►
From YouTube: IETF-SATP-20230314-1500
Description
SATP meeting session at IETF
2023/03/14 1500
https://datatracker.ietf.org/meeting//proceedings/
B
C
B
C
C
B
B
So
today
is
a
standard
workers,
Obsession
or
an
interim
session.
I
should
say
so
we'll
go
through
the
agenda
in
a
moment
when
everyone's
had
a
chance
to
log
in
it
was
made
available
on
data
check,
I,
don't
know
if
you've
had
a
chance
to
see.
It
probably
will
be
a
recap
from
the
last
month
of
the
authors
that
have
spent
time
talking
to
each
other
and
making
any
kind
of
advancements
to
the
documents
and
preparation
for
the
working
group.
B
D
B
D
E
A
D
A
D
A
D
A
D
A
Okay,
so
yes
go
ahead.
No,
please
go
ahead.
I
don't
know
this
is
by
the
way.
This
is
normal.
Every
time
there's
a
meat,
Echo
and
and
it's
even
more
complicated
if
the
chairs
are
in
the
room
physically
in
the
room
and
a
whole
bunch
of
people
are
like
a
whole.
You
know
remote
is
like
okay,
there's
always
like
technical
issues.
So
this
is,
this
is
normal,
I
think
I,
hope,
I,
hope,
bro.
You
know.
Claire
comes
back
soon.
D
A
And
I've
asked
the
chairs
that
if
we
have
extra
time
I'm
happy
to
like
you
know
given
like
overview
of
the
whole
thing
like
for
20
minutes,
because
they
might
be,
you
know
some
people,
you
know,
particularly
in
Japan,
who
are
new
to
this.
So
just
having
you
know
what
is
this
about?
You
know
just
basically
the
old
slides
that
we've
been
presenting
for
the
you
know
the
last
two
Buffs
and
before
that.
A
Okay
and
you
know
in
other
working
groups.
I
do
know
that
I
mean
if
you
have
you
know.
If
it
is
like,
you
know,
10
minute
slot,
then
somebody
might
present
on
just
you
know
a
rough
sketch
of
an
idea.
Just
to
get
immediate
feedback,
like
you
know
is
this
is
like
good
idea.
Bad
idea,
like
you
know,
you
know,
go
go
to
the
pub
you
know
kind
of.
A
C
A
C
C
A
D
A
Oh,
oh,
there
you
go
okay,
so
so
should
I
should
I
just
start
should
I
just
okay,
so
folks
welcome
G.
This
is
deja
vu.
Happy.
Have
we
not
been
here
before,
but
using
a
different
tool,
so
I
think
the
first
thing
is
to
show
the
slides
or
is
it
the
the
note?
Well
slide,
let
me
let
me
try
presentation
View
and
for
the
last
minute
on
how
to
do
that.
How
do
I
actually
how
to
actually
present
anything.
A
Oh,
dear
okay,
so
I'm
not
I'm,
not
I'm,
not
logged
in
this
I'm,
not
logged
in
as
the
chair,
so
I
I
have
to
have
to
request
sharing
with
you,
but
you
know
what
you
guys
know
what
the
ITF
notebook
slide
are.
But
so
let
me
let
me
get
going
I
think.
The
first
item
that
Claire
had
on
the
agenda
was
this,
the
the
updates.
So
let
me
describe
the
updates
done
to
the
architecture
document.
A
So
the
big
thing
is
that
now
both
the
architecture
document
and
the
SAT
core
document
matches
the
color
version
16
of
our
nice
color
flow
diagram,
except
for
the
messages
going
out
from
the
Gateway
into
the
network
right.
So
so
it's
just
it's
just
the
inside
Lane,
that's
being
recorded,
captured
in
the
architecture
and
in
the
core
sapi
core
protocol
document.
A
That's
that's
the
first
thing
so
no
surprises
there's
this
is
it
that
me
me
catching
up
the
documents
to
to
the
to
the
you
know
flow
diagram.
The
second
thing
is
that
in
the
architecture
I've
added
a
terminology,
there
context
ID
and
session
ID,
which
we
can
this.
This
is
the
topic
of
discussion
today
is
like.
Is
that
one
ID
is
it?
