►
Description
What's Next on OpenShift Panel at OpenShift Commons Gathering Red Hat Summit 2018 Q& A after Case Study Lightning Talks - moderated by Brian Gracely (Red Hat)
A
And
ask
each
one
of
you
so
the
challenge
that
you
had,
whether
it
was
you
know
getting
into
the
technology.
It
was
a
culture
change,
walk
us
through
just
for
a
minute
or
so
two
minutes.
How
do
you
start
that
change?
Because
we've
heard
about
you
know:
I've
got
to
get
executive
buy-in,
but
some
of
this
is
there
is
no.
There
is
no
white
paper
written.
There
is
no
best
practice
talk
a
little
bit
of
how
you
went
through
that
challenge
of
saying
like
how
are
we
going
to
take
on
managing
the
kitchen?
B
We
have
an
ecosystem
for
you
in
almost
four
years
and
we
we
started
early
to
put
that
onto
a
containerized
or
to
talk
of
containers,
but
what
we
felt
about
that
was
that
that
we
strictly
on
AWS,
and
we
have
it's
not
that
scaling
as
we
like
to
have
that,
and
we
all
looking
for
a
different
solution
for
that
and
how
to
make
that
better.
And
then
we
look
at
several
solutions,
of
course,
in
the
market.
B
C
C
A
Know
I'm
asked
Canon
and
sphere
sort
of
a
variation
on
that.
So
can
you
know
you're
dealing
with
Black
Friday
transactions,
which,
for
some
parents
buying
gifts,
feels
like
a
life-or-death
thing?
If
you
don't
get
the
right
thing,
it's
here,
you're
actually
dealing
with
life-and-death
stuff.
How
do
you
convince
your
team,
your
management,
that
some
new
technology
is
the
right
way
to
go
when,
when
you
know
that's
the
risk
that
you're
kind
of
balancing
it
against
had
walk
us
through
a
little
bit
of
that
thought
process
of?
How
do
we
make
this
work
so.
D
For
us,
we
we
actually
had
a
little
bit
of
a
perfect
storm
to
make
business
decisions.
Cost
is
usually
the
driver
and
to
build
highly
reliable
systems
is
extremely
expensive
and
one
of
the
things
that
I
was
faced
with,
as
the
chief
architect
was
to
reduce
the
cost
per
transaction.
Because,
as
we
scaled,
the
solution
and
more
and
more
tenants
were
coming
on
board.
The
the
cost
of
the
transaction
was
was
impacting
us
and
so
to
to
get
the
same.
D
What
we
call
non-functional
requirements,
the
same
availability
model
with
containers
was
actually
a
lot
easier
to
do
than
with
traditional
infrastructure,
and
we
have
a
geo
distributed
platform.
So
you
know
we
have
data
centers
throughout
the
world
and
being
able
to
run
multiple
OpenShift
clusters
and
be
able
to
scale
up
multiple
container
instances
and
be
able
to
kill
any
one
of
them
at
any
point
in
time
and
and
not
have
it
even
fazed.
D
The
system
was
pretty
compelling
and
then
to
to
show
that
we
could
do
all
of
that
and
lay
this
infrastructure
out
at
an
order
of
magnitude.
Less
cost
was
the
the
business
justification
that
we
needed
to
be
able
to
pull
the
trigger
on
it
that,
coupled
with
the
company
driving
towards
an
agile
model
and
DevOps,
coming
into
being
all
of
those
things
kind
of
blended
together
to
create
a
perfect
storm.
But
it
was
really
unique
in
that
this
wasn't
an
initiative
that
was
driven
by
operations.
D
A
D
D
E
A
E
We
need
sort
of
bridging
those
in
a
new
digital
era.
If
you
want,
and
they
could
see
that
we
need
to
change
from
from
the
legacy
systems
that
we're
running,
we
have
three
different
case:
handling
systems
and
and
I
think
the
newest
one
is
actually
mainly
built
by
Indians
before
they
got
structured.
The
first
automation
the
happen
was
an
Indian
guy
sitting
in
India
and
pushing
a
button,
and
then
it
was
automated
yeah,
that's
not
compliance,
and
it
seems
like
the
business
could
see
this
demand
for
new
technology
and
stuff.
E
So
we
made
a
couple
of
PCs
invited
the
right
people,
some
of
the
developers,
and
they
were
old.
You
know
they
got
a
Red
Hat
and
saw
some
new
toolbox
and
they
won
immediately.
So
it
was
more
the
culture
change.
That
would
be
the
issue
we
saw
that
immediately
and
and
the
first
problem
was
actually
our
general
management
because
they
were
used
to
just
going
into
the
IT
department
demanding
whatever
needs
the
Hat,
but
suddenly
the
process
is
changing.
