►
From YouTube: Kubernetes WG IoT Edge 20200506
Description
May 6 2020 meeting of the Kubernetes IoT Edge Working Group - discussion of upcoming conferences, Eclipse IoT Cloud2Edge
A
Yeah
regarding
the
online
conferences
last
week
was
the
first
red
hat
summit,
which
went
virtual,
so
I
didn't
look
too
much
after
it
didn't
have
time
to
to
to
participate.
I
participated
in
a
couple
of
like
a
community
chats
but
yeah.
It
wasn't
the
same.
Experience
like
chatting
with
people
at
the
booth.
B
A
B
A
To
me
also,
the
physical
events
has
that
that
particular
value.
So
that's
my
block
time
couple
of
days
to
just
go
and
watch
everything
you
didn't
get
to
watch
over
the
last
couple
of
months
and
talk
to
people.
Otherwise
you
will
watch
it
with
one
eye
and
try
to
do
something
with
you
with
another
eye
or
you
know.
B
B
B
B
D
A
B
Hey
moritz
and
cindy,
I
think
you
missed
the
last
one,
but
we
were
talking
about
potentially
trying
to
do
the
equivalent
of
what
we've
done
at
the
past.
Physicals
was
to
have
a
face-to-face
lunch,
and
we
were
thinking
we'd
like
to
simulate
that
with
the
birds
of
a
feather,
even
though
it
would
be
on
zoom,
and
even
when
we
had
those
lunches
they
weren't
officially
in
the
agenda.
B
But
I
don't
think
this
would
need
to
be
either
and
given
that
they're
virtual,
we
could
even
try
to
divide
up
users
into
use
cases.
Perhaps
where
I
think
the
industrial
iot,
the
telco
edge
and
maybe
retail
edge,
are
all
significant
groups
of
people
that
have
fairly
different
interests
and
rather
than
have
one
birds
of
a
feather,
it
might
be
worth
having.
C
D
Yeah,
I
just
realized,
I'm
muted,
so
I
I
know
like
for
zoom,
you
can
split
into
sub
groups
right.
I
I
really
like
the
idea
of
like
a
virtual
face-to-face
chat.
Potentially
we
can
like
a
meet
set
it
up
and
then,
if
we
see
more
people
and
then
we
can
vote
it
and
divide
it
into
even
further
groups
like
industrial
iot,
like
retail
telecom
or
even
connected
cars,.
B
And
I
think,
rather
than
divide
one
meeting
into
subgroups,
maybe
we
could
just
have
different
meetings
because
there
might
be
some
people
like,
even
though
he's
not
with
us
today.
Kilton
would
potentially
be
interested
in
multiple
groups
right,
so
okay,
it
might
be
worth
doing
that
and
I'm
thinking,
maybe
this
is
potentially
unlike
a
physical
going
to
grab
people
from
all
over
the
world.
B
Our
default
should
probably
be
since
this
is
a
replacement
for
europe
to
put
it
in
a
european
friendly
time
zone,
but
maybe
before
we
have
any
unofficial
face-to-face,
we
could
just
have
one
of
those
polls
in
slack
to
help
people
you
know,
have
the
people
actually
pick
a
time,
because
I
think
that
would
help
get
attendance
too.
If,
if
the
people
interested
in
that
use
case,
if
we
let
them
choose
their
own
time,
I
think
there's
a
a
better
chance
of
turnout.
A
B
A
So,
as
I
said
like
that,
that
meeting
that
we
had
for
north
america
an
hour
earlier
hour
later
those
three
three
hours,
I
think
and
that's
all
span
of
like
a
4
p.m,
to
6
p.m,
so
it's
reasonable
even
for
for
the
conference
hours
right
and-
and
it's
not
too
bad
for
for
us
folks
as
well.
Usually
I
mean
it's
it's
a
little
bit
early
in
the
morning,
but
it's
not
too
crazy.
C
B
B
B
C
B
B
B
B
Sort
of
let's,
let's
listen
as
much
or
more
than
we
talk,
that's
kind
of
when
you
think
of
it.
When
we've
had
our
lunches,
that's
kind
of
what
really
went
on
right.
We,
we
didn't,
have
any
presentation
at
all.
Really,
although
maybe
some
people
talked
about
a
certain
subject
spontaneously,
but
it
wasn't
a
deck
format:
yeah
yeah,
it
was
free
form.
B
If,
if
every
sig
and
working
group
and
user
group
went
there
and
asked
for
six
sessions
in
the
calendar,
the
calendar
is
just
going
to
be
so
overwhelming
that
it's
going
to
be
a
problem
and
they're
not
really
going
to
want
to
host
the
zooms
in
whatever
the
official
I
assume,
they're
doing
it
in
zoom.
