►
From YouTube: KubeEdge Community Meeting 20200226 (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
A
A
B
B
A
C
A
A
A
Mean
currently,
we
have
two
two
options.
One
is
that
we
we
reply
from
raid,
the
docs
to
to
see
if
they
can
enable
the
feature
building
preview
for
the
document
pro
requests,
or
we
can
just
the
move
all
the
document
content
to
the
website,
because
website
the
naturally
fly
already
had
preview
option.
B
A
Another
thing
I
want
to
discuss
is
about
the
structure
both
of
the
documentation.
So
if
you
have
to
look
at
the
document,
frankly
is
not
quite
clear
about
the
agenda
the
getting
started.
It
has
some
introduction
about
how
to
run
cube
age
and
also
how
to
contribute
and
the
set
up,
looks
like
part
of
user
manual,
and
there
are
many
section
about
the
concepts.
A
C
A
A
D
D
D
A
A
A
D
A
A
A
A
A
A
Think
we
we
can
do
one
but
just
translate.
We
have
not
yet
made
the
decision,
so
do
you
have
any
documentation
so
I
won't
relate
that,
like
studying
with
the
tropical
Europe,
so
people
can
directly
apply
or
tutorial
sessions.
There
will
be
two
hours
to
our
talking,
so
the
content
is
just.
You
can't
go
through
the
basics
and
teach
the
attendees
to
try
your
products,
something
like
that.
Yeah.
A
E
C
C
E
E
A
A
E
So
so
those
are
for
the
project.
I
think
it's
good.
We
get.
Those
and
I
would
encourage
people
to
submit
talks
around
edge,
especially
as
you
know,
intelligence,
a
trend
of
intelligent
edge
going
on,
and
maybe
you
can
even
combine
kubu
edge
with
volcano
or
any
other
AI
related
project
running
and
call
native,
so
that
kind
of
topics,
maybe.
D
E
Cip
I
mean
I'm,
just
brainstorming,
ideas
of
the
because
you
really
have
like
introductory,
and
we
dive
right
that
talks
about
the
project
itself.
But
what
did
she
keep?
The
momentum
going
we
want
to
kind
of
you
know,
incorporate
some
of
the
interesting
applications
right.
I.
Think
one
of
the
interesting
application
for
the
edge
is
intelligent
edge.
E
D
A
E
So
I
have
an
idea,
so
if
we're
targeting,
you
know,
maybe
q2
for
the
you
know
for
coop
edge
to
enter
incubation,
so
this
could
be
before
we
think
about
before
cook,
on
China
after
cook,
on
China
I
think
before
cook
on
China
I.
Don't
know
if
we'll
be
ready,
if
we
don't
have
enough
use
cases
in
you
know
in
production,
but
if
we
are
not
ready
before
cook
on
China,
then
that
means
Coke
on
China
will
be
extremely
important
for
us
to
to
do
the
last.
You
know
push
for
momentum,
so
maybe.
C
E
A
D
E
A
A
And
just
to
go
back
to
the
CFP
ID
years,
I
think
we
can.
We
can
invite
some
after
the
users
to
talk
about
their
use
case.
E
D
E
E
C
E
I
have
set
a
meeting
with
prett
are
picked
for
the
you
know:
qu
the
edge
the
edge
the
news
foundation
as
project
and
and
then
then
come
from
C&C
project
and
Chris
and
I
check
with
you
know,
CEO
of
CNCs,
so
I
put
a
meeting
and
we're
gonna
talk
about
how
we're
gonna
position
coop
edge
in
the
edge
ofh
landscape
because
they
have
a
landscape
and
I
want
to
see.
If
we
could,
you
know,
build
more
awareness
for
club
edge
in
that
community
and
also
in
their
marketing
material,
so
we're
gonna.
C
E
A
E
A
A
E
C
A
A
D
D
A
D
A
A
A
A
D
A
F
B
A
A
A
C
F
X-Tra
they
have
a
platform
cloud
edge
chat
forum,
so
they
they
are
entering
or
they
show
their
presence
in
a
cranial
community
in
form
of
blueprints,
and
that
way
it
helped
to
raise
the
visibility
of
their
platform
project
and
their
platform
project
by
it.
So
it's
still
a
a
upstream
project.
So
all
the
coding
development
is
fun
in
their
upstream
project,
but
they
they
present
themselves
with
blueprint
to
show
how
a
end-to-end
solution.
F
A
F
Don't
know
how
much
away
how
a
crane
or
blueprint
works.
It's
it's
pretty
much
a
focusing
on
the
blueprint
itself.
A
blueprint
can
be
a
set
of
configuration
files.
It
can
be
the
real.
You
know,
code
level
configuration
files
in
XML
or
it
can
be
in
any
documentation
format.