Is
it
well?
One
has
to
be
connected
to
the
other.
A
How
do
we
do
that
if
there's
two
IDs
and
so
on,
but
I've
put
it
into
the
terminology
section
of
the
architecture
document
with
the
expectation
that
that
it
will
be
copied
out
and
moved
route
into
our
a
separate
document?
I
believe
we
want
to
do
some
kind
of
a
glossary
glossary
document.
What
do
we
call
it?
Terminology
document
is
what
was
where's
called
it?
So
so
it's
there
it's
you
know,
I
think
the
last
two
entries
in
the
terminology.
A
A
Yeah
yeah,
okay,
so
so
maybe
maybe
we
could
just
compile
everything
you
know
in
one
like
Word
document
or
one
draft,
and
then
we
just
go
through
and
say
is
this?
Is
this?
Do
we
like
what
it
says
there?
You
know
like,
for
example,
the
word
like
the
term
blockchain.
We
don't
have
to
like
Define
that
just
literally
copy
whatever
somebody
else
says
like
the
nist
document.
It's
it's
got
its
own
glossary
at
the
end.
A
You
know,
so
we
don't
have
to
like
reinvent
criminologies
if
it's
already
there,
particularly
if
it
already
comes
from
nist,
which
means
it's
you
know,
people
have
seen
it
before
and
sort
of
have
you
know.
Was
it
capture
captured
it
sort
of
mentally,
probably
the
the
third
thing
or
the
fourth
day?
The
third
thing
is
that
in
the
core
document
now
every
flow
you
know
previously
did
not
do
this.
A
Every
floor
now
has
a
session
ID
right,
which
is
interesting,
that
that
I
guess
we
we
kind
of
I
I,
did
I
kind
of
did
not
add
that
so
now,
literally
every.
If
you
look
at
the
the
latest
version
from
two
days
ago,
anything
coming
from
the
client
Gateway
One
to
Gateway
two
to
the
server
has
a
session
ID,
which
is
independent
of
there's
a
thing
called
the
counter.
So
the
idea
here
is
that
the
the
client
and
the
server
has
has
their
own.
A
The
Gateway
has
their
own
local
way
of
tracking
I.
Think
we
just
called
it
the
counter
value,
but
the
session
ID
Remains,
the
Same
for
the
beginning
of
the
session,
and
so
the
session
ID
in
stage
zero
or
stage
one
identifies
the
asset
to
be
transferred
so
that
we
don't
have
to
repeat
what
asset
has
to
be
transferred
within
each
message,
and
we
could
do
a
hash
of
course.
But
but
you
know
what
I
mean
so
The
Binding
between
you
know.
You
know
I
want
to
move
this.
A
You
know
tokenized
asset
bound
on
this
session.
Id
occurs
in
stage
zero,
and
if
you
look
at
it
in
fact,
the
the
flow
doesn't
really
have
any
asset
being
being
moved.
It's
just
it's
just
literally
an
agreement
flow
that
proceeds
and
then
fourthly,
which
I
think
is
on
my
slide,
which
I
can't
show
is
that
we
may
need
to
define
the
lock
assertion.
A
A
So
back
again,
so
so
I'm,
you
know
I'm
in
confidence
every
time
I
talk
about
this
outside
the
ITF.
It's
this
three
times
now
I've
been
asked
well
like
who's
gonna
use
this.
How
do
we
use
this
and
I
kind
of
hand,
wave
and
say
like
well,
there's
an
application?
That's
going
to
drive
this
and
then
next
week
as
well.
How
does
the
application
know-
which
instance
you
know
is
it
referring
to,
and
this
is
why
we
ended
up
having
you
know,
so-called
context,
ID
and
session
ID,
but
I.
A
A
D
B
B
B
D
B
A
F
E
B
Guys
that
are
authors,
will
you
make
sure
so
Rama
and
Thomas
and
whoever
I
think
Alex
and
Dennis
are
co-authoring
the
terminology
document
and
we
have
the
final
draft
over
to
either
me
or
where's.
Ideally
the
Wednesday
before
our
session
on
the
Friday
just
so
we
can
make
sure
they
are
available
ready
for
the
working
group.
