►
From YouTube: CNCF TOC Meeting 2021-03-30
Description
CNCF TOC Meeting 2021-03-30
A
B
B
We
just
got
started
literally
just
opened
the
spreadsheet
and
mentioned
it
ingrained.
D
I
I
I've
seen
a
few
people
looking
at
red
bpf,
which
is
their
library
for
other
purposes,
which
I
thought
was
a
good
sign
that
they
were,
you
know,
had
designed
something
that's
useful
outside
their
specific
use
case.
So
it
was
quite
good.
People
seemed
like
people
are
recommending
that
as
the
best
library
to
use
for
rusty
bpf
at
the
moment,
I've
heard
seems
to
be
the
impression
I've
got
so
I
seem
quite
positive.
D
B
That
impression
yeah,
I
feel
like
rust
and
ebpf,
is
still
a
fairly
new
sport,
but
you
know
it's
pretty
exciting.
You
know,
I
think
it's
going
to
grow.
E
I
also
like
the
fact
that
they
listed
the
clear
reason
of
how
they're
different
from
falco
that
is
always
very
useful
during
the
reviews.
D
B
F
G
H
I
I
agree,
I
think
it's
for
sandbox.
The
bar
is
certainly
a
lot
lower
as
far
as
variety
and
diversity
of
contributors
from
different
places.
D
D
B
Oh,
I
guess
for
some
of
you,
this
might
be
the
first
time
you've
done
sandbox.
So
the
way
we
normally
do
this
is
we
put
votes
in
chat,
so
I
will
type
in
something
like
votes
for
ingrained
and
you
can
put
your
plus
ones
or
minus
ones
below
and
amy
keeps
track.
E
B
This
strikes
me
again
as
another
one
of
those
projects
that
I
yeah.
You
know
I
was
kind
of
started
thinking
about
as
like
ecosystem
projects,
where
it's
a
really
cool
thing.
It's
great.
It's
not
at
the
same
kind
of
scale
and
importance
as
something
like
kubernetes
or
prometheus,
or
you
know
blah
blah
blah,
but
that's
no
reason
at
all
not
to
put
it
into
sandbox.
D
B
B
C
B
B
Do
you
think,
there's
any
any
relationship
with
the
kubernetes
project
that
we
should
be
you
know,
does
it
make
sense
for
this
to
be
a
separate
project.
H
B
C
C
I
J
Yeah,
I
know
another
little
bit
background
of
this
project.
It's
actually,
I
think,
highly
related
to
the
machine,
learning
and
big
data
field,
which
is
kind
of
we
don't
have
similar
projects
actually
in
sensei.
For
now,
if
we
have
kobe
flow,
it
will
be
actually,
I
think,
can
drop
into
the
same
category.
I
Yeah,
I
mean
honestly
off
the
cuff
when
we
reviewed
it
in
the
storage
sig.
We
were
like
this
feels
more
like
an
apache
project
because
of
that
machine
learning
aspect,
but
I
do
know
that
that
is
growing
in
intensity
across
kubernetes
users.
So
I
think
it
could
go.
B
I
B
Chris
has
just
mentioned.
I
was
gonna
gonna
ask
this.
There's
another
linux
foundation,
another
linux
foundation,
foundation.
F
F
I
I
don't
know
if
it's
the
best
home,
because
that
was
our
first
impression
as
well
alex,
and
I
said
it
seems
like
it
would
fit
better
in
the
apache
foundation.
But.
B
Just
looking
at
the
comments
about
how
it's
aligned
with
cloud-native
computing.
D
D
But
do
we
not
want
this
kind
of
project?
I
mean
you
know
it's
like
large-scale
data
is
a
cloud-native
problem.
It's
just
I
mean
I
get,
we've
been
like
kind
of
neglecting
it,
but
I
don't
see
that
that's
something
that
we
necessarily
want
to
continue
to
put
into
other
places.
Necessarily
I
mean,
I
think
that
I
mean.
I
As
if
no
I
and
that's
kind
of
why
we
we're
on
the
fence,
like
you
know,
there
hasn't
been
a
strong
focus
on
actually
managing
and
using
data
at
scale.
