►
From YouTube: IETF99-CDNI-20170717-1740
Description
CDNI meeting session at IETF99
2017/07/17 1740
https://datatracker.ietf.org/meeting/99/proceedings/
A
A
Phil
you
want
go
next
slide
please,
so
this
is
CD
and
I
unable
to
be
there.
So
we're
going
to
try
something
new
we're
gonna,
try,
remote,
sharing
Francois
also
could
not
make
it
to
Prague
this
time.
So
Phil
has
graciously
agreed
to
sit
in
for
us.
There
I
think
blue
sheets
should
be
going
around.
Did
we
get
a
someone
to
take
minutes
and
a
volunteer?
C
A
A
A
A
A
A
A
Since
last
time
there
were
three
errata
filed
against
the
triggers
RFC,
the
streaming
video
Alliance
open,
cashing
group
has
been
looking
at
CD
ni
as
a
basis
for
their
protocols,
and
so
they've
been
reviewing
most
of
our
documents.
They
did
find
some
issues
in
metadata
which
we
caught
before
publication.
A
Uri
signing
is
our
big
thing
left.
If
everyone
recalls,
we
decided
last
time
to
go
with
JSON
web
tokens
and
those
going
to
talk
about
that
in
just
a
second
next
slide.
Please,
on
the
non-working
group
milestones
that
we've
been
tracking
rate
pacing,
we
still
haven't
gone
gotten
much
traction
on
yet
the
CD
ni
HTTP
delegation
Frederick's
going
to
talk
about.
He
has
an
updated
draft
which
is
talking
about
in
lieu
of
the
lark
discussions
and
ACME
discussions,
some
thoughts
he
has
on
how
to
do.
Http
delegation
within
the
CDN
I
context.
A
Next
slide,
please
and
then,
as
I
mentioned,
the
streaming
video
Alliance
has
been
looking
at
CDN
I
interfaces.
They
have
discussed
in
their
draft
some
additional
metadata
that
they
would
like
to
see,
and
so
they
have
not
submitted
anything,
but
we
are
expecting
that
they
won't
submit
a
draft
at
some
point
with
some
more
metadata
that
they
think
is
useful
to
everybody.
A
A
C
So
I
want
to
give
you
guys
all
an
update
on
your
I
signing
graph,
basically
what's
new
since
berlin,
that
was
our
last
meeting.
So
in
berlin
we
all
agreed
to
two
JWT
from
our
custom
format.
They
actually
made
document
20
pages
smaller,
was
leveraging
existing
standards.
So
it's
going
to
require
less
review.
We're
not
gonna
have
to
go
through
it,
for
if
no
review
on
it
and
there's
plenty
of
existing
implementations
of
JWT
that
we
can
use
that
we
don't
have
to
expect
people
to
write
new
parser
for
our
old
format.
C
One
thing
about
this
is
before
berlin.
We
were
told
that
we
can
only
have
one
query
parameter,
get
off
my
lawn
each
and
so
having
a
base64
blob
of
the
GWT.
It's
not
actually
a
regression,
although
that's
brought
up
a
lot
of
times
as
a
critique
of
the
draft,
so
that
was
the
major
change
from
9
to
10.
I
don't
remember
when
we
even
published
n.
C
C
As
Kevin
mentioned,
we
had
the
streaming
video
Alliance
review
as
well.
I'd
present
it
to
them
on
a
video
conference
and
they
gave
a
lot
of
good
feedback.
Some
of
the
things
we
changed.
We
added
a
URI
hash
container
for
sub.
So
the
subject.
So
when
you
make
a
JWT
for
a
particular
URL,
you
have
to
have
the
entire
URL
in.
E
C
Also
the
GWT
header
is
the
same
at
the
beginning
of
every
GWT,
or
it
can
be
as
long
as
you're
using
the
same
signing
parameters.
So
we
created
a
metadata
header
so
that
we
could
pass
that
out-of-band
to
make
the
the
JWT
even
smaller,
and
also
the
CDN
ivy
claimed
that
we
added
we
made
it.
We
changed
it
to
be
optional
and
assumed
it
to
be
one
of
absent.
So
for
the
least
for
the
first
revision
of
this
of
the
spec,
the
the
CDN
Ivy
won't
be
in.
C
And
then
we
did
a
draft
version
12,
which
is
the
current
draft
version,
and
the
big
thing
we
added
to
this
was
HTTP
adaptive
streaming.
This
was
something
that
we
had
in
a
separate
draft
for
a
while,
because
there
were
some
IPR
concerns
and
working
group
decided
that
that
was
resolved.
They
offered
us
to
use
it
under
good
terms,
and
so
we
merged
it
back
in
with
with
the
JWT,
we
had
two
new
JWT
claims
that
we
had
to
create
for
that.
So
there's
now,
CDN,
I,
ETS
and
CDN
is
TT.
C
Just
in
case
people.
Don't
remember
about
the
HTTP
adaptive
streaming.
