►
Description
Fast-moving forward to hybrid/multi-cloud stack, it’s inevitable for developers not only considering implementing serverless microservices once but also integrate with the multi-serverless platforms on Kubernetes as the smallest/fastest as they can. In this session, we’ll explore how Quarkus, Kubnertes Native Java Stack allows developers to write traditional microservices and make them deployable functions to multiple Serverless Platforms(i.e. Knative + Kubernetes, Amazon Lambda, etc.).
Presenter:
Daniel Oh, Principal Technical Marketing Manager @RedHat and CNCF Ambassador
A
Okay,
let's
get
started
hello
everyone.
Thank
you
very
much
for
joining
us
today
to
today's
cncf
webinar,
hybrid
serverless
development
using
quarkus
and
kubernetes,
I'm
jerry,
fallon
and
I'll
be
moderating
today's
webinar.
We
would
like
to
welcome
our
presenter
today,
daniel
o
principal
technical
marketing
manager,
at
red
hat
and
cncf
ambassador
just
a
few
housekeeping
items.
Before
we
get
started
during
the
webinar.
You
are
not
able
to
talk
as
an
attendee.
There
is
a
q,
a
box
at
the
bottom
of
your
screen.
A
Please
feel
free
to
drop
in
your
questions
in
there
and
we'll
get
to
as
many
as
we
can
at
the
end
there.
This
is
an
official
webinar
of
the
cncf
and,
as
such
is
subject
to
the
cncf
code
of
conduct.
Please
do
not
add
anything
to
the
chat
or
questions
that
would
be
in
violation
of
the
code
of
conduct.
Please
be
respectful
of
your
fellow
participants
and
presenters.
A
B
Jerry
yeah,
hello,
everybody
and
welcome
to
joining
you
cncf
regular.
So
let
me
kick
you
off
the
cool
video
first
here.
B
B
B
Okay,
so
what
a
fantastic
video
to
wake
up
our
new
webinar-
and
so
maybe
so,
maybe
you
can
remind
you-
know
how
the
java
was
born
and
how
to
get
started
and
how
it
was
developed
in
the
last
decades
ago.
So
actually,
this
year
is
a
25
years
anniversary
of
a
java
application.
B
So
there
are
a
lot
of
changes.
Maybe
you
people
already
playing
with
the
java
application
development?
Okay.
So
let's
jump
in
today
topic
hybrid
serverless
development,
using
focus
and
kubernetes,
so
my
name
is
danielle.
I'm
working
for
redhead
as
a
technical
marketing
major
with
the
specialized
clientele
application
development,
as
well
as
assigned
devops
practice,
and
gerry
already
mentioned
that
I
am
cncf
investor.
B
Also
devops
is
the
investor
in
order
to
evangelize
the
current
every
application,
architecture,
references
and
development
practice
with
the
along
the
devops
initiative,
and
then
just
I
wanna
encourage
you
all
guys
to
subscribe.
My
youtube
channel
the
beam
me
danielle
tv
and
you
will
find
a
lot
of
technical
demo
and
the
technical
tips
and
interesting
even
non-technical,
interesting
tutorials.
B
Okay,
you
can
also
follow
my
twitter
and
I
I
recently
published
my
ansible
book.
If
you
could,
if
you
are
interesting
as
a
reviewer,
just
let
me
know
by
twitter
or
my
youtube
channel,
all
right
move
on
so
just
just
take
a
step,
just
one
step
back
and
take
a
look
at
dad.
Why
we
needed
to
partners
to
develop
the
hybrid
subless
application
with
the
kubernetes
cluster.
B
So
just
take
a
look
at
that.
You
know
back
in
1995
java
was
born
and
unreached
upon
the
world
and
at
a
time
most
of
the
developer
really
loved
to
use
c
or
c
prosperous
rather
than
java.
But
we
have
a
such
a
great
change
in
around
the
world
like
internet
technology,
it
was
came
up
and
then
the
jabra
is
was
really
feeling
fitting
into
that
technology.
B
And
then,
after
a
couple
more
years
later,
so
the
all
enterprise
company
is
totally
change
their
infrastructure
layer
and
also
software
layer
to
maintain
and
run
their
business
application
for
the
end
users.
So
here's
some
example
to
speed
up
your
java
application
to
serve
your
business
capability
on
your
customer
or
end
user
customers.
B
So,
first
of
all,
almost
4
463
thousand
dollars
at
the
first
time
to
buy
software
hardware
and
specific
middleware,
etc
to
all
speed
of
your
by
application
as
well
as,
but
also
you
need
to
in
a
project
company
perspective.
They
needed
to
spend
more
than
eighty
thousand
dollars
every
single
year
to
maintain
software
hardware
etc.
B
But
this
is
actually
happened
and
many
25
years
ago
and
then
just
a
little
bit
narrowed
down
an
enterprise
architecture.
Point
of
view.
We
just
only
one
application
architecture,
also
known
as
a
monolithic
application,
so
main
priority
of
a
monitor
application
just
running
a
lot
of
the
application,
but
sometimes
the
fiber
application,
workflow
application
running
on
multiple
app
server
based
on
high
availability
architecture,
maybe
two
app
replication
or
thread
the
application
or
multiple
virtual
machines,
sometimes
a
bare
metal.
B
And
then,
if
you
want
to
change
your
business
functionality
or
even
office,
you
need
to
waiting
for
maybe
two
months
away
one
month
because
you
the
first
priority
day
application
server
is
scalability
in
the
reliability
rather
than
some
frequent
build
and
deployment
like
a
ci
city.
