►
From YouTube: IETF102-HACKATHON-20180715-1400
Description
HACKATHON Reports and Awards at IETF102
2018/07/15 1400
https://datatracker.ietf.org/meeting/102/proceedings/
A
A
B
Okay,
I
promised
in
30
seconds
I,
don't
know
if
I
can
make
that,
but
they'll
be
pretty
gun
close.
This
is
the
network
time
security
team.
We
can.
Those
of
you
been
around
the
IHF
awhile
we've
been
working
on
network
time
security
for
a
while.
Our
plan
specifically
was
to
improve
the
status
of
the
implementations
and
particularly
we're
looking
for
a
current
draft
12.
We
have
a
draft
Snorri
in
yes,
which
we
were
also
looking
at,
which
is
provide
providing
some
suggested
changes,
and
we
were
looking
at
some
interoperability.
B
We
had
sort
of
a.
We
did
a
really
good
job
of
remote
this
time.
Actually
we
had
a
we
had
G
mem,
which
was
Martin
and
his
implementation.
In
Germany
we
had
Daniel's
Python
implementation,
which
was
here
we
had
a
team
in
Sweden
that
was
working
on
an
implementation,
that's
in
hardware,
and
then
we
had
some
folks
from
the
network,
time
foundation
here
and
then.
B
Finally,
we
had
a
gentleman
who
showed
up
and
wanted
to
work
on
the
open,
ntpd
implementation,
which
was
very
exciting
so
on
the
left,
is
the
various
team
members
that
we
had
and
on
the
right
is
their
various
pieces,
both
the
documents
that
we
were
looking
at
and
also
some
of
it
code
that
was
used.
These
are
our
three
teams
and
the
pieces
that
we
had
scattered
in
various
places.
B
This
is
basically
what
we
got
done,
so
you
see
the
draft
that
it
currently
is
referencing
the
language
that
we
were
used
and
some
of
the
notes
associated
with
it.
And
finally,
this
is
a
chart
of
what
you
know
basically
kind
of
what
we
learned
and
so
across
the
top
or
some
of
the
server
implementations
that
we
have
down.
The
bottom
is
the
down
the
access.
The
vertical
access
are
the
clients.
These
are
some
of
the
pieces
that
worked
and
didn't
work.
The
bottom
line
is
that
interoperability
tests
are
really
important.
B
I
guess
we
all
know
that
the
NOC
is
very
helpful.
We
actually
turned
into
the
problem
child
for
the
NOC
yesterday,
which
is
generally
not
my
objective
and
that
wire
shock
really
helps.
So
that's
about
it.
It's
network
time,
security,
I
know
it's
not
the
newest
and
greatest,
but
it
is
really
important.
We
would
love
to
have
some
additional
help
in
the
future.
C
The
challenge
for
us
was
to
keep
the
project
as
simple
as
possible
and
focus
on
the
implementation
of
the
service
itself
and
not
get
bogged
down
into
the
network
implementation.
So
to
do
that,
we
essentially
took
four
switches
in
a
switch
network.
We
connected
them
up
so
that
we
could
realize
the
land
service
rather
than
trying
to
do
this
with
either
an
MPLS
or
anything
circle.
C
So,
in
trying
to
implement,
we
got
some
headway
into
getting
at
least
the
all
port
baseline
service
going
here.
This
is
the
EP
lancers.
Now
with
EVP
lan,
we
had
some
challenges
and
I'll
get
into
it
in
a
minute.
The
latest
set
of
code,
that
is
there,
will
be
updated
in
this
particular
github
location.
C
C
Vlans
can
be
defined
in
either
of
one
of
those
formats,
so
we
need
a
way
to
be
able
to
specify
a
range
or
set
of
values,
but
have
to
say
that
from
a
yang
perspective,
it's
a
nice
to
have
I,
don't
know
how
open
that
idea
would
be
to
reduce-
and
you
know
just
for
that
from
programmatically.
You
can
always
have
multiple
iterations.
E
Hello,
this
is
a
team
photo
from
Tulane
University,
the
I-20
sf4
framework
project.
If
your
pieces
are
tips
time
and
I
kept
the
person
project,
this
project
is
for.
We
want
to
bec
verify
the
frame
over
by
to
NSF
and
interfaces
based
on
young
Jeremy
them
like
this,
and
also
this
time.
Also,
this
is
a
student
of
project,
seven
great
student,
changing
band
to
grad
student
from
Chilton,
oh
yeah,
and
also
sleep
professors
and
researchers
and
Korea
and
Japan.
E
So
this
is
a
poster
for
this
as
well.
But
this
time
are,
we
approved
three
things.
The
first
one
is
a
neck
comm
supported
for
restoration
interface
for
I
to
NSF
framework
based
on
latest
the
capability
young
parameter.
The
second
one
is
I
to
NSF
scripted
policy
translation.
So
translation
is
very
important
because,
however,
a
policy
translated
into
more
about
skill
follows
automatically
another.
One
is
automatic
policy,
provisioning,
keepin,
a
low
level
policy
we
match
into
that
ability
of
functions
to
certify
out
required
the
lot
about
Curie
policy.
E
So
this
is
a
first
future
racial
age
interface.
You
can
see
the
holla
policy
given
to
skillet
controller.
It
request
are
the
higher
the
NSF
and
then
the
development
team
on
the
system
in
in
bulk
and
then
registered
into
spirit
policy,
and
then
we
can
help
certify
holiday
policy
using
appropriate
retrospective
our
functions.
So
this
is
a
theorem
level
and
also
the
second
one
is
the
security
of
policy
translation.
So
using
this
box,
this
is
a
translator.
So
high-level
policy
is
in
terms
of
our
audit
automatization
over
user.
E
That
one
is
a
translated
ease
of
procedure
and
their
low
level.
School
policy
needs
to
translate
to
the
righties.
So
whatever
policy
is
given,
this
one
is
a
decomposing
into
NSF
for
our
by
work
and
the
second
one
is
an
asset
for
Victor,
so
we
can
match
our
wait.
A
lot
ever
ask
a
policy
into
appropriate
to
NSF
conscience.
E
So
this
is
dr.
comparation.
We
use
the
mini
nap,
so
in
the
middle
we
have,
we
can
say
security
controller
for
RIT
to
another
framework
and
that
we
have
se
a
network
and
we
have
a
pact
with
NSF.
So
this
part
we
have
a
to
application.
