►
Description
Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon Europe 2021 Virtual from May 4–7, 2021. Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Public Technical Oversight Committee (TOC) Meeting - Moderated by Chris Aniszczyk, CTO, The Linux Foundation
Moderators: Chris Aniszczyk
Speakers: Liz Rice, Alena Prokharchyk, Justin Cormack, Saad Ali, Brendan Burns, Matt Klein, Michelle Noorali, Dave Zolotusky, Sheng Liang, Katie Gamanji, Xiang Li
The Technical Oversight Committee (TOC) provides technical leadership to the cloud-native community. The CNCF will host a public TOC meeting, inviting the community to discuss various agenda items along with holding an open Q&A for the community with TOC members.
A
Good
all
right:
okay,
we've
got
this
hollywood
square
thing.
You
know
we're
at
the
end
of
the
conference,
which
is
fun.
Thank
you.
Everyone,
for
you
know
attending
this
coc
panel.
Basically
to
you
know,
end
the
conference.
You
know
make
sure
that
the
toc
is
approachable
and
everyone
kind
of
asks,
questions,
and
you
know
things
you
can
learn
about,
what's
being
planned
for
for
the
next
year,
so
kind
of
kick
things
off
I'll.
Introduce
you
know
myself.
My
name
is
chris
anizek.
A
I
have
the
fun
job
of
serving
the
technical
community
in
cncf.
You
know
we
have
currently
10
toc
members.
You
know
that
will
go
throughout
introduce
themselves.
I
will
call
each
of
you
out
individually
to
kind
of,
say,
hi
and
talk
a
little
bit
about
your
you
know,
job
and
enroll.
So
to
start
things
off
I'll,
just
go
in
order
from
what
I
could
see
so
we'll
start
with
elena.
C
D
Hey,
I'm
dave
engineering
on
spotify's
platform
team
and
I've
done
a
bunch
of
work
around
our
adoption
of
kubernetes
and
grpc
and
envoy
work
around
the
perimeter,
and
things
like
that.
I
think
the
newest
toc
member
on
this
board,
so
I'm
still
kind
of
fig
figuring
out
what
I
do
here
and
how
to
do
it
but
happy
to
be
here.
E
F
Awesome,
I'm
michelle
nerely.
I
am
a
software
engineer
here
at
microsoft,
azure
I
work
on
open
source
tools
in
the
cloud
native
ecosystem,
focusing
on
making
containers
and
kubernetes
easier
to
use.
I
am
currently
a
core
maintainer
on
the
service
mesh
interface
project
and
the
open
service
mesh
projects,
and
previously
I've
worked
on
several
other
tools
in
the
cloud
native
space.
I'm
also
an
emeritus
maintainer
on
the
helm
project
happy
to
be
here.
A
G
H
G
A
I
A
B
J
K
A
A
A
You
know
one
thing
that
you
know.
I
constantly
highlight
that
a
lot
of
people
new
to
cncf,
certainly
don't
understand,
is
you
know,
as
a
community
we
have
come
up
with
a
set
of.
You
know,
guiding
principles
that
the
toc
has
produced,
that
we
publish
on
github
and
we
kind
of
you
know
evaluate
it's
been
a
while,
since
we've
done
an
update,
we're
probably
overdue
to
kind
of
take
a
second
look
at
it.
But
if
you
kind
of
go
through
some
of
these
principles,
you
know
we're
always
about
putting
projects.
A
A
You
know
we
kind
of
have
a
role
around
no
king
makers,
if
you
saw
from
you,
know,
liz's
wonderful
keynote,
just
earlier
kind
of
highlighted
that
point
around
no
king
makers-
and
you
know
one
size
not
fit
all
you
know,
even
though
we
host
things
like
specifications,
you
know
things
like
cni
smi
we're
not
a
traditional
standards
bodies.
You
know
we're
very
much.
A
You
know,
code
first,
practical
use
and
the
kind
of
spec
follows
and
really
you
know
at
the
end
of
the
day,
we
truly
want
to.
You
know
help
you
know
projects,
you
know
you
know,
above
and
beyond
what
they
could
typically
get.
