►
From YouTube: IETF110-HACKATHON-20210305-1600
Description
HACKATHON meeting session at IETF110
2021/03/05 1600
https://datatracker.ietf.org/meeting/110/proceedings/
A
And
welcome
to
the
the
hackathon
closing.
A
Hope
hope
you
had
a
good
week
and
we
have
thanks
to
all
you
for
joining.
I
see
we
have
a
good
list
of
presentations
lined
up.
I
want
to
wait
just
a
couple
minutes
for
those
people
who
are
still.
A
A
A
While
we
are
waiting
for
people
to
join,
I
want
to
just
cover
a
couple
things
that
I'll
cover
again
at
the
end,
but
I
figured
you
know
some
of
you
may
not
be
able
to
stay
for
the
whole
time,
although
I
hope
you
can
so
let
me
just
briefly
share
this.
First
of
all,
there
is
a
survey
for
the
hackathon
it
it.
The
link
has
been
added
to
the
wiki.
A
So
if
you
go
to
the
hackathon,
wiki
you'll
find
it,
but
I
also
put
it
here
in
the
slides
and
I've
uploaded
these
to
github,
so
you
can
grab
them
from
there
as
well,
then.
A
The
other
thing
I
wanted
to
mention
is
a
draft,
and
actually
first
shared
this
and
had
it
discussed
at
the
last
ietf
meeting
and
there's
been
some
updates
to
it
now
and
it'll
be
discussed
again
in
the
working
group
and
it's
basically
covering
the
hackathon,
how
it
works
kind
of
as
a
from
a
process
perspective
just
to
capture.
You
know
this
is
how
we
run
the
hackathon
and
why
we
do
things
the
way
we
do
and
also
part
of
the
thing.
A
The
reason
why
it
was
created
was,
and
the
whole
schmoon
working
group
has
to
do
with
you
know,
dealing
with
meeting
online
as
we
are
now
both
for
the
itf
meeting
for
the
hackathon,
and
so
the
draft
isn't
specific
to
that.
It
really
covers
the
hackathon
overall,
but
also
aspects
that
are
either
new.
B
A
Different
when
you're
running
it
online,
so
I
would
appreciate,
if
folks
here,
while
everything's
still
very
fresh
in
your
mind,
if
you
could
take
a
look
at
that,
it's
it's
not
very
long.
It's
maybe
you
know
10
pages
or
something
like
that
with
a
lot
of
kind
of
boilerplate
stuff.
So
it's
a
pretty
quick
read
and
if
you
have
any
thoughts,
I'd
love
to
hear
them
on
the
list,
a
pull
request.
It
is
on
github,
there's
there's
the
link
you'll
find
that
at
the
on
the
data
tracker.
A
Here:
okay,
so
just
a
reminder:
some
people
maybe
didn't
attend
the
beginning,
but
I
think
it's
always
good
to
you
know
attend
the
kickoff.
I
think
it's
always
good
to
just
recap:
why
we're
here
and
that's
really,
you
know-
for
the
hackathon.
The
whole
idea
is
to
advance
the
pace
and
relevance
of
of
ihf
standards
to
really
bring
together
the
that
spirit
of
collaboration
that
we
get
with
the
open
source,
community
and
and
work
on
code.
A
You
know
anything
really
related
to
ietf
standards
and
the
goals
of
that
are
both
to
you
know,
make
the
standards
better
right
because
we're
getting
feedback
from
actually
having
coded
things,
but
also
just
to
attract
developers
and
new
people
into
ietf
and
to
take
some
of
the
stuff
that
you
know
to
progress
really
from
just
the
specification
which
is
great
to
if
you
actually
have
some
code
that
can
help
you
get
started
with
implementing
or
adding
support
for
the
draft
or
with
testing
things
about.
A
A
Hopefully
it's
open
source
code
that
makes
it
easier
for
everyone
to
collaborate
and
share,
but
anyways
the
you
have
to
look
at
the
license
and
the
rules
and
restrictions
there
regarding
the
code,
but
when
it
comes
to
the
presentations
that
we'll
all
be
enjoying
during
the
closing
those
are,
you
know
your
your
normal
ietf
contributions
and
fall
under
the
note.
Well,.
A
Okay,
the
agenda,
so
we
have
two
hours.
I
don't
think
we'll
need
the
full
two
hours
but
based
on
the
number
of
presentations,
my
guess
is
we'll
take
a
bit
more
than
one
hour.
So
you
know
one
hour
plus
we'll
definitely
finish
before
the.
B
A
Of
the
two
hours
and
we'll
have
some
time
for
discussion
at
the
end
and
the
way
it's
going
to
work
is
we're
basically
going
to
run
through
the
list
of
projects
in
the
results,
presentation
schedule,
and
this
is
on
the
wiki.
If
you
want
to
take
a
look
at
it,
I
might
have
a
slightly
dated
version
here,
but
it's
pretty
close.
A
I
just
grabbed
this
maybe
an
hour
ago,
and
the
plan
is
to
just
run
through
an
order,
we'll
start
at
the
top
we'll
work
through
them
if
anyone's
not
able
to
present
during
their
slot,
we'll
try
to
go
back
again
at
the
end
or
if
someone
knows
in
advance
that
they
have
a
timing
issue
and
need
to
present
in
a
different
order.
A
A
Just
a
quick
reminder
here:
the
the
presentations,
as
we
mentioned
before
the
ideas.
These
should
be
quick
presentations,
short
presentations
just
covering
some
highlights
kind
of
up
level
to
present.
What's
real,
you
know
relevant
to
the
ietf
community
even
to
folks
who
aren't
deeply
involved
in
the
technology
you're
working
on
the.
A
You
can
this
is
just
to
give
people
an
introduction
to
what
you
did
and
share
some
of
the
highlights,
and,
and
then
you
know,
people
can
always
reach
out
to
you
for
those
more
detailed
discussions.
A
It's
a
way
of
expanding
the
scope
there,
because
we
know
there's
a
lot
of
great
code
all
over
the
place
that
isn't
just
relevant
to
the
ietf
community.
It's
relevant
to
other
communities
as
well.
So
this
is
a
great
way
to
engage
with
them
and
we
like
to
just
you
know,
be
aware
of
that.
That's
happening.
A
Okay,
all
of
the
projects
should
be
uploaded
to
this
repo
in
the
ietf
github.
If,
if
you
haven't
uploaded
yet
that's
okay,
you
can
present
just
you
know
the
version
that
you
have
on
your
laptop.
That's
fine,
but
just
you
know
afterwards.
A
Please
go
ahead
and
make
sure
it's
been
uploaded
and
if
you've
already
uploaded
it
that's
ideal,
because
that
way
people
can
follow
follow
along
with
your
presentation
if
they
want
to
so
this
repo
will
remain
open
after
the
hackathon
closing
is
done,
and
so
you
know,
if
you
find
things
if
you
continue
to
work
on
your
code,
the
next
few
days,
whatever
that's
great,
you
can
update
your
presentation
to
capture
that.
A
Okay,
so
with
that
we're
going
to
get
started
on
the
the
presentations,
this
again
is
a
snapshot,
I'm
going
to
pull
up
the
actual
list
and
go
off
of
that
what's
on
the
wiki,
but
it
should
be
quite
close
and
any
questions
before
we
begin.
