►
From YouTube: Kubernetes WG IoT Edge 20220420
Description
April 20, 2022 meeting of the Kubernetes IoT Edge Working Group. Discuss proposal to retire operation of this this group under the Kubernetes project, and apply to open a new group under the CNCF organization
A
Hi
welcome
to
the
April
20th
meeting
of
the
kubernetes
iot
edge
working
group
on
today's
agenda
unless
somebody
just
added
something
in
the
last
five
minutes.
The
only
item
on
there
is
something
I
put
on
and
that
was
to
discuss
a
graceful
retirement
of
this
group
underneath
the
kubernetes
project,
with
a
proposal
to
reconstitute
a
replacement
group
underneath
the
cncf
organization.
A
We
were.
You
know
this
group,
like
all
sigs
working
groups.
User
groups
under
the
kubernetes
project
has
a
recurring
obligation
to
give
a
report
to
the
kubernetes
steering
committee
and
some
people.
There
have
made
a
suggestion
that
maybe
that
the
operation
of
this
group
isn't
doesn't
quite
fall
under
the
intent
of
working
groups
and
that
maybe
a
more
appropriate
place
for
this
group
might
be
under
underneath
the
cncf
itself.
A
I
believe
that
part
of
the
reasoning
there
is
that
this
group
originally
got
involved
with
a
number
of
efforts
to
customize
kubernetes
for
iot
Edge
use
cases
and
had
a
in
the
beginning,
almost
a
one
for
one
overlap
with
kubernetes
itself.
But
over
the
years
a
number
of
collateral
projects
have
been
spawned,
I
mean
Kate's
on
the
call
and
has
acri
kilton
with
IO
fog.
A
It
isn't
the
cncf
project,
but
it
is
a
related
project
and
a
lot
of
this
group's
discussion
activity
actually
is
related
to
projects
that
are
related
to
kubernetes,
but
aren't
part
of
kubernetes
itself,
so
that
perhaps
rehoming
this
group
over
under
the
cncf,
where
a
lot
of
these
collateral
projects
are
already
operating,
is
a
more
appropriate
place
to
broaden
the
scope.
A
This
retirement
move
hasn't
happened
yet
and
I,
don't
personally,
don't
see
that
there's
any
rush
to
do
it,
but
we
have
had
some
informal
discussions
in
slack
suggesting
that
maybe
we
could
get
this
done
in
a
time
frame
of
perhaps
kubecon
North
America,
where
we'd
have
a
venue
to
pop
back
up
over
there
and
resume
operation,
perhaps
in
a
modified
form.
But
anyway,
that's
the
agenda,
item
and
I.
Think
Dion
and
Kate
have
had
some
conversations
with
people
on
the
cncf,
so
maybe
I'll
turn
over
the
microphone
to
them.
B
Sure
I
can
add
in
a
little
note
on
that
So
within
the
cncf.
There's,
a
bunch
of
tags
in
all
of
the
working
groups
fall
under
the
tags,
and
so
one
discussion
for
us
would
be
what
working
group
fits
the
best,
like,
obviously
the
edge
spans,
multiple
working
groups
and
that
doesn't
prevent
us
from
contributing
with
other
working
groups
and
other
tags.
B
But
one
tag
that
would
be
an
option
and
maybe
is
the
best
fit,
is
the
runtime
tag,
and
so
in
our
discussion,
one
of
the
co-chairs
of
the
runtime
talk
has
mentioned
that
we
could
fall
under
that
space.
If
that's
a
space
where
we
wanted
to
land
so
just
thoughts
about
that.
C
Yeah
I
was
going
through
all
this
from
from
our
recent
conversation
today
and
I.
I
think
that's
that
that's
the
best
space,
especially
is
runtime
Tech,
explicitly
says
that
that
agent
iot
is
is,
is
a
part
of
their
mission
and
the
scope,
so
so
I
think
it
will
be
a
a
good
fit
for
for
moving
the
group
there
right
foreign.
A
Yeah
I
think
that
if
it's
anything
like
what
the
process
is
under
the
kubernetes
project,
they
they
have
some
guidance
in
the
in
community
documents.
But
perhaps
the
easiest
thing
to
do
is
to
go
look
at
another
group
that
has
formed
recently
and
they
would
have
had
a
PR.
So
you
can
go
look
at
that
and
maybe
cut
and
paste
that
as
a
starting
point,
since
it
would
be
an
example
of
one
that
worked,
went
through
and
got
accepted,
I'm
not
sure
you
know
we
could
start
the
PR,
but
maybe
not
submit
it.
A
Yet,
if
that's
possible,
because
I
think
we
might
probably
want
to
have
some
discussion
over
it
before
we
throw
it
out
there,
perhaps
even
discussion
with
the
people
who
will
need
to
approve
it.
Not
just
people
involved
and
part
of
this
discussion
would
be
to
recruit
people
who
are
going
to
be.
You
know
on
the
leadership
and
participation
of
this
group
as
well.
So
you
know
I,
don't
think
you
want
to
start
a
working
group
that
is
empty
on
day,
one
that
you'd
like
to
have
people
who
got
informed
of
the
migration.
A
So
you
can
start
the
thing
on
a
good
basis
and
we
shouldn't
assume
it's
a
done
deal
I
mean
I,
I,
don't
know
if
kilton
and
Frederick
have
even
heard
of
this
before
so
we
should
just
open
this
up
for
comment.
We've
had
some
discussion
because,
as
part
of
this
report,
we
had
to
give
steering
they.
They
gave
communication
back
on
slack
on
where
this
might
be
headed.
But
this
isn't
this
isn't
firmly
resolved.
I
mean
if
somebody
really
wanted
to
push
for
continuing
this
group's
existence
under
the
kubernetes
project.
Maybe
that's
a
possibility.
A
So
far
we
haven't
really
pushed
and
tried
we've.
