►
Description
en Español from OpenShift Commons Gathering Buenos Aires August 2108
Openshift @ Itaú con Martín D'elía
A
I
now
want
to
introduce
a
very
special
guest,
because
Martín
d'Elía
is
the
architecture
manager
of
Banco
Itaú,
and
he
was
the
first
client
in
Argentina
who
opened
the
doors
for
us
to
be
able
to
do
an
openshift
proof
of
concept..
We
worked
with
him
for
a
long
time
and
your
team,
and
it
was
a
very
pleasant
experience,
and
it
is
an
honor
to
have
you
here
today.
Thank
you
very
much.
Martín
welcome.
B
B
First
I
will
tell
you
a
little
about
itaú
so
that
you
have
an
idea
of
the
context
and
taboo
in
argentina.
It
is
rather
rare
and
argentina
is
a
minimarkets
bank
that
has
approximately
500,000
clients
number
2
in
the
market.
Igor,
take
and
has
a
profession
of
1,600
employees
in
itaú
is
a
beast
so
easy
to
forgive.
Brazil
has
five
thousand
branches
and
close
to
60
million
clients.
Banco
itaú
in
brazil
is
bigger
than
the
entire
argentine
financial
market.
B
Together
then,
we
have
that
dichotomy
of
we
are
a
medium-sized
bank
here,
but
our
parent
company,
the
giant,
has
all
the
crest
technologies
the
wave
is
sometimes
playing
with
that
regional
and
local
vision.
It
is
difficult
for
us,
and
that
is
what
made
us
move
a
little
slower,
In,
this
type
of
initiative,
since
2016
or
so,.
We
undertook
a
quite
strong
digital
transformation
program,,
at
least
for
a
bank
of
our
size.
We
are
investing
approximately
30
million
dollars
in
three
years,,
which,
for
our
size,,
our
country
is
quite
very
strong,.
B
B
Just
to
give
you
a
broad
idea
of
what
the
bank
is
like
how
the
wind
from
our
katiba
is
distributed.
We
have
two
large
platforms.
All
our
core
is
at
s
400,000,
series,
basically
cabbage,
and
therefore
we
have
a
plant
quite
interesting
developers
that
we
do
not
have
approximately
150
developers
of
which
60
are
coleros,
so
that
already
puts
us
in
a
particular
situation.
With
these
new
practices
we
are,
if
you
want,
we
are
not
yet
in
the
cloud
world.
But
what
is
the
previous
world?
We
are
quite
good.
Basically,
we
are
100%
virtual.
B
Basically,
we
have
almost
no
physical
servers
except
for
whatever,
but
all
our
infrastructure
is
legal.
We
have
common,
waterfall
classes
drag
whatever
you
know,
we
have.
The
applications
are
deployed
in
ser
services,
basically
cluster.
I
felt
that
john
silver's
a
lot.
You
don't
mention
you
don't
have
a
tiny
0
application
for
each
application,
if
you
don't
have
giant
beasts
of
multiple
costs
and
that
we
support
a
lot
and
many
applications
that
that
is
great
with
cost
savings.
But
when
you
want
speed
and
agility,
it
gets
complicated.
B
We
have
some
automation,
but
not
so
much.
We
have
some
integrated
things
like
we
are
a
country
eminently
waterfall.
There
is
a
lot
of
manual
testing.
There
is
a
lot
of
regression.
There
are
a
lot
of
tests,
nothing
new
for
everyone.
What
are
the
problems,
because
haiti
is
starting
to
creak
with
the
demand
that
is
forcing
it
right
now.
Basically,
we
have
many
projects.
We
have
more
than
120
parallel
projects
at
any
time
in
large
and
small
projects.
B
That
makes
us
have
a
team
development
playing
various
pieces
of
the
same
application
for
different
projects,
beard
with
the
branding
all
very
nice
ssm.
It
reached
a
point
where
the
environments
begin
to
stop
and
What
release.
I
put
that
now,
according
to
regulations,
touch.
Take
out
of
the
code
smell
another
one
at
the
moment
that
this
is
stopped.
Lisa
has
reached
an
extreme
that
the
way
of
doing
things,,
even
though
it
is
efficient,,
then
a
stopper
for
the
laws
of
business,.