Yet
right,
we've
been
more
focused
on
the
infrastructure
components
compared
to
you
know
how
data
flows
in
and
out
and
managed
and
analytics
everything
else.
So
I
think
that
would
be
and
to
to
liz's
point
at
the
beginning
of
the
year,
focusing
on
what
our
strategy
should
be.
B
And
that
could
be
a
really
good
thing.
I
mean
if
we
had
a
like
momentum
towards
more
of
these
kind
of
projects,
then
maybe
that
would
well.
Is
there
always
going
to
be
a
giant
gaping
hole
in
the
fact
that
we
don't
have
kubeflow.
G
So
looking
at
the
getting
started
and
installation,
I
don't
see
anything
related
to
like
big
data
such
right.
It's
just
they
provide
an
api
and
they
give
an
engine,
and
so
it's
it's
an
object.
Manager
and
the
name
also
says
lib
vineyard.
So
it's
not
like
a
standalone
framework
for
ml
or
anything
like
that.
J
It's
actually,
I
think,
more,
like
a
plug-in,
for
example,
to
kubernetes,
so
in
that
case,
kubernetes
can
be
used
to
manage
the
large-scale
data
which
can
be
used
for
machine
learning
scenario.
That
is
my
understanding
on
this
project.
G
And
the
last
column,
they
they
goofed
up
a
little
bit.
They
they
have
references
to
fluid,
which
is
another
one
which
is
which
is.
B
B
H
But
but
it
does
seem
to
do
like
in-memory
data
set
bookkeeping
and
integrating
with
the
kubernetes
scheduler
for
right.
Some
sort
of
topology
aware
data
data.
B
B
Personally
speaking,
I
I'm
kind
of
open
to
the
idea
that
we
go
in
this
direction,
but
I'm
not
100
sure
whether
if
we
do
this
today
is
that
the
best
thing
for
this
project
you
know
will
having
this
project
in
the
sandbox,
be
the
best
way
to
make
it
successful,
and
maybe
it
is
maybe
I
don't.
I
don't
know.
I
just
don't
know.
Yeah.
C
G
G
D
Oh,
no,
that's
it's!
I
I
it!
Apparently
it
seems
to
be
more
than
that
because
they're
talking
about
using
s3fs
as
a
python
file
system
interface
for
s3,
so
it
presumably
provides
s3
bindings
and
pi
bind
for
seamless
operations
c,
plus
and
python,
which
is
so
it's
kind
of
it's.
It
looks
more
integrated
than
I
thought,
which
was
weird.
I
haven't
looked
at
it
in
detail.
I
was
just
looking
at
it
just
now.
I
was
thinking
that's
a
little
bit
weird.
B
D
G
B
This
has
no
stars
super
nude
that
created
march
I
mean
literally,
the
guy
hasn't
even
started
his
own
project.
K
B
B
B
C
D
I
I
I
took
a,
I
told
you
about
it
a
few
weeks
ago
when
it
came
up
on
the
application,
because
it
was
kind
of
interesting.
D
You
might
have
seen
that
it's
basically
a
kind
of
different
set
of
trade-offs
to
g
viser
for
doing
via
isolation
by
reflecting
cisco's
back
into
the
host.
D
So
you
basically
don't
run
linux
pretty
much
at
all
in
the
guest
you
just
use.
I
o
urine
and
shared
memory
to
move
cisco
filtered
syscalls
into
the
host.
I
didn't
look
at
it
in
a
lot
of
detail,
but
it
seems
definitely
kind
of
out
for
higher
performance
than
g,
vise
or
qmu,
potentially
at
least
according
to
their
initial
measurements
and
faster
startup
times.
D
I
think,
as
a
sandbox.
It's
I
mean
it's
not
got
a
lot
of
traction
and
it's
very
new,
but
I
think
it
probably
kind
of
passes
the
sandbox
bar.
D
E
Again
in
in
march,
it's
it's
very
new
as
well.
B
I
B
D
B
D
But
again
there
it
is
weirdly
all
in
one
commit
like
this.