A
Okay,
so
why
don't
we
get
started,
then
application
aware
gsr,
v6,
networking.
C
D
C
A
A
C
Good:
okay,
thanks
charles
hello,
everyone,
my
name
is
jaiway
mao,
I'm
glad
to
introduce
our
project
for
you.
It
is
application,
aware,
gsrv6.
Networking
our
plan
is
to
develop
gsrv6
on
the
linux
kernel
and
combine
gsrv6
with
apn6.
C
C
The
transmission
overhead
of
srv6
is
too
high.
For
example,
if
you
want
to
use
a
streaked
te
tunnel
with
10
seeds,
the
seeds
will
occupy
160
bytes
in
your
packets.
That
is
not
accepted
in
interactive
control
and
iot
scenarios,
which
package
length
is
under
128
bytes.
So
how
about
gsr
v6?
The
key
point
is
gsrv6
can
reduce
75
percent
transmission
overhead.
C
C
Next,
let's
have
a
look
at
apn6.
Avn6
allows
devices
to
identify
application,
related
information
and
use
ipv6
extension
headers
to
convey
the
information
along
the
packet
to
traverse
the
network.
The
information
contains
app
id
sla
level,
bandwidth,
delay
requirements
and
so
on
so
network
can
quickly
adapt
to
specific
sla
requirements
of
different
applications.
C
C
C
After
that
you
can
encapsulate
gsrv6
routing
header
to
packets
for
apn
6
we
developed.
We
developed
some
functions
to
encapsulate
apn
6
options
with
application,
specific
information
in
ipv6
hall
by
hub
options,
header
we
set
up
a
three
layer
topology
for
this
demo,
and
we
tested
the
performance
between
srv6
and
gsiv6.
C
C
The
three
applications
are
first
file
downloading,
with
security
tracking
in
a
service
function
chain.
Second,
interactive
control,
which
is
using
leave
and
short
messages.
Third,
high
definition:
video
on
demand
for
the
result
of
performance
test.
We
can
see
first
forwarding
rate
of
gsr
v6
is
55
higher
than
srv6.
C
C
C
A
A
B
Can
you
see
my
screen?
Yes
looks
good,
perfect,
excellent,
so
yeah,
I'm
representing
bmp
and
young
push
from
the
grow
and
net
conf
working
group.
So
here
our
plan
on
our
hackathon
plan
is
we
are
on
the
bmp
side.
We
were
focusing
on
on
performance,
so
basically
we
are
interested
just
one.
If
I
could
interrupt.
A
You
quickly,
at
least
on
my
side
as
pop
up
on
top
of
your
the
title
of
the
screen
about
sharing,
maybe
just
minimize
that
or
move
it.
Yeah
that'd
be
great
thanks.
Sure.
B
I
think,
let's
put
it
somewhere
else,
yeah,
okay,
great
thanks
a
lot.
So
on
the
bmp
side,
our
focus
is
basically
we
want
to
understand
what
is
the
impact
of
bmp
onto
the
bgp
process,
especially
if
jason
silip
in
out
with
local
vip
and
path
marking
is
enabled
onto
cpu
and
memory
consumption?
B
On
the
other
hand,
one
of
the
interesting
part
of
bmp
is
that
you
can
get
an
exact
copy
of
the
bgp
rip
state,
and
we
want
to
understand
here
if,
when
b,
bgp
congestion
is
here,
if
under
all
circumstances,
we
can
rely
that
there
is
no
loss
present
and,
on
the
other
hand,
when
bmp
is
enabled,
we
will
also
understand
what
kind
of
impact
it
has
on
the
bgp
route
propagation.
B
So
that's
about
the
bmp
part,
so,
on
yang
push
there
basically
we're
working
on
an
open
source
implementation
of
udp
node
event
distributed
distributed
notif.
So
we
have
a
data
collection,
implement
data
collection,
implementation
and
also
a
mockup
publisher,
and
we
are,
we
were
integrating
it
into
the
pmacct
open
source
network
data
collection
in
this
hackathon.
We
were
also
focusing
here
on
testing
the
efficiency
and
the
throughput
on
on
udp
notif.
B
B
B
These
two
routers
are
basically
route
reflectors
and,
on
the
very
right
hand,
side.
This
is
the
provided
edge
routers,
and
these
little
boxes
in
here
show
basically
the
bmp
coverage.
So
what
we
are
exposing
towards
the
data
collection
which
is
in
here.
B
So,
on
the
lab
environment
side
I
mean
this
is
not
the
first
time
we
were
participating,
the
the
idf
hackathon,
so
we
are
from
each
academe
steadily
improving.
So
in
this
time
we
were
focusing
on
test
ultimate
automatization,
so
we
are
in
basically
generating
sequenced,
pgp
vpne
for
unicast
routes.
B
B
Now
we
are
the
next
step
where
we
want
to
improve
in
this
test
that
we
use
this
time
we
were
focusing
on
huawei
vrp.
We
have
also
is6r
and
tuning
patchunos
in
the
testbed
present,
but
because
of
the
time
restriction,
we
were
not
able
to
use
it
so
in
the
next
itf
hackathon.
The
next
step
is
also
do
interoperability.
Tests
among
vendors
and
since
timestamps
are
important
on
detecting
delay.
B
B
It
will
be
soon
after
the
weekend
available
on
github
below
the
github
link,
and
we
are
working
now
since
several
months
towards
this
goal.
So
this
is
basically
the
first
udp
notif
message
which
we
captured
during
the
itf
hackathon
in
a
kafka
json
message,
so
this
was
like,
like
a
like
a
child's
birth
moment
for
us
that
we
finally
reached
this
goal
from
a
throughput
point
of
view
in
udp
notif.
B
To
be
honest,
this
was
like
really
a
big
thing
for
us,
because
one
of
our
big
aim
was
to
bring
performance
and
scale
into
young
push.
So
in
this
test
setup
we
had
one.
B
B
B
Now,
back
on
on
bmp,
so
on
our
test
cases,
we
were
focusing
on
the
route
reflectors
and
provider
edge.
Routers
were
json
cited
in
pre-policy,
localhip
and
adjacency
outpost
policy
was
enabled
with
path
marking
support,
and
one
of
the
key
findings
was
on
the
stress
conditions,
basically,
that
the
cpu
increased
after
the
bgp
was
converged
so
and
for
the
memory
part
there
was
a
slight
overall
increase
throughout
the
test
cases,
and
I
will
show
some
graphs
afterwards
and
what
was
pretty
impressing
and
was
also
a
big
surprise
is
even
on
the
stress
condition.
B
The
outbound
bmp
monitoring
was
always
be
very
consistent.
So
at
the
end
of
the
test
cases
the
state
mirrored
with
bmp
and
the
state
on
the
routers
were
always
matching,
so
we
did
not
lose
any
metrics.
B
Regarding
the
pgp
propagation
delay,
when
bmp
is
enabled
and
not
enabled
here,
we
we
found
out
that,
basically
that
the
testbed
was
not
accurate
enough.
The
delay
is
actually
in
a
in
a
very
low
sub
second
area
and
in
order
to
draw
a
final
conclusion,
we
need
to
improve
our
test
bed.
A
B
Then
this
is
just
basically
the
the
final
results
on
on
bmp
wherever
they
are
or
not
disabled,
and
here
basically,
you
see.