B
In
the
meantime,
we
have
a
such
a
great
opposite
technology
and
also
public
cloud
to
serve
a
lot
of
great
technology
to
many
enterprise
companies,
so
just
for
example,
back
in
2006.
B
The
amazon
first
of
all
served
the
public
instance,
which
means
many
companies
doesn't
need,
don't
need
to
worry
about,
purchase
new
hardware
and
maintain
dating
just
as
I
already
mentioned
earlier
yeah.
So
they
don't
need
to
spend
eighty
thousand
dollars
every
single
year
because
they
just
buy
and
just
consume
resources
from
a
public
cloud
provider
like
amazon
and
then
just
need
to
pay
for
that,
like
a
pay-as-you-go
policy
and
also
2009
the
huge
change
of
the
java
enterprise
application,
I've
been
there,
so
I
actually
enterprise
java,
developer
and
architect.
B
In
2009,
I
spent
a
lot
of
time
to
develop
a
business
application
like
a
cloud
and
a
data,
integration
and
integration,
external
server,
etc,
and
after
a
couple
more
years
later,
and
then
the
linux
container
was
born
and
the
kubernetes
born
is
totally
game
changer
as
an
impactor
layer.
So
now
microservice
is
a
plus
container
like
a
docker,
an
oci
enemy,
z
format
and
a
kubernetes,
also
known
as
immutable
infrastructure,
is
totally
changing
your
mind.
B
I
mean
how
to
optimize
your
application,
how
to
make
it
reliable
how
to
make
it
scalable
how
to
make
it
faster
and
higher
performance
on
top
of
that.
B
So
now
so
we
enter
project
companies,
don't
worry
about
the
purchasing
infrastructure,
but
they
really
more
prefer
to
use
manage
the
service
rather
than
manageable
service.
So,
as
you
discharge,
the
amazon,
google,
microsoft
and
a
lot
of
the
company
already
provide
manage
the
service,
and
then
customers
just
go
to
go
to
that
public
service
and
pick
up
the
whatever
they
need.
For
example,
I
just
need
to
storage.
I
just
need
to
memory
and
cpu
based
on
t-shirt,
size
condition
and
after
that,
just
needed
to
pay
for
that
hourly,
weekly
or
monthly.
B
This
is
a
change,
the
trend,
how
to
maintain
your
infrastructure
in
the
end,
how
to
run
your
application
so
in
this
trend,
so
we
need
to
think
about
okay,
small
application
in
a
more
efficient
application
with
a
less
memory
footprint
and
a
small
size
of
application,
which
means
the
low
bandwidth
or
network
traffic
is
more.
It's
gonna
be
reduce
your
cost
money.
So
so
that's
why
this
is
a
new
trend
to
consider
new
application,
not
only
java,
but
also
the
other
application
program
framework
like
a
gold,
python,
node.js
and
vertex,
etc.
B
So,
just
take
a
look
at
that.
The
architecture
standpoint
in
the
cloud
java
application
stack,
so
architecture
is
evolving
to
microservice
architecture,
which
means
you
can
run
single
application
functionality
on
top
of
the
single
runtime
like
a
spring
boot
and
node.js,
and
javascript
and
perfx,
and
even
dynamic,
core
etc
and
the
lifecycle
pretty
much
is
shortened.
So
now
we
can
deploy
maybe
every
single
day
or
two
time
twice
a
week.
B
You
know
pretty
much
a
faster
than
a
monitor
application,
because
you
we
have
some
kind
of
such
a
continuous
integration,
continuous
deployment
or
delivery
tool
like
jenkins
and
tech
tone
and
sucker
ci.
There
are
a
lot
of
practical,
great
cica
to
make
it
happen,
and
memory
footprint
are
pretty
much
less
than
like,
100
or
1200,
maybe
300
megabytes
with
a
memory
for
free,
and
you
just
need
to
a
few
seconds
to
start
up
your
application.
B
So
let's
try
to
understand
how
java
is
developed
and
why
we
need
to
use
that
for
developing
a
business
application
and
to
serve
world
wide
web,
also
known
as
internet
technology.
So
java
was
designed
for
high
performance
and
networking
and
then,
which
means
the
first
priority
of
the
java
application.
You
need
to
serve
your
application
all
the
time
and
with
your
high
performance,
rather
than
a
big
machine
or
the
high
cost,
to
maintain
your
hardware
or
a
multiple
layer
so,
and
also
the
java
also
was
designed
for
a
long
running
process.
B
Nobody
care
about
how
much,
how
many
memory
and
how
many
hardware
you
need
to
prepare
to
run
this
long-running
process,
but
but
they
just
really
care
about.
Specifically
it
operation
team.
They
really
are
interesting
about
how
many
times
this
application
have
some
error
exception,
or
some
kind
of
security
problem
so
generally
the
long-running
java
process.
Application
can
be
running
all
the
time
by
maybe
24,
7
or
365,
except
for
security,
patch
or
bug
fixed
critical
issue
or
even
operating
system
patch.
B
But
without
that
case,
your
application
over
the
end
of
project
app
by
application
should
be,
can
be
running
all
the
time
like
a
24
7..
The
problem
is
it's
a
huge
expense
making.
So
you
have
to
pay
lots
of
money
to
maintain
this
long
running
process
and
also
java
is
based
on
the
dynamic
behavior
characteristic
to
run
the
java
application
on
mutable
system
like
just
static
web
server
or
virtual
machine.