Then
this
is
already
It
is
not
Haiti,.
B
We
have
a
high
rate
of
conflict
between
basically
how
our
structure
is
all
virtual,.
Everything
that
is
configuration
changes
are
manual,,
administrator
going
and
touching
server
parameters.
As
it
happens,
they
are
not
scarce.
That
makes
the
state
of
understanding
of
this
server
increasingly
distant
from
what
initially
the
omo
environments
are
different
from
those
of
the
cross.
This
trajectory
known
to
us
is
brought
up.
That
starts
to
hit
the
group
performance.
B
B
B
Obviously
our
idea
is
to
start
having
new
practices
and
some
methodologies
if
you
want,
on
a
new
platform,,
but
we
believe
that
this
permeates
the
rest
of
the
organization
and
also
the
culture,
as
some
of
the
speakers
said.
Previously.
The
most
difficult
thing
is
not
the
technology
but
to
change
the
mainz
on
the
200
de
belo,
because
they
are
used
to
doing
the
same
thing
every
day.
They
are
looked
at
in
the
same
way
and
have
the
same
tools
and
for
years.
B
So
the
challenge
and
the
approach
is
It
is
difficult
and
it
is
in
Tucuman
where
it
is
most
difficult
to
think
more
than
anything
about
technological
regulations
after
basically
so
as
I
said
before,
because
we
are
on
this
path
because
we
want
to
bring
cloud
into
the
bank
when
I
say
cloud,
I
mean
it
from
a
concept
of
jazz
or
even
peace.
It
is
a
new
way
of
developing.
The
application
is
a
new
way
of
doing
things.
B
B
We
have
agile
teams
at
the
bank
working
with
pipeline
waterfall,
working
with
web
administrators,,
supporting
those
changes
that
collide
between
projects,.
So
supporting
agile
developments
without
box,
innova
methodologies
doesn't
work
if
it
works
for
a
time,
but
there
came
a
point
where
I
think
that
what
is
happening
then
the
box
and
that
culture
where
everything
has
to
be
automatic
and
the
own
ership
does
not
arrive
until
a
code
drone
arrived
and
it
was
moved,
becomes
vital
for
time
to
market.
B
B
Among
us,
immutable
containers
are
basically
living.
Containers
are
born
with
indubitable,
but,
as
I
told
you
before,
the
practices
of
management
configuration
can
be
applied
in
their
structures
that
are
not
with
themes.
That
principle
of
I
am
not
going
to
modify
a
code.
A
server
I
see
it
as
root
and
camila
with
fiction.
If
a
script
and
see
world
was
not
executed
so
that
it
goes
and
touches
that
server
It
can
be
applied
in
containers
today,
it
is
terrible
virtualized.
It
is
independent,
only
a
cultural
change.
B
That
is
not
easy,
but
the
concept
of
the
immutable
structure
makes
all
the
environment
fights.
Of
this
mind
to
be
the
same
as
before,
it
is
the
same
as
production,
really
development,
stop
making
sense
And.
If
we
can
achieve
that
those
guidelines
and
those
concepts
of
the
world
that
are
natural
in
the
world
of
webs
that
in
the
world
of
containers,
permeate
what
will
never
be
a
contaminate,.
What
is
it
and
from
there
is
the
objective
we
are
flying
with,
we
want
to
achieve
this,,
it
does
not
do
It
needs
to
discount
too
much.
B
Let's
say
we
want
applications
that
are
lowe
reid.
We
don't
want,
let's
say
the
applications
we
develop.
We
don't
have
to
care
where
it's
going
to
run
by
putting
between
my
sandy
banks,
sometimes
the
rope
in
the
private
cloud
and
tau
brasil,
no
matter
where
it
runs.
It
has
to
be
developed
in
the
same
way,
others
with
the
same
concepts.
That
is
key.
We
have
to
take
off
the
applications.
The
infrastructure-
that
is
this
concept
of
decoupling.
The
components
of
the
application
from
the
data
in
structure
is
what
makes
us
faster.
B
I
do
n't
care
where
I
run
I
know
it's
going
to
run
the
same
everywhere.