You
know
from
from
a
foundation,
so
you
know
during
this
whole
kubecon
with
you
know,
23,
000
or
so
people
are
attending
got
pinged
by
you
know.
You
know
at
least
probably
a
couple
dozen
people
like.
A
We
want
to
think
of
bringing
our
projects
to
the
you
know,
cncf.
How
does
it
work?
You
know.
Good
news
is
this
year
we've
documented
a
lot
of
improvements
when
it
comes
to
processes.
So
if
you
go
to
our
documentation,
especially
our
project
proposals,
doc
you'll
find
a
lot
of
that
information
there.
It's
generally
fairly
straightforward.
We've
made
a
lot
of
improvements
around
the
sandbox
process.
A
So
it's
a
lot
easier
now
to
you
know,
propose
projects
and
for
the
toc
to
kind
of
deal
with
the
inbound
of
requests
on
the
bright
side,
you
can
kind
of
see
all
the
existing
proposals.
So
if
you're
kind
of
new
to
this,
please
kind
of
you
know,
look
at
the
existing
proposals
out
there
and
kind
of
get
ideas
and
maybe
attend
one
of
the
public
toc
meetings
that
happen
every
couple
weeks
or
just
jump
on
slack
we're
super
friendly
and
happy
to
to
answer
any
questions
that
you
have.
A
You
know,
as
we're
kind
of
you
know,
wrapping
up
this
year.
You
know
some
of
us
on
the
staff
side,
we've
been
working
on
an
internal
report
or
you
know,
essentially
a
lot
of
data
that
feeds
into
our
annual
report
that
we'll
publish
in
january.
But
one
thing
that
kind
of
shocked
me
this
year
is
that
we
actually
accepted
35
projects,
the
you
know
super
majority
of
them
being
sandpox,
which,
if
you
kind
of
look
at
our
history,
it's
by
far
the
most
we've
ever
accepted.
A
A
We
added
two
new
sigs
this
year
around
observability
and
also
contributor
strategy,
which
has
been
great
in
kind
of
formalizing,
a
lot
of
ideas
to
help
projects
get
started
and
just
being
great
advisors
outside
of
just
the
the
staff
that
usually
handles
the
the
brunt
of
that
work.
As
I
mentioned
before,
we
simplify
the
sandbox
projects,
which
you
know
in
beginning.
I
think
I
had
some
hesitations
there,
but
generally
it's
definitely
improved
the
acceleration
in
which
we've
accepted
projects.
A
I
believe
so
it'd
be
great
to
kind
of
end
the
year
with
five
projects
hitting
that
and
in
that
level
the
big
thing
that
I
like
to
remind
folks
on
you
know
the
toc
is
really
staffed
with
community
members,
folks
that
kind
of
represent
you
know
the
community,
whether
they're
project,
maintainers
ambassadors
and
so
on,
nominations
for
essentially
six
slots
will
be
opening
up
in
january.
We'll
have
six
slots
and
they'll
represent
different.
A
You
know
kind
of
areas
and
and
designated
slots
that
the
toc
has
so
two
of
them
will
be
end
user
appointed.
So
this
is
from
our
end
user
community
that
cheryl
hong
runs
they'll,
be
working
to
appoint
those
three
of
them
will
come
to
the
governing
board
and
one
will
be
from
the
toc
itself
gets
to
appoint
them.
A
So
we'll
go
through
this
whole
process
and
we'll
be
super
public
about
it
remind
everyone,
so
just
pay
attention
and
if
you're
interested
in
running
feel
free
to
reach
out
to
me
or
any
other
toc
members,
and
they
could
kind
of
talk
about
the
you
know,
time,
commitments
pros
cons
and
all
that
wonderful
stuff
that
happens
with
the
roll
not
really
going
to
dive
into
prediction.
Since
liz
had
a
wonderful
keynote
on
this,
I
think
we're
going
to
you,
know
kind
of
move
back
to
hollywood
squares
and
essentially
start
asking.
A
You
know
the
toc,
some
questions
and
and
so
on.
So
let
me
go
stop
sharing
my
screen
here.