A
A
D
A
B
It
should
be
a
coordination,
but
we
do
need
to
make
sure
there
is
one
or
two
authors
who
are
the
ones
that
edit
it
based
on
everyone's
submission.
So
it
should
be
a
collaboration
as
with
all
our
documents,
but
we
do
need
to
as
with
again
with
as
all
documents
someone
needs
to
own
the
editing
and
making
sure
the
changes
are
committed
in
line
with
the
the
group's
agreement
and
I
know.
B
A
A
Okay,
so
so
folks,
that's
kind
of
the
end
of
of
my
sort
of
presentation.
It's
just
basically
reporting
that
and
the
fact
that
we
have
to
talk
about
the
session
ID
and
context
ID
and
you
know,
is
it
the
same
thing
or
is
the
session
ID
derived
from
you
know,
context
ID
and
so
yeah
happy
happy
to
have
that
discussion.
E
A
E
E
A
E
E
There
is
a
context
that
was
for
Global
transactions
among
different
sources,
and
in
there
we
could
define
various
things,
and
potentially
we
could
use
this,
because
when
we
initiate
a
transfer,
this
whole
context
about
which
gateways
I
mean
you
know
all
the
eventually
things.
The
networks
and
the
session
could
be
part
of
one
thing,
so
it
could
be
a
composite
thing.
It
could
be
even
a
you
know
like
a
small
Json,
something
that
is
the
context
that
contains
also
the
the
session
ID.
A
Yeah,
because
because
this
question
is,
is
if,
if
Network
one
uses
uses
either
some
bft
variant,
you
know
to
choose
to
elect
the
Gateway.
The
Gateway
would
not
have
known
anything
about
this
until
it
gets
elected
and
then
the
and
somehow
the
app
one
needs
to
now
talk
to
Gateway,
One
and
say:
hey
here's
my
context.
That's.
E
That's
where
the
the
line
between
what
we
communicate,
what
we
negotiate
in
stage
zero
might
have
an
impact
on
what
would
be
the
things
that
we
transfer
as
part
of
the
the
unique
session
the
instance
of
the
transfer,
including
the
session
ID.
So
maybe
there
is
a
chicken
in
the
egg
thing.
If
we
don't
discuss
you
know
stage
zero,
maybe
I
mean
we
are.
C
B
B
E
Right
so
the
thing
that
I'm
I'm
trying
to
capture
here
is
the
okay
you're.
Seeing
the
gray
down
in
the
bottom
here
sat
p
is
the
thing:
is
the
the
sequence
diagram
that
Thomas
has
version
2,
500
I,
don't
know
what
version
number
you
have
Thomas
now
in
the
your
sequence
diagram?
This
is
the
gray
box
down
here:
16
16,
okay,
so
the
idea
here
is
to
see
what
we
could
do
eventually
in
order
to
get
to
start
the
the
software
protocol.
E
E
So
they
think
the
idea
here
is
the
following.
The
client
goes
to
the
Gateway
and
say:
I
would
like
to
transfer
something:
I
mean
set
of
assets,
so
it
receives
back
the
context
and
the
the
transfer
context
is
something
that
is
identified
by
the
Gateway
think
about
something
that
could
potentially
be
signed
by
the
private
key
of
the
Gateway.
Something
so
Gateway.
One
can
recognize
that
this
transfer
context
belongs
to
to
to
to
itself.
E
B
B
C
E
B
E
This
is
the
the
phase
before
starting
a
SAT.
The
SAT
flow.
Again,
it's
it's
the
gray
box
down
here.
So
it's
everything
that
may
eventually
happen
before
the
idea
is
that
the
client
goes
to
the
Gateway
and
the
Gateway
gives
back
a
context
that
when
somebody
sees
it
can
recognize
that
it
comes
from
the
specific
Gateway
okay,
so
it
could
be.
For
example,
it
could
be
signed
with
a
private
key
that
is
owned
by
the
Gateway
or
any
other
mechanism.
But
the
thing
is
that
this
transfer
context
belongs
to
Gateway
One.
E
So
when
the
context
is
received
by
the
client,
then
the
client
can
go
to
the
network
so
again,
systems
here
are
networks,
dlts
or
could
be
other
systems,
and
it
binds
the
assets
to
the
specific
transfer
context,
which
means
that
I'm
saying
that
I
have.
