►
From YouTube: IETF105-SACM-20190725-1740
Description
SACM meeting session at IETF105
2019/07/25 1740
https://datatracker.ietf.org/meeting/105/proceedings/
A
B
C
E
C
G
B
A
B
B
B
I
J
K
B
B
D
B
D
G
I
have
a
very
small
number
of
slides.
Hopefully
this
will
be
super
fast,
so
I
have
done
some
updates
on
the
software
descriptor
extension
to
Rowley,
in
accordance
with
the
review
that
I
received
the
first
round
of
reviews
that
I
got
when
the
working
group
last
call
first
started
we're
all
integrated
into
version
7,
and
then
we
had
the
extended
working
group.
Last
call
and
I
got
an
additional
several
reviews.
That's
all
been
integrated
into
version
8,
which
I
have
posted.
G
It
was
like
at
the
start
of
this
week
like
Sunday,
or
something
like
that.
So
the
most
recent
version
reflects
all
of
the
review
that
I've
gotten
during
working
group.
Last
call
I
had
a
great
lot
of
really
great
comments
and
a
lot
of
really
great
fixes
and
little
things
like
that.
Next
slide,
please
the
only
thing
that
I
would
consider
a
major
change,
and
it's
not
even
really,
that
is
actually
a
typo
fix
in
a
normative
requirement.
G
The
required
application
media-type
for
the
coast
wit,
was
originally
Co,
Swift
plus
c4,
which
was
a
mistake
and
that
should
have
been
application.
Swit
plus
c
bore
it's
taken
right
out
of
the
coast,
would
draft
so
that
typo
was
fixed.
That
mistake
is
fixed
and
now
we're
in
alignment
with
the
coast
would
draft
and
everything
is
good
to
go
and
was
always
meant
to
reference
that
in
the
first
place
and
then
just
a
lot
of
minor
fixes
some
language
changes.
G
G
B
Okay
I
mean
based
on
the
mailing
list.
Traffic
I
didn't
see
any
opposition
to
the
dress.
I,
don't
does
anybody
here?
Have
any
further
comments?
I
based
on
the
working
group
last
call
nope,
okay,
I!
Guess
then,
with
that
we
will
exit,
let
working
group
last
call
and
progress.
It
onward
and
upward
excellent.
G
L
M
L
L
The
Coase
would
draft
is
based
on
similar
constructs
that
exist
in
iso/iec
1
9
7
7
0-2,
which
is
called
a
software
identification
tag.
So
one
way
of
thinking
about
Coast
wave
is
it's
the
maybe
younger
brother
of
house
with
tags,
so
we've
been
working
on
the
draft
for
a
while.
It's
been
in
working
group
last
call.
We
received
quite
a
few
comments
during
the
working
group.
Last
call
so
really
grateful
to
all
of
the
reviewers.
They
think
we
had
something
like
nine
nine
reviews.
L
B
L
L
L
We
had
a
text
about
the
use
of
utf-8
and
Coast
wit,
which
was
already
implied
because
when
you
see
Lauren
see
board
uses
the
utf-8,
but
we're
explaining
that
more
explicitly.
Now
the
one
normative
change
I
believe
that
we
made
as
far
as
the
model
was
adjusted
the
tag
ID
to
allow
for
a
you,
you
ID,
to
provide
it
as
a
16
that
should
be
a
16,
byte,
B
string
and
that
was
based
on
Karstens
feedback.
L
L
I
did
I
did
move
the
private
use
to
negative
integer
values,
which
is
more
consistent
with
some
other
see,
board
based
drafts,
which
also
then
makes
the
the
standards
track
and
specification
required
range
is
slightly
larger,
which
I
think
is
also
a
benefit
that
didn't
actually
change
any
of
the
initial
registrations.
So
it's
really
not
a
big
normative
change.
L
I
had
some
discussion
around
resolving
overlapping
value
spaces
for
version
schemes,
because
I
also
added
some
expert
review
guidelines
that
require
that
version
schemes
don't
have
overlapping
namespaces
and
some
of
the
initial
ones
do,
but
we're
stuck
with
those
because
we're
using
the
initial
values
that
come
from
sweet
sweet
tags.
Does
they
in
sync?
L
L
G
L
And
if
you
want
just
line
feeds
alone,
you
can't
get
that
based
on
the
current
definition,
but
that
didn't
seem
like
something
that
was
worth
actually
holding
up
the
publication
of
coastward
around,
because
you
know
the
the
line
endings
that
are
defined
in
the
current
RFC
seem
to
be.
You
know
enough
for
for
the
kinds
of
use
cases
that
we're
trying
to
support.
N
Carson
would
take
care
of
that
that
we
cannot
put
it
in
first
person
attract
signal
or
you
don't
know
how
this
evolves
reference
and
it's
not
a
real
cough
follow-up
work.
