sig
device
in
the
sheet
gateway,
that's
like
in
the
in
the
more
traditional
iot
setting,
and
then
when
we
have
the
sheet
peers,
such
as
in
the
pvp
in
the
ppp
document,
but
yesterday
on
six
law.
There
was
this
also
this
discussion
about:
okay,
how
we?
B
How
are
we
using
this
terminology
and
how
does
that
work
in
mesh
networks
where
there
is
no
uplink
and
downlink?
That
is
as
clearly
defined
you
know.
So
these
are
things
that
go
into
this
chic
architecture
and,
of
course,
how
all
these
things
bind
together
and
bind
together
with
the
yang
data
model.
So
these
are
things
that
are
discussed
and
that
will
be
discussed
more
in
the
sheik
architecture.
B
So
don't
be
also
one
thing
that
the
authors
of
the
of
this
document
would
like
to
express
is
to
not
to
not
hesitate
to
also
ping
us
to
if
you
have
like,
if
you
would
like
some
issues
and
some
questions
treated
before
the
other,
the
other
ones
right,
because
we
also
like
there
is
this
a
lot
of
things
that
we'll
need
to
describe
and
getting
some
input
on.
What
is
precious
for
you
on
the
first
time
is,
is,
of
course,
quite
important
for
us.
B
So,
yes,
I
mean
you
have
like
the
device
talking
to
the
network
application,
how
everything
works
together,
and
you
know
that
when
the
device
talks
to
the
to
the
she
gateway,
you
have
the
rule
manager
that
manages
the
rules,
and
then
we
have
the
the
compressor,
the
compressor
and
the
fragment
fragmenter
refragmenter,
and
we
have
the
same
setting.
But
you
know
like
if
it's
in
a
more
symmetric
view
and
if
it's
in
a
shake
peer
setting
such
as
the
ppp
document.
B
So
with
this
the
next
step
in
the
architectural
document
and
in
an
architectural
draft
and
with
pascal,
we
are
really
ambitious
about
this
document.
So
we
would
like
to
get
it
out.
I
would
like
to
to
to
get
it
fast
in
its
next.
Steps
is
really
to
go
through
the
life
cycle
of
how
how
the
the
whole
chic
life
cycle
works
right.
A
A
A
E
A
E
So,
yes,
I
I'm
gonna
present
the
new
version
that
has
been
uploaded
into
the
github
repository
and
so
it's
a
version
that
is
published
in
august
and
I
will
give
you
the
difference
between
this
version
and
the
previous
one
that
has
been
presented
at
the
during
the
past
etf
session.
So
next
slide,
please
all
right
so
here.
So
there
is
a
very
some
cosmetic
things
like
adding
a
dash
between
the
name
to
be
closer
to
be
better
understood
and
be
closer
also
to
what
we
find
in
rfc
8724.
E
E
F
A
E
Maybe
there
is
two
two
possibilities,
so
what
I
understood-
but
maybe
I'm
wrong,
was
to
ask
in
the
compound
arc
document
some
to
add
things
to
the
young
model,
but
the
other
approach
could
be
to
extend
the
pure,
the
young
model,
only
based
on
8724,
and
then
we
add,
we
augment
the
model
with
a
new
draft.
So
I
don't.
I
don't
know
what
what's
the
best
to.
E
B
E
B
A
G
A
A
D
E
So
what
is
new
also
is
that
we
put
relation
between
the
different
leaf
in
the
model,
for
example,
here
what
we
say
that
if
you
have
a
matching
operator,
so
if
it's
ignore
it
doesn't
need
a
target
value
and
any
of
the
other
cases.
A
target
value
has
to
be
defined
and
same
thing.
If
we
have
matching
operator
matching
opera,
if
we
have
sorry
lsb
msb,
it
requires
also
to
have
argument
after
that.
E
E
So
in
conclusion,
we
we
have
something
that
is
quite
stable.
Now
I
accept
what
you
say
about
componek,
but
it
will
it's
very
easy
to
to
change.
So
it
means
that
I
think
it's
very
mature
now.
It
has
also
been
tested
during
the
acaton,
so
we
introduce
the
young
model
into
open,
chick
and
now
we're
able
to
transform
json,
open,
cheek
format
to
the
young
model.