So,
let's
see
if
we
could
go
back
to
hollywood
squares
mode,
awesome
kind
of
see
everyone
here
for
the
most
part.
So
before
I
kind
of
kick
off
our
questions
out
there,
you
know
I'd
love
to
kind
of
see.
If
you
know
liz
made
all
these
kind
of
wonderful
predictions,
you
know
around
kubernetes
on
the
edge
webassembly
bf.
A
I
I
think
it's
just
too
hard
to
build
applications
and
we've
focused
a
lot
of
energy
on
operating
applications
and
that's
great
but
like
if
we
don't
make
it
easier
for
people
to
build
these
applications,
I
don't
think
we're
doing
anybody
a
service.
So
I'm
you
know
excited
about
bringing
things
like
no
code
to
kubernetes
and
other
kinds
of
you
know
developer
productivity
that
we
can
focus
on.
I
A
I
A
H
Anything
I've
got
something
I
think.
Extensibility
is
something
that
we've
been
driving
a
lot
on
in
this
community,
and
I
expect
that's
going
to
continue.
I
think
a
lot
of
the
standards
that
have
emerged
like
csi
have
become
well
established
and
have
allowed
kind
of
a
plethora
of
companies
or
solutions
to
be
able
to
integrate
together
with
the
ecosystem,
and
so
I
expect
that
those
extensibility
points
are
going
to
continue
to
expand.
H
Yeah
object,
storage
was
an
interesting
one
where
you
know,
datapath
has
not
really
been
standardized
and
we
don't
really
have
a
good
way
to
be
able
to
plug
in
an
arbitrary
object.
Storage.
Now
the
kubernetes
folks
are
starting
to
look
at
a
cob
container
object,
storage
interface,
and
what
would
that
look
like?
Is
it
you
know
going
to
be
as
extensible
as
csi
has
been
in
terms
of
being
able
to
plug
in
an
arbitrary
object,
storage
system
to
an
arbitrary
workload
on
kubernetes
on
any
environment?
Can
we
get
to
that
point?
A
I
will
move
on
to
our
set
of
questions
and
then
I'll
go
check.
You
know
in
slack
to
see
if
anyone
has
anything
in
particular,
but
to
kind
of
kick
off
our
kind
of
first.
You
know
question
you
know
I'll
I'll
start
with
essentially
the
the
service
mesh.
You
know
space,
so
you
know,
cncf
has
a
you
know
a
handful
of
projects
in
the
service
mesh
ecosystem.
You
know
kuma
smi,
linker
d
envoy,
even
though
matt's
completely
keeps
yelling
at
me.
That
envoy
is
not
a
service
mesh.
A
D
J
Is
not
a
service
mesh?
I
just
want
to
say
that
again,
envoy
is
not
a
server
smash.
You
know.
I
I
like
people,
ask
me
this
all
the
time
and
I'm
gonna
be
honest.
I
don't
know
what
the
answer
is,
and
I
I
think
I
don't
know
what
the
answer
is,
because
service
mesh
is
not
a
very
mature
space
right
now,
there's
lots
of
different
solutions,
I'm
I'm
personally,
obviously
very
biased.
J
I
think
users
are
going
to
come
to
expect
a
set
of
features
and
whether
that
comes
from
projects
like
k,
native
or
other
serverless
platforms-
and
you
know
I,
I
think
they'll
be
using
projects
like
envoy
and
kubernetes
under
the
hood,
but
I'm
actually
personally
very
skeptical
that
we're
going
to
continue
to
see
a
bunch
of
these
independent
projects.
I
think,
over
time
we're
going
to
see
some
of
them
go
away.
Some
of
them
are
going
to
get
merged
into
larger,
larger
projects
and
systems.
J
I
I
think
that
over
time
we
will
I'm
not
a
I'm,
not
a
big
believer
in
a
complete
rewrite
envoys,
hundreds
of
thousands
of
lines
of
code
and
with
millions
of
lines
of
dependencies.
So
I
don't
think
it's
realistic
to
do
a
complete
rewrite.
I
do
think
that
over
time,
via
things
like
web
assembly
and
embedding
rust-
and
I
I
think
we
will
see
envoy-
you
know
start
to
use
those
types
of
systems,
but
no,
I
don't
think
we'll
ever
do
a
complete
rewrite.