B
So
if
you
have
more
resource
resources
to
run
your
java
application,
we
just
need
to
one
more
cp,
just
scan
of
cpu
or
need
to
more
memory
resources
to
make
it
stable.
That's
the
the
dynamic
behavior,
because
the
dynamic
behavior
standpoint
so
java
java
is
the
running
on
java,
based
on
java
application.
You
can
just
compile
your
source
code
based
on
source
code,
also
known
as
the
buy
code,
and
then
your
media
code,
and
then
you
can
run
that
by
called
any
app
server
on
on
the
internet.
B
So
this
is
in
the
during
the
runtime,
your
job.
By
application.
There
are
a
lot
of
tasks
needed
to
be
done
behind
the
scene.
Let
me
talk
a
little
about
that
thing
just
a
little
bit
later,
but
this
dynamic
behavior
really
awesome,
because
when
you
create
your
java
application
and
compile
by
code-
and
you
can
run
any
application
server,
which
means
there's
no
racking
middleware
or
app
server,
you
can
buy
any
app
server
from
any
vendors,
any
open
source
project
tools
etc.
B
The
java
needed
to
have
some
the
dynamic
task
at
one
time
behind
the
scene,
but
when
you
scale
out
thousand
container,
which
means
you
have
to
dynamic
test
it
a
thousand
times
it's
not
a
productive
way.
So
that's
why
people
okay
java
could
be
can
be
a
heavyweight
and
one
container
platform
like
the
kubernetes.
B
So
why
do
you
think
it's
happening
and
let
me
explain
the
truth
of
java
plus
container
technology.
As
you
know,
java
was
born
proper
linux,
container
technology
or
kubernetes
technology.
So
that's
why
java
really
focused
on
world
wide
web
internet
technology
rather
than
immutable
infrastructure
so
needed
to
long-running
process
and
some
memory.
It's
a
little
bit
large
memory,
footprint
and
a
star.
B
So
that's
why
we
spend
maybe
a
couple
more
seconds,
even
30
seconds,
to
start
off
the
java
application
but
node.js
and
go
pretty
much
better.
Then
it's
a
pre
lightweight
and
a
small
memory
for
free.
As
you
see
here,
the
red
box
in
your
kubernetes
container
platform
and
the
blue
box
is
just
one
worker
notes.
Each
worker
node
has
a
same
resource
density,
but
you
can
see
there
are
only
full
paths
of
a
java.
B
Application
can
be
running
on
one
work
or
not,
but
node.js
may
8
and
then
go
it's
maybe
24
it's
two
times
three
times,
which
means
the
gold
and
node.js
a
small
size
application
in
a
prefab
pretty
fast
and
also
small
memory
footprint
rather
than
java.
B
So
this
is
another
reason
why
a
lot
of
developers
really
prefer
to
develop
the
serverless
application
and
running
on
tablet
serverless
platform
such
as
amazon
lambda.
So
new
really
survey
this
symmetry
just
already
at
the
beginning
of
this
year,
so
you
can
see
only
six
percent
java
application
was
taken
for
the
language
used
on
amazon
lambda
because
we
already
learned
about
the
java
of
hidden
truths.
B
So
a
lot
of
memory,
fulfilling
and
dynamic,
behavior
and
long
learning
process
is
definitely
does
not
fit
in
supportless,
even
kubernetes
cluster
or
the
optimization
perspective.
B
And
let's
take
a
look
at
that
suburban
landscape.
This
is
a
came
from
cncf,
so
you
can
see
there
are
a
lot
of
boxes
in
there.
So
there
are
a
lot
of
tools
framework
and
manage
the
platform
and
a
manageable
platform.
You
could
install
your
own
manageable,
manageable
platform
like
an
on-prem
system,
so
the
question
is
which
tool
which
framework
and
which
platform
would
be
perfect
to
my
serverless
application,
even
I.t
operation,
team
or
devops
team.
B
Okay,
they
need
to
make
a
choice:
the
best
tool,
best
framework
and
best
platform
to
manage
traditional
microsoft's
application,
as
well
as
serverless
application
at
the
same
time.
So
the
result
is
yeah
too
many
choices.
B
So
so
we
need
to
something
different
thing,
but
in
the
meantime,
java
actually
has
to
try
to
a
lot
of
effo
and
some
time
spent
a
lot
of
time
and
apple
to
make
it
better
even
try
to
evolve
to
catch
up
the
new
id
trend
and
technology.
So,
for
example,
try
to
optimize
the
java
runtime
based
on
linux,
container
packaging
technology.
B
Try
to
try
to
reduce
memory,
realization
or
cpu
utilization
also
create
a
new
the
garbage
collector
mechanism
like
shenandoah
and
also
also
make
it
distribute
the
different
type
of
operating
system
like
android
os
based
on
direct
system.
So
there
are
lots
of
apple
and
some
evolving
projects.
That
would
happen,
but
still
that's
not
good
enough.
So
we
need
more
something
different
to
make
java
make
it
greater
make
the
java
greater
along
with
the
container
technology
and
kubernetes
technology.
B
So
corkers
was
born
just
one
and
a
half
years
ago
the
progress
is
open
source
javascript
to
create
and
develop
crown
level
microservices
application
in
the
develop
serverless
application,
along
with
the
kubernetes
kubernetes
cluster,
and
sometimes
you
can
build
up
very
easily,
even
driven
application.