That's
if
you
want
to
see
some
of
what
it's
about
is
that's
decouple,
that's
the
magic
that
makes
this
work
and
work
everywhere
like
we're
going
to
do
so,
you're
not
going
to
see
anything
that
doesn't
know
they
are
all
products
most
are
open-source
products
that,
basically,
is
in
the
combination
of
infra
part
products
that
we
are
using.
B
This
is
a
difficult
decision.
That
is,
we
had
a
lot
of
time
to
take
what
we
do.
We
implement
something
from
scratch
without
looking
at
anything
from
what
we
have,
or
vice
versa.
We
look
at
everything
we
have
and
do
something
that
suits
all
the
needs
of
all
our
parks
and
now
that
we
decided
to
go
out
with
the
medium
and
implementation
of
content
in
its
being
greenfield
did
not
undermine
what
I
have
It
does
n't
matter.
B
If
I
have
liberty,
if
I
have
herpes
and
I
have
a
de
belo
because
they
use
it
or
if
it
is
from
0
that
it
is
going
to
be
done
with
the
best
standards,
with
the
open-source
tools
with
the
best
methodologies
nothing
legacy
goes
in
there.
Now
our
partner
dayne
m,
we
have
to
do
it,
brown
ping.
We
have
to
do
it
with
the
best
technologies,
but
with
one
eye
is
what
we
have,
because
if
not
our
challenge
that
this
allows
beyond
that,
it
will
never
be
achieved.
B
B
B
This
is
what
we
are
teaching
at
the
moment.
It
is
not
that
complex,
but
basically
it
is
two
clusters,
one
for
soon
and
one
to
prepare
on
our
site
on
our
main
site
and
then
obviously
through
good
war.
All
of
this
is
virtualized
on
purpose.
We
choose
not
to
go
and
not
to
do,
and
so
it
happens
at
the
same
time,
because
the
current
processes
that
we
have,
many
processes,
such
as
disaster
recovery
bank,
every
year,
we
have
to
do
the
test,
recovery
is
mounted
on
in
war
meter.
B
Now
one
thinks,
or
even
one
big
worry,
be
light
to
put
and
make
sure
it
would
be
too
much
change.
We
want
to
go
little
by
little
in
the
direction
so
that
we
have
operations
you
can
follow
us.
We
chose
to
mount
the
openshift
clusters
on
bing
guard,
at
least
for
now,
and
another
concept
that
I
also
mentioned
before.
Our
6-mana
pipes
are
designed
to
be
deployed,
anywhere,.
B
Many
integrations.
I
want
a
container
to
run
inside
the
bank
in
the
cloud,
Brazil,
oh,
of
course,
running,
I'm,
a
cp
anywhere,.
That's
the
concept,
then
say
those
differences
more
beyond
the
blocks.
Conferences
more
our
pipeline
is
not
at
all
clear.
Quite
common
I
did
not
put
all
the
stents
of
the
main
ones.
Basically,
we
are
training.
It
will
continue
for
all
the
supports
everything
you
are
here
is
using
git
low,
the
garages
I,
don't
know
I
guess
they
all
look
like
I
know
what
they
speak
to
me
printed,
excellent.
B
We
are
using
it
some
projects
shared
with
brazil.
It
is
that
it
opens
me
jeffers
the
mind
opens
the
doors
for
you
in
little
bocm,
you
move
away
the
code
repos
and
the
remittances
excellent
good.
They
are
here
we
bring
satisfactor
and
for
all
that
is
binary
of
free
content.
We
are
using
both
puppet
and
hansel
and
some
brain
for
testing.
We
are
exploring
some
things
that
have
changed
for
tv,
but
the
main
one.
B
B
It
belongs
to
be
achieved
on
42,
different
consoles,
to
know
where
the
error
was
without
may
come
puncture
because
it
lacks
a
library
without
knowing
that
an
error
in
anfield
with
any
of
the
scripts,
then
everything
is
going
to
the
layer
of
is
plan,,
which
is
white,
is
a
product
that
for
what
the
open
world
is,,
it
would
be
elastic
ivana,,
it's
all
the
same,
basically,.
It's
an
intake
of
events
that
correlates
them,.
This
makes
of
actions,.