That's
the
ETS
is
the
amount
of
time
that
gets
added
to
the
next
token,
when
it's
chained
and
stt
is
how
the
token
should
be
passed
and
right
now,
I
think
the
default
value
is
only
equipments,
so
we
already
have
some
feedback
that
we
got
about
changes
that
should
be
made.
So
this
is
going
to
be
another
provision
of
the
draft,
so
it
was
requested
that
we
have
a
CDN
is
CT
registry
so
that
we
could
have
other
values
in
there.
E
C
C
A
Alright,
good
I'll,
just
try
been
here,
I
think
it's
had
a
couple
of
good
reviews.
I
did
a
a
pre
Shepard
review,
which
was
large
as
as
Phil
mentioned,
the
SBA
guys
reviewed
it.
Ben's
reviewed
it.
The
authors
have
reviewed
it.
I
think
we're
pretty
good
I
think
once
you
get
the
updated
rev
out
with
the
last
changes,
you
want
to
make
I'll
start
a
shepherd
review
and
we
can
start
the
working
group
last
call.
F
B
B
B
E
B
We
have
two
new
objects.
First,
one
is
secure
delegation
which
indicates
between
the
between
two
CD
ends,
which
is
a
dedication
method,
so
here
we
have,
for
instance,
ik
mr.
delegation
innocence,
and
we
must
also
advertise
for
which
pass
button
and
as
much
so
delegation
will
will
apply
to
as
the
next
object
is
the
specification
of
the
ACMA
star
delegation
meter.
So
we
mainly
provide
some
information
about
how
to
trigger
the
select
mister
the
division,
so
those
tulips
objects
can
be
still
discussed.
B
B
E
B
F
A
Have
a
have
a
couple
of
comments:
I
sent
some
comments
to
the
list
about
the
metadata
objects
themselves.
I
think
I
think
we
there's
more
in
there
than
as
necessary,
and
but
we
can,
we
can
discuss
that,
but
with
respect
to
the
sub
certs,
do
you
see
that
as
something
that
should
be
combined
with
acne
or
is
something
that
is
completely
orthogonal
to
acne
and
thus
would
warrant
a
different
draft?
Or
did
you
see
all
these
methods
as
needing
to
be
together,
I.
B
A
A
A
H
G
I
G
Is
Sanjay
I
just
just
to
add
a
little
bit
more
onto
well
I.
Think
the
the
important
thing
is
that
the
sub
certs
deals
the
delegation
in
a
very
different
manner,
so
I
think,
based
on
what
has
been
published.
That
draft,
which
we
should
probably
you
know
in
the
next
iteration
here,
because
we
probably
will
not
wait
until
the
next
draft
comes
out.
E
A
Okay,
I.
A
Guess
so
so,
if
you
guys
can
can
do
another
Rev
and
then
we
can
look
at
it,
I
haven't
really
thought
a
lot
about
how
this
fits
into
our
Charter.
We
may
need
to
talk
to
Alexi
about
that
and
modify
as
necessary,
but
I
think
at
this
point
it's
looking
good
looking
in
the
right
direction,
but
let's
see
what
what
it
looks
like
in
the
next
rep.
H
A
So
we
have
the
agenda
back
perfect.
We
have
reached
the
end
of
our
scheduled
agenda
at
this
point.
If
anyone
else
as
long
as
we're
talking
about
the
new
work,
we
will
probably
have
to
reach
our
turf
or
anything
beyond
URI
signing.
So
if
anyone
has
anything
else,
they'd
like
to
discuss
the
mic
is
open.
C
E
As
we
all
know
that
the
video
industry
developments
the
4k
8kb,
are
in
a
are,
they
both
requires
much
higher,
throughput
and
bandwidth,
and
also
North
latency,
for
example,
by
choose
a
limited
like
500
million
family
users
using
4k
TV
and
also
for
the
optimal
VR
network.
The
performance
depend
was
requested
about
three
cigarettes
and
remittances
about
less
than
eight
minutes
ago,
and
also
that
video,
its
emerging
and
the
developing
very
fast.
So.
J
J
E
E
E
So
different
from
current
operators
CDN
it's
especially
for
a
tech,
the
IPTV
market,
for
example-
it's
typically
closed
and
vertical,
so
we
propose
some
kind
of
concept
to
horizontally
decoupling,
the
CDN
Network
into
CD
encore
network
and
this
idiom
access
network,
a
Cydia
core
network
folks
on
the
upper
advanced
CBL
ability
to
which
is
very
strongly
related
to
the
contents
and
also
the
Cydia
access
network.
It's
better.
It's
desk
ability
provides
basic
ability
which
is
really
irrelevant
to
the
contents.
E
So
terrorist
speaking,
what
we
are
proposing
is
a
three
layered
city,
an
architectural
CDN,
Czar's
TDM
parts
and
CDN
ours,
so,
which
is
a
very
much
similar,
sounds
nice
account
computing.
So
the
city
and
stars
folks
provide
providing
the
Advanced
tab
in
affinity
on
services
like
the
user
interest,
forecast
and
hot
content
delivery
and
the
past.
Never
it
provides
general
contains
delivery,
cabinet
E
and
for
us
it
provides
computing
and
storage
resources.
