►
From YouTube: CNCF Telecom User Group Meeting - 2020-02-03
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
In
order
to
be
able
to
show
that
the
CNF
are
following
best
practices,
and
so
I've
been
working
with
the
bulk
team
and
a
few
others
on
this
and
I'm
hoping
to
be
able
to
circulate
a
deck
in
the
next
few
days,
which,
needless
to
say,
should
generate
a
ton
of
interesting
feedback
and
and
given
the
intensity
in
this
community.
Probably
a
lot
of
criticism,
but
I
wanted
to
give
folks
a
heads
up
that
that
is
coming
and
it's
something
that
we're
actively
pursuing.
I.
Think
it's
going
to
particularly
flow
in
to
this
next
conversation.
A
That
Robbie
is
going
to
kick
off
around
re
to
where
I'm
very
hopeful,
but
that
this
test
suite,
if
we're
able
to
develop
an
initial
version
of
it
relatively
quickly,
could
be
a
very
useful
input
into
that
re
to
process
and
could
be
a
way
of
trying
to
represent
some
of
those
best
practices.
But
there's
a
huge
amount
of
open
questions
and
then
I
guess.
A
My
final
comment
would
just
be
that
if
the
the
test
suite
is
successful
and
we're
able
to
get
some
interest
in
traction,
such
the
goal
would
be
to
have
it
accepted
into
CFCF
as
a
sandbox
project
and
that
it
would
have
maintained,
errs
and
an
update
schedule
just
like
any
other
open
source
project.
But
I
would
point
out
that,
although
I'm
executive
director
I
don't
actually
get
to
control
what
projects
become
CFCF
hosted
projects,
we
have
this
independent
technical,
Oversight,
Committee
kind
of
like
our
Supreme
Court
and
they're,
the
ones
that
decide
and
the
answer.
A
If
they
look
slide,
deck
is
always
going
to
be
know,
they'll
evaluate
actual
code
and
then,
as
with
everything
in
the
telecom
world,
I
think
feedback
from
the
operators
is
going
to
be
a
huge,
huge
ly.
Impactful,
so
I
think
I'll
stop
there.
I
could
take
maybe
five
minutes
of
questions,
but
it's
just
it's
probably
just
worth
waiting
to
see
the
slides,
I'll
be
sharing
them,
be
a
slack
in
the
mailing
list
and
so
would
love
to
get
your
feedback
on
them.
But
it
doesn't
make
sense
for
me
to
read.
B
C
Don't
if
you
don't
mind
so
I
know
this
is
some
activities
happening
in
another
fan
around
ovp
phase,
two
I
just
wondering
how
that
relates
to
what
I
described
here
around
the
CNF
conformance
test.
Do
you
imagine
that
conformance
test
would
be
focusing
more
about
how
cloud
native
an
application
as
well
as
Aleph
Innova,
prefers
to
look
at
interfaces?
The
API
is
and
infrastructure
point
of
view.
How
do
you
see
the
split
or
is
there
any
relation
between
these
two
projects?.
A
Robby
I
think
we
should.
You
should
look
over
the
slides
and
we
should
do
a
call
about
it.
At
the
end
of
the
week,
I
honestly,
III
I
mean
the
one
thing
that
I'm
certain
of
is
that
I
have
no
interest
in
CN,
CF
kicking
off
a
competitive
effort
to
see
in
TT
and
re,
so
it
absolutely
focus
on
cooperating.
I
I
remain
a
little
bit
unclear
on
witches
upstream
how
decisions
get
made
the
kind
of
fault
lines
between
specs
and
and
implementation
and
those
kinds
of
details,
I
think.
A
Okay,
so
I
think
I'll
stop
there.
Hopefully
that's
enough
of
a
teaser
that
will
get
you
to
at
least
open
up
the
email
when
I
send
it
out
and
I
would
maybe
ask
Taylor
or
the
senior
to
just
mention
the
upcoming
events,
and
then
we
could
hand
it
off
to
Robbie
I
think
as
the
to
talk
about
Bari
in
the
meeting
in
Prague
and
then
out
next
steps.
There.
B
D
Yeah
we
have
some
upcoming
events,
the
schedule
for
cube
con
cloud
native
con
in
europe,
which
will
be
held
in
amsterdam,
the
last
few
days
of
March
and
first
studious
of
April.
The
schedule
has
been
announced
and
on
Tuesday
the
31st
in
the
afternoon
there
will
be
a
CNF
test
bed
session,
going
over
creating
a
cloud
native
user
plane
for
an
evolve
packet
core
and
on
Thursday.
There
will
be
a
birds
of
a
feather
for
this
telecom
user
group
on
in
April.
There
will
be
the
open
networking,
an
edge
summit,
North
America.
D
If
P
is
closed
today
and
the
schedule
will
be
announced
on
March
5th,
tubecon
China
it
the
see
if
these
are
open
for
July
and
those
clothes
on
the
21st
of
February,
then
in
September
will
have
open
networking
an
edge
summit
Europe
the
CFPs
are
open
for
that
as
well
and
closed
June,
7th
and
then
later
on.
In
the
year,
we've
got
Cuba
in
Boston
in
November,
and
the
CFPs
are
scheduled
to
open
up
in
April.
For
that
event,
are
there
any
other
events
of
note
that
you'd
like
to
add
to
the
agenda.
