►
From YouTube: KubeEdge Community Meeting 20200617 (Pacific Time)
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
Okay,
the
the
meeting
recording
is
on
so
hello.
Everybody
welcome
to
the
cube
age.
Community
meeting
today
is
june,
the
17th.
This
is
pacific
friendly
time.
A
Okay,
and
also
please
everyone
on
the
meeting
leave
your
name
here
in
the
attendees.
A
Section
all
right:
first,
we
can
go
through
the
release
updates
about
114.
A
So
I
want
to
highlight
it
regarding
our
release.
Cadence
three
months
is
a
release,
so
we
will,
by
default,
the
1.4
release
is
targeting
on
the
august.
The.
A
15Th,
so
counting
back
to
probably
two
weeks,
so
the
corner
phrase
would
be
july
july.
A
A
All
right,
let's
go
through
the
the
working
items
of
the
wonderful
releases,
so.
A
The
exec,
I
think,
that's
still
under
development,
no
updates
this
week,
the
matching
server
in
the
cloud.
Actually,
I
see
the
pull
request.
A
A
C
A
Right
because
upstream
have
some
issue
and
keep
refreshing
the
target
goal
on
data
target,
the
golan
version.
So
I
guess
we
still
need
to
wait
a
few
more
weeks
here.
A
The
gateways,
the
gateway
support,
gateway
support,
is
still
under
development.
No
no
particular
updates
this
week,
the
cloud
core
installation
document.
D
A
A
Yeah
yeah,
please
open
an
issue
to
track.
If
you
have
any
more
details.
B
A
A
A
Okay,
so
the
for
the
proposal,
any
refresh:
oh
no,
okay,.
C
Hello,
yes,
cass!
You
good
chinese.
C
A
A
C
A
Okay,
so
the
the
so
the
discussion
was
that
for
the
line,
the
row
13,
the
a
new
functionality
proposed
here
is
to
add
a
add,
a
command
to
the
installer
so
to
to
actually
output
the
formatted
data
from
the
the
age
core
database
sqlite
make
it
easier
for
the
developer
and
the
administrator
to
debug
when
any
situation
or
per
the
date.
B
B
A
Online
okay,
so
I
think
the
currently
the
the
feature
have
already
initial
version
of
the
proposal
and
the
implementation.
I
think
it's
still
waiting
for
more
review
comments,
so
please
everyone
join
review
and
give
your
feedback
under
the
pro
request.
A
E
A
Sure
so
for
the
debian
packages,
I
think
currently
it's
kind
of
stored.
We
have
the
pull
request
there,
but
it's
it
has.
It
hasn't
been
updated
for
for
a
couple
of
weeks,
so
we
still
want
this
feature
getting,
but
we,
I
think
we
need
more
help
from
the
developers.
So
if
anyone
interested
in
working
on
this,
please
please
reply
with
the
comments
under
the
pro
request
or
inside
this
document.
E
Yeah,
actually
I
put
one
item
we
can
follow
up
on
this
right
to
see.
What's
the
design
or
I
can't
take
the
ownership.
A
Yeah
we
can,
we
can
discuss
after
this.
We
go
went
through
this.
E
A
A
Okay,
so
the
reprovider
link
this
link
this
one,
this
one's
down
right,
okay,
so
the
list
watch
from
the
age
age
core
currently
is
still
under
planning
the
connection
management.
So
we
actually
have
a
chevy
already
have
a.
A
B
A
The
gc
for
sync
and
the
clustering-
I
think
that
one
we
know
you
have
the
ownership,
so
I
would
assume
there's
still
a
plan
under
planning.
A
Okay:
okay,
we
have
a
volunteer
now,
so
the
the
cooper
the
mean
per
key
at
the
main
purge.
Any
updates
on
this.
A
Some
no
add
some
notes
here.
A
Okay,
so
that's
all
about
the
release
development
of
this.
So
next
one
we
can
first
discuss
the
mesh
design
development.