E
So
it's
it
has
been
pushed
to
them
to
the
open
cheek
repository,
but
it
needs
to
be
finalized,
but
it's
in
a
good
way
and
finally,
next
slide.
So
the
next
slide
is
an
example
of
this.
I
think
yes,
so
here
is
a
cheek
data
model
on
next
slide
is
what
we
get
with
the
young
data
model
to
represent
part
of
the
rule,
and
the
next
slide
is
the
relation
between
all
the
the
draft.
E
B
B
E
H
B
A
A
A
C
Process
wise,
you
can
do
both
at
the
same
time
and
typically
during
the
working
group
last
call
you
can
ask
as
well
the
young
doctor's
review
if
we
are
confident
into
the
young
model.
That's
the
way
to
go
else.
If
you
are
unsure
about
the
young
model
itself,
then
it's
better
to
think
to
go
to
the
young
doctor
first,
because
if
they
need
to
change,
then
we
need,
of
course,
to
to
do
everything.
Good
class
called
other
change.
A
A
A
E
A
A
A
C
A
C
C
E
A
I
Just
as
a
reminder,
since
what
has
happened
since
ietf
111,
we
first
made
a
revision
from
five
to
six.
Where
we
added
the
sheet
compound
dac
mechanism,
then
we
had
the
discussion
about
what
to
do
with
this
compound
act,
to
make
it
generic
and
following
the
decision
from
the
group,
we
made
it
a
separate
document.
I
I
So
then,
after
doing
that
split
from
revolution,
seven
to
eight,
we
added
some
more
normative
references
to
the
compound
dac,
of
course,
and
the
usage
that
has
to
be
done
in
chico
versus
sigfox,
together
with
signaling
examples
and
abort
definitions.
So
that's
mainly
what
we
have
concentrated
in
this
last
version
of
the
sheikh
overseek
fox.
F
Hello,
we
did
a
small
hackathon
with
the
cheek
over
sigfox
draft.
The
idea
was
to
to
propose
a
real
case
scenario.
In
this
case
we
talked
about
a
mini
weather
station
and
the
objective
was
to
measure
at
least
temperature
and
humidity
and
send
it
via
a
json
file
to
well
while
using
check
over
stick
fox.
In
this
case
we
use
the
dilopi
and
the
pythons,
and
the
codebase
that
we
use
is
the
one
that
is
in
g
top
right
now.
That
is
the
cheek
over
sigfox
project.
F
Next
slide,
please,
in
the
results
that
we
got
well,
we
implemented
the
solution
using
two
threads.
One
thread
was
used
to
to
do
all
the
sensing
and
the
other
thread
was
used
to
handle
all
that
she
covered
sig
fox
communication,
all
the
fragmentation
transmissions
retransmissions
and
we
connected
both
threads
using
a
list.
F
In
this
case
there
you
can
see
the
json
file
format.
We
use
two
examples:
the
first
ones
the
first
one
also
send
temperature
humidity
battery
voltage
and
light,
and
the
second
one
on
this,
then
only
got
that
temperature
and
humidity
the
first.
The
first
case
we
got
seven
cheek
fragments.
That
is
exactly
one
window
and
in
the
second
case
we
got
four
cheek,
fragments
and
well
on
the
right.
B
I
The
document
I
mean
we
still
added
a
few,
a
few
more
details,
and
I
guess
once
we
have
the
other
two
documents:
stable,
the
young
and
the
compound
act
or
in
our
case
the
compound
neck.
Then
then
we
will
be
able
to
to
make
the
the
changes
and-
and
do
a
last
call,
so
we're
really
close
to
it.
We
just
want
to
make
sure
that
the
the
dependencies
are
are
stable
enough,
so
that
we
can
make
the
correct
adjustments.
I
I
I
A
Just
recently,
so,
oh
and
from
me
actually
in
my
review,
there's
this
discussion
about
the
data
model
and-
and
we
discussed
it
again
today-
what
I
would
really
like
to
see
is
is
lauren
to
actually
look
at
the
data
model
piece
that
you
have
in
the
component
hack
as
part
of
the
war
group.