J
Yeah,
I
think
less
less
lua
more
more
web
assembly.
I
think
we
will
end
up
using
rust
within
envoy.
I
think
over
time,
it'll
be
a
mix
of
c
plus,
plus
and
rust.
But
you
know
that's
that's
something.
That's
gonna
evolve
over
the
next
couple
of
years
I
mean
there's
hundreds
of
millions
and
billions
of
lines
of
c
plus
plus,
so
I
think,
there's
a
general
problem
out
there
that
we're
not
going
to
rewrite
the
world
and
we
have
to
figure
out
how
to
gradually
migrate
to
safer
systems.
B
Awesome
yeah:
we
are
seeing
more
rust
in
the
cloud
native
space,
which
is
exciting.
We've
at
the
security
space
has
really
adopted
rust
and
we're
seeing
things
like
parsec
and
sandbox,
and
there's
a
lot
of
interest
in
rust
there
and-
and
you
know
we-
you
know
so-
there's
a
bunch
of
rust
projects
and
there's
a
bunch
of
interest
in
rust
and
I
think
there's
a
lot
of.
I
think
we'll
see
a
lot
more
over
the
next
year,
because
a
lot
of
people
you
know
I
talked
to
in
the
space
are
adopting.
A
G
D
I
will
try
to
give
us
50
of
michelle's
awesomeness.
I
think
for
us.
The
bigger
thing
is
something
kind
of
the
point
that
matt
was
making
more.
I
think
about
envoy
that
for
us
talking
about
service
mesh
has
been
really
hard
because
most
of
the
company
doesn't
care.
They
just
want
some
feature
of
service
mesh.
So
for
us
it's
been
really
hard
to
go
around
the
company
and
talk
about
like.
Oh,
I
don't
know,
insert
service
mess
technology
here.
Do
we
care
and
very
few
people
across
the
company
building
features
care?
D
F
Let's
hope
this
works,
so
yeah
sig
networking
has
been
working
on
a
conformance
suite
for
smi
and
that's
a
really
exciting
thing,
because
whenever
conformance
comes
around
and
and
people
work
on
it,
that
means
you
know.
Multiple
people
are
trying
to
implement
one
thing
and
kind
of
converging
in
a
space,
and
I
think
in
the
service
mesh
space,
like
I
mean,
like
everybody
else's
echo,
like
it's
really
about
the
feature
set
not
really
about
the
actual
implementation
and
so
and
and
in
the
greater
cloud
native
community.
F
You
know
to
eventually
end
up
to
eventually
end
up
building
a
really
good
experience
for
end
users
and
eventually
getting
to
that
point
where
they
can
just
be
like
hey
like
I
just
want
retries
and
fault
tolerance
and
a
really
easy
and
seamless
a
b
experience
like
I
think.
Since
2015
people
have
been
asking
like
how
do
I
do
a
b
testing
in
cloud
native
or
kubernetes
environments?
How
do
I
do
canary
deployments
so
once
we
get
to
a
point
where
people
are
like?
This
is
easy.
A
Awesome
awesome.
Thank
you.
Thank
you,
michelle.
It
looks
like
we
finally
have
some
great
audience
participation
here
from
someone
throw
this
to
the
audience.
I'm
also
happy
to
answer
it.
Our
early
stage,
projects
built
by
individuals
likely
to
be
accepted
in
the
sandbox,
or
is
it
more
about
vision
or
project
maturity,
number
of
contributors
feature
sets
stability
and
so
on.
Anyone
want
to
take
a
you
can
see.
E
E
I
Yeah,
I
think
that
was
basically
what
I
was
going
to
say
as
well,
which
is
that
you
know
we're
not
github
right
we're,
not
a
project
hoster
and
we're.
What
we
are
about
is
facilitating
collaboration,
and
so
you
know
I
I
think
that
you
know.
While
we
have
projects
that
are
sort
of
led
by
single
companies,
they
tend
to
a
a
project.
That's
led
by
a
single
company
tends
to
have
a
higher
bar
in
terms
of
number
of
users.
