►
From YouTube: IETF95-T2TRG-20160407-1620
Description
T2TRG meeting session at IETF95
2016/04/07 1620
A
B
B
So,
in
particular,
please
have
a
look
at
the
IP
on
guidelines
of
the
a
RTF
which
are
essentially
identical
to
those
of
the
IETF
there's
a
brush
each
going
around
with
nori
cloths.
We
have
a
job
or
facilitator
the
mailing
this
and
we
have
this
habit
of
putting
up
atop
repositories
for
our
meeting.
So
if
you
want
to
look
at
the.
B
A
B
So
we
are
not
going
to
really
do
work
here
at
this
meeting
in
this
hour,
but
of
course
we
are
very
interested
in
getting
feedback
from
people
who
have
not
being
part
of
your
process.
So
far,
so
I
will
talk
a
couple
of
minutes
about
the
status
of
the
research
group.
Then
re
will
talk
about
one
of
our
deliverables
design
document
mortgage.
We
talked
about
one
of
other
deliverables,
a
secure
boot
strapping
the
survey
document.
Then
maybe
we
can
talk
about
the
outreach
and
cooperation
with
other
organizations.
B
Okay.
So
let's
just
start
with
overview
over
the
research
group,
the
researcher
was
relatively
new,
so
I'm
going
to
say
what
it
is
about.
It
is
about
making
the
Internet
of
Things
a
reality,
Internet
of
Things
the
term
that
is
used
for
many
different
ideas,
not
many
of
which
actually
contain
an
Internet
and
what?
Why
is
Internet
technology
not
used
more
for
the
Internet
of
Things?
There
are
some
stumbling
blocks
and,
and
we
have
to
work
on
those.
B
This
is
really
the
idea
find
out
where
those
some
big
blocks
are
and
maybe
find
ways
of
hearing
and
in
particular
those
ways
that
might
lead
to
IETF
standardization
and,
of
course,
that
there
are
so
many
things
that
belong
to
the
Internet
of
Things.
For
instance,
people
belong
from
the
internet
of
things,
but
we
are
mostly
looking
at
the
layers
from
the
Edit
area
adaptation
layer
up
through
the
application
layer,
and
that
includes
management
and
security.
B
We
have
three
documents
that
we
have
designated
at
it
as
deliverables
right
now
that
are
in
progress
in
various
stages.
We
talked
about
two
of
these
today.
One
is
only
just
now
being
started
being
worked
on,
the
other
ones.
I
already
talked
about
rest
and
secure
boot,
strapping
survey
and
finally,
last
but
certainly
not
least,
what
we
have
done
is
outreach
to
other
organizations,
so,
for
instance,
in
Tennessee
we
have
met
with
Bluetooth
people
and
we
had
a
joint
meeting
with
ocf
the
open
connectivity
forum
and
foundation
I'm,
sorry
and
with
w3c.
B
So
we
could
start
to
understand
what
those
various
organizations
are
doing.
Maybe
a
quick
peek
into
what
we
will
be
doing
and,
of
course,
the
work
that
I
just
described
continues.
We
are
looking
at
having
meetings
at
real
academic
research
menus.
We
kind
of
have
neglected
this
in
the
first
year
because
there
was
so
much
cooperation
with
other
standards
organizations
that
was
necessary
to
be
able
to
do
something
in
this
space.
B
But
that's
certainly
on
the
list,
and
we
also
want
to
co-locate
with
open
source
activities
and
there
we
will
have
an
opportunity
in
a
balloon
right
in
front
of
the
next
ITF.
There
will
be
a
summer
riot,
community,
riotous
and
research
operating
system
for
the
inter
other
things.
They
have
a
very
vibrant,
open
source
community
and,
finally,
why
we
already
have
worked
on
benchmark
and
reference
scenario
see
we
plan
to
formalize
this
work
a
little
bit
more
and
maybe
have
a
fourth
deliver
document
by
the
end
of
the
year.
He
introduces
on
me.
B
B
D
However,
sometimes
use
is
innovative,
at
least,
and
sometimes
are
violating
even
many
of
the
basic
principles
still
in
the
RFC's.
So
the
idea
would
be
to
collect
some
of
these
guidance
across
various
sources,
RFC's,
etc.
Who
didn't
want
document
that
you
could
be
able
to
easily
to
give
someone
who
is
designing
his
systems,
designing
api's
design
services
and
they
could
at
least
get
the
basic
rights?
D
Also,
the
idea
is
to
collect
terminology
there.
So
we
have
a
common
language
of
discussing
discussing
these
things
and
have
really
key
piece
of
information
and
pointers
for
details.
So
this
government
is
either
to
be
something
short.
It's
really
focus
on.
Ilt
have
a
few
examples,
the
quick
and
easy
to
read,
and
then
you
have
a
lot
of
pointers,
wait
and
find
find
more
information.
D
So
where
are
we
know?
We
have
the
second
version
out
about
a
month
ago.
Now
we
got
a
lot
of
feedback
on
the
previous
person.
It
was
presented,
for
example,
in
the
web
of
things.
