►
Description
JakartaOne Livestream is a one-day virtual conference for developers and technical business leaders that brings insights into the current state and future of Jakarta EE and related technologies focused on developing cloud-native technologies for enterprise Java applications.
Featuring Speaker: Lars Röwekamp
04:00-04:45 GMT-04
A
B
So
that
might
be
a
topic
later
in
the
interview
with
jürgen
sein.
That's
pretty
good
about
me.
I
come
from
the
company
open
knowledge,
that's
how
we
work
and
let's
get
started
right
away.
The
whole
world
is
up
to
date
according
to
native.
Yes,
we
also
have
our
own
conference
here
at
the
start,
and
enterprise
server
would
of
course
like
to
join
in.
B
So
we
want
to
be
part
of
the
party,
but
if
you're
really
honest,
then
at
least,
if
you've
been
involved
with
java
ee
for
a
little
longer,
then
or
kde,
then
it
seems
somehow
a
bit
very
heavy
weight.
The
application
server,
especially
all
the
libraries,
the
part
of
it
and
somehow
always
trying
but
a
bit
out
of
date
and
what
I
mean
out
of
date.
So
you
have
the
feeling,
although
a
lot
has
happened
in
the
specifications
in
the
last
few
years.
B
Http
two
sport
jason
at
ninja,
is
a
couple
of
76
it
with
reactive
clients,
so
there
it
is
so
contemporary
in
terms
of
developments.
But
if
I
do
n't
always
talk
about
it.
I
mean
something
like
that.
Well,,
that's
the
world.
We
currently
live
in,,
at
least
in
the
cloud
cities
clusters
that
somehow
dynamically
raise
pots
and
a
very
fast-moving,,
very
dynamic
world
where
you
just
start
up,
containers
be
restarted
and
you
have
the
feeling
that
it
doesn’t
really
work
with
enterprise
java,.
So
with
jacques
attali.
B
The
need
that
we
have
is
that
for
small
units
need
for
the
small
ones
in
the
sense
of
lower
memory.
With
may
so,
around
memory
needs
is
just
too
expensive
and
is
from
black
states.
So
we
need
something
that
is
fast
low,
startup
ze
iten
has
to
get
this
dynamic
in
there
and
we
need.
We
have
something
flexible
is
in
the
sense
of
modularization,
and
if
we
look
at
jacques
attali,
then
it
turns
out
to
be
exactly
the
opposite.
That's
great!
B
If
I
take
the
application
server,
plus
my
application,
I'm
fast
at
a
few
hundred
mb,
it's
slow
in
terms
of
start-up
time.
So
probably
the
server,
as
well
as
the
deployment
which
is
partly
in
the
nature
of
java,
that
I
have
to
add
a
bit
and
it's
relatively
inflexible
in
the
sense
of
all
or
nothing
so
either.
B
I
have
my
pros
that
are
defined
web
profile
or
the
full
pro,
and
I
just
have
to
use
all
apps
if
I
want
to
be
compatible
and
the
web
profile
wasn't
the
first
step
that
you
make
a
bit
smaller.
But
who
cares
about
that?
Look
at
the
web
profile,
so
half
intended
for
classic
web
applications
who
looks
at
the
web
profile
is
still
quite
powerful
because
ko
just
got
a
news
in
and
currently
jacques
attali
10.
B
The
next
generation
which
is
reading
prix
today
and
will
not
be
released,
will
not
cause
us
to
mess
something
up
right
at
this
point
attacked
and
said.
We
need
something
smaller,
which
is
much
better
suited
to
something
like
that,
how
to
build
microspheres
and,
of
course,
now
what
we
understand
as
a
micro
service
in
terms
of
size,
but
something
smaller
than
the
web
profile
and
has
specified
the
core
per
case
and
in
the
core
pro
case,
is
actually
the
basis
in
our
service
to
build
something
like
jacques.
B
It
just
to
build
up
the
interfaces
dax,
I.e.
the
rest
of
the
interfaces,
jason,
binding,
jason
pasing
to
be
able
to
deal
with
jason,
somehow
and
then,
of
course,
city,
but
pull
the
12
light
variant.
There
say
a
bit
later
what
you
so
not
quite
as
powerful
as
the
full
city
.
There
are
a
few
other
important
ones
that
I
just
mentioned,
be
those
are
the
essentials
to
expand
restful
services.
B
So
if
we
look
at
the
map
now
it
doesn't
seem
to
be
that
big
anymore,
so
the
problem
of
size,
you
get
a
handle
on
somehow,
but
we
still
have
this
application
server
problems.
I
really
don't
want
to
have
it.
I
want
to
have
such
a
self-living,
self-sufficient
micro
service,
and
it
also
applies
to
the
application
server
that
I
don't
actually
need
it
at
all,.
So
if
you
look
at
the
specification,
it
doesn't
say
anywhere
that
it
has
to
give
an
application
server.
B
That
represents
its
own
box
because
it
is
planned
in
rom
and
brings
certain
things
with
it
and
since
something
is
packed
or
several
things
can
be
packed
in.
