►
From YouTube: IETF111-QIRG-20210729-2030
Description
QIRG meeting session at IETF111
2021/07/29 2030
https://datatracker.ietf.org/meeting/111/proceedings/
B
Okay,
it's
10
30.
So
let's
begin
first
before
I
go
through
the
usual
notewell,
we
need
a
minute
taker.
B
So
looking
for
volunteers,
usually
we
would
also
have
a
jabra
scribe,
but
I'm
not
sure
we
really
need
one.
So
if
we
have
a
minute
taker,
that's
grand!
So
if
somebody
wants
to
step
up.
B
I'll,
let
you
guys
sit
on
that
until
I
finish
the
note
well,
so
I'm
just
going
to
show
a
few
slides.
The
iotf
note.
B
B
B
C
D
But
so
you
say
that
again
I'll
be
in
and
out
I
was.
I
have
a
couple
of
other
things
to
do,
but
as
long
as
I'm
in
I
can
take
it,
I
is
this
on
ether,
whatever
something
or
kodi
md.
D
D
Well,
it's
an
hour
I
can.
I
can
take
minutes
and
just
point
me
to
where
I
should
go
for
the.
B
D
D
Well,
it's
slightly
different
in
mine,
but
I
found
it
so.
Okay.
B
Cool,
like
I
said,
I
couldn't
proceed
thanks
a
lot.
Curitious
appreciate
it.
Okay,
so
today
it's
only
an
hour.
I've
got
a
very
small
agenda
as
some
administrative
I,
like
the
network
that
we
just
did
and.
B
We
gonna
do
a
bit
of
an
energy
status.
Then
there's
gonna
kaushik
is
going
to
present
the
use
case.
Draft
gonna
have
a
discussion
on
usually
the
usual
open
floor.
Just
open
discussion
questions
as
a
proposal
for
discussion
and
one
we
could
center
it
around
is
the
architecture
position
paper
that
brought
emailed
to
the
mailing
list
and,
at
the
end,
I'm
going
to
also
announce
the
next
seminar
that
we
have
scheduled
for
september.
B
So
the
current
document
status
is
that
the
principal's
draft
is
actually
not
present
for
the
agenda
of
this
meeting,
because
it's
currently
under
irsg
review
shepard,
who
was
completely
shipwrecked
right,
it
was
completed
by
david
earns.
I
would
like
to
thank
him
a
lot
for
that.
I
will
be
discussing
the
use
case
draft
today.
Hopefully,
it's
nearing
completion
we'll
see
after
today.
B
I've
had
a
look
at
the
update
and,
as
usual,
I
will
keep
reminding
people.
There's
three
expired
job,
because
every
now
and
then
somebody
comes
to
the
mailing
list
and
asks
about
it.
No
other
even
I'm
gonna
leave
this
for
later.
So
that's
about
it.
So
I'm
gonna
leave
the
floor
to
koshy.
Then
kashik
are
you
ready
to
present?
B
E
E
E
A
What
looks
like
you're
on
your
keynote
screen
and
the
so,
if
you're,
actually
in
presentation
mode,
we're
we're
not
seeing
the
presentation
mode
we're
only
seeing
we're
seeing
your
editor
screen.
F
E
Yesterday:
okay,
yes,
okay!
So
thanks
why
tech
and
hi
everyone?
So
I'm
koshik
so
today
I'll
be
talking
about
the
draft
on
this
application
and
use
cases
for
the
quantum
internet.
So
I
joined
this
this
project
like
few
weeks
ago
and.
B
C
E
I
pronunciate
the
name
for
properly
so
let's
see
what
we
have
been
doing
for
the
last
few
months,
so
this
is
like
a
small
background
like
on
version.
4
was
presented
like
during
this
ietf
110
on
march
10th,
and
then
we
had
like
lots
of
comments
and
then
that's.
Why,
like
the
authors
like
they
prepared
the
version
five
and
it
was
uploaded
on
march
29th,
and
then
they
prepared,
like
a
version
6
where
they
include
on
the
first
byzantine
negotiation
and
it
was
uploaded
in
may
3rd
and
the
version
7.
E
Actually
the
current
version
was
uploaded
on
july
12th
and
now
we
are
presenting
it
and
we
will
show
like
what
are
the
updates
see
here.
So
the
most
update
like
was
like
okay,
I
was
I
joined
in
that
draft
as
a
co-author,
and
then
we
added
like
three
application,
like
quantum
money,
quantum
imaging
and
quantum
chemistry.
We
try
to
be
fair
for
each
of
the
domain
of
this
quantum
technology,
like
quantum
money,
is
an
application
for
quantum
internet.
E
Quantum
imaging
mostly
application
for
quantum
metrology
and
and
quantum
chemistry
is
most
related
to
the
quantum
community
computing.
So
then
we
added
also
some
new
definitions
like
some
for
some
some.
E
Okay.
So
these
are
the
bullet
points
and
now,
let's
have
a
little
bit
closer
look.
What
are
the
updates
so,
on
the
right
hand,
side
you
see
all
these
red
points.
They
are
like
the
sections
where
we
mostly
made
the
updates.
