►
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
A
B
C
A
A
B
A
A
Well,
thankfully,
it's
code,
that's
behind
there
and
not
something
embarrassing
all
right.
So
if
you,
if
you
saw
the
last
time,
I
did
a
demo
for
whatem,
I
basically
just
did
a
demo
where
I
showed
putting
the
model
data
and
querying
the
model
data
and
getting
a
list
of
the
models
and
so
on,
and
I
made
a
little
bit
more
progress
than
that.
A
A
A
A
D
A
A
A
A
Anyway,
that's
not
really
all
that
important
I
get,
but
the
the
important
business
is
that
when
I
do
annapolis,
you
can
now
see
a
pet
clinic
if
it's
idle.
That
means
that
since
the
last
heartbeat,
it
has
determined
that
there's
no
actions
necessary,
obviously
there's
a
mountain
of
room
for
improvement
here,
and
this
is
just
sort
of
the
first
implementation
to
get
things
out
the
door.
The
goal
is
to
keep
iterating
from
here,
but
yeah.
I
can
now
use
wadam
to
deploy
pet
clinic,
which
is
pretty
huge.
D
B
Yeah,
I
have
a
couple
of
different
things,
but
I'll
start
with
maybe
the
one
that
I'm
most
curious
about.
Can
you?
I
know
that
when
you
did
wash
app
list,
the
pet
clinic
app
was
already
there.
What
is
that?
What
does
that
look
like
like?
What
did
you
have
to
do
to
get
that
pet
clinic
app
there
and
then
what
is
yeah?
What
does
that
actual
app
look
like.
D
A
Gonna
start
this
actor,
but
you'll
see
I've
got
a
link
definition
here
the
target
is
postgres,
which
is
the
name
of
the
provider
as
it
exists
in
the
file.
So
if
you're
used
to
manually
configuring
link
defs
with
wash,
you
know
that
you
have
to
go
and
look
up
the
public
key
of
the
capability
provider
in
the
public
key
of
the
actor,
and
then
you
have
to
figure
out
the
contract
and
then
inevitably
the
first
time
you
try
and
run
wash
you've
got
the
actor
and
the
provider
ids
in
the
wrong
order.
D
A
A
A
I
might
say
that
I
want.
You
know
four
replicas
and
if
I
do,
four
replicas
in
spread
equals
pet
click,
then
it
will
try
and
find
four
places.
To
put
it,
there's
only
one
host
that
one
host
will
get
four
of
them.
If
there
are
four
hosts,
those
four
hosts
will
get
an
even
spread
of
that
actor.
Same
thing
works
for
providers
where
you
know.
If
I
want
the
http
server
running
on
each
of
my
hosts,
I
can
do
that.
A
If
I
want
the
you
know
the
postgres
provider
running
on
each
of
my
hosts,
I
do
that
the
only
real
difference
between
the
the
way
the
spreadscaler
works
on
providers
versus
actors
is,
you
can
only
have
one
instance
of
a
provider
link
name
combo
on
any
ones
in
cloud
host,
so
it
takes
that
into
account.
So
if
I
tried
to
run
an
auction
and
there
were
insufficient
resources
when
I
would
then
right
when
running
the
the
deployment,
I
would
see
errors
come
out
where.
A
A
That
would
just
go
into
the
model
store
and
then
I
could
choose
to
deploy
the
other
one-
and
I
may
have
mentioned
this
in
the
last
one.
But
if
I
put
a
new
version
in
the
model
store
and
then
I
deploy
the
new
version,
all
the
the
previously
deployed
version
becomes
undeployed
and
then,
rather
than
taking
down
all
the
resources
waiting
for
it
to
calm
down
and
then
bringing
back
up
the
resources
for
the
new
one.
A
We
just
compare
the
new
one's
desired
state
against
the
current
state
of
the
lattice
and
only
reconciled
on
those
differences.
So
if
version
point
two
wanted
three
instances
of
an
actor
and
point
one
wanted
one
instance
of
an
actor
then
deploying
that
new
deployment
instance.
We
just
start
two
new
actors
and
that's
it.
B
A
A
A
A
B
E
A
E
A
Oem
is
open
application
model
and
all
it
really
is
is
a
very
loose
structure
for
defining
an
application
in
yaml
microsoft.
I
believe
it
was
originally
the
the
it's
an
open
standard.