B
That
cfp
list
there
at
the
top,
the
idea
is
to
make
sure
specific
talks
that
are
relevant
to
the
group.
Aren't
noted
since
there's
so
many
things
going
on.
So
if
there's
something
that
someone
is
presenting,
whether
you're
doing
it
or
want
you
think
someone
else,
some
someone
else's
talk
is
very
relevant
to
the
conversation.
Then
please
add
it
to
that
list.
C
C
Let
me
show
one
slide:
will
explain,
was
the
thinking
in
see
entity
and
and
why
we're
really
interested
in
doing
the
RI
discussion
and
just
to
clarify
that
the
slide
you're
going
to
show
now
is
all
about
OpenStack
vision,
and
this
is
not
the
communities
and
cover
and
container
base
infrastructure.
What
we
really
want
to
is
having
the
exact
same
picture
we
have
front
of
us,
but
to
apply
to
containers
and
component,
is
a
red
line.
C
C
C
Perfect,
so
this
is,
as
I
said,
this
is
a
few
locals
in
OpenStack,
so
we
do
have
a
difference
model
and
see
entity
and
the
reference
model
just
explaining
how
the
infrastructure
should
expose
the
resources
to
upper
layer,
and
then
we
do
have
the
two
different
architecture.
The
one
the
first
one
is
called
the
difference
architecture,
one
which
is
based
on
OpenStack,
and
this
is
learning.
How
do
you,
design
and
architect
your
infrastructure
based
on
OpenStack
technologies,
and
once
we
do
have
the
OpenStack?
C
We
do
have
the
reference
implementation
and
this
is
really
their
ri1
in
this
scenario-
and
this
is
really
looking
at
how
do
I
build
an
infrastructure
based
on
based
in
the
reference
architecture,
using
the
open
source
technologies
and
open
source
components
available
out
there
and
we
working
with
OPA
Navi.
This
is
why
you
see
open
AV
in
here,
which
is
reusing.
The
community
installers
available
and
open
AV
to
consume
the
artifacts.
C
We
generate
in
the
reference
implementation
and
create
an
installation
on
implementation
of
the
infrastructure
using
the
computer
lab
available
in
open
IV,
and
this
is
really
what
we
refer
to
as
the
reference
implementation,
which
is
really
the
realization
of
the
reference
architecture
and
building
the
infrastructure
from
open
source
component,
using
open
source,
installers
and
open
source
projects
and
using
also
community
labs
to
install
the
infrastructure
and
its
abilities.
Here.
A
part
of
the
reference
implementation
is
a
document.
It's
a
specification
documents
that
will
give
you
the
target
state
which
is
dealing
specifying
in
details.
C
What
is
are
the
artifacts
that
would
like
to
deploy
and
what
are
the
metadata
and
the
values
of
the
different
configuration
the
infrastructure
need
to
have.
Also,
we
do
have
some
requirements
when
it
comes
to
what
requirements
from
it
working
with
install
requirements
when
we
need
to
support
in
order
to
be
able
to
install
the
infrastructure,
we
do
have
also
lab
requirements
specifying
the
minimum
and
maximum
of
some
of
the
configuration
from
platform
perspective
CPU,
for
example,
RAM
and
and
sockets
and
NIC
cards,
and
so
on.
C
These
are
the
requirements
when
it
comes
to
a
lab.
So
finally,
we
have
the
cookbook
which
is
saying
in
details.
What
instruction
do
you
need
to
perform
in
order
to
allow
you,
as
a
consumer,
for
the
infrastructure
to
install
the
infrastructure,
whether
in
a
community
lab
or
in
your
own
lab
or
in
a
lab
a
service
environment?
C
Now,
once
the
infrastructure
is
installed,
then
we
do
also
work
with
ok
Navi
to
leverage
their
existing
tooling
and
ecosystem
to
do
testing
and
conformance
to
the
infrastructure
against
the
C
entity
specifications.
So
we
do
have
the
our
C
and
the
C
stands
for
conformance,
so
the
reference
conformance
one
in
C
entity
looks
at
what
other
this
cases.
We
need
to
allow
me
to
test
the
infrastructure,
Peston
requirement
and
specifications
set
by
C
entity.
C
So
if
C
entity
specify
a
particular
protocol,
for
example,
I
just
wanna
make
sure
that
this
infrastructure
can
actually
support
the
a
protocol.
If
the
C
entity
says
that
I
need
to
have
a
huge
page
enabled,
for
example,
then
there
should
be
at
least
case
here
to
test
if
a
huge
page
is
enabled
in
the
infrastructure-
and
this
is
really
as
you
can
see-
there
are
two
part
of
this-
is
the
under
bi
testing
and
there
is
the
the
vnf
testing.
What
comes
down
the
VI
is
all
about.
C
Is
the
infrastructure
table
of
exposing
the
resources?
I
am
promising
to
my
application,
based
on
the
reference
model.
Specification
existence,
the
entity.
It
comes
to
the
vnf
testing
once
the
infrastructure
is
validated
its
conformant
to
see
ntp
specification,
you
get
actually
a
badge
for
that
from
11,
and
this
is
why
you
have
the
11
open,
AV
verified
badge
for
your
infrastructure.
Once
the
infrastructure
is
ready
to
build
it
to
be
used,
we
put
a
V
enough.