On
top
of
that,
so
how
it
works.
Let
me
let
me
explain
that
very
quickly,
so
quarters
is
pretty
optimizing.
B
The
this
is
there,
so
we
just
limited
the
dynamic
behavior,
but
still
you
can
use
the
same
java
application
same
java
capability,
but
we
just
removed
that
thing,
and
so
now
we
just
follow
the
microservice
architecture,
but
you
can
deploy
the
traditional
microsoft
application
to
long-running
process,
as
well
as
serverless
applications
in
the
life
cycle,
maybe
sometimes
just
a
millisecond
and
still
you
can
deploy
maybe
just
one
per
day
or
just
twice
a
week,
just
like
a
ufc
icd
pipeline
strategy
also
start-up
time
pretty
much
faster
than
like.
B
Sometimes
your
millisecond.
Let
me
show
you
in
demo
quickly
just
a
little
bit
later
and
then
the
memory
footprint
is
pretty
much
a
small,
pretty
much
rust
then.
So.
This
is
how
the
traditional
java
framework
is
doing.
So
in
a
build
step,
you
can
just
create
a
buy
code
you're
after
the
your
packaging
tool
like
a
maven
grotto
and
then,
when
you
run
your
java
application
on
top
of
the
middleware
or
just
using
java
command
line.
So
there
are
a
lot
of
steps
to
do
under
the
the
hood.
B
So
first
of
all,
the
java
tried
to
load
the
configuration
all
file
and
I'll
try
to
parse
them.
Okay,
this
is
the
dev
environment
configuration.
Maybe
this
is.
We
don't
need
it
at
runtime,
because
we
use
the
lazy
corporation
strategy
and
also
try
to
load
classes
and
enable
disable
some
feature
and
also
scanning
annotation
and
parsing,
describer,
etc,
etc,
and
then
last
step
try
to
start
the
actual
application
like
a
thread
and
a
pool.
B
So
how
make
how
focused
to
make
it
different.
So,
as
you
see,
we
just
shipped
all
required
that
thing
to
the
left.
So
now
you
can
see
the
partners
try
to
that
at
the
build
time
all
require
a
task
it's
created
by
code
and
the
loading
file
and
a
scanning
and
a
parsing
in
a
packaging.
B
We
can
do
that
at
build
time
and
also
there
are
two
options
you
can
package
your
application.
First
of
all,
just
to
make
it
a
finja
like
a
jar
file
and
the
other
one
is
the
native
executable
file.
Like
a
gold
program,
you
can
run
directly
your
executable
file
and
running
on
gravie
and
rather
than
jvm,
based
on
ahead
of
time
ahead
of
time
strategy
com
corporation
strategy.
So
now,
when
you
see
the
runtime
environment,
we
just
run
this
application.
B
So,
first,
first
of
all
we
can
create
the
bin
jar
and
then
provisioning
and
create,
and
then
you
could
have
two
choices
you
can
create
enable
corporation
or
you
can
also
create
a
name
corporation
and
create
another
executable
file
and
just
run
it
on
and
also
the
serverless
standpoint
in
quarkx
provide
the
funky
annotation
so
annotate,
a
funky
extension
extension
is
part
of
the
capability
of
java
capability
for
situated
data,
transaction
or
less
less
import,
less
exposure
and
then
messaging
integration
and
cloud
integration
and
also
monitoring
observation
integration.
B
All
capability
is
so
we
can
call
the
quarkx
extension
once
you
add
quark
extension,
you
can
see
some
maybelline
dependency
on
your
maven
project,
I'm
going
to
I'm
going
to
show
you
in
a
live
demo
very
quickly,
just
a
little
bit
later.
So
here
is
the
funky
extension
to
make
a
java
standard
portable
api.
B
So
this
is
the
way
the
corkers
how
to
enable
java
developer
to
develop
a
serverless
application,
but
still
there's
no
learning
curve
to
catch
a
new
serverless
technology.
B
B
So
the
wonderful
beauty
of
the
partners
framework,
the
quarkx
provide
the
quarkx
2
as
the
extension
for
multiple
ids
2
like
a
vs
code,
just
like
mine
or
the
intellij
or
eclipse
or
netbeans.
You
can
just
install
or
add
a
plugin
or
extension
of
quarkx
2
on
your
vs
code
or
on
your
id
tool,
and
then
you
can
easily
create
a
progress
project.
B
So,
let's
try
to
create
a
focus
project
and
I'm
going
to
pick
it
up,
maybe
and
gradual
for
for
today,
I'm
going
to
just
add
the
maven
and
you
could
set
up
your
group
id
and
also
artifact
id,
and
I'm
gonna
just
leave
all
default
valuable
today.
So
this
is
the
project.
Budgeting
name
and
my
project
name
me
daniel
and
so
resource
name
of
the
partners
provides
cdi
capability
as
well
as
a
rest
endpoint
capability
by
default.
That's
why
we
just
there
are
one
resource
file-
will
be
generated
automatically.
B
So
just
name
is
it
reader
resource
here
and
then
I'm
not
going
to
pick
any
extension
at
this
moment
and
I'm
going
to
save
this
file
in
my
template
directory
now
you
can
see
here
so
palm
xml,
so
you
can
see
summer
dependency
automatically
generated
here,
pull
down
rest,
easy
and
unit
test,
and
if
you
don't
need
some
specific
dependency.
For
example,
the
qt
is
one
of
a
temporary
generator
for
your
prone
and
application
development.