A
A
G
G
G
Now
each
mess
is
worked
as
a
dns
server
and
do
proxy
for
requests
for
service
name
as
a
local
as
a
agent
load
button.
Now
it's
a
it's
very,
very
old
photos
for
something
such
as
the
dns
server
is
such
as
the
proxy
is
not
not
to
use
the
same
eye.
It's
only
a
limit
on
dock
on
dock,
so
we
need
a
docker
car
docker
knee
docker
docker
0
for
this
work.
We
want
to
use
this
for
all
with.
We
also
all
say
we
supported
losing
image
for
talk.
A
A
So
I
think
so
I.
C
C
B
G
Since
it's
a
now,
this
design
is
it's
owed
and
not
worked
for
for
all
cni,
and
we
do.
We
do
lot
of
lots
of
special
work,
but
not
use
the
exist.
G
A
So
I
think,
actually
we
it's
work,
it's
worth
to
clarify
the
scope
and
the
position
about
the
age.
A
A
There
are
some
offline
discussion
about
this,
so
actually
we
know
that,
for
each
mesh
is
targeting
until
thing
deal
with
the
the
actually
part,
two
part
and
the
part
to
service
communication
cases.
So
basically
there
are
kind
of
three
or
four
layers.
A
A
Name
resolution
right
basically
ds
and
then
we
have
the
cluster
ip.
So
actually
it's
a
name
to
dns
right.
A
G
A
Yeah,
basically,
the
these
are
the
four
pieces
we
need
to
take
care
so
for
each
in
the
each
computing
cases.
The
one
thing
we
need
to
to
deal
with
is
that
the
do
you
make
the
ass
work
offline
right.
A
Yeah,
so
maybe
node
level,
dns
or
or
kind
of
age
level,
dns.
A
A
Okay,
so
for
the
load
balancing
we
actually
need
to.
Basically,
we
need
to
support
l4
and
also
l7
right.
A
So
one
thing
here,
I
think
we
need
currently
the
edge
mesh
didn't
use
class
ip
but
introduced
a
another
ip
ip
wrench
kind
of
relevant
equivalent
to
class
ip.
So
maybe
we
we
need
to
go
back
to
reuse
class
type,
to
make
it
more
general
for
the
integrating
with
the
other
tools
or
services.
E
A
And
an
edge,
no
actually
so
the
so
currently
like
the
domain
name
to
the
service
name
to
class
ip,
that's
down
by
in
the
cloud,
that's
done
by
cube
dns,
so
in
the
offline
mode
in
the
disconnected
mode.
We
we
want
to
make
make
it
able
to
to
do
this
resolution
so
node
level,
dns
and
the
age
level.
A
Dns
is
very
useful
and
the
class
the
ip
to
endpoint
or
to
part
ips
is
actually
load
balancing
done
by
cube
proxy
or
the
load
balancer
or
like
if
you,
if
we
use
service
mesh,
like
you
still
that's
done
by
it's
number
envoy,
I.
A
E
So
yeah
yeah
sure
I
think
I
I
may
have
a
misunderstanding.
I
think
the
last
time
I
changed.
I
I
use
external
network
provider
totally
skip
the
cooper
proxy
everything
we
don't
change
the
ap
table,
but
yeah.
We
can
figure
that
out.
A
Yeah,
that's
possible,
you
can
you
can
when
you're
creating
the
service
you
can.
You
can
indicate
the
load
balancer,
so
I
I
think
here
so
each
mesh
today
there
are
a
there
is
a
built-in
implementation
to
to
do
this.
So
the
current
implementation,
actually
they
are
resolving
the
service
name
to
a
another
ip
range
and
then
use
that
ip
range
to
do
the
to
to
do
the
load
balancing.
A
So
what
I'm
trying
to
highlight
here
is
that
actually
we
can,
we
can
just
make
it
resolve
it
to
a
to
the
class
type
and
we
implement
the
equivalent
load
balancing
from
class
ib
to
a
real
endpoint.
There
should
be
okay.
