►
Description
Cloud Native in the Enterprise: Real-World Data on Container and Microservice Adoption
Containers and microservices are two of the fastest-growing trends in technology, enabled by DevOps. This talk will delve into the state of cloud-native prerequisites in the enterprise, the Docker and containers ecosystem including current adoption, and data on companies moving to cloud-native platforms. We'll close by looking at real-world examples of containers and microservices architectures at leading-edge companies.
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.
A
A
Whoo,
yeah
and
I
also
don't
have
to
sell
you
anything,
which
is
probably
a
good
thing
for
you,
I'm.
So
I'm
going
to
talk
about
cloud
native
in
the
enterprise.
It's
going
to
start
a
little
bit
high
level,
but
then
we're
going
to
work
in
pretty
quickly
to
data
on
containers
and
microservices
usage
and
a
few
examples
of
real
world
implementations
some
of
the
problems
they
ran
into
you,
some
of
the
reasons
they
made
the
shift
and
some
of
the
benefits
they
got
out
of
it.
A
So
I
give
this
stump
pitch
on
sort
of
DevOps
and
cloud
pretty
regularly
and
I
dumped.
Basically,
all
of
that,
because
if
you're
here
and
you
don't
know
what
cloud
is
you're,
probably
in
the
wrong
room,
but
I
did
want
to
start
out
with
a
couple
of
buzzwords,
because
every
conference
I
go
to.
Inevitably
vendors
mentioned
one
or
both
of
these
somewhere
in
the
keynote,
because
it's
all
about
doctor-
and
now
it's
all
more
recently,
it's
all
about
micro
services
and
2014.
A
As
a
data
point,
there
was
a
toss
con
one
of
my
favorite
open-source
conferences.
This
will
soon
be
one
of
them.
Next
year
right,
linux,
con
becomes
the
Evenstar
summit.
There
were
zero.
There
was
one
talk
in
2014
proposed
on
micro
services
and
a
2015
number
30.
So
it's
another
one,
these
examples
of
things
that
come
out
of
nowhere
and
suddenly
become
the
only
way
to
go
forward,
but
it
does
show
you
how
populist
this
stuff
is
getting
on
the
same
thing
glue
another
good
developer
event.
A
You
know
there
were
zero
talks
in
2014
and
it
went
to
12
one
year
later.
So
this
stuff
is,
you
know,
getting
a
lot
of
hype
and
whenever
there's
a
trend
happening
with
a
bunch
of
hype
going
on
I
always
want
to
try
and
figure
out.
What's
the
reality
behind
that,
how
can
we
make
this
a
little
bit
more
concrete
and
figure
out.
You
know
is
a
stuff
that
actually
has
value
and
the
people
are
really
implementing.
A
Or
is
it
just
something
that
you
know
bunch
of
vendors
are
going
up
on
stage
and
pitching
at
their
conferences,
and
so
some
of
the
high-level
reasons
for
this
you
know
you
heard
some
of
this
from
angel
and
I
just
wanted
to
show
you
a
little
bit
of
data
on
what
we're
seeing
about
the
I,
teashops
think
about
technology
and
think
about
their
priorities
and
the
important
part
of
this
slide
and
that
the
bottom,
the
n,
is
the
sample
size.
So
we
want
to
and
ask
more
than
1500
IT
professionals
and
IT
decision-makers.
A
What
are
their
key
priorities
and
the
important
part
is
at
the
top
that
blue
an
orange
thing
right
and
that's
the
shift
in
priorities
from
those
bottom
three
blocks
which
are
really
about.
You
know
cost-cutting
penny-pinching
to
an
approach.
That's
focused
much
more
around.
You
know
collaborating
with
the
business
and
investing
in
technology,
in
other
words
becoming
a
software-defined
company.
As
you
see,
you
know
this
dual
approach
of
coupling,
you
know
improved
agility
with
decreased
risk.
A
But
it's
true-
and
you
can
relate
this
back
to
math
too,
and
I'm
not
going
to
have
any
time
to
go
into
that.
But
if
you
go
back
to
the
math
of
things
like
batch
size
as
you
decrease
batch
size,
both
horizontally
from
an
architectural
point
of
view
and
vertically
from
a
continuous
delivery
point
of
view,
you
decrease
the
impact
of
any
of
the
risks
that
happen.