I
haven't
actually
heard
whether
that
is
the
technology
they're
going
to
choose
to
use
either.
B
B
B
Yeah,
if
we
had
I'm
not
expecting
us
to
have
a
thousand
attendees,
I
mean
I
think,
kind
of
even
the
free
levels
of
you
know.
I
I
have
a
zoom
pro
from
work,
but
I
think
we
all
we
can
get
enough
accounts
or
just
use
the
one
for
this
group
that
we
should
be
fine
with
the
level
of
the
attendee
count
that
I'd
expect
to.
A
A
A
In
the
chat,
so
there's
a
plow
to
edge
iot
package
done
at
the
eclipse,
iot,
which
basically
tries
to
provide
like
a
usable,
get
different
pieces
of
of
and
different
projects
from
the
eclipse
society
community
and
provide
them
as
as
a
as
a
as
a
usable
package
package
on
their
own.
So
this
first
cloud
to
edge
package
is
basically
a
kubernetes
packaged
to
install
the
the
you
know:
connectivity
gateway
device
to
win
and
and
the
device
management
in
into
a
single
single
package.
A
A
It's
a
pure
iot
use
case
no
no
edge
at
this
point,
but
that
that
should
be
the
next
one.
So
we
already
started
working
working
with
kilt
on
two
to
make
the
the
next
one,
which
will
be
more
edge,
oriented
and
also
try
to
provide
more
packages
that
are
basically
iot
applications
in
the
background
consuming
these
data
and-
and
you
know
getting
it,
the
whole
point
of
these
packages
is
to
provide
like
a
easy,
getting
start
to
experience
for
the
people.
A
With
a
with
a
single
command,
so
that's
that's.
That's
the
whole
point
so
yeah.
The
idea
to
come
is
to
have
more
something
to
do
with
io
folk
and
and
have
a
package
to
to
to
demonstrate
more
edge-like
deployment,
also
to
try
to
figure
out
what
we
can
do
for
like
a
kind
of
like
a
test
bed
or
something
so
the
the
real
application
that
that
will
basically
simulate
and
consume
data
and
then
show
people
how
to
actually.
You
know,
write
those
applications
and
reuse,
those
back-end
components
so
to
say.
A
It's
and
yeah
I'm
one
of
one
of
the
committers
there,
but
yeah
people
leading
it
are
trying
to
keep
it.
Keep
it
recent
with
all
the
recent
releases
and
yeah
yeah.
That
sounds
like
a
good
thing
to
talk
about
to
me.
Yeah
and
it
could
be,
it
could
be
like
basically
industrial
iot
use
case
because
it
it's
not
set
to
to
be
used
only
in
that
particular
vertical.
But
it's
it's
at
the
moment
mostly
done
it's
mostly
people
involving
into
industriality.
A
So
yeah,
I
just
wanted
to
wanted
to
put
it
out.
So
when
we
started
talking
about
this,
this
was
come.
To
my
mind.
This
is
one
of
the
topics
that
we
can
have
at
least
for
one
of
these
sessions.
If
not
for
those
sessions,
we
can
organize
always
a
presentation
in
the
regular
working
group
meeting,
but.
D
Development
deployment
operations,
so
I
think
last
time
hai
she
presented
about
home
or
dapper.
Potentially
we
could
have
a
targeted
session
around
like
targeted
for
developers
and
then
a
lot
of
the
survey.
Questions
were
for
that
like
to
understand
from
a
tooling
programming
model
or
app
model
perspective.
What
are
the
needs
from
people's
perspective?.
D
D
D
B
C
C
On
the
other
hand,
I
could
easily
do
talk
on
our
research
project
what
we
are
doing
there.
So
basically
we
are.
We
are
setting
up
a
qr
nethers
cluster,
for
multiple
research
institutes
for
manufacturing
and
also
try
out
microservice
style
architectures
for
controlling
manufacturing
machines,
so
that
could
probably
easily
fill
30
minutes
one
hour
or
whatever
so
yeah.
A
C
A
B
B
Too
I
just
realized
I
flipped
off
my
camera.
I
figured
out
why
I
couldn't
find
all
those
open
source
summit
links.
I
I
in
addition
to
my
standard
laptop,
I
use
a
chromebook
and
I
had
opened
all
those
links
there.
So
I
was
looking
through
my
browser
saying:
where
did
they
go
and
found
them
on
my
other.