E
F
Those
that
are
doing
microservices
development
or
cloud
native
development,
which
kind
of
goes
hand-in-hand
with
the
container
of
platforms.
How
did
you
guys
train
and
enable
your
app
teams
to
make
sure
that
they're
doing
it
the
right
way,
as
opposed
to
just
developing,
let's
say
the
old-style
big
monolithic,
apps
and
just
dumping
it
on
OpenShift
I.
E
Could
start
saying
that
we
started
out
hiring
a
team
from
Red
Hat
and
they
are
still
on
board
at
half
of
them.
We
are
hiring
in
developers
and
making
sort
of
you
know
training
on
the
job.
How
to
do
it.
We
have
a
complete
dev
of
teams
from
from
Red
Hat
helping
us
out.
You
know
getting
the
right
culture,
the
right
perspectives.
They
have
been
helping
us
with
our
architecture,
everything
since
we
only
knew
about
Microsoft
service.
So
that's
how
we
did.
B
Maybe
they
can
add
something
here,
weird.
For
fact,
we
we
are
a
company
of
playing
in
playing
people
playing,
maybe
children,
maybe
playing
children,
and
if
you
get
a
new
toy,
then
you
know
the
developers
are
allowed
to
play
with
that
and
that's
at
the
moment
we
are
there
to
stadium
so
that
they
can
play
around
with
a
test
cluster.
B
But
of
course
we
have
to
also
be
gdpr
compliant
and
we
have
a
lot
of
quality
assurance
and
we
are
currently
building
a
framework
for
for
over
shift
to
have
a
central
point
of
checking
compliance
checking,
security
and
so
on,
and
that
will
be
in
the
future
will
be
introduced.
And
then,
of
course,
the
the
children
cannot
play
anymore.
In
that
way
say
they
have
some
direction
where
they
can
go
and
some
marriage
and
where
they
can't
go,
and
it's
that's
the
way
we
are
currently
doing
that.
D
You
know
we,
we
spun
the
environment
up
by
doing
some
initial
training.
It
was
almost
like
a
grassroots
organization.
At
at
ACI,
we
started
by
presenting
to
the
tactical
Leadership
Council
this
the
concept
of
micro
services
and
how
to
build
them.
Then
we
created
small
development
teams
that
started
working
and
building
micro
services,
and
then
we
actually
expanded
them
into
the
the
docker
realm
after
they
were
already
working
with
micro
services.
So
we
didn't
start
out
of
the
gate.
D
The
unfortunate
thing
in
all
honesty
is
that
when
you
self
learn
like
that,
you
pick
up
bad
habits
and
when
knowledge
is
not
available,
you
try
to
figure
out
ways
to
work
around
the
knowledge
gaps,
and
then
you
apply
things
into
the
environment
that
are
not
necessarily
best
practices,
and
so
now
we're
in
a
little
bit
of
a
backpedal
and
we're
trying
to
fix
some
of
our
earlier
sins.
Where
we
made
compensations
for
lack
of
knowledge
and
where
we're
revising
things
now
so
it'll
be
an
even
better
environment,
I.
C
Last
year
we
choose
a
platform
and
we
integrate
it
in
our
ecosystem
and
later
we
we
choose
one
of
a
big
application:
intern
application,
all
the
application
with
more
and
more
feature
with
a
big
year
to
deploy
on
different
application
server
and
with
application
team
and
work
together.
We
we
try
to
integrate
the
first
pillar
of
macro
service,
12
factor
and
begin
together
to
separate
and
full
senility
and
extract
extensionality
from
main
application.
C
But
I
think
that
Cirie's
way
is
a
very
training
of
a
job
work.
Today,
today,
in
our
company,
we
have
a
main
table
for
composite
by
operator
team,
but
many
architect
and
many
application
team
who
debate
the
best
way
to
realize
right
pattern
for
internal
application
for
give
to
our
fabric
application.
This
pattern
for
make
internal
library
to.
C
F
A
You
know,
gentlemen,
4/4
time
I
think
we're
gonna
wrap
it
up
with
that.
That
was
a.
We
appreciate
the
input
folks
I
appreciate
the
question
just
for
the
lack
for
the
sense
of
time
we're
gonna
bring
up
some
of
the
p.m.
so
we
can
kind
of
do
an
ask
me
anything
and
then
we
can
get
to
kind
of
stretching
and
and
and
getting
out,
having
a
beer
or
having
a
drink
talking
with
some
people.
So,
gentlemen,
thank
you
very
much
for
the
time
tonight.