►
From YouTube: IETF102-LISP-20180719-1330
Description
LISP meeting session at IETF102
2018/07/19 1330
https://datatracker.ietf.org/meeting/102/proceedings/
B
C
F
It's
Sunday
actually
I
may
have
a
problem
in
Bank.
There's
a
separate
issue
get
into
Bank
on
Sunday
is
going
to
be
fun
because
I
have
supposed
to
be
in
Los,
Angeles,
Thursday
and
Friday
before
that,
and
it
turns
out
that
I
wanna
use
United
flying
United
from
LA
to
Bangkok.
You
either
connect
on
a
morning
flight
through
Narita
or
nighttime.
You
can
fly
through
San
Francisco
to
Singapore
to
Bangkok
I.
F
B
F
B
The
usual
pointers
to
the
Java
room
out
of
three
immutable
and
the
agenda
slides,
okay,
so
to
a
brief
update
on
the
states
of
the
documents.
So,
since
London
we,
we
have
a
non
newbies
document,
which
is
the
68th
for
this
yeah.
It's
about
map
versioning,
okay.
We
need
to
push
this
forward
as
we're
under
under
standard
track.
Okay,
because
all
the
sick,
38:33
beasts
they
refer
to
this
one
as
well.
Okay,
it's
kind
of
blurry
but
didn't
find
the
right
set
up
in.
B
So
then
we
have
the
usual
intro
and
SEC
documents
that
instant
by
different
stage,
we
have
the
vendor
al-kahf.
Actually,
that
was
in
working
group
last
call
in
the
last
two
weeks.
This
was
just
our
consensus
check
because
we,
this
document
was
presented
in
London.
It
was
asked
for
working,
two
plus
call
the
room
said:
okay,
there
was
consensus
then,
but
there
were
few
changes
that
I
asked
editorial,
so
they
came
in
few
weeks
ago
and
we
so
we
open
the
official
working
group
last
call.
We
had
a
few
emails
confirming
the
the
support.
B
The
saw
I
think
when
we
are
done
with
this
document
later
on,
I
will
send
our
formal
email
on
the
mailing
list
and
and
the
usual
process,
chef,
holding,
etc,
etc.
Now,
as
you
can
see
on
this
slide,
there
are
few
other
points
which
is
so
we
have
three
documents:
the
best
documents
plus
GB
that
are
gone,
really
done.
Okay,
even
though
the
right
app
is
ready,
but
I
don't
have
IP
our
disclosure
from
all
of
the
others
fabulou,
you
included.
F
Remember
this
IPR
disclosure
is
not
just
Cisco
hot.
This
is
to
you
know
of
IPR,
because,
presumably
you
know
all
the
things
you
should
know
of.
So
it's
not.
This
isn't
do
a
disclosure.
This
is,
do
you
know
of
something
that
isn't
disclosed?
If
so,
then
you
have
and
do
it
but
tell
us,
so
we
know
to
wait
for
that.
If
not
tell
us,
so
we
can
get
this
done.
This
is
not
an
onerous
thing.
B
Yeah
I
will
continue
pink.
The
other
out
was
on
the
these
documents
of
we
will
get
a
reply.
Otherwise,
we
will
write
down
in
the
write-up
that
we
don't
have
I
PR
disclosure
from
all
of
the
others.
We
will
send
anyway
for
forward
the
document.
Okay
to
the
is
G.
The
problem
will
kind
of
come
back
when
we
will
have
the
48
hour
out
of
confirmation
that
that
where
we
really
need
confirmation
from
the
others,
I
mean
and
we'll
see
how
to
deal
with
it.
B
Hopefully,
in
the
meantime,
we
will
get
in
touch
with
all
of
the
others.
Okay.
Well
as
there
is
still
concerning
IPR.
Actually
I
forgot
already
last
time
to
point
out
that
we
have
two
IPR
disclosure
about
Lisp,
AI,
D,
anonymity
and
lease
predictive
are
locks,
so
both
are
from
war
way.
You
can
check
on
the
data
trackers
I.
I
B
Do
you
can
have
more
information
about
this
IPR
disclosure
and
one
last
point
coming
back
to
the
B's
documents?
So
there
is
no
idea
of
disclosure
there
directly.
But
if
you
go
back
in
and
check
out
the
original
heart
of
C,
so
68
30
and
60
833,
they
are
alcoholics,
kosher
again
from
Y
way
so
which
means
that
basically
in
the
Indian
right
up,
we
will
state
that
because
the
B's
documents
are
based
on
the
original
documents
for
which
there
is
an
idea,
so
it
may
apply.
B
Then
it's
not
up
to
you
to
check
this,
but
just
point
out
these
things
may
be
done.
Do
not
I,
don't
know,
okay
and
that's
it
as
an
update
so
coming
to
the
agenda
today
we
have
an
update
on
the
young
model.
Is
the
only
official
document
we
will
discuss?
Rashad
can
Kali
come
here,
and
then
we
have
a
bunch
of
non-working
departments.
Mainly
Dean
will
talk
about
the
console,
plain
authentication,
authorization,
data,
Plymouth,
telemetry
Fred
should
stop
Polly
and
talk
about
gambe
still
lists
for
aeronautical
telecommunication
network
later
on
Alberto.
B
F
K
G
G
G
Just
okay,
thank
you!
Okay!
So
if
decap
swing
next
year's
participate
in
this
telemetry
feature
more
performance
data
is
available
and
we
can
make
better
informed
decisions
like
you
know
what
you
transmitted
on
ITR,
what
you
encapsulate,
but
you
don't
know,
what's
being
received
on
the
other
side,
so
you
can
measure
different
things
like
packet
loss
and
bandwidth
and
jitter.
G
So
this
was
first.
A
list
was
put
together
with
the
first
0-0
draft
is
doing,
is
just
identifying
what
sort
of
parameters
we
want
and
how
to
encode
them,
and
then
what
later
drafts
will
do
is
to
find
how
its
measured
on
the
ETR.
So
after
we
do
some
implementation,
we
could
say
how
we
measure
these
things,
but
things
like
packet
count
by
count
rate
packet
rate,
bitrate
bandwidth.
G
G
So,
like
I
said,
the
early
version
is
just
showing
that
we
send
these
things
in
our
load.
Probes,
we
oughta
look
probe
ascent
and
the
EGR
has
been
measuring
and
collecting
all
this
data
and
it
returns
back
in
an
hour.
Look
Pro
reply
an
extra
hour
look
record:
that's
the
JSON
elk
f-type
that
encodes
all
this
stuff
and
we
want
to
use
JSON,
because
if
we
want
to
add
parameters
later,
we
can
do
it
in
a
compatible
way
without
changing
the
protocol
and
arguably
without
changing
the
implementations
as
well.