Some
of
us
have
had
some
sidebar
discussions.
You
know
discussing
whether
we
should
push
for
continued
existence,
whether
the
cncf
was
a
good
home,
whether
perhaps
the
Linux
Foundation
lfedge
might
be
a
better
home
because
I
don't
believe
they
have
a
group
quite
like
this.
We
also
did
this
us
the
existence
of
the
edge
group
on
the
eclipse
Foundation,
which
I
don't
think
that
you
know
if
this
got
disbanded
altogether.
I
think
I
for
one
might
just
pop
up
over
there,
but
there's
no
need
to
create.
A
Another
group
I
think
that
you
know
that
the
kubernetes
topic
could
fold
underneath
the
one
that's
already
there,
and
it
is
kind
of
interesting
whether
the
world
even
needs
to
groups
but
I
think
the
justification
for
that
is
that
there
are
Edge
related
Projects
Home,
underneath
Eclipse
Foundation,
as
well
as
under
CNC
app.
So
maybe
that
that,
on
its
own
is
enough
justification
to
have
two
different
groups,
I
mean
because
both
these
organizations
are
running
their
own
events.
A
They
have
a
little
different
spin
on
where
they're
going
I
think
and
even
the
LF
Edge
I
mean
sometimes
it's
a
little
difficult
to
you
know
come
up
with
a
justification
for
why
three
existed
but
hey.
If
things
are
working,
don't
try
to
fix.
It
is
kind
of
my
attitude
that,
and
there
there
is
kind
of
value
in
the
redundancy
of
a
little
bit
of
duplication
in
an
ecosystem.
A
You
know
where
you
don't
have
just
one
biological
life
form
taken
out
by
a
parasite
or
something
if
you
have
a
little
bit
of
organ
of
competition.
Even
in
these
open
source
organizations,
who
maybe
maybe
that's
a
good
healthy
thing
for
the
world
to
have-
is
sort
of
my
observation
but
I
I'd
like
to
hear
any
comments
we
might
have
from
Frederick
and
Hilton.
D
Yeah
well,
in
my
case,
I
mean
I'm
fine
with
whatever,
in
the
sense
that
we
are
members
of
the
cncf.
We
are
members
of
felipege
and
elefedge
is
a
member
with
us
and
certainly
I.
Think
you
characterized
well
the
the
situation.
You
know
the
reason
why
we
have
our
own
Edge
walking
group
at
the
eclipse
Foundation,
which
is
essentially
to
provide
a
gathering
place
for
the
people
who
work
on
our
own
Edge
Computing
project,
so
that's
Eclipse,
IO
fog,
Eclipse
fog,
OS,
Eclipse,
Zeno
and
Eclipse
Kanto.
D
So
we
have
four
of
them
now
with
the
possibility
of
having
a
fifth
one,
maybe
later
this
year.
So
you
know
that
when,
when
we
get
her
there
and
you've
been
you've,
been
there
Steve,
we
have
a
tendency
to
to
focus
more
on
what
matters
to
our
projects.
But
in
any
case,
if,
if
you
lose
a
home,
we
are
certainly
more
than
happy
to
have
you
more
more
frequently
in
our
meetings
and
and
even
open
them
to
to
a
wider
Community.
D
You
know
it's
my
my
job
as
the
guardian
there
to
to
make
sure
that
we
we
reach.
You
know
whatever.
Whatever
audience
the
the
group
collectively
wants
to
reach
so
in
any
case,
I
wouldn't
be
certainly
opposed
to
create,
let's
say
a
new,
a
new
group
on
the
red
effect
or
under
whatever
place
you
feel
is
the
best
and
I
think
you
Steve
and
Dion
are
probably
best
place.
D
A
She's
she's
gotten
involved
in
the
last
year,
but
our
anticipation
is
actually
that
perhaps
the
leadership
will
start
out
with
Dion
and
Kate
and
I
I
have
been
frequently
seen
in
this
group,
but
I
am
not
one
of
the
chairs
of
the
group
I'm.
Just
you
know.
I've
I've
got
a
tech
lead
role
but
have
not
held
the
chair
role
and
Cindy
Zhang,
who
has
been
the
chair
change
jobs
in
the
past
year
and
is
working
on
something
unrelated.
A
Kilton.
Do
you
have
any
comments
at.
E
The
at
the
on
the
eclipse
side
in
the
edge
native
working
group
there
it
would
certainly
be
of
value
to
have
some
amount
of
time
dedicated
to
kubernetes
Edge
topics
so
that
that's
so
operationally
I
see
no
problem,
finding
a
space
to
say,
okay.
E
This
is
where
you
know
this
portion
of
this
meeting,
or
this
entire
meeting
is
dedicated
to
to
you
could
basically
call
it
cloud
and
Edge
right
stuff
where
kubernetes
is,
of
course,
the
you
know
the
the
ruler
of
cloud
and
and
or
just
strictly
kubernetes
right
and
so
yeah
I
think
what
you're?
What
is
done
in
this
group
has
a
home.
The
question
is:
where
is
it
going
to
be,
and
you'd
certainly
be
welcome
over
at
the
the
eclipse,
Edge
native
working
group.
A
Okay
well
well,
thanks
I
think
tentatively,
we
still
are
going
to
look
at
perhaps
cncf,
just
because
the
cncf
is
hosting.
Obviously
you
know:
Cube
Edge
I
think
Super
Edge
open
your.
There
is
a
fair
over.
A
A
Yeah
I
was
going
to
say
that
the
trouble
I
have
if
I
drop
is
I'm,
not
sure
that
anybody
else
has
logged
in
under
the
host
credentials.
Let
me
just
in
case
it
happens
again,
I'm
going
to
make
a
point
a
few
co-hosts
that,
as
a
backup
plan.
A
Yeah
I
don't
know
I've.