It
was
spearheaded
by
microsoft
and
it
kind
of
came
from
the
fact
that
kubernetes
had
so
many
different
yaml
files
with
so
many
different
formats.
People
wanted
a
a
little
bit
more
standard
way
of
being
able
to
parse
and
generate
those
files.
So
it
doesn't.
A
A
If
we
didn't
have
a
way
of
uniquely
labeling
the
stuff
that
belongs
to
the
brooks
deployment
versus
the
stuff
that
belongs
to
the
pet
clinic
deployment,
the
reconcilers
would
basically
argue
with
each
other.
Every
heartbeat
and
you'd
you'd
see
the
actor
count
constantly
going
up
and
down
and
up
and
down
this
way,
with
all
of
these
resources
being
labeled
as
part
of
a
deployment,
the
reconcilers
won't
fight.
E
D
A
A
A
A
E
A
A
E
D
A
It
will
it
will
never
be
required
if
you
want
this.
All
you
have
to
do
is
start
up
adam
on
the
same
nat's
infrastructure
that
you're
running
your
wisdom,
cloud
hosts
and
you
just
get
it
for
free,
but
there
are
aside
from
a
couple
of
things
to
make
the
host's
control
interface
accept
annotations
and
things
like
that.
A
A
Yeah
and
part
of
another
motivation
behind
this
particular
way
of
having
it
opt-in
is
with
wadam
not
being
part
of
the
host.
That
essentially
means
that
whatem
works,
regardless
of
what
host
runtime
is
up.
As
long
as
the
host
runtime
is
responding
to
the
control
interface,
it
can
be
controlled
by
watum,
which
means.
B
A
One
yeah
you
can
run
you
run
one
wadam,
you
run
widem
once
in
your
infrastructure.
It
supports
managing
app
deployments
on
multiple
lattices
at
any
given
time.
So
you
can
run
your
you
can
run
wadom
on
in
multi-tenant
mode
or
if
you
have
you
know
like
a
if
you're
you're
segmenting
your
lattices.
For
any
number
of
reasons,
you
can
still
use
the
same
instance
of
what
I'm
to
do
that
management
and
wadam
is
also
beam
clusterable.
A
So
you
can
start
five
instances
of
atom
and
they
will
all
be
failover
fault,
tolerant,
resilient
and
you'll
still
get
the
the
same
behavior.
So
maybe
next
time.
What
I'll
do?
I
think
you've
seen
me
just
demo
the
failover
behavior
before
but
like
I
can
start
a
deployment
monitor
on
node
one
kill
node
one
and
you'll
see
that
same
deployment
monitor
restart
on
node
two
immediately.
B
B
B
This
is
something
that
we've
known
and
we've
been
using
for
our
demos
like,
for
example,
if
you
saw
the
demo
that
taylor
and
I
did
at
kubecon
eu,
this
was
all
using
ngs
so
that
we
can
run
wasn't
cloud
hosts
across
a
bunch
of
different
cloud
providers
or
regions
or
wherever
and
just
kind
of
seamlessly
connect
them.
I
really
wanted
to
write
up
the
instructions
for
how
to
get
this
set
up
just
from
start
to
finish,
because
it
really
does
not
involve
very
much.
B
I
mean,
including,
like
giant
screenshots
of
the
ngs
page
and
formatting
this
as
a
pdf,
just
to
show
you
which
is
awful.
It
only
takes
up
a
couple
of
pages
like
all
of
this
stuff
is
really
not
difficult
to
set
up.
So
I
really
I
wanted
to
show
that
off,
so
that
should
be
coming.
That'll
be
coming
a
little
bit
later
this
week
on
this
ngs
blog
post.
B
So
I
wrote
a
little
example:
actor
called
dogs
and
cats,
and
if
you
take
a
look
at
the
code
here,
it
is
a
it
is
in
an
actor
that
receives
http
requests
and
makes
http
requests.
So
the
business
logic
for
it
is,
you
know,
only
a
couple
of
lines
here:
minus
a
a
single
d
serializer,
but
we
create
an
http
client.
We
generate
a
random
number
between
zero
and
one,
and
if
it's
zero,
we
get
a
cat,
and
if
it's
one
you
get
a
dog
picture
and
then
we
make
a
request.
B
This
is
all
using
there's
two
public
apis
that
will
just
give
you
brand
of
images
of
cats
and
dogs,
which
is
great
and
then
return
it
as
an
http
response.