You
could
make
an
errata
later
and
say
we
can
make
it
better
with
this
unique
version,
but
I
think
this
is
not
yeah
I
think
you
have
to
be
from
sort
of
all
right
now.
Yeah.
L
L
L
B
But
you
are
out
of
cycle.
Actually
what
we
had
just
discussed
was
because
the
the
draft,
despite
it
didn't
despite
the
fact
that
it
did
not
have
a
lot
of
a
lot
of
substantive
changes.
We
thought
we
would
give
like
a
two-week
period
for
Pete
for
people
to
have
a
chance
to
review
the
final
draft
before
we
forward
it.
As
far
as
we're
concerned,
we
have
consensus
of
the
working
group
to
forward
it,
and
people
have
one
last
chance
to
get
all
of
the
last
typos
fixed
set
work
for
everybody.
I
think
we've
got.
B
D
B
L
O
Yeah
I
should
I,
don't
have
to
be
a
doctor
for
mine,
okay,
so
at
I
know,
I
think
a
few
of
you.
It
may
be
not
I
think
was
it
just
Chris
and
I
at
the
Charis
workshop,
so
we
went
through
some
brainstorming
exercises
and
for
one
section
we
talked
about
adoption
and
lack
of
adoption
on
protocols
just
to
kind
of
get
a
feel
of
what
helps
and
in
thinking
about
this
work,
I
really
want
to
see
it
get
adopted
right
and
so
one
of
the
I
guess.
O
The
work
done
didn't
match
what
people
wanted
right
and
so
I
want
to
make
sure
and
I
know
it's
kind
of
late
to
do
that
that
we
kind
of
reassess-
and
this
is
after
I-
took
a
look
at
a
few
documents
recently
and
also
you
know,
seeing
that
we
might
be
splitting
off
separately
from
what
s
cap
is
doing.
Is
that
the
right
decision
you
know
who's
who's
listening
to
the
users
most
so
with
sticks?
O
People
aren't
as
interested
is
the
discussion
from
Kara.
So
maybe
it's
different
from
reality
for
other
people
they're
more
interested
in
using
it
as
a
format
to
manage
their
content
than
they
are
for
exchanges
and
IO
def.
It
described
more
than
what
people
wanted
to
exchange
so
that
might
have
been
one
of
the
hindrances
to
adoption.
So
in
thinking
about
this,
I
mean
I.
Think
it's
pretty
clear.
People
want
to
a
their
system,
integrity.
O
They
want
to
get
the
data
in
on
their
systems,
be
able
to
map
map
it
back
to
expected
policies
and
have
secure
systems.
So,
and
then
we
have
different
methods
of
exchange
that
we've
discussed.
We've
discussed,
XMPP
grid
we've
discussed
Rolly
I
personally
can
see
the
case
for
exchange
within
an
enterprise.
I
could
see,
you
know
out
to
a
managed
service,
that's
doing
assessments
and
you
may
have
different
exposures
than
I
do
and
so
I
kind
of
wanted
to
just
take
that
step
back
for
a
minute
and
see
you
know,
maybe
it's
customer
driven.
O
O
Anyone
anyone
yeah,
I
and,
and
then
I'm,
even
wondering
like
there's
some
brainstorming
exercises.
We
did
a
Karis,
it
might
be
worth
doing.
Some
of
those
I
might
do
have
tools
that
we
could
do
it
online,
where
we
could
collaborate
and
just
kind
of
reassess
just
to
see
if
we
could
maybe
we're
on
the
right
track
for
adoption.
Maybe
we
need
to
reassess
and
make
sure
we're
doing
the
right
thing
towards
adoption.
P
So
from
CIS
experiences,
he
is
right:
yeah
we
develop
best
practices,
controls
and
then
we
have
all
the
benchmarks
and
everything
they
offered
a
variety
different
tools:
software
right
that
helped
people
actually
implement
that
stuff.
A
lot
of
the
feedback
we
get
or
questions
I
should
say
that
we
get
from
users
of
the
CIS
controls
is
wow.
That's
a
lot
of
tools
to
use.
How
do
I
get
them
to
work
together?
P
Okay,
quickly,
right
right,
what
they
want
is
their
interoperability
out
of
the
box
says:
that's
are
expensive
so
and
and
what
comes
to
mind
is
a
blog
post
from
where
the
product
managers
at
tripwire,
okay,
he
recently
posted
I,
think
it
was
just
earlier
this
month
house.
They
have
this
idea
where
they
actually
want
to
create
a
community
of
open-source
tools
that
just
work
together
and
to
me
it's
kind
of
like
okay.
Yes,
let's
do
that,
but
let's
let
them
work
on
a
common
infrastructure
based.
