►
From YouTube: Lightning Talk: Wildly Distributed Programming: Wasm and the Future of Distribute... Brooks Townsend
Description
Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon Europe 2023 in Amsterdam, The Netherlands from April 17-21. Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Lightning Talk: Wildly Distributed Programming: Wasm and the Future of Distributed Computing - Brooks Townsend, Cosmonic
Distributed applications are notoriously difficult. The problems of running applications on multiple computers spanning physical and network boundaries have been studied for years. With all of the complexity of just deploying an application today, why go down the rabbit hole of designing it to be distributed? This is a demo-heavy talk where you'll see WebAssembly running all over the globe with wasmCloud. Brooks will showcase a real, and possibly ridiculous, distributed application using WebAssembly. You'll see an application whose components are spread across different cloud providers, different edge devices, and different regions. You'll see how, despite its multi-platform, distributed implementation, it can handle outages, failover, and migration without batting an eye.
A
All
right,
everyone,
hello,
welcome
to
widely
distributed
programming,
wasm
and
the
future
of
distributed
computing.
My
name
is
Brooks
Townsend
I'm,
a
lead
software
engineer
at
cosmonic
and
I've
also
been
a
maintainer
for
the
open
source
project
wasmcloud
for
the
past
two-ish
years
now,
love
programming
in
Elixir,
rust
and
webassembly,
which
puts
me
in
a
really
interesting
Venn,
diagram
of
programmers
and
serial
open
source
contributor
and,
of
course,
a
demo
Enthusiast
always
going
to
do
a
live
demo.
When
I.
A
A
So
when
we
look
at
the
Modern
Computing
environment,
we're
talking
about
this
trend
of
compute
getting
smaller
and
smaller
and
smaller,
and
you
may
have
seen
the
PC
Cloud
container
Evolution
we're
abstracting
more
and
more
away
from
the
individual
developer
for
the
benefit
of
the
developers,
mental
health
and
for
the
management
of
their
applications
right
so
with
Docker
we're
abstracting
away.
The
kernel
we
see
webassembly
is
going
A,
Step
Above
This
abstracting,
with
a
host
creating
a
truly
portable
piece
of
compute.
Now
this
is
great
and
with
wasm
Cloud
as
we're
looking
at.
A
What
we
can
do
is
we're
taking
advantage
of
webassembly,
there's
still
an
opportunity
to
abstract
away
the
non-functional
requirements
or
the
capabilities
from
applications.
So
to
give
you
a
concrete
idea
of
this
functional
versus
non-functional,
if
you
think
about
like
a
to-do
application,
your
functional
requirement,
you're
going
to
create,
read,
update,
delete
to
Do's
right.
That's
the
thing
that
you
as
a
developer
you
get
down,
and
you
want
to
write
that
code.
A
The
non-functional
requirements
are
things
like
I
I
need
to
receive
HTTP
requests
on
this
port
on
this
API
and
I
need
to
route
that
into
my
receiver,
and
then
I
need
to
be
able
to
store
those
to
Do's
in
some
kind
of
persistent
database
like
these
are
the
things
that
you
just
copy
and
paste
from
Project
to
project.
You
did
it
once
you
know
you
keep
doing
that
enterprises.
A
You
know
big
companies
do
this
across
thousands
of
different
applications,
and
that
makes
it
difficult
both
for
maintenance
and
when
vulnerabilities
arise,
to
do
these
kinds
of
patching.
So
this
is
where
we
see
you
know
ourselves
on
the
Computing
environment,
landscape,
now
I'm
going
to
go
through
quickly
to
describe
wasm
cloud
and
why
it's
perfect
for
this
distributed
use
case
and
why
it
uses
webassembly
in
excuse
me
five
different
layers
at
the
base
level
we're
using
webassembly
we're
running
webassembly.
This
is
the
code
that
you
as
a
developer
would
write.
A
So
we
get
all
the
benefits
of
portability
deny
by
default
security
and
scalability
that
that
webassembly
provides
the
wasmcloud
app
runtime
sits
on
top
of
that
removes
the
boilerplate
code
and
lets
you
securely
access.