What
is
a
malicious,
our
website,
the
filtering,
the
second
one
is
knowledge
as
the
email
to
trade,
so
we
use
the
open
source,
so
we
unloaded
the
frame
of
source
coding
book
it
up.
E
So
you
can
access
and
download
our
project
code
and
also
we
uploaded
our
cancer
temptation,
a
video
into
YouTube,
you
can
access
and
then
we
can
check
it
out.
So
are
we
learned
during
this
hackathon?
We
proved
the
youngbae's
registration
interface,
a
working
world
to
provide
automatic
security,
our
legislation
and
also
the
second
one-
is
security
policy,
translation
very
important
for
I
to
NSF
demo.
So
we
prove
using
our
custom
project
so
also
we
showed
a
based
on
pure
applesauce.
We
can
make
iTunes
sfm.
Oh
thank
you
for
your
listening.
Thank
you.
G
Other
one
is
a
microphone
for
weight
and
today
I
want
to
present
Jackson
projects
cast.
Okay,
our
motivation
is
to
pull
wide
it's
at
this
new
protocol,
which
we
defined
in
the
as
following
links.
It
can
works
real
and
it
can
validate
and
performance
to
satisfy
that
operators
requirement.
It
can't
build
it
separating
the
MJ's
architectures,
and
we
want
to
of
here
with
the
bun
to
benchmark
wanting
to
want
to
verify
that
the
king
children
can
communicate
to
the
interpreting
students.
What
was
its
new
Kotoko's?
G
G
It
can
look
real
to
lose
it
that
the
wheel
and
this
centralized
future
pink
and
distributed
user
user.
So
it
is
what
we
land
satisfied
and
the
perfect
at
the
nuclear
codes
worked,
feel
and
we
believe
eight.
Maybe
ten,
Dreiser
operators
with
Carmen
and
immediate
need
need
to
send
Rodrik
okay
here
and
once
you
send
our
team,
we
have
very
large
and
emergency
teams
who
are
hard
work
and
very
I'm
killed.
Some
Bella
building
in
Hilton
contributions,
and
especially
I,
want
to
sensory.
G
The
remote
participation
team,
who
help
us
to
assimilate
eight,
who
have
a
set
has
a
very,
very
rare
mention
and
also
I
want
to
mention
here-
is
that
intranet
violate
not
the
still
online.
Just
photons
support
us
even
Zell
today's
workout,
the
finals,
but
they
don't
heed
us,
don't
see
it
magic,
say:
condoms
photos
as
I'd
like
to
sense.
It
were
much
okay.
I
also
I
want
to
remember
that
if
we
will
show
our
demo
in
tomorrow,
Jackson
happy
hour,
and
if
you
have
interested
in
this
topic,
please
to
join
us.
Thank
you.
A
H
What
did
we
achieve?
So
we
have
two
running
implementations
insular
and
pilar.
We've
also
realized
that
we
want
to
go
through
the
formal
verification
of
that
protocol,
so
we're
doing
that.
Also,
the
link
to
the
github
on
different
key
servers
in
the
pilot
in
the
sealer
implementation.
What
else
we
did?
We
also
were
able
to
do
this,
a
CDHP,
implement
implementation
in
dialogue
and
also
some
updates
to
the
draft,
such
as
disabling
the
non
secure
configurations.
H
What
else?
What
did
we
learn
here?
One
of
the
key
things
that
we're
looking
to
do
is
to
be
able
to
integrate
with
the
nginx
OpenSSL
and
also
look
looking
for
the
trusted
environment
as
well
implementation
there.
We
also
look
to
implement
work.
Extensions
for
TLS,
1.2,
full
implementation
and
also
future
plans
include
for
us
in
the
next
hackathon
we're
looking
to
do
formal
verification
for
the
TLS
for
the
extension
for
the
extension
concealer
TLS
1.3.
H
F
I
Finding
some
color,
so
what
would
you
do?
We
did
something
called
drift,
which
was
the
first
time
we
had
actually
hecatomb
is
relatively
new.
Things
would
work
much
better
than
expected.
We
go
much
harder.
Actually,
who
was
here
about
five
people
were
hacking
code.
We
know
the
three
radical,
probably
the
most.
A
couple
of
people
have
challenges
no
cookies
in
the
company,
and
so
what
is
rift
so
rift
is
a
new
specialized
routing
protocol
just
for
IP
fabrics,
especially
data
centers.
I
It's
a
mixture
of
links
that
in
distance
vector
protocol,
if
it
means
anything,
it's
a
fairly
novel.
It's
completely
schema
based,
which
is
kind
of
you-
know
radical
purakh
in
protocol.
In
idea,
what
did
we
try
to
intend
to
build
in
this
hackathon?
We
try
to
get
the
Python
implementation,
the
first
piece,
which
is
basically
adjacency
formation,
which
is
quite
complicated
in
Python
to
interoperate
with
something
which
is
already
available
from
a
vendor,
though.
I
Basically
two
packages
packages
having
run
into
no
an
interoperable
test
and
bring
up
in
adjacency
and
then
see
where
we
can
get
further.
So
this
is
basically
the
output
of
the
adjacency.
We
got
up
from
the
Python
site
and
the
interesting
things
went
to
three-way,
which
means,
when
tore
the
whole
finite,
set
machine
and
exchanged
lot
of
information
and-
and
you
know,
arrived
where
we
wanted
it
to
be-
and
there's
a
lot
of
gory
details
involved.
I
One
because
you
know
Brune
already
need
a
good
amount
of
work
upfront
a
second,
because
we
found
that
all
this
schema
based
modeling
is
actually
saving
like
80
percent
protocol
work.
That
is
being
done
normally
like
know,
writing,
parsers
and
all
kind
of
quirky
code,
which
is
all
hand
woven
and
the
only
problem
we
really
found.
So
the
specs
seems
to
have
held
up
pretty
well.
The
only
problem
found
that
we
were
joining
multicast,
basically
on
different
interfaces,
no
on
the
laptop,
so
we
didn't
see
each
other.
I
At
the
moment
we
joined
the
same
interface
and
stuff.
Just
came
up,
so
we
ran
out
of
scope
of
three
hours.
So
then
we
attack
the
next
finite
state
machine
which
is
far
more
complicated
and
it's
basically
the
finite
state
machine
which
allows
a
complete
zero
touch
provisioning.
So
this
IP
fabrics
of
this
specific
property
that
it
is
highly
desirable
that
there's
simply
no
configuration
whatsoever,
maybe
have
to
be
there,
albeit
there
are
proper
doing
proper
IP
forwarding
to
be
there
behaving
like
switches.