So
I
deployed
that
locally
got
my
dogs
and
cats,
actor
http
server
and
http
client.
These
are
just
both
the
latest
versions
of
from
wasmcloud
and
I
set
up
a
little
listener
and
look
at
this.
If
you
hit
refresh
over
and
over
you
just
get
random
images
of
cats
and
dogs,
maybe
slightly
biased
towards
cats.
Oh
both
dog.
So
I
mean
I
don't
know
this.
B
One
was
just
kind
of
it's
fun
to
oh,
no.
I
broke
it
for
a
second,
this
one's
just
fun
too.
It's
fun
to
actually
just
get
down
and
write
a
little
example
or
actually
use
the
the
actor
frameworks.
So
I
wanted
to
show
this
one
off
and
I
actually
have
this
set
up
over
ngs
right
now.
I
was
doing
some
testing
for
the
the
blog,
so
I
have
another
host,
that's
actually
running
in
gcp.
B
So
if
I
want
to
do
something
like
take
the
http
client
and
run
that
in
gcp
instead,
I
can
totally
remove
it
from
my
localhost
started
in
gcp
and
then
come
back
and
you
know
we
keep
getting
pictures
of
of
dogs
and
cats,
even
though
it's
not
running
on
local.
So
that's
just
kind
of
a
sneak
preview
of
what
you
all
will
get
to
do.
B
As
you
run
through
these
instructions,
we've
been
really
excited
about
the
blogs
we've
been
putting
out
thus
far,
and
it's
it's
pretty
exciting
to
put
out
some
of
these
things.
That
really
really
makes
us
like
wonder
if
this
is
just
black
magic
when
you're
doing
things
over
ngs,
and
so
I'm
really
interested
in
seeing
what
what
you
all
think
of
it
happy
to
answer
any
questions
about
the
demo
that
I
kind
of
just
went
through,
but
I
just
wanted
to
show
off
dogs
and
cats
because
it's
that's
a
fun
one.
B
B
C
B
C
B
B
B
It
would
be
it's
exciting
to
be
able
to
show
these
things
off
and
then,
eventually,
depending
on
what
the
services
and
the
bandwidth
of
the
wazon
cloud
team,
we
love
to
kind
of
take
these
and
and
promote
them
as
like
official
capability
providers
that
we
distribute
things
like
the
file
system
provider
that
max
person
is,
is
pr'ing
into
our
capability
providers,
repo,
those
kind
of
things
we
would
love
to
show
off.
As
first
party
providers,
we
can
help
with
the
management
and
or
you
know
we
would
be
spearheading
the
management
and
the
upgrading
of
them.
B
I
think
a
lot
of
what
we
agree
on
so
far
is
that
it
would
be
great
in
the
capability
providers
repo.
If
we
have
a
you
know
down
here,
we
have
our
capability
providers
that
we
officially
support
to
also
have
a
community
provider
section.
This
is
the
first
step
for
showing
off
things.
You
know
in
your
own
repository
like
I
know
that
matt
gilbride
is
working
on
and,
I
think
finish,
the
key
value.
B
Implementation
for
dynamodb
that'd
be
a
great
thing
to
have
here
and
then
you
know
eventually
move
them
up
in
the
list.
I
think
what
we
really
want
to
figure
out
is:
what's
the
criteria
for
something
that
we
want
to
support?
Officially
it
would
you
know
if
you
look
at
some
of
the
services
we
already
support.
Http
server
and
http
client
are
pretty
standard
implementations.
B
B
All
right,
obviously,
it's
always
open
for
comments
there
and
then
one
other
rfc
that
I
wanted
to
bring
up
today.
Matt
wilkinson,
you
actually
wrote
and
are
on
the
call
today
yay,
but
you
kind
of
started
going
through,
and
I
think
you
presented
a
little
bit
of
some
options
that
we'd
like
to
see
in
a
wash
or
wasmcloud.tamil
file
that
kind
of
comes
with
the
actors
and
capability
providers,
for
example.
B
That
would
basically
let
wash
replace
some
of
the
makefile
components,
and
you
know
what
I
can
see
for
a
long
time
for
us
is,
is
not
not
completely
replacing
the
makefile
components.
We'd
like
this
to
be
completely
backwards,
compatible,
there's
no
harm
in
including
both
but
we'd
love
to
have
some
workflows
where
we
can
do
something
like
wash
build.