L
And
as
I
looked
at
the
stuff,
I
realized
that
oh,
we
should
probably
consider
this
colouring
thing
that
people
are
doing
an
IP
PM
right.
There's
this
way
one
way
of
measuring
loss
is
you
send
n
packets,
colored,
red
and
then
n,
packets,
colored,
blue
and
then
the
receiver
knows
that.
Okay,
it
only
received
98.
There
were
colored
red
and
there's
some
even
some
techniques
to
do
this
to
get
some
latency
measurement,
so
you
either
need
one
or
two
bits
somewhere,
but
I
find
a
place
for
those
bits.
Okay,.
B
I
B
I
can
ask
chemically
the
data
about
mappings
like
why
not
order
I
mean
if
you
are
a
next
year.
Why
somebody
else
cannot
ask
you
telemetry
data
about
one
mapping
now,
then
you
have
to
have
a
policy
that
says:
can
Victor
knows
about
the
the
telemetry
of
which
is
mapping
on
Ondina
or
not,
but
why
not
kids?
So
what.
G
You'll,
see
in
my
next
presentation
is
that
we
probably
don't
want
to
give
telemetry
data
to
anybody
who
just
requests
it,
so
we
want
to
actually
sign
map
requests
and
validate
them.
So
when
you
return
the
information,
you
know
it's
to
an
authenticated
source,
so
that's
one
thing
and
the
other
thing
is
is
if,
if
there's
a
source
that
wants
to
make
its
decisions
unilaterally
without
having
a
management
plane,
we
can
make
it
self-contained
in
Lisp
in
the
control
plane.
G
G
G
Ietf
did
one
update
two
updates
and
we
have
another
one
planned
for
the
summer
and
at
that
time
I
we
want
to
keep
it
as
an
individual
submission
and
what's
the
working
group
thinks
it's
good
to
make
it
a
working
group
document,
but
I'll
talk
about
that
next,
so
just
a
brief
over
you
know
what
the
draft
is.
Basically,
you
can
authenticate
and
authorize
XT
hours
using
the
mapping
system
and
then
in
the
way
it
does.
G
It
describes
how
to
sign
map
register
messages
and
how
to
sign
me
up
request
messages
and
the
map
server
actually
verifies
the
signatures
by
looking
up
public
keys
in
the
mapping
system.
So
we
use
or
introducing
this
concept
of
crypto
a
IDs
where
part
of
the
ipv6
Eid
is
a
hash
of
the
public
key,
that's
used
when,
when
a
message
is
signed
that
crypto
Eid
hash
is
used
as
a
lookup
to
the
mapping
system.
So
the
map
server
itself
can
retrieve
the
public
key
and
then
verify
the
signature.
G
If
the
signature
fails,
the
map
register
map
requests
is
dropped.
If
it's
succeeds,
then
the
next
levels
of
authentication
for
a
map
register
the
shared
key
authentication
to
see.
Are
you
authorized
to
register
this
particular
Eid?
In
this
instance,
ID
is
then
checked.
Do
you
know
if
I
may
ask
a
question
sure.
F
G
So
the
way
we
expected
out
was
the
hash
that
you
look
up
is
appended
to
an
ASCII
string
called
hash,
so
it's
actually
a
distinguished
name
as
the
Eid
in
the
Eid
record
and
DDT
can
build
hierarchy
on
any
ascii
character
in
any
string.
So
we've
played
with
this
before
so
you
know:
here's
the
delegation
for
Joel
and
here's
Joel
Halpern
a
helper
in
Prime,
and
it's
all
subordinate
to
that.
So
you
can.
You
can
break
the
hey,
a
hierarchy
up
not
on
bit
boundaries,
of
course,
but
on
character,
boundaries
right,
okay,
okay,.
G
So
well,
what
we
also
did
is
we
recently
introduced
signature
e
IDs.
The
signature
Eid
is
a
crypto
Eid.
Even
though
it's
not
one
that
you're
using
for
doing
ipv6
communication,
it
could
still
be
used
to
authenticate
other
things
that
you
want
to
send
to
our
register
the
mapping
system
or
request
from
the
mapping
system.
G
G
A
M
N
Yes,
so,
as
you
guys
know
very
well,
and
now
that
is
contraband
except
right-
that
we
have
taken
the
chance
to
to
see
how
we
can
apply
the
least
control
into
to
definitive
edge
the
last
time,
which
is
10
days
for
LA.
Now
that
I
know
these
4s
every
six,
the
the
tag
working
progress
really
applies
to
to
this
one.
Because,
as
we
speak,
we
are
discussing
changes
on
on
the
diaphragm
and
the
underwear
so
for
the
next
iteration
it
would
change
and
I
will
have
to
refine
as
we
go.
N
Offer
their
services
English
notes
information
about
where
they
have
to
send
the
traffic
so,
as
we
put
in
at
least
least
four
forward
to
go.
So
basically,
we
map
the
ladies,
the
point
being
is
heading
to
the
egress,
a
survey
six
node
and
the
VPNs
are
met
on
that
degrees,
not
where
the
traffic
has
to
go,
then
the
rest
of
his
stuff.
When
a
service
is
about
traffic
engineering,
far.
M
N
So
in
this
in
the
person
person
of
the
draft
in
the
zero
zero.
We
cover,
like
kind
of
the
foot
picture
right
where
you
have
an
overlay
with
mobility
and
defecate.
You
needn't
be
in
every
six,
but
becomes
already
about
that.
We
should
be
clear
on
that.
That's
not
necessary
and
I
would
plant
to
clarify
that
first
and
you
don't
need
traffic
engineer.
So
if
you
are
interested
in
sending
traffic
from
the
English
to
the
egress,
and
so
you
need
to
to
go
further
for
the
whole
thing,
but
for
now
we
have
covered.
N
So
these
are
the
basic
ideas
later
on
I
have
some
slides
with
examples
that
will
help
a
lot
to
clarify
how
things
work,
but
this
is
the
basic
idea
of
the
draft,
so
you
rely
on
the
IGAs.
A
services
knows
to
register
the
end
points
into
the
maps
solution,
and
then
we
use
a
new
aircraft
to
register
the
services
needed
information.
N
Thus,
the
two
points
are
covered
on
the
at
the
bottom
of
the
slide
and
and
then
how
do
you
resolve
the
end
points?
So
basically,
we
present
two
models
in
the
in
the
draft,
one
where
the
English
note
request
then
put
information
to
knowledge
system.