Those
non-functional
requirements
that
I'm
taught
that
I
was
talking
about
in
the
last
kind
of
section
now
the
capabilities
these
are
both
an
abstraction
that
makes
writing
your
application
easier
and
something
that
you
get
to
pick
at
runtime.
So
you're,
working
on
your
local
development
environment,
you
write
a
webassembly
application.
A
Now
these
composable
actors
just
again
a
fancy
way
of
saying
webassembly
modules
for
wasm
cloud
Implement
that
business
logic.
This
is
where
you're
writing
your
code,
they're
coded
in
terms
of
contracts
and
because
they're
so
tiny.
It's
really
easy
to
scale
these
out
to
different
clouds,
different
edges
and
what
really
comes
to
be
a
challenge
with
that
is
webassembly
can
really
run
in
a
ton
of
different
places,
but
it
doesn't
really
matter
very
much
for
composable
applications
when
you're
running
them
on
the
edge.
A
If
you
can't
connect
to
it,
if
you
can't
have
your
compute
communicate
with
each
other
and
that's
exactly
what
the
lattice
network
does
for
wasm
cloud,
it's
a
it's
a
mesh
Network
enabled
by
Nats
under
the
hood,
essentially
lets
your
application
seamlessly
talk
to
each
other,
no
matter
where
they're
physically
running.
No,
no
concept
of
you
know
discovering
nodes
or
running
in
a
specific
location,
and
you
know
I
work
for
a
company
called
cosmonic
cosmonic.
A
The
platform
is
essentially
a
managed
platform
for
webassembly
applications
went
into
develop
a
preview
about
a
month
ago,
and
you
know
obviously
accepting
everybody
here
to
come
sign
up
during
kubecon.
But
it's
the
painless
experience
to
develop
those
server-side
webassembly
apps
easy
to
go
across
platform
across
Cloud
you'll.
Get
to
see
that
today
here
in
about
a
second
and
the
really
cool
thing,
is
everything
the
cosmotic
platform
is
is
built
on
wasmcloud,
so
we're
running
the
cosmonic
platform
using
wasmcloud
and
webassembly
in
production.
A
Now
so,
when
you
go
play
around
with
the
platform,
it's
all
webassembly,
and
so
that
you
know,
I
would
love
to
talk
about
more,
but
actually
Taylor
Thomas,
my
co-worker
and
I
we're
doing
a
full
talk
about
it
at
the
main
conference
on
Wednesday.
So
you
know
if
you're
sticking
around
hope
to
see
you
there
love
to
talk
more
about
it.
Now,
that's
enough!
A
So
this
is
a
real
real
to-do
list,
real
distributed
application.
But,
of
course
it's
really
simple
when
you
look
at
it.
So
what
what
is
the
actual
architecture
of
this
application?
Look
like
I
have
no
idea
if
this
is
going
to
be
legible
at
all.
That's
not
bad,
so
I'm
running
an
HTTP
server
or
a
webassembly
module,
that's
accepting
HTTP
requests
on
cosmonic
and
what
this
webassembly
module
is
it's
a
it's
like
an
API
Gateway
for
to
do
so.
A
It's
receiving
my
API
request,
I'm,
sending
a
message
out
on
a
message
broker,
specifically
using
Nats,
but
like
you
can
swap
that
out
for
Kafka
or
for
red
panda,
which
is
kind
of
just
Kafka,
anyways
or
whatever,
and
then
I
have
another
webassembly
module.
That's
storing
those
things
in
a
key
value,
store
now
I'm
running
this
app
actually
on
AWS,
gcp,
Azure
and
Oracle
Cloud.
All
of
those
things
are
seamlessly
connected
into
one
network.
A
So
to
show
you
what
this
looks
like
I
can
go
over
to
cosmonic
I
have
five
Cloud
hosts
that
are
running
you
can
see
like
this
is
the
Oracle
one.