Group
I
was
represented
in
that
article
also
shared
with
some
other
organizations.
So
we
added
there
quite
a
bit
more
technology
was
there
were
some
terms
for
the
leasing
and
somewhere,
maybe
dr.
thought,
both
or
the
clear
and
in
particular,
to
move
issues
around
I'd
impotency,
an
application
state,
but
seemed
to
be
confusing
this
faith.
They
were
clarified.
D
D
A
D
Rest
consistency,
so
those
three
issues
were
make
perfect
case
the
current
person,
but
we
still
have
at
least
three
key
topics
that
you
want
to
discuss:
war
ii
straps
one
is
resource,
immediate
back
design.
So
how
do
you
go
about
designing
your
problem
in
our
kinetic
as
resources
as
media
types
of
interactions
and
also
hypermedia
different
applications?
So
how
do
you
think's
and
forties,
etc,
really
design
your
application?
D
So
I'm
putting
this
way
too
short
to
short,
slow
to
actually
discuss
it.
If
it's
open
issues
but
I
recommend
you
all
Rita
crap
reviews
and
comments
as
few
so
are
very
welcome,
and
also
we
would
like
to
hear
from
the
community
from
you
guys
and
also
from
organizations
that
you
are
working
outside
of
the
IEP.
So
a
lot
of
these
things
are
like
well.
E
Terminology
in
our
struggle
way
that
the
definition
of
rest
is
that
our
state
is
maintained
in
the
client
and
not
in
the
server,
but
in
some
I'll
team
partners,
highly
constrained
environments,
stays
became
the
server's
world.
The
fight
with
they
actually
minimum
data
transfer
so
run
stuff
is
basically
flying
the
face.
If
you
say
this
is
going
to
be
restful
and
Yuri.
Look
at
that
distressful!
It's
that
restful,
so
our
week
is
kind
of
fuzzing
the
words
around
here
or
we're
saying
that
these
systems
just
are
not
restful.
I.
G
On
this
topic,
yeah,
we
had
a
discussion
on
this
during
one
of
our
coats
anyway.
So
so,
usually
the
way
is
in
breast
yeah.
That's
resource
stated
on
the
server
side
that
can
also
be
modified,
and
it
also
is
part
of
the
overall
state
that
we
have
in
our
application.
So
pensive
I
had
a
problem
with
this
application
state
definition,
because
it's
not
only
what's
it
been
climbed.
If
you
have
people
who
don't
didn't
read
this
think
the
application
state
should
be
the
whole
thing.
G
So,
what's
on
the
time,
and
also
what
I
changed
on
the
server
so
sometimes
so
we
have
to
make
sure
should
we
stick
to
this,
but
it's
written
in
the
thesis
or
should
we
use
something
that
is
common
language
right
now
so,
and
there
was
a
lot
of
development
recently
and
we
haven't
gotten
enough
feedback
actually,
which
way
we
should
go.
So
apparently
you
recommend
to
go
with
what
people
or
the
language
people
use
right
now
and
use
this
right.
Okay,
I
see
a
not
your
I
think
this
is
good
input.
Yeah.
D
And
also
I
mean
I
recommend,
so
we
have
now
one
definition
of
the
whole
applications,
a
state
issue
in
the
draft,
but
that
is
really
we've
been
going
back
and
forth
on.
What
do
we
think
is
right
way
to
go
and
as
muddy
as
are
saying
more
more
impetus
more
than
welcome,
so
above
please
and
everyone
else
too.
Yeah.
H
Useful
to
define
what
you
mean
by
breast
and
I
think
the
comment
that
fortress
made
a
sort
of
reflected
that
at
that
India,
if
you
have
it
already
than
99
get
up.get
of
thing,
I'm
sort
of
put
a
basic
communication
are
like
everything
is
fine
and
everything
is,
is
a
pretty
uncontroversial
I
think.
But
then
there
are
these
other
aspect
that
we
stumbled
them
upon,
which
was
topics
like
you
trigger
an
action
and
the
result
doesn't
come
back
immediately.
Instead
of
things,
did
you
do
this
custom
in
the
document.
D
H
And
also
I
think
some
of
the
other
think
that
were
discussed
in
the
context
of
the
public
subscribe
that
Mike
Ross
has
been
working
on
we're
cast.
You
refer
them
as
the
Buddhist
still
like
that,
and
and
also
with
the
notion
of
device
sleeping
sleeping
devices
and
some
of
those
requirements
that
come
from
the
iod
word
that
are
not
necessarily
a
sort
of
it
waits,
not
necessarily
immediately
obvious
on
how
to
map
them
to
the
rest.
Design.
I
think
those
would
also
be
interesting
to
discuss
there
because
that's
where
things
get
interesting.
D
Right,
I
think
that's
a
valid
point
and
that
will
here
we
go
into
their
design
patterns
far
because
we're
thinking
on
writing
something
along
the
lines
of
okay.
This
is
your
problem,
and
quite
communicants
of
the
funny
result
will
be
a
better
kind
of
way
of
doing
that.
So
it
thinks
leaving
notes.