My
system
is
managed
by
my
organization's
I.T
and
it
got
upgraded
to
Windows
11
fairly
recently,
and
there
have
been
a
few
glitches.
Maybe
it's
maybe
they're
unrelated
and
just
coincidental,
but
I've
I've
noticed
some
odd
behavior.
So
maybe
this
was
one
of
them.
I,
don't
know
so
yeah.
A
The
other
thing
where
I
think
this
makes
a
lot
of
sense
is
we've
had
this
very
successful
kubernetes
on
the
edge
day
hosted
by
the
cncf
and
they've
been
co-locating
it
at
kubecon,
Cloud
native
cons,
I,
haven't
heard
attendance
figures
for
the
one
in
Europe,
but
in
its
the
the
last
two
that
actually
took
place
had
over
600
attendees
each.
So
that
is
a
pretty
big.
You
know
the
the
attendance
at
this
working
group's
own
meanings
have
not
been
that
great,
but
for
those
events
that
there
there
is
a
pretty
big
following
and.
E
I
was
going
to
ask
I
was
going
to
ask
how
that
the
attendance
and
how
that
event
has
been
viewed,
because
the
event
of
course
arose
from
this
working
group
and
has
now
grown
to
become
something
more.
But
this
working
group
has
always
been
the
home
for
looking
at
the
paper
submitted
did
right,
the
talks
submitted
and
all
that
so
I
would
think
that
the
success
of
that
that
sub-conference
would
indicate
that
yeah
that
this
group
really
just
needs
to
be
re-structured
under
the
cncf
right
and
so
yeah
yeah.
A
A
Think
that
that's
a
natural
progression
really
when
you
get
down
to
it,
look
at
all
of
these
open
source
things
that
spawned
off
the
Linux
Foundation,
but
how
much
of
them
you
know
how
much
of
a
talk
or
a
conversation
in
a
meeting
really
is
related
to
Linux
I
mean
when
you
get
down
to
it,
kubernetes
itself.
A
Technically,
there
are
windows
containers,
but
I
mean
the
vast
majority
that
I
think
is
Linux
Centric,
but
if
that
underlying
infrastructure
project
did
its
job
in
the
long
run
it
drops
down
to
invisibility,
you
know
it
falls
between
behind
abstraction
layers
so
that
the
typical
users
don't
end
up
realizing
it's
even
there.
If
you
did
a
really
great
job
of
abstracting
out
the
complexity
so
that
it
just
works
and
they
never
have
to
touch
it
sooner
or
later
it
becomes
invisible
and
maybe
that's
part
of
what's
Happening
Here
I,
don't
know.
C
Yeah
I
mean
from
the
start
the
kubernetes
working
group.
How
they
are
structured
is
is
to
have
a
limited
lifespan,
to
tackle
one
area
technically
and
and
try
to
solve
one
one
issue.
We
started
with
that,
but
it's
outgrown
its
original
purpose
and
and
became
more
like
a
user
user
group,
but
for
things
much
more
than
just
kubernetes,
as
you
said,
like
it's
all,
Cloud
native
for
the
edge
right,
there's
a
lot
of
talks
about
observability
about
you
know,
cicd
or
all
the
all,
all
that
kind
of
things.
C
A
One
of
the
things
I'd
like
to
suggest-
and
this
came
about
even
in
the
discussions
with
the
kubernetes
steering
committee-
and
you
can
tell
it
in
these
reports-
we're
obligated
to
put
in
they're
focused
on
metrics,
like
code
contributions,
which
really
the
way
this
group
was
operating,
didn't
make
any
sense.
You
know
we
weren't,
we
weren't
checking
in
or
approving
code
to
the
kubernetes
project
itself
and
over
time.
A
In
the
early
days
when
kubernetes
was
talking
about
Edge,
there
were
people
entertaining
things
like
changing
the
scheduler,
changing
the
cubelet
for
Edge
use
cases,
but
when
you
get
down
to
it
over
time,
I
think
kubernetes
realized
that
the
thing
the
amount
of
entry
code
was
getting
too
big
and
too
complex
and
even
they
even
started
kicking
out
things
like
storage
plugins
from
day
one
I
think
Network
plugins
were
sort
of
divorced,
but
they
wanted
a
structure
where
kubernetes
was
just
a
basic
framework
for
plugging
in
options
to
implement
particular
things
and
those
options
would
live
out
of
the
kubernetes
tree
itself.
A
Now,
if
we
re-home,
we
could
perhaps
Envision
some
scope
where
we're
code
related,
but
I
have
a
I'm
thinking
that
we
are
really
more
of
something
akin
to
a
user
group,
because
when
you
look
at
the
talks,
both
at
kubernetes
on
the
edge
day
and
at
meetings,
a
lot
of
this
exists
to
expose
kind
of
best
practices
and
share
ideas
for
users
trying
to
deploy
Cloud
native
Solutions
on
the
edge,
some
of
it
being
kubernetes.
Some
of
it.
A
You
know,
I,
think
that
many
of
us
have
Come
Away
with
an
attitude
eventually
that
kubernet,
sometimes
kubernetes.
You
know
kubernetes
can
be
a
great
hammer,
but
not
every
application
for
a
tool
calls
for
a
hammer
and
in
some
cases,
with
resource
constrained
Edge.
A
Maybe
some
other
solutions
for
running
containerized,
apps
or
maybe
apps
that
aren't
even
containerized
are
good
and
kubernetes
can
be
part
of
the
big
picture,
but
it
isn't
the
solution
for
every
problem
you
might
have,
and
you
know
a
lot
of
this
comes
down
to
either
a
user
Focus
or
you
could
all
maybe
call
it
a
systems.
Integration
Focus,
where
what
you're
trying
to
do
is
hook
together,
a
bunch
of
projects
but
you're
not
doing
code
in
any
one
project,
you're
just
kind
of
attaching
the
building
blocks.