C
C
Is
expecting-
and
this
is
where
we
write
these
cases
in
C
entity-
to
specify
what
are
those
testing
we
need
to
do
for
vienna's
to
make
sure
they
are
compliant
with
the
entity,
and
the
idea
really
here
is
still
under
development
once
that
vnf
is
running
on
top
of
the
infrastructure,
conformant
identity
interfaces
and
api's,
you
get
an
a
badge
as
well.
That
will
make
sure
that
will
certify
that
you
vnf
is
actually
can
run
against
the
entity
compliant
infrastructure.
Also,
we
think
about
some
performance
testing
so,
and
this
is
a
little
bit
tricky.
C
So
this
is
still
in
the
roadmap,
and
this
is
really
looking
at.
How
do
we
characterize
the
workload
in
two
different
categories
to
allow
us
to
test
the
performance
for
that
group
of
VN
apps
and
figure
out
how
they
measure
and
score
against
other
the
reference
implementation
for
the
golden
vnf?
We
also
would
like
to
to
have
any
entity
now,
I
guess.
C
The
question
is
one
trying
to
understand
when
it
comes
to
communities
and
containers
workload,
we
do
have
a
reference
architecture,
difference
architecture,
and
this
is
really
specifying
the
cloud
platform
in
container
based
and
using
Cabana
these
technologies,
and
this
is
really
where
we
specify
what
components
would
like
to
see.
We
specified
the
networking
details.
We
specified
the
lifecycle
measurement
of
infrastructure,
as
well
as
the
workload
and
the
question
becomes.
C
If
I'm
a
CNF
developer
and
I
would
like
to
say,
okay
I
wanna
build
my
application
using
I,
remembered
my
CNF
against
the
infrastructure
that
see
entity
is
approving
off.
So
what
is
the
infrastructure?
Where
is
they
exist?
How
can
I
actually
have
an
instantiation
of
the
infrastructure,
using
open
source
components
and
open
source
elements
and
tooling,
to
allow
me
to
actually
install
in
my
lab
and
run
CNF
on
top
of
that,
and
maybe
do
some
testing
around
this
now?
This
is
very
whether
the
request
and
ask
from
in
my
initial
email
about.
C
We
need
to
start
thinking
about.
What
is
the
difference.
Orientation
is
what
does
it
live
and
what
to
an
existing
open
source
components
and
tooling,
you
would
like
to
leverage
it
take
advantage
of
to
build
the
reference
implementation
once
we
finish
that,
in
my
view,
and
once
we
have
the
reference
orientation
being
able
to
be
instantiated,
we
need
to
think
about.
Okay
homogeneous
fell
well,
and
then
we
need
to
think
about
opening
the
rc2
is
based
and
targeting.
C
For
example,
if
we
specify
in
C
entity
that
the
interface
will
infrastructure
has
to
be
based
on
memory
interface,
just
an
example,
then
if
the
CNF
is
not
using
him
interface,
then
it's
not
considered
conformant
to
see
entity
and
hence
we'd
like
to
make
sure
that
anyone
in
CNF,
build
and
design
is
using
the
same
interfaces
and
open
api
towards
the
infrastructure,
which
c
NT
and
c
entity
is
approving
of,
and
hence
they
need
to
have
that
kind
of
ecosystem.
Around
testing
and
certification
is
really
important
to
make
sure
the
c
entity
story
is
complete.
C
Ok,
so
one
of
the
things
we
also
start
to
kick
this
off.
One
of
the
things
we
start
to
look
at
is
we
start
having
an
issue
in
github
and
I'm
gonna
shorten
the
screen,
and
this
is
Daddy
saying
we
need
to
really
figure
out.
How
do
we
do
this
difference
in
augmentation
to
person
difference
architectural
specification?
We
do
have
an
ether
bad
as
you
can
see
in
in
the
in
the
github
issue
and
I
guess.
C
The
question
I'm
would
like
to
ask
the
community
here:
what
are
the
existing
thinking
around
project
exists
within
CNC,
F
or
the
open
source
in
general?
That
will
allow
us
see
entity
to
be
able
to
use
those
open
source
projects
to
instantiate
that
reference
implementation
to
have
something
conformal
to
see
entity
or
RA
to
the
communities
based
one.
C
F
Think
anything
our
dad's,
so
Tom,
Karen
I'm,
sorry,
is
very
much
an
iterative
approach,
so
I
think
it's
everyone.
Everyone
agrees
that
our
a2
is
not
in
a
position
where
we
can
test
an
implementation
against
sick
in
a
complete
and
robust
way,
and
but
we
want
to
get
started
on
on
understanding
what
the
constraints
are
of
the
implementation
or
the
requirements
we
have
or
any
feedback
that
there
might
be
into
RA
based
on
actually
implementing
something.
So
that's
why
we're
starting
our
I
before
our
a
is
it's
kind
of
complete
or
near
complete.
G
Yeah,
so
this
business
hotel
from
juniper,
so
the
bigger
question
I,
would
ask
the
the
CN
CF
Tugg
community
is
that
is
everybody
familiar
with
RA?
What
is
what
is
happening
in
order
to
have
they
had
an
opportunity
to
look
at
it?
So
perhaps
that
could
be
the
starting
point
so
share
that.
Take
a
look
at
it,
whether
well,
how
does
it
map
to
the
CN
CF
Todd
work
which
we're
doing
and
what
are
the
commonalities?