Would
it
in
that
matter?
B
I
think
I
would
summarize
this
as
the
objective
of
this
works.
Not
so
much
do
you
find
what
rest
is.
There
is
a
nice
PhD
thesis
you
can
read
about
that,
but
to
find
out
what
can
rest
bring
to
the
table
or
four
or
the
Internet
of
Things.
So
what
is
the
right
way
to
use
rest,
maybe
even
to
give
up
some
of
the
principles
if
you
find
out
they,
they
hurt
us
in
some
places
or
to
add
once
if
we
find
out,
we
need
so
rest
will
not
entirely
emerge
unchanged
from
from
this
exercise.
B
F
Good
afternoon,
I
move
it
and
I
will
be
presenting
our
draft
on
secure,
IOT
bootstrapping
and
to
give
you
some
context
in
the
Yokohama
meeting,
we
look
at
the
security
considerations
draft
and
we
have
this
general
feeling
that
I
OT
bootstrapping
is
kind
of
very
important
and
there
was
sufficient
interest
in
the
group.
So
we
decided
to
move
it
to
a
separate
document
and.
F
One
of
the
first
challenges
that
we
faced
is:
how
do
you
define
secure
boot
strapping?
It
means
a
lot
of
different
things
to
know
of
different
people,
and
this
is
the
definition
we
came
up
with
right
now.
So
it's
the
process
by
which
an
IOT
IOT
device
in
our
OT
network
security
becomes
operational
at
a
given
location
and
point
of
time.
So
so
the
goals
of
of
secure
boot
strapping
can
be
many
things.
It
can
be
authenticating
a
pre-existing
identity
or
or
creating
a
new
identity.
C
I
F
H
F
H
F
Centered
on
the
list
incorporate
so
going
back
to
the
goals
of
secure
boot
strapping.
It
could
be
for
a
author
ization
of
network
access.
It
could
be
for
creating
these
groups
or
domains
and,
of
course,
for
hearing
one
IOT
device
with
another
device
or
keyring
11.
I
would
securely
appearing
one
I
would
be
device
to
a
cloud
service
and
this
definition
is
kind
of
very
broad,
but
it's
on
a
purpose
because
I
OT
itself
is
so
broad.
F
So
we
try
to
categorize
these
methods
or
classify
this
method
so
that
it's
more
structured
and
the
first
is
managed
methods
where
the
general
assumption
is
that
you
have
reestablished
trust,
relation
and
authentication
credentials
and
managed
methods
can
be
centralized
or
federated.
To
give
you
an
example,
if
you
look
at
EAP,
if
you
use
EAP,
psk
or
TLS,
you
have
a
triple
is
over,
but
you
could
also
have
something
like
I
do
wrong
there.
F
So
the
next
category
we
identified
when
we
started
looking
at
these
bootstrapping
methods.
Was
this
p2p,
your
ad
hoc
methods,
and
here
the
assumption
is
that
devices
don't
have
any
clear,
stablish
credentials
and
you
have.
You
typically
do
an
authenticated.
If
you
have
an
exchange
and
to
prevent
the
man
in
the
middle
attack,
you
have
to
use
out-of-band
channel,
so
the
previous
example
of
peering
smartphones
over
bluetooth
or
using
Wi-Fi
Protected
setup.
F
If
you
were
in
socked
in
the
previous
session,
you
might
have
listen
to
the
talk
about
using
Oh,
be
with
EAP
moving
on
then
there's
the
opportunistic
or
leap
of
faith
methods,
and
here
the
point
is
not
so
much
as
as
as
the
initial
authentication
itself,
but
rather
to
check
that
the
same
initial
identity
is
used
throughout.
So
the
kind
of
the
continuity
of
the
identity
of
connection
and
if
some
methods
might
assume
that
the
attacker
is
not
present
at
the
initial
setup.
F
And,
of
course,
most
methods
actually
use
some
components
from
managed
and
ad-hoc.
So
that's
why
they
are
hybrid
and
categorization
is
not
always
easy
or
clear,
and
choice
of
bootstrapping
method
heavily
depends
on
the
business
case.
So,
as
honda's
pointed
out
in
the
sog
meeting,
of
course,
business
cases
governs
what
bootstrapping
methods
are
used
most
of
the
time.
So
what
kind
of
third
parties
are
available
and
who
wants
to
retain
the
control
or
avoid
the
work
done?
And,
of
course,
why
are
we
doing
this?
It's
to
learn
the
designers
options
and
trade-offs?
F
Our
goal
is
not
to
review
all
the
existing
bootstrapping
methods
and
no
prepare
100
page
document
that
no
one
will
ever
read,
and
instead
we
want
to
help
iod
application
developers
to
look
at
the
document
and
decide
what's
suitable
for
them,
and
end-of-life
andrew
bootstrapping
is
actually
a
very
complex
topic.
So
if
you
have
been
following
the
news
about
the
revolve
smart
hub
and
how
that
has
been
great
now,
it's
actually
quite
tricky
and
we
would
like
to
actually
have
a
discussion
on
on
that
in
the
document.