That
will
build
your
actor
with
the
language
tool
chain
and
and
then
sign
it
automatically,
and
then
things
like
wash
run
that
will
automatically
run
the
actor
in
a
wasm
cloud
host.
Things
like
that.
B
We're
looking
at
little
way
of
of,
I
guess
little
ways
but
big
ways
to
improve
the
developer
experience
so
matt.
I
know
you
proposed
this.
A
few
weeks
ago-
and
we
had
a
couple
of
discussions
here-
and
please
don't
read
this
at
all
as
and
ask
if
you're
like
still
working
on
this
or
like
gonna,
have
time
to
work
on
this.
F
F
I
think
it
was
the
config
crate,
with
the
rust
to
just
kind
of
quickly
parse
the
tomo
file
and
then
for
initial
concept,
even
just
use
the
callout
to
the
makefile
and
have
that
work,
the
the
like
the
project
make
files,
I
guess
so
yeah,
that's
pretty
much
where
I'm
at
with
that
right
now.
If
people
want
to
leave
comments
on
their
thoughts,
please
go
ahead.
B
Awesome
well,
matt
I'll,
probably
probably
ping
you
then
to
see.
Well,
you
know
what
kind
of
work
you've
done
if
you've
pushed
anything
up
to
a
fork
or
whatever,
and
we
can
probably
maybe
we
can
do
like
a
pairing
on
this.
For
the
like
the
first
implementation
or,
I
feel,
like
you,
normally
end
up
picking
up
some
awesome
feature
of
wash
and
just
like
sneaking
a
pr
in
on
the
weekend
that
just
like
blows
our
mind
when
we
come
in
on
monday,
which
is
pretty
fun.
So
thank
you
for
this.
F
B
All
right!
Well,
that
was
let's
see,
that
was
what
I
wanted
to
share
from
the
community
side
of
things.
Oh,
I
probably
should
mention
we
had
one
more
blog
actually
published
to
the
watson
cloud
site,
which
was
over
the
weekend
or
a
little
earlier
this
week.
Kevin.
Do
you
want
to
say
anything
about
the
the
anniversary
blog
post.
A
A
I
mean
yeah,
there's
a
bunch
of
reflections
in
there,
but
I
think
the
the
really
mind-blowing
part
is
that
we've
been
doing
wasm
cloud
for
three
years
and
we're
just
getting
started,
and
we
have
a
ton
of
really
cool
things
that
we're
we're
still
planning
on
doing
so.
Yeah,
it's
even
after
three
years,
we're
barely
getting
started.
C
B
G
A
B
A
B
A
B
A
Know
about
the
transition
between
rust
and
elixir,
so
I
guess
probably
the
maybe
the
first
place
to
start
there
would
be
the
the
blog
post
that
brooks
just
highlighted,
and
then
we
have
an
architectural
decision
record
specifically
covering
that
decision.
So
does
it
probably
does
a
better
job
of
explaining
it
than
I
can
off
the
cuff.
B
Yeah
totally
would
recommend
the
blog
post
for
the
the
journey
kevin.
It
looks
like
on
our.
It
looks
like
that
rfc
for
moving
the
dispatch
layer
to
otp
is
actually
just
an
rfc
in
the
loss
of
cloud
repo
and
it's
not
in
our
adr.
We
should
probably
get
that
moved
over.
I
think
that's
a
significant
enough.
One.
A
C
B
B
G
Not
only
that
it
was
great,
it
was
great
to
see
all
the
folks
that
showed
up
at
open
source
summit.
That
was,
you
know,
was
a
little
lightly
attended
compared
to
years
past.
You
know,
we
typically
would
see.
You
know
five
to
seven
thousand
and
yeah
that
one
had
about
twelve
hundred.
You
know
we
are
all
excited
about
detroit
and
we'll
be
out
in
detroit
in
force.
G
You
know
on
the
cosmonaut
side
anyway,
I'm
certainly
out
representing
wilson
cloud
and
as
always,
if
anybody
from
the
community
wants
to
volunteer
for
kubecon
eu,
we
had
folks
show
up
and
and
volunteer
and
work
the
work,
the
booth.
We
should
have
a
wasn't
cloud
project
booth,
as
well
as
a
nice
big
cosmonautic
booth
out
at
for
kubecon
us
so
not
too
early
to
start
planning
for
folks
that
want
to
come
out
and.