B
Well,
I
was
I
was
going
to
say
kind
of
the
continuation
of
a
little
bit
earlier.
I
think
we
were
we
kind
of
grew
out
of
the
kubernetes
working
group
because
we
weren't
kubernetes
specific,
like
you
were
saying
Steve
we
were.
We
wanted
to
push
beyond
that,
whether
it
was
to
more
of
non-kubernetes,
specific
Edge
workloads
and
then
also
web
assembly
stuff.
B
So
we
kind
of
grew
out
of
the
kubernetes
working
group
and
I
think
growing
into
a
cncf.
One
makes
sense
just
because
it
has
a
bigger
scope
and
there's
kind
of
a
need
for
someone
to
fill
that
space
there
like
a
place
for
people
to
locate,
to
talk
about
Edge
initiatives,
which
is
why
Edge
day
is
so
popular.
B
It's
getting
pretty
popular
at
kubecon,
but
with
the
discussion
of
User
Group
being
a
focus
of
it
or
like
an
infrastructure
group
I
think
if
we
were
to
go
cncf
route
I
think
we
still
need
some
sort
of
byproduct
of
the
group.
It
says
on
there
so
whether
that's
like
a
white
paper
or
that
is
reports
of
some
sort,
I
think
a
byproduct
could
be
Edge
day.
B
But
I
do
think,
alongside
being
a
user
group
or
another
approach
we
could
take
is
kind
of
by
being
like
a
surveying
group
like
a
group
that
can
kind
of
provide
standards
is
something
that
we
may
we
brainstormed
a
little.
So
maybe
this
is
a
group
where
people
bring
in
their
expertise
and
Define
what
what
are
some
ideal
footprints
for
what
scenarios
and
can
people
pass
certain
operational
tests
and
kind
of
getting
people's
Edge
Solutions
certified
or
providing
maybe
some
Scopes.
B
E
E
E
B
B
Better
I'm
back
yeah,
yes,
I,
think.
B
B
On
the
edge
and
then
categorizing
projects
based
on
where
they
fit
talking
about
something
the
differences
there
I
think
there's
a
lot
of
exciting
work
that
could
be
done
there.
A
B
C
We
had
a
little
bit
of
discussion,
a
lot
of
demos
as
well
right,
so
so
that's
something
I
mean
well
more
more
these
days,
you
know
Building
Systems,
with
all
these
projects
that
happen
and
showcasing
how
how
actually
this
can
this
can
be
done,
but
one
thing
I
wanted
to
say
is:
is
that
one
thing
that
I
liked
about
Tech
runtime,
that
that
is
proposed
as
a
new
home
is
that
kubernetes
itself
is
is
under
under
that
text?
So
there's
nothing
to
stop
us
to
to
basically.
C
You
know
propose
improvements
to
to
kubernetes
as
as
a
part
of
this
working
group,
because
it's
it's
it's
under
the
same
same
Tech,
runtime
there
right
so
whatever
we
need
or-
or
you
know,
play
more
with
the
device
plug-in
and
and
what
K
is
doing
with
accurate.
That's
all
I
think
perfect
match
for
for
for,
for
that,
take.
A
So
Kelton,
going
back
to
the
reference
architecture,
thing
I
think
we
had
a
little
bit
of
discussion
about
this
in
the
last
meeting
and
what
I
proposed
there
was
that
I
would
just
as
soon
not
sign
up
for
this
group
to
be
sort
of
the
over
our
all
authority
on
what
the
best
architecture
is
or
not
partially,
because
you
know
I,
don't
know
that
we're
qualified.
But
what
scares
me
even
more
is
anytime.
A
You
do
that
you're
signing
up
for
a
lifelong
obligation,
because
the
best
reference
architecture
today
is
probably
going
to
have
a
sell-by
date
of
six
months.
The
way
things
move
so
rapidly
in
this
area
that
you're
signing
up
for
a
lifelong
amount
of
work
and
it's
going
to
go
stale
realistically
I,
don't
know
that
we'd
be
able
to
keep
up,
and
the
thought
I
had
is
more
along
the
lines
of
coming
up
with
a
suite
of
metrics
and
test
cases
where
people
would
self-certify.
A
So
if
somebody
wants
to
propose
a
reference
architecture-
and
there
might
be
many-
I
mean-
let's
face
it-
there
are
so
many
use
cases
at
Edge
that
the
best
reference
architecture
for
smart
cities
might
be
quite
different
from
a
reference
architecture
for
small
retail
versus
device,
Edge
industrial
control
and
there's
no
reason
that
you
can't
have
multiples.
But
if
we
could
have
a
consistent
set
of
metrics,
like
you
know,
for
what
you're
outlying
here,
how
much
resource
does
this
need?
How
many
CPU
cores
of
what
type,
how
much
memory
you
know?
A
What
would
I
expect
if
I'm,
a
prospective
user
trying
to
go
down,
follow
this
roadmap?
What
do
I
need
to
make
it
happen?
And
anyone
who
wants
to
put
this
on
the
list
kind
of
owes
a
self-certification
and
all
we
do
is
host
a
mechanism
to
publish
these
things
and
we
aren't
the
arbitrator
of
the
one
correct
religion.
If
you
will,
we
just
are
a
sounding
board
to
put
these
things
up
now.
A
It
gets
to
another
thing,
a
point
I
want
to
make,
and
that
is
I
use
the
term
user
group,
but
I'm
kind
of
hesitant
here,
because
I've
seen
attempts
to
do
user
groups
in
these
communities
and
the
trouble
is
that
most
users,
I
would
contend,
don't
really
have
their
backing
of
their
organization
to
be
spending
several
hours
per
week
on
this
kind
of
activity.
