►
From YouTube: IETF110-NVO3-20210308-1430
Description
NVO3 meeting session at IETF110
2021/03/08 1430
https://datatracker.ietf.org/meeting/110/proceedings/
A
Okay,
hello,
everyone,
I'm
just
gonna,
wait
a
couple
more
minutes,
I'm
still
waiting
for
our
secretary
to
join
and
we'll
give
folks
plenty
of
time
to
to
join
the.
A
A
Okay,
hello,
everyone
welcome
to
mvo3
for
this
actually
fairly
short
meeting.
We've
got
a
very
short
agenda
for
ietf
110.
My
name
is
matthew
bocce,
I'm
here
with
my
co-chair
sam,
and
we
also
have
our
secretary.
Is
you
I
think
it's
going
to
take
the
minutes?
Thank
you.
A
Okay,
so
this
is
the
the
notes.
Well,
I'm
sure
you
are
all
very
familiar
with
this
by
now.
This
basically
says
any
anything
you
say
within
this.
The
frame
of
this
ietf
is
considered
to
be
a
contribution
to
the
itf
I'll.
Just
leave
that
up
on
on
the
on
the
screen
for
a
few
seconds.
A
Apply:
okay,
blue
sheets
are
done
automatically,
which
is
great
and
azure
is
taking
our
minutes.
Thank
you.
A
So,
as
I
mentioned,
we
have
a
very
short
agenda
today,
just
a
quick
working
group
status
update
and
then
an
update
on
the
draft
for
bfd
for
geneve.
A
Okay,
right
milestones
so,
as
usual
need
an
update.
Many
of
these
are
completely
out
of
date,
and
but
there
are
possibly
one
missing,
which
is
we
don't
have
a
milestone
for
our
young
model,
so
we
probably
ought
to
help
to
add
one
for
the
young
models
in
terms
of
completion
of
the
working
groups,
charter
or
completion
of
as
much
as
we
can
do
in
the
working
groups
charter
and
on
our
milestones.
A
A
A
Any
comments
on
that,
so
I
think
the
approach
should
probably
be
that
sam
and
I
will
will
update
the
milestones
and
we
put
a
proposal
to
the
list
of
what
we
believe
needs
to
be
completed
before
we
can
say
that
the
working
group
is
is
done.
B
C
Correct,
yes,
thank
you
good
question
and
thank
you
for
bringing
it
up.
I
think
that
the
oem
draft
is
don't
was
adopted
last
year,
but
it's
a
stable
and
I
would
can
ask
a
working
group
to
consider
adoption
poll
or
something
like
early
review.
C
Definitely
adoption
poll
will
produce
the
comments
and
then
we
can
address
the
comments
and
I
will
we
are
committed
to
work
on
them
right
away.
So
I
think
it's
reasonable
to
expect
that
we'll
have
it
complete
by
next
meeting.
So
if,
for
example,
adoption
poll
will
not
adoption
fall
so
last
call
I
apologize
last
call
are
issued,
then
our
comments
will
come
and
we'll
address
them
by
next
meeting.
A
So
so
we
have
mechanisms
as
well
to
to
get
wider
ietf
review,
such
as
a
routing
area
review
for
the
driver.
C
A
Document
status,
so
we
have
one
new
rfc
since
the
last
last
time
we
met
as
a
working
group
in
the
iatf,
that's
rfc
8296,
which
is
geneve.
So
that's
our
first
standards
track
rfc
for
the
generic
network,
virtualization
encapsulation.
A
So
congratulations
to
the
authors
and
thank
you
for
your
hard
work
on
addressing
the
fairly
lengthy
and
detailed
set
of
of
comments
for
that.
We
also
have
a
encapsulation
draft
draft
itfmvo3ncap.
That
kind
of
I
think
that
was
the
one
that
describes
the
output
of
the
the
encapsulation
design
team.
We
did
try
and
try
and
last
call
that
last
year
didn't
get
a
lot
of
responses
on
the
list.
A
We
did
at
some
point
promised
that
that
would
be
that
we
would
try
to
publish
that
document
and
we
felt
that
as
a
working
group,
it
was
valuable
at
the
time
to
to
publish
it
as
really
as
a
way
of
documenting
documenting
the
experience
of
one
working
group
in
picking
an
encapsulation
when
there
were
multiple
candidates
on
the
table.
A
But
obviously
we
can't
move
forward
without
sufficient
review
from
the
working
group
and
consensus
that
this
is
actually
something
that
the
community
is
interested
in.
So
I
think
probably
the
way
forward.
This
is
to
run
another
working
group
last
call,
but
we
would
obviously
like
to
see
much
better
buy-in
from
the
working
group
on
on
progressing
with
this.