A
While
you
decrease
the
cues,
as
you
push
things
out
to
production
and
you
increase
the
agility
you're
able
to
have-
and
you
know
I
love
the
site-
the
new
stack
glad
they
have
that
pancake
breakfast
this
morning,
good
stuff,
but
when
I
think
about
the
new
stack
I,
don't
think
about
it
as
a
single
stack.
Instead,
it's
an
infinite
array
of
possible
stacks,
because
I
think
fragmentation
is
one
of
the
most
important
forces
in
technology.
Today
it
used
to
be
back
in
the
mid-90s.
A
If
you
wanted
to
pick,
if
you
wanted
to
say
hey
what
languages
will
cover.
Ninety
percent
of
all
the
code
that
gets
written
you'd
come
up
with
about
three,
but
if
you
fast-forward
today,
and
you
say
all
right
how
many
languages
do
have
to
support
if
I
want
ninety
percent
of
all
the
code
to
work
on
my
platform,
suddenly
you're
at
something
more
like
10
right,
it's
no
longer
C,
C++
and
Java
is
good
enough.
A
That
has
to
be
done
and
then
segment
those
across
with
a
pis
and
so
I'm
going
to
spend
a
little
bit
of
time
on
this
dev
ops
stuff,
like
I,
said
I'm,
assuming
that
you
mostly
know
and
are
deeply
familiar,
in
fact
with
most
of
it.
But
what
I
did
only
show?
You
was
a
little
bit
of
data
on
where
things
are
right
now
out
there
in
the
real
world.
A
Instead
of
inside
of
this,
you
know
high-end
tech
bubble
that
we
live
in
and
one
of
the
big
trends-
I'm
sure
you've
probably
heard
this
metaphor
before-
is
this
pets
versus
cattle
idea
of
moving
from
you
know,
single
servers
that
are
handcrafted
and
art
is
only
maintained
to
something
that
is
thought
about
in
terms
of
herds
and
fleets
and
any
individual
server
or
container
or
vm
is
fungible
can
be
replaced
at
any
time.
Everything
moves
toward
policy-driven
infrastructure.
A
What
I
think
something
that's
leg
behind
in
this
when
we
think
about
DevOps,
I,
think
about
culture,
automation,
measurement,
we've
had
the
automation
part
of
that
mostly
solved
in
terms
of
moving
toward
the
cloud
native
approach
of
cattle
or
of
goats.
If
you
prefer
not
to
think
about
cattle
that
way,
we've
got
them.
The
automation
part
mostly
solved
with
config
management,
with
container
orchestration
with
all
these
different
tools,
but
where
we've
leg
behind
most
of
the
way
until
now
has
been
the
monitoring
side
of
things
all
right.
A
What's
the
health
of
the
overall
service,
I,
don't
care
about
any
individual
server
and
if
you've
heard
Adrienne
cockroft
talk
about
this
he's
been
pushing
this
for
a
while
now,
but
most
of
the
monitoring
tools
of
leg
behind,
but
finally,
that's
starting
to
change
now
between
tools
from
vendors
and
tools
from
the
open
source
community
like
prometheus,
and
so
it's
good
to
see
the
monitoring
side
of
DevOps
finally
catching
up
with
the
automation
side
of
it.
So
this
is
an
interesting
example.
How
many
of
you
are
familiar
with
what
happened
in
night
capital?
A
Eight
production
servers
to
a
live
trading,
environment
managed
to
lose
half
a
billion
dollars
in
the
span
of
45
minutes
and
went
out
of
business
and
I
contend
that
wasn't
a
failure
in
terms
of
documentation
or
policy,
or
any
of
that.
That
was
a
failure
because
they
didn't
think
about
it,
like
cattle
right
had
they
deployed
to
their
production
environment
in
a
cattle
driven
manner,
they
would
have
been
fine.
A
Well,
we're
still
not
too
far
right
how
far
we're
getting
in
terms
of
automation,
most
I
tea
shops
today
are
still
mostly
manual,
which
is
kind
of
terrifying,
because
it's
just
going
to
be
more
problems
like
that
night
capital
example,
although
that's
a
kind
of
an
extreme
case,
there
are
many
many
changes.
Many
many
downtime
issue
is
caused
by
failure
to
fully
understand
and
failure
to
fully
automate
at
the
production,
environment
and
one
process
to
it.