G
What
are
the
differences
and-
and
we
take
that
so
that's
an
RA-
is
driven
from
the
resource
model
RMS
so
which
is
which
is
fairly
defined.
So
we
take
that
and
then
we
want
to
take
it
to
the
RI
which
implementation,
which
is
what
Robbie
is
talking
about.
So
so
I
think
that
this
is
the
best
way
to
connect
all
the
dots
and
get
to
a
joint
collaboration
effort
from
on
both
communities.
E
Get
things
so
that
they're
a
little
bit
easier
to
to
work
with
and
negatives
in
the
cloud
data
space,
the
the
thing
that
I
did
I'd
like
to
try
to
flag
for,
for
you
all
is
to
when
you're,
when
you're
writing
up
the
the
specs
and
the
the
best
way
I
can
describe
kubernetes
versus
previous
stacks
and
technologies.
That
we've
seen
is
that
kubernetes
itself
is
an
emergent
system,
an
emergent
Pat
form.
E
Where
do
you
put
your
your
the
thing
that
you
want
and
then
the
system
attempts
to
converge
to
it
as
opposed
to
having
the
top
central
controller,
push
down
details
to
all
of
the
things
underneath
of
it
and
trying
to
centralize
all
that
logic
into
into
one
place,
and
so
so
bill.
So
be
cognizant
of
that
as
you
as
you
write
the
specs
out,
because
it
has
some
its.
It
sounds
like
a
like
a
detail,
but
it
has
some
interesting
properties
that
that
greatly
affect
how
you
build
things.
E
It's
just
put
out
the
just
put
the
thing
out
saying
that
this
system
must
be
capable
of
having
multiple
interfaces
injected
into
a
kernel,
and
then,
if
you
want
to
you,
want
to
list
and
those
particular
techniques,
then
you
can.
You
can
create
a
sub
spec
or
something
similar
that
links
back
to
that
original
definition.
So,
in
order
to
went
this
thing,
these
are
the
these
are
the
recommended
paths
or
so
on
for
for
making
that
happen.
E
But
what
it
means
is
that,
when
a
new,
if
a
new
platform
comes
around
or
three
days
itself,
changes
that
we
don't
have
to
the
the
specs
are
not
baked
in
with
implementation
assumptions
and
that
we
can.
It
gives
us
a
migration
path
from
from
one
two
to
the
other
and
is
something
that
allows
people
to
come
in
and
and
enter
into
the
space.
E
And
it
does
add
a
little
bit
of
work
onto
the
CNF
testing
side,
because
we
have
to
make
sure
that
those
CMS
will
continue
the
work,
whether
you
pull
them
in
using
once
I
think
versus
another.
But
what
that
will
forces
to
do
is
come
up
with
really
good
interfaces
in
order
to
make
sure
that
that
cooperation
works
and
will
end
up
in
a
much
better
place
in
the
long
run.
C
We
witness
in
the
implementation
back
to
the
architecture
and
vice
versa,
and
although
in
see
entity,
we
always
say
that
we
we
need
to
be,
as
we
staggered
as
possible
to
the
actual
implementation.
In
some
scenarios,
when
it
makes
sense
to
be
very
specific,
I
think
we
should
be
not
shy
to
specify
things
as
we
feel
there
is
a
need
to
do
that.
Now.
There's
this
compromise
line
a
sweet
spot
we
need
to
find.
C
Where
do
we
need
to
specify
more
and
where
we
need
to
be
completely
restricted
now,
to
really
better
understand
that
and
the
impact
of
this
we'd
like
to
have
the
Afghan
segmentation
I
started
to
to
appear,
so
we
can
actually
a
stall
experiment
with
it
and
feedback
order.
Learning
back
to
the
RA.
Now
there
has
been
some
comments
and
in
github
I
saw
around.
We
should
mature
RA
to
first
before
thinking
about
doing
the
RI.
C
Now,
although
I
do
understand
the
rationale
behind
this
one
thing,
we
believe
that
our
I
will
allow
us
to
think
about.
This
problem
now
will
allow
us
25
those
project
that
we
need
to
put
more
resources
into
and
more
focus
into
now,
similar
to
talk
about
openness,
technologies.
We
have
open
a
V,
we
have
a
ship,
we
know
we
have
put
some
resources
in
there
and
some
focus,
and
there
I
would
like
to
do
similar
things
in
here.
C
So
we
could
think
about
how
the
ecosystem
would
look
like
in
to
end
now.
I
know
there
are
some
comments
here
around
things
like
okey,
D,
cube,
cube,
one
car,
basically
metal
three
so
under
is
dissipate,
for
example,
right,
so
I
do
think
at
this.
My
opinion
that
when
things
are
our
testbed,
this
is
where
the
difference
conformance
is
more
relevant
to
this.
When
we
talk
about
the
foreign
segmentation,
maybe
one
of
those
tooling
can
be
used
alongside
some
of
the
labs
available
to
install
the
infrastructure.
C
E
So
I
think
we
can
break
the
infrastructure
down
into
into
multiple
parts,
because
one
of
them
is
how
do
you
instantiate
the
actual
container
infrastructure
itself
like
how
do
you
essentially
kubernetes
and
versus?
How
do
you
install
a
variety
of
components
that
your
that
your
CNF
will
need
then
versus?
How
do
you
install
the
CNF
itself
and
so,
rather
than
driving
into
implementation
details
right
now?