B
So
these
were
two
of
the
three
documents
that
we
are
working
on
right
now
and
I
want
to
quickly
talk
about
the
outreach
component.
I
already
said
that
we
did
a
little
bit
of
work
there.
So
what's
the
point,
the
point
is
that
we
actually
want
to
do
mutual
education
with
other
stos
in
in
this
other
standards,
developing
realizations
in
this
space.
So
we
want
to
educate
each
other
about
the
way
that
other
body
works.
B
So
we
want
to
understand
how
ocf
works
and
all
Ciara
wants
understand
how
we
work.
But
of
course,
we
also
want
to
know
what
what
are
the
interesting
products
they
have.
What
what
are
these
already
doing?
How
do
they
fit
in
the
into
the
overall
picture?
And
how
are
these
completing
the
picture
or
maybe
raising
additional
research
questions?
That's
really
the
reason
why
it's
a
good
idea
for
research
group
to
have
these
outreach
activities.
B
We
have
had
one
meeting
with
the
open
connectivity
foundation
now
right
there
and
we
probably
want
to
increase
the
involvement
and
that's
a
typical
example
for
a
problem.
We
have
there
because
they
kind
of
have
been
working
in
a
bunker
for
a
year
and
then
at
some
point
came
out
with
their
specification
1.0
and
then
finally,
we
could
find
out
what
they
were
by
doing
and
maybe
places
where
we
might
have
had
useful
info
input.
If
they
had
asked
us
earlier
down.
We
have
had
some
discussions
with
a
brutal
brutal,
of
course.
B
It's
an
interesting
connectivity
technology,
but
Bluetooth
also
has
some
application
layer
components
that
are
very
interesting
to
go
interface
through
an
internet
of
things
that
are
involved
now.
These
are
just
three
organizations
and
there
are
about
150.
We
are
probably
not
going
to
cover
all
those
150,
but
I
would
very
much
like
to
hear
from
you.
J
J
A
formal
presentation
from
the
thing
to
think
this
is
looking
to
buy
API
others,
heroes
that
are
from
the
active
in
iod
and
four
ways
very
much
involved.
There
is
the
one
MVM
s
do,
and
also
the
osea
Alliance
the
European.
During
the
recent
arrests,
then
I
can
also
try
to
get
the
SAV
conduction
and
arm
also
corporation
of
those.
B
K
A
K
A
A
N
That
feeler
I'm
vice
chair
by
Tripoli
802.
We
have
a
an
act
of
coordination
effort
with
the
ITF
a
through
the
joint
I
Tripoli
802
IETF
leadership
coordination
group
and
which
I
also
co-chair.
We
have
a
fair
amount
of
work
going
on
in
I
triple
e
802,
a
relative
to
things
not
so
much
I
OT
but
network
of
things.
You
know.
N
Basically,
since
we
are
at
layer,
2
areas
like
addressing
and
end
quality
of
service
for
the
time
sensitive
networks,
which,
of
course,
is
related
to
the
debt
net
working
group,
which
I
co-chair
here
so
yeah.
I
don't
think
there's
much
overlap
but
they're.
Certainly,
I
need
to
coordinate
for
the
things
we
do
it's
becoming
more
likely.
A
N
We
we
have
a
good
relationship.
We
maintain
a
ordination
list
to
keep
track
of
any
any
issues
that
arise
that
seem
to
particularly
need
coordination.
I,
don't
think
we
have
as
much
of
a
relationship
with
the
I
RTF
end
of
things
potentially
and
and
so
I'm,
not
sure
what
what
what's
needed
to
get
that
looped
in
but
but
yeah.
We
have
a
very
active
coordination
effort.
O
B
O
Yeah
so
yeah,
so
I
do
participate
with
zigbee
Alliance,
but
I
cannot
speak
for
their
lines
here,
but
I
think
they
mentioned
yesterday
about
the
semantic.
The
data
models
are
degree
ln
participative
fine.
I
don't
believe
I
don't
know,
but
I
don't
believe
there
is
an
ileus
awake,
I,
RTF
and
zigbee,
but
just
a
data
point
I
can
take
an
action
item
and
talk
to
that.
O
You
know
organization
and
this
members
meeting
is
happening
in
June
first
week
in
Amsterdam,
so
I
can
basically
enough
me
touch
with
you
and
like
if
there
is
an
opportunity,
sometimes
you
know
other
organ.
They
come
and
present
what's
going
on.
If
that
is
one
of
the
objectives,
also
right
because
it's
happening
in
Europe-
and
you
know
where
this
I
rdf
work
activities
and
we
present
it.
That
could
be
one.
You
know
action
item,
but
I
can
definitely
check
with
that.
The
organization-
and
you
know
get
back
to
you
because
I
don't.
C
L
B
P
Meet
adjustability,
so
I
think
I
just
have
a
few
questions.
The
outreach
cooperation
is
this
something
a
part
of
the
charter
of
this
working
group.
Oh,
and
is
this
some
something
we
like
other
working
group
in
IETF,
for
example,
debt,
net6,
low
six
fish?