A
But
people
who
are
getting
backed
by
a
vendor
often
do
and
because
of
that,
it's
very
tough
to
recruit
users
who
actually
show
up
and
kind
of
do
the
chopping
wood
carry
water
kind
of
tasks
to
keep
a
group
going
because
there's
real
work.
There
I
mean
the
the
work
of
holding
meetings
recruiting
speakers
I,
don't
know
reviewing
cfps
for
conferences.
That
I
think
everybody
on.
A
This
call
is
aware
of
how
much
time
that
could
take
and
it's
something
where
you
want
to
have
the
backing
where
whoever
is
giving
you
a
paycheck,
considers
that
part
of
your
job
and
often
in
user
organizations
it
either
wouldn't
be
considered
that
or
some
big
users
really
have
a
lot
of
skin
in
the
game,
but
they
consider
it
a
competitive
advantage
and
aren't
necessarily
that
much
in
into
sharing.
If
this
is
really
an
edge
that
you
know
where
some
executive
in
the
organization
would
consider
it
to
be
hey.
A
We
invested
a
lot
of
money
in
learning
all
this,
and
do
you
propose
that
you're
going
to
go
out
in
a
public
forum
and
just
explain
to
our
number
two
or
three
competitors
how
to
do
exactly
what
we
spent
millions
of
dollars
doing
and
it
just
doesn't
come
off
so
I.
A
If
we
give
ourselves
the
official
label
User
Group,
somebody
is
going
to
ask:
why
aren't
users
in
the
leadership
and
that's
a
tough,
ask
and
I'm
kind
of
curious,
because
kilton
and
Frederick
you
are
operating
similar,
similar
groups
in
other
organizations
and
what
what
has
been
your
experience
with?
Actually
getting
users
as
consistent
contributors
to
the
operation
of
a
group?
Can
it
be
done
and
any
advice
you
might
have?
If
we're
going
to
Charter
a
new
one
would
be
welcome
cool.
E
There's
I'm
going
to
touch
on
several
several
points
that
stacked
up
first
is
a
Dan
I.
Think
the
runtime
tag
is
awesome.
I
think
that's
a
I
think
that's
a
very
because
it
really
does
also
hit
on
the
need
for
something
that
operates
at
the
edge
that
links
with
your
kubernetes-based
back
end.
That
might
be
expansive.
It
might
be,
you
know
a
webassembly
at
the
edge,
it
might
be
Docker
containers
it
might
be
whatever,
but
it
really
does.
E
Think
that
your
description
of
what
the
not
reference
architectures,
but
call
them
more
like
demonstration,
architectures
right
or
example,
architectures,
that's
much
more
aligned
with
what
I
was
thinking
I,
don't
think
that
we
would
want
to
be
putting
out
architectures
that
are
meant
to
be
the
right
way
to
deploy
something,
but
rather
something
you
can
use
for
recreating
it
yourself
or
aligning
for
a
particular
use
case.
E
Some
of
the
items
right
I
love
the
idea
of
having
metrics
and
being
able
to
certify,
because
those
are
the
things
that
would
then
require
the
the
Brain
Trust.
That
is
this
group
and
the
people
that
come
in
and
make
presentations
right
in
order
to
get
those
things
right.
But
then
those
things
are
independent
of
how
you
implement
and
they
should
hold
over
time,
and
if
they
need
to
be
adjusted,
it
would
be
more.
E
It
would
be
less
frequently
right
than
actual
implementation
details
and
which
gets
to
then
the
the
idea
of
user
groups
or
being
a
user
group
yeah
I
think
that
if,
if
you
it's
very
natural,
they
have
of
vendors
whenever
you're
talking
about
like
demonstration
architectures.
Oh
vendors,
love
that,
because
let
me
do
one
that
shows
how
to
use
blah
as
the
database
backend
great.
E
I,
don't
think
that
qualifies
as
something
that
we'd
like
to
publish
so
we'd
have
to
figure
out
the
right
way
to
walk
that
line,
and
then
it
comes
to
the
final
question
which
is:
do
do
users
actually
get
the
chance
to
dedicate
and
so
on.
I
would
say
that
for
a
minute
there,
that
minute
being
probably
2018
and
19
it
looked
like
yes
and
into
2020..
E
The
world
is
in
a
state
right
now,
where
it
seems
that
those
who
are
using
Edge,
Computing
Technologies,
don't
have
the
time
or
the
budget
support
to
come
and
talk
to
the
world
about
what
they're
doing
they're
just
getting
down
to
work,
and
so
at
the
edge
native
working
group,
we
have
noticed
that
people
are
not
able
to
come
as
consistently
or
at
all
like
they
used
to
be
able
to
to
just
talk
about
their
use.
E
So
no
new
exploration
for
right
now,
given
that
the
best
thing
that
we
can
probably
do
to
serve
the
edge
Computing
Community
from
a
kubernetes
perspective
would
be
to
make
sure
that
we
keep
doing
something
that
draws
together,
the
Brain
Trust
that
is
needed
in
order
to
produce
value
and
in
a
couple
of
years
time
I
think
more
users
will
start
to
come
back
to
it
and
could
focus
at
that
time
around.
What
are
you
doing
with
this?
How
is
it
going?
What
are
your
concerns
in
your
interests?
E
We
have
the
users
of
the
The
Edge
Computing
Technologies,
hosted
by
Eclipse,
so,
for
example,
iofog
we
have
them
occasionally
still
show
up
and
talk
about
what
it
is
that
they're
experiencing
or
what
they
want
to
see
and
so
on.
But
the
thing
that
seems
to
be
working
for
us
right
now
most
consistently
is
to
be
able
to
represent
the
interests
that
people
have
in
Edge
and
then
be
a
a
knowledge
group
that
can
filter
and
say
yes.
