►
From YouTube: IETF94-TCPINC-20151104-1300.webm
Description
TCPINC meeting session at IETF94
2015/11/04 1300
B
B
D
To
Tippie
wait:
two
are
the
new
chairs,
so
my
name
is
MIA
coup
de
vent
and.
D
So
welcome
to
Yokohama
and
first,
of
course,
I
show
you
the
knot,
note
well.
That
applies
to
the
session.
So
please
read
it
carefully
and
be
aware
that
everything
you
say
applies
here
to
the
node
well
and
then
I
have
this
very
nice
light
with
some
information
for
those
people
who
are
participating
over
media.
We
have
to
sweat
your
queue
system,
so
we
would
try
to
look
at
it
in
monitor
and
try
to
get
here
and
if
you
want
to
say
something
and
there's
also
scream
big
screen
in
the
room.
D
D
So
the
plan
is
not
to
standardize
both
of
them,
because
our
charter
aims
basically
for
one
solution
and
if
we
going
to
standardize
two
incompatible
solutions,
then
there's
like
the
goal
of
reaching
up
putting
opportunistic
encryption
is
going
off
our
way,
because
then
they
might
not
be
able
to
talk
to
each
other.
So
there's
clearly
go
to
come
to
one
kind
of
solution
and
for
my
point
of
view,
there
are
three
ways
to
reach
it:
either
both
approaches
converge
into
each
other.
D
There
was
like
one
proposal
which
was
around
for
a
while
had
implementations
and
like
all
the
details
and
another
proposal
where
the
description
was
very
rough,
but
they
are
catching
up
so
there's
more
information.
They
are
working
on
implantation,
so
maybe
there's
a
chance
that
we
can
make
a
decision
at
some
point
and
there
is
a
third
way
where
we
kind
of
standardized
both
of
the
proposals,
but
we
make
one
of
them
mandatory.
So
we
have
a
default
way
to
talk
to
each
other
and
the
other
one
would
be
something
like
a
version.
D
E
To
put
another
perspective
on
this,
interoperability
is
a
mandatory
non-negotiable.
There
is
more
than
one
way
to
get
there
from
here,
as
shown
on
the
three
options
in
the
screen.
One
of
the
IHF
are
modular
principles
is
rough
consensus
and
running
code.
We've
had
quite
a
bit
of
revenge
with
rough
consensus
is
time
to
lay
a
little
more
emphasis
on
running
code
that
actually
solves
a
problem.
D
So
what
we
are
planning
to
real
to
reach
our
goals
from
here?
If,
first
of
all,
we
want
to
assign
Shepherd
to
all
of
the
working
documents
very
soon,
so
we
get
a
little
bit
more
feedback
here
and
then
we
are
basically
looking
for
a
second
editor
for
the
TLS
solution,
because
having
more
than
one
eye
on
there
and
what
it
more
than
one
person
working
on
the
document
could
speed
up
the
process.
D
So
if
you
like
to
volunteer
here,
let
us
know-
and
the
third
thing
which
is
important
is
we
would
like
to
have
expert
reviews
like
in
February,
mid
or
end
of
februari.
So
we
would
like
to
see
both
proposals
in
the
state
where
it
actually
makes
sense
to
have
expert
reviews
and
get
some
more
feedback
from
those
experts,
and
those
experts
are
will
be
experts
from
the
change
foot
area,
as
well
as
from
the
security
area,
and
we
work
with.
Has
this
with
our
80s
to
get
those
experts
in.
D
There
is
a
very
good
or
the
TCP
crypt
already
has
been
played
implantation
around
for
quite
a
while
for
different
operation
systems
and
user
space,
and
they
updated
the
implementation
and
I
believe
they
are
working
on
the
kind
of
implementation
as
well.
So
there
is
also
an
old
and
outdated
current
implementation
I
believe
they
try
to
update
it
and
you
guys
might
just
go
to
the
mic.
If
you
want
to
say
something
on
this,
but
I
think
you
also
give
some
information
on
this
later
on
your
talk.
D
F
Eggert
I'm
going
to
get
to
my
point
edema,
but
I
want
to
speak
here,
so
we
are
interested
in
implementing
something,
but
we're
not
going
to
commit
resources
on
something
that
we
might
eventually
throw
away.
So
we
would
I
would
like
one
thing
that
I
can
tell
my
people
follow
this
implement
this.
You
know
this
is
what
the
ITF
will
do.
Two
is
one
too
many:
I
have
a
technical
preference,
it
doesn't
matter
overarching,
Lee
I
want
one
and
I
want
one
now,
because
otherwise
this
is
irrelevant.
F
We
can
mean
we
can
pretend
a
wonder
will
happen
and
they
will
merge,
but
but
I
don't
see
it
happening.
I,
don't
I
really
care
how
we
make
a
decision.