Basically,
if
bmp
is
flapping
or
not,
there
is
no
impact
on
it.
So
the
good
thing
about
what
about
this
hackathon
was.
It
was
good
well
prepared.
We
had
a
good
test
automation.
B
We
were
thanks
to
slack
and
teams
very
well
connected
same
again.
As
last
time
we
were
missing
the
beers
and
cocktails
and
thank
you
for
all
our
colleagues
who
participated
this
hackathon.
A
Thanks
thanks
for
the
presentation,
hopefully
we're
getting
closer
to
the
cocktails
and
beers.
A
I
did
have
one
question
on
the
the
time
stamps
I
didn't
quite
catch.
This
is
that
a
change
that
you
would
also
need
in
the
draft,
or
is
that
just
your
test
bed
to
test?
What's
going
on
with
the
draft.
B
Know
that,
basically,
the
draft
supports
basically
timestamps
in
seconds
and
also
in
milliseconds,
but
in
the
vendor
implementation.
It
was
not
a
q
and
it
was
not
precise
enough
to
measure
the
delay.
So
we
need
to
improve
basically
the
vendor
implementation.
A
B
A
B
This
is
this
was
one
of
the
key
findings
that,
in
order
to
measure
such
a
propagation
delay
such
time
stamp
in
such
final
candlelight
disease
is,
does.
F
A
I
don't
see
vladimir.
A
He'd
warn
me:
he
may
be
a
little
bit
late:
okay,
let's
let's
go
on
to
the
next
one,
then
and
we'll
come
back
to
vladimir
at
the
end
tom,
if
you're
here
ns3.
A
F
A
Okay,
actually,
vladimir,
I
can
see
you're
you're
here.
Are
you
ready
to
present
or
would
you
rather
we
come
back
to
you.
A
All
right:
well,
then,
we
will
go
on
to
sebastian.
G
E
E
E
Okay,
are
you
seeing
the
slide?
Yes?
Okay.
Thank
you
very
much.
This
is
my
first
ietf
hackathon.
It
might
be
a
little
bit
different
from
some
of
the
other
projects.
Big
bang
is
a
is
a
software
toolkit,
that's
for
studying
collaborative
communities
in
general
and
especially
standard
setting
groups.
So
what
it
does
it
is
it
ingests
email
get
repositories
it
ingests
stuff
from
the
iutep
ietf
data
tracker.
E
As
of
this
hackathon,
it
ingests
listserv
data
which
some
other
standard
settings
groups
use
as
a
communications
tool,
and
then
it
combines
those
with
data
science
tools
from
you
know
the
scientific
python
stack
for
entity
resolution,
social
network
analysis,
natural
language,
processing,
time
series,
analysis,
we're
expecting
to
start
working
with
a
group
that
does
information
extraction
soon
and
it's
a
little
bit
different
from
arcos
rfc
stats,
which
we
love
and
are
inspired
by
because
it
brings
in
a
wider
range
of
data,
sets
and
supports
different
kind
of
research
questions.
E
So
the
big
bang
research
community
has
both
social
scientists
that
are
interested
in
standardization
collaboration
and
the
ietf
in
particular,
and
also
computer
scientists,
developing
new
kind
of
data
science
methods
for
using
this
kind
of
data.
E
So
it's
got
a
number
of
different
groups
contributing
article
19
is
an
ngo
there's.
Several
university
researchers
that
are
involved
not
on
this
list
is
some
participation,
that's
very,
very
vision
from
the
university
of
glasgow,
we're
building
a
python
interface
to
the
data
tracker,
and
for
this
sprint
we
saw
a
number
of
new
participants
growing,
including
you
know,
a
new
user
and
a
new
software
contributor
who's
made
a
lot
of
very
great
improvements.
E
We
did
a
lot
of
just
sort
of
basic
maintenance
and
updating
of
sort
of
the
installation
procedure
to
keep
up
with
changing
underlying
dependencies.
We
produced
an
instructional
video
about
installation.
Basic
usage
did
some
debugging
and
built
in
this
listserv
sort
of
scraping
tool
which
we
want
to
apply
later
to
other
organizations
like
3gpp.
E
We
also
did
some
more
substantive
sort
of
analysis,
or
even
and
science,
with
the
data
one
about
one
project
that
nick
doty
did
was
about
attendance
analysis
and
there
was
some
other
work
about
looking
at
organizational
involvement
via
email
data,
so
this
is
from
nick
doty.
This
is
a
plot
of
ietf
meeting
attendance
by
country
over
time
and
what
he
was
interested
in
in
particular,
was
the
last
three
meetings
had
been
remote.
E
You
can
see
they've
got
some
of
the
red
x-axis
tag
and
he
was
pleased
to
see
that
there
was
increased
attendance
relative
to
prior
meetings,
because
the
meeting
was
remote,
though
not
quite
as
high
as
when
at
the
peak
of
ietf
attendance,
which
was,
I
guess
around
2004
or
so,
but
the
other
thing
that
he
noticed
was
that
you
can
see
say
in
let's
see,
meeting,
76
or
94,
where
there's
a
higher
higher
percentage
of
of
people
attending
from
japan.
That's
because
of
the
location
of
the
meeting.
E
So
we
can
see
swings
in
the
nationality
of
attendance
based
on
the
location
of
the
meetings,
the
remote
meetings.
It
seemed
to
lose
that
locality
effect,
which
I
suppose
is
expected,
but
is
something
to
keep
in
mind,
I
suppose
going
forward.
E
The
other
thing
we
really
dug
into
in
terms
of
data
analysis
was
some
of
the
research
groups
that
are
involved
in
the
project
are
very
interested
in
which
organizations
are
which
are
influential
in
which
working
groups,
and
while
the
there's
author
stats
work
that
definitely
supports
this
sort
of
stuff.
It
doesn't
generalize
to
other
standards
groups
that
don't
have
as
wonderful
a
tool
as
data
tracker,
so
we're
trying
to
explore
this
using
email
domains
for
participation
in
mailing
lists.
E
But
there
are
some
challenges:
there's
a
lot
of
people,
especially
in
the
itf
community
that
have
personal
email
domains
and
there's
also
generic
email
hosts
like
gmail
or
gmx,
which
you
know,
encompass
several
different
individuals
and
we're
trying
to
work
out
data
cleaning
methods.
To
make
this
a
little
more
useful,
we
were
experimenting
with
a
kind
of
entropy
threshold
of
email
just
on
the
distribution
of
email
addresses
per
domain,
and
we
found
this
quite
effective
at
filtering
out
personal
email
domains.
E
We're
still
trying
to
figure
out
what
to
do
about
these
generic
email
hosting
issues,
but
so
far,
we've
been
pretty
good
at
successfully
isolating
when
an
email
domain
is
coming
from
a
corporation
or
organization,
as
opposed
to
representing
like
an
individual,
which
is
what
we're
kind
of
trying
to
get
at.
So
for
the
future.
E
We
plan
to
release
some
improved
documentation,
hopefully
make
a
containerized
environment
so
that
people
could
interact
with
the
data
through
an
interactive
notebook
as
opposed
to
having
to
sort
of
download
and
install
the
whole
thing
we
want
to.
We've
got
some
plans
to
refactor
the
core
code
for
better
encapsulation,
and
we
want
to
develop
a
kind
of
standardized
process
for
looking
at
organizational
involvement.