We
can
take
advantage.
How
will
the
process
work?
Then?
You
know
if
we
get
get
to
know
about
other
stos
technologies
and
their
deployment.
P
B
First
of
all,
I
think
we
are.
We
are
pretty
good
connected
with
the
ITF
Merkel
said
that
I've
working
in
di
okie
state,
so
I
think
we
have
that.
Of
course,
the
idea
is
not
to
replace
the
IAB.
The
IETF
has
a
defined
J'son
model,
which
involves
the
the
IAB
where
we
need
formal
relationships
and,
of
course,
most
of
the
useful
work
is
done
in
informal
relationships
are
people
who
work
in
both
organizations,
so
we
are
not
trying
to
change
that.
B
D
And
I
would
also
comment
on
the
working
who's
versus
the
research
group,
as
our
Charter
says,
the
things
that
are
in
ours,
who
are
the
ones
that
have
potential
for
standardization,
so
East
annotation,
it's
already
happening
in
a
working
group.
Of
course
that's
a
natural
place
to
discuss
some
issue,
but
if
there's
some
issue
and
address
do
is
working
on
that
is
currently
not
in
many
of
our
working
groups
and.
A
B
Right
so
looking
at
the
space
that
the
research
group
spends,
we
probably
won't
do
application
specific
data
models,
so
we
are
not
so
interested
in
what
exactly
is
alight.
How
do
you
define
the
color
space
of
a
life
that
there
are
domain?
Specific
organizations
who
do
this
very
well
and
our
job
is
probably
more
or
to
find
out?
Where
are
these
organizations
how
it
can
we
collect
information?
B
O
G
A
semantic
interoperability
requires
self-describing
into
action.
Modeled
step
was
a
position
paper
that
we
handed
in
to
the
IEP
workshop
on
semantic
interoperability
kind
of
summarizing,
a
part
that
is
part
of
the
work
that's
what's
going
on
in
the
T
to
T,
RG
and
yeah.
So
in
general,
it
was
well
accepted
that
there
are
information
models
required
for
interoperability,
so
that
you
can
make
use
of
the
data
that
is
produced
by
the
IOT
devices.
G
This
is
mainly
about
the
what
what
does
it
mean
our
position
or
that
what
we
try
to
accomplish
is
that
we
also
have
a
model
that
this
machine
readable
to
describe
the.
How
so
how
can
machines
actually
access
this
information
in
describe
this
in
a
good
way,
and
so
what
we've
found
out
so
far,
it's
quite
straightforward
to
include
descriptions
for
this
on
the
server
side.
G
But
then
the
consumption
of
these
description
on
the
client
side
is
really
hard
and
also
for
programmers
finding
that
the
right
abstractions
is
hard
and
that's
why
we
often
ended
up
with
the
hard-coded
into
action
models,
and
we
don't
have
this
kind
of
machine,
readable,
flexible
interaction
model.
So
why
do
we
actually
need
it?
We
needed
to
handle
change,
and
it's
specifically
interesting
for
the
IOT,
because
we
will
have
devices
and
things
that
are
out
there
for
decades.
So
the
basic
idea
is,
we
can
interact
or
can
bind
two
things
together.
G
So
a
light
switch
can
work
with
a
lot
like
up,
and
we
also
want
you
to
replace
this
with
new
things
so
that
they
can
control
alternative
things.
It's
like
a
point.
We
want
to
add
new
things
that
can
also
be
controlled,
but
we
also
want
to
make
sure
that
these
kind
of
new
things
work
still
with
the
old
things
and
the
critical
thing
is.
G
We
actually
also
want
that
the
old
things
can
work
with
the
new
things
that
actually
are
designed
later,
and
for
that
we
need
a
kind
of
an
interaction
model
that
more
describes.
How
can
you
use
these
things
and
do
not
hard
code
it
that
it
just
works
with
the
two
things
we
know
already,
so
to
summarize
for
semantic
interpretation,
of
course
we
need
the
information
model,
so
this
is
describing
the
information
that
is
exchanged
for
that.
We
need
some
vocabulary
so
that
we
agree
on
that.
G
We
know
the
meaning
what
exchange
there's
some
connection
to
the
w3c.
What
working
group
they
try
to
have
the
semantic
model
rooted
in
rdf
to
have
these
kind
of
descriptions
and
be
working
on.
We
are
working
on
the
interaction
model
so
describing
the
possible
interactions
that
also
needs
vocabularies.
So
there's
a
quite
a
similarity
here
and
we,
and
particularly
I,
want
to
use
in
to
look
into
hyper
media
driven
applications.
So
what
comes
from
the
hypermedia
as
the
engine
of
application
state
constrained
from
rest?
G
A
G
Look
like
a
quick
summary
how
our
approach
looks
like
so
we
compose
things
or
model
things
for
multiple
resources,
so
yeah
resource
model
in
rest,
and
then
we
have
these
atomic
attack
steps
or
a
request-response
exchange
that
shape
the
processes,
how
yeah
the
whole
application
should
work
and,
of
course,
in
a
committee
you
use
links
in
forms
to
describe
how
a
machine
can
formulate
these
requests
and
we
need
some
vocabulary.