My
recommendation
would
be
to
is
to
point
us
at
where
we
can
help
discuss
these
type
of
things
and
I.
E
Think
through
the
approach
you're
taking
is
right,
like
let's
put
something
down
on
paper:
let's
go
build
something
in
order
to
to
see
it
work
and
then
let's
come
back
and
discuss
what
we've
built
and
then
changes
as
we
go
on
now.
One
thing
that
was
really
interesting-
and
this
is
what
the
cig
networking
group
did
occur,
that
is
that
working
group
is
when
they
were
developing
their
specs.
E
The
this
stuff
may
be
too
large
to
do
that
for
the
whole
solution,
but
for
the
smaller
components
we
certainly
can
can
can
do
that
for
some,
some
of
them
as
well
and
and
also
but
even
just
getting
like
the
the
multiple
ideas
and
specs
down
even
and
you
would
just
think
and
do
those,
because
we
now
have
a
better
understanding.
We've
been
around
long
enough
that
we
can
get
a
better
understanding.
E
Have
some
things
will
work
and
thought
experiment
to
get
some
but
yeah
I,
think
you're,
I,
think
you're
heading
in
the
right
direction,
with
that
I
would
definitely
recommend
like
getting
a
place
where
we
can
discuss
like
what.
What
are
what
are
the
solutions
that
we
currently
see
in
heavy
use
within
the
within
the
enterprise
space?
And
my
my
recommendation,
you
don't
have
to
follow
this,
but
my
strong
recommendation
would
be
to
try
adopting
those
tools.
E
First,
like
if
you're
working
in
kind
of
Ironman
about
things
like
terraform,
if
you
want
to
install
infrastructure,
you
use
things
like
bike,
helm
or
operators
or
mixture
of
the
two
of
them,
and
in
that
way
the
we
can.
We
can
see
whether
those
meet
your
needs
and
if
they
do,
then,
then
it's
easy.
We
can
just
slap
them
in
and
you
get
the
benefit
of
the
community
working
on
them,
as
opposed
to
something
that
has
only
worked
on.
C
Absolutely-
and
this
is
one
reason
why
I
actually
brought
we
came
to
the
CN
CF-
this
bit
is
trying
to
get
the
opinion
from
the
community
here
now
you
mentioned
you
wasn't
to
have
a
meeting
to
discuss
that
I
was
thinking.
Maybe
we
can
use
the
CN
CF
talk
meeting,
which
is
this
one
to
kinda
sign
to
go
into
the
details
and
see
which
one
which
tooling
is
more
appropriate
to
install
quality.
C
So
you
mentioned
here,
in
my
view,
are
key
points
so
how
to
install
Copernicus-
and
this
is
an
question
I-
would
like
to
maybe
start
finding
answers
for
now,
installers
right,
I
guess
this
is
what
it
goes
to
so
my
question
community.
Do
we
consider
okay,
do
use
an
installer
cube
one
gravelly
metal
three?
Is
that
the
choice
we
need
to
pick
one
of
those
and
you
mentioned
helm
chart.
C
E
Before
we
even
jump,
this
far
like
I,
would
literally
just
like
put
a
Google
Doc
and
be
in
this
channel
like
being
for
anyone
right
now
and
have
people
put
suggestions
down
for
for
all
of
them,
because
we
can
also
shop
it
around
and
say:
hey,
which
would
you
use
to
to
install
and
try
to
get
a
more
complete
list,
and
so
so
before.
Even
jumping
into
the
details
like
I
would
recommend
just
like.
E
Let's
give
a
survey
of
what's
out
there
first
and
if
Dan
and
and
others
are
fait
with
spending
some
time
on
here,
I'll
be
happy
to
I'd,
be
happy
to
do
that
as
well.
But
I
think
one
that
once
a
month,
cadence
is
probably
not
fast
enough
for
for
what
you
for
what
you're
looking
for.
But
but
we
can
definitely
do
a
lot
of
this
on
a
sink
and
at
least
start
getting.
E
Those
initial
questions
done,
because
right
now,
when
with
the
concern,
is
that
we
have
a
lot
of
which
we
have
some
tools
that
that
are
being
described.
But
but
it
would
be
good
to
get
a
survey
of
the
larger
set
of
tools
that
are
that
are
out
there,
so
we
don't
earn
nothing
up
abilities
of
of
all
of
them
like
that,
I,
don't
I,
don't
think
we
understand
the
problem.
Well
enough
to
say
this
is
the
tool
video
did.
You
have
to
use
so.
A
I
might
push
back
a
little
bit
on
that
Frederick,
which
is
even
if
we
had
did
that
survey
and
had
perfect
information.
I,
don't
know
that
we're
ever
gonna
want
to
say.
Oh,
you
must
use
cube
spray
to
install
your
your
kubernetes.
In
order
for
it
to
be.
You
know,
CMT
tra,
to
compliant
I
mean
that
we
have
this
whole
certified
kubernetes
program,
where
the
basic
premise
of
it
is
that
you
should
be
able
to
modify
the
software
and
modify
the
installer
and
come
up
with
your
own
installer
and
and
installed.
E
H
Is
purely
about
the
reference,
implementation
and
I
think
one
of
the
characteristics
that's
fairly
important
to
use,
for
instance,
the
size
of
the
community
activity,
and
so
on,
so
that
we
pick
a
tool
that
is
fairly
common,
I
guess,
and
that
would
just
be
the
reference
implementation
that
see
entity
uses
to
validate
synapse
against
on
the
commercial
side.