E
So
you
can
find
out
more
about
the
project
on
our
github
page
and
there's
a
mailing
list
there,
which,
which
you
know
you're,
happy
to
join
in
and
check
out
and
join
the
discussion,
and
that
is
all
thank
you.
A
Oh
great
well,
thank
you
very
much.
That's
that's
some.
Some
new
information
I
look
forward
to
seeing
is
that
more
analysis
comes
out
of
that.
Are
you?
Are
there
plans
for
like
sharing
any
of
the
data
that
you're
generating
in
it?
I
don't
know
in
the
ihf
plenary
or
some
other
format,
it's
cool
the
hackathon
community
from
the
tooling
perspective,
but
a
lot
of
people
if
you're
interested
in
the
results.
E
So
that's
that's
a
great
that's
a
great
point.
As
I
said,
this
is
my
first
itf
attendance
and
I've
been
working
on
the
big
bang
software
for
some
time,
there's
others
in
the
big
bank
community
that
are
weren't
tied
yet,
but
I'll
bring
that
up
with
them.
I
think
we'd
love
to
present
this
at
the
plenary
cool.
A
F
I'm
here,
but
I
was
wondering
if
I
could
ask
still
a
question
on
the
previous
one:
oh
sorry,
sorry,
yeah,
yeah,
so
just
really
great
stuff,
very
interesting
and-
and
indeed
we
should
hear
it
in
a
more
general
setting
in
the
itf
plenary
or
something
I
I
did
have
one
comment
on
on
this
generic
email,
stuff
and-
and
there
are
ways
to
deal
with
that.
F
One
of
the
things
that
worked
for
author
stats
was
that
you
could
by
hand
basically
list
the
most
important
generic
email
domains
and
by
hand
at
least
the
most
interesting
companies.
Or
most
you
know,
companies
that
many
people
work
for
and
from
that
you
can
actually
override
the
email
domain
information.
So
you
could
actually
analyze
a
little
bit
of
who
that's
what
because
people
sign
their
company
instantly
in
the
in
the
documents
with
the
company
or
provide
the
company
name
in
the
documents.
So
that's
what
that's
one
way,
perhaps
that
you.
E
E
That
that
is
awesome,
and
thank
you
for
your
question.
Of
course,
we
we're
paying
attention
to
your
work.
If
there's
any
way,
we
could
reach
out
to
you
and
get
you
know
your
list
of
hand-edited
email
addresses
of
dreams.
That
would
be
incredibly
valuable,
but
I
think
that's
something
we're
thinking
about
doing
down
the
line.
It's
just
a
lot
of
work.
F
Yeah
so
so
this
team
worked
on
quick
measurements
and
quick,
of
course,
is
an
encrypted
transfer
protocol,
which
means
that
the
measuring
it
is
is
much
harder
than
it
was
for,
say,
tcp,
where
you
could
just
look
at
the
x
and
so
forth.
F
The
quick
standard
does
have
some
features
for
enabling
some
specific
or
explicit
measurements,
whereas
the
spin
bit
allows
some
things
to
be
measured
and
we've
built
some
software
in
earlier
times,
two
or
three
years
ago
already
started
work
on
something
called
spindam,
for
instance,
that
I
can
look
at
quick
flows
and
understand.
What's
right.
F
Okay,
should
I
see,
do
you
really
want
to
share
your
screen?
Yes,
I
really
want
to
share
my
screen.
A
Now
do
you
see
my
screen?
I
it
shows
your
arc
of
the
screen,
but
it
says
it's
all
black.
I
don't
know
if
anyone
else
sees
more
than
that.
No,
it
looks
like
it's
next
screen
same.
F
F
Yeah
now
I'm
seeing
a
black
now
yeah
perfect!
Thank
you.
So
we
built
the
software
called
spin
dump
to
help
analyze
some
of
these
encrypted
traffic
flows
of
the
text
that
we
can
and
we've
been
working
on
that
and
expanding
that
in
various
directions
in
in
several
hackathons
actually
over
the
last
year
and
a
half
or
so
this
time
it
was
these
people
from
telecom,
italia,
huawei,
university
of
hasselt
and
ericsson
who
participated
there
for
ticker.
The
next
slide.
F
And
so
we
worked
basically
on
four
things
this
week.
The
first
one
is
like
new
research
on
different
types
of
measurements
that
we
could
do
with
encrypted
protocols
and
there's
a
there's,
a
draft
and
there's
an
implementation
of
the
draft
in
in
spin
dump
integrated
into
the
software.
You
can
get
go.
Take
a
look
at
the
draft.
I
won't
go
into
details
of
that
today
to
save
some
time.
F
Instead
of
having
the
network
do
this,
these
measurements
and
one
one
thing
that
we
actually
were
surprised
by
a
little
bit
in
this
effort,
was
that
that
it's
it's
quite
possible
to
measure
quick
flows,
at
least
when
there
is
the
spin,
but
spin
bit
is
being
used,
but
it's
quite
difficult
to
use
or
to
observe
tcp
for
the
reasons
that
some
of
these
platforms
hide
some
of
the
some
of
the
traffic
from
from
us
on
the
phone
platforms
and
so
forth,
and
and
it's
it's
not
possible
to
see
all
packets
even
with
this
like
packet
capture
interfaces
or,
alternatively,
one
would
need
more
permissions
to
to
see
everything,
but
so
it's
been
kind
of
an
interesting
side
effect
of
this
encrypted
protocols
that
they
actually
go
through
everything
on
the
points
and
internet
work
and-
and
you
can
see
what
you
can
see.
F
F
We
also
do
quite
a
bit
of
sort
of
monthly
maintenance
of
the
software,
so
we
need
to
update
it
usually
with
new
versions
of
quick
now.
Fortunately,
the
quick
protocol
is
being
released
as
rfc,
so
the
version
has
been
set
as
one
and
and
that's
it
for
the
moment
at
least,
but
we
did
integrate
a
few
pr's
and
fixed
a
number
of
issues
added
some
features.
Some
interrupt
testing
and
one
of
those
features
that
we
added
is
is
q,
log
and
q.
F
Lock
is
like
this
logging
format
that
is
supported
by
many
endpoint
implementations,
so
quick,
it's
mostly
used
initially,
at
least
for
for
debugging
purposes.
There's
a
couple
of
drafts.
You
can
go,
read
them
and
we
chose
to
add
q
logos
as
a
possible
output
format
from
from
spin
dump,
and
the
idea
is
basically
that
you
would
get
information
from
several
directions:
you'll
get
it
from
the
host
implementations,
you'll,
get
it
from
network
tools
such
as
spindom.
You
would
get
it
from
the
server
implementations
and
then
you
can
put
all
of
this
thing.
F
F
So
this
is
kind
of
sounds
may
be
complicated,
but
it's
quite
simple:
quilo
clock,
q
lock
is
jason
essentially,
and
it
consists
of
a
basically
an
array
of
events,
client
events
and
then
there's
information
about
the
protocols
and
and
so
forth,
and
what
we've
been
able
to
add
there
is
is
measurement
events
and
there's
various
kinds
of
measurements
that
spin
dump.
Does
this?
What
is
shown
on
the
screen
is
just
one
example.
F
Well,
that's,
that's,
basically
it
and
looks
interesting.