P
P
O
O
Q
The
more
complex
a
standard
is
the
less
likely
we
as
vendors
are
going
to
implement
unless
we
have
extreme
compelling
reasons
from
customer
demand,
so
it's
usually
easier
when
they're
simple
and
easy
and
they're
iterative
and
we
add
to
them-
and
we
can
punctuality
and
feature,
but
if
I
give
this
to
product
management
and
then
to
architects
to
look
at
and
they
come
back
and
tell
me:
oh
that's
9
months
or
12
months
before
we
can
implement.
It's
not
gonna
happen.
Just
let
me
finish.
Q
Okay
and
if
things
come
back
and
they
say
oh
yeah,
I
can
do
some
of
this
over
the
weekend.
You
know
it
may
not
take
over
the
weekend,
but
it
does
get
done
so
I've
ranted
about
this
before
in
standards
not
just
here,
but
in
lots
of
other
places
that
there's
a
PR
marketing
problem
that
we
all
suffer
from,
and
it
really
takes
a
lot
of
work
to
get
the
market
to
understand
it.
To
get.
Q
You
know
the
various
communities
to
know
what
it
is,
how
it
works,
why
they
should
be
asking
their
vendors
for
it
and
boiling
it
down
into
bite-sized
chunks
that
people
can
grok,
and
so
I
was
joking
earlier,
like
who's
going
to
start,
writing
like
that
to
get
started
guide
for
some
of
these
things
like
they
help
people
and
I
know,
we've
done
some
of
it.
You've
done
a
lot
of
excellent
work
with
this,
but
a
Stanford.
You
know
design
school
exercise
with
post-it
notes.
My.
P
So
again,
as
a
vendor,
then
right
so
respect
if
I
have
I
think
is
that
none
of
this
work
here
the
code,
components
that
that
you
would
implement
as
a
vendor
right
would
be
considered
your
secret
sauce,
your
core
stuff,
that
you
need
to
get
done
right,
I,
think
that
would
be
true
for
any
better.
So
the
question
I
have
is,
if
there
were
collaborative
leak
about
open
software,
that
you
could
integrate
with
your
tools
so
that
you.
O
Q
Q
I
won't
speak
on
sticks,
but
just
purely
as
the
vendor
side
of
it.
It
there's
things
like
this.
That
can
help
it.
You
also
have
to
distill
these
hundreds
of
pages
of
documents
into
things
that
product
management
and
architects
can
understand.
Because
quite
house
is
there
nothing.
So
we
have
to
delete
it
once
there's
a
commitment
by
the
business
unit
to
implement
and
so
I
like
along
these
technologies.
That's
why
I
come
that's.
Why
I
try
to
support
them?
I
try
to
meet
as
much
of
it
as
I
can.
O
Q
N
This
is
hang
and
talking
about
this
never-ending
cycle
of
what
are
we
actually
going
to
do
that
customer
need
this,
and
is
it
opera
might
be
very
happy
about
the
results
that
we
presented,
build
from
the
hectare
from
kasnia
actually
were
sick
and
tired
of
talking
about
stuff.
We
started
implementing
for
three
hectic
plans
now,
and
we
know
tangible,
really
good
things
to
show
that
amazingly
there's
light
at
the
end
of
the
Sun.
Oh
I
think
this
might
push
some
things
that
you
would
actually
wanted.
Okay,.
L
We
started
this
process
what
seven
years
ago
and
and
when
we
originally
were
chartering,
you
know
I
showed
up
as
long
as
condition
to
a
bunch
of
other
books
from
the
SKF
community,
because
we
wanted
to
work
on
sort
of
the
next
versions
of
some
of
the
core
s
kept
specifications,
things
like
X,
CCDF
and
opal
and
and
those
kinds
of
things.
Then
we
we
ran
into
some
problems,
organizing
that
work.
P
L
P
L
O
L
D
O
B
B
H
D
H
Okay,
so
the
hackathon
so
Hank
approves
of
the
hackathon,
which
is
good.
Everybody
wants
that
again.
I'll
mention
karl-heinz
a
couple
times
on
here,
but
he
joined
the
hackathon
remotely
from
Germany
and
throughout
the
weekend
and
provided
a
lot
of
help.
So
so
we
we
had
a
number
of
objectives
going
in
kind
of
wanted
to
start
talking
a
little
bit
about
how
to
represent
what's
information,
to
collect
from
an
import
and
from
an
endpoint
how
to
represent
what
was
collected
kind
of
dovetailing
a
little
bit
on
some
of
the
s
cap.
H
Stuff
I'll,
probably
talk
a
little
bit
about
that
on
other
slides,
but
I
used
some
adaptations
of
some
s
cap
content.
In
there
we
wanted
to
implement
a
simple
collector
actually
to
interpret
that
that
what
to
collect
information
and
actually
get
it
from
that
endpoint.