A
And
even
if
we
go
back
to
ideas
that
are
20
years
old,
like
a
gel
I
am
shocked
by
how
many
people
I
talked
to,
who
still
don't
do.
Agile
development
and
I
was
just
at
a
conference
recently
where
somebody,
through
and
agile,
alongside
things
like
cloud
native
and
containers
as
a
new
paradigm
and
I'm
like
what.
What
have
you
been
smoking
because
I
really
want
some
of
it.
A
And
so,
when
we
look
at
agile,
it's
still
limited
to
something
like
two-thirds
of
IT
shops.
Today,
it's
crazy
and
I
was,
at
a
conference
last
fall
running
this
little
DevOps
roundtable
and
to
the
last
one,
every
single
company,
the
came
to
it
was
still
doing
waterfall
development,
and
so
you
know
for
some
of
the
world,
there's
a
long,
long
ways
to
go
on
the
journey
to
cloud
native
and
it's
an
idea
that
they
might
hear
about.
A
But
it's
going
to
take
in
at
least
five
years
to
get
to
that
point,
and
even
if
we
look
at
DevOps,
you
know
things
are
getting
farther.
Things
are
doing
a
little
bit
better.
I
mean
something
like
sixty.
Five
percent
of
the
shops
are
doing
agile
for
DevOps,
it's
more
like
forty
percent
on
average,
so
we're
making
progress
toward
getting
there.
It's
finally
I
think
getting
to
the
point
you
know,
eight
years
in
since
we
came
up
with
the
term
DevOps,
it's
finally
reaching
the
mainstream.
A
Finally
getting
past
the
early
adopters
which
changes
the
discussions,
we
have
right.
Suddenly
it's
a
discussion
of
business
values
instead
of
like
this
is
a
really
cool
tech
and
we
should
totally
go
with
it.
Suddenly,
you
get
to
a
whole
different
level
of
conversation.
Talking
to
you
know,
CIOs
and
VPS,
and
all
that
stuff
about
things
like
I
showed
you
the
beginning
like
agility
and
risk,
and
so
on.
Instead
of
hey,
we
should
definitely
automate
the
servers.
A
We've
got
to
go
with
ansible
on
this
one
or
whatever
it
is
so
there's
still
a
long
ways,
I'm
still
a
pretty
broad
spread
in
terms
of
tools
and
I.
Don't
want
to
get
all
semantics
about.
These
are
or
not
DevOps
tools,
because
those
kinds
of
arguments
are
just
waste
of
time,
but
it
does
you
to
give
you
a
sense
of
a
very
broad
spectrum
of
adoption,
even
within
whatever
you
might
want
to
call
DevOps
right,
which
to
me
is
just
taking
as
when
applying
that
to
iterate
between
the
business
and
the
customer
more
quickly.
A
All
of
this
stuff
falls
somewhere
in
the
middle,
whether
it's
basic
project
management
or
whether
its
paws
or
whether
it's
config
management
or
whatever
and
there's
a
very
broad
spread,
and
you
know,
there's
not
really
a
point
at
which
I
think
you
get
to
DevOps,
but
I
do
say
that
devops
from
my
perspective
is
a
prerequisite
for
cloud
native
right
and
that's
why
I'm
talking
about
this
stuff?
You
can't
do
cloud
native
effectively
if
you
don't
have
the
right
processes,
the
right
structure
in
place
to
support
it.
A
Otherwise,
you
know
you
end
up
in
this
place
where
the
processes
don't
enable
you
to
do
any
of
the
possible
benefits
you
could
have
gotten
out
of
it,
and
all
you've
done
is
created
more
work
for
yourself.
I
was
talking
to
accompany
a
a
couple
years
back
and
they
were
really
proud
of
themselves
because
they
had
moved
to
a
private,
but
what
they
did
was
they
wrote
their
own
private
cloud
from
scratch,
which
first
off
is
a
terrible
idea.
A
I
mean
if
you
want
to
pick
a
way
to
get
a
broken
crowd,
broken
cloud
and
no
support
from
anybody.
Just
write
your
own
and
they
were
super
excited
because
you
know
what
they
had
done.