E
So
combining
paths
and
us
we
can
provide
some
kind
of
a
basic
platform
which
is
operator
crossing,
which
is
which
means
we
can
have
some
kind
of
information,
resource
sharing
and
open
interface
or
open
idea,
API
between
them,
so
post
CDN
service
providers
and
the
CDN,
and
also
the
operators.
They
can
get
some
kind
of
benefits
from
from
this
kind
of
architecture
for
the
CDN
service
provider.
E
Like
him
folks
much
on
the
upper
layer,
advanced
service
and
for
operators,
they
can
utilize
the
center
office
resources,
which
is
currently
not
so
utilized
high,
hiney,
efficiently
and
also
for
operators.
They
can
reduce
the
reduce
the
backbone
investment
which
is
caused
by
the
very
strict
requirement
as
for
k8k
or
the
uhd
industry
development.
E
So
it
is
some
kind
of
post
the
technology
in
innovation
and
also
business
model
innovation.
It
needs
tight
cowork
with
CDN
providers
and
also
telecom
operators.
So
this
that's
why
we
are
here
to
present
this
kind
of
idea.
Since
all
of
you,
you
are
just
the
the
you're
just
right
stakeholders.
So
thank
you.
Thank
you
for
your
attention
at
questions.
G
E
We
think
that
you're,
currently
the
CD,
CD
and
I
you
previously-
you
have
some
kind
of
work
to
try
to
do
some
interconnection
between
CDN
service
providers,
but
we
think
that
it's
not
sufficient
enough,
so
I
think
to
image
the
whole
the
acedia
industry
space
so
such
as
the
new
emerging
application
needs
like
the
ultra
high-definition
video
industry
and
also
like
IOT,
smart
city
etcetera.
So
we
can
discuss
together
how
to
bring
how
to
emerge
the
market
space
and
also
to
utilize
many
current
resources
and
also
to
discuss
how
to
a
corporation.
E
So
maybe
we
can
consider
for
CDI.
Our
suggestion
is
to
is
that
we
maybe
we
can
consider
this
kind
of
a
wizard.
This
is
meaningful
for
all
of
the
CDN
stakeholders.
If
that's
in,
if
you
are
interested
in,
we
can
have
a
further
discussion,
and
maybe
it
maybe
some
new
topic
or
new,
maybe
new
new
direction
of
this
city,
so
s4c
denied
working
group.
Maybe
it's
a
it's
like,
maybe
the
the
next,
the
next
page
or
maybe
we
can
reach
other
the
see
didn't
like
working
group
yeah.
That's
just
our
suggestion.
H
E
E
A
I'm
good
answer:
I'm
gonna,
insert
myself
here
as
chair,
so
I
don't
think
designing
business
models
is
what
were
what
we're
here
to
do,
but
if,
if
we
have
a
technical
need
or
a
solution
that
we
can
develop,
that's
that's
something
different
in
terms
of
whether
or
not
this
applies
to
CD
and
I.
I
think
what
we
need
to
see
more
of
yes,
we
can
recharter
if
we
find
something
that
we
want
to
work
on,
but
if
this
is
all
intra
CDN,
it's
clearly
out
of
scope
of
CD
and
I.
A
We
have
always
said
that
the
CDN
I
interfaces
can
be
used
internally,
but
we
are
not
focused
on
that
usage.
But
if
you,
if
you
wanted
to
put
together
a
draft
with
a
use
case,
that
that
describes
how
to
use
CD
and
I
that
I
think
that
we
give
everyone
a
much
better
idea
of
what
it
is
you're
trying
to
achieve
and
how
again
like
Emile,
said
how
the
how
it
works
into
the
existing
CD
a--
and
I
interfaces
or
what
enhancements
may
be
needed,
similar
to
what
we're
looking
at
with
HTTP.
A
These
are
based
off
of
the
existing
api's
and
we're
extending
them
to
do
something
else.
We
can
certainly
extend
them
to
do
something
else,
but
I
think.
What
we
need
here
is
is
a
more
concrete
understanding
of
what
that
interface,
you
think
is
going
to
look
like
or
what
it
is
that
CD
and
I
can
provide.
E
E
E
A
I
think
it
would
be
interesting
if
you
wanted
to
put
together
a
draft
and
submit
a
draft
with
a
use
case
and
and
how
it
relates
to
CD
and
I
existing
or
what
enhancements
you
would
be.
Looking
for.
That
can
help
push
this
particular
model,
and
that
way
the
the
working
group
could
evaluate
it
in
a
more
concrete
manner.
E
Write
a
draft
and
you
the
internet
draft
or
yeah
yeah.
So
yes,
a
Michael.
K
E
K
A
draft
to
hear
to
that
CDI
to
check
is
there
anything
you
needed
to
do
in
this
working
group,
and
what
for
here
are
single
CDR
I
we
need
to
in
the
following.
You
know
we
need
to
focus
on
the
technology,
for
example,
to
check
a
weather.
Is
there
any
something
new?
Besides?
What's
the
existing
protocol
is
defined
here?
You
know,
okay,.