So
we're
not
going
to
use
this
one.
B
So
I'm
going
to
remove
that
and
go
to
source
file
and
the
one
java
file
grid
resource
generate
automatically
here.
So
maybe
I'm
gonna
make
it
bigger
for
you
all
attendees
here.
So
as
you
see
just
a
hello
world
example,
so
endpoint
hello
and
return
hello
and
pretty
simple.
So
now
I
am
java
developer,
let's
say
and
then
I
needed
to
develop
new
function,
but
in
for
that
I
needed
to
make
sure
this
application
totally
working.
So
let's
try
to
open
a
new
terminal
here.
B
So
maybe
parkers
and
dev,
so
this
command
line
is
amazing
plugin,
but
you
can
run
quarkx
application
as
a
development
mode
for
developer.
Once
you
run
your
application,
you
will
see,
live,
cooling
activated
and
also
the
install
features
is
the
cdi
bean
rest
easy.
This
is
a
default
feature
by
quarters.
It's.
This
is
an
awesome
feature
for
developer,
because
the
cdi
means
rest
easy
to
require
functionality
to
develop
your
modern
microservice
application.
B
B
So
let's
try
to
hello
what
about
welcome
cncf
webinar
and
I
just
saved
my
file
and
let's
try
to
invoke
the
endpoint
one
more
time
now
we
have
the
result
here
when
we
come
cncf
webinar
here,
but
the
carcass
automatically
hardly
replaced
by
application,
which
means
the
developer
doesn't
need
to
rebuild
repackaging
and
redeploy
and
restart
my
runtime
environment,
corkers
or
the
wired
and
make
it
happen
for
developer
automatically
magically.
This
is
what
live
coding
means
exactly
in
the
end,
this
the
this
is
a
maybe
boring
task
for
every
single
time.
B
You
change
the
your
code,
but
this
is
a
mandatory
task.
When
you
develop
your
application,
it's
happening
every
single
day
for
developer
the
daily
life.
So
in
the
end,
this
feature
will
increase
the
developers
probability
specifically
working
from
home,
even
working
in
your
company,
all
right,
so
I
just
changed
your
written
code.
Maybe
some
people
will
be
curious
about,
if
you
like,
add
a
new
java
class,
even
umass
and
the
new
class.
Is
this
still
walking
automatically
or
the
wire
and
the
ready
pro
re-running
without
any
stop
any
manual
stuff?
B
Let
me
show
you
that
thing.
So,
I'm
gonna
develop
new
function,
a
new
microservices
method
here,
in
order
to,
let's
try
to
add
a
new
cdi
bin.
Here,
I'm
gonna
add
a
new
file,
glitter
service
java
file,
and
then
this
is
a
class
and
I'm
gonna
add
this.
Is
the
applications
called
javabeans,
so
one
single
method,
just
the
greeting
and
with
some
parameter
like
just
name
and
I'm
gonna-
add
written
code
as
the
let's
say,
welcome
and
then
the
name
and
then
just
like
hybrid
serverless
development,
using
partners.
B
That's
it
just
simple:
I
just
passed
down
parameter
and
I
put
in
my
return
code
and
then
go
back
to
resource
file
and
first
thing
I
needed
to
inject
my
cdi
bin
here
reader
service
and
it's
just
service.
B
So
it's
here
and
then
I'm
going
to
create
a
new
one,
just
copy
dc1
make
it
faster
here
and
then,
but
I
need
to
differentiate
the
endpoint
path
here,
like
a
green
and
I
need
to
dynamically
mapping
my
parameter
like
a
name
and
then
change
the
method.
Name
like
a
greeting
and
mapping
the
pet
parameter
here,
just
the
easiest
way
to
make
a
good
readability,
I'm
just
using
the
same
name
here
and
I'm
not
going
to
return
the
static
code
here.
B
But
I'm
going
to
invoke
my
cdib
service
and
greeting
that's
all
so,
just
to
recap,
I
just
add
a
new
and
endpoint
method
and
I'm
using
the
cdib.
I
just
created
a
new
class
here
and
then
let's
try
to
existing
application.
B
B
Okay,
greedy,
oh
yeah,
this
is
the
typo
find
out
and
let's
try
it
one
more
time.
Okay,
now
we
are
hello:
daniel,
oh
and
hybrid
serverless
development
using
parker's,
pretty
cool.
I
don't
even
change
my
I
don't
even
rebuild
repackaging,
video
pro
and
restart
my
application,
none
time,
but
still
it's
totally
working
on
my
local
machine.
This
is
pretty
awesome.
B
So
what
I'm
going
to
do
next
now,
I'm
going
to
packaging,
I'm
going
to
start
my
live
encoding
and
I'm
going
to
packaging
my
application
and
I'm
going
to
skip
the
unit
test
here,
because
the
quarkx
already
includes
the
unit
test
based
on
juni
by
default
and
then
the
the
generator.
The
test
scenario
is
expected:
written
called
hello,
but
as
already
you
know,
we
just
changed
that
something
like
the
wacom
cnc
webinar.
So
in
reality
you
have
to
add
or
model
file
to
generate
a
test
scenario
along
with
your
application.
B
But
in
this
demo
I'm
going
to
skip
that
and
I'm
going
to
run
my
application
using
java
command
the
java
and
the
in
the
target
directory
and
the
hybrid
and
run
on
java
file.