So
these
r
l
types
or
link
relation
heights
and
formulation
types
to
put
some
meaning.
A
G
Does
it
mean
to
invoke,
is
intellectual
right
now
actually
and
because
the
server
can
publish
these
links
and
forms
during
run
time,
we
also
can
influence
something
at
runtime
and
something
is
happening.
So
here's
a
nice
illustration
of
the
whole
thing
what
is
going
on,
so
the
idea
is
that
the
client
gets
an
entry
point,
affordable
thing
and
entry
point
your
eye
off
mrs.
McCall,
for
instance,
pointing
to
a
resource
directory.
It
could
be
your
google
search
engine
and
from
there
we
follow
links.
G
We
get
a
couple
of
links
back,
we
select
the
right
link
and
we
follow
it
and
from
there
we
follow
again
links.
So
this
could
be
on
the
same
resource
on
this
air
on
the
same
thing
could
also
point
to
other
things.
The
second
component
is
we
can
in
these
documents
we
also
learn
about
forms,
so
we
can
invoke
actions
and
we
submit
the
form
it
could
create
a
new
resource,
but
it
also
can
produce
something
like
an
actual
result.
G
If
we
process
something,
for
instance,
and
his
pattern
continues.
So
the
only
thing
we
do
is
follow,
links
and
submit
forms
and
those
need
to
be
described
to
to
formulate
the
whole
process
that
is
going
on.
So
sometimes
what
you
see
is
you
have
to
follow
a
lot
of
these
links,
which
might
be
inefficient
and
so
on.
So
there
are
a
lot
of
optimizations
that
are
possible.
G
One
obvious
here
is
to
use
bookmarking
that
is
known
from
from
the
web,
so
once
I
found
the
exact
resource
where
I
constantly
want
to
get
the
sensor,
reading
I
can
bookmark
this
and
interact
with
that.
One
and
just
read
it
out,
so
I
don't
have
to
have
all
the
discovered
process
every
time
and
the
nice
thing
is
because
we
can
publish
this
at
runtime.
We
can
decide
at
a
later
point:
hey.
We
need
a
new
mechanism
here
at
this
resource
before
instance,
want
to
have
some
access
control
there.
G
So
we
can
redirect
the
link
and
insert
during
runtime
a
new
kind
of
process
step,
and
this
way
handle
change
and
the
challenging
part.
We
want
to
look
into
how
to
find
the
programming
abstractions
for
that
another
ongoing
work
is
about
okay.
How
do
we
convey
all
this
information
so
year's
draft
by
Kawasaki
and
to
have
a
more
concise
representation
of
links
and
forms
leverages,
sip
or
two
to
encode?
All
this,
and
but
there's
also
a
more
generic
work
going
on.
G
So
what
kind
of
links
and
forms
to
be
actually
need
for
the
IOT,
so
the
summer
is
we
want
to
achieve
semantic
entrepreneurial
ability
with
this?
Where
change
handling
change
is
a
high
objective.
Of
course
we
need
information
models
for
that,
and
there
are
some
relation
to
the
w3
but
working
group.
So
we
have
these
joint
meetings
because
they
want
to
build.
G
This
is
information
model
and
we
focus
on
the
interaction
model
ever
decide
immediate,
written
application
approach
and
which
we
also,
by
the
way,
try
to
establish
at
the
definitions
at
the
w3c,
so
that
this
is
kind
of
the
benefit.
We
think
this
is
useful.
They
hopefully
also
go
see.
This
is
useful
and
then
had
some
alignment
across
different
standardization
organizations.
G
B
Thank
you.
I
hope
know
that
the
terms
information
model
and
interaction,
what
does
mean
a
little
bit
more,
so
the
problem
we
are
running
into
is,
of
course,
that
the
world
is
complicated
and
there
are
lots
of
things
out
there
and
there
will
be
many
things
out
there
that
have
some
characteristics
that
we
need
to
operate
on
in
this
information
and
interaction
models,
and
there
was
a
workshop
a
couple
of
weeks
ago,
organized
by
the
IAB
under
the
charm,
semantic
interoperability
and
one
hypothesis
that
was
formed
at
this
workshop.
B
Is
there
won't
be
one
schema
that
describes
the
world,
I
mean
their
sources.
Projects
like
you
might
might
have
heard
about
that.
It's
writing
this
email
for
the
world,
but
it's
kind
of
difficult
and
worse,
there's,
probably
not
even
going
to
be
one
language
where
use
different
domains
are
using
different
ways
to
describe
the
world
and,
however,
to
build
systems
where
we
have
lots
of
things
communicating
with
each
other
things
that
are
cut
that
are
coming
from
different
domains.
B
However,
we
need
to
interface
to
movies
things,
so
we
are
very
interested
in
in
technologies
that
can
be
used
to
collect
this
information
to
integrate
this
information
in,
but
in
particular,
also
to
translate
between
different
modeling
languages
and
just
to
give
a
peek
into
what
we
did
in
Tennessee
in
that
workshop.