E
For
example,
in
section
2,
we
added
those
new
definitions
and
in
section
3,
for
example,
in
section
3.2.1,
3.2.2
and
three
point
two
point:
three:
we
added
the
application
new
applications
of
the
quantum
network,
like
quantum
money,
quantum
imaging
and
quantum
chemistry,
and
in
section
four
is
more
specifically
like
section
4.1,
we
added
the
description
about
measurement
device,
independent
qkd
and
continuous
variable
qkd.
E
Those
are
the
two
variants
of
the
original
quantum
key
distribution
protocol
that
with
a
high
probability,
it
will
be
like
those
two
technologies
will
be
very
popular
in
the
near
future
for
for
the
implementation,
and
then
we
also
made
some
other
for
the
changes
in
section
five,
so
today
I'll
be
just
going
through
all
this
section,
two
and
three
and
four
now
talk
about
those
small
changes
we
made
in
section
five,
okay.
So
let's
start
with
section
two,
so
in
page
4
we
added
the
definition
of
epr
pair.
E
E
You
should
say
it's
a
like
the.
If
your
pairs
are
like
four
orthogonal
states,
all
of
them
are
orthogonal
to
this
state,
and
so
what's
the
like
interesting
fact
about?
Is
that,
like
those
two
qubits,
they
have
some
kind
of
special
correlation
that
cannot
be
observed
in
the
classical
information
theory,
and
it
turns
out
that
this
kind
of
special
correlation
can
be
a
very
useful
or
maybe
a
crucial
point
or
crucial
resource.
B
E
For
quantum
communication
and
quantum
computing
and
most
of
the
time,
although
it's
it's
not
exactly
clear,
but
it
it's
a
strong
belief
that
this
kind
of
correlation
gives
a
quantum
advantage
over
the
what
the
classical
one
okay.
So
then,
let's
talk
about
something
called
entanglement
swapping.
We
also
added
this
definition
in
page
four,
so
it's
a
very
interesting
tool
or
like
a
process
for
sharing
entanglement
between
two
distant
parties
via
some
intermediate
nodes.
E
Suppose
we
have
like
two
apr
pairs.
Let's
say
a
b
and
c
d
where
a
and
d
they
are
a
bit
a
distant
party
and
b
and
c
they
are
sitting
on
the
same
lab
and
a
one
two
share
an
epr
pair
with
2d.
E
They
can
perform
some
measurement
called
bell
state
measurement,
and
if
it
is
successful,
then
the
correlation
between
a
b
c
d
will
be
destroyed
and
there
will
be
a
correlation
between
bc
and
there
will
be
a
correlation
between
a
and
d.
So
that
means
like
if
the
bell
state
measurement
is
successful,
a
and
b
will
share
an
epr
pair
and
that's
a
very
basic
tool
for
for
the
untrusted
quantum
repeater
technology.
So
we
use
this.
This
kind
of
entanglement
swapping
operation
to
create
an
epr
pair
between
two
distance
end
nodes.
E
E
So
it
means
that
if
I
want
to
send
a
qubit
data
qubit
to
a
distant
party
instead
of
sending
it
physically
to
the
communication
quantum
communication
channel,
I
can
just
use
some.
B
E
Of
local
operation
and
classical
of
communication,
so
more
precisely
suppose,
alice
and
bob
they
share
an
epr
pair,
which
looks
like
this
and
alice
wants
to
send
a
qubits.
I
unknown
qubit
side
to
bob
to
do
this
like
what
alice
can
do
ellis
can
perform
the
bell
state
measurement,
the
same
kind
of
measurement
we
have
seen
in
this
entanglement
swapping
operation.
E
So
after
the
measurement,
what
will
happen?
The
correlation
between
alice
and
bob
will
be
gone
and
alice
will
get
some
measurement
outcome.
Sigma
and
bob
will
instantly
get
the
state
like
sigma
of
psi,
but
he
will
not
have
any
information
about
the
sigma
sigma
is
like
a
unitary
operator.
We
call
them
pauli
operator
and
to
get
the
psy
back
from
sigma
of
psi.
What
alice
needs
to
do
and
is
to
send
this
classical
sigma
to
bob
and
from
there
bob
can
retrieve
the
the
state
side.
E
E
Okay.
Now,
let's
focus
on
the
application
on
section
two.
So
sorry,
section
three
in
section
three
point:
two
point:
one:
we
add
like
an
application
for
quantum
cryptography.
It's
called
quantum
money,
so
quantum
money
is
like
one
of
the
first
discovered
applications
for
for
quantum
information
theory.
It
was
discovered
before
even
bb-84
so
because
the
idea
was
that,
like
intuitively
at
that
time,
there
was
no
no
cloning
theorem.
I
think
here.
So
the
intuition
was
that
if
we
are
given
a
single
qubit,
we
cannot
create
a
copy
of
it.
E
So
that's
a
perfect
candidate
for
creating
a
bank
node,
because
you
cannot
forge
those
nodes,
but
the
problem
during
that
time
was
that,
like
okay,
no
one
could
create
a
copy,
but
even
no
one
could
verify
those
banknotes.
So
that's
why
that
idea
was
not
that
popular.