That
is
relatively
vaguely
described
in
the
specification,
and
if
you
want
to
do
yourself,
a
favor,
a
little
deeper
just
want
to
look
at
this
post
read
the
principle
says
the
specifikati.
Actually
only
what
this
server
has
to
be
able
to
do
how
it
does.
It
is
up
to
the
server
itself,
and
there
are
things
that
are
definitely
possible.
B
What
makes
up
the
server
together
with
its
own
application
as
a
year,,
I.e.
with
its
own
application
logic
as
a
year,,
is
also
often
referred
to
as
an
ambassador
or
something
else,
and
the
bankruptcy,
then
as
a
standalone,
running
application,
and
that
is
what
you
get
from
the
white
rock
profi.
But
there
is
also
this
possibility
for
almost
every
application,
server
manufacturer
and
then
I
no
longer
have
this
problem
with
the
application
server
that
someone
always
has
to
clear
the
vip
and
have
to
wait
for
him,.
So
it
seems
it
is
with
the
micro
service.
B
Yes,
it
is
really
cool
to
give
so
on
the
picture
here,
three
micro
services,
the
al
le,
live
in
their
own
jvm,,
which
means
they
do
n't
interfere
with
each
other,
and
everyone
has
their
own
business
logic
and
their
own
part
of
the
application
server..
It
would
be
nice
if
it
were
that
easy
in
practice,,
but
unfortunately
that's
not.
B
The
problem
with
these
matches
is
that
I
am
alone
there,
but
that
there
are
several
of
them
and
because
there
are
several,
there
is
also
communication
between
these
accession
services
and
I,
somehow
had
to
take
care
of
the
fact
that
I
get
Syria
patterns,
structure
and
error.
Tolerance
structure
I
have
to
security
over
these
micro
services
propagate
across.
So
I
don't
want
to
have
to
explicitly
register
with
everyone
as
a
user,
but
somehow
there
has
to
be
a
mechanism
for
them
to
exchange
their
security
information.
B
So
I
would
like
to
somehow
be
able
to
know
what
I
have
built
up
over
20
years
and
also
know
what
the
community
can
do
and
who
use
the
many
positive
experiences
and
everything
that
has
gone
into
the
apps
and
their
development
and
do
n't
have
to
reinvent
the
wheel
just
because
I'm
suddenly
driving
a
new
architectural
style,,
so
I
could
of
course,
also
say
I'm
doing
something
different,
but
on
the
enterprise
principle
java
is
simple,
stable,
extremely
stable
is
not
come
t.
It
is
no
coincidence
that
all
large
companies
rely
on.
B
It
is
very
widespread,,
so
the
knowledge
is
also
very
widespread.
There
are
simply
people
who
are
familiar
with
it
and
who
know
the
problem,
and
it
is
standardized
now
the
ejot.
The
academy
stands
for
enterprise,.
It
is
no
coincidence
that
it
is
simple,
is
designed
for
large
systems,
and
it
doesn't
really
matter
if
they
are
distributed
or
not,,
but
what
we
just
saw
in
the
picture
is
still
missing
for
the
world
of
dispute
runtimes
and
what
is
written,
there,
that
's,
something
like
monitoring
distributed
is
monitors,
namely
health
checks.
Metrics.
B
These
are
things
that
the
application
server
has
previously
taken
over
for
itself
for
the
monolith
in
a
central
location.
Suddenly
it
is
distributed.
I
have
to
take
care
of
it,
something
like
distributed
racing,
I.e.
a
common
tracy
that
goes
beyond
a
use
case,
and
they
all
use
these
tracy
die
and
somehow
pump
their
information
into
one
mean
seed
checkers
distributed
configuration
management.
I
have
configurations
that
are
to
be
shared
across
multiple
services
and
I
need
a
mechanism,
a
central
office
where
the
hotel
industry
we
just
talked
about
resilience
patterns.
B
What
happens
when
a
service
may
not
be
available,
namely
that
multiple
attempts
are
automatically
made
with
the?
If
that
doesn't
turn
out
to
be
a
success,
I
want
to
address
a
service
other
than
Frauenberg
and,
of
course,
security
also
just
talked
about.
Then.
I
would
like
to
be
able
to
propagate
this
security
context
through
the
various
services
through
all
that
is
missing
in
jakarta
and
the
hands
devices
has
already
said.
Then
too,
there
is
something
else
and
this
clips
community,
namely
the
motto
that
is
actually
only
the
best
addition.
B
If
I
want
to
go
towards
the
microserver
system
originally
a
bit,
I
don't
want
to
say
it
came
about
as
a
countermovement,
but
as
a
compromise,
because
the
g
Anze
java
ee
world
not
progressed
so
quickly.
I
consciously
used
java
ee
because
years
ago
it
didn't
progress
quite
so
quickly
and
above
all,
not
this
focus
on
macho
services
has
meanwhile
merged
the
whole
thing
very
closely.
And
how
do
you
have
to
imagine
that
my
profile?
Five?
