►
From YouTube: wasmCloud: wasm Primer Discussion, wasmCloud on the Edge Demo, Community Callout! and Updates-16Feb
Description
wasmCloud is a platform for writing portable business logic that can run anywhere from the edge to the cloud, that boasts a secure-by-default, boilerplate-free developer experience with rapid feedback loop.
https://wasmcloud.com
B
Yeah,
so
I'm
not
gonna
say
it.
I
had
a
little.
I
had
some
some
buffer
time
built
in
a
little
spiel,
so
it's
cool
I'll.
Just
do
that
and
then
nothing
happened
so
today.
What
I
wanted
to
show
off
is
something
that
we
part
of
what
we
showed
at
our
dot:
50
launch
demonstration,
our
pet
clinic
microservices
demo,
but
kind
of
with
a
twist
I
wanted
to
show
it
distributed
running
on
the
edge.
B
I've
got
a
host
running
on
my
laptop,
which
I've
got
a
couple
labels
here,
but
things
like
which
I
guess
first
of
all,
before
I
get
started,
is
sharing
my
whole
monitor.
Is
that
a
fine
screen
size?
Are
you
all
able
to
see
it?
Fine
cool?
B
Let
me
know
if
you
want
me
to
zoom
in
or
anything
so
I
have
a.
I
have
a
host
running
on
my
laptop,
which
you
can
see
by
this
label.
M1
mac.
I
have
the
pet
clinic
app
label
on
it
and
for
location.
I'm
calling
this
like
on-prem
like
this
is
a
server
that
we're
running
on
on
location.
B
Now
I
have
this
connected
to
two
other
hosts
with
nats,
which
you
can
kind
of
see
the
log
streams
for
in
the
terminal
on
the
side,
but
I
have
this
one,
which
is
a
google
coral,
which
I
have
right
here,
running
lawson
cloud
on
it
and
I
have
its
location
set
as
the
edge
and
the
other
one
is
a
raspberry
pi,
which
I
have
right
here,
which
its
location
is
also
set
as
the
edge.
You
know,
these
are
just
arbitrary
host
labels,
but
you'll
see
why
that
is
important
in
a
little
bit.
B
So
what
I
wanted
to
start
with,
which
everything
is
just
running
on
prem
right
now
on
my
machine,
is
our
pet
clinic
application
and
if
you
haven't
gotten
a
chance
to
look
at
this,
it's
based
off
of
a
popular
microservices
demonstration.
It
was
a
java
spring
boot
example
and
we
kind
of
linked
to
that
in
our
documentation.
B
B
This
is
like
our
api
gateway,
so
you
make
an
http
request
to
the
pet
clinic
api,
that
actor
or
microservice
forwards
requests
over
to
the
applicable
other
microservices
like
whether
it's
meant
to
be
handled
or
meant
to
handle
customer
requests
or
or
querying
visits
or
adding
visits
or
vets,
or
things
like
that.
So
we've
kind
of
separated
this
out
into
a
couple
different,
distinct
components,
and
this
is
the
wasm
cloud
version
of
it.
B
B
So
that's
what
you
can
kind
of
see
reflected
here
in
the
washboard.
We
have
http
server
and
our
postgres
sqldb
provider
and
then
our
four
actors
or
microservices
in
this
example
and
they're
all
linked
together
and
if
we
go
one
tab
over,
I
actually
have
the
pet
clinic
application
pulled
up.
So
we
can
see
a
couple
of
different
vets
and
pet
types
that
we
can
handle.
B
I
added
myself
as
an
owner,
and
I
can
do
things
like
add
my
dog
doug
in
here
just
putting
in
a
couple
things
like
that
he's
a
dog.
This
is
a
real
working
like
microservices
example
web
that
is
powering
a
web
application.
B
So
what
I
want
to
do
now
is
show
you
what
this
looks
like
on
the
edge.
So
we
we
have
everything
on-prem.
We
want
to
move
to
keeping
maybe
our
core
services
like
http
server
and
and
our
postgres
on
on-prem
or
in
the
cloud,
but
we
want
to
move
the
computing
power
out
to
the
edge.
B
So
what
we
can
do
here
I'll
go
to
another
tab,
and
I
have
I
have
wash
connected
to
this
nas
instance
and
the
way
that
I
do,
that
is
with
awash
context,
which
I
kind
of
demoed
a
few.
I
guess
a
few
months
ago
now
it's
been
a
little
bit,
but
basically
I
have
this
host
config
this
configured
to
connect
to
nats.
B
We
can
start
a
couple
of
different.
We
can
start
more
instances
of
our
clinic
api
with
a
constraint
that
we
want
this
actor
to
run
on
the
edge
and
then
we
can
start
maybe
three
instances
of
this
actor.
B
Now,
when
you
do,
this
wash
is
going
to
automatically
auction
for
a
host
that's
running
on
the
edge
so
either
my
coral
or
my
raspberry
pi
can
respond
to
this
and
we'll
start
three
actors
on
that
host.
B
So
when
we
run
that
we
won't
see
anything
right
away
because
I'm
focused
on
the
specific
hosts,
but
if
we
go
to
poke
around
at
our
different
hosts
here,
we
can
see
that
we
have
three
instances
of
the
clinic
api
running
on
the
raspberry
pi.
Now
we
can
do
the
same
thing
for
vets,
which
is
another
microservice
that
contributes
to
this
example.
B
We
can
add
in
additional
constraints.