The
I
think
the
editors
have
also
not
been
so
so
active
on
this
as
well.
A
So,
yes,
we
definitely
need
a
an
active
and
active
editor
or
two
on
this
draft.
If
it's
going
to
move
forward,
we
also
have
some
informational
documents
on
alternative
encapsulations.
A
I
think
that
the
one
that
had
interest
in
progressing
last
time
we
discussed
this
was
for
the
excellent
gpa,
and
so
we
probably
need
to
work
in
group
last
call
that
as
well
but
again,
I
think
we're
going
to
be
we're
going
to
have
to
look
for
fairly
significant
interest
from
the
working
group
and
some
comments
and
proper
level
of
review
from
a
working
group
in
order
to
progress.
This
document.
A
A
We
that
went
through
working
group
last
call
it's
now
with
martin.
It
is
held
in
his
queue
waiting
for
a
companion
document
in
the
best
working
group
that
shows
the
bgp
extensions
required
for
geneve
that
so
really
we're
kind
of
waiting
on
on
the
progress
of
that.
A
The
virtual
machine
mobility
documents,
so
that
was
sent
back
to
the
working
group
after
after
review
by
martin.
There
are
some
comments
from
martin
about
it
being
not
really
specific
to
nvo3.
I
don't
know
if
martin,
if
you
want
to
say
anything
more
about
that
at
this
time,.
C
D
Okay,
yes,
on
that
specific
document,
I
think
I
I
had
communicated
back
to
the
working
group,
my
expectations.
D
I
know
that
donald
had
tried
to
work
a
bit
help
the
authors
to
rework
a
bit
the
document,
but
still,
I
think
there
are
a
number
of
questions
that
the
the
working
group
needs
to
needs
to
bring
an
answer
to
and
what
would
be
great,
if
would
be
that
you,
the
chance,
helped
drive
that
discussion
to
see
where
it
leads.
A
Okay,
thank
you
yang
models,
so
we
have
a
configuration
model
for
nvo3.
I
believe,
there's
also
a
young
model
for
geneva
as
well,
which
isn't
listed
on
here,
sam.
I
think
you
were
looking
at
shepherding
these
these
drafts.
Do
you
want
to
say
anything.
B
Yes,
that's
correct,
so
I
have
received
email
from
the
authors
yesterday,
so
I
will.
I
have
two
action
items
for
that.
The
first
one
is.
B
I
will
have
to
get
review
comments
from
young
doctors
and
I'll
also
try
to
prepare
for
last
issuing
the
last
call.
A
Okay,
great
thank
you
and
we
have
two
oem
related
working
group
drafts.
We
have
the
bfd
genevieve
draft,
that's
on
the
agenda
and
we
also
have
the
geneva
m
draft,
which
we
just
discussed.
A
Okay,
so
first
up
is
vfd
for
geneve
right.
I
need
to
change
my
sharing
to
do.
A
E
Ahead,
hello,
it's.
D
E
Speaking
yeah,
this
presentation
is
bfd
virginia.
This
draft
was
adopted
by
his
working
group
in
november
last
year.
I
present
the
01
version
of
the
working
group
job
next
slide.
Please.
E
The
new
tunnel
encapsulation
has
been
published
as
rc
8926
and
the
bfd
4x10
has
been
published
as
rc
8971,
and
the
second
main
update
is
to
resolve
the
open
issues
with
pfd
over
internet
over
genevieve
encapsulation,
including
number
one.
If
the
vip
of
the
original
mve
has
no
gs,
then
what
are
you
just
to
be
used
for
ipv4
or
ipv6
number
two,
if
the
vip
of
the
terminating
me
has
no
ip
address,
then
what
idea
just
to
be
used
for
ipv4
or
ipv6?
E
E
E
Otherwise
there
are
two
main
differences
between
bfd,
4gb
and
obviously
89
71.
E
number
one
pfd
for
gnu
specifies
non-management
to
bni
solution.
Nevertheless,
rc
8971
specifies
management
vi
solution
for
completeness.
E
We
plan
to
also
the
management
vi
solution
in
this
job
by
referencing
to
geneve
oem
working
group
dropped.
E
That
chapter
provides
specifics
for
the
requirements
on
both
non-management
bni
solution
and
the
management
vi
solution,
as
well
as
how
to
achieve
management
of
ai
solution.
In
the
case
of
genevieve
number
two
bfd
for
genevieve
use,
ipo
genevieve
as
well
as
ethernet
over
genevieve.
E
Nevertheless,
rc
rc
8971
use
only
isn't
it
over
vxlan
due
to
vaccine
characteristics.
E
E
E
Is
this
is
the
opening
issue
two
and
its
resolution?