E
Yep,
I
think
I
think
we
don't
need
to
go
to
detail,
but
our
goal
is
to
clarify
the
scope
and
to
recruit
people
to
draw
the
effort
to
to
make
it
happen.
Yeah
I'm
happy
to
take
ownership
of
this
and
I
we
can
have
because
we
only
have
about
30
minutes
left.
I
have
other
topics,
so
let's
clarify
the
scope
and
to
see
who
want
to
set
up
and
to
do
the.
A
Yeah,
so
for
the
headline
service.
Currently
I
don't
have
much
idea.
If
we
use
a
class
ip,
maybe
we
can
also
allocate
one
cluster
ip
for
each
handler
service,
then
it
should
be.
The
implementation
should
be
much
easier
for
the
part
to
partner
network.
I
think
the
the
major
consideration
here
is
that,
especially
in
the
cases
that
both
cloud
nodes
and
edge
nodes
are
managed
in
the
same
cluster,
we
want
to
make
each
mesh
able
to
coordinate
with
actually
any
third
party
scene
eye.
A
The
the
background
is
that
a
user
may
want
to
start
with
a
managed
or
a
a
commercial
kubernetes
distribution
from
a
vendor,
so
the
in
that
case
the
ci.
There
are
already
some
cni's
plugin
stored
in
a
traditional
cluster,
so
curvage
itself
is
kind
of
adding
plugins
to
a
to
a
cluster.
So
I
think
the
best
idea
is
we
can
collaborate
with
the
existing
c9
instead
of
replacing
it.
A
So
something
to
highlight
here
is
that
inside
one
subnet,
we
can
still
rely
on
the
ci
plugin
to
do
the
part
communication.
A
A
Gateway
can
be
one
implementation
when
doing
a
cross-subnet,
but
also
some
some
other
technology,
like
p2p,
eternal
or
or
even
like
sd1,
I
think,
are,
are
also
options
that
one
is
kind
of
relevant
to
the
underlying.
A
A
Okay,
so
that's
just
some
basic
ideas
and
we
think
I
think
we
need
to
define
the
the
development
plan
for
the
follow-up
view
releases
and
we
can
define
different
maturity,
phrases.
A
E
Yeah,
let's
make
a
goal
say
by
in
two
weeks:
we
should
have
a
finalize
the
design
and
we
should,
in
order
to
catch
up
the
coffee
right.
If
we
finish.
A
All
right,
so,
let's
move
to
the
next,
the
cncf
incubation
preparation.
So
let
me
let
me
find
out.
A
Here,
actually,
I
have
created
the
the
pull
request
to
the
cncf
toc
ripple,
applying
for
kewage
to
moving
to
the
incubation
stage.
A
And
also
another
thing
we
need
to
spend
more
effort
is,
oh
sorry,
what's
this.
A
A
A
Okay,
so
another
thing
currently
we
need
to
spend
more
time
on
is
the
sincere
duty
leaders
document.
A
So
this
one
is
kind
of
the
report
about
the
the,
including
the
details
about
the
project
growth
over
the
already
last
year
and
also
it
it
need
to
include
some
details
about
the
real
world
and
user
use
cases,
and
also
some
of
the
other
consecutive
considerations
like
project
and
code
code
quality,
sort
of
things.
A
A
We
yeah
we,
we
we
need
to
update
the
like.
We
have
the
sync
controller,
it's
not
included
in
this.
A
E
So
I
think
we
should
do
is
a
label
which
one
is
to
do
so
then
we
can
just
mark
down
which
one
is
not
finished,
so
we
can
finish
this.
One
is
the
update
to
the
update.
A
Okay
yeah
this
one
to
be,
I
can
add
more
details
about
this.
D
A
E
E
A
That's
okay!
I
I
don't
think
it's
there's
any
fixed
requirement
about
this.
A
Yeah
and
also
the
versioning
scheme,
I
think
it's
okay
and
we
can
also.
A
Yeah,
we
don't
need
at
least
the
releases
here.