E
This
would
be,
for
example,
how
Global
sign
fits
into
Edge
appropriately
because
of
what
they
do
with
identity
signatures,
blah
blah
blah
and
here's
where
it
slots
in
and
to
help
the
rest
of
the
world
understand
what
is
and
is
not
relevant,
but
then
as
a
vendor
that
that
is
interesting
for
them.
That's
just
like
one
light
example:
yeah.
D
Yeah
I
said
it's
only
second,
everything
that
you
said
kilton
one
thing
I
would
add:
we've
seen
a
bit
more
and
users
show
up
in
our
spot,
plug
working
group
and
then
Spa
plug
is
a
specification
and
implementation
on
the
top
of
mqtt
blah
blah.
Whatever
I
mean
it's
not
directly
relevant
to
this
crowd,
but
as
as
a
group,
it
attracted
more
consistent
participation
from
users
of
the
technology,
because,
essentially
with
with
this
open
specification
and
the
compatibility
program
that
we
will
deliver
around
it,
then
you
know
some
end.
D
Users
are
are
compelled
to
come
to
the
table
to
provide
requirements
so
that
essentially,
what
the
spec
and
its
implementations
will
deliver
is
something
that
they
want
to
buy.
Okay,
so
in
in
in
the
case
of
kubernetes
and
marginally
speaking,
Edge
and
kubernetes
together,
I,
don't
think
we
can
replicate
that.
However,
I
see
potential,
you
know
you
alluded
to
the
fact
you
want
to
Define
metrics
and
things
like
that,
so
that
people
can
can
see
if
they
conform
to
that
or
not.
D
So
maybe
if
you
can
get
a
requirements
so
to
speak
for
those
things
from
actual
then
users,
maybe
now,
after
that
they
will
want
to
have
a
stake
in
that
thing
right
and
and
start
showing
up
just
to
show.
If
there's
new
goodies,
they
can
buy
that
that
conform
to
that
or
new
architectures
that
Implement,
whatever
they
provide
you
requirements
for
So,
just
an
idea
like
that.
A
A
I
think
we
might
have
trouble
recruiting
people
to
to
fill
those
jewel
shoes,
at
least
for
now,
but
obviously
we're
not
going
to
ban
them,
and
the
whole
purpose
of
the
group
really
is
to
serve
those
if
it's
just
nothing
but
vendors
talking
to
other
vendors
I,
don't
think
this
is
a
this
wouldn't
even
be
of
interest
to
the
vendors
themselves.
I
mean.
Ultimately,
the
group
exists
for
users.
D
B
Is
it
too
much
to
be
a
group
for
multiple
categories?
So
can
we
be
a
group
for
end
users
and
then
also
a
group
for
Engineers
who
are
trying
to
architect
their
Solutions
and
want
some
example,
reference
diagrams
and
also
a
group
for
vendors
to
come
together
and
establish
Norms?
Is
it
too
much
to
try
to
be
all
of
that
foreign.
A
I,
don't
think
so,
I
think
to
some
extent
it
makes
sense.
You
know
if
you've
got,
we've
got
communities
like
people
who
make
kubernetes
distros
that
are
targeting
Edge
need
to
talk
to
these
other
collateral
projects
that
say,
do
device
management
that
aren't
kubernetes
per
se
and
there
are
clear
interfaces
there
and
then
kind
of
It's,
The
End
user
organizations
who
would
deploy
them
in
a
use
case.
That
would
kind
of
be
the
overall
systems
integration
point
that
would
kind
of
use
the
whole
broad
category.
A
But
it's
you
know
with
regard
to
to
vendors
or
even
project
managers
for
open
source
projects.
They
need
to
worry
about
interfaces
to
other
open
source
projects
and
getting
those
well
architected.
So
I
think
we
have
a
pretty
broad
potential
audience
and
when
we
get
started
we
we
should
put
out
a
charter
that
makes
it
clear
that
we're
open
to
all
of
those
viewpoints
and
that
when
we
try
to
get
this
thing
going,
we
want
to
make
sure
we
recruit
people
from
all
of
those.
A
Frankly,
if
we
can
bring
these
experts
between
the
interface
between
a
couple
of
Open
Source
projects
to
come
and
give
a
talk
about,
what's
involved
there,
that's
of
Interest
not
only
to
them,
but
people
who
would
use
you
know
those
combinations
of
projects,
so
the
bigger
the
tent
we
put
up
I
think
the
better.
This
group
will
be.
B
So
it
seems
like,
through
this
conversation,
we've
kind
of
all
it
seems
like
come
to
a
place
where
maybe
cncf
within
the
runtime
might
be
a
good
place
to
start
approaching,
and
then
that
comes
with
what
is
our
Charter?
Do
we
want
to
start
a
working
document
for
looking
at
a
charter
and
what
that
would
look
like
and
kind
of
yeah
sorry
go
ahead.
C
That
that's
what
I
wanted
to
propose,
maybe
start
a
Google
doc
and
start
putting
ideas
there,
all
the
things
that
that
we
talked
about
today.
You
know
where
we're
coming
from
where
we
want
to
go.
What
we
want
to
address
I,
think
that
that's,
at
least
for
me
sounds
like
a
like
a
good
plan
and
see
where,
where
is
to
lead
us.
A
Yeah
I'd
like
to
suggest
that
getting
started
with
a
doc
is
great
but
clearly
label
it
draft,
because
I
think
we're
at
a
stage
where
we're
open
to
all
ideas
still
and
we
have
an
opportunity
given
that
kubecon
Europe
is
coming
up
in
a
month.
Maybe
we
should
even
try
to
do
some
sort
of
attempt
there
with
people
who
might
be
present
physically
to
hold
further
discussions,
and
you
know
don't
don't
be
in
a
rush
to
get
the
final
document
resolved
in
the
next
few
weeks.