But,
but
for
that,
so
I'm
going
to
open
my
activity
monitor
in
this
my
local
machine
and
there
are
cpu
memory
utilization
and
go
to
memory,
and
I
just
try
to
check
it
out
how
many
jvm
is
running
on
my
logo.
You
know.
Actually
I
have
a
lot
of
jbm.
I
don't
even
know
that.
B
Okay,
so
just
take
a
look
at
the
here,
the
last
two
one
203
megabyte
or
jvm,
and
the
other
one
60
and
then
go
back
to
id2
and
just
try
to
run
this
application.
B
B
So
now
you
can
see
just
pulling
a
second
to
start
up
pretty
much
faster
than
any
other
existing
java
framework,
because
you
already
know
maybe
spring
boot
and
the
other
javascript
use
takes
up,
maybe
two
and
three
seconds:
even
five
seconds
shut
up.
But
this
is
under
one
second
and
then
try
to
access
the
endpoint
so
same
result,
which
means
this
applicant
totally
working
after
eye
packaging
binger
rather
than
fetcher,
because
then
you
can
see
target
directory
and
then
we
just
need
to
hear
just
227
kilobytes.
B
This
is
a
finger
and
I
go
back
to
memory
activator
and
now
we
have
here
100,
18
or
just
more
than
100
megabytes
as
a
memory
for
free.
Just
remember
this
number,
because
I'm
going
to
show
you
very
fantastic
stuff,
so
go
back
to
here
and
I'm
going
to
stop
and
then
what
are
going
to
be
next
step
to
deploy
this
application
and
suffer
less.
B
So,
first
of
all,
on
top
of
the
kubernetes
stuff,
you
need
to
packaging
your
application,
like
a
linux
container,
using
the
docker
cli
or
using
a
zip
command
line,
a
jeep,
a
zip
google,
a
container
tool
or
some
approach
to
s2i
bill.
There
are
multiple
choices
to
containerize
this
application
and
push
this
container
image
in
external
registry.
So
I
already
make
that
happen,
but
there
are
two
options
I
already
mentioned.
You
can
package
your
application
as
a
job
file
or
a
sql
file.
B
If
you
want
to
go
to
palm
xml,
you
can
see
here
profile
native
and
then
in
this
profile,
create
your
application
after
make
the
three
jar
and
then
try
to
create
the
new
executable
file,
just
like
a
container
images.
So
all
necessary
dependency
is
already
packaging
in
this
one
single
executable
file.
That's
why
the
packaging
time
takes
a
little
bit
longer
than
the
general
maven
packaging
create
the
finger,
and
but
it's
not
too
much
long.
B
Maybe
just
a
few
minutes
but
depends
on
definitely
depends
on
the
size,
the
how
the
complexity,
your
application
you
developed,
and
once
you
create
the
new
executable
file,
you
don't
need
to
jvm
to
run
this
application
any
longer.
B
So
now,
after
that,
we're
going
to
containerize
this
application
using
docker,
sierra
etc,
but
I'm
going
to
skip
that
thing.
Just
save
the
time
today,
because
I'm
already
packing
this
application
using
docker
cli
and
I'm
gonna
push
that
img,
my
docker
home
and
create.io
or
as
external
containeristic.
B
So
we
just
finish
that
and
go
to
just
target
directory
again
we
have
the
new
executable
file
here
now
how
to
run
it
just
target
and
just
run
file
like
a
shared
script
or
a
batch
file
in
a
windows
system
like
a
go
program,
languages
and
your
linux
operating
system.
And
now
we
can
see
you
know
one
thing
awesome
thing
is
we
just
need
25
milliseconds
to
start
off
exactly
same
application
when
you
start
when
you
try
to
start
off.
Second
time,
it'll
be
a
little
bit
faster,
so
20
milliseconds
to
start
off
this
application.
B
So,
as
you
see,
I
told
you
just
pulling
a
second
is
really
faster
than
any
other
java
application
stack,
but
now
we
have
20
milliseconds,
which
means
almost
maybe
50
times
less
than
existing
application
java
application.
What
about
the
memory
proof
ring
so
go
back
to
here
and
then
try
to
find
the
runner.
B
Now
we
have
8.2
megabytes
previously
same
application:
100
training,
but
now
80
is
almost
25
yeah,
almost
the
12
times
less
than
memory
for
free
and
the
story,
time
20
times
faster
than
startup.
So
this
is
the
reason
why
the
partners
make
the
change
the
truth
of
the
java
and
then
more
more
java
developer
have
a
choice
to
develop
microsoft's
application
based
on
quarkx
and
then
just
change
it
to
then
microservice
application
as
a
function
and
then
deploy
multiple
serverless
platform.
B
B
So
here
is
my
kubernetes
cluster
based
on
operation
container
platform
and
let's
write
a
new
project
here.
Let's
say
hybridized
suburbs,
something
like
that
and
then
I'm
gonna
pull
down
my
container
image.
B
I
already
push
that
thing
just
bubble
this
webinar
and
then
one
thing
you
need
to
do:
deploy
this
application
as
serverless
on
operation
container
platform.
Just
click
on
k,
name
service:
this
is
automatically
deploy
your
microsoft
application
as
k
name
service,
which
means
just
like
you
have
serverless
capability
and
once
you
click
on
that
and
then
your
application
is
already
almost
running
and
you
go
to
raw
okay.
I
need
to
do
60
milliseconds
to
start
off,
go
back
to
the
topology
view.