I
A
number
of
sort
of
important
importance.
I
mean
I
think,
there's
ncd
is
a
great
example
of
a
very
important
project
where
there
was
relatively
few
contributors,
and
it
was
basically
from
one
or
two
companies,
but
it
still
was
important
to
have
it.
So
it's
not
like
it's
a
hard
and
fast
rule,
but
it
I
think
that
the
you
know
if
you're
talking
about
fast
path
or
easiest
way
to
do
it,
build
up
a
bunch
of
people
who
want
to
collaborate
on
something
and-
and
it's
pretty
much
becomes
an
obvious
thing
to
add.
H
A
Definitely
so
I
guess
to
kind
of
summarize
that
is
you
know,
I
don't
think
it's
a
no.
If
you
wanted
to
try
to
submit
a
project
that
was
only
one
person.
You'll
get
feedback
eventually
from
the
tocs,
though,
but
I
think
in
general
you
know
a
discussion
behind
each
kind
of
project
is
looked
at
individually
on
on
its
merits.
A
Other
questions,
so
you
know
I've
been
you
know,
meeting
with
you
know
telcos,
fortunately,
or
unfortunately,
you
know
heavily
for
the
last
couple
of
months
and
there's
kind
of
a
lot
of
discussion
around.
You
know
carinity's
on
the
edge
kubernetes.
You
know
you
know
in
smaller
form
factors
you
know
I,
you
know
what
are
people's
thoughts
on
the
toc
you
know.
Are
we
gonna
see?
A
You
know
something
like
what
happened
with
the
linux
kernel,
where
over
time,
just
lots
of
different
people
stretching
in
interesting
directions
and
moving
into
new
form
factors,
and
so
on?
You
know
what
what
role
do
you
think
kubernetes
will
have
you
know
on
the
edge?
We
have
multiple,
let's
call
them
tiny,
distros,
k3s
now
and
k0s.
I
saw
this
this
this
week.
Things
like
cube
edge.
I
think
microsoft
announced
something
called
acry
recently.
There's
a
bunch
of
these
things
popping
up
all
over.
C
A
B
G
C
Yeah
from
the
end
user
perspective
speaking,
all
users
are
looking
for
they're,
looking
for
a
universal
way
of
delivering
their
applications
to
to
different
locations,
whether
it's
a
platform
in
your
organization
or
something
something
on
the
edge
and
that's
where
kubernetes,
on
the
edge
really
strikes
as
for
implementation,
so
we
have
qbh
open
yard.
We
have
k3s
some
challenges,
they
share
some
challenges
and
some
are
unique
to
the
project.
C
So
I'm
just
curious
to
to
look
at
them
and
just
you
know
see
for
the
shared
challenges,
what
what
can
be,
what
can
be
improved
and
for
the
new
ones
like
for
every
new
project.
It's
always
important
to
look
and
see
like
what
what
differs
it
from
other
water
edge
projects
are
already
part
of
cncf.
What
unique
it
brings
to
the
ecosystem.
K
K
K
I'd
be
you
know,
as
toc
would
always
ask
you
what
makes
you
different
and
what's
unique,
but
but
a
lot
of
these
things
you
don't
really
know
until
until
you
start
working
on
it
and
engage
the
users
engage
the
community,
so
I
definitely
would
engage
would
encourage
folks
to
to
come
up
with
with
new
distros
or
on
your
projects,
because
this
there's
just
so
many
ways
to
to
do
the
same
thing
and
it's
just
a
standard.
It's
just
not
there.
Yet.
A
Awesome
we
have
a
few
minutes
left.
I
kind
of
want
to
get
the
last
question
for
us.
You
know
most
of
our
end
users,
but
you
know
other
other
folks
on
the
teams
that
can
answer
this.
You
know
I've
been
working
with
a
lot
of
different
end
users.
Companies
recently,
I
helped
bring
an
organization
called
the
finnops
foundation,
the
linux
foundation,
which
is
mostly
dedicated
to
you,
know,
wrangling
cloud
financial
management.
For
some
of
us,
you
know
who
may
work
at
a
cloud
provider.
