►
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Wait, People Run Kubernetes on Mainframes? - Elizabeth K. Joseph, IBM
When you think of container orchestration mainframes probably aren't the first thing that come to mind. But modern mainframes run Linux as a first class citizen and KVM can be used for virtualization, opening a whole world of open source tooling integration via libvirt and related virtualization tooling. The careful observer may have already discovered that the mainframe architecture (s390x) is one of the architectures that's built for every Kubernetes release. How did this come to be? Who uses these mainframe builds of Kubernetes? Why would you run a distributed container orchestration service on a platform that's a symbol of the monolith we're looking to leave? Drawing upon my work with distributed systems and containers, including time spent on OpenStack, Apache Mesos and Kubernetes, and my new experience with mainframes, this talk answers all of those questions and more.
https://sched.co/UaeG
A
All
right
so
I'm
gonna
get
started.
There's
someone
here
from
Crais
I
saw
your
on
Twitter
Oh
brethren,
alright,
so
I'm
here
to
talk
to
you
about
a
very
unusual
topic
for
cube
con.
So,
first
of
all,
thank
you
for
making
the
walk
over
here
either
you
really
want
to
be
here
or
you're.
Super
lost
so
either
way.
I
definitely
appreciate
it.
A
So
I
want
to
get
started
so
once
upon
a
time,
I
worked
on
distributed
systems
and
I
spent
about
six
years
doing
this
working
on
all
kinds
of
different
distributed
systems,
and
if
you
had
asked
me
a
year
ago
what
a
mainframe
was.
There
were
a
few,
a
few
words
that
marketing
doesn't
like
me
to
say
that
would
have
come
to
mind
old
legacy,
outdated.
They
probably
have
spiders
in
them.
A
A
A
We
would
work
with
customers
and
I
would
talk
to
community
members,
and
they
generally
would
bring
in
a
new
DevOps
team,
and
that
would
be
the
people
who
would
come
to
modernize
the
platform.
So
the
company
would
have
a
mainframe
already,
but
they
wanted
a
new
team
to
come
in
and
make
things
all
shiny
and
new
and
use
all
those
weak,
whiz-bang
features
and
everything.
So
they
brought
in
this
external
team
and
the
mainframe
team
would
be
sequestered
off
into
their
own
little
basement
somewhere
in
a
different
team.
A
They
wouldn't
really
be
talking
to
this
new
DevOps
team,
but
they
were
pretty
much
separate
from
the
process,
and
this
was
very
unfortunate,
but
down
the
road
we'd
be
working
with
them
and
I'd
see
a
story
at
a
conference
or
something
and
eighteen
months
later
into
their
modernization
effort.
They
did
a
lot
of
micro
service,
they
split
up
their
monolith,
but
in
the
end
they
didn't
replace
the
mainframe
and
in
some
of
these
talks,
especially
when
I
went
to
ones
at
maces.
Con
they
sort
of
saw
this
as
a
failure
they're
like
listen.
A
We
got
to
a
certain
point,
but
we
just
couldn't
replace
the
mainframe,
because
it's
doing
things
you
know
you
can't
we
can't
take
it
down.
We
can't
break
it
up.
We
don't
understand
it
and
there'd
be
a
whole
bunch
of
reasons
why
they
got
to
this
point.
So
I
already
mentioned
a
little
bit
about
my
background,
but
I've
been
doing
systems
administration
for
about
15
years.
All
of
that's
been
on
Linux,
so
I
joined
IBM
back
in
April,
and
this
was
an
interesting
conversation
with
them
because
they
contacted
me
like
December.
A
But
I
talked
to
some
of
the
engineers
and
they
actually
built
a
really
compelling
case
around
open
source
and
the
direction
they
were
going
and
I
eventually
joined
on
so
I
joined
the
IBM
C
team,
as
a
developer
advocate
I've,
also
written
a
couple
of
books,
so
I've
been
in
the
a
bunch
of
community
for
years
and
I
also
wrote
a
book
about
OpenStack
when
I
was
working
on
that.
So
definitely
my
background
is
like
a
deep
distributed
systems.
A
A
So
the
first
mainframe
was
the
s3
60
I
got
to
visit
one
when
I
went
to
pick
up
see
a
couple
months
ago
up
in
New
York,
one
of
the
they
have
a
whole
like
Museum
outside
of
their
assembly
plant
up
in
the
pickup
scene,
IBM
office.