I
think
we
we
are
going
to
have
to
stop
the
coding
for
a
moment
and
think
about
some
of
the
drafts
going
forward.
The
class
are
being
discussed.
I
think
tsv
wg
in
the
coming
week
by
robin,
but
but
since
they
started
from
this
logging
debugging
perspective
and
not
from
the
measurement
perspective,
then
building
those
measurements
capabilities
into
the
formats
will
probably
take
a
little
bit
of
effort.
So
we
sort
of
identified
opportunities
for
improving
the
drafts
or
writing
the
awards.
Perhaps
so
that's
it.
A
F
You
you
can
identify
a
particular
quick
connection
or
a
pair
of
connections
going
between
two
parties
and
but
like
what,
in.
F
A
Oh,
it's
me
okay,
so
karsten,
unfortunately,
couldn't
be
here
as
no
one
from
the
asdf
wishy
team
was
able
to
be
free
at
this
particular
time
slot.
So
I
just
have
a
very
quick
update
from
them.
Let
me
pull
that
up
and
I'll
share
it
with
all
of.
A
D
A
So,
on
behalf
of
the
asdf
wishy
team,
just
wanted
to
give
you
a
a
brief
update.
A
So
this
project-
it
was
it's
a
joint
project
between
the
asdf
group
and
then
the
chairs
from
that
group
running
it
along
with
the
tdtrg
group
and
that's
really
where
wishy
is
being
discussed,
and
so
the
goal
for
them
was
to
validate
sdf
1.1.
I
believe
at
the
last
hackathon
it
was
1.0.
They
were
working
with
the
semantic
proxying
trying,
mapping
files
etc
a
link
to
their
github
code.
Wiki
can
get
more
details
a
rather
exploratory
hackathon
is
how
karsten
phrased
it.
A
But
you
know
two
pieces
of
code
did
come
out
of
that
that
that
will
aid
with
model
transformation,
so
the
good
news
is
these.
The
results
of
the
hackathon
are
already
expected
and
planned
to
be
discussed
in
more
detail
in
the
actual
working
group
and
research
group
session.
So
I
have
them
listed
they're
both
on
thursday.
A
And
with
that,
I
think
we
will
go
back
to
our
next
one,
which
paul
we
have
a
couple
presentations
from
you
and
great.
I
see
you're
here.
G
A
Okay,
we
can
see
it
we're
on
your
wrap-up
slide,
though,
and
now
it
just
switched
to
presentation
mode,
which
you
know
may
not
be
the
best
format.
A
A
That's
good
and
then
just
one
more
thing:
there
is
a
little
pop-up
near
the
bottom,
which
maybe
you
can
just
move
that
it's.
D
H
I
will
share
our
hackathon
project
I
to
nsf
this
time,
so
this
is
our
poster.
H
So
this
time
we
want
to
implement
itunes,
f,
monitoring
data
model
and
also
we
want
to
show
the
approval
concept
of
security
management.
Automation,
new
draft-
so
especially
this
draft
proposed
application
interface
for
delivering
feedback
information
from
itunes
to
analyzer
to
security
controller.
H
So
basically,
this
is
a
typical
i2
nsf
framework.
So
this
time
denial
of
service
attacker
generate
the
tcp
plot
traffic
to
web
return,
network
security
function
a
and
then
analyzer
collect
the
data
from
nsf
a
and
then
using
a
machine
learning
ml.
It
detect
nsfa
is
affected
by
dios
attacker
and
then
using
application
interface.
H
H
So
basically
what
we
got
done.
So
we
implemented
the
monitoring,
interface
and
then
analyze
the
dos
attack.
Also,
we
deliver
the
pwa
information
over
application
interface
and
then
security
controller,
translate
high
level
security
policy
into
a
lot
of
very
skilled
policy
and
then
deliver
that
new
skill
policy
over
nsf
phase
interface
and
then
nsf
using
firewall,
and
then
it
can
drop
the
tcp
flat.
H
So
this
time
we
show
the
feasibility
of
application
interface.
This
is
nearly
a
proposed
by
itunesf
working
group
for
security
management,
automation.
So
next
step
we
have
five
items.
We
want
to
extend
monitoring
young
data
model,
monitor
packet
flows
and
also
we
want
to
use,
as
pro,
along
with
the
itunes
monitor
interface,
to
observe
that
traffic
monitoring.
H
Now
this
is
our
ito:
nsf
open
source
project
at
github,
so
this
is
a
new
repository
for
at
this
time.
So
this
is
my
team
okay.
So
this
is
a
collaboration
work
with
the
sunshield
university
from
professor
young
kim's
lab
and
from
actually
doctor
zhang,
chobak
and
mr
ninja
choi,
and
we
have
a
student
so
patrick
he
is
a
main
player.
This
hakason
and
dr
jinyoung
kim
and
jung
young,
kim
and
yosef
ann
are
my
student
and
my
phd
and
also
I
have
more
moji
school.
H
He
is
an
intern
from
liberty
university
in
the
united
states
and
also
I
have
a
visitor
student
gyeongshi
kim
from
kyungbung
national
university.
So
we
work
together
with
ipwave
and
vmware
working
with
teams.
So
those
are
our
sponsors.
Thank
you
for
listening.
Listening,
your
mi
presentation.
Thank
you.
Any
questions
or
comment.
A
I
A
H
H
H
So,
okay,
charles
can
I
move
to
next
presentation.
A
Yeah,
I
think
so,
unless
did
anyone
else
have
any
other
questions,
otherwise
yeah
we'll
move
to
the
next
one.
I
don't
see
anyone
in
the
queue
so
go
ahead.
Okay,
great.
H
Okay,
this
is
a
if
you
wave
a
basic
protocol
project.
My
name
is
jam
for
zhang,
I'm
from
skk
or
sungyeong
university,
so
this
is
a
foster,
but
basically
previously
we
had
three
hackathons.
The
first
one
is.
We
demonstrated
the
ip2006
to
packet
transmission
over
ocb,
wi-fi
modules
and
also
the
108
case.
We
implemented
simulation
of
context,
aware
navigation
protocol
and
last
hackathon
109.
H
We
started
the
implement
context
of
a
navigator
protocol,
so
this
time
we
use
the
r1
i1
lobotics
robot.
It
can
be
controlled
manually
or
autopilot
mode.
So
this
time
he
manually
controlled
next
time
we
tried
to
autopilot
mode,
so
it
sport
the
linux
corner
in
terms
of
real-time
linux.
So
it
also
like
to
be
a
sport.
H
H
Those
are
our
previous
components
and
the
internet
draft.
So
basically
our
contact
over
navigation
protocol
based
on
a
neighbor
discovery
message
option
is
called
the
beaker
mobility
information
option
vmi.
So
this
is
a
and
the
option.
So
we
support
the
corporation
context
messaging
and
the
emergency
context
message.
H
So
basically,
we
implemented
and
tested
using
a
robot
r1
a1
robotics,
so
they
exchanged
a
message,
ccm
and
the
ecm.
But
this
time
we
you
implemented
the
udp
instead
of
i76.
This
hackathon
will
try
to
implement
icmp
version
6..
So
we
try
to
extend
the
ubuntu
linux
corner
for
nd.
It
turns
out.
There
is
some
high
complexity,
so
we
just
yeah
designed
how
to
extend
the
neighbor
table.