I
think
ip
policy
ip
policy
with
we
need
to
update.
A
A
Yeah,
so
it's
kind
of
mixed,
because
we
here
there's
another
question
about
the
cdc.
E
Okay,
okay
yeah
this
this
one,
we
should
add
the
to
try
to
say
our
location
yeah,
let's
okay.
As
long
as
we.
A
A
The
the
runtime
members
chairs
ricardo
claus-
I
think,
okay.
E
Yeah,
let's
ping,
ricardo
and
quinn
after
we
finish.
A
E
A
Okay,
who
had
this
where
this.
A
Yeah
actually
so
regard
yes,
jen.
Please
go
ahead.
H
Yeah
the
the
reason.
A
H
Two
additional
channels-
one
is
a
cranial
specific,
because
we
have
edge-related
cranial
projects.
We
might
start
to
to
prepare
for
the
release
validation
process,
so
that
probably
needs
some
coordination
that
that
that
a
criminal
channel
is
for
that
purpose
edge
ai.
H
I
I
thought
it's
it's
better,
just
to
separate
separate
it
from
the
cranial
channel,
because
it
it
it
it's
not
limited
to
the
amount
of
floating
we're
doing
in
ukraine.
It's
more!
It's
more
general.
A
H
It's
my
proposal.
You
guys,
discuss
and
see
whether
that's
it's
working.
A
Yeah
actually
so,
regarding
the
agi
ai,
I
think
maybe
we
can
also
try
to
start
set
up
a
sig
or
working
group
for
this,
because
there's
kind
of
a
big
scope
and
we
definitely
want
to
identify
the
relevant
requirements
to
enhance
cubah
so
to
format
a
sig
or
working
group.
We
also
need
to
have
few
funding
members
funding,
sig
members
funding
working
group
members
about
this,
and
and
also
we
need
to
clarify
the
folk
areas
of
focus
and
the
scope
sort
of
things
for
a
cranial.
A
E
A
separate,
totally
separate
project
in
in
slack.
A
Yeah,
I'm
thinking
about
whether
we
can
so
one
idea
is:
we
can
create
something
like
cross-community
works,
so
acrino
is
one
big
one
and
also
in
the
future.
Maybe
we
will
add
more.
H
Yeah,
I
think
either
way
is
okay.
Maybe
we
start
small
just
to
add
two
channels
to
two
for
discussion:
purpose.
Yeah
and
later
you
know,
if
we
need
to
promote
any
of
these
to
a
bigger,
you
know,
scope,
then
we
can.
We
can
do
that
so
right
now.
I
think
it's
it's
there's
a
almost
a
immediate
need
for
a
more
you
know,
more
frequent
communication
among
people
working
on
this.
A
H
Just
like
what
you
did
as
a
general
or
a
sigma
on
those
just
add
a
channel
yeah,
so
that's
at
least
that's
one
place
we
can
discuss
and
in
the
future.
If
that
need
to
be
promoted
to
a
bigger
slack
group,
and
we
can
do
that
later.
H
But
I
I
do
think
it
takes
the
slack
owner
to
add
the
channel,
though,
because
today
I
tried
to
see
whether
I
as
a
as
a
you,
know,
user,
I
I
don't
think
I
can.
I
think
it
does
take
the
the
group
owner.
A
Okay,
I
will
double
check
so
just
to
clear
what
well
so
well,
what
kind
of
topics
are
relevant
for
using
the
acronym
channel
so
see
guys
see
ai,
I
mean
it's
clear:
we
can
put
all
the
ai
relevant
topics
there
so.
H
Yeah,
it's
it's
for
it's
for
the
cranial
project,
to
release
planning
and
also
validation.
So
it's
more
or
less
anything
related
to
the
ukrainian
community
itself
say,
for
example,
it
may
not
be
a
particular
technology
like
ai,
but
something
like
if
we
have
to
go
through
a
certain
process
of
a
cranial
in
order
to
release
the
the
blueprint
project.