You
can
basically
pick
any
to
all
you
want,
and
that
needs
to
be
differentiated.
A
Sorry
so
I
mean
you're
looking
for
that,
I
certainly
would
recommend
the
CNF
testbed,
which
uses
cube
spray
and
is
general
generally
a
relatively
vanilla
installation
of
it.
In
you
know,
cube
spray
is
part
of
the
kubernetes
project
and
makes
use
of
cube
ADM,
which
is
the
official
recommended
set
of
tools
for
installing,
but
I
would
just
be
hesitant
on
saying:
oh,
that's
the
correct
way
of
installing
it
a
particularly
known
way.
I
guess
I
would
have
a
semi-related
question
for
Robby
in
particular,
which
is
your
document
talked
about.
A
What
test
centers
could
be
used
to
validate
it
and
I
guess.
I
did
check
in
on
that,
whether
you
have
reviewed
the
way
that
certified
kubernetes
works,
where
we
don't
require
or
allow
the
use
of
third-party
test
centers.
One
of
the
may
be
semi
novel
aspects
of
our
approach
is
that
it's
an
open
source
test
suite
the
vendor,
runs
it
and
and
uploads
the
logs
showing
their
performance.
A
But
then
one
of
the
key
ideas
is
that
there,
rather
than
just
needing
to
take
the
work
us
just
the
community,
just
need
to
take
the
vendors
word
for
it.
Any
user
of
that
platform
is
able
to
come
back
later
and
rerun
that
test
suite
and
show
that
they're
still
getting
valid
results,
and
so,
in
that
sense
we're
crowdsourcing,
the
validation
or,
or
you
know,
lack
of
cheating
by
participants
rather
than
having
essentially
a
paid
contractual
relationship,
and
so
I
guess
I
just
wanted
to
check
in.
Are
you
aware
of
that
model?
A
And
is
there
a
reason?
I
mean
I
understand
that
that's
not
the
model
that
ovp
uses,
but
do
you
feel
like
it's
so
for
re?
One
I
definitely
understand
that's
the
direction
you're
gonna
be
going,
but
for
RA
to
that
there
might
be
an
openness
and
C
NT
T
to
trying
a
different
approach
toward
conformance
to
certification.
Sure.
C
So
for
for
RA
one
either
even
for
other
one,
we
don't
say
you
have
to
go
to
specific
lab
to
get
certified.
We
do
support,
as
I
mentioned
the
self
certification,
meaning
be
all
open-source
components,
open,
source
tools.
You
run
the
test.
You
create
the
log
file
of
the
result
of
your
testing.
You
submit
that
result
into
the
portal,
which
is
adoption
open
AV.
C
We
have
the
CVC
community,
look
at
the
logs
and
and
the
we
require
two
pluses
from
the
community
and
once
we
have
that,
and
we
agree
that
the
results
are
genuine,
then
we
give
the
badge
for,
for
that
particular
vendor.
Now
we
do
also
have.
The
third
party
lab
scheme
is
not
more
than
some
companies
that
have
some
experience
into
helping
vendors
running
those
certification,
but
they
don't
certify
on
behalf
of
or
see
entity.
Anybody
the
certification
happening
is
the
same
way
as
if
you
run
it
in
your
own
lab
and
I
guess.
C
C
The
expectation
is
to
have
something
similar
more
open
to
your
point,
where
anybody
can
install
the
tools
required
specified
by
entity,
and
once
you
store
the
tooling
that
this
case
is
random.
The
requirements
required
these
cases
to
run.
You
submit
your
results
in
some
way
or
another
and
I
know
we
haven't
talked
about
ovp
FS
yet,
but
the
intention
is
to
have
something
similar
to
that
some.
Some
group
of
people
reviewing
those
results
and
deciding
based
on
that.
C
If
the
platform
is
conforming
to
the
entity
or
not,
I'm
not
know
exactly
how
the
kubernetes
do
conformance.
This
for
governor
is
a
bad
form
today,
but
it
is
something
we
certainly
can
learn
from
if
it
if
their
model
is
better
than
the
one
we're
thinking
about.
We
certainly
are
flexible
I
doubt
that
one.
I
Hi,
this
is
Mark
Shaw
stack,
I've
got
a
question
for
the
tug,
which
is
kind
of
an
extension
of
that
from
Robby
and
also
something
Frederic
said
earlier,
which
is
what
is
the
strategy
that
you
guys
are
employing
or
plan
to
employ
in
areas
where
there
are
multiple
choices
for
a
technology?
C
and
I
comes
to
mind.
You
know
first
right,
how
do
you
plan
to
certify
us
or
qualify
a
CNF.
A
A
But
if
each
CNF
that
you
purchase
the
the
vendor
says
you
know
vendor
at
me,
says:
oh
here's,
our
amazing
firewall,
you
know
and
it
only
works.
If
you
install
Acme
CNI
on
this
cluster
and
by
the
way,
if
you
install
Acme,
C
and
I,
then
you
can
install
always
see
an
eye
or
air
sensor
or
anybody
else's.
Then
all
we've
done
is
recreated
a
walled
single
vendor,
walled
garden.