Thankfully,
I
work
on
a
tool
that
does
stuff
like
that.
H
So
we
wanted
to
do
some
cool
things
with
XMPP
as
transport,
we
wanted
to
I
tried
to
get
the
ability
to
use
a
number
of
different
XMPP
features
to
actually
shuttle
the
data
around
and
cause
good
things
to
happen
so
that
I
q--
stanzas.
The
information
query
allows
a
request
and
response
sort
of
model
to
tell
and
then
a
collector
to
actually
grab
in
from
from
the
endpoint
and
send
it
back
to
the
entity
that
requested
it.
There's
some
other
entities,
other
stanzas
out
there
and
other
functionality.
H
We
use
publish/subscribe
capabilities
of
XMPP
to
do
that
as
well.
To
move
that
information
around
and
actually
ended
up
being
able
to
shuttle
it
around
using
just
basic
XMPP
messages.
So
just
as
I
would
send
a
chat
message
to
somebody.
I
can
send
the
payload
to
to
tell
that
entity
to
collect
and
not
really
care.
H
H
Currently,
the
the
current
ovale
data
models
actually
a
tightly
coupled
collection
and
evaluation
into
a
very
large
document,
centric
kind
of
format,
so
you
kind
of
have
to
send
all
of
the
things
that
you
want
to
test
and
evaluate
and
collect,
and
all
that
kind
of
stuff
in
one
big
format,
I
actually
extracted
out
all
of
the
collection
piece
to
make
it
its
own
sort
of
separate
namespace.
So
all
you
have
to
actually
pass,
as
are
the
what-what.
H
Ovale
calls
objects
which
really
allows
you
to
just
indicate
what
to
collect,
there's
some
other
things
in
there,
but
we
didn't
get
into
the
more
comp
educated
things
that
you
can
do
with
oval
like
sets
and
filters
and
variables,
and
things
like
that.
They're
they're
modeled
in
there,
but
we
didn't
test
any
of
them.
H
H
There
are
some
cool
XML
that
shows
what
that
is.
There's
a
family
object
which
is
pretty
gives
the
operating
system
family
and
we
did
some
environment
variables.
Those
were
they.
Those
were
basically
the
two
objects
we
tested
with
the
system
characteristics
that
would
come
back.
If
anybody
wants
to
know
what
my
IP
address
was.
Okay,
that's
not
on
there,
but
the
computer
name
environment
variable
on
my
machine
is
called
Cisco
to
have
it's
just
in
case.
Anybody
wants
to
know
that.
H
So
again
we
were
joined
by
Carl
Heine's.
He
stayed
up
nicely
till
to
three
o'clock
in
the
morning
on
Saturday
in
in
Germany
two
to
keep
working
on
stuff.
He
he
is
the
expert
on
the
concise
map
stuff
and
a
Java
guy
as
well.
So
we
had
a
good
time
conversing
through
Hanks
sort
of
video
chatting
capabilities
and
were
able
to.
If
anybody
saw
me
with
headphones
on
the
entire
time.
I
was
talking
at
Caroline's,
so
we
created
the
XMPP
extensions
to
handle
those
collection,
requests
actually
perform.
H
The
collection
generate
the
system
characteristics
and
actually
shuttled
that
data
back
via
XMPP.
Again
we
did
it
through
standard
request
response
capabilities
in
the
XMPP
iqs
stanzas.
We
did
it
through
publish/subscribe.
We
did
it
right
through
direct
XMP
feed
messages
as
well.
On
karl-heinz
his
end,
he
received
the
collected
system
characteristics
via
an
XMP,
an
XMPP
adapter
that
he
built,
so
it
basically
just
listens
for
the
messages
and
is
able
to
handle
the
payloads
that
we
sent
back
translated.
H
So
we
kind
of
had
the
previous
slide
kind
of
showed
how
the
the
work
was.
Divvied
up,
I
was
kind
of
basically
responsible
for
the
XMPP
infrastructure,
the
Oval
related
stuff
in
the
actual
collection
which
looks
the
pub/sub
looks
like
this,
so
the
orchestrator
would
request
the
information
to
be
collected.
The
collector
would
collect
it
and
send
it
back,
and
then
the
orchestrator
would
publish
that
to
the
to
the
pub
subtopic
and
then
the
the
german
side
of
the
of
the
equation
was
the
XMPP
adapter
and
translating
and
saving
all
that
stuff
into
that.
H
H
The
direct
XMPP
message
again
sort
of
does
not
include
the
the
response.
Oh
I
didn't
update
all
of
that
slide
so
that
that
IQ
response
containing
ovale
system
characteristics
should
just
be
that
should
be
deleted
from
that
slide.
Sorry,
and
it
just
has
the
the
collector
sending
sending
the
message
stanza
with
the
with
the
system
characteristics.