B
You
can
see
a
couple
of
the
1
here
and
the
interesting
thing
is
there
is
exactly
the
ip
or
exact
price
for
the
problem
cases
I
just
listed,
so
we
see
something
like
open,
tracing
ip
or
open,
die
conficker
or
photo
lorenz,
eklat
or
david
popper
games,
so
the
provocation
11.
This
is
james,
webb
dead,
holds
and
photos
and
below.
So
these
are
exactly
the
points
we
just
saw.
B
I
saw
I
got
it
a
bit
wrong
one
and
hidden
here,
and
that
other
is
interesting
if
let's
take
a
look
at
the
lower
ones,
for
example,,
these
are
components
from
the
environment,
well,
that
I
have
in
the
environment
anyway
and
where
it
is
in
the
micro
profile
is
called,
and
we
also
need
it
in
this
market,
per
case.
You
don't
see
cdi,.
I
just
saw
that
I
put
it
in
wrong
and
hidden
it,,
so
cdu
would
also
be
in
there,
and
if
you
take
a
closer
look,
then
that's
it.
B
What
in
jakarta
roughly
does
not
agree
with
every
corner,
but
roughly
what
was
declared
as
core
profile
file
in
jakarta
the
ten
and
it
is
relatively
certain,,
at
least
according
to
discussions.
I
had
yesterday,
that
in
the
next
generation
of
mentor
it
will
be
micro,
profi
60
is
said,
and
that
is
based
on
the
core
pro
case.
So
then
I
actually
only
have
this
vision
that
we
saw
in
the
fire.
So
this
one
here
and
plus
core
pro
case,
and
that
is
a
very
important
element
so
that
the
two
just
work
much
more
closely
together.
B
B
So,
let's
assume
the
service
has
already
started,
but
has
to
initialize
a
cache
or
provide
a
connection
to
a
system,,
then
that's
sorry,
but
not
yet
trained,
and
you
use
that,,
for
example,
in
this
climate
nice
world
we
just
saw
in
the
health
checks
against
that
system.
So
then
I
can
specify
a
whole
health
and
point
here
and
then
I
could
say.
Please
look
at
the
container
after
that,,
I.e.
B
a
double
container,
where
the
whole
thing
runs
out
for
the
following
health
check
and
I
know
the
services
are
alive
or
may
have
to
be
new,
be
started
because
after
three
health
checks
there
is
still
no
positive
answer.
Jackson
stops
him
to
simply
determine
you
are
fine.
Are
you
not
well?
So
it
is
a
very
binary
answer
or
seminars
decision.
I
often
don’t
want
to
know
but
I’m
already
against
them
wall,
but
I
want
to
proactively
take
countermeasures,.
B
User
return,
in
my
use
case
and
above
this
tankt
annotation,
from
line
4
to
line
8
and
the
shares
annotations,
simply
indicate
that
a
magic
should
be
automatically
generated
here
and
it
should,
so
to
speak,.
Look
at
how
long
a
run
of
the
call
is
running
and
they
are
then
aggregated,,
which
means
not
only
this
one
run
is
measured,,
but
the
run
is
related
to
measured
the
others
and
then
I
can
monitor
at
any
time
whether
my
calls
last
longer
how
long
the
average
call
lasts.
B
How
long
25
percent
of
the
calls
last
whether
there
are
any
super
exceptions
in
the
negative
or
in
the
positive
case,
and
then
my
infrastructure,
then
automatically,,
e.g.
up
or
down
scale.
So
if
I
notice
the
colts
are
taking
too
long
or
there
are
too
many
calls,
rhein
can
also
add
a
counter.
Then
I
simply
escalate
my
infrastructure
automatically
up
again
via
the
picture
we
just
saw
about
cuba,
nice
or
somehow
in
the
cloud.
Other
men
must
also
have
a
very
important
photo.
Lorenz
So,
it's
a
topic,
so
to
speak,.
B
Implementing
resilience
patterns
can
be
done
in
the
application,,
but
you
can
also
do
it
via
the
coup
of
the
cluster,
for
example.
The.
Basic
idea
here
is
I
determine
a
communication
partner
with
whom
I
work,
so
another
service
king
calls
who
doesn't
react
and
then
steps
in
system
does
not
react,
and
you
can
see
that
in
line
3
and
4
there
I
look,
so
to
speak,.
He
reacts.
So
there
is
a
time-out,
yes
or
no,
and
if
there
is
a
place
think
in
advance.
What
I
have
decoratively
indicated
here,
so
a
different
method.
B
In
my
case,
would
the
method
fall
back
on
a
cache,
I.e.
data
that
was
served
up?
I
can
also
do
something
completely
different.
I
can
specify
classes
so
in
my
case,
I
have
a
method
with
which
I
can
specify
the
complete
class.
That
does
something,
and
there
are
various
others
photo
lorenz
pattern,
so
parquet
and
otherwise
braker
and
other.
So,
like
ent
speaking
puberty
libraries
knows
what
exactly
do
we
have
last
thing?