B
The
idea
is
that
there
is
only
one
place
where
all
the
events
of
the
pipeline,,
no
matter
what
it
generates
for
us,,
are
there
available
to
do
analytics
and
so
that
it
reveals
them,.
They
practice
a
single
place
to
mine,
whether
it
is
an
element
with
classroom
deployment,.
It
says
that
everything
failed
up
to
the
same
place.
Those
critics
as
we
started
this
way
as
we
want
to
change
the
objective
beyond
the
infraero
objective
of
having
containers
and
being
cloud
ready
towards
the
day.
B
We
want
to
convince
the
drinkers
people
to
move
to
this
new
world
I.
As
a
writing
front,
I
could
tell
you
from
now
on.
Everyone
has
to
do
this.
It
will
achieve
that.
This
form
does
not
work
very
well
from
experience
in
the
league,
so
what
we
are
doing.
Basically,
we
are
following
the
same
steps
as
the
container
to
the
are
network
brain
house
program.
B
We
chose
a
development
team
that
followed
that
it
feels
like
it
has
problems
and
we
had
to
say
with
so
many
problems
that
it
hurts
you
today
with
syrups
and
it
hurts,
because
you
have
release
management
code
management.
Tell
me
all
your
pain,
pons
that
hurts
you
today,
so
it
can't
be
efficient
enough
for
what
they
ask
you
business
to
mount.
We
grab
the
same
with
the
times.
Look
what
They
have
less
paintball,
because
the
truth
is
that
the
infra
is
quite
oiled,
but
let's
also
say
that
the
base
administrator
is
called
a
lot,.
B
He
is
the
number
one
expert
to
cut
all
the
applications
and
sleeplessness
because
they
don't
even
know
how
to
do
it
through
interaction,
because
it
is
based
on
duty,
one
of
the
best
ones.
Well,
baines
tell
me
what
hurts
you
based
on
that?
We
prioritize
an
option
scheme
choosing
applications
and
how
to
deploy
the
6
cd
pipeline
to
cover
your
pain
points.
Iteratively.
We
are
not
going
to
do
a
double
of
ivan,
it
never
works.
We
give
it
Everyone
who
goes
to
the
development
zones,
well,.
These
projects
also
have
to
be
agents,.
B
They
have
to
be
iterative,
test,
these,
they
work,.
Not.
If
there
is
a
quick
look
back,,
then
we
went
with
interactive
printing,,
but
with
real
examples
and
with
real
times
that
have
raise
the
platform
quality
that
these
are
our
ambassadors.
Then
these
are
the
tips
that
will
later
go
around
the
bank
and
add
more
people
to
this
new
world
because
it
works
because
faster
and
not
solving
problems.
There
is
no
more
magic
than
this
if
he
should
achieve
note
that
a
new
way
things
are
takes
away,
problems
and
others
in
a
way.
B
You
can
do
things
faster
and
you
can
deploy
your
application,
but
it's
going
to
go
faster.
It's
a
matter
of
adding
use
cases
and
adding
times
the
great
old
acquaintance
word
of
mouth.
We
are
At
this
time,
another
conscious
decision
that
we
made,
and
it
is
a
little
slower
for
that
matter,.
We
are
training
the
two
teams
that
are
in
pilot
parts
on
the
platform
before
crossing.
B
People
who
are
going
to
participate
in
this
are
going
to
know
about
these
technologies
before
the
first
server
is
installed.
Ideas
that
they
do
it
with
network
support,,
but
that
they
do
it,
as
we
know,.
They
have
had
other
previous
experiences
that
the
provider
from
Khartoum
is
coming,
well,.
Now
we
have
to
use
it,
but
they
have
to
be
the
same
teams
that
are
leading
the
entire
pilot,
the
owners
for
this
project
to
come
to
light
and
be
successful,
because
for
that
they
have
to
know
how
to
do
the
training
after
the
transfer.
B
Everything
is
very
nice,
but
you
have
to
understand
what
they
are
about
speaking
if
he
tells
you
how
many
master's
degrees,
as
he
replied,
you
have
to
have
what
you
are
talking
about
and
how
to
adapt
that
habitat.
That
is
why
training
is
key.