They
had
moved
from
a
provisioning
time
for
a
new
vm
from
60
days
down
to
30
days,
because
it
was
all
blocked
by
the
process
right.
It
was
all
human
approvals
of
stuff
where
that's
no
longer
required
once
you're
in
even
a
cloud
world,
let
alone
a
cloud
native
world
of
containers
and
functions
as
a
service.
A
And
so,
if
you
can't
fix
the
processes,
you're
never
going
to
get
any
of
the
benefits
out
of
it.
And
that's
why
I
say
DevOps
is
such
an
important
prerequisite
and
it's
not
even
worth
having
the
cloud
native
conversation
unless
you're
at
a
point
where
you
can
always
do
continuous
delivery
and
where
you
have
easy
ways
to
move
quickly
without
being
blocked
on
other
parts
of
the
organization.
A
And
if
we
look
at
typical
release
speeds
again,
people
are
still
moving
pretty
slow
right,
we're
a
long,
long
ways
from
the
euphoria
of
continuous
delivery
for
most
of
us-
and
this
is
another
one
of
those
things
where,
if
you
can't
do
continuous
delivery,
you
can't
really
and
effectively
have
the
cloud
native
conversation,
and
so
the
only
people
who
can
really
have
these
conversations
right.
You
can't
the
weekly
folks
they're
on
their
way,
mostly
folks,
they've
got
some
changes
to
make
the
quarterly
folks
and
slower
than
that.
A
You
know
it's
again,
one
of
those
things
where
it's
a
year
is
out
for
them
to
get
here,
and
so
it's
unfortunate,
but
until
we
can
get
most
of
those
people
in
the
middle,
much
farther
to
the
left
toward
at
least
that
daily
release
pace,
which
means
they
fix
the
processes
right.
If
you
can
release
daily,
you
fixed
all
the
processes
that
are
blocking
you
from
speed,
and
then
you
can
start
having
that
cloud
native
conversation
being
able
to
move.
You
know
as
quickly
as
you
want
to.
A
And
so
now,
let's
move
on
to
the
container
discussion
as
we
move
farther
into
the
cloud
native
concepts
and
prerequisites
and
ideas
are
to
show
you
some
data
from
a
few
different
data
sources
on
what's
going
on
in
terms
of
container
adoption
and
I.
Don't
expect
you
to
be
able
to
read
the
key
or
even
really
see
the
lines
on
this
I'm
just
making
a
couple
of
key
points.
A
So
don't
worry
too
much
if
you
can't
read
it
so
I
dug
into
Stack
Overflow
a
pretty
popular
developer
discussion
forum
and
they
make
this
cool
data
Explorer
available.
So
you
can
go
basically
right,
sequel,
queries
and
look
for
anything
you
want
to
on
it.
So
I
want
to
look
for
tagged
questions
per
month
over
time
over
I
mean
the
important
part
is
the
three
or
so
years
on
the
right
part
of
it.
I
mean
you
can
see
that
you
know
we've
had
containers
as
an
idea
for
decades.
A
At
this
point,
it's
not
a
new
thing,
but
what
changed
was
it
suddenly
became
really
easy
to
use
the
barrier
to
entry
got
a
lot
lower
because
of
the
usability
crisis,
and
because
of
you
know,
the
ability
to
combine
a
bunch
of
different
kernel
features
into
something.
That's
pretty
straightforward
in
terms
of
using
it,
and
so
doctor
did
a
great
credit
to
the
community
for
doing
so
and
enabled
all
this
stuff
to
start
moving
on
naval
the
conversation
to
move
forward
and
as
you
can
see
here,
it's
blown
up
right.
A
The
container
conversation
has
gone
from
zero
to
a
million
in
the
span
of
the
past
few
years
and
the
other
interesting
point
is
and
we're
going
to
get
it
more
into
this
leader
is
what's
happening
in
terms
of
dealing
with
those
containers.
It's
the
at
scale
and
the
second
most
popular
line
on
here
is
covered
eighties
right,
llego
cloud
native
computing
foundation,
and
so
we're
seeing
you
know,
emerging
traction
for
the
container
orchestration
part
of
things
too.
But.
A
It's
not
just
developers-
and
this
is
the
thing
I
always
get
concerned
about
when
we
start
hearing
high
pine
trenzas,
all
right.