A
I
think
that
it
would
be
better
to
have
a
timeline
of
months
rather
than
weeks
here
and
take
advantage
of
some
of
these
opportunities,
as
the
world
resumes
normality
to
solicit
viewpoints
from
everybody
and
kind
of
the
upcoming
conferences
are
going
to
be
a
great
idea
to
do
that,
even
if
it
ends
up
being
hallway
track.
You
know
at
kubecon.
We
should
go
out
there
and
take
advantage
of
that
opportunity.
B
I
think
it's
always
good
to
go
into
those
kind
of
interactions
with
something
to
point
to
and
so
I
do
think.
A
draft
Charter
would
be
helpful
and
we
can
amend
and
extend
and
or
maybe
it's
a
not
a
chart
draft
Charter
and
instead
it's
a
draft
possibilities
for
collaborating
in
a
new
working
group
setting
but
being
able
to
point
people
to
that.
To
say
like
this
is
kind
of
where
our
ideas
are
going.
B
Would
love
you
to
like
leave
some
comments
on
this
and
so
I
think
that
does
provide
a
little
motivation
for
maybe
getting
something
together
before
that.
C
Yeah
because
it
will
be
I
used
to
exercise
for
us
as
well
just
to
put
all
our
radius
in
one
place,
and
you
know
just
work
out
what
we
as
a
group
want
to
do
right,
no
matter
where
we
land
or
or
where
we
go
just
just
collect
all
these
things
that
we
have
been
talking
about
in
a
single
place.
So
that's
it's
written
down.
A
In
terms
of
you
know,
we
it
can't
be
totally
open-ended
and
I'd
suggest
that
maybe
the
ultimate
date
would
be
like
kubecon
North
America,
where
we're
going
to
have
another
kubernetes
on
the
edge
day.
If
we
have
a
new
home,
we
could
announce
it
there
if
we
get
it
done,
and
if
people
weren't
aware
I
believe
that
date
is
slated
for
October,
so
we've
got
a
pretty
decent
Runway
to
get
things
done
and
I
could
Envision
that
once
that
draft
moved
to
kind
of
our
view
of
the
final.
A
C
Do
you
know
who
actually
is
responsible
for
organizing
kubernetes
on
edge
day
I
mean
it
started
kind
of
organically?
And
now
it's
just
a
part
of
the
process,
so
so
that
we
can.
Actually,
you
know.
A
Yeah
I'm
I'm
involved
with
it,
but
you
know
my
principal
contact
I
know
who
that
person
is
and
exactly
it's
it's
the
point
but
inside
the
cncf
itself.
I,
don't
know,
there's
that
thing
about
Conway's
law,
where
they
say
software
resembles
the
org
chart.
I
have
no
clue
what
the
org
chart
for
that
event
is
inside
the
cncf.
A
I
have
seen
other
names,
including
kind
of
the
some
of
the
key
Executives
from
the
cncf,
have
been
on
some
of
those
email,
distribution
lists,
and
some
of
them
have
been
shown
up
at
as
speakers
at
some
of
those
events
and
referenced
it
in
the
Keynotes
for
kubecon
cloud
native
con
itself.
So
certainly
it's
apparent
that
they're
well
aware
of
it,
but
who's
in
charge.
I,
don't
know
yeah,
okay,.
C
I
mean
just
one
more
thing
to
to
keep
in
mind
if,
if
that's
gonna
fall
under
one
of
the
things
that
that
the
group
will,
you
know,
will
be
responsible
for
you
know
who
are
we
communicating
with
and,
and
you
know
beyond
the
program
committee
you
know,
is
there
any
influence?
We
can
do
about
that
event?
Yeah.
A
It
just
strikes
me
that
maybe
it's
an
inversion
of
authority
to
declare
that
our
group
is
responsible
for
that,
because
if
it's
part
of
a
larger
conference
like
Cloud
native
con,
we're
not
in
charge,
so
they
could
strikes
me
that
they're,
the
ones
who
decide
who
is
useful
to
conduct
these
things,
whether
the
event
even
continues
to
go
on
Etc
and
I'm,
not
sure
that
that
belongs
in
our
Charter
other
than
a
kind
of
a
mansion.
In
an
annual
report
that
hey
we
happen
to
be
doing.
A
This
is
fine,
but
whether
we
officially
declare
ourselves
the
gardenia
that
I
don't
know
if
that's
appropriate
or
not,
but
we
can
put
it
in
the
document
and
open
it
up
for
discussion.
But
I.
Think
I
would
like
to
to
go
talk
to
some
of
the
people
at
the
cncf
to
get
their
view
on
whether
that
would
be
an
appropriate
item
to
have
listed
or
not.
C
Yeah
that
that's
exactly
what
I
want
to
do
to
to
see
like
you
know,
we
were
involved
into
that
and-
and
you
know
not
now
part
of
the
program
committee
for
for
the
last
two
years,
but
but
you
know
beyond
that,
is
there
anything
else
we
can
do
to
help
I'm,
I
I,
don't
know
yeah
so.
A
So,
regarding
transition
of
the
group
you
know,
have
we
talked
this
one
out
for
the
purposes
of
today
or
does
anybody
have
any
further
comments
or
directions
you'd
like
to
go
I.
E
Think
Kate's
I
think
Kate's
strategy
of
getting
something
down
on
paper
with
the
what
seems
to
be
the
the
the
most
the
most
likely
good
home
is.
It
sounds
great
to
me
because
when,
with
that
in
hand,
then
discussions
will
be
kind
of
like
what
do
you
think
of
this,
which
is
kind
of
what
you
say.
It's
set
case,
it's
better
to
have
something
to
show
people
and
and
not
to
say
what
do
you
think
period.
But
what
do
you
think
of
this
I
like
that
idea?
A
lot.
A
B
A
Know
I
can
what
I've
been
doing
is
creating
documents
in
my
personal
Google
Drive
account,
which
I'm
not
sure
is
the
best
thing
to
do.