It
gets
the
exit,
pipian
seeds,
so
like
the
the
final
service
expansion
that
the
traffic
has
to
go
to,
and
also
that
little
information
that
we
saw
in
the
previous
slide
like
the
they,
the
Eagers
know,
the
traffic
is
still
intact
and
Alda.
N
N
So
first,
first,
what
I
call
him
a
slice,
this
general
solution
so
going
to
the
map
server
and
then
going
to
the
PC
right.
So
this
is
the
basic,
a
scenario
that
we
consider
you
can
see
their
doors.
A
and
C
are
the
English
and
aggressive
notes
and
on
the
left,
you
can
see
the
information
in
the
English
note
a
as
well
as
the
information
they
on
the
map.
Server
and
you'll
see
how
the
information
gets
gets
populated
as
we
go.
N
If
you
look
here
at
the
table,
you
can
see
how,
for
instance,
a
has
two
important
right
and
it
has
two
recognition
functions
that
point
to
the
two
graphs
that
that
guy
is
in
parenting.
So
when
you
want
to
send
traffic
to
to
this
valve,
you
will
use
this
one
and
so
on,
and
you
can
imagine
a
similar
table
for
for
C
all
right.
N
So
this
is
happy
to
start
a
mob
register,
so
they
just
send
information
about
the
local
endpoints,
the
localities
to
the
server
as
usual,
and
what
the
history
is,
the
the
information
about.
If
you
want
to
go
to
this
guy,
that
is
an
instance
91.
This
is
the
ingress
Sr
function
that
you
need
to
capture
the
traffic
towers,
and
this
is
the
traffic
is
didn't
act,
the
color
that
you
actually
with
that
particular
Empire.
So
you
can
use
that
later
on
to
communicate
the
PC
to
compute
the
appropriate
path
through
there
to
the
under
map.
N
So
now,
let's
see,
let's
say
that
you
have
traffic
going
from
from
V
to
X,
as
you
can
see
this
or
destination
source
added
another's
payload.
So
what
happens
is
commonly
stuff?
My
request,
my
reply,
so
you
get
the
this
line
from
the
from
the
map
request,
molecular
change
where
the
Ino
knows
that
to
go
to
X,
it
has
to
encapsulate
traffic
towards
C
column,
D
33,
and
that
it
can
use
the
direct
product
to
to
find
the
appropriate,
not
in
the
under
day.
N
So
that
is
what
is
happening
here.
Yeah
I
know
they're
not
going
to
the
details
of
how
they
sell
encapsulation
go
straight
case
that
you
guys
can
take
a
look
at
that
on
the
spring.
Go
through
the
plenty
of
presentations
around
on
this
basically
captures
a
traffic
through
the
services
underlay
to
until
be
the
capsule
eight
in
Si
and
deliver
as
usual,
so
now
mobility,
which
is
what
we
covered
in
the
drive
right.
N
So
let's
assume
that
X
moves
to
a
new
egress
point
e,
so
this
is
regularly
spaced
off
your
cinema
blister
and
this
my
register
well
I've
put
both
options
in
in
there,
so
I
consider
the
mobility
and
the
ATF
this
pops
up.
Some
of
you
may
argue,
the
only
one
is
needed
and
I
can
I
could
agree
with
that.
So
after
us
how
you
want
to
deal
with
mobility,
so
you
can
use
one.
The
other
are
both,
so
they
are
all,
as
you
know,
interval
so
I
put
them
both.
G
N
In
the
next
model,
so
in
this
model
we
assume
that
the
SR
part
is
like
computed
by
the
one,
the
computer,
it's
really
from
the
PC
directly
from
the
English.
So
it's
like
this
joint
from
the
front
dailies
retrieval
of
the
final
decapsulation
point,
but
in
the
model
where
we
call
okay,
these
two
options.
That
is
certainly
the
case.
So
if
the
SL
passengers
in
definitely
fine
so.
G
In
the
EIT
mobility
draft
we
never.
This
is
AB
new
setting
a
precedent
which
is
ok.
It.
The
ID
mobilities
draft,
said
whenever
the
our
log
changes
or
has
a
new
binding
and
we
trigger
map
registers,
we
do
SMRs,
we
do
pub/sub,
we
do
data
trigger
SM,
ours,
whatever
the
case
for
all
those
mobility
cases,
but
I,
don't
remember
if
the
text
says
if
the
our
load,
if
the
contents
of
the
our
local
record
changes
versus
the
are
locating
well.
F
F
F
N
G
Is
not
a
criticism
by
any
means
on
segment
routing,
but
let's
think
of
this,
let's
look
at
just
a
ID
mobility
earning
over
an
MPLS
cloud
and,
let's
say
the
label
switches
for
the
underlay.
That's
the
same
case
here
where
the
our
look
doesn't
change.
Just
the
label
changes
because
you
have
a
new
te
path.
I,
don't
know
if
we
have
language
that
talks
about
the
characteristics
of
the
underlay
have
changed
the
how-to
part,
not
the
wear
part
science.
O
F
N
Let's
maybe
go
through
this
second
model,
because
I
think
your
question
applies
more
to
the
second
mother
than
to
this
one.
So,
let's
just
finish
with
this
once
the
English
has
been
notified
of
the
new
attachment
point
for
the
endpoint
it
will
find
if
needed,
and
us
our
policy
to
go
through
the
through
the
underlay
traffic
goes
there
okay.
So
now
the
second
model
is
I
think
what
universe
as
never
in
this
model,
the
map
server,
is
the
one
that
interacts
with
the
PC
to
find
the
paths
to
the
underlay.
N
So
things
are
like
in
a
previous
one.
Let's
see
if
that
everything
has
been
released
already
and
whatnot,
so
traffic
appears
on
the
ingress
map.
Request
is
trick.
What
happens
here
is
that
the
MSM
arrow
goes
to
the
PC
or
maybe
firstborn
previously
on
has
something
for
the
store
there.
So
the
map
server
goes
to
the
to
the
see
and
okay
I
want
to
know,
which
is
the
SL
policy
to
go
to
the
study
series
because
they
have
someone
that
is,
that
is
asking
about
an
endpoint
attached
to
that
aggressor.
N
I,
understand
the
whole
path
to
to
that
guy.
So
the
luxury
gets
up
and
sends
back
a
map
reply
with
the
with
the
complete
path
right.
So
in
this
case
the
note
a
will
have
a
Marcus
entry
that
says
that
okay
I
want
to
send
traffic
to
X,
so
I
have
to
use
that
s
a
b6
path
that
goes
through
B
and
then
C
using
the
decapsulation
function.
N
The
the
traffic
course
like
in
the
other
example
mobility
is
the
same,
and
in
this
case
it's
worth
the
the
mobility
event
triggers
also
the
depart
for
the
SL,
which
seeks
English
now
to
reach
the
ETS,
not
say
this
as
well,
and
this
is
I
think
that
the
case
that
we
know
was
mentioning.