It's
a
bunch
of
vendors,
pushing
something
nobody
were
using
in
the
real
world
or
its
bunch
of
developers
playing
around
with
toys
and
no
companies
are
really
implementing
the
stuff
in
production,
and
so
that's
where
I
went
back
to
our
data
where
we
serve
a
bunch
of
IT
shops
and
start
looking
at
things
over
time
and
start
looking
at
things
that
scale
talking
to
the
sort
of
senior
IT
crowd.
A
In
addition
to
being
able
to
look
at
the
developer
side
of
things-
and
this
is
a
first
time,
I've
shown
the
latest
data
set
on
here,
which
is
looking
at
container
adoption
over
time.
So
you
go
from
early
2015,
where
I
think
one
of
the
interesting
parts
of
this
is.
If
you
look
at
the
initial
and
broad
production
right,
you
can
combine
those
and
say
all
right
how
many
IT
decision
makers
say
their
their
company
has
containerized
absent
prod
about
6%,
pretty
small,
but
reflects
relatively
closely
on
things
like
the
release
speed.
A
A
We
got
up
to
something
like
just
over
fourteen
percent
of
shops
able
to
have
containerized
absent,
prod
and
a
bunch
more
who
were
had
of
NPO,
CS
and
dev
test
and
somewhere
around
there.
But
what's
starting
to
be
interesting
was
this
change
that
you
can
see
happening
at
the
discovery
in
eval
phase
and
the
no
plans
phase
which
give
you
some
early
clues
of
the
community
and
the
population
of
users
starting
to
segment
themselves
out
in
a
couple
of
different
camps?
A
And
if
we
fast
forward
again
to
the
latest
data,
you
see
even
more
of
that
right.
So
we're
kind
of
stuck
right
now
at
about
fourteen
percent
of
the
shops
that
have
containerized
absent,
prod,
I
suspect.
This
is
related
to
things
like
a
release.
Speed
like
not
getting
all
the
prerequisites
in
place
to
be
able
to
do
containerized,
apps,
but
what's
interesting
is
this.
This
shift
that's
happening
at
the
discovery,
neevalle
phase
in
the
no
plans
phase
and
how
I'm
interpreting
this
is.
A
A
Why
they're
not
seeing
the
value
and
we've
actually
asked
about
this
in
the
survey
see
what
are
the
technology
blocker
is
where
the
technology
driver
is
all
that
good
stuff
business
blockers,
business
drivers,
I'm
going
to
show
you
a
little
bit
of
that,
but
not
a
lot
of
it.
One
of
the
big
issues
is
lack
of
expertise
right.
They
just
can't
find
the
people,
they
can't
train
the
people.
A
What's
going
on,
I
call
out
the
Silicon
Valley
bubble,
but
it
goes
beyond
the
valley
right.
It's
the
global
tech
community,
of
sort
of
the
elites,
if
you
will
compared
to
everybody
else,
and
it's
critical
that
we
spend
more
time
sharing
knowledge
outside
of
this
community.
All
right
like
go
home
if
you
live
somewhere
besides
the
valley
go
home
and
go
speak
at
local,
like
IT
meetups
and
that
kind
of
thing
about
what's
cloud
native,
why
is
the
value
about?
A
And
so
let's
pops
are
all
that
good
stuff
again,
so,
like
I
said
about
fifteen
percent,
or
so
maybe
one
in
seven
ish
shops
have
containerized
absent
prod
about
a
third
of
them
are
at
the
point
of
playing
around,
which
is
pretty
good.
Considering
some
of
the
graph,
as
shown
you,
this
has
gone
from
like
I,
said
zero
to
60
in
the
span
of
just
a
couple
of
years,
and
it's
very
rare
for
any
kind
of
new
technology
to
get
into
this.
A
Many
enterprises
that
fast
and
so
just
a
little
more
data
on
the
kinds
of
things
that
are
happening
out
there
with
containers.
We
spend
a
lot
of
time
thinking
about.
What's
the
right
place
to
run
a
lot
of
workloads,
but
what's
the
multi-cloud
hybrid
cloud
feature
all
that
jazz,
and
so
we
go
there
and
ask
people
where
are
you
running
your
containers
today?
Most
of
them,
it
turns
out
today,
are
running
them
at
home.