It
can
include
the
installation
right
so
basically,
once
the
blueprint
is
accepted
and
the
verified
published.
A
F
This
is
but
the
blueprint
project
has
its
own
process,
so
it
has.
You
know,
we're
going
through
this
process
right
now,
so
it
has
the
proposal
face
and
then
eventually
you
you
will
have
to
allow
them
to
to
access
your
your
end-to-end
system,
or
you
can
look
at
their
lab.
They
may
have
a
hardware
platform
which
you
may
be
able
to
leverage.
F
So
eventually
they
have
a
verification
test
process.
It's
it's.
It's
kind
of
you
do
in
a
request,
a
response
form
factor
to
verify
your
system
can
do
certain
things
once
that
passed.
Then
then
it'll
it'll
be
beautified
and
then
for
a
crane
of
blueprints.
There
are
usually
two
releases
every
year,
the
current
time,
one
the
release
date.
One
is
May
30th
and
the
other
is
November
30th
mm-hmm.
F
Well,
these
are
the
two
things
and
then,
if
we
want
to
use
their
underlying
infrastructure,
say,
for
example,
coupage
is
actually
not
gambling
the
hardware
right,
so
we
may
be
able
to
load
into
their
hardware
platform.
If
we
will,
if
that's
the
case,
they
also
have
some
CITV
process
for
you
to
to
to
leverage.
But
my
understanding
is
that
that
is
not
necessary
in
general,
so
the
the
blueprint
has
its
own
process.
D
F
F
What
we're
doing
so,
the
key
thing
again,
the
key
thing
is
really.
This
blueprint
is
to
bring
cube
edge
into
you
know
and
a
crane.
Oh
it's
interesting
enough.
A
crane,
oh,
although
it
started
off
as
a
telco
edge
community
initiated
by
AT&T,
but
somehow
the
visibility
itself
is
getting
bigger
and
bigger
people
are
looking
for
edge
solutions,
especially
a
tech
which
may
not
be
used
in
a
typical
telco
edge
scenario.
They're
still
looking
for
looking
into
crane,
no
blueprints
and
see
whether
there
is
anything
they
can
leverage
so
so
this
is.
F
A
A
A
F
A
F
In
a
future,
some
of
the
additional
further
training
may
be
offloaded
to
even
to
the
cloud.
So
this
is
the
diagram
left
side
is
device.
This
is
edge
where
the
cooperage
will
be
in
the
picture,
and
the
cloud
side
is
the
central
piece,
because
I
think
a
cool
badge
actually
also
has
the
the
cloud
core
part.
So.
C
F
See
that's
why
I
like
the
right
now,
the
green
pieces
I
just
none,
some
of
them
with
all
under
Coolidge,
but
I
understand
that
there
are
actually
layers
of
these
because,
for
example,
is
may
not
be
cool
by
just
responsive
a
we
can.
We
can
write
that
later,
but
the
whole
idea
is
I.
Think
there's
one
thing
sometimes
it
may
be
confusing
is
because
that
the
term
up
device
here
the
device
is,
is
corresponding
to
the
coop
edge.
Not
the
coop
edge
edge
know
that
this
device
is
like
a
cell
phone.
So
it's
it's
different.
C
F
F
Now
we
actually
look
at
the
curvature
code.
I
think
there
are
some
container
edged
container
management,
I
think
at
the
edge
edge
cord
is
it's
doing
so
I
I,
don't
know
how
much
we
we
can
integrate
that
so
I.
Actually,
based
on
my
my
understanding,
I
drew
a
diagram
like
this,
so
the
right
side
is
the
cube
edge.
The
left
side
is
the
offloading
framework.
F
Okay,
and
if
you,
if
we
see
from
the
bottom,
the
bottom
like,
if
you
look
at
that,
the
device
in
the
cube
edge
will
correspond
to
the
device
on
the
the
offloading
framework
side
and
the
edge
piece
is
so
for
the
edge
we.
We
actually
rely
on
a
edge
pass
container
based
to
to
load
our
simple
applications
right.
So
there's
a
service
management
part
to
discovery.
All
these
things,
I,
don't
know
how
much
we
we
can
integrate
with
with
the
edge
core.
F
So
that
is
the
piece
we
probably
can
is
discuss
in
detail,
probably
not
today,
but
that
is
an
area
we
can
discuss
further
and
then
the
other
thing
is
for
the
ml
floating
part.
There's
there's
a
some
storage
requirement,
so
I
I
just
saw
a
cougar.
Has
a
datastore
here,
I,
don't
know
whether
that's
a
that's
a
storage
exposed
to
the
application
level
know.
F
Because
if,
if
I
draw,
if
I
make
an
analogy,
so
the
dhole
mo
unloading
service
part
is
it's
very
similar
to
the
right
side.
The
app
yeah.