So
if
we,
if
we
add
a
constraint
for
device
equals,
I
think
coral
is
the
one
for
for
the
google
the
edge
tpu
yeah.
We
can
say
that
we
specifically
want
to
run
this
on
the
coral
and
on
the
edge,
and
then
this
specific
host
is
going
to
respond
to
that
and
start
scheduling
more
instances
of
bets.
B
B
We
can,
we
can
say
specifically
to
run
this
on
raspberry
pi,
so
vets
visits
customers
so
pretty
much
all
I'm
doing
here
manually.
You
know
this
is
something
that,
like
an
application
like
wadm,
the
application
deployment
manager
would
take
care
of
you
like
scheduling
these
on
specific
hosts,
I'm
just
showing
that
you
can
manually
push
workloads
onto
hosts
that
you
want.
B
B
Now,
if
we
go
back
to
our
pet
clinic
application
and
we
click
around,
it
looks
the
same
right.
But
why
is
that
special?
If
this
looks
the
same,
when
we've
started
up
additional
instances
of
these
actors,
you're
actually
seeing
the
exact
same
application?
It's
just
automatically
load
balancing
those
requests
to
our
different
devices,
and
you
can
see
that
in
action.
If
we
go
back
to
our,
if
we
go
back
to
our
terminal
here,
I
have
the.
B
I
think
this
is
the
coral
set
up
with
with
debug
logging
on
which
is
a
little
more
verbose,
but
you
can
actually
see
when
invocations
come
across
and
hit
the
coral
you'll
see.
B
You'll,
see
invocation
succeeded
here
in
the
logs,
for
when
a
invocation
succeeds
in
the
lattice
in
general,
not
necessarily
on
this
device.
But
if
you
click
around
a
little
bit,
you
can
see
that
we
receive
an
invocation
here
for
the
wasn't
cloud
or
for
the
the
vets
actor
and
this
host
or
the
the
actor
running
on
this
host
is
actually
getting
that
request.
B
So
if
we
keep
poking
around,
you
know
sometimes
we'll
see
an
invocation
actually
get
handled
by
the
coral.
Sometimes
we
see
implications
succeeded
anyways,
it's
just
a
way
to
show
that
these
requests
are
being
live,
load
balanced
across
all
instances
of
these
actors
without
any
additional
configuration
on
just
using
wasmcloud
starting
additional
actors.
That's
great!
B
Now
another
application
of
this
live
load.
Balancing
or
you
know
these
things
that
we
get
with
with
nats
and
wasmcloud
is
we
can
actually
go
back
to
you
know
our
on-prem
device
take
all
the
workload
off
of
it.
So
we
can
get
rid
of
all
of
the
instances
of
which
is
just
one
of
our
actors
or
our
microservices,
and
so
now
all
of
the
actors
are
running
on
the
edge
and
by
the
edge
I
mean
on
one
of
these
two
little
devices
right
here
right
now.
B
If
we
go
back
to
our
pet
clinic
application,
we
can
click
into
owners.
Take
a
look
at
bets,
all
the
same
things
that
we
were
doing
before,
but
now
all
the
workload
has
been
spread
onto
our
little
edge
devices
nothing's
on
none
of
the
actors.
The
business
logic
are
on-prem
anymore.
It's
all
automatically
sending
invocations
requests
to
these
little
edge
devices.
B
Now
what
makes
this
really
powerful?
We
touched
a
little
bit
on
this
in
our
dot
50
demonstration.
If
we
go
back
a
little
bit
and
we
look
at
like
what
the
spring
boot
pet
clinic
looks
like
this
is
the
this
is
the
equivalent
architecture
diagram.
It
has
all
kind
of
config
servers
and
service
discovery
and
load
balancing
on
http
requests.
All
of
that
is
kind
of
built
into
the
application,
but
with
wasmcloud.
B
These
things
in
the
in
the
red
boxes
here
are
essentially
what
you're
getting
for
free,
like
what
you're
getting
out
of
a
better
developer
experience
with
this
automatic
failover
automatic
load,
balancing
and
even
cooler.
You
know
we're
running
this
on
a
little
one:
gigabyte
single
chip
computer,
so
the
raspberry
pi,
the
the
edge
tpu.
B
I
feel
like
I'm
gonna
overuse
those
those
words
automatic
load,
balancing
and
failover,
but
what
I
really
wanted
to
show
is
this
true
this
this
demonstration,
we
have
that's
a
real
microservices
application
running
on
these
little
computers,
because
I
know
that
this
is
going
to
have
some
good
implications
for
our
machine
learning
effort
that
we've
got
going
on
right
now
running
things
on
the
tpu
and
in
general
there
really
wasn't
too
much
setup
that
went
into
this.
B
I
I
took
the
host
release
from
the
arm
docker
image
and
put
it
on
put
it
on
the
coral
in
the
raspberry,
pi
and
they're
all
configured
to
you
know
once
you
configure
it
to
talk
to
the
same
nats
instance.
They
all
talk
to
each
other,
no
problem.
So
as
a
supplement
to
this
demo,
I'll,
probably
publish
a
couple
different
commands
for
just
getting
this
to
all
connect
together.
If
anybody
else
wants
to
try
this
too,
so
I
saw
a
couple
things
in
chat.
B
I
just
stopped
sharing
so
that
it
wouldn't
all
show
up,
but
does
anybody
have
any
questions
or
want
me
to
dive
into
anything
else?
On
on
this
demo,.