Now,
that's
your
usual.
You
have
n
things
and
you
need
to
translate
between
them
picture,
and
so
you
need
n
square,
minus
M
translators
and
that's
the
future
of
the
intent
of
things.
B
Well,
maybe
there
is
something
that
can
be
put
in
the
middle
and
that
can
serve
as
a
hub,
but
yeah.
This
would
be
how
to
form
and
most
likely
it
will
not
be
one
such
a
red
star,
but
there
will
be
number
of
them
and
they
will
differ,
for
instance,
in
the
amount
of
data
loss
you're
getting
when
translating
from
one
email
to
four
different,
but
it's
also
important
to
think
about
two
different
levels.
Here
we
can
translate
data
from
different
data
models,
so
trivial
example,
of
course,
is.
B
We
have
a
temperature
sensor
that
generates
Fahrenheit
temperatures
and
we
want
to
know
that
information
insurgencies,
but
we
also
need
to
be
able
to
translate
the
models
themselves
in
the
first
place.
So
there
is
someone
something
that
says
this
is
the
Fahrenheit
base
temperature
sensor
and
we
want
to
use
this
same
information
in
a
different
place.
That
is
using
a
different
model
to
describe
the
same
information,
and
it's
not
just
about
information.
It's
not
just
about
sensors,
but
it's
also
about
interaction.
B
So
we
want
to
have
effects
on
the
real
world,
and
these
effects
often
lead
to
we
actuated
through
interactions
through
multiple
steps
before
my
coffee
machine
can
make
cover.
For
me
that
needs
to
know
what
the
strength
of
the
coffee
that
I
want,
and
so
on
done
so
looking
more
closely,
it
actually
turns
out.
This
is
an
area
where
we
of
course,
have
lots
of
prior
work.
B
So,
just
in
the
information
model
space
we
probably
have
three
layers:
we
have
the
information
model
layer,
we
have
the
data
model
layer
and
we
have
something
like
a
serialization
and,
of
course,
each
of
these
layers
has
about
10
different
terms
that
are
not
entirely
compatible
with
each
other.
So
just
taking
all
the
existing
state
of
the
art
is
not
trivial,
so
what
we
need
to
find
out
is
how
far
can
we
get
here?
Can
Can?
We
maybe
even
build
little
systems
that
demonstrate
some
capability.
B
So
one
of
the
next
things
we
want
to
take
up
is
looking
at
this
very
wide
field
and
find
those
areas
that
the
research
group
should
be
taking
up
research
that
we
should
be
encouraging.
That's
one
of
the
things
a
research
group
can
do
just
describe
the
challenges
that
we
are
having
in
such
a
way
that
researchers
can
can
go
ahead
and
work
on
these
challenges
and
then,
of
course,
find
that
research
and
take
it
in
to
to
this
organization.
H
So
that
so
what
I
would
like
to
see
is
if
someone
could
look
at
some
of
the
popular
sanitation
results
and
try
to
actually
create
the
mapping
between
the
different
data
models.
There
will
be
challenges
first,
because
some
of
the
organizations
haven't
actually
are
written
down
their
data
model
in
a
machine-readable
form,
so
it
would
be.
H
H
They're
all
seen
Alliance
guys
had
sent
me
a
link
to
their
Microsoft
wishers
with
your
Express
plugin,
where
they
use
their
describe
information
and
information
communication
model
who
then
generate
code
that
is
sort
of
like
as
a
b-side,
which
was
always
pretty
cool
and
but
I
haven't
seen
any
others.
But
then,
once
we
have
this
first
exercise
of
the
map,
we
could
potentially
get
actually
do
automatically
generate
code
from
one
to
the
other
or
can
use
the
results
that
have
been
worked
out
and
then
do
that
more
automatically.
H
Even
four
gateways,
I
would
be
very
interesting.
I
haven't
seen
that
work
even
trying
to
do
the
mappings
I
cannot
find
anything
like
someone
who
had
tried
that
say
the
likely
emblem
object
model
to
the
OIC
one
or
something
else,
but
maybe
there's.
If
some
one
of
you
knows
I
would
be
really
interested
to
look
at
those
recite.
The.
B
Q
Pocket
I'm,
Gabrus
crap
just
respond
to
that
the
comments
at
the
ifc
workshop.
I
actually
agree
with
our
members
comments
they
were,
but
that
you
can
generate
code
for
stoves
for
any
two
things,
but
you
actually
have
to
write
code
to
do
the
methane,
because
the
data
models
aren't
and
aren't
automatically
compatible.
You
actually
have
to
write
conversion
code,
so
I
don't
think
you
can
do
that,
and
that
was
just
a
point
that
that
was
a
product
at
sea
workshop.
I
think
those
generally
I'm.
H
Respond
student
I
remembered
at
this
world,
but
I
think
it
was
set
in
a
context
where
it
was
mentioned
of
like
zombie.
The
bluetooth
sig
low-energy
work
where
they
use
not
a
restful
style,
but
instead
just
an
RPC
based
mechanism
and
you
there
is
no
formal
description
of
how
the
interaction
what
you
actually
have
to
read
through
the
specification
to
figure
this
out.