Clearly,
this
commit
called
ad
source
code.
D
With
542
117
000
editions
is,
it
is
a
is
just
a
little
bit
weird
like
there's
no
previous
game
history,
but
it
is
a
one
gigantic
commit
of
a
lot
of
stuff.
B
That's
excellent:
would
anyone
particularly
feel
strongly
that
we
should
go
to
a
vote.
B
D
B
Trickster
tricksterproxy.I
o
this
is
about
as
far
as
I
got
in
the
in
the
spreadsheet
beforehand,
so.
B
B
G
Had
was
they
list
four
primary
functions
of
their
stuff,
and
I
wasn't
too,
I
couldn't
figure
out
like
how
those
four
are
related
to
each
other
or
whether
they
put
all
the
four
in
the
same
bucket.
F
D
They
do
say
that
cortex
and
thanos
might
have
cache
layers
modeled
after
this,
which
doesn't
make
the
question
of
like.
Could
this
become
a
generic
caching
there
for
those
projects,
so
they
don't
have
to
implement
their
own
and
would
that
make
sense
yeah?
That
would
be
interesting,
because
that
would
that
would
definitely
be
an
interesting
outcome
if
this
was
to
become
a
just,
a
generic
caching
layer
for
other
projects.
C
F
B
This
seems
to
consolidate
that
opinion
700
odd
stars.
B
H
I
didn't
attend,
but
it
was
presented
at
siegfried
and
early
march.
I
didn't
attend
specifically.
B
I
don't
particularly
remember
it
being
one
of
the
ones
that
I
know:
ricardo
aravina
mentioned
a
few.
You
know
at
the
tfc
meeting,
but.
B
B
E
Came
later
they
came
after
ricardo
presented
at
the
toc
okay,
which
is
you
know,
it's
fine.
G
G
So
one
observation
on
this
one
was
in
the
product
risk
description.
It
says
that
they
are
trying
to
define
two
extensions
and
they
would
like
to
use
cncf
to
be
able
to
drive
them
to
specification
as
default
implementations
or
turn
it
into
specification,
which
is
fine.
It's
just
that
that
that's
what
they
are
here
for
they
are
here
to
so
they
can
use
their
cnc
of
presence
to
be
addition
of
some.
D
B
Okay,
so
maybe
that's
that's
two
questions
for
them.
One
is
the
name.
The
second
would
be
exactly
what
they're
contributing
I
mean
I
I
feel
quite
positive
towards
the
general
idea
of
the
project,
but
I'm.
B
F
F
Yeah,
I
mean
there's
some
work
going
on
there,
but
I
don't
know
where
that
that
is
going
to
end
up.
You
know
there's,
like
I
view
ssvm
as
a
runtime
right,
so
whether
what
specs
they
implement
and
so
on,
I
think,
is
a
whole
other
problem
and
that
will
prob
the
specs
will
probably
be
defined
in
another
body
anyway.
From
my
perspective,.
C
F
B
D
F
So
yeah
there's
a
couple,
so
I
don't
know
you
know
we
have
to
be
sensitive
time
because
people
are
dropping
off,
so
I
could
reach
out
to
them
and
ask
for
a
rename
and
talk
to
our
lawyers.
The
question
is:
if
you
want
to
do
you
want
to
vote
or
wait
until
that
topic
is
done,
because
we
could
vote
and
say,
like
hey,
we're
comfortable
at
as
long
as
they
change
your
name
or
come
back.
You
know,
two
months
later,
whenever
we
meet
yeah.
B
B
C
B
A
could
hold
a
vote,
but
perhaps
it'd
be
better
to.
F
B
All
right,
I
think
at
this
point
I
should
probably
drop
off
because
I'm
not
sure
we'll
get
through
the
next
one.
So
how
many
did
we
get
through
amy.
A
Eight,
so
we
we
can.
A
Another
one
probably
going
to
look
for
the
27th
of
april.
B
Spreadsheet,
all
right
apologies
to
the
to
the
folks
we
didn't
get
to,
but
I
guess
we
held
an
extra
meeting
this
month.