C
I
think
the
problem
with
machine
learning
on
kubernetes
is
that
basically,
the
application
isn't
any
different.
You
know
yeah
it's
just
like
like
some
kind
of
binary
or
some
kind
of
python
script,
you'll
deploy
and
it
could
be
anything
you
know
so
might
be
quite
hard
to
to
distinguish
that
because
you
can't
really
see
from
the
outside
if
it's
run
on
kubernetes
or
not
yeah,.
B
C
I
just
read
an
art
article,
like
I
don't
know,
two
or
three
weeks
ago
from
tesla
that
they
are
having
this
crazy
machine
learning
framework
where
you
can.
Basically,
let's
say
you
have
like,
like
a
data
set
where
you
want
to
like
detect,
I
don't
know
bicycles
or
bicycle
drivers,
and
they
have
like
this
framework
that
you,
you
can
run
your
your
tests
or
your
your
algorithms
and
when
you
see
something
like
a
like
a
car
with
a
bicycle
on
the
back,
which
of
course
isn't
really
a
bicycle
driver.
C
But
still
your
your
network
is
going
to
detect
it.
You
can
like
select
this
kind
of
picture
and
then
you
kind
of
like
select
multiple
of
them,
and
then
you
create
a
new
network
which
is
basically
detecting
these
kind
of
images
in
a
live
video
stream,
and
then
this
one
gets
downloaded
to
the
cars
and
the
start
with
the
cars
start
basically
pulling
similar
pictures
during
during
drive
or
doing
driving
and
kind
of
collect
this
new
data
set
for
for
the
for
the
user
and
then
sends
it
off
to
the
cloud.
C
B
C
Yeah,
but
what
I
found
really
cool
about
this
idea
is
that
they
have
like
this
live
camera
running
on
a
on
a
car,
and
then
you
kind
of
just
upload
the
model
on
the
go
and
start
picking
your
your
frames.
Depending
on
what
your
your
model
on
the
car
says,
it
could
be
quite
interesting
to
see
how
they
actually
deploy
these
models
if
they
have
something
custom
or
if
they
have
kubernetes
edge
running
on.
B
B
D
A
D
A
D
A
B
B
D
No,
I
I
I
know
like
in
my
org.
There
are
few
people,
because
this
is
a
major
microsoft,
organized
group.
Yes,
yeah,
we
are
a
big
contributor
there,
I'm
pushing
for
the
industrial
4.0
standard
yeah,
I
think
for
the
for
the
industrial
iot
subject
when
they
are
meeting,
would
like
to
hear
like
how
people
are
thinking.
A
A
B
A
B
D
Yeah,
I
I
think
down
the
similar
topic.
It
will
be
good,
like
for
our
group,
to
write
a
paper
or
something
summarize
like,
what's
going
on
in
the
open
source
community
around
all
the
edge
iot
area
like
I
I
think.
Originally,
we
have
the
white
paper
we
list
out.
All
the
scenario
and
potential
architecture
now
like
there
are
a
lot
of
open
source
projects
occurring
like
exist,
are
started
in
multiple
places.
We've
heard
about
them,
I'm
thinking
we
could
like
start
writing
another
one.
B
Yeah,
I
I
think
that
my
perspective,
I'm
ready
to
hear
arguments
with
this,
but
I
see
a
lot
of
those
that
are
sort
of,
in
my
mind,
too,
vendor
dominated
where
I
think
that
the
world
could
use
more
of
those
things
done
by
users.
But
then
the
users
often
don't
have
enough
interest
from
their
employer
to
pay
them
to
spend
a
lot
of
time
on
it.
D
B
B
B
Want
somebody
to
have
spent
the
time
to
actually
do
kind
of
a
a
bake
off
where
you
try
them
all,
maybe
even
live,
and
you
know
give
the
audience
a
kind
of
a
first-hand
view
of
what
the
experience
is
like
when
you
shop
and
compare
against
these
things
that
I
haven't
seen.
But
it
would
take
quite
a
bit
of
effort
to
do
it.
D
B
C
A
B
E
B
D
B
You
know
the
other
way
to
do
it,
where
you
wouldn't
need
people
getting
permission
to
participate,
but
somehow
you'd
have
to
fund
it
or
recruit
these
people.
If
we
could
get
a
bunch
of
college
students
like
interns
to
go,
take
this
on
where
you
just
divide
them
up,
and
you
give
them
these
and
give
them
like
a
mission
like
a
hackathon,
but
they
get
assigned
what
technology
they
have
to
attempt
to
use
and
then
come
back
and
see
what
the
results
are.
Now
it
could
be
just
by
random
chance.