A
So
I'm
very
cognizant
of
in
fact,
I
have
a
slide
here
that
I'll
be
sharing
the
few
days
called
the
CNI
trap,
and
one
possible
way
out
of
it
is
to
try
and
define
the
characteristics
of
a
CNI
that
have
given
CNF
needs,
but
I
don't
mean
to
imply
that
it's
a
simple
problem
or
has
a
necessarily
a
simple
solution.
You
didn't.
I
Think
I
was
going
to
give
you
an
easy
question:
did
you
but
yeah?
So
that
and
again
that's
just
an
example
question
really
what
I'm?
Looking
for
and
I'll
look
in
that
material
that
you
said
will
be
forthcoming
is
what
the
approach
is
not
just
for
CNI,
but
in
general
for
areas
where
there
are
multiple
options.
So
you
just
can't
you
know
wire,
can't
wire
your
testbed
one
way
or
demand.
Everyone
uses
one
paradigm,
so
I'll
just
leave
it
at
that
and
an
exercise
for
for
follow-up.
C
And
this
is
one
example,
then
we
said
within
see.
Entity
would
like
to
be
as
we
staggered
as
possible.
So
to
your
point
is
what
are
the
feature
said?
The
CNF
would
like
to
acquire
from
a
CNI
plugin
if
we
canvass
tag
that
we
would
to.
But
if
scenarios
where
we
find
ourself,
we
have
to
go
into
the
details
and
start
specifying
more.
We
will
go
as
far
as
it
takes
to
be
to
make
sure
the
CN
have
scam
run
together
in
the
same
infrastructure
without
having
to
dictate
and
requirements
underneath.
C
Okay,
so
I
think
I,
don't
know
I
mean
what's
the
agenda
for
today,
I
don't
want
to
see
the
whole
time
of
the
meeting,
but
to
be
good
to
understand.
What
do
you
guys
feel
the
next
step,
so
I
heard
about
survey
I
heard
about
a
cube
spray
being
the
the
tool
used
by
CNF
this
bed
and
I
heard
about
having
a
dedicated
meeting
around
that.
So
what
is
the
feeling
our
CNC
after
goob,
when
this
call
happy
to
spend
the
time
of
this
meeting,
to
discuss
the?
C
All
right
so
may
I
ask
so
we
did
have
a
question
here
right,
which
is
his
helm.
Is
the
the
format
we're
going
to
use
a
to
which
told
we
can
use
I
know
some
people
mention
the
Q
spray,
the
lab,
what
the
lab
would
be
I
know
we
have.
There
was
a
demo
in
Red
Hat,
the
one
shown
in
coop
con
in
in
San
Diego,
and
that
potentially
could
be
the
starting
point.
If
we
is
time
to
think
about
how
they
did
it
and
what
tools
they
use,
maybe
does
where
we
can
start
from.
C
F
F
But
for
me,
the
more
important
thing
that
we
need
to
get
at
the
moment
is
is
those
two
things
so
one
is
is
where
we
can
deploy
a
kubernetes
platform,
so
the
lab
location
pack
it
on
there,
whatever.
So
what
what
people
know
is
available
for
us
to
start
using
imminently,
because
I
think
that's
important.
The
second
thing
is
the
installer,
so
with
whether
it's
okay
do
coop
spray.
Whatever
you
know,
what
do
people
have
the
skills
in
using
because
I
think
as
a
community
we
don't
want
to.
F
We
don't
want
to
go
and
choose
something
that
no
one
knows
I
use,
so
I
think
that's
what
I
thought
Frederick
was
getting
at
going
to
understand
what
what
the
community
knows.
How
to
use
already,
because
the
bit
about
helm
and
installing
the
CNF
for
me
is
is
not
really
what
we're
after
at
the
moment,
that's
a
kind
of
that's
the
secondary
part
about
managing
the
workloads
within
kubernetes,
rather
than
managing
the
platform
itself.
Yep.
E
E
I
need
a
kubernetes
cluster.
How
do
I
install
it
versus
second
stage
would
be
like
how
do
I
install
the
components
my
infrastructure
needs
and
another
stage
may
be.
How
do
you
install
the
CNS
and
so
on
so
break
them
up
into
logical
stages
and
that'll
like
now
giving
you
a
starting
point
we're
already
having
kubernetes
cluster
okay?
What
do
I
need
to
add
in,
and
the
second
thing
is,
there's
a
whole
effort.
That's
already
done
a
lot
of
work
towards
this.
E
It's
also
ran
by
Taylor,
carpenter
and
and
those
colleagues
where
there's
it's
:
well,
you're
familiar
with
the
CNF
testbed
and
the
CNF
testbed
is
designed
to
be
a
repeatable
lab
that
you
can
spin
up,
and
it's
been
things
up
like
kubernetes
as
stanwyk
said
that
uses
cube
flow
and
uses
which
uses
cube
adamant
under
underneath,
which
is
the
closest
thing
to
a
career
that
is
installer.
It's
that
it's
they
recommend
the
path
and
just
start
to
start
with
the
tooling
from
from
there
and
that'll
actually
prevent
you
from
repeating
a
bunch
of
work.
H
Maybe
one
more
thing
to
consider
is
also-
and
maybe
they,
the
CNF
testbed
guys-
can
elaborate
on
that
a
little
bit
more
cui
acd
integration.
So
it's
from
my
perspective,
maybe
that's
step
2.