A
You
kind
of
have
this
notion
of
cloud
privileges
like
everything's
like
free,
pretty
much,
but
you
know
if
you're
working
at,
maybe
you
know
spotify
or
you
know
smaller
company,
this
stuff
could
get
fairly.
You
know
expensive,
you
know
quickly
and
the
tools
to
kind
of
manage,
built
or
predicting
the
cost
of
spinning
the
service
just
seems
not
to
be
there.
Does
anyone
have
any
thoughts
of
you
know
how
this
space
may
have
all
the
work?
What
could
be
done
here,
because
we
don't
have
any
projects
at
all
in
cncr?
C
Yeah
capacity
planning
is
definitely
is
definitely
challenging
and,
to
be
honest,
I
don't
know
like
what:
what
are
the
tools
out
there
that
can
solve
everything
like
we
are
building
something
internally,
but
I
would
be
happy
to
see
more
tools
evolving
in
the
space,
because
it's
a
shared
problem
right,
no
matter
how
big
your
company
is,
you
need
to
spend
your
money
wise
and
do
the
and
do
the
planning
carefully
so
yeah.
I
agree.
That's,
I
think,
that's
a
that's
a
good
point.
J
J
I
was
going
to
say
quickly
that,
from
my
experience,
one
of
the
biggest
problems
that
people
face
in
this
space
is
not
only
the
tooling
but
actually
tagging
people's
tagging,
people's
resources
to
figure
out
how
to
do
cost
attribution
and
there's
a
couple.
Companies
that
are
in
this
space
right
now
and
what
I'm
actually
hopeful
of
is
that
we
can
develop
a
shared
standard,
something
something
like
cloud
events
but
actually
related
to
cost
attribution.
J
And
I
I
think
that
having
that
type
of
api
will
make
it
much
easier
to
have
tooling
across
platforms
like
kubernetes
and
other
cloud
providers
that
can
help
us
figure
out.
You
know
first
how
to
tag
everyone's
resources,
because
without
having
that
resources
tagged,
it's
virtually
impossible
to
make
any
progress
here
and
then
once
that's
in
place
and
I'm
hopeful
that
can
be
an
open
standard.
Then
I
think
there
can
be
companies
and
there
can
be
tooling
on
top.
That
can
help
people
actually
with
with
cost
cutting.
D
Sure,
I
guess
the
biggest
thing
I'll
say
is
the
same
thing
that
matt
just
ended
with
that
like
there's,
always
so
much,
I'm
gonna
say
recorded,
but
I'm
totally
happy
to
talk
to
anyone
about
this,
and
just
like
elena
said,
spotify
has
built
a
bunch
of
internal
tools
for
things
like
this,
whether
it's
predicting
or
even
just
understanding
and
labeling
or
cost
attribution.
All
of
that
stuff.
D
We've
tried
to
open
source
some
pieces
of
that,
but
I
would
love
to
turn
as
much
of
that
possible
into
an
open
standard
like
matt
was
saying
and
like
instead
of
each
company
building
their
own
thing,
that
does
10
percent
of
the
job
poorly.
Let
us
come
together
and
build
something
that
does
much
more
of
it
better.
A
Awesome
yeah.
Definitely,
I
think
you
know
that
could
be
an
exciting
area
that
if
people
have
projects
in
that
space,
please
contribute
them.
We'd
love
to
kind
of
hear
from
you.
So
with
that
you
know
wrapping
up
our
30
minutes
or
so
thank
all
of
our
panelists.
You
know,
you
know,
folks
are
definitely
you
know
busy,
but
everyone
is
very
approachable,
so
feel
free
to
you
know,
jump
in
on
toc
meetings.
A
We
nearly
have
you
know
you
know
70
to
sometimes
even
100
people
on
these
things,
so
please
jump
on
them
and
then
submit
any
projects.
You
have
we're
happy
to
help
you.
So
thank
you,
everyone
for
taking
the
time
you
know
cons
and
wild
and
please
you
know,
follow
us
on
on
slack
twitter
and
whatever
communication
meeting
refers.
So
thank
you
all
right.
Take
care.