I
So
you
do
know
before
our
v6
and
all
the
stuff
is
included
and
they
have
to
exchange
a
lot
of
information
to
find.
Where
are
they
are
to
follow
G
wise
in
the
fabric
and
there
we
basically
wrote
a
lot
of
code
being
not
too
distracted
by
you
know
the
ongoing
soccer
game
I
think
that
completely
killed
our
productivity
for
a
while
and
we
ended
up.
You
know
with
very
serious
amount
of
comments
on
the
models.
I
You
know
certain
quirks
where
the
model
wasn't
completely
orthogonal,
where
it
wasn't
clearly
specified
some
finance
and
machine
bands
being
better,
no
specified
rough
procedures.
I
mean
the
spec
is
like
70
pages
and
growing.
It's
a
food
routing
protocols.
It's
quite
quite
a
sizable
thing.
Yes,
actually
we
got
much
more
out
of
the
circuit
on
that
we
did.
We
expected
you
know
yeah.
J
Okay,
so
this
is
presentation
of
the
spin.
So
for
those
who
don't
know,
there's
been
bit
it's
a
post
mechanism
that
will
allow
passive
observer
to
do
measurements
of
round-trip
time
of
quick
clothes
by
observing
in
the
middle,
using
a
single
paper
for
me
so
yep,
the
relevant
drafts
are
basically
all
quick
drafts.
So
what
we've
been
focusing
on
in
this
hackathon
is,
if
we
can
use
heuristics
to
get
more
robust
measurements
using
this
single
bit.
J
So
what
got
done
was
that
we
took
a
crappy
mailbox
that
we
had
started
implementing
last
time
and
included,
so
we
could
handle
multiple
parallel
connections.
We
couldn't
we
implemented
new
heuristics
to
reject
bad
art,
these
samples
and
we
also
improved
our
met
measurement
system.
So
we
couldn't
include
a
sequence
of
Nathan
links
where
we
could
apply
different
impairments
and
different
latencies,
especially
looking
at
delay
and
reordering.
So
we
could
try
a
lot
of
different
conditions.
J
So
if,
if
impairments
occur
below
the
point
of
measurement,
then
they
have
virtually
no
effect
on
the
how
well
they
spin
bit
works
if
they're
above
the
aspect
is
been
bitten,
there's
where
we
need
to
apply
our
heuristics,
we
also
looked
at
the
effect
of
the
the
sample
frequency
comparing
the
this
spin
bit
sample
frequency,
which
is
just
one
spur
RTT
with
with
the
sample
frequency
that
the
server
has,
which
is
basically
for
every
AK.
We
could
see
that
it
had
basically
no
effect
on
on
the
estimate.
J
The
arti
it's
just
some
little
result
we
got
here
is
this
is
an
example
of
the
results
we
got.
This
is
measuring
a
5
percent
reordering
happening
above
the
spin
bit.
It's
been
that
measurement
point
we
track
here.
There
are
TT,
as
observed
by
the
spin
bit
of
server
and
the
server
at
full
frequency,
where
the
RTT
is
collected
for
every
AK
and
the
server
collecting
the
RTT
measurements.
J
A
K
So
how
do
we
build
a
global
IOT
out
of
all
these
parts
and
there
are
already
several
stos
out
there
that
have
found
out
that
maybe
they
want
to
be
the
glue
that
binds
all
the
other
stos
together,
and
so
we
decided
we
are
going
to
be
the
solvent
the
sooner
that
goes
into
all
that
group.
So
this
is
essentially
what
we
are
trying
to
do.
K
So
there
are
lots
of
existing
data
models
like
lightweight
m2m
and
so
and
ocf
and
so
on,
and
there
are
also
Semantic
Web
technologies
that
actually
turn
out
to
be
useful.
This
time,
and
so
we
don't
have
several
forms
of
self
description
once
defined
here
in
the
ITF
like
RFC,
66
90,
or
the
thing
description
that
is
being
done
by
w3c.
K
Where
are
things,
and
we
have
various
forms
of
directories
where
self
description
information
can
be
stalled
and
by
now
we
actually
have
lots
of
code,
but
the
point
is
we
need
to
get
all
this
come
together,
and
so
what
we
did
this
time
was
lots
of
infrastructure
work.
So
one
of
us
closed
about
40
issues
in
the
node
watch.
Implementation
will
fix
the
resource
directory.
We
integrated
with
IKEA
lights
and
made
finger
aqueous
accessible
on
the
cloud.
Did
lightweight
m2
finca
scription
generators?
K
What's
variant,
it's
on
these
are
all
it's
all
a
bit
super
if
you
are
not
familiar
with
the
technology,
so
I'm
not
going
to
read
all
of
that.
So
in
the
end,
what
we
learned
is
one
and
a
half
days
is
too
short
to
send
all
that
glue.
So
we
probably
will
do
more
echelons
in
online
form
since
we're
at
the
application
layer.
We
can
easily
do
that.
K
On
the
other
hand,
lots
of
things
are
now
coming
together,
so
when,
when
every
one
of
these
lights
goes
on
in
our
table,
usually
you
have
about
five
different
stos
working
together
to
make
that
happen.
This
is
now
happening
and
starting
to
work.
On
the
other
hand,
we
also
identified
you,
given
that
we
are
research
group.
We
have
happy
to
have
identified
a
number
of
research
topics
that
will
need
work
to
get
a
more
comprehensive
solution.
K
L
L
Improvised
like
this,
so
you
may
or
may
not
know
whenever
you
make
a
video
call
voice,
video
call
through
blue
jeans
or
WebEx
or
jitsi,
or
whatever
the
conference
exists.
Music
conferencing
provider
can
see
everything
you're
doing
and
deliver
it's
whatever
intelligence
service.
It's
collaborating
with
data
services,
its
selling
off
your
metadata
and
all
that
stuff,
not
very
private.
At
all.
So
we've
had
this
working
group
called
perk
to
add
some
privacy
to
that.
L
If
these
three
were
listed
here
and
hack
them
into
a
bunch
of
open-source
product,
so
live
SRTP
is
very
widely
used
for
protecting
real-time
media
and
SS
is
a
library
it's
used
for
DTLS
and
a
bunch
of
key
management,
stuff,
Firefox,
I,
assume
you've
all
heard
of,
and
then
we
hacked
up
a
media
distributor
and
a
key
distributor
interest
and
the
infrastructure
bits
one
and
go
and
one
in
C++,
graphing
and
SS.