H
This
is
what
the
concise
map
model
look
like.
So
it's
a
series
of
identifiers
and
and
metadata
that's
sort
of
links,
those
identifiers,
so
the
I
guess
I
can't
really
walk
over
there,
but
the
system
identifier.
There
is
one
of
the
identifiers,
the
the
collection
identifier,
which
is
the
one
on
the
far
right
hand,
corner
identifies
the
actual
active
collection,
the
different
identifiers
for
the
items
that
are
right
next
to
the
collection,
one
identify
the
basically
the
family
item
in
the
environment,
variable
that
we
collected
again.
H
So
we
we
actually
learned
a
ton
in
this.
We
we
were
able
to
again
sort
of
use,
all
of
XMP
peas,
capabilities
to
both
to
published,
subscribes
and
the
request
response
and
send
a
direct
message.
Stanza
containing
all
that
payload
to
actually
trigger
the
collection
to
receive
the
system
characteristics
send
them
over
to
the
other
side,
we
were
able
to
again
translate
that
map
data
or
translate
the
system
characteristics
into
the
map.
H
So
my
collector
was
actually
able
to
you
know
it's
kind
of
it's
fully
sort
of
s,
cap
compliance,
so
it
can
collect
a
number
of
different
platform,
specific
items,
but
the
the
concise
map
implementation
on
the
other
side
could
really
only
handle
the
family
and
environment
variable
items
that
we
decided
on
before
the
hacks.
The
one.
H
Were
able
to
really
kind
of
help
define
some
of
the
operations
that
and
interactions
between
components,
so
we
want
to
start
moving
them
to
get
them
included
in
as
part
of
the
architecture.
A
lot
of
the
information
from
the
various
data
models
can
be
used
and
then,
hopefully,
to
again
sort
of
start
defining
that
core
set
of
operations
that
core
set
of
interactions
between
the
different
components.
H
H
Don't
necessarily
have
a
favorite
place
where
that
that
that
migration
could
take
place,
I'm
happy
to
continue
to
flesh
out
the
the
oval
collection
models
and
that
decoupling
of
collection
and
evaluation
and
take
it
to
the
S
cap,
2.0
working
groups,
some
of
the
endpoint
data
collection
and
and
oval
authoring.
You
know
groups
to
to
see
how
that
flies
again.
H
Yeah
last
big
thing
again
for
karl-heinz.
Hopefully,
some
folks
named
Hank
will
take
those
messages
back
of
thanks
to
him
and
again
for
staying
up
really
late.
Hank
I'll
say
was
also
pretty
instrumental
in
enabling
some
of
our
collaboration
and
getting
going
to
really
get
this
hackathon
off
the
ground.
So
I
really
appreciate
that.
So
thanks
to
Hank
and
I
think
that's
it.
G
G
N
So
this
is
the
thing
thanks
for
the
excellent
overview
and
I
think
this
is
a.
This
is
repeating
too
a
lot
of
the
things
we
were
storming
at.
They
mentioned
operations
like
the
extra
things
we
need
the
architecture
to
define
how
to
interact
with
other
components.
We
have
to
start
leading
different
data
models
that
we
could
show
interoperability.
Okay,
two
items
on
the
city
data
side
yet,
but
it
is
a
proof
of
concept.
So,
if
you
can,
we
can
all
it's
baby.
The
concept
is
dumb
system.
N
I'll
know,
there's
a
select
work
of
creating
be
missing:
CD
dl-44,
the
symmetric
metric
and
connection
works
like
common
areas
and
recipe
from
XML
to
CD
d
ln
back,
and
there
was
no
loss
all
the
time.
So
basically,
we
have
its
operability
plugfest
achieving
right
now,
and
that
is
I
think
the
a
very
important
thing
here.
N
But
there
is
actually
something
coming
out
of
this,
only
with
respect
to
code
and
yeah,
it's
plugging
and
working,
but
we
have
found
the
basis
to
create
the
actual
data
model
that
it
will
be
able
to
view
multiple
protocols
and
that
we
can
now
understand
better
what
orchestration
really
looks
like
and
if
we
do
this
again
and
then
maybe
again,
probably
aware
that
going
to
be
compete
because,
let's
stop
working
at
this
between.
Yes,
this
will
be
ongoing
work
and
therefore
we
will
not
rely
on
the
echo
fonts
alone.
Thank.
H
You
I'd
say
that
might
also
be
worth
exploring
just
from
my
perspective
and
not
knowing
necessarily
T
too
much
about
how
the
the
CDL
would
be
built
to
maybe
see
if
it's
there's
a
more
generic
way
to
represent
it.
So
we
don't
have
to
keep
expanding
it
and
we
can
represent
the
oval
data,
for
example,
in
some
you
know
something
that
doesn't
necessarily
always
need
a
lot
of
modification.