But
very
recently
there
are
like
more
sophisticated
protocols
like
it's
called
public
key
quantum
money
and
everything
where
the
anyone
can
verify
the
bank
notes
locally,
but
it
so.
E
The
security
is
like
under
some
kind
of
computational
rebounded
adversary,
okay,
so
so
this
is
a
very
interesting
application
for
quantum
cryptography
and
then
we
added
another
application
for
quantum
sensor.
It's
called
quantum
imaging.
E
So
here
what
we
can
do,
like
is
the
network
of
quantum
sensors
actually
can
measure
the
magnetic
fields
generated
by
the
flow
of
current
through
some
neural
assemblies
in
the
brain.
While
the
subject
is
moving
and
this
kind
of
thing,
we
cannot
do
using
currently
classical
strategies
so
and
then,
because
of
that,
actually
it
can
reveal
the
dynamics
of
the
networks
of
neurons
inside
the
human
brain
on
a
millisecond
time
scale,
so
it's
showing
as
a
huge
potential
for
the
for
the
application.
B
E
Quantum
algorithm
to
somehow
determine
the
molecular
energy
energies
of
the
small
molecules
when
they
then
chemical
accuracy.
So
the
interesting
part
is
that
for
to
implement
such
algorithms,
we
don't
need
a
full
scale.
Quantum
computer
with
the
niska
devices
like
these
near-term
quantum
devices
we
can.
We
can
realize
those
things.
Of
course
you
know
for
a
very
small
number
of
molecules,
but
if
we
want
to
increase
the
number
of
molecules,
then
we
need
a
bigger
computer,
but
at
least
they
are
like
a
proper
applications
for
for
the
near-term
quantum
computers.
E
B
E
Holy
grail,
we
cannot
have
a
secrecy
more
than
that
better
than
that,
but
the
problem
is
that
all
of
those
things
are
in
in
pain
and
paper.
So
if
you
come
to
the
practice
like,
if
you
want
to
build
those
things,
all
the
measurement
devices,
all
the
like
detectors,
they
have
the
flaw
and
by
exploiting
those
flaws
like
people
in
first,
I
think
in
2009
the
people
were
like
come
up
with
the
like
side
channel
attack
on
the
con
commercial
qkd
devices.
E
They
call
like
quantum
hackers
so
after
that,
like
people
are
thinking,
okay,
so
how
to
mitigate
those.
These
kind
of
side,
channel
attacks
one
possible
way
to
avoid
the
side.
Channel
attacks
which
are
coming
from
the
detectors
is
like
this
measurement.
G
E
Independent
qcd,
okay,
so
using
this
kind
of
protocol
you
can
even
if,
like
the
attacker,
if
you
use
this
kind
of
protocol-
and
even
if
that
tries
to
manipulate
the
detector,
then
also
the
they
will
not
be
able
to
generate
the
key
or
case
the
key
which
is
shared
by
alice
and
bob
after
the
successful
completely
successfully
completing
the
protocol.
E
So
that's
a
good
advantage,
so
there
are
like
more
sophisticated
version
of
this
measurement
device.
Independent
protocol
is
called
device,
independent
quantum
key
distribution
where
you
don't
alice
and
bob
doesn't
need
to
trust
any
of
your
their
like
hardware
devices,
but
these
things
are
bit
away
from
the
practical
implementation,
whereas
this
measurement
device
independent
qkd
like
many
experimental
demonstration
for
them.
So
another
protocol,
which
I
would
like
to
know
another
version
of
the
qkd
or
like
qkd
technology,
we
added
hearing
in
page
13,
is
called
continuous
variable.
E
Qkd,
it's
a
it's
comparatively
new
technology
so
where,
instead
of
sending
like
a
district
discrete
quantum
state
here,
what
we
do
we
send
gaussian
modulated
like
coherent
state,
and
so
the
advantage
is
that,
like
you,
can
use
the
classical
communication
technology
right
away
to
fall
for
this
kind
of
continuous
variable
qkd.
So
it's
very
engineering
friendly
and
implementation
friendly.
E
But
the
problem
is
that,
unlike
the
discrete
variable
qkd
here,
the
security
proofs
and
everything
they
are
not
that
mature,
but
the
the
community
is
working
very
hard
like
maybe
in
the
couple
of
years
we
will
have
very
concrete
security
analysis
and
it
can
be
used
for
in
in
practice.
So
mostly
so.
Okay,
another
disadvantage
for
this
one
is
that,
like
we
cannot
have
a.
We
still
don't
have
a
quantum
repeater
for
them,
but
they
are
good
for
short
distance
communication,
even
also
for
the
free
space
communications.
E
Okay.
So
these
are
the
all
these
updates
and
let's
now
talk
about
the
what
are
the
next
steps.
So
this
is
our
version
seven
and
it
seems
that
it's
a
bit
stable
and
we
also
it's
stable,
because
we
take
into
account
all
kind
of
comments
we
get
from
this
ietf
109
and
one
ietf1110
and
yeah.
We
are
still
waiting
for
for
new
comments
if
it
is,
if
there
is
not
enough
comment,
then
it's
stable
and
then
we
can
check
whether
it's
ready
for
the
last
call
or
not
so.