So
the
short
answer
for
what
we
accomplished
here
is
we
made
this
like
whole
crazy,
key
dance
work.
L
This
actually
does
tell
the
story
of
Park
if
you
sit
and
stare
at
it
for
a
while
and
you're,
not
colorblind.
Sorry
about
that
inclusion,
point
of
view,
but
basically
there's
this
whole
dance
that
you
go
through
involving
DTLS
and
SRTP,
and
some
extensions
to
those
and
passing
a
bunch
of
keys
and
parts
of
keys
around
and
encrypting
things
and
decrypting
things
that
allow
the
media
distributor
this
conferencing
server
to
do
the
little
bit
of
modification
it
needs
to
do
without
also
getting
to
see
the
contents
of
the
packets.
L
L
We
I
think
have
audio
working.
We
have
you
can
do
audio
conferences
through
this
now
with
full
privacy,
there's
just
something
in
the
operating
system.
It's
fragmenting
your
video
packets
of
positivity,
so
video
is
not
quite
working
but
work.
We
got
kind
of
you
know
audio
working
and
it
proved
out
this
this
new
architecture,
lessons
learned,
were
pretty
much
things
work
as
specified.
We've
done
a
bunch
of
clarifications
as
you
as
you
can
might
gather
from
that
previous
slide.
L
There's
a
bunch
of
complicated,
slicing
and
dicing
in
this
key,
and
this
have
to
keep
that
happy
as
long
as
you
keep
all
those
balls
in
the
air
and
you
have
them
in
the
right
place
at
the
right
time
works
fine,
but
we
need
to
clarify
these
guys.
How
did
you
all
that
I've
done
a
couple
minor
bugs
about
which
happy
news
where-
and
this
is
all
thanks
to
it's-
a
team
between
Cisco
and
Missoula
folks,
bunch
of
people
contributing
different
expertise.
This
crypto
folks
media
folks
that
need
all
come
together.
D
D
M
M
So
one
of
the
projects
that
we
had
quite
a
large
team
on
was
a
multi
provider
Dana
spec
project.
So
today
you
can
do
multi
provide
the
DNS,
which
is
great
for
defending
it
against
DDoS
attacks,
or
you
can
do
dinner
sack
which
is
great
for
integrity
of
your
data,
but
you
cannot
do
today.
Operational
readiness
is
both
together,
so
there's
a
draft
that
proposes
how
you
can
do
this,
and
the
idea
behind
the
draft
is
that
each
of
the
vendors
independently
sign
the
zone.
M
There
are
two
models
proposed
and
they
differ
in
the
way
that
they
manage
the
keys.
The
work
that
was
done
here
this
weekend
was
that
both
of
those
models
were
deployed
and
validated
interesting.
The
first
thing
that
popped
up
is
that
some
of
the
diagnostic
tools
couldn't
handle
this,
because
it
just
went
expecting
these
configurations.
We
had,
though
they
did
develop
a
minimal
API,
so
the
two
vendors
can
synchronize
a
zone,
the
keys
that
I
come
between
each
other
and
some
brand
new
tools
to
develop.
M
So
it
could
check
the
consistency
of
the
zones
that
are
signed
by
the
separate
vendors.
It
really
nice
piece
of
work.
One
of
the
other
big
operational
issues
we
have
in
the
dynast
at
the
moment
is
the
fact
that
there
is
no
standard
way
to
have
a
cname
from
parent
or
the
apex
per
zone.
I've
been
having
a
huge
fun
fighting
against
opt
lists
about
how
we
solve
this
properly.
M
M
One
of
the
other
projects
I've
got
worked
on
is
oblivious
DNS.
If
you
haven't
come
across
this,
it's
a
privacy
technique
that
changes
the
communication
path
between
a
stub
client
and
the
recursive
resolver
by
inserting
an
oblivious
tin,
a
stub
in
between
them.
The
idea
here
is
that,
by
inserting
syrup,
neither
the
stub
nor
the
recursive
can
see
both
the
client
IP
address
and
the
content
of
the
query.
So
no
one
entity
can
correlate
the
two
and
there
was
a
draft.
An
existing
go
line
prototype.
M
What
happened
here
was
that
an
en
I'll,
let
lab
folk
started
working
on
porting
that
into
CE,
and
the
first
things
that
came
out
of
that
was
a
huge
amount
of
feedback
to
the
authors
of
the
draft
into
how
they
could
change
how
they're
doing
it
in
DNS.
All
the
better
to
make
it
work
better.
The
protocol
and
also
the
implementers,
were
very
impressed
with
the
crypto
that
was
defined
in
the
paper.
M
Well
small
piece
of
work
was
looking
at
doe
myself,
with
reviews
from
a
whole
bunch
of
people
too.
Many
to
put
on
here,
I
started,
looking
at
trying
to
defined
and
I'm
an
anonymity
profile
to
use
with
those
so
that
clients
who
want
to
use
that
can
look
very
standard,
not
stand
out
and
to
help
us
understand
whether
this
will
work
or
not.
We
developed
a
a
very
lightweight
proxy
that
you
can
point
a
doe
client
at
and
I'll
expose.
M
The
headers
that
are
being
sent
out
on
top
of
the
dough
requests
and
blast
side
is
that
it's
also
a
piece
of
work
to
take
a
rather
old
patch
and
pull
it
to
the
latest
version
of
NSD,
and
this
therefore
puts
the
capability
to
be
DNS
ever
tear
less
into
an
authoritative
server,
and
this
worked
it
got
tested.
We've
got
some
performance
numbers
it,
no
real
surprises
is
just
it
just
seems
to
work,
and
lastly,
there
was
also
work
on
implementing
a
a
draft
to
which
it's
based
around
zone
digests
for
integrity.
N
All
right,
hello,
everyone
so
for
those
of
you
who
may
or
may
not
know
sni
privacy
problem
isn't
something
a
flake
tales
for
a
very
long
time
now.
We
always
just
send
this
particular
piece
of
information
in
the
clear,
basically
revealing
to
the
entire
world,
where
we're
going,
and
so
we've
wanted
for
a
very
long
time
to
encrypt
this,
and
so
this
particular
idea
for
a
draft
and
set
up
to
actually
do
so.
N
This
was
along
with
some
various
other
people,
fastly
CloudFlare,
Mozilla
and
Apple,
and
various
TS
libraries,
particularly
Pico
to
us
and
SS
and
Bourne
SSL,
and
the
results
were
pretty
good.