You.
N
Know
if
you
don't
have
a
CDA
parlors
of
the
code
has
to
represent
the
specification.
All
the
boils
down
that
we
needed
a
CD
deposit
that
then
creates
classes.
You
can
instantiate
for
the
code
that
is
basically
right,
yeah
yeah.
We
have
that
on
multiple
fronts.
We
could
have
fun
see,
unfortunately,
not
in
Java
first,
but
it
is
almost
sweet.
I.
P
Hey
how
much
time
do
I
have
as
fast
as
I
can?
Okay,
you
got
it
cool,
sacrum
architecture.
This
actually
probably
won't
take
very
long.
There's
some
pending
changes
to
the
draft
that
bill
and
I
are
making.
We
want
to
remove
the
XMP
base,
xmpp
based
solution
from
the
draft,
pretty
much
entirely
right.
Now,
in
the
unpublished
like
in
the
github,
my
we
moved
it
to
an
appendix
depending
on
what
people
say
it
belongs,
or
it
doesn't.
If
nobody
has
an
opinion
will
yank
it.
P
It's
supposed
to
also
mean
you
know
anything
that
the
enterprise
has
to
operate,
whether
they
actually
own
that
infrastructure
or
if
it's
shared
infrastructure
or
what-have-you.
So
it
was
one
clarification.
We're
gonna
make
another
clarification.
This
one's,
probably
more
important,
is
talking
a
little
more
about
capability
versus
interface
versus
operations
before
I
get
into
those
three
things.
P
Essentially,
for
an
interface,
we
want
to
be
able
to
say
configuration
assessment
of
Windows
10
using
oval
5.11.
That's
an
interface
that
basically
implements
that
to
some
degree
right.
Does
that
make
sense,
we
want
to
be
able
to
specify
interfaces
at
that
level
eventually
and
right,
probably
in
the
solution
level,
and
maybe
not
the
high
level
architecture
operations
I
want
to
gather
system
characteristics
from
Windows
10,
basically
you're
telling
something
go
collect
this
information
right.
That's
essentially
an
operation
we'll
put
some
clarification
in
the
draft
around
these
things
and
people
can
comment
and
complain.
P
Maybe
I
don't
know
the
self
architecture
is.
We
did
want
to
talk
about
that,
a
little
bit
too
just
to
make
it
clear.
It's
it.
The
overall
sacrum
architecture,
kind
of
does
have
this
dual
nature
in
terms
of
sub
architectures,
that
there's
collection
subsystems
right
so
like
the
EPC
e
vcp
right,
EPC,
P,
draft
ACP,
formerly
easy
feature
is
a
collection
of
sub
architecture,
and
this
sack
of
architecture
supports
that.
It
also
supports
the
idea
of
gang
push
or
using
an
oval
based
system
that
doesn't
go
through
EPC
P
right.
P
It's
it's
intended
to
support
all
of
these
things.
So
we
kind
of
call
those
the
collection,
architectures,
collector
subsystems
and
then
above
that,
so
to
speak,
is
more
of
the
cooperative
tooling,
which
is
what
we
talked
about
at
the
beginning
when
Kathleen
I
was
talking
about.
You
know
what
we're
doing
here,
basically
right,
so
the
the
co-operative
tooling
architecture
is
essentially,
how
did
all
these
different
components
in
a
standard
standardize
away
start
to
talk
to
each
other
without
a
lot
of
operator
intervention,
if
any,
if
as
possible?
P
Yes,
we
want
to
clarify
components.
So
I
mentioned
we're
initially
focusing
on
configure
configuration
management.
We
will
propose
specific
components
and
we'll
provide
some
descriptive
details
for
each
of
those
components
into
have
their
capabilities,
the
interactions
they
have
between
each
other
and
even
the
different
types
of
interaction.
So
in
the
hackathon,
for
example,
they
had
pub/sub,
they
used
IQ,
which
is
I,
believe
query
response,
and
then
they
had
message,
which
is
just
you
send
it,
and
unless
you
hear
something
bad,
you
assume
it
went
well
right.
P
P
There
are
two
open
questions
in
the
draft.
I
should
workflows
be
documented
in
the
draft
or
in
separate
drafts.
We're
gonna
start
in
the
draft
and
if
we
need
to
we'll
move
it
out
a
same
thing
for
interfaces
or
interactions.
Basically,
we'll
start
in
the
draft
and
move
them
out
when
it's
necessary,
so
basically
the
more
detailed
it
becomes
I
think
then
it
becomes
more
of
a
solution
and
once
it
crosses
that
line
or
ever
there'll
I
might
be.
We
can
move
that
into
a
sub
draft.
N
P
N
P
R
R
R
Sorry
so
now
more
seriously,
you
you
have
a
slider
about
some
kind
of
expressions
see
you
have
a
slide
with
some
kind
of
expressions.