H
That
is
the
neighbor
cache
and
also
it
has
entry
for
neighbor
using
neighbor
structure.
So
we
try
to
take
advantage
of
the
current
implementation
and
we
want
to
extend
the
and
the
option
so
still
ongoing
work.
Hopefully
next
hackathon
we
try
to
finish
so
next
step.
We
will
continue
developing
and
debugging
a
new
option
called
the
speaker,
mobility,
information
option
in
linux,
corner
and,
and
also
we
also
will
export
context
over
navigation
protocol
based
on
this
new
option
and
then
we'll
test.
H
This
ip46
neighbor
discovery,
and
also
we
will
sport,
the
autopilot
mode
of
robot
cars,
such
as
cmp
contacts
or
where
navigation
protocol
max
is
changed
such
as
there
is
some
accident
over
optical
ahead
of
it.
Speakers
can
exchange
the
message
and
they
can
avoid
that
obstacle
or
the
collision
with
each
other.
So
this
is
a
ip
wave
open
source
project
at
guitar.
H
Also,
this
also
collaboration
work
with
the
social
university
professor
young
game
and
my
pitch
student
vienna
me
and
chris
and
don
they
are
my
phd
students
and
son
from
sunshine
university,
and
you
know
national
university.
In
the
same
way,
we
work
together
with
the
itunes
sf
and
the
vm
working
group
teams.
H
A
Okay,
great
next
presentation,
ratson
company.
A
D
Okay,
okay,
let
me
try
and
see
okay,
so
we
came
here
with
a
verizon
project.
The
reason
project
is
a
project
we
launched
four
or
five
months
ago,
with
the
goal
of
building
a
bunch
of
open
source
packages
that
can
be
composed
into
another
station
verification
service.
These
packages
tend
to
cover.
You
know
several
different
facets
of
data
station
verification
pipeline.
D
So
we
have
data
formats
we
have
apis
to
vacations
to
with
each
other.
We
have
evaluation
of
the
appraisal,
policy
endorsement
and
trust
anchor
configuration
storage,
query
whatever
all
that
stuff.
So
this
is
the
second
microphone
we
do
one
verizon
and
last
time,
itef
109.
D
The
other
thing
we
wanted
to
do
was
bringing
up
to
speed
our
8
implementation
with
with
the
latest
version
of
the
working
group
draft-
and
this
was
very
important
to
us,
because
the
latest
version
contains
a
new
claim
that
is
critical
to
our
workflow
and
we
wanted
to
play
with
it
a
bit
and
see
if
it
worked
right.
D
So
the
thing
so
worked
on
four
drafts:
three
working
group,
crafts
and
one
individual
draft,
which
is
the
draft
strategic
psa
token,
which
we
want
to
progress
through
the
independent
stream.
D
We
did
a
considerable
amount
of
development
and
integration
in
four
different
packages
of
the
verizon
project,
and,
apart
from
that,
we
spent
a
lot
of
time
discussing
new
ideas
about
you,
know
missing
tools
and
and
and
stuff
like
that
or
evolution
of
the
api
implementation
to
cover
other
scenarios.
We
have
you
know
pr's
that
have
been
merged.
D
We
have
issues
that
have
been
opened
and
all
that
stuff,
wherever
demo
we've
put
together
a
demo
based
on
docker
that
you
could
run
if
you
wanted
to-
and
it's
there
at
that
url
at
the
bottom
of
the
page-
and
we
did
not
do
any
interrupt
for
now,
but
we
chatted
with
another.
It's
stack
to
to
decide
the
sort
of
inter
test
which
we
wanted
to
perform
and,
and
we
agreed
with
them
to
to
schedule
a
an
interrupt
test
soon,
maybe
a
virtual
one
instead
of
waiting
for
the
next
cycle.
D
But
what
we
learned
is,
as
I
said,
we
wanted
to
have
this
profile
claim
implemented,
and
we
did
it
thanks
to
yogesh
and-
and
that
was
very
interesting
because
we
realized,
while
doing
that,
that
the
spec
as
it
as
it
stands,
is
not
a
hundred
percent
correct
and
we
need
to
have
more
precision
around
interesting
bits
of
it
and
so
there's
a
there's,
a
new
issue
that
has
been
raised
and
there's
some
discussion
already
ongoing
and-
and
I
think
that
this
would
be
a
will-
be
in
fact
some
kind
of
feedback
that
needs
to
be
taken
into
the
youtube
discussion
next
week
to
wrap
up
yeah.
D
The
team
members
is
jokers
myself
in
sergey
and
yogesh
was
a
first
timer
and
I
think
he
enjoyed
it
very
very
much
as
we
did
that's
it.
Thank
you.
A
Okay,
well
great
glad
you
found
that
issue
in
the
draft
now
rather
than
afterwards.
Yes,
exactly
what
we
want
to
do
so
all
right,
I'll
talk
to
the
team
for
that,
any
anyone
with
questions.
A
A
J
I
A
Yeah,
let
me
let
me
do.
A
A
All
right,
I'm
just
looking
for
the
the
right
one
here.
I
J
Please
yeah.
So
the
main
goal
of
this
hackathon
is
to
figure
out
the
container
talking
performance
impact
by
the
various
resource
options
so
based
on
our
draft,
so
we
set
up
two
main
features.
The
first
is
implementing
containerized
infrastructure
with
the
various
network
models
and
second,
is
verifying
performance
impact
depending
on
the
various
configuration
settings
next
slide.
Please.
J
Yeah
so
the
last
two
are
the
hackathon,
so
we
set
up
the
environment
and
then
test
some
new
technologies
and
network
networking
models,
and
then,
at
the
time
in
this
hackathon
we
first
our
first
vlog
is
to
discuss
about
the
current
status
of
ventrac
for
in
the
containerized
infrastructure,
benchmarking
area.
J
So
the
we,
the
survey
and
then
checking
the
other
walks,
not
only
our
our
works,
so
the
one
of
them
is
the
the
open
source
activities
which
such
as
the
bs
puff
in
the
opnfp
and
another
part,
is
that
the
industry,
the
benchmarking
test
results
such
as
the
fdiu
and
then
the
pdk
performance
report.
J
So
based
on
that,
so
we
are
mapping
our
defined
networking
model
category
and
the
test
scenario
to
map
to
the
to
to
the
other
activities
and
then
based
on
that,
we
want
to
generalize
our
test
architecture
and
then
some
the
test
cases
and
the
next
slide.
Please.
J
Benchmarking
works
into
the
our
lab
environment,
so
first
is
using
the
obs
dpdk
and
then
ppp
memory
and
sriv,
and
we
set
up
this
environment
and
then
we
measure
the
through
food
and
the
latency,
and
the
package
drops
in
the
some
single
port
to
the
external
traffic
generator
and
the
other
is
that
the
multi-post
test
scenario,
which
is
supported
in
the
vpp
memimp
so
and
then
based
on
this
multiple
port
case
and
then
size
function,
chaining
case.
J
We
follow
that
scenario
and
and
we're
validating
that
and
at
the
same
time
we
using
the
the
other
or
the
technologies
such
as
ostpdk,
with
the
same
testing
scenario.
So
we
we
are
comparing
with
the
existing.
The
test
result
for
us
compare
their
performances,
so
these
works
are
still
on
progress
and
then
we
will
get
the
results
soon.