I
believe
that
there
are
mechanisms
with
groups
to
have
a
a
document
owned
by
the
kubernetes
project
itself,
but
I
don't
know
what
that
is,
and
it's
just
I've
heard
sidebar
that
such
a
mechanism
exists
now.
A
Part
of
this
is
that
I'm,
technically,
not
a
chair
of
the
group,
so
maybe
credentials
for
that
were
given
to
chairs
I,
don't
mind
creating
a
doc
on
my
own
account,
but
historically
there
have
been
things
in
the
project
where
that
has
come
to
a
bad
end.
If
somebody
goes
away
and
ended
up
owning
the
dock,
it's
kind
of
kind
of
rough,
but
it.
B
C
B
A
And
then
the
typical
thing
that's
done,
I
know
is
once
you
create
it,
that
just
is
basically
who's
paying
to
host
it,
but
the
permissions
are
independent
of
that.
So
there's
a
way
that
if
you
give
added
authority
to
the
membership
list,
which
is
the
Google
Groups
for
this
group,
then
anybody
who's.
A
group
member
can
go
edit
or
comment.
So
that's
kind
of
the
right
way
to
do
it.
In
some
cases
the
kubernetes
project
they
have
other
lists,
like
just
members,
is
much
more
open.
A
Where
you
don't
have
to
be
a
member
of
this
group,
you
could
be
a
member
of
any
others,
and
maybe
that's
even
more
appropriate
to
do
here
on
this
document.
The
downside
of
making
it
world
editable
is.
The
experience,
unfortunately,
has
been
that
trolls
go
in
there
and
deface
things
and
entertain
themselves
by
putting
inappropriate
comments
in
so
I.
A
Don't
think
you
want
a
world
editable
but
yeah
if
you
can
create
it
Beyond
and
have
some
kind
of
open-ended
thing
and
ultimately,
I
think
we'll
want
to
share
it
over
on
the
cncf
side,
which
might
have
their
own
Alias
account
to
to
to
do
that,
or
maybe
the
list
of
participants
out
of
the
CNC
have
the
smaller.
We
can
add
them
one
at
a
time
just
using
their
emails.
C
It
shouldn't
be
a
problem:
I
mean
we
managed
the
much
bigger
white
paper
Dock
and
fine,
and-
and
this
is
The
Limited
in
in
time
right.
So
in
a
couple
of
months
it
should
be
resolved.
So
I'll
do
it
tomorrow,
because
I
need
to
get
off
after
the
meeting.
But
but
tomorrow,
I'll
I'll
start
the
talk
and
and
share
with
all.
C
I
wanted
to
have
a
discussion,
maybe
for
the
for
the
next
meeting,
I've
been
playing
a
lot
with
Bluetooth
mesh
lately
and
and
using
it
from
the
containers
and
using
it
from
from
the
kubernetes.
So
we
have
someone
open-ended
things
that
that
is,
you
know
going
through
our
minds,
so
so
would
like
to
pick
other
people's
people,
people's
brains,
maybe
at
the
next
meeting
of
of
you
know,
present
the
the
current
state
and
and
what
people
think
would
be
what
would
be
the
best
best
way
forward.
A
Sounds
good
just
be
cautious,
I
haven't
looked
at
when
that
drops
down,
but
sometimes
with
these
kubecons.
We
have
problems
where
the
meeting
overlaps,
the
conference
itself,
yeah.
C
B
A
Yeah
one
other
thing,
I'd
like
to
throw
out
there
when
we
do
or
if
we
do
re-host
under
cncf,
is
to
consider
meeting
times,
because
I've
kind
of
over
the
past
couple
years
come
to
the
conclusion
that
maybe
once
every
two
weeks
is
too
often,
but
other
people
could
have
different
attitudes
and
then
I
don't
live
in
Apex.
A
So
I'm
I'll
declare
my
bias
right
up
front,
but
that
that
time
is
perhaps
a
little
awkward
and
I
think
our
attendance
has
been
a
little
bit
reduced
in
that
and
the
ability
to
recruit
speakers.
A
Once
again
it
might
be
just
that
I.
My
speaker,
contacts
are
primarily
North,
America
based,
but
people
often
aren't
crazy
about
that
the
hour
of
the
Apec
time.
So
we
should
perhaps
re-evaluate
the
the
meeting
dates
and
times
when
we
re-charter.
If
we
do.
A
Other
comment
I'll
have
having
watched
other
groups,
is
you
can
say
well
we'll
just
try
this,
but
moving
a
meeting.
Time
is
awful
in
my
experience,
because
every
time
I've
ever
seen,
it
moved
people
keep
showing
up
with
the
for
the
old
one
for
sometimes
years,
and
it
I
think
it's
important
to
get
the
time
make
the
time
a
good
choice
from
the
from
day.
One.
A
Okay
last
call:
we've
got
one
minute
before
the
end
date,
but
any
any
last
comments
come
go
for
it
now.
D
Okay,
I
would
I
would
only
say
that
you
know
kilton
kilton
opened
the
door
to
maybe
more
regularly
put
kubernetes
on
the
agenda
at
our
Eclipse
meetings,
so
gladly
glad
to
discuss
this
offline
to
say
what
would
be
the
right
way
to
to
achieve
that,
but
certainly
I
think
that
was
a
good
suggestion.
A
Yeah,
if,
if
you're
looking
for
somebody
to
promise
to
show
up
and
cover
that
yeah
I,
think
I'll
I
can
only
speak
for
myself,
but
just
let
me
know,
and
I
I
usually
try
to
make
those
meetings
anyway,
but
sometimes
I
can't
of
course,
okay.
Well,
thanks
everybody
for
attending
and
in
a
couple
weeks
we'll
have
at
least
Dion's
proposal
on
the
agenda,
and
perhaps
some
other
things
too
so
see
you
then
bye.