Certainly
that's
step
2
after
step
one,
but
at
the
end
of
the
day
we
probably
want
to
regularly
run
the
tests
we'd
like
the
reference
implementation
to
validate
that
the
tests
are
suitable
and
that
the
platform
still
passes
and
all
those
things.
A
Is
another
big
strength
of
the
the
CNF
test
bed?
Is
that
because
it's
open
source
and
continually
updated,
we
would
encourage
any
vendor
that
wanted
to
incorporate
it
into
their
own
internal
development?
To
do
so,
and
so
it's
also
very
feasible
if,
for
example,
if
there's
open
source
CNF
that
are
referenced,
CNF,
that
we
can
set
that
up
with
an
ongoing
CI
process,.
H
Does
that
maybe
that's
a
stupid
new
question,
but
in
the
CNC
of
context,
does
it
run
already
in
a
CI
CD
fashion?
Wait
after
MIT
I'm
wearing
an
opening
v
hat
right
now,
because
that's
where
I've
come
from
but
like
open
V
has
see
ICD
tooling
to
deploy
stuff
and
then
run
tests
and
so
on.
Just
similar
setup
exist
in
the
open
on
this.
Since
you
have
side
as
well
for
the
testbed
and
for
others.
H
A
H
A
H
A
Testing
system-
that's
used
by
certified
kubernetes
to
run
the
test,
suite
the
conformance
test
suite
on
top
of
kubernetes,
and
essentially
we're
not
doing
that
today
in
the
CNF
test
bed.
Just
because
it
would
be
duplicating
a
ton
of
work,
that's
happening
upstream,
but
it
is
something
that
we'd
be
very
happy
to
to
build
in.
It
seems
like
that
would
be
valuable.
C
B
Right
now
we
do
have
a
as
some
CI
efforts
around
deploying
kuberan
Nettie's
and
a
specific
use
case
or
set
of
use
cases
that
we're
running
and,
like
dan
says
it's
really.
The
CI
CD
pipeline
should
be
focused
on
whatever
you
want
to
validate,
and
so
there
was
some
specific
things
that
we
wanted
to
keep
verifying
I.
B
Think
going
back
to
what
Frederick
said,
though,
knowing
what
you're
wanting
capabilities
lies
and
what's
driving
it
would
be
a
good
place
to
start.
So,
if
you
have
the
stages-
and
you
say
we
want
to
be
able
to
deploy
to
whatever
the
lab
location
and
we
want
anyone
to
be
able
to
deploy
that
and
and
then
going
with
what
Mark
was
saying
as
far
as
people
will
want
different
options,
you
may
have
seen
ice
and
other
things.
So
you.
C
B
If
you
have
those
laid
out
the
capabilities
and
requirements,
then
the
tooling
that
the
tools
can
fit
into
that
I
think
it'll
make
a
little
bit
more
sense
and
that
ties
back
to
that
I
think
original
concurrences.
This
is
more
of
a
prusik
concept
thing.
Here's
here's
an
example
of
a
platform:
they
can.
Anyone
can
refer
back
to
the
reference
architecture,
which
should
be
this
capabilities,
and
then
another
platform
can
be
implemented
by
anyone,
and
ideally
test
and
CNS
that
are
deployed
to
that
platform.
B
Yeah
I
would
say
requirements
and
and
what
you
would
like
for
people
to
be
able
to
do.
It
looks
like
in
the
chat
at
least
Thomas
and
there's
been
some
responses.
I've
been
asking
some
things,
but
if,
if
you
want
to
be
able
to
deploy
to
it
wherever
that
location
is
so
if
it
was
something
like
packet
or
there's
a
public
lab,
that's
fine
and
you
want
to
be
able
to
deploy
to
your
own
labs.
B
Well,
those
are
requirements
that
I
think
will
drive
some
of
the
tooling
and
then
you
can
look
at
stuff
like
do.
You
need
lifecycle
management
of
the
cluster
right
now,
what
whatever
things
you're
wanting
that
may
or
may
not
be
necessary
for
something
that
for
this
reference
and
implementation.
But
if
you
have
that,
then
that
will
help
drive
conversations
about
the
tools.
B
B
Really
important
one
that
I
want
to
bring
up
is
the
underlying
hardware
infrastructure
and
configuration
if
I
would
at
least
have
that
noted.
It's
often
skipped
that
someone,
someone
set
up
the
underlying
hardware,
any
type
of
network
configuration
VLANs
everything
else,
and
then
the
software
is
installed
on
top
of
that
and
I
think
that's
important
to
at
least
note.
B
If,
if
that's
expected,
okay
and
probably
at
least
document
the
expected
configuration
to
run
the
tool,
run
the
software
to
bring
up
the
cluster
and
then
otherwise,
is
there
a
desire
to
be
able
to
automatically
do
that
for
a
lab
and
testing?
Do
you
want
to
be
able
to
configure
the
network
dynamically.
C
Yeah,
that's
a
good
point.
Actually,
okay,
all
right
I
think
come
on.
Tell
me:
I
didn't
have
much
time.
I
think
we
have
a
good
plan.
Now
we
can
spend
some
time
the
requirements
and
maybe
in
the
next
cnc
after
group,
we
can
have
a
quick
look
at
that
and
then
once
we
finalize
the
comm
understand,
what
does
it
was
it
needed?
We
can
think
about
the
tooling
and
the
physical
location
and.