B
D
B
A
I
said
bets
or
visits
brooks,
would
would
you
mind
sharing
the
sort
of
high
level
architectural
diagram
and
then
just
maybe
just
doing
a
little
bouncing
icon
over
it
kevin
yeah,
just
basically
what
kevin
dropped
into
chat,
if
you
wouldn't
mind
just
sort
of
visually
explaining
the
flow
of
the
http
request
from
your
browser
to
and
how
it
sort
of
navigates
through
the
system
here.
B
Yeah
yeah
yeah,
so
essentially
the
the
arrows
are
in
the
correct
direction.
Here,
what
happens?
Is
you
have
the
application
on
the
front
end?
That
makes
a
request
of
the
api
and
what
actually
receives
that
request
is
the
http
server
provider.
Here,
that's
the
thing:
that's
the
thing!
That's
running
on
my
machine
and
actually
able
to
get
network
requests.
B
It
forwards
that
request
to
the
clinic
api
actor,
which
has
the
watson
cloud
http
server
capability.
So
we
know
that
it's
allowed
to
get
http
requests
and
then
from
there
you
know
we
send
off
the
request
to
the
applicable.
You
know
I
keep
calling
them
microservices,
but
the
actor
that's
designed
to
do
that
type
of
query.
C
B
Pretty
much
yep,
oh
okay,
all
right!
I
missed
that
part.
Sorry
and,
and
the
reason
you
know
we
we
would
have
something
like
the
clinic
api
actor
directly,
listen
to
http
requests.
You
know
we
with
wasm
cloud.
We
would
want
that
all
to
be
federated
so
that
random
actors-
or
you
know,
potentially
untrusted
actors-
couldn't
just
receive
and
make
http
requests.
B
But
as
of
right
now,
webassembly
modules
can't
make
http
requests
or
or
listen
for
http
requests.
You
know
open
up
sockets,
so
the
provider
is
just
a
way
to
to.
You
know,
listen
to
and
and
give
that
to
the
api
cool.
Oh
yeah,
that's
a
great
question.
A
Nikki
was
that
is
that
clear
now,
where
all
the
components
were
and
what
you
saw
in
the
demo
was
everything
the
the
four
boxes
in
blue
started
running
on
brooks's
mac.
Then
he
replicated
them
to
the
raspberry
pi
and
to
the
google
coral
tpu
edge
device
shut
down
everything
running
on
his
mac
and
there
was
no
interruption
to
the
application
whatsoever.
It
was
completely
seamless,
self-forming,
self-healing,
auto-discovery
and
obliviates
the
need
for
all
of
the
sort
of
scaffolding.
You
would
normally
have
to
spin
up
in
order
to
manage
that.
C
Yeah
well,
that
part
was
clear.
The
part
raw's
falling
down
was
he
was
showing
his
web
page
and
it
seemed
to
me
the
web
page
was
reflecting
the
state
of
the
provider,
which
is
why
I
thought
that's
where
I
got
all
turned
around.
I
didn't
realize
that
he
was
basically
using
the
actor
as
an
api
router
back
into
the
provider.
C
B
When
I
yeah,
when
I
added
my
dog,
we're
persisting
all
of
our
state
in
the
postgres
database,
so
when
we
do
things
like
query,
for
you
know
who
what
dogs
does
this
individual
customer
have,
that
request
hits
the
http
server
provider
gets
forwarded
to
the
clinic
api,
which
says?
Oh,
this
is
a
request
pertaining
to
customers
and
then
that
customers
actor
actually
executes
the
sql
query
to
get
that
yeah.
C
So
so
the
picture
I
had
imagined
was:
if
you
take
that
pink
box,
that
says
server
provider
and
you
made
it
blue
and
you
moved
it
to
the
left
side.
That
was
my
my
image
right.
You
were
coming
in
into
the
provider
and
the
provider
was
going
up
into
the
after
the
actor
was
going
into
the
provider.
I
didn't
realize
that
there
was,
I
don't
want
to
say,
call
back,
but
this
I
didn't.
I
I
just
assumed
that
the
http
server
part
was
on
the
actor,
which
is
where
I
was
trying
to
get
turned
around.
B
C
A
I'm
gonna
argue
with
you
over
the
least
person
on
the
call
nikki,
but
that's
neither
here
nor
there.
Everybody
brings
different
perspective
to
this
call
and
very
succinctly:
everybody
adds
a
ton
of
value.
You
know
like
to
your
point.
A
Your
just
emerging
familiarity
with
the
with
the
space
is
an
awesome
opportunity
for
us
to
check
how
we
explain
things,
and
I
really
appreciate
your
a
desire
to
continually
ask
questions
and
help
us
to
explain
things
better,
because
it's
that's
one
of
the
biggest
challenges
that
we
have,
and
I
know
that
may
actually
be
a
good
segue
if
there
are
any
more
questions
to
what
brooks
said.
A
One
of
the
things
nikki,
I
think
you
were
working
on
is-
is
bringing
some
different
voices
to
the
table
and
trying
to
help
explain
the
material
a
different
way.
Would
you
like
to
maybe
share
a
little
bit
of
some
of
the
things
that
you've
been
working
on
and
what
you're?
What
you're
thinking.
C
Yeah,
actually,
this
is
actually
kind
of
a
good
point.
So
let
me
let
me
try
and
sort
of
give
like
the
30
second
primer.
C
So
you
know
I've
been,
as
you
guys
know,
I've
been
trying
to
learn
the
stuff
trying
to
orient
myself
to
it
and
kind
of
sort
of.