We
got
complete
interrupt
between
all
supporting
clients
and
servers.
There
are
two
servers
they
have
up
right
now
that
support
this
that
have,
you
know
fixed
records
containing
these
sni
keys.
N
O
We
did
some
work
this
weekend
on
dinner
service
discovery.
Probably
half
the
people
in
this
room
have
heard
of
DNS
service
discovery,
and
you
think
you
may
think
that's
the
thing
too
is
multicast
DNS,
that's
link-local
multicast
and
that's
part
of
what
it
does
now
for
a
long
time.
It's
also
run
over
unicast
and
in
fact
that's
what
we're
doing
here
at
the
ITF.
O
If
you
tap
the
AirPrint
button
on
your
iPhone
or
your
iPad,
you
to
the
terminal
room
for
into
show
up
that
is
doing
a
serve
discovery,
but
you're
not
in
the
terminal
room.
That's
the
other
side
of
the
building
up
a
couple
of
floors,
so
that
works,
but
it
requires
the
experienced
ITF
network
operators
to
know
the
right
records
to
put
in
the
DNS
server
to
configure
that
manually.
So
the
DNS
SD
working
group
has
been
working
on
automating
that
process.
O
This
is
important
these
days
because
there
are
wireless
mesh
networks
where
multicast
is
not
very
efficient.
Enterprise
Wi-Fi
networks,
frequently
disable
multicast
for
the
same
reason
that
multicast
is
very
wasteful
on
Wi-Fi
and
home
networks
getting
more
complicated.
So
it's
becoming
more
common
to
have
more
than
a
single
network
segment
in
the
home.
O
Now
these
the
documents
were
looking
at
67-63
is
the
base
service
discovery
specification
that
defines
how
to
client
does
the
queries
are
beginning
after
DNS,
what
we're
adding
is
automating
populating
the
namespace
Ted
Landman
has
been
working
on
a
relay
which
is
analogous
to
a
dhcp
bootp
relay
at
home.
You
may
have
DHCP
server
on
the
network,
but
in
an
enterprise,
it's
common
to
have
the
DHCP
server
in
the
data
center
and
lots
of
little
bootp
relay
Agron's
scattered
around
all
the
links
in
that
enterprise.
O
So
taking
that
inspiration,
ted
has
been
working
on
a
similar
thing
for
service
discovery,
where
you
can
have
a
discovery
proxy
in
the
data
center
and
it
in
effect,
has
remote
virtual
interfaces
on
various
links
around
the
organization
and
I
Ted.
Now
how's
that
running
on
open
wrt
and
some
raspberry
pie
boxes.
He
has
it
running
on
one
of
his
boxes
at
home.
O
So
this
is
not
a
greenfield
development.
He
implemented
that
in
Python
tested
that
against
bind
9
with
DNS
update,
enabled,
and
his
code
is
now
in
github
little
screenshot
there
is
showing
I
was
looking
on
my
web
browser.
This
was
in
fact
a
remote
screen
sharing
connection
to
my
house
in
Santa
Cruz
in
California.
You
may
be
able
to
see
there.
I've
got
four
of
my
access
network
security
cameras
are
showing
up
when
Tim
used
the
registration
protocol
to
advertise
another
web
service.
That
fifth
one
appeared
on
the
list
at
my
house
in
California.
O
P
Hello,
my
name
is
Vivek.
I
am
tell
me
about
the
work
we
did
in
the
ACE
product,
so
ace
is
for
authentication
and
authorization
constrained
environments.
So
what
we
are
trying
to
do
in
the
working
group
is
define
profile
of
all
that
is
usable
on
devices
and
networks
that
have
real
resource
constraints
and
we're
now
at
the
stage
where
we
can
try
and
have
codes
talk
to
each
other
and
that's
what
we
basically
plan
to
do.
So
we
had
different
implementations
of
these
two
drafts
and
we
wanted
to
get
them
to
talk
to
each
other.
P
We
had
a
interrupt
test
specification
which
I,
wrote
and
thought
was
really
really
really
really
detailed
turns
out.
You
can
never
be
detailed
enough.
It's
like
you
can
always
get
the
key
identifier
wrong
or
whatever
else
you
can
find
to
shoot
yourself
in
the
foot
I'm
going
to
get
that.
That's
what
we
did
was
we
got
our
implementations
talk
to
each
other.
You
can
see
various
repositories
from
our
implementations
that
we
have
in
place.
Actually
all
the
SDI
ones
are
the
same
implementation,
but
they
split
it
up
in
different
registers
for
different
functions.
P
So,
in
the
course
of
the
interrupt,
we
did
a
number
of
updates
to
discipline
tations
because
we
discovered
a
lot
of
problems
in
the
code.
Luckily,
we
didn't
discover
new
problems
in
the
protocols,
so
looking
good,
we
had
some
questions,
though
mainly
about
key
identifier
and
how
they
are
supposed
to
be
handled.
What
we
learned
is
yeah.
We
can
get
the
basic
protocol
interactions
to
work.
You
can
get
the
different
implementations
to
talk
to
each
other
on
unconstraint
devices.
P
P
People
make
so
if
anyone
feels
like
the
need
to
do
an
extra
good
thing
for
ITF
I.
Think
starting.
This
kind
of
work
would
be
immensely
helpful
for
future
hackathons
and
people
trying
to
do
interrupt.
Some
new
protocol
drafts,
so
team
members
we
had
Jim
shot
this
lewis
sebastian
and
sherea
and
unclimbed
in
I
hope
I'm,
not
mangling.
Your
names
too
much
I
should
have
asked
before
and
myself
Jim
and
I.
We
have
been
at
several
IDF's.
P
F
Q
Q
So
our
goals,
where
our
in
general
to
provide
an
open
source
implementation
of
sheet-
and
the
draft
is
listed
here
and
right
now,
the
the
draft
is
in
version
16.
Some
changes
have
been
made
recently.
The
draft
is
entering
its
second
working
group
task,
oh,
and
so
our
goal
was
to
advance
the
state
of
the
code
to
reflect
that
the
16
version,
so
mostly
by
reading
sharing
our
understanding
of
the
draft,
the
last
changes
and
coding
making
sure
the
code
reflects
at
last
version.
So
what
got
done?
Q
Also,
we
have
dated
a
few
contributors
on
the
recent
advances
of
the
draft
because
then
followed
they
all
the
details
of
the
discussions,
so
we
updated
the
Sheik
has
two
parts.