B
Okay,
already,
your
container
is
already
oh
ooh
super
fast
and
click
on
the
the
route
url
and
you
can
see
the
right,
the
landing
page
or
the
quarkx.
You
can
find
some
resources
how
to
get
started,
how
to
set
up
your
idt
id
tool
and
let's
try
to
access
amp
in
hello,
and
then
we
gotta
the
same
result
that
we
got
at
the
end
user
and
also
you
can
maybe
differentiate,
because
in
reality
you
can
deploy
multiple
pog
using
multiple
runtime
and
then
maybe
sometimes
okay.
B
I
need
to
make
him
differentiate.
My
runtime
need
to
find
out
easily.
Where
is
my
focus
application?
So
here's
a
you
can
add.
Actually,
a
label
here
so,
for
example,
f,
openshift,
dot,
io,
slash,
runtime,
equal
corpus,
and
now
you
can
see
this
is
a
quickest
application
and
then
this
application
automatically
goes
down.
I
mean
scale
down
to
zero,
just
like
always
the
expected
subless
capability
in
30
seconds.
This
is
a
default
time.
Maybe
you
can
change
that
the
configuration,
maybe
one
minute
with
20
seconds
in
one
second.
B
B
Here,
oh
maybe
I
needed
to
do
one
more
thing,
but
for
moving
on
the
produce
application
so
because
in
my
application,
so
I'm
going
to
change
this
application
as
a
function,
because
I
already
packaging
that
feature.
B
So
let's
try
to
do
that
so
in
order
to
make
it
this
application
as
a
function,
so
I
needed
to
add
a
new
extension
here.
Just
like
you
mentioned
the
phone
key.
There
are
multiple
funky
extensions,
funky,
hdd
binding
and
I'm
going
to
add
a
new
extension
here
and
then
I'm
going
to
run
quickest
application
once
again
here.
B
And
one
thing
I
need
to
that:
adjust
using
a
phone
key
phone,
cue,
annotation
here.
So,
for
example,
I'm
going
to
delete
here
a
new
annotation
on
queue
and
when
you
use
a
function
annotation
your
the
function
name
is
will
will
be
same
as
your
message
name.
So
in
this
case
your
function
name
is
hello,
but
sometimes
okay.
I
need
to
define
a
name
when
you
use
a
function,
so
you
can
actually
do
that.
B
Like
a
name
like
a
name,
let's
say
hybrid,
hybrid
and
also
you
don't
need
to
get
measured
because
the
function
use
the
pulse
method
by
default
and
automatically
map
to
your
parameter
parameter.
And
then
you
don't
need
to
pass
any
longer
and
I
needed
to
delete
a
needed
to
packaging
file
here,
all
right.
B
So
pretty
simple:
even
your
application,
less
code
is
pretty
simple
and
then
I
just
try
to
access
endpoint
to
make
sure
like
a
hello
endpoint
here
so
now,
you're
saying
ambassador,
and
then
you
need
to
try
to
endpoint
you
and
you
function
like
a
hybrid,
so
mine
is
then
o
and
then
pass
down
this.
The
parameter
name
and
the
endpoint
is
a
hybrid.
B
And
now
we
have
the
wizard:
hey:
welcome
daniel
hybrid
serverless
development
using
focus,
so
I
just
exactly
a
packaging
application
with
this
snapshot
and
pushing
my
container
registry.
So
we
already
deployed
application
on
the
kubernetes
cluster,
such
as
operation
container
platform
and
go
back
to
here,
I'm
going
to
copy
log
url
and
then
change
it.
This
endpoint.
A
B
Besides
my
application,
all
right,
so
I'm
ready
and
now
let's
try
to
invoke
endpoint
and
then
what
happened
next
and
once
you
run
your
application.
As
you
see,
your
policy
automatically
goes
up
super
fast,
just
like
one
second,
and
then
you
got
the
same
measure
here.
B
Welcome
dental
hybrid
stabilization,
same
research,
so
just
imagine
some
application
can
be
a
startup,
maybe
five
seconds
or
even
ten
seconds,
so
that
application
definitely
doesn't
fit
in
a
suburban
application,
though,
just
like
the
the
new
release
of
a
result,
that's
why
a
lot
of
developers
really
prefer
to
use
no
js
go
super
fast
startup,
but
now
quarters
totally
gain
changes.
B
So
now
using
quarkx
application
on
kubernetes
and
serverless
environment,
you
can
have
the
same
benefit
but
and
but
also
you
can
still
develop
the
same
java
interface,
which
means
you
don't
need
to
run
any
new
technology.
Okay,
I
need
to
run.
I
don't
need
to
run
node.js
or
go
or
python
to
develop
sync,
which
is
a
functionality
for
serverless,
still
same
application,
but
change
that.
So
I'm
going
to
show
you
one
more
thing
today.
B
And
then
back
to
here
and
I'm
going
to
close
the
terminal
window
and
then
start
my
dev
environment.
So
let's
try
to
deploy
the
same
application
without
code
change
into
another
serverless
platform
such
as
amazon
render.
So
in
order
to
do
that,
let's
try
to
add
a
new
extension
quark
extension
here
and
a
funky
extension.
B
You
could
see
here
the
amazon
render
extension
and
then
you
can
just
add
that
thing
and
automatically
will
be
pulled
down
necessary
dependency
on
your
maven
project
and
then
just
one
thing
you
need
to
do
to
deploy
amazon
render
you
need
to
define
your
specific
function
name
here,
so
you
know
that
the
quarkx
funky
export.