E
This
open
issue
is
that
if
the
vip
of
the
terminating
me
has
no
ip
address,
then
what
I'd
be
addressed
to
use
for
ipv4
or
ipv6
in
zero
zero
version
job
two.
In
this
scenario,
the
iph
must
be
chosen
from
the
column,
current
ffff
column.
E
Nevertheless,
indeed
in
the
third
one
version
dropped
in
this
scenario,
the
ip
address
should
be
set
to
column
column,
one
slash
one,
two,
eight
for
ipv6,
that's
because
column
column,
one
slash
one.
Two
eight
for
ipv6
is
a
loopback
address.
E
E
E
E
So
with
the
open
issues
and
the
resolutions.
E
E
Number
three:
last
but
not
least,
the
authors
are
in
favor
of
vfd
much
more
than
other
oem
mechanism.
E
Firstly,
the
out
of
band
the
exchange
of
bfd
discriminator
is
an
optional
add-on,
which
is
not
necessary
because,
as
stated
in
this
job
as
the
receiving
side,
the
bfd
control
package
can
also
be
demultiplexed.
With
your
discriminator
equals
to
zero.
E
If
out-of-band
exchange
of
bfd
discriminator
is
used,
then
there
are
multiple
methods
that
can
achieve
it.
For
example,
echo
request,
reply
can
be
used,
bgp
evpn
can
be
used
and
open
flow
can
be
used
too.
E
A
Okay,
I
think
that's
with
the
the
oem
geneve
draft
it'll
be
worth
getting
a
routine
area
review
of
this,
I'm
wondering
if
we
shouldn't
also
get
a
ask
for
a
review
from
the
bfd
working
group.
I
appreciate
it.
It's
the
same
group
of
people,
mostly
so.
B
E
So,
for
which
encapsulation
do
you
mean
the
bfto?
Isn't
that
over
genevieve.
B
E
E
So
for
ipv6.
In
this
version,
in
this
zero
by
version,
we
use
the
column
current
one
slash
one.
Two,
eight.
B
So
the
destination
host
or
nve,
whichever
is
processing
these
bfd
frames-
needs
to
open
a
udp
socket
for
127
address.
Is
that
right.
B
Is
it
done
similarly
for
other
vxlan
as
well,
or
is
it
purely
for
geneve.
E
C
It's
very
interesting
question
so
when
we
discussed
the
deal
with
vxlan,
so
it
was
pointed
out
that
ipv4
loopback
addresses
mapped
in
ipv6.
They
don't
have
a
special
meaning.
So
the
only
loopback
ip
address
that
exists
in
ipv6
is
1
28.,
so
functionally
for
ipv6.
C
If
we
use
a
loopback
address
range
for
ipv4
addresses,
then
in
ipv6
we
can
use
only
single
address,
one
slash
128
and
that's
what
recommended
in
geneve
oem
document
that
ip
encapsulated
control,
packets
use,
addresses
from
ipv4
loopback
range
or
one
slash
128
for
ipv6.
C
So
in
that
regard,
bfd
document
and
genevam
documents
are
consistent.
B
C
This
this
goes
to
nbe
so
which
terminates
geneve.
C
B
C
Oh
okay,
right
again,
that's
that's
a
little
bit
because
basically,
the
premise
of
this
is
that
it
terminates
on
that
law
because
it
wants
to
test
as
much
as
possible
of
their
infrastructure.
C
Yeah
again
that
that's
that's
the
case.
If
that
doesn't
have
ip,
then
I
would
imagine
that
ethernet
encapsulation
will
be
in
place.
E
Another
is
for
the
non-management
of
ai,
so
this
job
to
only
describes
the
non-management
of
vmi
solution.
So
for
now
management
bi
we
provide
two
encapsulations
first
is
ip
over
eft
over
ip
over
geneve.
Another
is
isn't
that
a
bfd
over
ethernet
of
a
geneve.
E
So
for
the
second
one,
there's
a
scenario
that
the
web
vip
of
the
originating
mae
or
the
web
of
the
terminating
me
has
no
ip
address.
E
So
for
for
this
specific
scenario,
we
provide
resolution,
so
there
are
openings,
one
and
obviously
to
adjust
for
this
specific
scenario
I
just
mentioned.
A
Okay,
so
with
that
that
was
our
first
and
last
session
presentation.
Thank
you
for
for
joining.
Please
look
out
for
requests
on
the
list
for
a
review
of
the
various
drafts
that
we
need
to
complete
and
last
call
at
some
point
in
the
near
future
and
with
that
any
other,
would
anybody
like
to
raise
any
other
issues
with
the
working
group.
A
Okay,
well,
I
think
that
closes
the
meeting.
Thank
you
and
hopefully
see
you
see
you
at
the
next
ietf
wherever
that
might
be
online
or
or
somewhere
else.
Thank.