I
have
to
kind
of
the
way
my
brain
thinks
I
have
to
see
it
from
the
top
down,
because
I'm
pretty
good
with
details,
I
can
understand
details,
but
if
I
don't
understand
where
the
details
sit
in
the
overall
structure,
I
get
very
discombobulated.
C
So
I've
been
thinking
about
this
technology
kind
of
sort
of
from
that
point
of
view,
and
so
what
I
said
to
liam
last
friday
actually
was
like
hey.
You
know,
I
think
I
think
there's
some
fundamental
concepts
that
I
think,
if
you
were
to
air
them
out.
C
I
think
a
lot
of
these
discussions
about
lattice
actor
provider.
All
that
stuff
would
just
fall
into
place,
or
at
least
it
would
fall
into
place
for
me.
So
what
I've
done
is
I've
started
to
work
on
a
little
bit
of
a
video
and
if
it's
okay
with
you
guys,
I'd
like
to
show
it
to
you
guys,
I'd
like
to
do
it
twice:
okay,
the
first
time
so,
where
I
am
at
this
process,
is
I've
done
all
the
animation,
but
I
haven't
done
the
voiceover.
C
So
I'd
like
to
just
play
it
without
saying
anything,
it's
literally
two
minutes
and
then
what
I'd
like
to
do
is
play
it
again
and
kind
of
sort
of
tell
you
what
I
think
the
content
should
be
on
the
slides
and
then,
if
you
guys
could
give
me
feedback
good,
bad
and
different.
I
think
that
would
be
very
fruitful
for
me
at
least
so.
Can
I
share
my
screen.
A
C
C
Yeah,
so
I
want
to
play
it
because
I
haven't
done
the
audio,
the
audio
okay,
the
way
this
software
works
is
this
is
two
minutes,
but
this
is
many
many
hours
of
my
life,
so
let's
just
go
through
it
and
tell
me
if
the
video
visuals
make
sense
and
then
I'll
walk
through
it
again,
and
I
will
just
tell
you
what
I
think
needs
to
be
on
each
screen.
Okay,
so
it's
going
to
be
a
little
bit
of
an
exercise,
but
let's
see
how
this
goes,
you
guys
can
see
this
right.
D
My
first
thought
is
this:
is
amazing,
did
you
did
you
draw
all
these
things
yourself.
C
I
have
nothing
better
to
do
with
my
life.
No,
this
is
a
as
you
can
see.
I've
got
a
trial
version
of
this
stuff.
Okay.
Well,
so
let
me
walk
through
it
with
a
little
bit
of
narration
okay,
but
I'm
trying
to
tell
a
story,
but
this,
but
the
pictures
have
to
stand
without
the
narration.
Now,
obviously
you
know
you
know,
images
is
half,
but
narration
is
the
other
half,
but
I'm
hoping,
if
I
put
them
together,
I
get
something
good
right.
C
So
let
me
just
quickly
work
through
this
really
fast.
Just
the
high
level
points.
This
is
obviously
the
intro
slide.
You
know
nothing
special
here
so
here
on
this
slide.
What
I
want
to
talk
about
is
you
know
what
it
is
and
why
we
think
it's
exciting.
Okay,
so
this
slide
is
what
is
it,
which
is
the
as
I
narrate
it,
I'm
only
going
to
draw
one
person.
So
when
I
start
talking
about
you
know,
let's
talk
about
what
wasm
is
and
you'll
just
see
the
person
right
and
I'll
start
talking
about.
C
Well,
it's
a
standard.
Then
it
showed
the
browser
with
the
books
on
it.
The
person
walking
up
is
means
it's
a
little
process,
and
you
know
the
the
globe
with
the
play
icon
is
that
it's
ubiquitous
technology?
It
can
run
anywhere
any
place
anytime,
okay,
so
this
is
obviously
you
know
the
evolution
of
computers
right.
So
obviously
you
have
the
cpu
right,
and
the
point
I'm
trying
to
draw
out
here
is
application.
Management
continues,
continues
to
be
a
pain
in
the
rear
right.
So
let
me
finish:
let
this
one
dry
out.
C
Okay,
so
basically
you
know,
obviously
you
have
cpu
the
cpu
evolved
right.
You
went
from
vacuum
tubes
to
whatever
billions
of
transistors
you
want
it
from.
You
know,
punch
cards
to
ssds
all
that
stuff.
Interesting
enough,
I
actually
looked.
This
up.
Vmware
was
founded
in
1998.
If
I
don't
pull
correctly,
it's
just
a
point
of
information,
so
in
the
1990s
you
get
virtualization.
C
C
C
C
So
this
is
like
the
money
slide.
Okay.
This
is
where
I
like,
if,
if
you're
going
to
grab
someone's
attention,
if
I
have,
if
I
lost
them
before,
I
totally
lost
it
now.
This
is
like
the
money
slide.
So
I'm
going
to
talk
about
well,
here's
the
idea
behind
wasm
right.
It
runs
in
you
know
in
any
browser,
runtime,
it's
a
sandbox
secure.
C
C
D
D
The
other
thing
is,
I
don't
want
to.
D
I
don't
want
to
diminish
the
slide
here,
because
it's
just
amazing,
but
if
we,
if
we
use
wasm
down
at
the
bottom
and
then
on
the
same
slide,
we're
talking
about
actors,
providers
and
lattices,
it
may
give
people
the
impression
that
those
things
are
part
of
the
wasm
standard
and
not
part
of
wasm
cloud.
Okay,.
D
I
think
for
clarity.