So,
let's
assume
that
yes,
the
SR
policy
changes
so
yeah.
In
this
second
model,
the
map
Shara
will
notify
like
okay.
I
I've
sent
this
information
to
you
in
the
past
and
others
in
Fircrest
change.
So
here's
the
new
information.
G
E
N
N
N
And
then
right,
this
is
really
working
progress,
so
things
may
change
quite
a
bit
for
for
personal
one
and
at
some
point,
while
this
is
more
stable,
we
went
over
to
to
Experian
Camp
list
and
there,
and
we
also
plan
to
have
this
listed
limitation
on
on
the
open
source
router.
So
we
can
play
a
bit
with
it.
So
that's
all
I
have.
R
So
Victor
Cisco
question
on
on
the
second
model,
where
you
have
two
single
look
up
once
you
get,
the
update
from
the
PC
is
the
thought
to
actually
push
that
update
to
the
ITRs
or
are
you
gonna,
send
them
up
notified
to
the
ETRS
involved,
and
then
let
the
SMR
mechanisms
take
place
or
are
we
restrict?
Are
we
thinking
about
restricting
this
to
pops
up
as
saneras
only
yeah.
N
So
if,
if
we
follow
it,
the
universe
end
before
right,
so
once
this
had
changed
on
the
policy
PC
will
notify
the
map
server.
Yes,
the
map
server
will
see
a
change
from
the
airlock
information
and
it
will
candy
pop
server.
It
can
do
pawns
half,
we
can
do
also
in
mobility.
Let's
say
that's
observe
so
it
sends
Marvin
notifies
to
whoever
has
request
that
information
in
the
past
right
or
has
been
subscribed
to
that
information.
N
G
Look
at
the
ID
mobility
draft
to
see
what
the
text
looks
like
to
seem.
It
generically
covers
this
case
without
being
specific,
mm-hmm.
Okay,
so
I
just
want
to
make
a
reference
to
Luigi's
question
and
listen
to
telemetry
talk
about.
Why
should
the
ETR
give
the
information
back
to
the
ITR?
This
is
a
good
example
here,
where
an
ITR
may
have
to
our
low
chrec
ERDs
to
either
the
same
our
local
different
ones.
G
But
the
point
is
as
if
there's
different
say
it's
for
the
same
one,
because
it's
kind
of
interesting
but
there's
different
segment
routes,
so
you
may
want
to
our
load
probe
twice
because
you're
going
on
a
different
path
and
then,
when
you
get
the
information
back,
you
want
the
ETR
to
tell
the
idea,
and
it
knows
with
path
that
is,
since
these
segments
routes
are
source
based
or
relative
to
the
ITR.
You
need
to
you.
Have
this
directionality
problem.
I
B
P
I
can
respond
on
that
officer
side
of
that
anyway,
you
can
I
think
it
was
at
the
front
of
this
document
where
you'd
mentioned
over
to
where
you
know
we
went
for
kind
of
the
whole
enchilada
right
off
the
bat.
We
need
to
break
it
down
into
into
what
happens
if
we
are
not
doing
te.
I.
Think
your
question
of
doing
lisping
cap
and
doing
s
our
traffic
engineering
as
well-
maybe
one
that
needs
to
be
looked
at
in
there
as
well.
We
could
we
could
expand
on
that,
because
I
think
you're
right.
G
Okay,
so
can
you
go
back
to
the
ABC
slider,
so
I'm
glad
that
Luigi
brought
this
up
because
it
begs
another
question:
why
can't
a
B
and
C
be
our
TRS
and
you
encapsulate
at
each
point
and
re
encapsulate,
and
if
you
do
that,
you
get
our
look
probing
between
each
components
of
this
source
route.
This
is
exactly
what
list
te
is
doing
with
LPS
right.
They
can.
I
G
So
one
argues
when
I
use
the
question:
why
do
you
need
segment
routing
v6?
Because
if
you
put
that
our
TR
topology
in
there
those
links
the
the
link
between
or
the
underlay
between
a
and
B
and
B
and
C
could
be
a
combination
of
ipv6
and
ipv4?
Now
I
know
the
answer
to
the
question,
but
I
just
want
I'm
glad
he
brought
it
off
because
I
didn't
want
to.
M
R
So
you
mentioned
an
interesting
scenario
where
you're
trying
to
basically
break
things
down
and
show
that
you
could
actually
have
SR
v6
with
Lisp
and
no
traffic
engineering,
and
in
that
case,
what
are
your
thoughts
on
basically
subsuming
the
pc
functionality
into
the
msmr?
Would
you
need
a
PC
at
that
point.
B
Actually,
my
name
is
second
question
kind
of
related
I
mean
why
PC?
Let
me
finish
so
you
you
have
a
two
models:
integrated
and,
and
it's
pretty
fine,
why
you
need
to
specify
PC,
because
you
can
just
describe
the
two
models
in
an
abstract
way
and
say:
yeah
I
need
the
traffic
engineering
information
I
either
the
mapping
system
does
it
for
me
all
I.
B
S
I'll
just
total
mic,
so
this
is
a
talk
that
I've
been
giving.
This
talk
to
the
last
couple
lists.
Working
groups
was
also
given
the
writing.
Research
rowdy
working
group
this
this
meeting,
a
simple
BGP
based
mobile
routing
system
for
the
aeronautical
telecommunications
network.
So
some
background,
the
International
Civil
Aviation
Organization
is
building
an
IP
based
networking
system
for
worldwide
air
traffic
management
and
for
those
that
don't
know,
ICAO
is
walking
distance
from
this
building.
S
Their
headquarters
were
just
down
the
road
from
here,
so
they're
developing,
what's
called
the
aeronautical
telecommunications
network
with
Internet
Protocol
services
or
etienne
IPs,
it's
a
new
development
in
the
ICAO
working
group
I,
which
is
like
the
IETF
versatile
aviation.
It's
an
ipv6
based
and
mobility
capable
and
it's
going
to
accommodate
aircraft
with
multiple
data
links.
So
an
aircraft
might
have
some
satellite
links.
All
Dax
is
a
cellular
link,
that's
being
developed,
currently
VHF
links
and
Inmarsat
iridium,
you
name
it.
S
It's
based
on
a
hub
and
spokes
arrangement
with
regionally
distributed,
stub,
ASPRS
and
centrally
located
core
ASPRS,
the
stub
autonomous
system,
boy,
routers,
advertising
withdraw
mobile
network
prefixes
that
belong
to
planes
and
the
core
autonomous
system
border
routers
form
a
hub
autonomous
system
between
two
four
packets
between
the
stub
autonomous
system,
where
routers
proxies
or
X
TRS
connect,
datalink
sub
networks
to
the
overlay
and
clients
that
are
aircraft
that
may
connect
multiple
sub
networks.