I
would
like
to
go
into
again
about
security,
so
there
is
much
much
more.
B
I
just
want
us
to
get
an
impression
of
how
the
whole
thing
is
structured,
namely
usually
based
on
annotations,
and
at
least
it
fits
for
me.
It
was
a
relatively
seamless
transition
into
the
ee
jakarta
world
and
you
don't
have
the
feeling
when
you
work
with
the
micro
profile
that
you're
suddenly
moving
in
a
completely
different
world
or
working
with
a
completely
different
framework,.
But
it's
a
really
well-
coordinated
addition,
and
you
can
see
that
particularly
here
with
the
topic
of
security
and
context,
propagates
that
here
the
line.
B
Now
I
have
to
do
lines
myself
there.
I
let
myself
a
job
dead
insects,
so
to
speak,
so
how
we
assumed
the
call
that
ran
is
with
the
car
recession
harder
with
the
barry
atom.
So
with
the
jogging
called
I.
Let
the
insects
here
a
single
line
and
then
nine
and
ten
actually
two.
If
you
write
it
and
then
I
have
access
to
the
whole
token,
so
to
speak,,
or
I
can
also
have
individual
frames
issued.
That
is
lines
12
to
14
he'll.
B
Tell
you
I
read
the
claim
with
the
insects
right
away
as
a
type,
so
I
can
also
typify
the
whole
thing
accordingly
and
access
it.
If
I
use
the
microphone
in
the
world
user,
I.e.
in
a
context,
then
this
claim
is
automatically
taken
in
the
background
and
the
roles.
So
there
is
not
the
small
one
but
jogging
taken,.
B
There
are
certain
fixed
plans
that
have
to
be
filled,,
something
like
bobst,
for
example,,
and
then
these
groups
are
taken
and
the
whole
thing
is
automatically
translated
into
the
marriage
world
and
I
can,
for
example,
use
the
blinds
according
to
the
annotation,,
so
that
's
a
pure
e
annotation.
It
has
nothing
to
do
with
micro
profile
and
the
translation
takes
place
under
the
hood
and
that's
really
nice
r,
very
good
integration.
At
the
point.
B
Good,
what
does
that
mean
at
runtime?
You
can
see
it
here,
so
I
started
the
whole
thing
and
then
you
see
something
comes
out
so
a
stand
alone,
micro
service-
that
just
has
this
micro
profile
in
it,
so
the
the
the
libraries
this
requires
it
would
be
like
taking
this
corps
profile
from
jakarta
later
+,
the
micro
professionals,
something
comes
out
about
80,
mb
and
a
start
time
of
1600
milliseconds.
That’S
good.
Is
that
not
good
it’s?
B
Definitely
better,
isn’t
it
faster
and
smaller
than
if
I
would
take
spring
in
or
take
an
application
server
in
and
transform
the
whole
thing
with
it,.
But
if
we
look
again
at
this
cloud-
native
world
with
juventus
and
the
containers
just
fly
around
there,,
so
scale,,
it
just
seems
so
through
the
area
that
still
isn't
the
best
option,.
Let
me
give
you
numbers
for
comparison,.
B
What
have
we
achieved
now,
so
we
are
always
on
the
subject
of
micro
services
or
building
something
that
can
survive
well
in
the
cloud
nice
field
with
jacques
attali
and
macro
profile
and
in
this
case
micro
services,
and
what
have
so
far
we
have
achieved
on
the
plus
side.
Certainly
we
move
in
the
standard.
That
is
always
good.
I
think
at
least
we
don't
have
to
work
with
any
of
their
apps.
We
form
the
micro
and
the
macro
architecture
side,
so
to
speak,
micros
what
runs
within
a
service,
I.e.
B
what
I
do
there
and
macro
were
these
overarching
things
when
services
communicate
with
each
other,,
for
example,
like
tracey,
I
w
ie
long
I
comprehensive
that
this
security
magazine
I,
think
of
both
and
also
on
the
plus
side.
What
comes
out,
how
narrow,
fast
and
flexible,
relatively
seen,
so
we
just
heard
numbers
from
proprietary
foreign
are
then
clearly
ahead
are
relative
to
the
pulse:
six,
certainly
it
But.
B
It
is
also
the
case
that
the
more
I
take
from
this
circle,,
the
less
narrow
and
fast
the
whole
thing
becomes,
and
also
the
smaller
disadvantage
is
that
boost
wrapping,,
so
I
said:
I
move
in
the
standard
of
the
boot
trapping
is
not
standard,
but
that
only
as
a
comment,
ok,
small
interim
conclusion:
what
have
we
achieved
so
far?
We
have
achieved.
We
come
from
heavyweight
and
a
bit
slow,
but
super
stable,
classic
enterprise
business.
B
I
the
players
at
some
point
and
that
should
run
for
a
quarter
or
so
that
also
runs
smoothly
for
a
quarter
and
are
now
thanks
to
jacques
attali
and
micro
pro
case,
since
that
we
are
using
zeiss
artifacts,
the
can
be
started
relatively
quickly.