Yes,
this
one,
so
I
am
new
at
completely
new
here,
so
I'm
really
completely
naive.
Are
you
referring
to
a
very
formal
expression
on
which
year
that
should
be
at
all
no.
P
R
Actually,
my
point
is
that
it's
not
pedantic,
it's
really
important
I,
think
I
said
I,
because
I
am
chairman
and
I
knew
as
g17,
and
we
recently
had
a
person
who
came
is
a
new
work
item
for
a
universal
syntax
for
secret
expression.
Then
that
was
quite
remarkable.
We
start
to
work
on
that
and
an
associate
is
he's
a
lot
of
benefit.
I
think.
Actually
he
started
to
present
to
21
cos.
He
and
others
to
see
is
organization
as
well.
So
maybe
there
are
dots
to
connect.
R
Really
mattered,
because
it's
only
when
you
stop
to
go
at
that
letter
that
you
can
add
many
other
benefits
in
terms
of
you
know
when
you
have
to
do
your
audit
is
your
stink,
because
we
can
really
make
this
professional,
so
I
really
appreciate
the
link
as
well.
Knowing
that
is
so
network
certainly
be
interesting
to
that
mix.
Happy
to.
Perhaps
if
the
new
item
gets
approved,
maybe
miss
enemies
of
statement
with.
Is
it
a
weight
that
you
guys
work
here
or
sorry
like.
R
Okay
right.
So
that's
that's!
That's
the
really
interesting
point
here.
The
second
piece
related
to
does
anything
we
just
add.
So
it's
just
a
remark
to
myself
that
now
I
understand
that
really
I
need
to
put
on
my
work
program
to
understand
better
what
you
guys
are
doing
here,
because
now
I
start
to
feel
the
ring
possible
between
your
work
and
endpoint
and
and
and
I
really
an
intuition.
There
is
more
wish
I
mean
there
is
motive
locally
proposed,
but
let
me
know
what
you
guys
are.
B
B
Updated
milestones,
so
the
only
milestone
we
actually
have
associated
with
this
on
our
milestones
right
now
is
to
submit
an
initial
draft
which
we
have
met.
Do
we
have
any
sense
of
when
we'd
be
ready
to
to
how
close
do
we?
How
many
III
mean?
No,
the
only
reason
if
you'll
recall
the
reason
why
the
only
milestone
we
have
is
because
we.
P
Where
what
what
the
intent
of
the
draft
is
is
to
specify
I,
don't,
okay,
it's
describing
an
abstract
architecture
that
is
intended
to
provide
the
messaging
infrastructure
necessary
for
sending
some
payload
right
at
this
point,
I
care
very
little
about
what
that
payload
is
I'm.
Just
saying,
okay
I
mean
so
it's
so
like.
When
you
look
at
the
the
XMPP,
the
actual
solution
right,
we
shoved
ovale
through
there,
then
you
know
CD
deal
whatever.
It
was
right
and
it
came
back
out
that
could
easily
be
io
def.
P
Necessary
things
for
like
even
configuration
management
right
so
that
if
you're,
an
Orchestrator
and
you're
telling
a
collector
I
want
you
to
collect
something
you're
saying
you're
telling
it
collect
this
piece
of
information.
That's
one
way
to
do
it.
You
could
also
say:
watch
this
information
from
the
architecture
perspective
at
this
level,
I,
don't
care
which
one
they're
saying
that's
inside
the
payload.
They
understand
that
by
capabilities
right,
so
we're
trying
to
say
how
do
they
each
one
of
each
component,
one
of
the
components?
How
do
they
express
their
capabilities
to
each
other?
N
B
I
I
I
B
K
F
F
This
is
there's
one
interesting
thing
in
here,
but
I'll
come
back
to
it.
Well,
I'll
come
back
to
it
right
now.
So
what's
in
the
draft
right
now
is
only
a
handful
of
elements.
It's
really
aimed
at
kind
of
what
Adam
and
Hank
just
agreed
to,
which
is.
How
do
you
enable
kind
of
saying
here's
an
endpoint?
F
F
I'm
gonna
keep
doing
this
from
the
chairs
desk
for
a
second
Wow.
Sorry,
yeah,
so
I
mean
there's
one
vote
to
keep
it
it's
that
would
really
be
about
future
proofing.
This,
but
I
could
imagine
its
usefulness,
especially
if
you
you
know,
have
any
kind
of
version
where
you
want
to
do
a
compound
version
of
a
host,
but
I
haven't
done
anything
like
that.
F
Yet
so
another
idea,
that's
in
there
is
the
idea
of
opaque
an
opaque
information
field,
and
so
this
is
really
kind
of
the
basic
description
and
so
I
have
communicated
with
Bill
some
I
needed
to
kind
of
finish
this,
but
which
is
right
now.