F
C
F
F
Right
so
so
these
are
the
details.
We
probably
will
spend
some
some
time
to
discuss
more
details
together
and
then
also
we
have.
We
have
a
edge
cloud
communication
part
right
so
right
now,
I
just
look
at
the
edge
hub
on
clown
hub
I
think
that's
a
communication
channel.
We
can
leverage
but
I,
don't
know
whether
you
know
the
edge
hub.
What
is
what
are
the
api's
right?
It's
not
a
edge
cloud,
but
we
can
directly
access.
I
assume,
that's
also
a
internal
confusion
channel
when
you,
when
you
say.
A
A
So
so
country,
the
country,
the
the
trainer,
is
for
other
metadata
exchange
and
we
already
set
up
plan
to
support
especially
streaming
data
in
103.
Please,
the
the
next
release,
and
we
want
to
expose
that
channel
for
the
mac
metrics
and
also
for
the
data
data
exchange
between
applications
on
the
cloud
and.
F
Yeah
so
I
think
eventually,
because
you
do
have
this-
this
communication
channel
already
established
all
right,
so
you
may
packages
and
expose
it
in
a
form
of
API.
So
then
we
don't
have
to
you
know,
set
up
these
communication
channel
ourselves,
so
we
can
just
leverage
whatever
edge
core
can
offer.
In
fact,
an
answer.
That's
a
very,
very
attractive
feature
for
the
edge
applications,
so
that
part
now
I
think
for
this
is
the
overall
integration
architecture
I.
Just
true
briefly,
I
think
there
are
lots
of
details
we
can
spend
some
time.
F
I
can
I
can
actually
write
lists
of
questions,
put
agenda
or
info
for
next
meeting.
So
we
can.
We
can
have
a
better
detailed
topics
to
discuss
and
overall
for
for
the
blueprint
projects.
I
think
we
try
to
I
talk
to
in
so
we
want
to
take
it
slow.
The
purpose
of
the
first
release
is
at
least,
let's
bring
coop
edge
in
the
door
into
the
community,
and
then
we
will,
we
can
do
you
know
several
releases
and
each
release.
F
F
So
should
we
use
I,
know
ku
badge
it
doesn't
doesn't
dictate
or
it
doesn't
have
a
preference
one
or
the
other,
but
for
blueprint
project
was
the
easy
way
to
do
it.
With
this
require
some
community
support
or
or
cloud
lab
can.
Can
you
know
make
this
happen
on
their
own,
so
these
are
the
I
put
a
resource,
put
it
I,
don't
in
the
agenda.
So
that
is
just
something
we
would
probably
need
to
discuss
it
over,
make
sure
all
these
pieces
there
are
people
working
on
it.
Yeah.
A
So
just
justice
on
physics
from
the
two
page
community,
currently
developing
the
community,
have
several
options:
trying
to
learn.
Hardware:
one
is
the
wrestler
raspberry
pi,
because
that's
almost
the
easiest
get
one
and
the
many
developers
have
one
of
them,
but
the
other
limitation
that
raspberry
pi
is
not
powerful
enough.
So
some
of
people
they
just
Iran,
run
virtual
machines
in
different
region
to.
A
To
simulate
the
networking
across
cloud
and
age
and
because
the
machines
is
on
the
cloud,
so
it
can
be
powerful
yeah
for
the
other
out
there,
a
hardware
platform
like
we
a
thought.
We
also
did
some
investigation
like
hardware
like
how
we
address
I
think
we
may
need
some.
We
may
need
to
recompile
the
binary
from
source
source
club,
but
generally
it's
it's.
Okay
and
therefore
the
AI
mission,
any
framework
I
think
community
members
have
tried
the
tensorflow
mmm-hmm.
F
A
F
That
that's
true,
so
I
was
thinking
for
at
least
for
first
release.
So
we
we
can
just
get
a
standard
server,
a
typical
rack
server
and
then
make
sure
we
we
won't
have
any
resource
hardware
resource
limitation,
because
we
we
definitely
don't
want
to
spend
time
on
that
for
the
really
first
release
we
just
wanna.
You
know
the
functionality.
Is
there
right?
So
so
then
we
can
talk
about.
You
know
how
to
how
to
move
the
whole
thing
into
a
non
server.
More
customized,
embedded
hardware
environment
think
that
that
would
be
Austrian
school.
F
That
that
will
take
some
effort
on
you
know:
optimizing
the
the
ML
framework
itself.
There
can
be
a
little
challenging,
but
I
guess
we
have
time
to
fulfill
future
needs
to
do
that,
but
first
release.
If
no
disagreement,
then
we
can
shoot
for
a
standard
server.
Cloud
lab
seem
to
have
a
few
servers
available.
If
that's,
if
that's
the
case,
I
think
in
probably
can
can
help
too
coordinated
a
few.