It
probably
would
be
better
that
way,
because
people
already
think
wasm
on
its
own
does
too
much
so
yeah
having
like
wasm
and
then
why
some
cloud
on
top
and.
A
B
Oh,
no,
it's
fine!
I
just
threw
it
up
right
when
you
got
to
the
slide
and
you
took
a
pause.
I
was
gonna
say
the
same
point
of
feedback
that
kevin
was
going
to
or
that
kevin
said
just
that.
This
is
the
point
in
the
presentation
where
it
becomes
wasn't
cloud
specific,
and
so
I
I
think
that
same
feedback
of
like
wasn't
cloud
on
top
of
wazin
is
perfect.
Okay,.
C
C
You
know
green
cloud
does
kaput
and
it
just
reroutes
right,
I
think
you
know,
and
obviously
the
heterogeneous
actors
running
on
heterogeneous
devices.
C
C
This
is
a
if,
from
the
point
of
view
of
a
cso
or
a
seat
as
cio
or
cto,
this
is
profoundly
important.
D
Right
yeah,
I
was
going
to
say
that,
like
the
the
visualization
here
with
the
things
attached
to
the
mesh
is
amazing
on
its
own,
but
the
animation
of
showing
one
device
attempting
to
reach
a
provider
and
then
finding
a
fallback
path
that
one
animated
illustration
does
more
than
I
could
do
in
a
half
hour
worth
of
you
know,
powerpoint
slides.
A
Nikki,
I
think
my
I
would
like
you
to
say
that
you
know
on
your
and
you
should
actually
write
the
you
know
write
a
little
bit
of
that
around
you
know
the
automated
you
know
resilience
and
you
know
just
note
those
points
I
mean
hearing
your
words
now
make
this
slide.
You
know
one
plus
one
is
a
hundred.
You
know
you
know
it's
great
to
hear
that
it's
even
better
together,
yeah
sorry
go
ahead,
though,
back
to
the
back
of
the
wall.
C
Yeah,
so
at
this
point
I
mean,
I
think
the
rest
of
this
slide
is
pretty
obvious.
You
know
this
guy
goes
this
guy
takes
over
again.
I
didn't
put
the
pause
in
here.
I
need
to.
I
need
to
narrate
this,
and
I
put
I
need
to
time
it
right,
but
and
then
obviously
oops.
Sorry,
then
this
is
just
like
okay,
you
know
we're
still
we're
still
children.
C
We
believe
this
is
the
path
forward.
Come
join
us
right,
but
what
I'm
trying
to
get
out
of
this
meeting
is
like,
from
your
point
of
view,
does
this
visualize
and
obviously
I
need
to
fix
a
slide
with
the
wazing
cloud
and
the
contract
and
put
in
the
right
place.
Those
are
all
simple
things
to
fix,
but
from
a
visualization
point
of
view,
is
this
the
story
that
you
guys
want,
because
this
is
a
story
that
I'm
receiving
now
that
may
or
may
not
be
the
story
that
you
had.
A
Look
nikki.
I
think
I
think
the
bigger
point
when
we
met
on
this
one-on-one
that
I
wanted
to
really
encourage
was
your
point
of
view.
Here
is
what
matters
right
and
you
and
the
sort
of
what
you've,
how
you
perceive
this
elephant
of
of
our
distributed
fast,
is
really
hearing
it
played
back
in
your
own
words
is
going
to
connect
with
a
whole
different
population
of
personas
that
share
your
world
view
and
that
you're
going
to
speak
to
directly.
A
So
I
think
it's
really
not
important
what
we
think
on
the
call,
and
I
also
still
love
it
and
I
think
the
generally.
I
think
it's
great
what
I
would
suggest
as
a
next
step
is.
I
have
worked
with
this
community
now
for
a
couple
years,
and
it
was
amazing
when
we
were
preparing
like
the
keynotes
for
kubecon
eu,
and
you
know
some
like
the
big
web
assembly
stuff-
was
that
the
people
on
this
call
helped
me
write
those.
A
So
what
I
did
is,
I
think
I
shared
a
couple
of
those
docs
with
you.
You
know
take
if
you
have
12
slides
here,
that
get
drawn,
take
12,
slides
and
put
them
in
a
google
doc
and
then
start
typing
your
script
under
it
and
then
invite
folks
here,
if
you
want
to
help
really
dial
in
that
language,
so
we
get
it
really
precise
and
crisp.
I
was
blown
away.
A
Justin
gave
all
kinds
of
great
feedback
on
you
know.
Even
grammar
and
stuff,
like
that
on
some
of
my
talks
in
the
sort
of
power
of
the
community
to
bring
these
bring
these
together.
G
Yeah,
did
I
fix
my
mic?
Can
you
hear
me
cool
nicky?
The
one
thing
I
will
say
the
only
recommendation
I
have
is
you
said
at
some
point.
This
is
really
clear
to
everyone
on
this
call.
I
will
have
you
know.
I've
been
riding
this
train
for
almost
a
year
and
a
half.
Now
I
only
last
week
grasped
the
power
of
the
lattice
is
going
to
evolve.
G
So
I
would,
I
would
say,
like,
even
though
my
my
introduction
to
web
assembly
was
wasn't
cloud,
so
I
would
say
that
the
assumption
that
we
all
understand
it
greatly.
I
wouldn't
take
that
because
I
even
learned
something
just
now
from
yours
that
I'm
you
know
there
are
some
of
us
on
here
that
are
still
writing
and
it's
still
a
funky
bucking
bronco.
If
you
will
so
just
wanted
to
give
that
feedback.