H
Yeah
thanks
for
watching
and
thank
you
kochie.
That
was
really
interesting.
I
I
just
have
one
a
one
word
suggestion
and
it
relates
to
section
3.2.2
of
the
draft
I
yeah
so
quantum
center
applications.
I
wondered
if
you
had,
because
because
you
used
the
word
metrology
when
you
were
introducing
the
the
deck
and
I
wondered
if
you
would
consider
replacing
quantum
sensor,
applications
with
quantum
metrology
applications.
E
H
I
mean
the
simple
reason
being
it
wouldn't
exclude
any
of
the
use
cases
that
you're
looking
at,
but
it
would
also
give
the
opportunity
to
include
specifically
time
quantum
clocks,
which
I
think.
E
A
The
metrology
versus
sensors,
the
I
have
for
several
years,
been
using
the
term
sensors
very
broadly,
to
include
metrology
and
clock
synchronization,
and
things
like
that,
including
some
of
the
papers
we've
published.
A
H
I
I
kind
of
do
because
well
the
the
session
I
was
in
they,
they
described
quantum
clocks
as
a
subset
of
their
metrology
work,
whereas
I
wouldn't
typically
think
of
a
clock
as
a
sensor.
E
C
H
So
yeah,
okay,
it
may
just
have
been
a
a
a
different
taxonomy
in
use
by
what
was
actually
I
mean.
It
was
specifically
a
uk
group
of
academics,
and
maybe
they
just
have
a
slightly
different
taxonomy.
Kaushik
is
probably
much
better
placed
to
to
gauge
that
than
I
am,
but
it
was.
It
was.
B
E
E
B
D
Basically,
I
think
robin's
point
is
to
include
all
the
applications.
Now
you
can
choose
a
name
that
is
more
inclusive,
but
let's
just
include
the
applications
and
then
pick
a
name
that
everyone's
okay
with.
B
Yeah,
I
agree
with
that
sentiment
and
I
have
one
question
for
you:
karshik,
which
is,
if
you're,
adding
quantum
money.
One
of
the
obvious
things
to
at
least
contrast
this
by
saying
is:
how
does
it
compare
to
cryptocurrency
is?
Is
there
a
value
in
quantum
money
now
that
cryptocurrency
is
a
thing.
E
Money
yeah
I
mean
so
otherwise
you
exploit
the
the
no
cloning
theorem
that
you
cannot
copy
your
your
qubit,
something
like
that
right.
So
so
I
think
this
quantum
cryptocurrency
would
be
like
something
something
different
and
then
the
current
definition
of
quantum
money,
so
right
yeah.
But
I
think
it's
a
bit
different.
G
E
A
My
hand
up
and
just
sort
of
jumping
in
and
interfering
like
I
usually
do.
I
have
a
few
comments
on
on
the
draft,
some
some
that
are
directly
related
to
what
caution
just
presented,
some
of
which
aren't
one
that's
been
on
the
back
of
my
on
on
the
back
burner
for
a
long
time
I
haven't
gotten
around
to
mentioning.
A
I
don't
think,
is
this
draft
uses
the
term
epr
pair,
but
why
tech's
architectural
principles
draft
uses
the
term
belt
pair
and
I
think,
as
qrg
we
would
be
best
off
in
picking
one
consistent
term,
and
in
this
case
I
think
I
think,
we've
already
sort
of
settled
on
belfair
belfair
has
the
advantage
that
it's
easier
to
talk
about
talk
about
bell
states
as
a
basis
set.
A
B
Sorry
and
based
on
the
content
of
the
chat
from
robin
it
actually
seems
there
may
be
some
people
confused
as
to
whether
epr
pair
is
the
same
thing
as
a
valve
power,
so
just
to
clear
that
out,
yes,
epr
pair
is
the
same
thing
as
belter
same
thing
as
entangled
pair.
There's
lots
of
names
floating
around
for
them.
Anything
that's
to
do
with
a
quantum
internet
and
pair
is
most
likely
a
bell
pair.
A
A
I
think
pointing
out
that
there
appears
to
be
this
sort
of
non-local
correlation
in
quantum
mechanics
and
therefore
this
was
one
of
the
things
that
sort
of
bugged
einstein
about
about
quantum
mechanics,
whether
or
not
it
was
actually
complete
and
correct
and
whatnot,
but
their
paper
was
actually
relatively
qualitative
and
certainly
doesn't
use
modern
terminology
and
notation.
But
in
the
1960s
john
bell
said
you
know
if
I,
if
einstein's
objection
is
really
right,
here's
an
actual
here's,
an
actual
experiment,
here's
an
actual
mathematical
way.
A
We
could
sit
down
and
do
it
and
here's
a
term
you
could
connect
to
to
doing
it
and
he
and
part
of
what
he
defined
is
now
called
bell
pairs,
and
you
can
actually
also
use
them
as
a
basis
set
for
for
actually
discussing
all
two
qubit
states.
Just
like
you
could
talk
about.
You
know
combinations
of
zero,
zero,
zero
one,
one
zero
and
one
one
so
in
in
the
engineering
sense.
It's
it's!