Then
route
optimization
removes
the
autonomous
system
where
routers
from
the
path.
S
So
what
it
looks
like
here
is
that
it's
think
of
it
in
terms
of
concentric
circles.
In
the
center
circle,
you
have
those
purple
routers
that
are
core
autonomous
system,
border
routers
that
are
within
the
same
autonomous
system
and
they
talk
use
an
eye
bgp
to
keep
full
topology
information
and
the
next
outer
circle.
You
have
those
green
routers,
those
are
the
stub
autonomous
system,
border
routers
and
they
connect
to
aircraft
mobile
network
prefixes
into
the
core
and
then
out
in
the
next
circle.
S
S
S
So
the
details
on
the
BGP
arrangement
here
is
that
each
the
Hanauma
system,
border
router,
is
a
stub
autonomous
system
unto
itself.
All
core
autonomous
system
border
routers
make
our
members
of
the
same
core
autonomous
system,
the
stub
it
advertised
their
associative
mobile
network
prefixes.
That's
the
prefix
is
owned
by
the
airplanes,
so
the
core
using
the
ebgp
and
the
core
autonomous
system
border
routers
at
originate
default,
but
they
did
not
advertise
any
mobile
network
prefixes
to
the
stub
autonomous
system
or
routers
by
the
court
system.
S
South
tunnel
system
border
routers
discover
all
mobile
network
prefixes
in
the
system
using
ibgp
and
they
can
connect
the
overlay
to
the
global
public
Internet,
in
which
case
they
would
advertise
only
a
short
aggregate
like
a
2001
DBA
:
font
size
32.
Instead
of
dumping.
All
the
airplane
mobile
network
prefix
is
into
the
BGP
routing
system.
We've
seen
in
past
experiments
that
that
doesn't
scale.
S
This
is
kind
of
reiterating
what
I've
already
said
about
the
hub
and
spokes
model,
and
this
is
the
diagram
that
comes
from
the
draft
so
changes
since
last
version.
We
removed
the
discussion
of
route
optimization
and
mobility
in
multi-link,
because
these
are
covered
under
other
drafts.
We
have
the
draft
that
that's
that's
the
Lisp
ground-based
Lisp
that
talks
about
one
proposal
there
and
I've
got
another
draft
on
arrow
that
talks
about
a
second
proposal,
but
this
draft
is
now
generic.
It
doesn't
apply
to
either
lisper
arrow
directly
and
either
one
can
be
fit
in.
S
We've
had
some
discussions
how
the
multiple
systems
can
come
into
play
in
different
partitions
of
the
network,
and
the
document
focuses
on
the
BGP,
topology
and
route
interactions,
and
it
emphasizes
that
in
earth.
Autonomous
system,
mobility,
events
and
quality
of
service
updates
are
not
propagated
into
the
BGP
routing
system.
So
that's
that's
pretty
much
it
it's
it's
a
it's
again.
It's
think
of
it
in
terms
of
the
concentric
reigns
where
you
have
your
mobility
and
your
multi
only
can
happen
in
the
outermost
ring
so
that
the
core
doesn't
get
affected.
By
that
you.
G
S
Think
that's
it.
Let
me
just
check.
Oh
okay,
here's
the
draft
again
it's
been
presented,
I
can
working
group
by
mobility
subgroup
and
it's
and
reactive
consideration
there
and
it
has
been
presented
in
the
routing
working
group
that
ITF
99,
101
and
also
one
or
two
is
just
presented
on
Monday
and
also
presented
at
the
list
working
group
at
IES,
101,
101
and
also
now
now
this
one
at
102,
so
I
think
that's
all
I
have
any
any
further
questions.
G
S
B
S
J
Brian
with
Nexus
actually
just
was
wondering
if
you
were
considering
using
the
planes,
is
sort
of
a
mesh
network,
so
they
could
relay
each
other's
packets.
That's
a
very
interesting
idea.
So
air-to-air
links
are
you
thinking.
S
B
T
Well,
nobody
it's
my
first
presentation
so
well
hope
it
goes
well
make
sure
that,
should
questions
get
comments,
let
me
know
anything.
I
may
not
be
seeing
correctly
on
this,
but
this
is.
This
is
essentially
something
that
you
know
what
I've
been
working
on
kind
of
speculate
on
for
a
while
a
little
background,
I
work
on
peer-to-peer
networks,
cryptographic
distributed
ledger
so
as
we've
kind
of
had
some
interesting
interactions,
figuring
out
how
we
can
come,
integrate
some
network
level
on
components
to
distribute
ledger.
T
So
that's
kind
of
where
the
berthing
place
is
the
decentralized
mapping
system
came
in.
Essentially,
if
you
have
a
period
of
your
network,
valued
in
hundreds
of
billions
of
dollars,
let's
say
Bitcoin
for
an
example
and
it's
running
over
Lisp,
you
need
to
basically
mitigate
central
points
of
failure
which
mapping
servers
can't
potentially
be
which
could
essentially
bring
down
the
entire
network
and
cause
significant
damages
monetarily.
T
So
what
if
lists
STRs
did
rely
on
pre-configured
map
servers
and
what,
if
map
servers,
could
also
allocate
to
their
own
shards
when
they
came
online?
Now?
What
if
the
abstraction
layers
on
top
of
the
OSI
could
actually
improve?
Key
management
and
authorization
in
map
servers
and
what,
if
we
could
mitigate
map
server,
DDoS
attacks,
the
partial
free
image
collisions
of
zero
bits.
T
So
what
this
basically
means
is
we
can
distribute
the
X
TRS
and
the
map
servers
and
make
them
auto
allocating
as
the
type
of
distributed
consensus
mechanism
so
that
they
can
allocate
specific
shards
within
one
another
and
then
from
that
they
can
create
a
higher
level
of
reliability.
Now
this
is
gonna
mitigate
DDoS
attacks.
I'll
get
to
that
about
the
partial
pre
image
collisions.
T
That's
essentially
finding
a
number
of
zero
bits
that
are
partial,
pre-minced
collision
on
any
type
of
hash
with
Iran,
and
once
the
little
sensor
would
be
a
proof-of-work
that
would
create
an
asymmetric
type
situation,
which
means
that
the
requester
would
require
to
fulfill
thousands
more
hashes
than
the
actual
verify
or
the
map
server,
which
will
may
be
required
to
fulfill
one.
So
you
could
draw
packets
and
I'll
get
to
that
later,
so
I'm
going
on
a
tangent,
so
starting
a
map
server.