That
would
be
a
scenario
where
you
say
I
might
have
daily
deployments
weekly
the
dream,
and
then
it
should
run
reasonably
stably.
So
I
don't
have
the
situation
that
somehow
an
instance
is
permanently
added.
B
B
If
we
had
the
chance
to
become
much
faster
and
smaller
at
all,
and
to
understand
whether
that
is
possible
and
if
yes,,
how
it
is
possible,,
you
have
to
keep
the
actual
problem
in
mind
that
actually
a
problem
is
not
the
packages
are
somehow
too
big
or
so,
but
the
real
problem
is
that
the
start
time
takes
too
long.
I
have
something
and
it
starts,
and
these
states
just
take
way
too
long.
If
I
want
to
start
new
services
regularly,
we
don't
want
to
lose
and
I
escalate,
because
I
have
too
high
a
load.
B
So
I
ca
n't
here,
because
I
have
a
new
version,
the
connecting
rods.
That
was
the
scenario
that
we
have
already
covered
before,,
but
I
would
like
to
escalate
because
I
notice
my
load
is
too
high
and
I
want
now
in
this
millisecond
or
in
the
second
at
least,
to
start
something
that
takes
this
load
off
me
and
then
relatively
easy
to
drive
down
again
later
and
I
once
brought
a
quote.
You
can
read
through
later
in
peace
or
now
it
is
important
to
have
the
boots
locked
on.
B
That
has
suddenly
become
a
very
important
benchmark
in
this
new
world,
and
it
is
just
terrible
in
the
ii
and
my
chocolate
world
in
the
java
world.
You
don't
have
to
say
anything,
and
why
is
that?
I
brought
you
here
that
me,
the
meter
dating
process
for
the
start
up
phase
does
not
need
to
be
looked
at
in
detail,,
nor
understood
in
detail,.
It
should
only
make
it
clear
that
this
is
what
happens
when
you
start
the
planner,.
B
So
only
for
this
start-up
phase
do
I
need
a
relatively
large
amount
of
java,
code,,
a
relatively
large
number
of
java
classes
and
a
java
library,
and
the
question
is
whether
you
can't
somehow
optimize
that
and
that
is
it
the
next
step,
so
to
speak,.
So
I
set
the
course,
for
example,
because
it
is
currently
or
not
only
time
but
really
very
established
in
the
community,
but
there
are
also
other
franconian
frameworks
that
take
a
similar
approach.
B
That
is
exactly
the
approach
of
this
framework
to
say,
we
are
trying
to
attack
and
optimize
at
the
point
of
things
that
normally
go
to
the
detroit
z
eit,
to
be
dissolved
already
to
the
image
time
to
dissolve
and
thereby
simply
throw
out
all
these
necessary
classes
for
the
dynamics.
The
idea
is
to
have
someone
else
with
the
goal
of
cloud
made
it
from
cloud
made.
It
means
to
take
the
container
first
approach,.
That
means
narrow
packages,
narrow
package,
so
just
the
what
should
come
out
will
be
smaller,
because
I
have
smaller
container
images.
B
I
would
like
to
have
faster
boot
times
to
enable
an
immediate
scale
up
and
lower
memory.
Rss
memory
it
wants
sims
adsiz,
I.e.
the
entire
space
that
communicates
through
the
process,
consumes,
is
simply
to
be
able
to
run
more
containers
for
the
same
money
and
we
are
important
to
the
cloud.
The
markets
consume
resources,
the
more
expensive
it
would
be,
and
how
achieved
it
complained
that
krokus
has
set
up
a
few
rules,
so
I
called
them
mobutu
rules.
B
There
are
a
total
of
four
or
five
I'll
put
three
before
the
ever
now
are
relevant
to
our
scenario
and
the
first
rule
is
wild
to
the
people
so
be
like
that,
be
quick
and
narrow
and
quick.
So,
let's
see
if
the
people,
what
does
that
mean?
As
I
just
said,
a
large
part
of
this
dynamic
I'm
going
through
the
annotation
I,
don't
need
it
at
runtime.
That
would
be
classic
reflection,
but
I
actually
only
need
it
at
detroit
time
and
therefore
it's
just
as
possible.
So
at
runtime
it's
static.
B
It's
easily
possible
to
move
wild
time
to
the
image
time
this
here
as
an
example.
So
how
does
it
work?
Normally?
I
have
my
java
code
with
its
the
penalties,
make
comcom
clear
and
that
I
run
at
the
world
championship
and
when
we
looked
at
the
diploma,
so
to
speak,,
which
dependencies
are
there,
and
here
there
are
now
an
additional
optimization
step
and
the
optimization
step
solves
that
before
open
and
you
see
the
property
100
that
should
signal
the
package
which
ends
up
coming
out.
B
Mmt
is
smaller
and
can
then
be
started
correspondingly
faster
at
the
world
championship
and
this
whole
pasing
process,,
I.e.
annotation
pasing
process,,
had
to
take
place.
What
does
that
mean
in
comparison?