It's
more
useful
and
more
mathematically
defined
in
a
modern
sense.
E
Like
more
more
like
a
general,
they
talk
about
this
entire
correlation.
Whether
epr
pair
is
mostly
looks
like
exactly
like
this,
so
so
that's
also
like
a
built-in,
more
general
sense.
Yeah.
A
Great
that
was
one
comment.
I've
got
one
or
two
others
here,
let's
see
so
kaushik.
You
just
mentioned
that
the
mdi
and
cvqkd
you
mentioned
them
they're
now
added
to
the
draft-
and
you
said
oh
by
the
way,
we
don't
really
have
a
good
way
to
do
this
over
repeaters.
E
A
Part
yeah,
because
particularly
for
cv,
because
because
repeaters
are
they're,
essentially
digital
you're,
abstracted
away
digital
level
stuff
and
whereas,
whereas
cv
is
you
know,
it's
literally
continuous
variable
stuff.
So
I
don't
know
how
you
do
that
over
health
pairs.
Mdi
also
is
sort
of
a
physical
level
concept
as
opposed
to
and-
and
you
know,
abstracted
away
at
sort
of
the
digital
level
right,
but
maybe
maybe
a
little.
B
No
so
I'll
say
my
comment,
so
I
read
the
draft
and
I
think
I
think
it's
not
especially
with
rodney's
comments
being
where
says
she
just
cleaned
up
stuff
and
all
the
comments
being
cleaned
up.
So
I
think
it's
getting
near
it's
basically
there.
I
I
did
notice,
there's
quite
a
few
typos
in
the
latest.
Edits
though
typo
isn't.
B
Things
I'd
improve,
but
I'll
send
that
to
the
mailing
list
to
the
list
of
what,
but
after
those
last
I
I
think
it's
basically
ready.
I'm
happy
with
the
draft
as
it
is.
I
think
so
I
think
I'd
be
ready
for
rg
last
call
and
comments
are
slowing
down.
So
that's
just
my
comment
on
it.
I
You
this
thing
asks
you
for
for
permission:
every
time
you
try
to
talk
or
to
be
singing.
No,
it
is
just
a
a
thought.
Looking
at
what
has
been
presenting
and
how
the
document
has
evolved,
I
was
just
wondering
whether
calling
it
use
cases
is
a
little
bit
limited
because
it
has
to
be
more
about.
I
know
I
don't
know
how
to
call
it
application
scenarios
or
something
like
that.
I
It
goes
beyond
a
collection
of
concrete
use
cases
simply
that
it's
not
that
I'm
I'm
pretty
happy
with
how
the
the
document
is,
but
it
might
be
that
if
we
want
to,
I
mean
in
the
future
when
this
goes
further
and
people
may
be
interested
in
having
a
look
at
the
documents-
probably
use
cases,
it's
a
little
bit
limited.
That's
all.
B
I
B
B
Cool,
so
the
next
point
on
the
agenda
was
open
flow
discussion.
We
tend
to
always
have,
and
there
always
seems
to
be
enough
to
discuss
some
people
ask
questions
that
we
do
actually
fail
at
the
time.
This
time
I
suggested
rodney
and
I
suggested
to
add
a
point
about
the
mail,
the
architectural
prince
paper.
B
What
is
it
called
architectural
position
paper
that
rodney
sent
out
some
time
ago?
So,
if
there's
any
questions
about
it
or
discussion
to
the
advocate,
there
was
one
on
the
mailing
list
some
time
ago.
B
It's
a
good
time
for
you
to
ask
questions
any
clarity
about
that
or
rodney
to
just
ask
questions
to
the
floor.
A
Let
me
preface
this
with
a
couple
of
comments
on
the,
so
the
primary
discussion
on
the
mailing
list
was
between
alex
and
joe
touch,
I'm
not
sure
if
either
of
them
is
actually
here
today,
I
didn't
see
them
on
the
attendee
list.
Last
I
looked
the.
A
I
was
not
totally
ignoring
the
conversation
going
on
about
about
that,
but
I
didn't
feel
the
need
to
step
in
because
because
because
joe's
replies
I
thought
were
actually
pretty
good
joe
and
I
actually
agreed
on
a
lot
of
things
that
blog
posting
was
put
together,
sort
of
in
response
to
well
sort
of
not
not
particularly
response,
but
the
it
has
become
time
to
start
thinking
about
all
of
the
pieces
that
are
required
for
complete
architecture
and
to
start
putting
one
together,
and
my
research
group
over
the
last
you
know
almost
15
years
has
has.
A
Worked
at
different
parts
of
it
we
worked
at
at
you
know
the
the
recursion,
which
we
think
is
the
right
approach
for
doing
for
doing
quantum
internet
working
we've
worked
on
how
nodes
communicate
about
what
the
sequence
of
operations
is,
that
they
ought
to
that.
They
ought
to
be
executing
which
we
call
the
the
rule
set
based
approach.
A
We've
worked
on
routing
most
recently,
we've
worked
on
security,
and
so
it
seems
like
it's
time
to
start
putting
all
of
that
together
and
I
will
say
we're
actually
in
the
process
of
starting
from
that
blog
post
and
actually
turning
it
into
a
a
full-scale
actual
research
paper
to
be
submitted
to
one
of
the
the
top
networking
conferences
with
a
deadline
coming
up.