This
is
basically
the
thing
that
would
have
an
envelope
of
some
other
type
of
data
in
the
in
the
information.
F
So
if,
if
I
allow
any
kind
of
opaque
element
and
let
you
to
transfer
anything,
as
you
know,
just
kind
of
a
blob
within
in
the
structure,
then
there's
very
little
motivation
to
standardize
more
elements
on
the
flip
side
to
kind
of
Brett's
comment
from
before
this
potentially
lowers
the
bar
very
much
to
doing
any
kind
of
implementation,
because
there's
only
a
handful
of
fields
you
need
to
put
in
and
then
you
can
put
in
your
blob
as
long
as
the
other
person
knows
what
the
blob
is.
Hank.
N
F
You
know
some
kind
of
version
of
how
you
do
those
negotiations
I
envision
some
kind
of
enumerations
that
would
be
I
in
a
register
of
you
know
the
types
of
data
formats
that
could
be
exchanged
in
those
ok
cap
envelopes.
I,
don't
think
that's
a
complete
list
of
commands.
I
don't
want
I'm,
incapable
of
getting
too
far
ahead
of
of
say
bill
in
and
karl-heinz
and
anybody
else
who
wants
to
participate.
So
I
will
by
necessity,
lag
them
somewhat.
Hopefully,
by
not
too
much.
H
F
Know
that
that
is
absolutely
the
the
goal
and
and
I'm
tracking
what
you
guys
are
doing
as
to
what
I
put
in
here
so
I
obviously
do
want
the
information
model
to
align
and
I'm
paying
attention
good.
For
me,
what
I
am
a
chair
as
it
turns
out
yeah,
which
means
you
know:
I,
don't
have
to
beat
myself,
although
maybe
I
do
to
read
everything.
F
This
is
the
only
place
where
I'm
using
a
list
currently
so
there's,
at
least
in
the
current
version
of
the
information
model.
There's
one
thing:
that's
the
command
and
then
there's
an
Associated
list
which
could
be
the
parameters
in
the
return
types
I
don't
know.
If
that's
the
same
way
to
do
it,
I
don't
know
if
that's
great
in
the
information
model,
but
that's
what
it
is
right
now
and
I'm
happy
to
accept
people's
better
wisdom.
So.
F
L
F
F
Yeah
and
that
was
kind
of
it,
and
so
this
is
the
kind
of
feedback
that
I'd
like
to
get
from
you
guys
and
obviously
I
mean
I,
know
bill
on
Adam
and
I
have
talked
and
I
will
continue
to
talk
and
follow
what
you
guys
are
putting
in
there.
But
that's
the
that's
the
gist
of
where
this
is
heading
and
I'm
happy
to
have
people
help
that.
N
Which
is
because
vibha
really
was
cold
and
have
proof
of
concept,
and
it's
actually
by
over
then
again
I
see
your
start
with
addresses
first
and
always
I'm
a
big
fan
of
lecture
on
amis.
So
maybe
have
the
categories
again
and
saying
this
is
addresses
intended
to
be
used
as
identifiers
isn't:
yada
yada,
yada
baby.
So.
N
S
P
M
D
B
B
B
B
B
B
G
P
H
H
G
So
let
me
say
that,
because
I
think
that
if
I
were
to
get
the
bulk
of
that
text
down
and
then
maybe
come
and
talk
to
you
guys
instead
for
just
that
technical
filling
in
those
technical
fields
that
detail
like
once
I
say:
oh
we're,
gonna
put
X
CCDF
in
it,
come
and
talk
to
you
and
say
well
what
fields
or
what
things
inside
of
X
EDF
would
be
important
to
take
out.
I
think
that's!
G
C
K
B
G
L
L
P
C
L
B
K
S
N
Push
is
done.
The
only
thing
I'm
making
this
faces.
I
didn't
know
that
we
planned
for
that
I
totally
forgot
about
it.
Was
that
so
so
little
me
yes,
it's
a
valid
producer
of
information
and
I
assume
that
we
can
revive
that
when
we
are
at
the
place
in
the
architecture
that
has
the
appropriate
set
of
operations
to
initiate
that
so
I
would
assume
after
the
next
meeting.
We
can
start
working
on
that
effectively
before
that.
It
is
just
wishful
thinking.
So.
B
N
B
N
F
B
C
F
B
N
It's
actually
if
it
sees
the
current
scope,
we
are
pretty
fine
and
you
just
have
to
revive
it
when
the
time
is
right
and
the
architecture
has
a
place
to
fit
with
the
characterization
profile.
Yet
above
you're,
starting
on
that,
so
I
think
that
the
architecture
is
done.
We
should
be
able
to
complete
the
terminology
iteration
later
so
give
me
four
months
after
architecture.