T
So
what
if
each
sex
er
was
a
map
server
allocated
to
a
DHT
sharra
to
a
distributed
hash
table
and
what,
if
eat,
sexy?
Our
could
actually
map
register
to
each
X?
Y
are
based
on
deterministic
modulus.
So
essentially,
what
you
could
do
is
every
time
a
new
map.
Server
came
online,
they
could
replicate
one
another
and
then
they
would
be
acting
through
a
shared
type
of
address
registry
that
will
get
to
that
with
the
DNS,
which
essentially
would
be
able
to
create
an
Alok
eight
new
shards.
T
So
it
would
in
a
sense,
share
some
of
the
cache
and
new
notes
coming
on
that
are
aware
of
new
map.
Servers
would
essentially
be
able
to
start
asking
or
talking
to
these
in
charge
within
caches,
and
then
the
map
servers
would
eventually
cache
and
timeout,
since
some
of
the
records
would
go,
and
so
it
would
automatically
be
able
to
locate
and
create
shard
4506
as
needed.
So
this
would
be
something
that
would
be
determined
by
the
map
servers.
T
It
would
probably
require
a
little
OSHA
machinery,
but
it
would
allow
them
to
essentially
Auto
synchronize
and
Auto
manage
themselves,
so
the
DDoS
protection
we
always
have
the
needle
in
a
haystack
problem
with
the
us
attacks.
What
do
we
do
if
a
central
map
server
goes
down
under
it?
It
is
attacked
me.
Does
this
the
network?
Now?
What?
T
If
the
server
low
determines
the
location
on
the
DHT
charge
or
map
servers,
and
also
access
to
require
requesting
party
to
compute
a
proof-of-work
by
iterating
and
wants
to
find
a
value
that
meets
a
member
of
zero
bits
making
require
work
for
the
requester
if
the
server
happens
to
be
under
heavy
low?
So
if
you
have
some
sort
of
minimum
threshold
that
the
map
servers
are
located
near
shards
at
let's
say,
30
percent
over
a
moving
average
window
you'll
be
able
to
reasonably
determine
if
the
server
load
goes
above,
50
or
60
percent.
T
That
you're
under
some
form
of
DDoS
attack,
which
you
can
throttle
requests
from
all
parties
by
requiring,
let's
say
100,
milliseconds
or
500
milliseconds
of
actual
proof
of
work,
which
would
be
easy
to
verify.
On
the
map.
Server
side
it
would
slow
down
request
the
requesters
side
and
it
would
allow
the
map
server
to
resolve
that
load
and
reduce
it,
which
would
improve
the
actual
accessibility
for
the
end
user.
T
Now
the
downside
for
the
user
is,
they
would
require
computing
cycles
if
the
server
did
get
under
heavier
load,
but
if
you're
on
ddos
attack
and
the
service
completely
down,
that's
completely
service
outage.
This
would
just
be
a
slow
the
service
which
would
require
you
to
essentially
pay
your
dues
to
get
access
to
the
system
so
how
to
start
the
mapping
system
so
as
Joel
brought
up
last
time,
we
went
over
the
lifts
decentralized
draft.
There
was
a
bootstrapping
problem,
so
the
DNS
record
would
simply
be
something
that
is
pre-configured.
T
They
would
talk
to
a
DNS
type
of
map
server.
That
would
just
be
keeping
a
running
record
of
all
of
the
actual
distributed
map
servers.
So
this
would
be
the
way
that
you
get
through
the
bootstrap.
You
can
figure
that
in
and
then
the
DNS
server
is
essentially
what
you
get
the
a
record
for
which
then
you
talk
to
it
over
a
predefined
message
set
and
ask
it
directly
back
for
the
list
of
money,
map,
servers
or
shards
and
then
you'll
know
exactly
ones.
Now.
T
If
a
shard
is
in
bootstrapping
phase,
which
could
be
a
24-hour
phase,
creating,
let's
say
shard
4,
then
you
would
be
essentially
talking
to
modulus
3
in
modulus,
for
in
order
to
bootstrap
and
help
that
map
cache
fill
up
to
eventually
get
through
that
bootstrapping
period
to
where
the
new
char
can
allocate.
So
the
exit
er's
that
are
part
of
the
mapping
system
resolved
the
first
DNS
records
to
obtain
the
initial
DNS,
see
the
send
resolves
return.
T
T
So
the
DDoS
request
throttling
one
of
the
map
servers,
can
essentially
use
the
AED
as
a
part
of
the
throttling
mechanism
to
do
the
individual.
Now,
when
we
have
a
DDoS
attack,
we
have
a
problem
where
somebody
could
be
essentially
manipulating
the
source,
the
destination
and
completely
spoofing
all
the
IPS,
the
problem
that
we
can
have
there's
the
needle
in
the
haystack
problem.
T
So
we
resolve
that
through
the
CPU
collective
usage
and
use
that
to
essentially
say
well,
if
you're
making
me
work
more
I'm
going
to
you
work
a
little
bit
more
and
then
that's
the
global
to
do
the
DDoS.
Now,
if
it's
just
a
DOS
attack,
then
you
could
easily
throttle
that
by
saying,
if
you're,
making
too
many
requests
per
second,
then
we're
gonna
slowly
increase
the
amount
of
zero
bits
that
you're
required
to
do,
which
is
going
to
increase
the
computational
requirements
for
you
to
send
a
message.
F
There
seem
to
be
mixed
did
at
least
two
different
sets
of
requests
in
this
which
come
from
different
parties
and
I'm
a
little
confused
about
that.
There's
map
registers,
which
come
from
other
etrs
while
waiting.
The
ETR
is
one
of
your
DHT
elements,
so
it
wouldn't
be
sending
map
registers.
So
there
probably
are
no
map
registers
and
their
map
requests
which
come
from
ITRs
to
map
requesters,
not
map
servers,
okay,.
U
F
Now,
who's
going
to
prove
work,
because
the
request
has
to
go
across
the
the
DHT
to
the
guy.
Who
has
the
authoritative
thing
he's
going
to
ask
somebody
to
prove
work?
But
it
looks
like
he's
asking
the
trusted
party
to
end
up
doing
work,
because
the
trusted
party
is
getting
besieged
with
requests,
which
he
has
no
way
to
ask
for
proof
of
work.
T
F
That
doesn't
answer
the
question
at
all.
My
point
is
map
requests
come
from
ITRs,
yes
to
the
edge
of
the
mapping
system,
those
don't
you!
You
haven't
talked
about
those
entries
entities,
a
missed
slide
at
all,
so
obviously,
apparently
those
aren't
asking
for
proof
of
work.