A
But
even
on
top
of
that,
whether
that,
whether
that's
accepted
to
the
conference
or
not
on
down
the
road
it
seems
like
working
on
similar
areas
in
qirg
is
probably
a
good
way
to
go
forward
for
from
here.
And
so
if
people
have
either
comments
on
the
technical
content
of
that
particular
blog
posting
or
the
discussions
in
the
mailing
list
or
want
to
discuss
whether
or
not
it's
that
that's
the
kind
of
direction
we
want
to
take
qirg
and
going
forward,
I'm
open
to
discussion
on
all
of
that.
B
So
are
there
any
questions
from
the
floor
based
on
that
email
chain,
slash
position
paper?
I
have
one
comment,
which
is
that
I,
like
the
rule,
set
based
approach.
I
call
it
under
a
different
name:
that's
more
familiar
to
the
classical
networking
audience
which
is
matching
action
yeah.
B
I
like
that
approach
because
yeah,
it's
good
and
it's
good
foundation
for
actually
building
up
a
functional
control,
plane
et
cetera,
also
just
to
make
this
link
for
the
audience
when
rodney
says
we'll
set
based
approach,
it's
based
on
a
condition
plus
action
clauses
which
you
can
just
replace
the
name
condition
with
the
name
for
match,
and
you
get
the
same
thing.
A
Yeah,
so
it's
much
like
a
software-defined
networking
architecture,
except
that
we're
viewing
it
as
the
rule
sets,
is
being
constructed
and
distributed
in
a
distributed
fashion
rather
than
depending
on
a
single
centralized
controller.
So
we
think
this
is
a
this
will
actually
work
across
a
global
internet,
and
we
also
think
it
will
work
reasonably
well
with
with
the
actual
recursive
architecture,
which
is
the
other
sort
of
key
element
of
what
we're
proposing.
A
But
the
my
group
is
actually
you
know
we're
actively
working
right
now
on
a
lot
of
the
details
of
how
how
the
rule
sets
actually
interact
with
each
other
and
trying
to
figure
out
issues
like
how
you
manage
resources
inside
a
quantum
repeater.
You
know
how
many
are
you
going
to
define
design?
Are
you
going
to
assign
this
set
of
qubits
to
to
connection
a
and
that
set
of
qubits
to
connection
b?
A
How
do
you
deal
with
synchronization
issues?
How
do
you
avoid
deadlock?
How
do
you,
how
do
you
maximize
the
the
the
autonomy
of
the
nodes
and
minimize
the
minimize
the
number
of
round-trip
or
one-way
latency
weights
and
things
of
that
issue
without
having
to
to,
without
potentially
sacrificing
correctness?
A
So
all
of
those
you
know
all
of
those
kinds
of
issues
are
exactly
the
kinds
of
things
where
qirg
participants
really
have
the
expertise,
because
the
particularly
for
these
1g
networks,
the
first
generation
of
networks
using
the
entanglement
swapping
and
purification
they're,
very
much
a
distributed
computation
building,
building
end-to-end
entanglement,
involves
actions
and
decision-making
at
nodes
all
along
the
path,
so
so
actually
doing
that
that
level
of
distributed
computation,
that's
something
the
physicists
are
not
good
at,
and
the
people
in
in
ietf
and
irtf
are
good
at
and
so
that's
the
kind
of
thing
where
we
would
like
to
have
more
contribution
from
the
qirg
community.
D
A
A
Yeah,
the
so
again,
there's
there's
one
where
we
probably
ought
to
pick
one
term
and
use
it.
The
wojtek
tends
to
use
distillation
began
and
I
tend
to
use
purification
and
we're
really
probably
out
to
to
standardize
on
one
term.
They
actually
have
a
slightly
different
algebraic
meaning,
but
I'm
not
enough
of
a
mathematical
physicist
to
sort
of
keep
them
straight
wojtek.
Any
other
comments
on
that.
B
Yes,
so
I
spent
probably
a
day
once
figuring
out
which
term
to
use-
and
I
ask
somebody
who
knows
maths
better
than
me-
and
they
explained
purification
to
me
and
I
did
not
understand
it
and
I
went
for
distillation.
Then
it's
basically
the
long,
long
story.
Short
purification
has
some
very
particular
meaning
to
do
with
interpreting
quantum
state
in
the
context
of
a
density
matrix
or
something
like
that,
and
I
think
it
actually
is
incorrect
to
use
the
term
purification.
A
I
did
not
create
the
term
I
I
picked
it
up
from
somewhere
else
and,
and
I
think
a
lot
of
the
the
early
papers
on
on
the
concept
actually
use
the
term
purification
when
you're
talking
about
taking
using
multiple
entangled
states
to
put
together
you
to
essentially
perform
tests
on
one
entangled
state
and
create
the
and
improve
the
fidelity
of
the
end-to-end
connection.
But
the
postdoc
in
my
group
mikael
is
actually
more
more
of
a
he's,
a
theoretical
physicist
and
I
think
I
think,
given
a
blank
sheet
of
paper.