The
teams
have
to
know
about
the
technologies
before
the
projects
start
and
then,
before
the
end
of
the
year.
Let's
go
to
be
with
the
launched
pilot.
B
This
is
our
interactive
roadmap.
We
can
add,
features
both
the
platform
and
the
role
of
the
box,
the
number
one
paintball
in
tillage,
so
what
we
said-
and
we
are
going
to
put
that
it
inflated
at
the
beginning
for
everyone
from
there.
We
are
adding,
for
example,
The
next
one
we
are
going
to
put
in
the
tour.
We
are
going
to
put
everything
that
is
metrics
and
agile
development.
B
On
top,
we
are
also
going
to
use
it
for
bar
tracking
to
tie
the
stories
and
that
the
canvas
cells
can
use
it
too,
because
it
was
the
second
request
back
every
one.
These
things
have
an
order
based
on
the
combs,
the
people
who
are
going
to
use
it.
This
solves
life
problems
for
them,
so
they
are
going
to
be
the
main
ambassadors
and
promoters
of
technology
through
the
fungus.
This
is
key.
This
is
not
done.
It
is
not
done
by
decree,,
it
is
done
by
adoption.
B
Later,
when
our
roadmap
ends,,
we
are
going
to
replace
all
the
LM
packages
we
have
with
the
new
one.
That
is
why
it
is
key,
and
that
is
why
what
I
said
before,
the
LM
Brownfield,.
We
have
to
take
into
account
all
cases.
of
the
current
m
law
so
that
it
can
replace
it
once
and
letter
dénia
lm.
It
is
because,
if
the
battery
sinks,
the
developers
have
to
start
organizing
as
its
organized
moment
in
the
infra
platforms,
we
have
as
a
defense
the
way
we
get
code.
B
B
Some
lessons
like
they
said
no
We
are
not
in
production
yet,
but
we
already
have
some
cool
tecos,
at
least
for
what
we
entered
to
achieve
and
where
we
are
standing
now.
This
is
not
a
technological
change.
This
solves
business
problems
that
is
perhaps
the
most
screwed
up.
Tx.
We
try
to
solve
a
technological
problem.
We
have
had
a
tilde
I,
have
containers
how
nice
the
question
is:
what
business
problem
am
I
solving
business
stakeholder
I
am
giving
it
more
value
faster.
B
Well,
that
is
key
to
this
adoption
program
because
we
have
determined
it
needed
very
nice.
You
have
an
application
set
up
running
happy,
it
is
said
of
Everything
is
going
great
and
now
that
it
solves
and
wins,
it's
good
to
be
clear
that
business
problems
solved
are
very
important
so
that
this
later
becomes
a
flagship,
strategy,,
obviously,
as
people
who
tried
to
lose
as
a
strategy,
but
the
only
way
for
a
strategy
to
become
the
blood
and
change
the
culture,
a
bank
is
to
solve
business
problems.
B
B
B
We
are
going
to
bake
the
monitoring
within
the
platform
and
that
of
belo,
but
you
will
be
able
to
see
if
your
application
works
hand
in
hand,
that
change
of
chip
takes
a
long
time,,
so
training,
not
only
technical,,
but
also
that
we
want
to
achieve
what
is
the
role
of
each
one
in
this
new
way
of
working
is
key,
and
it
is
what
most
it
costs
if
they
take
us.
Something
is
this
training.
Evangelization
is
the
number
one
thing
you
need
to
make
these
changes.
B
Technically,
you
can
hire
experts
and
well,
but
after
that
knowledge,
and
that
new
way
to
hope
to
see
me
and
be
part
of
the
day
to
day
and
the
blood
of
a
Haitian
nation
takes
a
lot
of
effort
security,
barbaric
with
containers.
Everyone
is
a
mess
security
that
is
there.
Two
virtual
machines
roles
profiles,
4609
for
which
are
banks.
We
have
a
lot
of
regulations,
the
security
aspect
and
it
is
key,
You-
have
to
bring
in
the
security
and
compliance
people,
there
aren't
any,
but
also
why,
because
they
also
have
the
chip,.
B
So
I
was
talking
to
the
guys
from
the
box,.