So
those
kind
of
thing
will
be
discussed
there.
H
So
so
that's
why
I
think
that
the
name
is
a
cradle.
It's
it's
a
little
bit
similar
to
what
you
guys
used
to
describe
the
incubation,
but
that's
the
cncf
right.
So
that's
that's!
That's!
That's
the
cncf
specifically,
because
the
whole
cool
badge
is
cncf,
so
you
don't
need
a
big
channel
to
discuss
to
discuss
that.
But
cranial
is
a
smaller
scope.
H
So
in
order
to
concentrate
the
the
communication,
so
I
think
a
dedicated
channel
will
be
easier.
A
Yeah,
so
I
I'm
trying
to
understand
so
like
in
the
arcadino.
There
are
also
the
separate
setup
like
the
individual
room,
zoom
link.
So
I'm
trying
to
understand
the
difference
of
the
snack
channel
under
cubase
or
a
standalone
qhbp.
H
A
Or
something
product
under
the
acronym
workspace,
so
what's
the
difference.
H
Okay,
so
here's
what
I
thought
the
the
items
discussed
here
usually
it
would
require
more
quebec
community
to
be
aware
of
or
to
support
or
kubeji
specific.
The
ukrainian
blueprint
project
itself
has
other
aspects
like
say.
For
example,
it
may
involve
the
offloading
pieces.
Those
can
be
discussed
in
the
cranial
side
and
also
some
cranial.
H
You
know
solution
discussion
that
might
be
discussed
in
a
cranial
bp
project
here
is
this
channel
will
be
used
more
anything
from
the
blueprint
we
need
to
make
it
aware
to
the
to
the
cool
badge
community
or
need
the
community
support.
So
things
like
like
this,
so
it's
dedicated
to
the
cool
badge
audience.
H
Because
the
chances
are
not
all
the
kubernetes
community
member,
they
will
attend
a
cranial
blueprint
or
project
weekly
meetings
or
go
to
a
crano
website
to
check
to
constantly
check
status
or-
or
you
know,
some
progress
and
that's
actually
not
not
needed,
not
required.
This
channel
is
is
served
as
more
as
a
a
notification
or
a
bridge
to
highlight
the
the
to
highlight
the
the
relevant
event
or
or
anything
related
to
kuwait,
specific
to
kuwait,
from
the
cranial
blueprint.
E
E
A
Okay,
next,
the
it
has
the
ability
of
key
enemy
like
if
hctl.
A
Okay,
okay,
so
this
is
just
the
placeholder
item
here
to
track
the
analysis
issue
to
do
the
the
more
detailed
proposal
will
be
done
in
the
next
few
weeks.
Next,
two
weeks,
okay,.
E
Then
csf
webinar,
so,
as
we
discussed
last,
I
think
a
couple
times
ago
we
said
we
should
have
won
after
1.3
release
and
yeah.
We
should
prepare
one
and
I
think
tina
gave
us
a
very
good
suggestion.
We
should,
if
we
could,
we
can
bring
some
and
user
to
demo
their
real
case.
That'd
be
great.
A
E
A
Yes,
and
also,
I
was
told
by
the
foundation
guys
that
the
webinar
channel
is
busy,
so
we
need
to
apply
as
early
as
possible.
A
Yeah,
I
actually
got
the
email
just
not
yet
right.
There.
A
F
A
B
A
Oh
yeah,
so
actually
one
thing
we
not
yet
did
is
that
they
send
a
a
email
to
the
to
the
mailing
list
to
call
for
adding
adopters
right.
A
And
also
we
can
send
a
send
message
to
the
other
channels,
like
wechat
slack.
A
A
All
right,
so
I
think
we
can
start
start
the
call
the
meeting
the
time
is
almost
up.
A
Okay,
so
the
next
meeting
will
be
june,
the
24th,
it's
the
europe,
your
friendly
meeting,
dexter
time.
So
thank
you.
Everybody
for
attending
the
meeting,
take
care
and
see
you
until
next
time.