You
know,
let's
try
to
hybrid
the
function
name
here
and
then
let's
try
to
maven
packagey
and
I'm
going
to
skip
unit
test
here.
B
If
you
already
have
some
experience
to
deploy
your
java
application
to
amazon
render
you
need
to
use
some
command,
the
amazon
command
line
and
some
configuration
etc.
So
once
you
pack
in
your
application
and
go
to
target
directory,
all
you
need
to
deploy
this
application
into
amazon
render
it's
already
there
like
a
the
bash
screen.
All
including
the
amazon
render
command
line
like
amazon,
lambda,
etc,
and
also,
if
you
want
to
test
this
application
on
your
local
machine
and
the
device
yemo
file
on
the
based
on
the
same
command
right
here.
B
So
I'm
not
going
to
test
my
local
machine.
I
just
try
to
deploy
amazon
lambda
directly.
So
this
is
my.
The
amazon
render
the
function
and
then
there's
no
the
function
here,
but
in
order
to
deploy
amazon
lambda,
you
need
to
create
your
own,
the
low
file,
to
make
sure
to
do
that.
Just
make
sure.
Okay,
I
already
created
amazon
download
here
and
I
just
need
to
load
resource
name
to
deploy
my
application
as
a
function
and
go
back
to
render
and
there's
no
function.
B
Name:
okay,
just
remove
using
the
menu
the
batch
script
here.
Just
define
your
resource
name
here,
and
this
is
my
resource
name
and
go
back
to
here,
and
I
just
run
manage
this
grid
file
and
create
it
takes
a
couple
of
seconds
to
deploy
your
application
in
general.
And
then
you
can
find
the
result
here.
Just
in
a
few
seconds
here.
B
Okay,
we
got
a
successful
return
and
go
back
to
amazon
and
the
console,
and
then
now
you
go
to
the
hybrid
cloud
is
already
deployed
and
let's
try
to
let's
try
to
test
this
application
so
configure
application.
The
name
is
screening
such
like
that
and
I'm
gonna
be
changing.
My
name
like
like
a
daniel
all
and
save
file
test
it
and
then,
once
you
deploy
now,
you
can
see
the
same
result
for
different
parameter
variable
here.
B
So
this
is
a
really
a
cool
way
and
once
you
testing
what
even
you
don't
need
to
any
longer
this
application,
you
just
need
to
delete
the
application
with
the
same
command
line
and
go
back
here.
It's
already
gone
and
if
you
change
your
for
example,
okay,
I'm
gonna
change,
different
function,
name
and
you
need
to
maybe
package
it
once
again,
but
in
the
meantime
your
menus
bash
script
is
already
deleted
and
regenerated.
B
So
you
can
also
add
your
resource
name
once
again
or
you
can
just
the
add
the
resource
name
as
a
parameter
in
your
manager.
When
you
run
the
base
the
basis
click
file
here.
B
Seconds,
okay,
we
got
here
successful
and
just
reload
application,
so
the
application
function
is
deployed
and
the
creator
test
event
record.
Just
hello
and
there's
no
parameter
actually
so
just
create
that
thing
and
test
and
then
process
this
event.
And
now
we
got
a
here
welcome
cncf
webinar.
B
So
now
I
don't
need
to
this
application
any
longer,
which
means
you
just
delete,
save
your
money
and
function
here.
The
desk
go
yeah,
so
this
is
the
I
prepared
for
the
today
demo,
I'm
gonna
just
recap:
the
today
webinar,
so
redhead
runtimes
actually
fully
support
the
carcass
environment
as
well
as
multiple
cloud,
every
long
time
like
spring
boot,
node.js
and
and
etc,
and
also
there
are
memory
in
memory,
data
grid
and
a
messaging
broker
and
a
single
sign-on.
B
This
is
all
piece
of
your
cloud:
every
application
architecture,
a
few
more
interesting
about
the
quarkx
environment.
Here
is
the
couple
more
url.
So
there
are
idc
quarkus
study
the
idc
published
this
report.
You
can
find
that
some
performance
comparison
between
parker's
caucus
and
quarkx
and
the
difference
javascript
and
also
there
are
bin
url
try
to
focus.
You
can
just
go
to
this
url
with
your
web
browser
and
then
there
are
interactive
sample
experience
engine
running
portal.
There
are
lots
of
scenario
you
can
run
it.
B
Maybe
10
million
and
50
million
for
each
scenario
in
a
project
generator
called
our
focus
io.
You
can
actually
use
the
focus
tool,
just
like
I
did
in
my
vs
code,
but
also
you
can
go
to
this
website
and
then
just
click
on
whatever
you
need,
as
it
dependency
or
extension,
and
just
click
on
that
and
generate
your
project,
and
you
can
just
download
g
file
and
extract
that
file
and
just
import
your
id
tool,
all
right,
yeah.
I
think
it's
I'm
done.
B
You
know
if
if
there
are
any
questions
and
it's
your
time,
ask
me
or
you
can
just
follow
my
twitter
or
subscribe,
my
youtube
channel
and
just
add
your
comment
and
question
whatever
you
are
interesting
about
this
agenda.
A
Thank
you
very
much
daniel
for
that
wonderful
presentation.
We
are
out
of
time,
unfortunately,
for
questions.
Thank
you
very
much
for
your
presentation
today
and
for
everyone
for
attending
have
a
wonderful
rest
of
your
day
and
have
a
wonderful
weekend.