One
is
a
compression.
The
compression
one
is
augmentation
recently,
so
we
take
the
fragmentation
code
as
a
retro
and
hit
her,
and
also
we
take
it.
The
compression
part,
and
especially
in
completely
network
side,
which
one
is
between
end
device,
implement,
improves
the
net
website
in
communication.
Q
What
we
had
learned
last
time
at
the
101
with
that
it
takes
a
long
time
to
get
all
these
little
devices
and
the
nor
one
network
working
in
this
room
and
we
can
easily
waste
a
full
day
having
ever
to
be
working
with
lovely
devices.
So
we
knew
that
so
and
now
we
decided
to
have
three
versions:
three
targets,
one
being
a
full
Python
free
implementation,
which
means
my
computer
server
side.
Q
So
as
a
wrapper,
we
had
six
people
locally
around
the
table,
contributing
one
remotely
in
Japan
with
11
hours,
time
difference
so
stay
awake
until
4:00
a.m.
local
time
to
group
and
to
among
those
were
newcomers
to
the
ITF
end,
and
so
we
got
the
code
advanced,
but
that
is
the
finish
that
will
probably
carry
on
working
this
week.
R
When
you
got
the
one
slide
here,
we
were
a
group
of
people
kind
of
all
working
on
our
ending.
The
Gemma
project
is
a
replacement
protocol
for
IMAP
and
for
email,
and
so
we
have
a
category
that
we're
working
on
a
client
and
PHP
for
PHP
web
mail
systems.
We
had
a
group
working
on
the
gemma
Chester,
which
the
tests
wait
for
server
implementations.
I
was
working
on
the
Gemma
pearl
proxy,
which
proxies
from
Joe
map
I'm
at
Kelvin.
R
I've
have
servers
so
that
you
can
run
your
existing
servers,
then
from
Jim
up
on
top
of
it,
and
we
also
did
some
work
on
the
terrace
IMAP
server
and
found
some
inconsistencies
in
the
spec
which
we
have
pushed
some
change
requests
back
for
we're
meeting
tomorrow
afternoon.
If
you
want
to
come
along
and
find
out
about
it
and
we'll
all.
S
S
Disclaimer
for
the
whole
rest
of
the
presentation
here,
so
we
wanted
to
basically
get
a
common
understanding
of
what
these
drafts
in
the
body
there's
two
different
key
exchange
mechanisms,
one
music
diffie-hellman
one
using
a
public
key
mechanism-
and
we
just
wanted
to
figure
out-
can
we
compute
the
same
trees?
Can
we
compute
the
same
stuff?
The
effect
my
group
is
back,
there
still
hacking
away,
and
they
just
achieved
a
couple
common
views
of
of
the
same
thing
as
using
a
bunch
of
different
implementations.
We
have
some
new
implementations.
S
Everyone
is
sort
of
comfortable
using
different
things
and
they
sort
of
want
to
use
these
in
different
applications
down
the
line.
So
there's
a
C
implementation.
There's
a
J's
node
one,
there's
a
most
of
these
people,
ported
Richard
Barnes's
code
code
to
whatever
they
wanted,
John
didn't,
because
what
got
done
a
whole
lot
of
just
whiteboarding
and
just
trying
to
figure
out
if
we
all
have
this
common
understanding
of
what
the
heck
this
stuff
is
and
how
it
works.
S
There
are
two
sets
of
code
here
that
are
achieving
some
notion
of
a
common
view
of
the
tree.
There's
a
couple
of
trivial
pull
requests
to
MLS
drafts:
we
have
live.
What
do
we
learn
and
it's
hard
to
describe?
We
kind
of
better
understand
this
stuff.
The
drafts
are
in
the
early
state
or
in
a
better
state
to
contribute
to
this
kind
of
conversation.
Actually,
can
you
prove
some
of
this
code?
I
do
think
we
I'm
not
sure
how
to
work
this
I'm,
not
an
expert
in
this
stuff.
S
T
So
hello,
we've
been
working
on
the
coming
project,
for
which
is,
you
know
the
lingual
yang
for
constrained
devices
for
for
this
hackathon
and
so
yeah
just
a
position
where
we
are
well.
It
is
an
addition
to
the
to
the
yang
family,
so
we
have
that
count
and
unrest
conf
and
come
on.
You
can
say
that
it
is
a
set
of
technologies,
so
there
is
a
way
so
bringing
the
idiot
the
aim
to
ecosystem
to
constraint
devices.
T
So
we
have
a
technology
to
turn
app
coming
to
Seaborn,
so
efficient,
a
payload
representation,
and
then
we
have
how
to
use
actually.
How
do
we
have
the
bindings
to
a
co-op
to
the
co-op
protocol
and
something
that
we
can
call
like
a
compressed
IDs
but
they're
not
really
compressed
its
structured
identifiers?
So,
like
your
numbers,
so
that
we
can,
we
are
actually
able
to
send
these
identifiers
so
very,
very,
very
constrained
links.
So
that's
on
what
we
have
been
working
on
what
exists?
T
Well,
we
have
couple
of
drafts
that
are
working
with
items
and
that
have
been
stable
for
some
time
now.
We
have
running
example,
registry
for
these
numbers.
That
is
out
there,
so
you
can
go
and
you
can
play
with
your
own
yang
modules
if
you
wish-
and
there
has
been
an
implementation,
so
I'm
saying
example
of
implementation,
but
it's
released
full
implementation
of
the
content
server
over
the
F
interrupt
platform.
So
in
case
you
would
like
to
do
interrupt
testing.
We
would
strongly
advise
you
to
actually
approach
the
F
interrupt
guys.
T
It's
really
something
very,
very
interesting,
very
powerful
that
allows
you
to
learn
all
kind
of
interpretive
interoperability
tests.
So
this
is
the
existing
and
what
we
wanted
to
achieve
during
this
hackathon
is
actually
to
provide
some
open
source
code
and
we
wanted
to
help
people
bootstrap.
Their
implementations
so
provide
examples
of
how
to
actually
interact
with
with
the
F
interrupt
implementation
that
is
out
there,
and
we
wanted
also
to
to
start
the
implementation
for
who
python-based
open-source
combine
clients-
and
you
know
in
the
future,
come
my
server.
T
D
T
The
basic
the
baseline
of
interactions
with
with
with
these
implementations,
so
anyone
can
actually
say:
okay,
well,
I'm,
interested
in
in
Python
implementation
or
a
C
or
C++
or
other
well.