C
Yeah
I
mean,
and
that
was
the
other
thing
I
wanted
to
ask
like
not
the
people
that
work
on
wasn't
cloud
directly.
I
was
just
wondering
if
anyone
else
in
the
audience
got
something
out
of
that,
that
presentation
that
they
hadn't
realized
or
hadn't
quite
gropped
or
maybe
presented
the
same
thing,
but
in
a
different
way
that
they
had
never
thought
about.
C
Okay,
I
mean,
I
guess
the
final
comment
is
again.
I
think
the
key
to
this
is.
I
really
want
this
to
be
less
than
five
minutes.
I
think
if
it's
five
minutes,
it's
too
long
with
the
animation
that
I
have
right
now,
that
was
like
a
little
over
two
minutes.
I
don't
think
I
can
narrate
it
in
the
time
that
it's
it's
there.
I
think
it's
gonna
blue
to
I'm
trying
to
get
it
to
like
three
or
four
minutes
is
my
my
target.
C
But
this
you
know
what
I'm
looking
for
is
like
anyone.
Who's
got
a
technology
background
at
some
level.
Right
can
take
three
or
four
minutes
out
of
their
out
of
their
life
and
watch
this
and
be
able
to
start
having
a
meaningful
conversation
about
what
this
is
and
isn't
right,
because
I
think
a
lot
of
times.
C
You
know
what
I
said
to
liam
last
week
is
that
you
know
whenever
you
hear
a
new
technology.
It's
all
this
mumbo-jumbo.
Everyone
says
the
same
mumbo-jumbo
and
then
you
know
you
just
hear
it.
You
become
numb
to
it
right
it's
nothing
to,
but
to
you
know,
to
not
use
big
complex
words
and
not
use
the
same.
You
know
mumbo
jumbo,
but
but
just
show
very
simply,
okay.
This
is
why
it's
different
you've
got
a
sandbox.
C
The
sandbox
is
protected
memory,
blah
blah
blah
blah,
and
you
know
very
simple,
english
right,
not
not.
Oh,
it's
a
homogeneous
array
of
memory,
not
nothing
like
that.
You
know
it's.
You
know
and
using
simple
terminology
that
you're,
you
know
you're
kind
of
sort
of
targeting,
like
maybe
a
sophomore
in
college,
in
computer
science,
or
something
like
that,
and
they
should
be
able
to
get
several
things
out
of
this.
That's
that's
the
target.
I'm
shooting
for.
A
It's
great
nikki,
I'm
excited
to
see
where,
where
it
evolves
to
over
the
next
couple
weeks,
as
you
dial
this
in
and
get
this
get
this
going
any
other
questions
for
nikki
or
feedback
on
this
and
nikki.
I
would
say
you
know
as
you.
If
you
want
to
pull
together
the
google
doc
and
get
feedback
or,
as
you
want
to
collect
more
feedback,
just
feel
free
to
engage
folks
directly
in
slack
too,
you
know
start
your
own
threads.
You
know
drop
stuff
in
general.
A
I
think
that's
a
good
way
to
get
feedback.
We
do
have
a
number
of
users
around,
I
don't
know
50
or
so
that
can't
make
the
meetings,
but
that
do
follow
up
on
youtube
and
other
places.
So
I
think
you'll
get
some
some
good
follow-up
engagement
there.
Just
a
final
call,
any
other
questions
for
nikki
on
this
one.
A
Okay,
great
job
nikki,
I'm
super
impressed
and
I
really
appreciate
all
the
hard
work
you're
putting
in
on
that.
It's
amazing
it's
not
something
that
we
called
out
and
that
we
asked
for
it's
an
area
that
you
just
saw
an
opportunity
and
you're
just
filling
it.
So
that's
amazing-
and
I
think,
brooks
you've
got
a
community
call
up
for
today.
Is
that
right.
B
B
Alrighty,
so
two
community
call-outs
this
week
just
good
first
issues
across
our
couple
repositories.
We
found
an
issue
over
the
last
week.
I
guess
it
was
exactly
in
the
last
week
six
days
ago,
where,
when
you
stop
a
capability
provider,
what
happens
in
the
wasn't
pod
host?
Is
we
send
the
provider
a
message
to
shut
down?
It's
automatically
listening
for
that
message
because
of
wasmbus
rpc.
So
when
it
gets,
the
shutdown
message
says:
okay,
like
I'll
clear
up
all
of
my
resources
and
then
close,
my
nas
connection
and
then
I'll
go
away.
B
Now
you
know
providers
are
automatically
listening
for
this
and
you
know
most
of
the
the
capability
providers
that
we've
that
have
been
written
are
like
the
first
party
ones,
but
you
know
in
the
case
where
cleaning
up
cleaning
up
resources
can
take
a
little
bit
longer
than
two
seconds.
I
think,
is
how
long
we
give
providers
to
clean
up
the
runtime
host
automatically
will
try
to
like
kill,
dash
nine,
the
capability
provider,
and
this
is
just
to
make
sure
that
we
don't
have
running
providers
that
we
don't
have
control
over.
B
But
if,
in
the
way
that
it
does,
this
is
by
you
know,
essentially
shelling
out
the
kill
binary.
If
that
binary
isn't
installed,
then
we
can't
actually
kill
the
provider
in
that
way.
So
what
we
really
want,
you
know
there's
there's
some
additional
design
works
to
do
here
to
give
better
granular
control
over
executables.
B
D
B
Restart
and
then
it'll
start
another
instance
of
the
provider,
which
is
like
a
great
feature
of
elixir.