So
the
proof
of
work
is
coming
from
the
guy.
Who
has
the
authoritative
information
well
by
the
time
you've
gotten
it
to
him.
T
F
T
G
F
Is
a
major
change
not
to
the
mapping
system,
but
to
the
interface
which
we
have
defined
as
stable?
Now
we
can
evolve
it,
but
it's
that's.
What
I
was
trying
to
get
out
is
the
difference
between
mapping
system
operation.
We've
structured
this
so
that
putting
in
a
new
mapping
system
is
transparent
to
the
external
interfaces,
yes
and
changes
to
those
external
interfaces.
It
looks
like
this
takes
changes
to
those
external
interfaces
right.
T
T
T
G
So
the
idea
would
be
Joel
that
the
ETR
would
do
the
proof
of
work
and
the
hash
would
be
the
hash
doesn't
have
to
be
sent
in
the
map
register.
Only
the
nonce
does
that
nonce
we
already
sent
today
now.
What
is
said
on
the
other
side
is
that
the
guy
takes
the
nonce
hashes
the
packet,
and
then
he
he
gets
the
hash
needles.
If
he's
good
or
bad.
Now,
as
you
want
to
change
the
level
of
difficulty,
we
know
we
have
a
map
notify
that's
an
acknowledgement
for
the
map
register.
G
V
R
R
T
Okay,
so
I
apologize
for
I'll
do
better
next
time,
but
okay,
so
anyhow,
so
then
we
essentially
what
you
can
do
is
you
can
have
certain
types
of
attacks?
Let's
say
EC,
they
said
verify
unless
crypto
that
would
be
done
through
a
CPU
exhaustion
attacks.
So
by
being
able
to
throttle
and
create
the
asymmetric
proof
or
work,
you
would
be
able
to
essentially
require
the
end
party.
That
means
doing
the
requests
or
any
sort
to
basically
require
to
do
a
proof
of
work
which
would
throttle
them
now.
The
benefits
are
the
XE.
T
Ours
only
depend
on
each
other
and
today
already
if
they
want
to
talk
to
each
other.
Now
no
third-party
trust
or
dependency
exists.
Master
Cup,
locos
has
low
latency
and
map
servers
have
McDonough
C
and
scalability.
Those
are
important
components
to
get
into
a
completely
centralized
environment
to
utilize
lisp,
with
something
such
as
a
distributed
ledger.
T
So
the
distributed
ledger
is
an
abstraction
layer
on
top
of
lisp
or
on
top
of
the
LSI
reference
stack
can
actually
handle
the
removal
of
need
of
third
parties,
such
as
key
management,
authorization,
access
control
schemes,
since
the
distributed
ledger
is
actually
a
very
secure
data
layer
that
is
built
through
a
consensus
mechanisms
globally.
There's
a
lot
of
things
that
you
can
actually
be
implementing
on
list.
T
So
the
prerequisites
defined
in
this
document
are
for
their
prerequisites
of
line
to
make
the
decentralized
mapping
system
function
in
a
case
that
would
eventually
allow
a
distributed
ledger
to
sit
on
top
of
it,
which
also
allows
you
to
create
reputation,
systems,
mathematically
verified
on
Eid,
so
that
you
know
exactly
who
you're
talking
to
and
then
you
can
have
certain.
Let's
say:
selection
bias
from
the
list.
Selamat
read
data
on
the
distributed
ledger
and
vice
versa.
T
F
Sounds
like
there's
a
real
assumption
here.
That's
very
important
and
valid
for
some
of
the
use
cases
you're
describing
but
I
want
to
make
sure
I'm
understanding
you.
Your
assumption
seems
to
be
that
this
is
an
environment
in
which
not
only
are
the
xtr
highly
distributed,
but
any
X
T.
It
is
perfectly
okay
for
any
X
TR
to
claim
to
be
serving
any
ID
as
long
as
nobody
else
is
claiming
that
Eid
or
Eid
prefix
know.
T
When
you
have
to
distribute
a
ledger
on
top
of
that,
we
actually
create
a
reputation
system.
The
series
of
events,
so
that
somebody's
claim
is
not
just
a
claim
out
in
the
open
abyss.
You
actually
have
a
masterful
verifiable
trail
of
somebody's
trustworthiness
so
that
you
can
have
a
distributed
trustless
system
by
mathematical,
verification.
F
T
F
F
I
F
T
R
F
T
T
G
O
G
T
There's
a
lot
of
implications
on
it
and
I
mean
I
believe
that
the
OSI
as
we
have
it,
works
to
a
point,
but
the
need
for
so
many
third
parties
is
kind
of
showing
that
there
needs
to
be
more
layers
under
that
stack
and
I.
Think
the
distributed
ledger
is
one
of
the
layers
that
could
sit,
possibly
even
in
the
applications
presentation
of
the
OSI
and
then
so
on
and
so
forth,
create
kind
of
the
auto
executing
environments
and
tie
financial
instruments
into
it.
I
W
If
they're
following
their
intent,
try
that
the
intent
is
to
get
away
from
the
relying
on
the
third-party
elements
as
much
as
possible,
but
then
you
still
have
to
rely
on
us
for
the
seeding
processes.
I
guess
right
so
does
that
somewhat
reduce
the
efficacy
of
the
world
solution
mean
here
the
way
we
wanted
to
be
centralize
it,
but
we
are
going
back
to
the
third
party,
if
not.
T
Necessarily
if
you
were
passed
an
IP
address
of
your
friend
that
was
already
connected
to
the
network,
and
they
said
you
can
use
my
IP
address,
you
can
hard
code
that
directly
inter
put
in
your
configuration
file
and
it
would
function
the
same
way.
All
it's
doing
is
functioning
to
give
you
the
gateway
and
the
network
once
you
get
in
the
network,
address
discovery
happens
automatically,
so
the
DNS
is
kind
of
a
convenient
user
way
in
order
to
get
access
to
the
network.
T
T
That's
fine
generally
you're,
not
gonna
want
to
rely
on
one
domain
name
so
and
Nexus.
What
we
do
right
now
is.
We
have
a
set
of
twenty
or
thirty
different
DNS
records.
That
act
is
the
first
nodes
that
a
node
ever
talks
to
is
then
gets
the
seeds
of
all
the
other
nodes
that
are
accessible
and
then
it
goes
and
it
starts
connecting.
So
you
do
that
through
redundancy
by
having
multiple
DNS
records.
T
T
Possibility
this
malicious
and
that's
the
beautiful
thing
about
distributed
consensus.
Is
you
don't
have
to
just
talk
to
one
node?