Okay,
next
slide.
Please.
J
So
the
based
on
the
discussion
and
then
validator,
so
we
need
to
update
our
related
draft
to
rearrange
our
chapters
and
figures
and
words
of
it's
more
need
to
be
more
generalized
and
then
up-to-date
technologies
and
adding
the
other
the
benchmarking
activities
and
then
for
the
next
hackathon.
So
we
are
finding
a
remaining
area
for
the
benchmarking
containerized
benchmarking
area.
J
So
the
first
we
choose
is
we
choose
multiples
in
a
single
host
case
or
with
the
different
acceleration
techniques.
J
So
there
is
some
two-part
one
partition,
multiple
to
market
report
and
apparently
we
send
to
the
packet
to
the
multiple
two
multiples
and
then
another
one
is
the
training
scenario
which
is
traffic
should
be
forwarded,
sequentially
to
or
through
the
multiple
ports
and
the
other.
The
testing
scenario
case
can
be
the
port
or
with
multi
multi
interfaces,
so
the
using
when
the
port,
using
the
multi
interfaces
and
then
going
traffic
simultaneously
through
the
multi
interfaces.
J
So
we
we
want
to
the
the
venture,
came
for
them
and
and
want
to
find
some
issues
of
which
impacts
to
the
performance
of
the
containerized
infrastructure.
Okay,
next
slide.
Please.
J
Okay,
so
here
is
our
team
members,
so
the
professor
young
kim
and
then
some
guys
in
the
iistrc,
the
research
center
and
then
professor
jayang
poor
zhang
from
sushi
san
jung
university
and
then
dr
henshi
yang
from
the
kt.
So
this
is
our
member
and
then
thank
you
for
listening.
A
Okay,
well
thanks
for,
for
sharing
and
for
the
presentation,
any
questions.
Thank
you.
G
Problem,
let
me
see.
G
So
just
like
previous
time
we
used
this.
I
used
the
metalmost
chat
surfer
of
the
dns
operations
in
analyzers
research
center,
because
quite
a
lot
of
open
source,
dns
developers
already
there.
So
it's
a
convenient
and
appropriate
place
to
communicate,
and
I
it's
also
open
to
anyone.
So
anyone
can
join
in
a
chat
on
about
dns
hackathon
projects
they
might
have.
G
I
have
a
link
on
the
hexon
wiki
as
well,
but
so
one
month
ago
I
changed
the
channel
name
into
itf
110,
dns
hackathon,
and
just
when
I
did
that
people
started
to
complain
a
little
bit
right,
because
it's
a
pity
that
we
have
to
do
online
hackathon
again
and
it
would.
It
would
be
better
to
have
the
in-person
hackathon
and
you
know,
with
online
hackathon.
I
only
have
15
minutes
three
times
over
the
week
to
spend
on
this.
G
So
I
decided
well,
why
don't
we
organize
in-person
hackathon
day
at
the
netherlab's
office,
and
I
proposed
that
on
the
channel
and
most
people
were
hesitant,
but
at
least
one
person
was
enthusiastic
about
it,
which
is
the
dice,
and
so
we
will
have
an
in-person
hackathon
day
tomorrow
at
the
email
network's
office.
G
G
G
It's
now
in
ad
review
state
about
having
the
zone
the
dns
zone
transferred
over
tls,
and
so
there's
a
lot
of
things
that
are
involved
with
that.
Like
authentication
that
you
don't,
if
you
really
want
your
zone
to
be
private,
you
may
not
leak
it.
G
In
in
clear
text
basically,
so
there
are
some
things
you
have
to
watch
out
for
and
related
to
salt.
This,
the
transfer
of
tls
is
also
the
extended
dns
errors,
which
is
a
part
of
salt.
If
you're
doing
zot,
they
assume
you're
doing
extended,
dns
errors
too,
but
actually
not
that
many
implementations
yet
of
extended
dns
errors,
at
least
not
in
our
software
and
also
not
in
bind
I've
heard
today.
G
Sarah
dickerson
set
up
this
excellent
hackathon
planning
document,
in
which
you
know
everything
surrounding
salt
was
mentioned
and
other
other
steps
and
plants.
Unfortunately,
sarah
herself
got
ill
this
week,
so
she
could
could
not
participate
but
chiffon
palavi
and
han.
They
were
a
unit
that
worked
from
east
coast
united
states.
I
think
they
had
a
mini
group
in
which
they
talked
over
a
google
meet
at
an
elnett
lobs.
We
have
new
guy
named
tom
capari.
He
worked
on
ede
and
all
the
work
from
pallavi
hung
and
chifan
was
reviewed
by
walter
weinhardt.
G
G
So
what
got
done
tom
capari?
He
made
extended
dns
errors
for
anestheda
our
authoritative
name
server
from
analyte
lobs,
and
this
was
quite
interesting.
He
found
some
missing
error
codes
in
the
rfc
so,
for
example,
a
specific
error.
If
your
d
name
expansion
overflows,
if
you
end
up
with
a
too
large
domain
name,
this
is
possible.
It
has
a
very
unspecific
response
card.
Currently
pallavi
worked
on
access
control
lists.
G
So
what
did
we
learn?
Learn
we
all?
We
still
miss
all
miss
the
in-person
hackathons
that
will
have
a
mini
one
tomorrow,
yeah
just
me
and
matthias.
Basically,
but
that's
I'm
very
much.
Looking
forward
to
that
from
ede,
maybe
we
need
a
update
on
8914,
or
maybe
we
come
up
with
some
more
codes
that
are
missing
from
that.
So
that's
nice
and
swot
just
works
rocks,
so
that's
all
good.
G
G
A
Yeah
thanks
willem
tom
you're
you're
in
the
queue
is
that
for
the
next
presentation,
or
do
you
have
a
question.
A
Otherwise,
william,
I
thanks
you
made
dns
and
privacy
and
security.
Look
like
a
lot
of
fun.
I'm
glad
you
have
all
that
infrastructure
in
place.
I
think
that'll
help
with
your
your
in-person
hackathon,
I'm
a
bit
jealous.
I
have
to
say-
and
I
hope
that
you
know
we
don't
know
about
ihf
111,
yet
I
guess
there's
still
a
chance
to
do
something
in
person,
so
we'll
see
we're
working
on
it.
H
A
Okay,
okay,
great
thanks
and
we'll
turn
it
over
to
tom
tom.
If
you
request
screen
sharing,
you
should
be
able
to
share
your
presentation.
K
A
A
A
A
Yeah,
let
me
let
me
find
your.
A
Okay,
let's
see
what
am
I
looking
for
here,
all
the
ns3
models.
A
K
A
K
K
K
Okay,
so
yeah
sorry
for
the
technical
dif
difficulties,
so
my
name
is
tom
henderson
and
I'm
here,
representing
typically
the
group
that
works
on
ns3
simulation
models
in
the
context
of
the
iutf
hackathon,
and
I
I
think
we've
been
doing
this
for
five
or
six
ietfs
now
in
a
row.
K
K
K
K
So
I
would
share
the
links.
Maybe
after
the
after
the
presentation
people
are
interested,
can
download
the
slides.
All
of
our
code
is
uploaded
to
git
lab.