I,
don't
really
care
what
the
outcome
is
anymore,
but
I
don't
want
to,
because
otherwise,
unless
there's
one
we're
not
going
to
implement
it
and
yeah.
D
F
D
There
are
two
points.
First
of
all,
you
can
still
implementing
Supino,
because
that's
a
common
thing
and
that's
part
of
the
implementation
effort
and
maybe
changing
from
one
to
the
other
solution.
Is
you
know
not
such
a
big
effort?
Actually,
but
that's
your
decision.
The
other
point
is
also
that
we
are
not
we
won't.
We
won't
hold
up
this
group
for
one
of
the
solutions.
G
Yeah
I
wanted
this
is
David
Mazouz
from
Stanford
I'm,
one
of
the
co-authors
of
TCP,
crib
and
and
I
want
to
kind
of
further
the
point
that
that
the
existence
of
another
draft
makes
it
makes
it
very
hard.
So,
first
of
all
the
protocol,
that's
not
implemented
is
always
going
to
sound
like
the
best
protocol,
because
you
can
just
say
things.
You
haven't
worked
out
all
the
details,
so
we've
you
know
been
working
on
this
for
years
and
like
in
response
to
feedback.
G
So
one
of
the
things
that's
happening
to
us,
for
example,
is
like
our
funding
is
over
because
we've
had
we've
been
working
on
this
for
years
and
we're
still
not
done
so.
If
this,
certainly,
if
there's
anyone
in
the
room,
who's
interested
in
who's
employer
might
be
interested
in
making
a
tax-deductible
contribution
to
Stanford,
don't
support
this
work
like
that
would
be
fantastic
and
I'm.
That
means
funny,
but
I'm
actually
not
joking
about
that,
but
but
the
other
thing
is
yeah.
G
And
then
you
know
if
in
a
month
there's
a
implementation,
the
other
one
well
great,
then
we
can
go
back
to
where
we
were,
but
if
there
isn't,
then
it
won't,
if,
like
at
least
would
want
to
like
slowed
us
down
because
what's
happening
is
where
we're
just
where
it
is
really
slowing
our
progress
down
to
have
this
thing
dragged
on
and
on.
It
was
like
you
know
two
meetings
ago,
it's
like
what
we're
gonna
have
implications
by
the
next
meeting
and
then
the
last
meeting
or
really
gonna
have
implementations.
G
But
you
know
it
was
even
even
supposed
to
be
this
meaning.
This
must
be
like
a
couple
months
in
or
thing,
and
you
know
well
now
we're
like
dragging
on
too
like
well.
Maybe
next
meeting
will
really
have
to
have
implications
well,
I
mean
next
meeting.
I
have
to
find
you
no
money
to
like
even
send
people
for
the
next
meeting
right
I
mean.
D
D
We
want
to
hold
those
milestones
honestly,
really
because
there
is
a
time
pressure
in
this
working
group,
because
otherwise,
if
we
come
out
to
a
solution,
very
student,
nobody
will
need
it
anymore,
and
so,
if
like
one
of
the
solution,
can
make
this
time
skating
in
the
other?
That's
a
decision.
I
don't
want
to
put
this
as
a
race,
but
because
it's
not
a
race,
but
there
is
some
time
pressure
and
if
we
have
a
solution,
that's
working
that's
implemented.
There's
no
need
for
me
to
wait
for
anything
else.
Lars.
F
Eggert,
so
keep
it
waiting
for
one
of
the
horses
to
die,
isn't
productive
right
there.
They
seem
pretty
healthy
to
me.
So
we
need
to
kill
one
and
I.
Don't
know
how
you
kill
it.
So
my
personal
favorite
is
a
sushi
eat
off
between
the
2i,
editing
teams
or
something,
but
so
having
two
is
not
helpful
here.
Right
and
I
really
don't
care
anymore,
which
one
it
is,
but
I
just
want
to
have
one
so
that
we
can.
Actually,
you
know
make
that
one
really
good.
D
I'm
not
killing
any
horses
and
and
like,
and
we
did
this
so
if
say,
I
mean
it's
hard
for
us
to
say
how
much
people
are
actually
working,
how
much
time
name
they
spend
on
this
and
how
much
money
they
spend
or
whatever.
But
if
the
case
would
be,
we
would
be
a
small
group
and
like
we
would
not
have
the
energy
to
work
on
two
proposals.
Then
that's
very
true,
but
there
are
so
many
people.
We
got
so
many
feedback
that
people
are
interested
to
implement.
D
F
Laws
are
gonna
promise.
This
is
the
last
of
them
comes
in
back
on
this,
but
so
clearly
having
two
proposals:
splits
the
attention
of
the
working
group,
but
because
you
need
to
documents,
you'd
understand
them
both
you,
they
have
the
boat
in
your
head.