Let's
go
laura
if
you
could
please
do
that.
Go
ahead.
Just
just
speak.
Low,
yeah.
I
A
E
I
I
D
Yeah,
I
don't
know,
do
you
want
to
discuss
it
now?
Do
we
have
time
and
did
you
have
time
to
read
the
comments?
Otherwise
we
can
take
that
offline,
but
I
was
curious,
for
example,
if
you
use
the
compressed
version
of
the
bitmaps
in
your
sigfox
implementation,
because
I
don't
quite
understand
this-
how
this
would
work.
F
I
wanted
to
to
comment
yeah.
Actually,
I
think
that
the
compressed
bitmap
is
not
supported
for
the
intermediate
windows
in
this
case,
as
the
bitmap
should
be
a
fixed
field
size,
so
yeah.
We
need
to
to
fix
that
to
to
make
sure
it
says
that
only
the
last
bitmap
can
be
compressed.
Actually,
we
were
looking
at
that
the
last
bitmap
can
be
compressed,
but
intermediate
bitmap
should
be
off
right.
A
In
my
review,
there's
also
this
question
about
nxd.
Are
you
the
famous
straightforward
xd
of
the
chic
rfc,
which
says
what
you
need
to
provide
in
the
profiles
and
in
the
document
you
seem
to
indicate
that
there
are
some
things
which
should
be
in
the
profile,
but
then
there
is
no
no
such
list.
As
you
know,
here
is
the
extension
to
an
xd
and
it
would
be
nice
to
see
it
more
explicitly.
A
A
A
A
B
Okay,
so
thank
you
very
much,
juan
carlos,
so
yeah
that
that
will
be
the
the
next
thing.
The
respond
to
the
comments-
and
you
know,
integrate
the
part
that
comes
from
the
on
the
yang
yang
model.
Yes
and
eric
would
like
to.
Maybe
you
can
add
a
point
on
on
this.
So
do
we
keep
the
working
group,
the
open
class
call
open,
or
do
we.
C
B
B
C
B
H
Thank
you
alex
hello.
Well,
I
will
present
this
new
version
for
nviot
next
slide.
Please
we
received
some
comments
from
the
mailing
list,
principally
from
evo
and
lauren,
so
we
have
put
all
these
new
inputs
in
this
sixth
version.
Next
slide,
please,
the
the
most
important
changements
are
the
terminology
where
we
use
now
the
terminology
from
ietf.
H
So
instead
of
user
equipment
we
put
device
and
for
the
node
v
we
put
radio
gateway
and
then
for
the
network
gateways.
I
leave
the
I
put
network
gateways
together
with
the
name
of
the
get
get
the
network
gateway
of
the
nviot,
because
if
not,
I
have
to
put
network
guideway
one
two,
three,
four,
five
and
so
forth,
so
it
was
better
to
leave.
I
think
it
was
better
to
leave
the
name
together
with
the
network
gateway
for
these
devices.
H
H
So
I
the
architecture,
now
is
very
flat,
but
this
means
that
each
network
gateway
will
be
a
very
principal
device
in
the
architecture,
and
so
the
network,
mme
and
the
network
gateway
csgw
that
normally
are
in
the
same,
are
in
the
same
device
and
now
2d
device
different
devices,
because
their
connections
are
different.
One
will
be
for
a
very
small
bandwidth
with
the
mme
and
the
other
is
a
normal
radio
connection
with
data
for
the
csgw.
H
H
B
B
B
H
So
one
point
that
we
were
discussing
with
edgar
this
monday
is
that
this
new,
this
version
six,
is
really
ready
for
rascal.
But
if
we
want
to
work
with
light,
color
and
transfer
objects,
then
it's
not
ready.
We
need
to
work
this
part,
so
we
work
in
a
new
document
in
a
version
two
documents
or
we
do
it
in
this
version.
Six.
J
J
J
A
D
D
D
Yeah
with
girls
does
that
work,
yeah,
so
yeah
after
years
of
working
on
chickens.
This
is
a
happy
time
where
you
see
adoption,
and
so
I
wanted
to
report
about
that
so
that
we
can
celebrate
so
we're
not
it's
not
quite
complete,
but
pretty
good
shape
at
the
royal
lines
and
the
the
first
driving
use
case
is
dlms
over
lower
one.