You
generally
talk
to
multiple
ones,
so
if
they
do
give
you
a
list
of
malicious
addresses,
that's
why
you
do
the
full
validation
of
the
ledger
on
your
local
node
and
verify
all
the
cryptography.
So
even
if
they
try
to
send
you
malicious
data,
you're
gonna
be
able
to
detect
it
and
it's
gonna
be
about
assuming.
T
I
T
Requires
an
incredible
computing
power
through
proof-of-work
and
also
other
mechanisms
to
proof
of
stake,
validation
processes
to
actually
make
those
cryptographic
blocks.
So
you,
wouldn't
it
wouldn't
matter.
If
you
talk
to
an
invalid
land,
they
could
give
you
a
bad
blotching
or
a
bad
set
of
blocks
and
you're
gonna
be
able
to
verify
with
your
neighbor
the
correct.
Thank.
G
So
I
asked
III
look
I
mean
many
people
have
asked
this
question
before:
what's
the
decentralized
internet
and
I
like
to
think
of
it
of
a
way
is
if
you
use
DNS,
are
you
depending
on
a
third
party?
Well,
you
may
be
depending
on
a
third
party,
but
you
could
still
be
decentralized
because
it
depends
where
the
third
party
is
and
who's
managing
it.
So
I
like
to
think
of
it.
G
This
way,
if
we
use
something
like
list
decent
and
every
xtr
is
a
mapping
system
or
a
map
server
of
the
same
mapping
system,
I
like
to
think
of
it
as
a
neighborhood
of
devices
that
have
local
connectivity,
that's
an
IP
packets,
each
other.
They
want
to
be
self-sufficient.
They
want
to
run
themselves.
They
have
no
connections
to
the
Internet,
but
maybe
if
you
don't
have
multicast,
which
is,
if
you
have
multicast
locally,
then
you
can
do
everything
dynamically.
Just
like
we
see
how
a
bonjour
works
in
our
house
with
Macs
and
stuff
like
that.
G
But
if
you
don't
have
that
multi
guess
for
some
reason
you
can
still
depend
on
dns,
but
the
DNS
is
all
local
and
trusted.
Among
that
trust
group
in
the
in
the
in
that
neighborhood.
So
even
though
you're,
depending
on
this
protocol,
that
looks
centralized
it's
being
deployed
in
a
decentralized
trusted
environment
and.
F
Mitigate
my
far
as
I
know,
none
of
the
arguments
for
our
deployments
we
describe
for
a
lisp
match
what
you
just
described.
Email
agree.
They
tend
to
be
scattered
around
the
internet,
not
locally
connected
or
they
tend
to
be
under
if
they're
serving
a
single
data
center,
which
is
the
opposite
case,
where
they
are
clustered
they're
under
a
single
administrative
control.
And
so
you
don't
have
a
distributed.
Trust
problem
right.
I
F
T
Supply
chain
management
and
auditability
and
immutability
of
the
data.
So
that's
where
you
create
trust
where
it's
mathematics,
it
becomes
the
verify
of
truth.
So
in
a
sense
you
could
call
it
blockchain
or
distributed
ledger
a
truth
layer
of
the
Internet,
which
is
something
that
we
don't
really
have
now.
The
truth
is
determined
by
global
consensus
and
the
more
people
you
have
validating
and
that
consensus
the
more
you
could
consider
that
truth.
So
it
essentially
gives
you
just
layers
of
abstraction
that
help
give
an
improved
frame
of
reference.
T
I
T
Why
do
centralized?
Well,
it's
important
the
Internet,
the
IETF
was
designed
to
be
to
decentralize.
The
internet
was
designed
to
be
decentralized,
it
makes
it
more
robust
and
it
removes
it
from
central
points
of
failure
and
also
four
central
points
of
control.
So
peer-to-peer
networks
are
we
seen
their
star
forces
and
they've
proven
high
levels
of
robustness.
You
have
100%
uptime
and
you
don't
have
to
have
let's
say
any
any
point:
essential
failure.
No,
you
always
have
fall
backs
if
a
map
server
gets
put
under
a
DDoS
attack.
T
Somebody
is
somebody
and
they
can
prove
who
they
are
by
signing
from
the
specific
key
I'm
a
signature
chain
and
I
can
look
up
not
just
from
that.
Eid
I
can
look
up
their
whole
track
record
if
they
validate
it.
If
they've
had
any
spends,
if
they've
created
this
a
pet
loans,
a
bad
escrow
contracts
that
failed
and
so
on
and
so
forth,
so
it
just
gives
you
a
better
auditable
trail
of
events
that
you
can
use
to
create
better
selection
bias,
whether
from
yeti
user
map
servers
or
anything.
So
my
questions,
actions.
G
I
just
want
to
add
it
to
your
last
point.
Yes,
so
the
idea
here
is
that
crypto
e
IDs
give
you
identity
at
the
network.
Layer
and
cryptocurrency
wallets,
give
you
this
identity
that
can
that
is
very
changeable,
because
you
have
different
set
of
parameters
when
you
want
to
change
it
at
the
application
layer.
The
question
is:
is:
can
the
security
of
this
application
layer
help
the
network
being
more
secure
and
vice
versa,
yeah.
T
I
B
G
B
Are
not
unlocking
the
best
documents
because
there
is
no
reference
to
the
document
which
is
not
existing.
I
sent
an
email
to
you
to
you,
I
bet
about
where
we
are
with
this
dog
meat.
So
it's
up
to
you
to
answer
the
question,
so
you
you
personally
said
I,
don't
want
a
variety
because
I
I
don't
think
it's
the
right
solution,
so
I
don't
care
basically
yeah!
Oh
I,.
B
B
G
F
F
F
I
F
It's
I
believe
it's
a
chartered
work
item,
but
it's
not
doesn't
have
to
be
part
of
getting
to
PS.
We
want
it,
that's
not
the
question.
We
didn't
think
it
will
fit
well
where
it
was,
and
so
that's
not
where
it
is
and
frankly,
if
nobody
cares
to
put
in
the
work
to
put
it
back
together.
That
says
something
about
the
work
groups.
Actual
interest
in
that
work.
That's
a
problem,
but
it's
a
different
problem.
G
F
I
Let
me
try:
let's
try.
Let
me
try
to
have
a
different
perspective.
Okay,
I
think
at
that
text
we
seem
the
right
context,
hot
value
and
the
right
context
was
the
the
rest
of
the
text,
the
rest
of
their
absolute.
Now.
What
you're
telling
us
is,
you
can
take
that
debt
and
put
it
somewhere
else,
but
I
think
that
when
you
do
that
it
loses
the
value.
So
it's
not
that
no
one
cares
about
that
text.
It
is
not
true.
That
seems
no
one
is
willing
to
do
that.