One
of
the
reasons
why
we're
using
it
is
because
of
its
reliability,
but
here
we
don't
want
the
crash
and
burn
to
trigger
starting
another
provider.
So
this
good
first
issue
is
just
around
taking
this
command
and
making
sure
that
we
kind
of
try
catch
it
and
if
it
fails,
then
we
should
issue
issue
a
warning
log,
because
it
doesn't
necessarily
mean
that
the
provider
failed
to
stop.
B
It
could
just
be
taking
a
little
bit
long,
but
we
certainly
don't
want
the
process
to
to
fail
to
exit,
and
then
you
know
restart.
So
this
really
only
affects
a
couple
of
really
unique
scenarios.
If
you
start
a
host
on
a
machine
that
doesn't
have
the
kill
binary
installed,
this
happens,
and
that
usually
is
like,
if
you're
on,
like
a
scratch,
docker
image
or
something
really
slimmed
down,
so
so
that
led
to
us
essentially
finding
this.
B
So
that's
the
first
one
it's
in
our
otp
runtime,
so
awesome.
If
anybody
wants
to
start
start
playing
around
with
that
feel
free
to
reach
out
in
the
help
channel
and
then
also
another
thing
about
our
github
github
actions
we
want
to
you
know:
we've
been
publishing
new
versions
of
our
capability
providers
over
the
last
few
weeks.
We
kind
of
do
this
all
the
time
and
what
we
want
to
be
able
to
do
is
come
to
the
capability
providers
repo
and
see
essentially
at
a
bird's
eye
view.
B
If
I
want
to
start
the
newest
http
server,
what's
the
oci
reference
for
that,
and
this
isn't
too
difficult
to
find
right.
If
we
go
to
our
tags,
we
can
scroll
down.
I
guess,
for
example,
like
http
client,
we
should
see
we
should
be
using
the
zero
three
nine
version,
but
would
really
be
awesome.
You
know,
I
know
that
shields
dot
io
has
like
a
configurable
api
for
for
making
like
the
badges
that
we
have
on
wasmcloud,
but
at
the
very
least
I
the
the
action
that
we
used
to
create
a
github
release.
B
We
could
list
the
oci
reference
there.
So
if
anybody
has
some
experience
with
github
actions
and
wants
to
take
a
stab
at
this,
this
is
a
great
place
to
contribute
to
something:
that's
really
going
to
have
a
huge
impact
on
the
community,
because
it'll
be
so
much
easier
to
find
the
latest
version
of
these
capability
providers,
so
that
is
the
community
callout.
For
this
week,
we've
got
one
on
the
capability
provider's
side,
which
would
be
some
good,
old-fashioned,
yaml
editing.
B
If
you
love
doing
that
and
then
one
on
the
otp,
the
elixir
side,
anybody
have
any
questions
while
we're
on
the
call.
If
you've,
you
know,
heard
it
and
you're
super
interested
in
taking
one
of
them
and
if
there's
any
questions
feel
free
to
ask
them
now.
C
B
As
far
as
the
the
capability,
it's
a
kill,
dash
9
on
the
capability
provider.
So
you
know
really
the
the
biggest
bad
thing
that
could
happen
is
the
provider
is
legitimately
trying
to
clean
up
its
resources.
You
know
flush
message,
cues
finish,
writing
files
and
it
just
takes
a
little
bit
longer
than
than
two
seconds.
You
know
that
process
gets
interrupted,
but
it
does
have.
B
It
has
2.3
seconds
to
respond
and
clean
some
things
up.
That's
configurable!
You
can
make
that
longer.
So
if
you
know
that
you
have
providers
running
that
need
to
make
sure
that
they
have
time
to
flush,
then
yeah,
that
that
is
what
that
that
would
be
a
way
to
adjust
that,
because.
C
When
you
do
a
dash
nine
you're
actually
telling
the
operating
system
to
pull
the
rug
out
from
underneath
the
process
right
so
depending
upon
what
it's
doing
again,
it's
more
about
like
what
it's
doing
like.
If
you
guys
are
comfortable
that
pulling
the
rug
out
is
okay,
then
that's
great
I'm
just
but
the
dash
nine
is
you're
going
to
the
operating
system
itself
and
say
goodbye.
B
Yeah
yeah,
our
our
intent
here
is
that
you
know
this
is
the.
This
is
the
point
where
we've
told
the
provider
to
clean
up,
and
you
know
this
is
you.
This
happens
after
the
after
the
capability
provider
responds
saying:
hey,
I'm
done
with
doing
my
cleanup
or
until
a
timeout.
B
So
if
you
can
figure
that
to
be
longer,
then
you
know
it's
not
like
it'll
take
60
seconds
every
time
but,
like
you
know,
kevin
is
starting
to
kind
of
say
in
chat,
we're
looking
at
a
better
way
of
interacting
with
these
processes,
instead
of
just
spawning
with
with
erlang
like
spawning
a
port
and
then
kill
dash
nining
it
when
it's
time
to
time
to
die.
That
would
give
us
a
way
to
do
things
like
you
know,
say,
shut
down
check
in
with
the
process.
B
You
know
ask
how
it's
doing
things
like
that,
but
this
is
mainly
intended
to
prevent
capability
providers
from
just
continuing
to
run
in
the
background
after
they
they
should
have
cleaned
up.