H
H
D
So
the
proposal
was
set
to
use
check
over
r1
to
compress
udp
and
ipv6,
and
that's
what
was
adopted,
and
so
the
nice
thing
is
that
you
hardly
have
anything
new
to
specify
everything
is
already
in
place,
just
stitch
the
pieces
together
and
so
a
technical
speed
specification
is
being
written
right
now
at
the
lower
lines
and
just
to
say
how
this
works,
and
basically
the
spec
is
just
a
few
pages.
It
says
you
shall
use
rfc8724
rfc9011
and
that's
about
it.
D
Document
because
the
lore
alliance
does
certification,
which
the
atf
doesn't
do
so
that's
an
expectation
before
this
gets
in
place
and
future
work.
I'm
not
sure
how
much
how
far
we
will
go
into
that
direction,
but
would
be
to
investigate
if
lauren
can
be
used
with
matter,
and
you
know
the
connected
smart
homes
under
the
csa
alliance
and
so
chic
would
then
also
be
used
to
carry
ipv6
over
r1
in
that
context.
D
So
that's
the
work
at
the
lower
alliance.
I
forgot
to
mention
the
editors
of
the
technical
specification.
I
think
it's
important
to
highlight
the
contribution
by
clicking
aclio
semtech
people
that
you
know
some
of
the
archers
of
this
working
group,
for
example,
and
contributors
by
many
companies
and
institutions
and
family
to
to
most
of
you
in
this
working
group.
B
A
Yes,
I
will
present
chicago
ppp,
so
basically
we
we
we
discussed
how
we
can
do
shake
about
pvp,
so
we
can
do
shake
over
the
ethernet
using
pppoe
and
see
if
we
have
ethernet,
we
have
wi-fi
so
basically
doing
pvp
gives
us
gives
us
all
the
serial
links
it
gives
us
point-to-point
ethernet
and
gives
us
wi-fi,
so
that
was
kind
of
cool
to
do.
Shigeru
rpp,
but
the
prime
was
ppp,
is
not
part
of
the
charter
as
one
of
the
technologies
that
we
support.
A
So
we
we
decided
to
try
our
luck
at
interior,
so
we
did,
and
the
document
has
spent
pretty
much
a
year
at
that
interior
and
I
failed
to
get
any
command
feedback
return
from
from
interior,
not
much
interest
in
check
and
probably
not
much
interest
muslim
ppp.
That's
this!
Even
if
interior
is
supposed
to
do
everything
I
mean
this
is
this
looks
like
a
thing
of
the
past.
A
C
I
simply
want
to
second
what
you
said
and
that's
correct,
so
the
expertise
in
shake
is
obviously
in
this
group,
so
why
this
document
cannot
be
adopted
by
the
working
group
because
it
doesn't
fit
the
charter.
The
experts
are
still
in
this
working
group,
so
I
count
on
this
working
group
to
make
a
review
and
then,
when
you
are
basically
done,
the
kind
of
working
group
classical
doesn't
say
its
name,
because
there's
not
a
real
group
last
call,
but
everyone
is
happy
with
it.
C
Pascal
then,
just
contact
me
over
email,
and
I
I
I
will
do
the
follow-up,
so
basically
eddie
sponsored
his
ietf
stream.
So
it's
pretty
much
like
any
working
group
rfc,
except
that
there
is
no
group
last
call,
but
the
rest
is
the
same
right.
There's
an
itf
white,
that's
called
isg
evaluation
and
you
can
be
a
standard
document.
A
That's
that's
perfect,
so
so
that's
pretty
much
where
we
are
so.
Yes,
I
will
post
again
on
the
mailing
list
asking
people
to
review
the
draft.
I
think
it's
it's
expired
or
getting
expired.
It
is
expired.
So
I
will
republish
a
note,
4
and
ask
you
know
post
to
the
rp1
mailing
list
to
ask
for
for
reviews
and
with
this
we
are
concluding
this
meeting.
So
oh
question
by
the
way:
go
ahead.
Dominique
yeah.