A
That's
not
the
case,
I'm
the
same
thing,
if
you
ask
people
where
they
think
they're
going
to
be
in
a
couple
of
years
and
I'm,
I
was
a
little
bit
leery
of
this
stuff
because
everybody
is
more
optimistic
than
they
have
any
right
to
be
services.
Oh
yeah,
we'll
we'll
be
microservices
in
two
years.
The
problem
it's
going
to
be
everything
all
of
our
apps
are
going
to
be
rewritten.
A
This
is
another
one
I
find
interesting.
You
know
I
had
this
preconception
with
the
way
people
would
run
containers
that
pretty
much
everybody
would
just
slap
one
in
a
vm
right,
because
we
have
a
lot
of
discussions
about
containers
and
security
right
now,
and
a
lot
of
people
are
unclear
on
whether
containers
are
sufficiently
secure
compared
to
VMS
or
whatever
else,
and
so
you
know
we
wanted
to
go
there
and
ask
how
are
you
handling
this
stuff?
What
are
you
doing?
A
The
container
is
relative
to
VMS
and
give
them
three
options,
which
number
one
was
basically.
This
is
a
new
workload.
Has
nothing
to
do
with
VMs,
isn't
stealing
an
app
that
used
to
run
a
vm
before
number
two?
Was
you
slap
it
right
into
a
vm
number?
Three
was
you're
migrating
stuff
over,
and
I
was
really
surprised
to
see
that
it
was
a
pretty
notable
split
across
those.
There
wasn't
one
dominant
approach.
A
People
are
running
containers
in
a
lot
of
different
ways,
and
some
of
them,
you
know,
like
I
said,
are
essentially
a
cloud
native
style
workload.
Some
of
them
are
basically,
let's
lift
and
shift
into
a
container,
and
now
we're
cloud
native
right
turns
out.
It's
not
quite
that
easy
and
in
fact,
you've
just
made
your
problems
worse,
because
you've
moved
something
that
requires
hardware
redundancy
to
an
environment.
That's
software
redundancy!
A
Now
you
just
have
more
downtime
and
in
terms
of
what
people
are
doing
with
the
containerized
apps,
I'm
from
my
perspective.
It's
mostly
infrastructure
led
right
now,
and
we
ask
people
well
like
20
or
so
different
kinds
of
workloads,
but
it's
it's
mostly
the
stuff.
That's
like
web
servers
apt
of
that
kind
of
thing.
No
huge
surprises
here
in
retrospect,
but
it's
always
easy
to
have
twenty-twenty
hindsight.
There
aren't
many
people
putting
their
essay
p
in
a
container
right
now.
A
Let's,
let's
put
that
way,
I
saw
them
throw
a
tweet
into
the
cloud
native
day
hashtag
about
we
can
run
on
cloud
like
well
yeah,
but
I'm
not
sure
your
cloud
native
yet
and
so
bring
any
of
that
together.
The
discussions
we've
had
over
the
course
of
this
talk
so
far,
you've
got
fragmentation
being
a
really
important
driving
force.
From
my
perspective.
Combine
that
with
companies
that
have
meet
met
the
prerequisite
of
getting
DevOps
in
place
on
taking
those
containers
out,
and
you
end
up
with
microservices
and
I.
A
What
I
did
on
this
slide
is
I
just
dropped
docker,
because
it's
so
far
off
the
scale
that
you
can't
even
really
tell
what's
happening
with
anything
else
and
left
kept
everything
else
in
and
what
you
can
see
is
again
the
same
thing.
Cuba
nadeyus
is
doing
pretty
well
among
those
container
orchestration
style
tools,
but
the
other
ones
are
by
no
means
dead.
Right
may
Zeus
is
still
doing
reasonably
well
down
there
in
that
kind
of
brownish.
Yellow
line.
Dr.
A
swarm
is
that
yellow
bright,
yellow
line,
starting
to
pick
up
pretty
quick
now
that
it's
come
out,
there's
a
lot
of
other
options,
kind
of
floating
around
too-
and
I
put
this
up
here,
I'm
especially
I
wanted
to
point
out
that
dr.
swarm
part
to
make
the
point
that
there's
never
a
point
at
which
the
game
is
won
and
it's
over
and
we
could
say
our
Cuban
eighties
has
one
we're
good
now
and
there's
always
going
to
be.