The
security
agent
doesn't
want
it
to
be
obtained
from
the
servers
and
for
the
dt,
the
prayers
and
everything
is
technically
orchestrated
by
a
manager
edition
buffer,
obviously
to
the
extreme,
but
the
security
people
and
the
people
with
plans
also
have
to
accompany
this
change.
They
have
to
open
their
minds.
We
have
to
evangelize,
the
estates
are
still
further
from
the
community
than
we
are
The
way.
These
new
technologies
work,
the
concept
of
the
oppressor,
is
rare,.
B
B
B
What
I
said
before
this
is
imperative:
it
is
not
a
danger
the
operation
also.
Today
we
are
going
to
do
this
tomorrow,
only
one
other.
We
passed
the
other
this
when
you
want
to
remember,
you
are
in
a
totally
different
place
where
you
started
to
make
a
creative
process.
We
have
to
be
very
clear
what
I
said
before.
What
is
the
business
objective
that
we
want
to
achieve?
We
want
this
to
be
the
best
platform.
B
Only
the
best
container
platform
with
improvement
of
m
is
one
thing:
I
want
restoration
to
be
permeated
is
another
in
a
cultural
change
is
another,
so
keep
in
mind,
Of
course,.
The
final
objective
is
key
because
each
step,,
if
it
is
not
a
false
step,,
you
have
to
be
very
clear
that
each
step
adds
value
and
check
that
hypothesis
to
see
if
I
can
achieve
it,
be
careful
with
the
amount
of
data,.
B
This
type
of
platform
pipeline
generates
a
quantity.
The
data
is
absurd,,
the
data
is
useless,,
the
information
must
be
transformed,.
That
is
why
he
said
the
thing
about
es
pla,
pnc
or
the
elastic
bands.
Before,.
You
have
to
be
very
careful
so
that
the
information
reaches
the
right
people
and
in
an
digestible
way
and
give
it
a
ingestion.
Log
of
one
is
plan
k
where
he
watched,
peri
say
take
here:
you
have
your
blog.
They
serve
us
on
air.
You
have
to
give
him
information,
so
how
is
it
to
report?
How
that
data
is
transferred?
B
Information
must
also
be
part
of
the
design
of
the
platform
in
each
step.
Yes,
It
does
not
work
for
us
later
when
you
have
a
problem
due
to
being
with
the
best
platform
in
the
world.
20
minutes
are
through
a
staff
that
there
is
a
closed
port,
because
the
cdn
did
not
open
it.
Whatever
everything
has
to
be
in
the
same
place
and
everything
It
has
to
be
understandable
for
all
people
in
a
clear
way.
B
What
I
said
before,
great
with
the
world
of
containers
but
other
worlds
that
are
not
sentences,
I
will
never
put
an
application
bought
from
a
provider
that
has
12
cuts.
48,
follow
the
branch
of
the
container
train,
no.
It
never
happens,
so
you
have
to
take
these
cases
into
account,,
obviously
not
put
them
on
day
1
in
the
complexity,
because
you
can't
cover
all
the
cases
it
may
be.
B
It
never
ends,
but
you
have
to
have
management
say
well,
one
day
see
you
to
clarify,
or
it
is
not
left
out,
but
first
you
have
a
lot
of
the
forest,
then
decide
what
battles
you
are
going
to
play,
but
never
forget
to
see
everything
around
you
and
the
latest.
What
you
have
done
before
the
regulatory
concept,
I
love,
the
idea
of
content
deployment
does
not
have
it.
I
need
a
little
guy,
a
user
in
somewhere
for
49.
Tell
me
if
ok
I
study
what
is
with
this
deployment.
B
Well,
we
were
up
to
the
content
of
this
book,
but,
like
that,
a
thousand
things
cloud,
the
issue
of
roles
and
responsibilities
and
creation
of
functions.
Each
platform
is
barbaric,
but
the
regulation
of
each
country,
the
cei
The
only
thing-
is
to
download
the
implementation
and
the
standards
and
the
rules
and
the
best
practices
in
the
world
to
each
organization
in
each
country,
because
that
is
key,.
Otherwise
you
can
get
a
very
bad
surprise
or
they
do
not
want
to
go
out
and
produce,
and
that
is
all
14.