You
know
this
specific,
for
example,
token
that
I
would
like
to
transfer
to
client
two
and.
D
E
E
So
some
part
of
the
context
should
be
standardized.
For
example,
to
my
view,
it
might
contain
the
session
ID.
For
example,
what
we
see
a
session
ID
could
be
part
of
the
context,
so
I
initiate
a
transfer
and
as
the
Gateway,
which
is
connected
to
the
client
that
initiates
this
transfer
I
create
a
session
ID
which
should
be
unique.
You
know
across
all
transfers.
E
D
E
E
Oh
okay,
if
we
see
if
we
see
the
local
session,
for
example,
we're
having
the
Gateway
definition,
we
have
a
bunch
of
assets
that
are
locked
by
the
DLT
and
there
is
also
somewhere,
if
I,
remember
the
results,
also
session
ID
somewhere
there.
So
what
what
it's
proposed
here
is
to
have
a
transfer
ID
that
refers
to
the
context.
The
context
could
be
a
Json
file,
for
example,
with
many
things,
including
the
session
ID.
E
Correct
and
I
would
like
to
bind
this
specific.
For
example,
say
you
have
a
token
to
transfer
a
non-fungible
talking
to
transfer.
This
asset
is
bound
to
this
specific
transfer
context.
So
when
later
on
inside
satp,
the
Gateway
goes
to
the
network
and
say
I
would
like
to
lock
it.
It
will
pass
the
transfer
context,
so
the
system
would
know
that
it
comes
from
the
Gateway
and
is
consistent
with
the
initial
step.
Here.
E
That's
also
is
you
know
answering
to
your
comment
about
the
optimistic
locking.
You
could
have
many
clients
that
are
binding
the
asset,
but
there
is
one
I
mean
one
Gateway
that
will
do
the
transfer.
At
the
end,
there
is
nothing
preventing
clients
to
bind
assets
to
transfer
contacts.
It
doesn't
mean
that
you
know
the
the
first
gateway.
The
first
transaction
context
will
transfer,
it
could
be
optimistic,
could
be
the
one
that
comes
first
or
whatever
condition.
D
E
This
could
be
yes,
it
could
be
anything
that
the
network
can
understand.
For
instance,
I
mean
if
we
take
the
distributed
transaction
processing,
the
the
network,
the
DLT,
can
intentionally
pull
out
from
the
transfer.
At
any
moment,
they
don't
have
to,
you
know
continue,
and
it
could
be
a
roll
back.
Then,
in
that
case,.
E
It's
it's
a
I
mean
all
the
things
that
I'm
discussing
here
are
conditions
for
all
back
inside
the
SAT,
so
I'm
not
discussing
about
the
specific
conditions
for
rolling
back,
but
I'm,
saying
that
here
there
is
a
binding
of
a
specific
asset
to
a
context.
So
let
me
go
through
that
in
two
minutes
and
then
we
can
have
the
discussion
just
for
you
to
have
all
the
the
flow
and
see
if
it
makes
sense
or
not.
E
E
So
what
would
be
that
moment
when
Bob
receives
the
context?
It
goes
to
the
local
system
and
say
there
will
be
a
transfer
that
is
going
to
happen
for
the
moment.
I
mean
this
might
be
also
something
that
we
that
doesn't
make
I
mean
there
is
no
effect
on
the
DLT,
because
there
is
no
asset
yet
in
the
DLT.
E
The
important
thing
is
is
number
seven
is
when
the
Bob
goes
to
its
Gateway
and
say:
I
have
a
context
that
comes
from
a
remote
application
that
I'm
passing
to
you
and
because
the
transfer
context
contains
information
about
the
Gateway,
the
Gateway
2
will
go
back
to
Gateway,
One
and
say:
okay,
I
am
the
other
party
and
I'm
giving
you
back
the
context
that
you
have
created.
There
is
a
typo
here.
Also
it's
a
it's
a
transfer.
E
It's
not
a
transaction
context
in
eight,
so
as
a
Gateway
2
I'm,
connecting
with
you
Gateway,
One
and
I'm,
giving
you
the
the
context
that
you
have
created.