Our
finalized
proposals
have
been
uploaded
to
the
mainline
ns3
gitlab.com
instance.
K
And
I
would
share
some
preliminary
results
of
alignment
between
our
code
and
in
linux
results,
but
I
guess
that
will
have
to
wait
or
you
could
download
the
presentation
at
a
later
time
in
terms
of
what
we
learned
this
week.
I
think
it's
really
there's
nothing
that
would
change
any
of
the
existing
drafts
or
rfcs.
K
So
it's
just
a
capability
to
be
advertised
to
the
transport
area,
working
group
for
people
who
wants
simulation
models
of
this
work
and
it's
just
two
credits.
We
did
have
one
students,
sachin
nayak,
who
was
his
first
hackathon
and
I
worked
with
him
a
lot
on
wednesday
and
otherwise
it
was
just
me
working
this.
A
Week
any
questions
for
tom.
A
All
right,
well,
tom,
thanks
for
sharing
and
we'll
we'll
get
the
you
know
where
you're,
oh
it
looks
like
there
is
a.
Is
there
a
question
and
a
chat?
Yep
yuri
was
just
telling
you
good
stuff
tom.
Thank
you
yeah.
A
I
agree
we'll
we'll
get
it
straightened
out
where
your
presentation
is
and
make
sure
it's
it's
it's
been
uploaded
if
it's
not
there
already,
so
that
everyone
else
can
find
it,
and
with
that,
let
me
see
if
we
have
a
vlad,
because
we
skipped
over
vlad
earlier
as
well,
for
a
yang
model
for
network.
B
A
Get
things
working
and
I
met,
and
that's
what
he's
doing
now
so
we'll
prob
he'll
probably
have
something
to
share
later
on
I'll,
encourage
him
to
just
send
what
he
has
to
the
list
and
an
update
to
the
list,
letting
people
know
what,
if
he
uploads
a
presentation
that
they
can
go
and
grab
it
from
there.
A
Reloading
me
protecto
and
coming
back,
it
seems
like
I
can
share
now,
but
sorry
I
was
gone
for
a
couple
minutes
there,
so
I
think
we
got
through
all
the
presentations.
Is
there
tom?
I
still
see
you
in
the
queue,
but
is
that
from
your
previous
presentation
or
did
you
have
something
to
add.
F
H
A
Great
well,
I
believe
we
got
through
all
the
presentations
just
a
reminder.
They
should
all
be
here
in
the
110
project
presentations.
If
any
of
you
have
updates,
because
you
continue
to
work
on
your
project
over
the
course
of
the
next.
You
know
week
or
two
or
you
know
all
the
way
up
until
the
next
hackathon.
That's
fine!
You
can
always
update
what's
here.
If
there's
something
you
think,
there's
useful
to
share
in
terms
of
your
results
from
from
this
hackathon.
A
A
They've
supported
us
throughout
this
whole
virtual
online,
only
hackathon,
and
we
really
appreciate
that
and
look
forward
to
having
them
and
welcome
new
sponsors.
I
don't
know
if
any
of
you
follow.
I
don't
know
if
it
was
on
the
ietf
list,
there's
a
new
structure
for
how
sponsorship
works.
They
specifically
call
out
things
around
running
code,
of
which
the
hackathon
is,
of
course,
one,
but
there's
other
things
too,
like
the
code
sprint
and
you
know
other
ways
to
get
involved
with
code.
A
So
hopefully
that
will
help
and
we
certainly
welcome
more
sponsors,
especially
when
we
get
back
to
in-person
meetings.
Those
are
a
bit
more
expensive
for
us
to
run
so
we
can
have
the
food
and,
of
course,
the
beer
and
cocktails
that
everyone
mentioned.
A
I
just
want
to
remind
you
again,
please,
for
those
of
you
who
aren't
here
at
the
very
beginning
of
today's
closing.
We
do
have
a
survey
for
the
hackathon
and
it'd
be
great.
If
you
could
take
a
few
minutes.
This
is
just
for
the
hackathon.
I
think
it's
like
10
questions,
maybe
even
fewer,
depending
on
how
you
answer
the
first
few
and
that'll
help
us
we'd
love
to
have
feedback.
A
That'll
help
make
future
hackathons
better,
whether
they
be
online
or
in
person,
and
then
also
I
mentioned
the
the
draft
that
I've
put
together
that
covers
the
hackathon
really
from
a
perspective
of
how
to
run
it
so
that
you
know,
I
don't
anticipate
being
a
hackathon
co-chair
forever
and
I
think
barry
has
other
things
that
he's
probably
going
to
do
so.
A
In
any
case,
it's
like
everything
else
in
the
itf
we
like
to
have
drafts
that
cover
what
we
do
and
this
one
specifically
the
reason
it
came
about
now
and
it's
in
the
schmoo
working
group
is
because
you
know
with
the
nature
of
having
to
meet
online
because
of
the
current
circumstances
and
other
circumstances.
That
may
happen
in
the
future.
A
It
forced
us
to
rethink
a
bit
about
how
we
run
ietf
meetings
and
certainly
the
hackathon,
and
so
this
isn't
specific
to
running
it
online.
It
really
covers
all
facets
of
the
hackathon,
and
so
especially
for
those
of
you
who,
just
you
know,
you
know,
participated
in
this.
One
things
are
fresh
in
your
mind:
it'd
be
great
to
get
some
feedback
on
that
draft.
A
From
from
that
perspective
as
well,
because
I
have
my
view
of
how
it
works,
but
but
would
love
for
you
to
point
out
things,
especially
things
that
you
think
are
important
and
that
aren't
mentioned
in
there
at
all
that
I
need
to
call
out
so
that'd
be
great.
It
will
be
discussed
in
the
shmu
working
group
meeting,
which
I
forget.
What
exactly
which
day?
That
is,
I
think,
it's
wednesday
or
thursday
next
week.
A
And
with
that,
we're
going
to
wrap
it
up
and
certainly
hope
to
see
all
of
you
next
time
and
hopefully
in
person,
as
we've
discussed,
also,
if
anyone
has
any
questions,
any
thoughts
happy
to
to
take
those
now
as
well,
and
actually
I
have
one
question
for
the
group-
I
see
some
people
have
already
had
to
drop
off,
but
we
had
a
new
template
this
time
that
people
could
use.
We've
had
a
powerpoint
version
of
that
template
that
I
know
many
of
you
have
used
in
the
past.
A
I
think
sarah
put
that
together
for
us
initially
and
it's
been
very
helpful
and
this
time
benson
oh
yeah,
I
see
benson's
on
he
helped
put
together
or
he
did
put
together
an
html
version
of
that
template,
which
you
know
we
talked
about
the
kickoff
just
wondering
if
anyone
did
use
that,
if
you
did
love
to
hear
about
how
that
experience
went
and
as
with
everything
with
the
hackathon,
we
look
to
make
it
better
next
time
so
enhancements
to
the
templates
new
templates.
A
Appreciated,
okay,
well
I'll
close
them
by
by
thanking
all
of
you,
and
you
know,
hope
you
have
a
good
weekend,
rest
up
and
hopefully,
a
busy
but
productive
ietf
week
next
week.
I
look
forward
to
seeing
everyone
there
and
thanks
a
lot
for
participating
in
the
hackathon,
really
appreciate
it,
and
good
luck
with
the
rest
of
ietf
110.