G
Hey
brooks,
I
have
a
quick
question,
so
the
the
the
github
action
one
right,
it
can
be
done
in
a
bunch
of
ways,
but
what
I'm
thinking
is
utilizing
something
like
web
hooks
is
a
way
to
do
it,
but
it
also
offers
us
a
really
unique
chance
to
dog
food.
Wasn't
cloud:
is
there
any
interest
in
like
a
bigger
collective
to
have
like
a
running
wasm
club,
long
term
that
we
can
do
things
like
push
whip?
G
G
A
Well,
what
one
of
the
things
we're
working
on
at
cosmonic
as
a
fun
project
is
doing
wasm
wasmcloud
on
these
lcd
displays.
So
my
what
I
want
to
do
is
be
able
to
like
gift
somebody
irl
and
like
have
it
pop
up
on
their
board,
but
there's
all
kinds
of
hijinks
that
I
think
that
we
could
have
with
with
the
distributed
with
the
community
lattice.
A
So
that's
a
really
solid
idea,
jordan,
and
probably
something
we
need
to
think
about
controlling
too,
because
I
think
I
have
this
little
law
that
anything
free
will
be
abused
at
some
point.
Just
from
you
know
my
years
monitoring,
networks
and
stuff
like
that,
well
phenomenal
sidebar
discussion.
We
still
have
a
couple
other
things
to
cover
today.
If
I
could
just
do
a
couple,
quick
announcements
here,
the
first
is
just
a
call
out
about
the
let's
see
a
few
upcoming
deadlines
and
events.
A
That
is,
there's
also
a
channel
on
slack
if
you
haven't
seen
it
yet-
and
I
do
another
call
out
that
the
kubecon
eu
wasm
day
cfp
is
is
approaching
here.
So
it's
still
open
until
may
16th,
I'm
sorry.
The
event
is
on
may
16th,
but
it's
still
open
until
the
28th
and
we
would
love
to
get
some
more
talks.
A
So
any
ideas
you
have
wasn't
cloud
related
or
not,
we'd
love
to
get
you
in
consideration
for
a
speaking
slot
and
then
I'd
also
call
out
that
there's
still
plenty
of
time
to
get
registered
for
red
badgers.
We
love
wasm
event
on.
We
love
tech,
which
is
march,
the
2nd
at
6,
30
gmt,
so
around
1,
30
or
2
30.
A
Stuart
is
there
anything
that
you
need,
or
do
we
tweets?
Can
we
all
maybe
try
to
retweet
your
event?
Is
there
anything
we
can
do
to
help
support
you
in
the
event
you're
putting
on.
A
H
That
wants
to
register
is
more
than
welcome
to
come.
Taylor
and
brooks
are
doing
a
talk
very
kindly,
which
is
great
and
there's
also
talk
from
single
store
about
how
they're
extending
their
they're
bringing
their
processing
close
to
the
data,
which
is
great
yeah.
I
mean
just
come
along
virtually
or
if
you're
in
london
in
person
it'd
be
great
to
see,
but
yeah.
A
Okay,
so
quickly
just
turning
to
open
floor,
I'm
not
sure
kevin.
Are
there
any
architectural
decisions
or,
I
think
steve?
Are
you
on
today
anything
coming
up
on
the
wasm
cloud
side
that
we
would
want
to
chat
about.
D
Sure
there's
some
stuff
that
isn't
merged
yet
that
once
it
does
merge,
we'll
we'll
get
some
documentation
and
some
samples
and
so
on,
but
the
the
short
version
of
it
is
that
wasm
cloud
is
going
to
support
a
decentralized
configuration
service
so
that
when
you
start
a
wasm
cloud
host,
it
will
ask
the
config
service
for
the
configuration
that
applies
to
that
host,
and
so
that
host
will
then
get
a
list
of
actors
to
auto
start
and
capability
providers
to
auto
start,
and
it
will
also
receive
registry
credentials
so
that
it
can
then
be
told
how
to
pull
from
private
or
secure
oci
and
bundle
registries.
D
So
there's
there's
quite
a
bit
of
potential
for
being
able
to
use
that
in
a
bunch
of
different
scenarios,
but
yeah
that
should
add
some
power
and
flexibility
to
to
wasn't
cloud
deployment
scenarios.
A
Kevin
that's
super
exciting
from
from
perspective
of
what
that's
gonna
enable
us
to
be
able
to
do
as
far
as
bootstrapping
things
and
you
know
getting
applications
to
scale
better
that'll
be
incredible,
jennifer.
Anything
on
your
end.
I
know
you've
been
working
away
on
on
a
couple
of
features.
I
A
Okay,
super
all
right
great,
well
open
floor
just
around
the
around
the
community.
Anybody
have
any
topics
they
wanted
to
raise.
A
All
right:
well,
I
think
this
was
an
amazing
meeting
and
brooks
that
demo
today
was
just
phenomenal.
I
mean,
I
know
you
just
like
wing
these
sometimes,
but
they
always
just
look
so
amazing,
we're
gonna.
I
think
we
pulled
that
one
out
and
put
that
one
up
as
a
special.
Like
you
know,
individually
playable.
It
was
just
very
well
done
and
nikki
your
video
is
coming
along
awesome.
A
I
can't
wait
to
see
where
it
goes
and
it's
so
great
just
to
have
some
different
voices
in
the
community
show
up
and
share
their
perspective,
because
I
think
it
really
does
help
to
make
the
community
a
better
place.
If
there's
nothing
else,
I
wouldn't
stop
recording
and
we
can
hang
out
as
usual,
but
happy
wisdom
cloud
wednesday
folks
see
you
next
week.