We
get
an
artifact
out
remind
you
that
at
some
point,
it’s
somehow
80,
mb
and
1600
milliseconds,
so
it’s
always
up
measured
on
my
computer,.
These
are
not
standardized
benchmarks
or
anything,,
but
it's
now
about
the
comparison,
and
what
comes
out
is
about
twice
as
fast,,
so
it
starts
twice
as
fast
starts
twice
as
fast
means
until
the
first
request
can
be
processed.
B
That's
also
a
bit
tricky
So
that
other
states
are
not
that
slow
either,,
but
until
the
first
class
of
the
proxy
accesses
the
real
implementation,,
which
is
initialized
and
then
the
path
can
be
processed.,
However,,
that's
counted
in
again,.
That
means
8,100.
Milliseconds
is
my
kind
of
fact
only
twelve
kb
big
before
what
80
that
egg
n
bit
tricky.
It's
not
really
12
kb
in
size,
but
almost
20,
so
a
quarter
smaller
than
what
we
just
saw.
It
has
something
to
do
with
the
structure
of
how
the
imk
warkus
is
stored
in
production.
B
Let's
look
at
it
again
in
comparison.
These
are
now
benchmarks
which
I
pulled
from
the
quarks
side.
What
do
I
like
about
that
so
bacchus
with
the
normal
jvm
in
this
case,
open
danny
kaye
versus
for
tuesday
between
tuesday
would
be
spring
spring
spring,
good
or
jakarta,
plus
with
prof,
and
we
see
the
difference
is
that
it
takes
up
about
half
the
space,.
That
is
half
e,
the
size
doesn't
matter
and
if,
if
I
add
additional
libraries,
then.
B
It's
just
fine
in
this
case,
then
it's
not
quite
half
anymore,
but
at
least
25
30
percent.
That
makes
a
lot
of
difference
is
crucial,
but
this
we
said
this
boudgoust
raps
an
is
relatively
important
and
we
see
here
in
both
variants,
one
with
and
one
without,
I.e.
with
database
access
and
without
us
being
significantly
better.
Now,
the
variant
without
a
factor
of
four
is
better
and
the
variant
with
is
still
even
more
than
a
factor
four
knives.
So
that
makes
a
lot
of
difference
at
the
point
you
can
have
it.
B
B
Just
take
a
look
at
the
browser
window
and
then
you
can
find
out
relatively
quickly
at
which
points
this
optimization
has
a
st
found.
So
we
already
have
something
that
is
at
least
half
the
size
or
a
quarter
of
the
size
of
the
fire.
That
starts
relatively
quickly
that
you
can
imagine
a
little
better
in
this
cabinet
is
the
world
where
I
could
not
imagine
a
little
tip.
Who
does
something
like
that?
B
How
server
lets
so,
where
I
put
the
kol
kol
against
such
a
construct,
and
this
construct
then
has
to
start
up
first
and
then
have
to
answer
so
that
means
my
complete
response.
Time
has
to
be
mapped
over
it.
I
could
n't
really
imagine
that
here
in
the
scenario,
but
also
there
are
possibilities,
that's
the
second,
where
you
usually
point
moorer
in
faster
than
what
we
just
saw
and
that's
achieved
via
head
of
time
compilation.
So
he
often
has
a
compilation.
B
Really
means
nothing
other
than
pre-deployment
or
in
a
final
compression
step,
to
translate
the
whole
thing
again
to
a
native
sims.
We
see
that
here.
So
we
h
atten.
This
optimal
code
is
already
step
2
and
instead
that
is
then
sent
to
the
wm,
because
I
clicked
wrong
wrong
direction.
A
native
image
is
generated
again,
so
an
executable,
an
excel
thingy
in
the
windows
environment
or
stop
runaway
accordingly
in
the
linux
environment,
mac
os
environment.
So
I
can
raise
the
target
platform
accordingly.
B
Ok,
if
we
look
at
it
now,
the
25
mb
package
has
become
a
bit
smaller
than
what
we
just
saw,
but
the
start
up
time
is
11
milliseconds.
So
if
you
can
imagine,
I
want
it,
server
server
doing
send
chorus.
On
the
other
hand,
the
processing
time
is,
of
course,
added
plus
these
eleven
milliseconds.
That
is
quite
realistic.
The
motto
is
almost
extremely
fast.
The
whole
thing
we
can
see
here
again
in
comparison.
The
green
is
what
is
new.
B
Now
we
already
had
the
other
two,
so
I
suddenly
only
have
12
mw
in
the
background,
so
that
we,
the
re
alised
I,
have
n't
said
that
yet
about
the
gm,,
the
straight
translates
to
requiem,
and
that
means
the
whole
thing
is
translated
into
native
in
magic,,
so
to
speak..
You,
don't
do
that
all
the
time
as
a
developer,,
so
the
step
takes
a
relatively
long
time,.
That
means
in
development.
You
would
actually
choose
the
variant
that
I
showed
before
and
at
some
point
you
are
from
saxony
rallye.