This
is
the
Azure
one
they're
on
a
variety
of
different
architectures
operating
systems,
and
you
know,
regions
and
clouds,
which
is
awesome
because
I
wrote
this
whole
app
on
my
local
MacBook
and
you
know
obviously
I'm
not
come.
You
know,
compiling
and
deploying
to
Mac,
but
I
never
had
to
worry
about
any
of
that.
A
When
I
was
writing,
my
application
I
deploy
it
on
my
local
and
it
works
all
the
same
when
I
deploy
it
to
the
cloud
and
when
you
look
at
what
this
actual
application
looks
like
I've
got
my
HTTP
server,
that's
getting
those
requests
and
forwarding
them
to
the
to
do
webassembly
module.
That's
that's
sending
out
a
message
to
the
distributed
key
value,
webassembly
module
and
that's
storing
that
in
a
redis-backed
data
store,
but
each
individual
one
on
each
different
cloud
is
configured
to
talk
to
on
AWS.
A
It's
elastic
cache
on
Azure,
it's
a
cloud
mem
store
or
whatever
on
gcp.
It's
Cloud
memory
store,
so
I'm
not
just
like
distributed.
Storing
this
in
one
cloud
like
across
a
few
different
regions,
actually
storing
it
in
four
different
clouds
in
four
different
databases,
and
is
this
the
architecture
that
you
want
for
your
production?
App?
No!
No!
You
probably
don't
want
this
right.
It
is
a
ridiculous
demo
on
purpose.
But
when
you
hear
the
tagline
like
that's
something
that
someone
would
talk
about
for
a
cloud
native
application
today,
you
know
the
cross
Cloud
distributed.
A
Now,
why
would
I
do
this?
Why
would
I
over
engineer
the
heck
out
of
something
that
can
just
be
a
to-do
application
and
it's
for
two
reasons,
one
to
show
that
the
future
of
applications
is
distributed
both
because
it's
motivated
by
The
Edge
and
it's
creating
the
edge?
But
people
have
figured
out
like
running
a
container
on
a
cloud
host
pretty
well
like
if
you
just
need
to
do
that,
you're
pretty
golden,
but
the
edge
capabilities
and
what
we're
seeing
with
people
making
applications.
A
Today,
people
are
trying
to
get
the
fastest
response
times
running
it
as
close
to
the
user
as
possible.
People
are
trying
to
get
the
lowest
cost
so
not
running
this
massive
cluster
running
things
as
close
to
the
data
as
possible,
too,
to
save
on
network
costs
and
the
reason
why
this
is
a
perfect
use
case
for
webassembly
is
because
webassembly
brings
that
portability
and
security
wherever
you
go,
you
write
your
application
when
you
develop
on
your
laptop
great,
but
your
Enterprise,
you
know
when
you're,
when
you're,
creating
this
app
for
your
company.
A
A
It
removes
the
management
from
all
those
things
and
so
really
at
the
end
of
the
day,
webassembly
is
perfect
for
this
type
of
distributed
computing
thing,
because
it
brings
that
that
portability
and
security
and
and
connecting
it
all
to
the
edge
is
the
challenge
that
we're
trying
to
tackle
now.
I
have
a
couple
couple
references,
please
feel
free
to
join
our
slack.
Everything
that
you
saw
today,
even
if
it
was
through
the
lens
of
the
cosmonic
product,
is
empowered
by
the
open
source,
all
open,
Core
Company.
All
of
this
code
is
open
source.
A
A
That's
a
that's
a
great
question,
so
the
kind
of
capabilities
model
that
we
are
chasing
with
wasmcloud.
This
is
actually
something
we've
been
working
with
since
2019,
so
so
something
that's
been
around
for
like
pretty
long
before
the
component
model.
However,
looking
at
what
components
and
what
world
files
are
bringing
this
is
actually
a
perfect
place
for,
for
this
is
exactly
where
wasm
cloud
is
going
and
though
it
doesn't
have
support
for
the
component
model
today.
This
is
exactly
our
plan
right.