So
you
could
recognize
that
this
comes
from
you.
So
we
have
the
two
gateways
connected
and
since
we
have
these
two
gateways
connected,
Gateway
One
can
then
initiate
sat
so
sat
happens
and
at
the
end
we
can
have
also
the
client
that
goes
back
to
the
Gateway
and
say
what
happened.
What
is
the
completion
of
sat?
Is
it
commit?
F
Sorry
for
that
so
Dennis
one
question
I
had
on
the
number
four
is
here:
we're
assuming
the
client
apps
are
different
entities
right
and
Alice
is
trying
to
send
an
asset
to
Bob,
but
for
all
intents
and
purposes
it
could
be
Alice
trying
to
transfer
an
asset
to
themselves
on
another
Network.
So
you
know,
a
lot
of
this
stuff
is
very,
very
volatile.
E
F
E
To
do
the
thing
is
that
in
the
remote
in
the
remote
side
there
might
be
four
that
goes
directly
to
Gateway
2,
for
example.
If
this
is
the
case
or
but
the
thing
is
that,
on
the
other
side,
on
the
other
end,
okay,
there
might
be
shortcut
if
client,
one
I
mean
Alice
is
Alice
on
the
other
side.
But
if
you
take
it
in
the
generality
of
what
we
are
trying
to
do
here
is
that
we
are
having
two
parties
that
they
would
like
to
transfer
an
asset.
B
Yes
and
we
we
have
assumed
as
part
as
part
of
the
satp,
we
have
assumed
that
client,
one
and
client
two
person,
one
person
to
system
one
system,
two
have
had
a
method
of
communicating
with
each
other
I.
Just
don't
want
us
to
be
spending
too
much
time
as
a
working
group,
two
lost
in
the
detail
of
what
happens
before
the
SAT
is
initiated
as
I
feel
like
we.
E
And
so
there
is
not
so
much
about
extending
it's
more.
The
fact
that
there
are
things
that
are
not
specified
in
the
in
the
protocol.
For
example,
we
don't
know
how
Gateway
One
and
Gateway
2
are
going
to
be
somehow
connected.
So
there
are
some
things
that
are
still
not
necessarily
100
specified,
yeah.
B
E
B
E
I'm
not
mistaken.
Yes,
if
yeah,
if
they're
not
not
mistaken,
there
is
somewhere
in
the
original
architecture.
Maybe
it's
no
doubt
not
an
SAT
but
in
the
early
versions
on
what
happens
in
the
beginning.
So
somehow
the
bootstrapping
protocol
of
essay
and
bootstrapping
part
of
the
protocol
should
be
defined.
So
you
know
I'm
very
happy
to
to
to
write
a
draft
on
that.
B
C
B
A
As
much
as
possible,
so
just
so
Dennis
in
the
in
the
sad
core,
it's
still
there,
there
is
a
type
so
there's
three
API
types
which
is
type
one
which
is
to
go
from
the
application
to
the
Gateway.
But
we
never
talk
about
that.
It's
there
in
the
diagram,
but
not
no
text.
All
we
talk
about
is
the
type
2
which
is
the
one
that's
supposed
to
be
implemented
by
the
two
gateways
back
and
forth,
but
it's
still
there
from
the
it's
you
know
hasn't
in
fact,
I,
don't
think.
A
D
D
So
so
what
I
understood
from
the
protocol
preceding
satp
was
it's
kind
of
a
recommendation,
but
what
It
ultimately
achieves
is
it?
It
puts
the
appropriate
state.
It
puts
the
gateways
in
the
appropriate
state
where
they
are
ready
to
start
on
satp
right,
correct,
so
and-
and
we
can
also
Define
the
format
of
that
data
structure
that
they
both
are
initialized
with.
So
I
think
to
that
achieve
that
we
need
to
show
that
and
I
think
what
state
they
start
off
with.
That
is
part
of
the
sap
scope.
E
Yes,
I
mean
the.
There
are
a
few
things
that
okay,
we
don't
know
how
the
binding
between
the
gateways
happens.
So
we
need
to
establish
this
and
also
we
need
somehow
to
communicate
the
the
transfer
context,
okay
to
the
to
the
networks.