B
Then
you
translate
that
in
the
native,
so
we
give
about
several
minutes
that
it
takes,
and
here
again
in
comparison,
so
we
have
reached
completely
different
dimensions
here
and
especially
here
fast
food
again
and
first
response.
Here
too,
we
see
that
we
are
suddenly
on
the
move
in
completely
different
areas
than
what
we
had
before,.
B
Ok,
then
we
all
come
from
the
java
world
and
one
of
the
Oktoberfest
highlights
in
the
java
world
was
write
once
run
anywhere
and
we're
losing
that
now
so
now
it
suddenly
means
let's
get
out
of
what's
actually
everywhere
lukewarm
fen
could
suddenly
do
something
active
again,
so
we
both
lose
the
huge
advantage
we
had
in
japan
or
what
we
peddled
for
the
group
that
is
no
longer
relevant
today.
So
it
is
relevant
that
I
have
something
that
amounts
to
preaching.
B
So
if
I
If
I
know
the
target
environment
for
which
I
am
building
it,,
then
I
want
it
to
always
work
there,
and
that
is
why
it
is
possible
today,
because
I
am
moving
in
a
container-
shaved,
world,,
so
I
know
the
target
platform,
and
the
target
platform
is
on
my
computer..
The
container
target
platform
is
on
my
computer
Exactly
the
same
one
that
I
also
have
in
production
and
that's
the
trick,.
So
that's
why
it
makes
sense
to
go
this
route
and
to.
B
Simplify
it
a
bit
again
offers,
according
to
the
character
framework,
android
framework,
but
also
details
that
make
just
take
care
that
different
variants
of
the
images
can
then
be
generated
immediately.
Okay,
what
does
all
this
have
to
do
with
micro,
profil
and
jakarta
in
that
somehow
seems
to
be
something
very
unique
that
we
just
saw.
I
kept
a
bit
secret
and
we'll
get
to
that
in
a
moment
that
I'm
not
easy
can
take.
B
Any
code,
can
take
any
layman
ries
and
then
through
locks
can,,
so
to
speak,
can
optimize
through
this
last
step,
but
they
have
to
fulfill
a
few
basic
conditions
and
basic
conditions
means
they
have
to
be
prepared
for
them
and
in
this
case
redhat
or
they
have
the
quarks
community
and
decisively
made
redhat
relatively
cool
said.
If
we
want
to
be
successful,
then
we
have
to
actually
support
the
whole
zoo
of
libraries.
B
Jawala
million
people
are
so
popular
right
from
the
start,
so
we
have
to
prepare
them
and
then
hand
them
over
to
the
users
and
that's
a
basic
idea.
Industry
standards
as
a
basis
use
with
simultaneous
possibility
of
proprietary
extension.
So
there
is
a
kind
of
extension
mechanism
and
you
can
see
that
here
in
the
picture,
so
that
was
this
optimal
step
and
you
see
a
lot
of
logos,
so
something
like
kafka,
but
about
nice
things
and
co
also
becomes
abs
So.
B
B
One
of
these
logos
is
also
the
micro
case,
so
mike
rupp
incident,
I
can
work
quite
normally
micro
per
case
so
with
special
implementation,
namely
prepared
implementation,
and
then
this
free
optimizations,
followed,
start
f
and
now
think
about
it
again,
micro
profile
and
the
jakarta
of
the
corps
profile.
What
is
to
come?
I
said
there
is
nothing
in
this
basket
profile
but
city,
sorry.
I
said
that
in
passing
and
that
will
be
relevant
later,
and
that
is
exactly
the
decisive
trick,,
which
is
why
the
kkd
core
pro
case.
B
everything
that
has
to
do
with
reflection
and
further
and
is
therefore
ideally
suited
for
this
head
of
time.
Compilation
so
my
benefit
as
soon
as
the
profile
is
out.
It
will
not
be
long
before.
We
can
also
use
that
in
a
circle
or
just
with
our
own
body,
above
it
or
just
as
a
full
micro
case,
then
in
version
6.0,
with
the
corresponding
incident
with
the
basket,
including,
ok,
conclusion.
B
B
Open
up
new
pots,
every
second
and
scale
them
down
again
and
somehow
permanently
form
a
cluster
move,
but
have
a
relatively
stable
picture
in
my
dreams
every
day,
or
maybe
also
in
a
higher
frequency,
but
a
second
doesn’t
matter,
there,,
I.e.
a
second
or
two
seconds.
If
something
holds
up,
and
we
have
also
seen
that,
with
markus
so
markus
and
head
of
time,
compilation
even
smaller
and
no
ch,
much
faster
artefacts
can
be
formed
and
I
could
also
play
along
with
the
cloud
native
service
world,,
for
example.
B
Very
briefly,
I
already
said
at
the
beginning:
we
will
also
talk
about
it.
I
need
that,
so
I
need
this
very
right
picture,,
so
to
speak.
I
'm
back
to
this
one
again
right
picture
or
is
that
in
the
middle
enough
for
me
in
most
cases,
so
in
my
experience,
the
picture
in
the
middle
is
enough.