I
can
do
this
to
validate
actually
that
it's
actually
working
and
then
move
on
to
automatic
automated
code
generation
based
on
YT
k.
So
ydk
is
hearing
augment
kit,
so
it's
basically
a
way
for
people
to
develop
their
applications
based
on
Ian,
it's
open
source.
So
what
got
done
well?
T
We
developed
base
examples
on
various
operating
systems,
Linux
and
Mac,
and
particularly
part
about
this
one
back
to
this
one
foundation
and
we
actually
clarify
to
read
this
helped
us
also
dive
into
the.
How
do
you
develop
a
car,
my
client
and
server
by
the
way?
So
basically,
it
boils
down
to
developing
independently
the
yang
to
see
formatting
and
the
coop
bindings,
which
is
actually
pretty
great
thing,
because
this
is
this
allows
you
to
split
your
work,
and
you
know
there
are
no
big.
T
They
are
not
interdependent
specifics
and
that's
actually
a
process
that
is
compatible
with
commercial
and
even
open
source
net,
compress
converse
servers
and
clients.
So
you
can
get
your
existing
net
compress
con
server
and
you
can
add,
come
out
pretty
easily
and
we
actually
identified
also
the
steps
necessary
to
do
this
and
we
started
our
ydk
based
common
limitation.
T
Of
course,
the
the
normal
difficulties
when
you
mix
up
Python
with
different
versions
with
Mac
with
a
36,
and
we
chose
library
several
times
where
a
very
good
pointer
is
to
virtualize
your
hackathon
environment.
If
you
can
use
docker
or
stuff
like
this,
and
also
why
the
case
user-friendly
for
the
users.
But
you
know
for
people
that
try
to
extend
it
a
little
bit
less
so
and
I'd
like
to
thank
our
participants.
T
So
this
is
our
first
hackathon
actually
that
we
started,
but
we
had
a
person
that
was
a
remote
participant
who
had
a
first-timer,
and
we
had
also
a
cross
participation
with
the
Quarian
group
so
that
we
really
had
very
good
discussions
about
this.
And
you
have
the
docs
that
are
on
the
ether
pad
and
you
have
some
code
on
github
Thanks.
U
Hi,
hello,
everyone,
Kahneman,
JISC,
accommodations,
yeah,
okay,
talk
about
the
update
of
thousands
of
adoptability
50
he's
a
hawk
complan
thought
it
the
details
of
a
great
shooting.
It
is
for
a
standardization
of
shooting
for
DDoS
protection.
Actually,
are
we
attending
the
pakistan
from
IETF
99
our
concept
TV
dinner?
At
a
time
there
was
only
one
implementation,
but
this
time
now
we
have
mainly
to
implementation,
and
then
we
did
first,
the
channel
interrupted
phenomena.
U
Vaneeta
code
switches
are
always
running
on
AWS
from
my
side,
then,
on
the
other
side,
from
implementation
from
in
secret,
which
is
running
on
the
production
anti-ddos
box,
then
our
adult
gray,
on
top
of
each
other
talk
over
this,
you
know
channel
on
the
italian
affected
data
channel
if
the
first
time
interoperability
easting
each
other,
then
on
the
each
other.
We
do
the
help
helping
function
to
the
DDoS
mitigation,
like
iteration
of
gaia,
of
dry,
anthony
shield
and
also
capability,
is
cover.
Your
company
then,
are
either
interrupt
result.
U
You
may
see
that
green
cells
are
company's
accomplishment
in
the
IETF.
101
and
blues
has
accomplished
complete
accomplishment
in
the
I'ii
taik
to
them.
You
may
see
that
the
test
space
is
enhanced
significantly
and
others
did
date,
each
other.
It's
a
saucepan,
but
the
lowest
idea
is
the
testing
over
the
data
each
other,
and
all
of
this
was
successful
and
found
issues.
We
will
report
you
to
the
working
group
and
also
we
did
the
soon
interrupts
with
the
latest
draft
of
Junichiro
draft
they're
not
here
to
take
away
the
cost.
U
Education
of
dots
is
now
major
enough.
So
we
did
a
successful
internal
testing
in
several
issues
fun,
but
that
is
not
so
significant
on
total
will
be
reported
to
the
party
group.
So
a
documentation
of
a
needs
Oasis
and
one
is
proprietary
and
also
more
implementations-
are
welcome
and
at
the
reference
implementation
or
it
is
code,
is
available
on
github,
so
fitting,
read
and
use
these
horses.
U
F
V
This
time
one
of
my
co-workers
worked
on
a
hardware
dedicated
for
our
interoperability
and
hackathons,
which
you
can
see,
on
the
left
hand,
side
the
mighty
microcontroller
board
with
separate
flash
and
secure
element,
and
we
try
to
get
this
the
code
working
on
this
new
board.
The
goal
was
to
generate
this
manifest,
which
is
the
metadata
for
the
firmware
cluster
security
wrapper
encoded
in
Seaborg,
sign
it
with
cosy
and
then
verify
it
on
the
IOT
device
and
so
for
the
generation
we
use
Python
but
of
course,
on
the
IOT
device
we
used
on
this
board.
V
This
is
our
group.
Actually,
over
there
we
had
a
couple
of
people
who
were
the
first
time
participants
of
the
of
the
suit
hackathon.
We
also
had
first-timers
in
the
ITF
in
red,
and
and
for
this
me
and
pay
me,
what
did
we
do
since
we
had
a
couple
of
new
folks
working
on
this
topic,
we
spent
some
time
on
setting
up
the
development
environment
for
this,
for
the
board
getting
distributing
the
ports
and
getting
getting
everything
working.
V
We
use
to
embed
OS
as
a
operating
system
for
the
for
the
device
we
successfully
generated
the
manifest
encoded.
It
sine-theta
turned
it
all
the
keying
stuff
and
also
wrote
the
parser
for
the
verification
of
it
under
on
the
device.
We
also
produced
a
detailed
write-up
like
last
time,
so
we
can
actually
memorize
things
till
the
next
event
and
which
can
be
found
here
on
this
link
a
couple
of
items
we
learned
during
the
activities
we
managed
to
break
a
couple
of
the
boards.
That's
obviously
we
need
to
work
on
that.
V
We
also
would
be
nicer
to
have
a
way
to
set
up
everything
in
a
in
a
much
faster
way
for
first-timers,
because
there
are
lots
of
tools
involved
in
in
the
imperative
development
in
setting
those
up
appropriately
for
different
operating
systems.