Also,
so
this
communication
between
the
the
client,
the
Gateway
and
the
network
somehow
has
to
be
established.
So
that
was
this
stage.
Zero
kind
of
flow
that
I
I
shown
here
is.
D
Now
the
details
of
that
I
think
were
kept
out
of
scope
in
the
charter,
though
the
outcome
of
that
is
in
the
scope,
right,
I
think.
If
we
make
that
fine
line,
then
we
can
probably
satisfy
player
that
yeah.
Maybe
we
need
a
little
bit
of
discussion
on
that
because
at
least
the
end
result
of
the
protocol
is
part
of
the
satp
score.
E
Thinking
about
the
way,
if
you
go
to
implement
sat
there
are
several
unknowns
there,
and
so
that's
what
what
I
was
trying
to
capture
is
all
the
the
steps
that
are
necessary
in
order
for
SATs
to
you
know,
to
start
when
you
say
begin
transfer,
there
are
a
lot
of
magic
things
that
are
happening.
I
mean
Gateway,
knows
about
the
network,
I
mean
the
assets
somehow
are
abandoned
and
are
locked
how
all
these
things
get
somehow
you
know
put
all
together.
So
that's
that's
the
the
thinking
behind
it.
A
E
E
A
E
F
E
F
We
should
focus
on
the
assumptions
and
what
stage
zero
means
like
what
are
the
preconditions
that
should
have
happened
at
stage
zero.
How
does
you
know
in
that
state?
Look,
but
not
necessarily
how
you
know
the
gateways
or
client
applications
get
there.
It's
just
okay.
What
are
the
hard
requirements
of
information
that
needs
to
be
available?
You
know
one
Gateway
or
the
other.
If.
E
F
E
F
B
E
B
B
A
Not
for
the
ATS
specific
but
I
have
I've
been
asked
by
a
couple
of
people
independently,
particularly
people
in
Australia.
Is
it
possible
to
have
the
interim
group
meetings
flip-flopping
time
zones
so
that
the
people
in
Australia
can
try
to
dial
in
so
and
and
also
people
from
Pacific
United
States,
which
I
think
it's
okay,
but
particularly
people
in
Australia
can
dial
in
so
one
meeting
current
time
zone.
The
next
meeting
a
different
time
zone,
yeah.
B
Yeah,
so
any
slides
that
you
want
to
present
any
documents
that
are
being
worked
and
I've
spoken
to
the
document,
authors
separately,
so
Dennis
yourself
and
Thomas.
If
I
could
have
those
documents,
the
last
draft
ready
by
Wednesday
the
29th
I,
think
it
would
be
that
way.
We
won't
have
the
the
situation
we've
had
here
in
terms
of
trying
to
share
it
on
the
the
meat
Echo.
B
We
have
Thomas
has
got
a
couple
of
documents
that
he's
lead
author
on
I'll
speak
to
Martin
separately
is,
is
I
work
with
him
and
Rama
I
know.
You've
been
working
a
bit
more
on
the
use
cases,
so
the
final
vote,
or
the
most
up-to-date
version
of
that
document,
if
we'd
like
to
discuss
them
so
we've
we've
assigned
some
time
in
the
working
group
to
the
discussion
of
these
documents
is
that's.
B
What
we're
here
to
do
is
get
these
documents
out
just
to
set
the
scene
a
little
bit
for
the
guys
that
are
the
main
authors.
It
would
be
five
minutes
of
an
introduction
of
the
topic
and
then
open
floor
discussion.
So
you
don't
need
to
do
too
much
in
terms
of
slides
to
present
it
would
be
more
introducing
and
then
the
open
working
group
doing
what
they
do
best.
B
Fabulous
so
with
that
said,
make
sure
everyone's
got
their
tickets,
Friday
9
A.M,
Japan
time
we've
got
two
hours.
We
have
quite
a
bit
to
go
through.
We've
been
fairly
conservative
with
our
time
slots,
but
know
that
if
we
do
go
over,
we'll
still
have
enough
time
to
get
all
the
agenda
covered.
So
we're
very
excited
about
that
and,
like
I
said
any
last
agenda
points
reach
out
to
me
or
where's
or
both
of
us.