Now,
we've
done
this
for
a
while
now,
but
in
that
and
I,
don't
think
going
forward
in
this
mode
longer.
Is
it's
gonna
be
very
good?
It's
so.
E
E
Look
Jen,
look
at
the
gentleman
behind
you,
I
mean
I've.
Take
a
look.
What
I've
I'm
new
to
this
I've
looked
active
on
the
mailing
list?
I've
watched
the
adoption
call
there's
clear
support
support
for
both
these.
It's
not
clear
to
me
that
these
two
protocols
are
actually
plug
compatible.
That
they'll
hear
different
requirements
from
different
people
and
different
opinions
about
the
importance
of
different
different
requirements.
I'll.
F
Let
the
two
people
behind
me
answer
that
Martin
stepping
down,
so
he
really
doesn't
care
anymore
and.
A
F
D
Because
if
you're
talking
about
energy
and
repeating
there's
a
lot
of
energy,
but
what
I
would
like
is
people
who
would
be
interested
in
implementing
one
of
solution
than
any
point
of
time
that
might
be
no
or
they
might
be
like
at
the
end
of
the
schedule
or
that
might
be
they
might
have
the
same,
pose
the
not
saying
if
you
pick
one
I
would
implement
it.
Otherwise,
I,
don't
please
come
and
talk
to
us
or
send
us
an
email.
So
we
actually
get
an
overview.
You
know
how
much
energy
they
actually
is.
C
There's
a
lot
of
energy
in
this
room,
it's
more
heat
than
light,
so
but
sorry,
I'm
aaron,
falk,
Akamai
I
am
I,
am
very
suspicious
of
arguments
to
take
action
in
the
ITF
that
are
based
on
speed,
because
the
ITF
does
not
have
a
great
track
record
for
speed
it.
There
were
four
most
of
my
experience
in
the
idea
for
the
last
20
years.
It
takes
between
five
and
ten
years
to
get
a
draft
published
as
a
proposed
standard.
C
B
Well,
let's
not
go
down.
Okay
now,
I
have
a
concern
about
deployment
of
ether
solution
and
a
for
example,
suppose
that
we
stick
either
TCP
crypt
or
the
TRS
injection
on
the
ftp
I
were
not
in
my
home.
It
goes
to
the
ftp
energy
in
the
nut
and
the
ftp
algeria
says
what
heck
that
doesn't
look
like
a
FTP
and
scratches
it.
So
that's
just
one
example
of
a
potential
deployment
issue.
I'd
like
to
see,
since
our
goal
is
to
deploy
opportunistic
encryption.
B
I'd
like
to
see
a
couple
of
target
deployment
scenario
and
say:
hey:
let's
deploy
the
solution
on
I,
don't
know
when
a
FTP
HTTP
or
some
specific
applications
or
physical
against
the
application
for
which
we
want
to
have
opportunistic
encryption
and
then
and
let's
pick
that
as
their
the
deployment
cares
and
says,
we'd
like
to
take
that
as
a
benchmark.
Do
that
and
show
me
that
you
can
in
fact,
but
it
in
the
server
but
in
the
clients
and
get
it
done.
E
G
So
I
think
that
the
like
one
obvious
one
would
be
NFS.
A
lot
of
people
are
still
using.
You
know
like
NFS,
three
and
stuff.
That's
that's
totally
unencrypted.
G
Yeah
ftp
is
problematic
because
you
know
precisely
because
of
these
because
of
the
the
porch
that
are
being
transmitted
and
and
interactions
with.
Not.
But
but
there
is
I'll
talk
about
this
a
little
bit
of
an
API
talk
like
we
have
a
proposal
for
how
how
this
can
get
rolled
out
and
part
of
it
is
that
you
know
there
should
be
like
dhcp
hooks
and
your
DCP
client
that
probe
the
network
a
little
bit
before
deciding
what
to
enable.
So
we
can
go
through
an
intermediary
phase
where
it's
something
so
we're
like.
G
E
So
this
part
part
of
my
response
to
your
question
as
opposed
to
trying
to
say
in
advance:
what
are
we,
what
is
going
to
be
used
for?
Let's
take
a
look
of
people
out
there.
Interest
is
uniform
because
of
inversions
of
TCP
crypt
code
available
for
download
for
quite
some
time,
so
information
on
what
people
downloaded
them
use
them
for
might
provide
some
insight.
Your
question
about
about,
where
is
where
are
the
likely
early
targets
for
adoption.
B
B
B
E
E
Things
Christian
points
out
is
actually
generalized.
What
he
basically
said
is
that
ftp
through
an
app
requires
an
alg
given
the
requirements
for
encryption
and
authentication
that
we're
pursuing
here.
The
no
protocol
coming
out
of
his
working
group
is
likely
to
work
well
through
any
alg
period.
So
ftp,
sorry.