It's
a
hassle
using
IDs
for
debugging,
specifically
if
it's
more
complicated
crypto
would
also
be
nice,
but
requires
some
work
and
in
we
obviously
struggled
a
little
bit
with
the
implementation
in
this
this
time
with
SIBO
our
last
time
with
cosy.
V
So
we
managed
to
deal
with
that
in
between
the
Alana
ITF
meeting
in
this
meeting.
But
now
we
are
the
issues
with
cozy
libraries
and
there's
also
some
feedback
for
that
for
that
for
the
draft
work,
yeah
and
finally,
getting
everything
back
together
in
a
reasonable
size
for
a
boot.
Loader
is
still
going
to
be
a
challenge,
but
let's
sort
of
let
the
same
exercise
for
for
future
hackathons
and
future
work,
and
that's
it.
Thank
you.
A
W
W
For
the
idea,
902
hackathon
focused
on
two
paths:
application
integration
and
the
other
is
in
probability.
First.
Can
you
just
move
ones
like
that?
Nice?
So
all
that's
that's
forgivable,
different
motions
and
the
application
integration
that
we
had
work
on
the
implementation
of
both
28
and
upstream
open-source
projects
and
vampire.
In
the
case
of
PHP,
it
is
not
possible
to
specify
in
LS
1
and
3
connection
and
we
using
the
PHP
client
library.
So
we
have
to
make
a
PHP
trail
again
with
Els
one
hand,
free
compatible
and
test
it
using
PHP
code
itself.
W
As
regards
to
the
interplay
we
do.
There
are
some
challenges
to
make
Els
mailboxes
more
compatible
with
Railway
deployment.
We
have
the
changes
to
make
sure
that
all
unit
tests
work
accordingly,
so
the
application
concern
a
little
smoother
and
slide.
So
we
have
HP
continuously
did
not
have
its
care
functions
of
all
50
RS
120.
So
if
we
invest
were
sent
to
PHP,
which
is
a
2d
measuring
work
for
humans
will
be
percent
and
happy
to
measure.
W
We
will
machine
metrics
work
around
it
with
a
website
to
the
microwave
enemies,
so
cacnea
center
developer,
and
it's
where
box
contactable,
we
still
as
1.3
support,
but
we
can
use
eval.
You
may
have
Magus
proteins
and
we
already
merge.
So
ever
in
websites.
We
have
been
able
to
accomplish
several
envelopes
tasks.
We
serve
all
TLS
features
such
as
full
handshake,
resumption
and
zero
article.
There
are
dysfunctional
open,
SSL
growing
as
a
self
yella
slide.
W
W
This
sweats
times
a
half
as
the
time
with
him
or
in
this
move
one
knows
life,
so
we
also
have
our
first
time
that
spends
in
the
IDF
TLS
welcome
to
happen.
I,
remember,
sweat,
sweat
is
actually
15
years
old,
so
everywhere
and
with
us
wonderfully
except
the
issue
and
anyhoo
made
an
amazing
nothing
an
HTTP
program,
so
I
do
pictures,
always
use
idea
family
to
happen.
W
That's
us
and
they
will
just
when
you
that's
emotions,
I
use
artifacts
needs
women
and
the
TLS
weapon
feeder
team,
where
NZ
and
everyone
comes
in
with
security,
Charles
and
very
frustrating
the
hackathon,
and
we
also
diversified
routed
users.
So,
thank
you,
everyone
and
you
people
over
the
next
item.
X
All
right
so
there's
the
link
again
for
for
voting
will
give
you
maybe
what
time
is
it
now
3:30?
So
we'll
give
you
maybe
10
minutes
to
get
your
votes
in
again,
please
don't
vote
for
your
own
project
out
of
your
peers,
which
other
project
best
accomplished.
What
you
think
we're
here
to
accomplish,
and
and
I'll
just
shut
up
now
for
about
10
minutes,
we'll
let
that
sink
in
and
get
your
votes
in.
You
see
anything.
A
That's
all
I'll
leave
that
up
on
the
screen
for
those
of
you
who
need
to
get
the
link
just
one
more
reminder
of
a
couple
of
things.
We
have
the
the
code
lounge
that
will
be
going
on
all
week,
Monday
through
Friday
I,
encourage
you
to
use
that
space.
If
it's
helpful,
you
don't
have
to
sign
up,
there
is
a
sign-up
sheet.
A
Signing
up
is
really
just
to
help
you
coordinate
with
other
people,
so
you
can
kind
of
put
your
name
down
your
project
and
then
everyone
knows
that
you're
planning
to
be
there
at
specific
times
now
for
the
happy
hour
that
the
hack
demo
have
the
hour
that
you
do
need
to
register,
because
we
need
to
know
how
many
people
were
aren't
going
to
how
many
teams
are
going
to
be
there.
So
we
get
the
right
number
of
tables
the
right
number
of
signs,
and
all
of
that
so
I
just
see
a
show
of
hands.
A
How
many
teams
you
have
one
person
from
your
team
raise
your
hand
if
you're
planning
on
demoing
something
tomorrow
at
the
happy
hour,
one
two
anymore:
okay,
three
yeah!
So,
okay,
before
so
yeah
again
it's
just
it's
a
way
for
you
to
share
what
you've
done,
get
some
more
feedback
on
what
you've
done,
share
it
with
more
people.
It
would
be
great
to
see
some
of
some
of
these
teams
you
showing
other
people
what
you've
done.
A
It's
it's
also
a
real
good
way,
just
to
get
the
word
out
about
the
hackathon
in
general
and
what's
going
on
here,
it's
really
nice
to
see
so
many
of
you
here,
but
there's
obviously
a
lot
more
people
who
maybe
have
been
intimidated
or
didn't
quite
understand
the
spirit
of
it
and
you
know
showing
your
project
there.
It's
a
good
way
to
get
that
word
out.
X
Okay,
we're
ready
for
results
here.
The
the
votes
have
stopped
coming
in
so
I
think
we're
ready
to
to
do
results
thanks
everybody
for
doing
this
participating
this
way.
I
think
this
was
a
fun
way
to
do
it.
I,
like
I,
like
the
way
it's
worked
out.
We
have
a
clear,
clear,
sorting
out
here.
The
number
one
group
is
the
DNS
DNS
SEC
DNS
privacy
group.
That's
got
the
most
votes
and
tied
for
second,
we
have
perc
and
suit
so.