The
other
one
was
cool
and
nice,
but
uncomplaining
brings
with
it
a
few
restrictions
of
things
that
can't
be
done
with
it.
You
have
to
read
up
on
the
website
there.
B
You
just
see
how
much
trade
that
often
means
for
yourself
and
that's
why
I
ask
this
question,
so
usually
we're
we
don't
have
any
constructs
like
amazon
or
netflix,,
so
we're
not
amazon
or
netflix
or
twitter,
and
what
they
say
is
what
you
have
to
do
does
not
necessarily
apply
to
our
our
world.
So
what
you
see
here
is
a
runtime
view
of
the
netf
lix
user
management,
so
that's
a
service,
so
to
speak,
with
all
its
web
services
and
all
the
nodes
at
runtime,
and
I
don't
know
which
of
you
has
user
management.
B
That
looks
like
that.
I
mean
we
don't
have
a
customer
big
that
complex
or
that's
a
different
view.
That's
amazon,
liveview
the
the
panels,
the
topology
2008.
Well,
that's
twelve!
Fourteen
years
ago
it
looked
like
this:
they
just
looked
which
services
are
running
where
and
how
are
they
connected
to
each
other
and
you
could
still
zoom
in
and
I
predict
them
most
of
you
who
have
me
this
picture,
so
you
are
not
amazon,
twitter
or
netflix
is,
unless
you
are
always
are
right,
but
now
not
from
that
means.
B
Usually
this
average
is
enough
that
we
have
seen
and
is
called
the
average
view,
actually
think
in
containers
right
from
the
start
and
see
how
I
can
optimize
the
container
and
for
a
while,
just
as
m
icro
per
case
tasks
then
said
again:
okay,
we
no
longer
want
the
application
servers
down.
There
were
the
trees
or
closer,
but
we
want
everything
as
a
hit
have
pictures
robots
whatever
you
call
them
and
actually
a
really
cool
thing
that
just
has
a
small
problem.
B
This
feature
is
just
very
big
and
every
time
I
change
a
line
of
code.
So
the
firm
commitment
of
my
application
logic
plus
the
laimer
is,
I
need
+
the
components
of
the
application
service
and
every
time
I
change
a
line
of
code.
I
have
to
stop
rebuild
this
feature
and
pack
new
ones
into
the
container
for
the
container.
It's
just
that
I
can
fall
back
on
the
guests,
because
it's
something
completely
new
for
him.
That
means,
please
say
goodbye
to
patches
works
with
late
containers.
B
You
can,
of
course,
relatively
schma
I
keep
these
layered
containers
in
which
I
thought
about
what
the
space
image
is,
so
to
speak,,
and
you
can
keep
the
whole
thing
even
smaller,
with
so-called
displaced
container
images,
which,,
so
to
speak,
only
take
the
parts
from
the
operating
system
that
you
really
need
to
run
this
service
and
then
for
a
long
time.
You
can
do
an
incredible
amount
of
optimization
with
that.
You
don't
necessarily
have
to
take
that
last
step
with
markus,
but
it's
just
a
good
option
in
the
interaction
in
a
very
short
time.
B
C
B
Yes,
so
I'll
repeat
the
question
again
very
briefly:
how
do
you
see
it
future?
The
frontend
of
java
or
jakarta?
I
can
from
the
experience
with
our
customer
n.
Most
people
talk
about
it
being
relatively
strict
that
they
just
take
a
single
page
application
or
whatever,
or
engine-
based,
the
whole
thing
and
then
via
restful
calls
or
which
or
web
shops
communicate
with
the
pool,
servers
and
events,.
Whichever
mechanism
is
always
nice.
B
Nonetheless,
there
is
still
a
relatively
strong
sap
community
and
if
you
look
at
the
development
of
the
chips
beck,,
which
are
also
always
more
progressive
on
the
positive
side,,
I
spoke
briefly
to
ed
burns
yesterday
because
he
used
to
be
very,
very
strong
in
this
whole
jsf
environment.
On
the
way
it
gave
me
mvc.
B
Meanwhile,
mvz
anchored
in
there
jaycees
as
basic
technology
for
james,
it's
not
just
the
chain
but
disposed
of
from
the
latest
version,
and
you
work
with
face
net
so
to
speak
in
atriga
technology,
which
is
100
percent
tailored
to
josef.
So
just
the
ones
not
so
the
future,
to
put
it
mildly,
josef
wird
e
s.
I
think
I
always
give
for
a
specific
target
group
and
server
now
anyway,
because
I
need
that
as
a
basis
for
communication,
regardless
of
the
visualization.
C
B
Yes,
otherwise,
I
thank
my
says
now
also
in
45
minutes
perfect
to
the
second,
you
can't
expect
more.
Yes,
then
thank
you
for
listening
have
a
great
day.
Anyway.
Great
goals
are
coming
I'm,
looking
forward
to
the
panel
afterwards
and
to
the
pre
release
party,
so
to
the
millions
out
there
at
the
screens,
even
nicer
conference.
We
see
hear
each
other.