So
it's
that
has
some
inherent
limitation
where
this
Arctic
obviously
doesn't
work
out
of
etiquette.
A
lot.
B
D
L
I
was
going
to
pass
one
Alice's
comment
and
then
Mike
similar
comment
to
Tate,
but
I.
Think
if
the
IIT
workshop,
we
also
had
discussion
about
what
returners
atoms,
which
were
certain
compatible
elements
between
all
the
data,
different
profiles,
data
models,
and
then
these
could
be
put
together
to
form
sort
of
molecules
and
more
complex
systems,
so
encourage
people.
They
have
a
look
at
the
slides
around
that
too
I
think
that
was
a
really
helpful
for
discussion.
We
had
at
the
workshop
at
none.
G
Epic,
since
it's
on
the
same
topic,
mckisco
watch,
I
would
say,
generating
code
is
not
you
and
I.
Think
it's
even
pointless.
It
doesn't
help
if
I
have
to
reprogram
my
device
to
make
it
work
with
a
different
kind
of
device,
and
if
there's
not
one
I
have
to
reprogram
under
guys
again
doesn't
help
in
the
whole
interoperability.
R
Mike
Amundson
with
CA
Technologies
I
would
just
say
that
interoperability
should
only
be
at
the
message
and
nowhere
else
we
shouldn't
concern
ourselves
with
anything
behind
the
interface.
We
can
be
interoperable
in
message
format,
and
that
includes
the
semantics,
because
that's
the
hard
work,
but
what
happens
behind
that?
We
should
not
know
about
data
models
or
anything
patient.
They
should
be
severed
thanks.
H
In
response
to
what
Medea
said
yeah,
there
seems
to
be
some
controversy
about
what
the
different
goals
are
and
some
people
who
believe
in
some
of
the
goals,
but
not
any
others
and
and
I
definitely
agree
with
you
that,
like
the
stuff
that
you
previously
presented,
has
obviously
more
powerful
concepts
in
there.
But
unfortunately,
this
is
all
this
body
of
work
that
has
been
done
in
the
past.
H
K
You're
opening
in
on
this
slide
says
what
should
we
be
encouraging
and
one
of
the
things
that
is
difficult,
as
Dave
villagers
said
about
there's
code
for
the
mapping,
so
that
you
can
have
all
the
semantic
tagging
in
the
world.
But
if
the
fundamental
modeling
constructs
are
different,
then
you
can't
match
them
up
without
some
human
interaction,
for
example,
units
is
either
a
flat
list
of
enumerated
units,
or
it's
like
powers
on
the
fundamental
meter
time
lumens.
You
know,
etc.
K
I'm,
yet
completely
different
ways
of
expressing
the
units,
that's
not
something
that
can
be
matched
up
by
submitted
k.
So
that's
what
so,
what
I'm
suggesting
with
that
is
not
go
as
me,
but
that
a
good
research
topic
would
be
to
look
for
these
like
design
patterns
book
for
these
mappings.
That
will
have
to
be
hard-coded
and
let's
create
a
list
of
these
mappings
that
are
needed
for
the
IOT
space,
and
then
we
can
all
hard
coat
these
mappings
and
then,
with
proper
semantic
tagging.
I
can
find.
Oh,
you
have
those
kinds
of
units.
K
E
Good
example
of
that
is
I
chipley
11
0
73,
which
has
a
lot
of
data
modeling
arm
embed.
No
lobbying
on
my
work
has
been
done
and
it's
really
depending
on
what
they're
dealing
with
their
that
for
most
for
medical
devices.
But
if
you
look
at
what
they've
done
there,
it's
really
much.
What
can
you
get
over
the
network
get
over
the
media
and
so
forth,
and
that's
that
make
one
example
see
how
energy
can
be
I.
B
B
B
So
I
have
one
question
you
in
this
room
is
a
professor,
andrew
loura,
so
I
think
what
we
really
useful
to
discuss
about
ways
to
get
academic
researchers
involved
in
this.
So
if
you
can
can
talk
to
me
after
this
segment,
I
would
be
really
interested,
so
we're
going
to
meet
in
balloon
in
the
context
of
ITF
96.
You
probably
also
will
use
the
w3c
meeting
enlist
ball
at
the
end
of
sep
tember
for
another
meeting
and
we
are
looking
at
good
ways
to
use
the
solar
meeting.
B
B
One
of
the
problems
that
we
are
having
is
software
updates
and
a
general
things.
Devices
need
to
be
updated
or
you
have
to
throw
them
away.
Nope
there
are
lots
of
people
are
doing
that.
So
it's
at
rest
side.
It's
not
a
research
problem,
but
then,
if
you
look
a
little
bit
closer,
there
are
a
lot
of
things
going
on
around
software
updates
that
raise
new
questions
where
maybe
the
existing
mechanisms
don't
quite
work.
So
the
idea
is
to
have
a
workshop
in
dual
to
discuss
these
new
issues
or
extended
issues
around
software.
It's.