You
know
new
options
emerging
that
mean
everybody
has
to
continue
to
keep
up
and
innovate
right.
A
You
can't
say
like
all
right
we're
done
with
Cuban
eighties.
We
finished
Dave
day
/
job
done.
You
got
to
keep
moving
forward,
otherwise
you
know
you
become
the
one
who
gets
disrupted
and
then,
if
we
tie
that
into
the
survey
data
you
know
went
out
there
and
like
I,
showed
you
earlier
about.
Fifteen
percent
of
IT
shops
have
containerized
apps
and
prada
cording
to
the
survey.
A
A
And
the
interesting
thing
the
thing
I'm
showing
you
on
this
graph
is
there's
two
lines
on
here.
One
of
them
the
red
one,
is
the
cost
per
job
they're
running
in
there
and
they're
cluster,
and
the
blue
line
is
the
number
of
total
jobs
completed
and
you
may
or
may
not
be
able
to
see,
but
there's
a
vertical
line
right
in
the
middle
there,
which
is
when
they
made
that
flip
over
to
micro
services.
And
the
interesting
thing
was
that
interesting
improvements
in
efficiency,
getting
rid
of
single
points,
of
failure
being
able
to
move
quicker.
A
They
saw
economic
improvements
to
write
their
cost
per
job,
that
red
line
went
down
dramatically
after
they
made
the
move
because
they're
able
to
make
more
efficient
use
of
the
infrastructure
they
already
had,
instead
of
set
having
it
sitting
around
right.
So
it's
just
another
stepping
that
iterative
process
of
moving
more
toward
the
cattle
style
model,
the
cloud
native
model,
even
the
functions
of
service
model
and
just
to
kind
of
flip
through
the
architecture
you
know
they
used
to
have
the
good
old,
classic
stack,
move
to
something.
A
Almost
everybody
who
is
implemented
microservices
ends
up
with
a
big
mess.
Complexity
is
the
new
normal,
and
the
challenge
is
coming
up
with
the
right
sets
of
tooling
to
be
able
to
cope
with
that
things
like
distributed
tracing
tooling
things
like
modern
service
level,
monitoring,
policy-driven
infrastructure.
All
this
stuff
is
critical
for
companies
who
are
trying
to
do
this
so
that
they
can
effectively
deal
with
all
the
crap
I'm.
A
This
is
more
like
a
two-year
journey.
They
did
a
talk
while
back
where
they
published
how
many
micro
services
they
had
successfully
implemented
over
time,
and
it
took
them
two
years
to
get
to
the
point
of
getting
to
60
microservices
up
there
two
years
to
get
to
the
point
of
fully
replacing
everything
they
had
already
had.
So
it's
not
an
overnight
process.
It's
you
know
more
of
a
strategic
investment
and
when
you're
talking
to
your
companies
or
your
clients,
company
is
about
this
stuff.
A
I
think
it's
important
to
emphasize
that
this
is
a
strategic
investment
that
will
take
a
while
to
show
the
dividends.
This
isn't
a
flip
it
in
a
quarter
type
thing
we're
not
date
rating
with
micro
services
I'm.
The
final
example
I'm
not
going
to
spend
a
lot
of
time
on
this
because
I'm
running
out
of
time
here
is
basically
that
Chinese
kayak
right
they
moved
to
microservices
the
difference.
That's
interesting
in
the
way
they
did
it
was.
A
They
took
kind
of
an
inverse
Conway's
law
approach,
in
which
they
said
we're
going
to
take
our
existing
architecture
and
essentially
segment
it
by
business
unit
by
the
business
units
we
already
have,
and
instead
of
trying
to
reorganize
everything
and
do
micro
services.
That
way,
so
you
can
see
they
had
the
old
monolith
broken
into
four
main
chunks
now
exactly
the
same
microservices
and
for
man,
shocks
and.
A
A
There
are
a
lot
of
companies
that
haven't
met
the
prerequisites,
yet
that
are
doing
much
more
legacy,
driven
things
with
their
cloud
platforms,
running
olds
to
old-school
workloads,
web
apps,
migrating
legacy,
workloads
to
the
cloud,
but
not
ready
for
cloud
native
yet,
but
it
is
just
in
its
starting
days
and
I
think
it's
going
to
go
far
and
with
that.
Thank
you
for
your
time,.