A
I
think
he
would
probably
lean
toward
the
term
distillation
as
well.
He
described
to
me
once
why
why
he
thinks
that
it's
a
more
appropriate
term,
but
the
explanation
was,
you
know
I
got
I
sort
of
nodded
along
and
then
afterwards
went.
Okay,
I'm
not
sure.
I
understand
it,
but
I
understand
mikel's
position.
E
Okay,
so
I
think
I
mean
I
would
also
prefer
distillation
over
over
purification.
I
think
it's
more
appropriate
in
a
sense
that
purification
and
somehow
we
say
like
okay,
if
we
have
a
density
matrix,
a
matrix
representation
of
a
cubit
like
by
purification
we
mean
like
okay,
we
go
to
the
larger
hilbert
station
and
then
consider
the
pure
purest
version
of
that
of
the
density
density
matrix.
So
in
the
distillation
we
somehow
basically
says
okay.
E
Here
we
have
multiple
epr
pairs
with
or
multiple
state,
which
is
low
fidelity
and
we
want
to
distill
the
the
entanglement
out
of
it.
So
it's
just
like
that's
why
I
prefer
the
distillation
over
or
purification
so
yeah.
B
Good
question
any
other
questions.
Basic
questions
you
can
see
are
fine
because,
whilst
perhaps
me,
rod,
koshik
and
a
few
others
might
understand,
what's
going
on,
it's
a
it's
a
good
chance
for
us
to
actually
make
sure
everybody's
on
the
same
page.
A
Yeah
in
voidtech's
draft
on
the
the
architectural
principles
draft
section,
4.4.3,
error
management,
the
first,
the
first
sub
sub
subsection
of
that
is
distillation
and
it
says
through
a
process
called
distillation
parentheses,
sometimes
also
referred
to
as
purification
and
parentheses.
A
So
I
think
we
have
sort
of
converged
on
on
the
term
distillation
I'll
do
my
best
to
to
to
use
distillation
rather
than
purification.
Although
the
decades
have
it
is
sort
of
hard
to
break.
G
Hello,
can
you
all
hear
me?
Yes,
yes,
yes,
good
good,
so
I'm
gonna
ask
the
really
stupid
question,
because
I
don't
even
pretend
to
understand
anything
about
quantum
internet.
G
So
so
you
know,
you've
got
the
principles
drift
and-
and
if
I
understand
correctly,
that's
sketching
out
a
sort
of
set
of
concepts
and
ideas
which
you
would
expect
different
quantum
quantum
internet
to
to
obey,
and
this
this
blog
post
is
sketching
out.
One
concrete
realization
of
those
principles
in
a
in
a
a
particular
network.
G
A
In
the
end,
well,
sir,
we'll
certainly
need
to
converge
on
one
inter-networking
architecture,
but,
but
I
I
would,
I
fully
expect
that
there
will
be
multiple
network,
architectures
and
so
far
ours
is
is,
is
you
know,
despite
over
a
decade's
work?
It's
very
embryonic,
and
yet
it's
still
the
only
real
inter-networking
proposal
I
know
of
as
opposed
to
network
proposal.
A
There
will
be
there
will
be
many
different
physical
implementations
for
sure,
and
there
will
also
be
different
philosophies
for
how
to
control
those
networks
and
different
philosophies
for
how
errors,
within
those
net
those
networks,
and
then
those
networks
will
certainly
want
to
interconnect
in
in
the
classical
inter-networking
sense,
and
so
something
needs
to
happen
at
that
boundary.
A
We
also
we
are
proposing
that
that
our
rule
set
based
approach
can
actually
be
used
all
the
way
down
to
to
essentially
the
link
level,
but
in
an
internet
working
sense.
It
would
not
have
to
be
if
you
want
to
do
something
different
for
for
your
for
your
link
and
network
architecture.
That's
fine!
As
long
as
as
long
as
at
the
boundary
you
could,
you
could
meet
the
right
requirements
for
for
for
doing
the
internet,
working.
G
B
That
test,
but
so
you
mentioned
test
beds,
yes,
so
there
are
because
so
the
united
states
have
published
this
document
about
their.
I
think
they
call
it
blueprint
as
well.
Then,
in
europe
there
are
demonstration
networks
built,
in
fact
the
netherlands
actually
happens
to
allocate
615
million
euros
to
quantum
technology,
and
some
of
that
is
going
to
quantum
networking
and
actually
building
and
this
one's
actually
a
very.
This
is
not.
This
is
not
just
for
pure
research.
B
C
B
In
fact,
actually
let
me
just
find
you
very
quickly
a
link
after
I
answer
this
because
actually
would
be
very
good
to
share
yes,
the
plan
is
to
have
very
much
have
within
the
next
few
years.
I'm
just
trying
to
be
careful
too,
that
I
only
say
what
the
public
information
is
and
the
plan
is
to
actually
have
networks
that
are
open
to
industry,
for
use
cases
to
open
to
net
researchers
for
researching
protocols.
The
idea
is
to
actually
have
networks
open
to
people
who
want
to
use
them
and
explore
them.
B
A
Okay,
great,
thank
you,
while
you're
doing
that
robin
is
also
in
the
queue
here
and
we're
down
to
just
a
few
minutes
left
on
our
time
schedule
here,
robin.
H
Okay
is
mike,
live
now,
yeah
great,
so
yeah.
Thank
you
colin,
because
you've
kind
of
given
me
permission
to
ask
my
own
my
own
noob
question.
So
following
on
from
colin's
question,
I
think
in
all
in
all
the
qirg
sessions
that
I've
been
in
any
quantum
network
has
always
had
to
be
accompanied
by
a
classical
computing
kind
of
control
plane.
H
A
Yeah
and
in
that
sense,
the
the
work
that
we're
primarily
doing
right
now
in
in
that
in
that
terminology
would
actually
probably
be
control
plane
type
work
rather
than
the
actual
data
plane
related
work.
Although
when
you
talk
about
data
plane,
there
there's
both
the
need
for
the
actual
physical
photons
and
memories
if
the
photons
flying
around
and
memories
sitting
still
and
the
the
and
also
you
know
sort
of
the
low-level
control
of
that
which
is
still
also
ultimately
classical
control.
A
H
H
Would
might
one
of
those
be
that
in
one
quantum
network
you
choose
one
particular
entanglement
property
to
measure
and
in
a
in
an
adjacent
quantum
network,
the
the
other
guys
might
have
chosen
some
completely
different
property
of
their
pairs
to
measure.
H
A
D
Yes,
when
rodney
said
that
you
know
the
connection
between
quantum
and
classical
networks
will
primarily
be
in
the
control
plane
that
got
me
perking
up,
because
I
had
written
a
draft
on
that
which
is
long
expired,
but
you
know
that
that's
more
about
sharing
in
an
ospf
connection
in
a
classical
network,
that's
adjacent
to
the
quantum
network,
sharing
properties
of
the
nodes
in
the
quantum
network,
like
how
many
qubits,
how
many
storage
qubits,
how
many
traveling
qubits
you
have
and
so
on.
D
I
don't
know
if
that
I
don't
know
if
people
have
had
a
chance
to
actually
look
at
that
and
if
that's
the
foundation
for
at
least
some
of
the
control
plane
work
that
reflects
the
quantum
network
in
the
classical
network.
So
people
can
make
useful
decisions
about
how
they
want
to
create
new
entanglements
or
distill
them
or
whatever.
A
Not
particularly,
I
was
actually
just
looking
for
the
expired
draft,
I
haven't
pulled
it
up,
you
pulled
it
up
yet,
but
the
if
you
want
to
revive
it,
I
would
say
by
all
means:
let's
do
it,
but
you
know
in
light
of
the
work
we've
done
in
the
last.
You
know
what
almost
almost
two
years
in
qirg
here
it
might
mean.
You
know
we
might
look
at
it
and
decide
you
know
it
either
needs
to
come.
B
I
look
forward
to
discussion
around
it
myself.
So
I'm
sorry,
sorry
for
cutting
rodney,
but
we
do
have
four
minutes
and
we
take
a
will
cut
us
off
so
so
I
would
say
interesting.
I
think
I
think
the
one
the
one
issues
is
being
a
very
practical
draft.
It
would
be
cool
to
actually
see
something
in
action,
but
I
don't
know
I
shoot
an
email
on
the
mailing
list.
To
be
honest,
I
found
it.
I
found
it
an
interesting
topic
myself.
So
we'll
see
what
we.
A
B
Yeah,
okay,
thanks
kiriti,
so
let
me
just
quickly
share
my
screen.
B
B
So
so,
basically,
we
will
announce
it
again
at
the
qrg
mailing
list.
It's
actually
already
officially
announced,
but
I
will
just
keep
sending
reminders
and
encounter
them
and
everything
so
yeah
everybody's
welcome.
Hopefully
we
have
a
good
attendance
and
hopefully
be
very
interesting.
Paul
banks
said
he'll,
be
talking
about
control,
plane
topics
which
should
be
of
interest
to
a
lot
of
people
in
their
qrirg.
B
B
So
I
think
you
can
actually
just
stop
there
and
yes,
so
thank
you.
A
A
Okay
hope
to
see
you
all
in
november.
It
would
be
nice
to
do
it
face
to
face,
but
if
not
the
we'll
probably
do
another
session
sort
of
like
this,
where
we
sort
of
run
through
the
drafts
sort
of
sort
of
work,
you're
pretty
direct
working
stuff
on
it.
If.
G
A
November
end,
in
the
september
september
seminar,
your
seminar
should
be
good.
Please
come.
H
H
B
A
A
The
because
the
ietf
and
irtf
are
really
composed,
primarily
of
professionals
for
for
whom
you
you.
The
the
focus
of
of
what
gets
discussed
at
the
meetings
is
really
their
their
their
their
their
full-time
day.
Job.
And
for
me
for
me
that
that
serves
as
incentive
to
try
to
keep
up.
A
B
Yeah
now
we'll
see.
B
I
it'd
be
good
to
actually
keep
get
the
use
cases,
application
scenarios,
draft
or
use
cases-
yes,
they're
called
yeah.