But
that's
like
sort
of
what
I
thought.
They're
old,
but
it
turns
out
they're,
not
IBM,
releases
the
new
mainframe
roughly
every
two
years,
so
that's
the
one.
On
the
other
side,
we
released
a
new
one
in
September.
We
had
whole
big
fancy
release
events.
A
It
was
pretty
cool
but
they're
they're
much
smaller.
Now,
but
essentially
it's
a
big
computer
but
they're
not
as
big
as
they
used
to
be
the
one
here
closer
to
me.
It
fits
into
a
19-inch
rack
spot,
so
it
uses
like
standard
connections
that
you
get
into
a
data
center,
and
so
it's
pretty
much
just
like
drop
it
in
the
one
on
the
far
side
there.
So
this
is
one
that
runs
z/os,
a
bunch
of
other
operating
systems.
A
The
orange
one
is
the
Linux
one
machine
that
only
runs
Linux
and
that
is
an
example
of
a
four
frame
model,
so
it
takes
like
for
these
19-inch
ones,
reconfigures
all
the
bits
inside
and
in
like
a
fully
decked
out
four
frame
model
of
one
that
came
out
this
year.
You've
get
40
terabytes
of
RAM,
a
bunch
of
PCIe
control
units
across
like
12
drawers,
and
these
things
are
huge.
I'll
show
you
a
picture
in
a
minute,
and
then
you
have
like
eighty-five,
logical
partitions.
A
A
big
thing
here
is
that
they're,
not
x86,
depending
on
who
you
talk
to
the
architecture,
is
IBM
Z
or
Z
architecture
on
Linux
you're,
most
likely
to
see
the
term
s/390
X,
but
they're
the
the
fastest
commercially
available
processors.
So
super
computers
are
faster,
but
these
are
the
ones
that
you
can
buy
so
5.2
gigahertz
processors.
There
are
like
a
hundred
and
ninety
of
them
in
a
fully
decked
out
system,
and
they
have
a
twelve
cores
per
chip
which
you
can
sort
of
see
on
there.
A
A
This
is
what
the
inside
before
frame
one
looks
like
this
is
the
air-cooled
one
there's
also
a
water-cooled
one,
but
you
can
see
like
in
the
center
that's
hard
to
see.
There's
like
no
good
color
to
put
on
this
chart
the
middle
ones.
Those
are
all
your
processors
and
then
your
all
your
I/o
drawers,
so
that
has
like
your
HSMs
and,
like
all
your
connection,
to
storage,
because
this
does
not
have
storage.
This
is
like
one
big,
massive
computing
machine,
oh
yeah
and
there's
like
Network
and
stuff
on
the
top.
So
this
is
a
storage.
A
This
storage
unit
was
released
last
letter
in
September
alongside
the
new
Z
15.
This
one
has
almost
6
petabytes
of
storage,
which
is
a
lot
there's
like
different
storage
models
somewhere
like
SSD,
and
some
of
them
are
spinning
disks.
But
this
one's
like
the
top-of-the-line
s,
d1
with
six
petabytes.
It
has
like
its
own
operating
system
that
manages
the
disks
on
this
machine,
and
so
traditionally
this
is
what
they
run.
So
z/os
is
kind
of
like
the
mainframe
operating
system.
If
you
think
of
mainframes,
that's
what
you're
gonna
think
about
there's
also
zvm.
A
A
Zp
TF
is
the
one
that
all
of
your
airlines
will
use,
except
for
Southwest
they're,
like
the
one
in
her
line
that
doesn't
use
mainframes,
but
all
the
rest
do
so
it's
for
like
special
purpose
for
like
high-volume
transactions
and
then
there's
Linux.
So
Linux
has
been
running
on
the
mainframe
for
about
twenty
years.
A
So
I'm
sure
some
people
just
came
here
to
satisfy
morbid
curiosity,
but
about
half
of
you
said
you've
actually
seen
more
mainframes
and
have
them
in
your
organization.
So
you
have
a
mainframe,
but
you
want
to
do
all
that
DevOps
e
stuff
right.
You
want
containers,
you've
heard
of
this
thing
called
kubernetes
and
you
want
to
modernize
everything
and
make
everything
awesome.
A
A
Looking
back,
if
someone
had
just
explained
to
me
that
mainframes
sit
in
this
space
between
like
racks
of
x86
hardware
and
supercomputers,
like
that,
would
have
been
a
good
description
for
me,
but
that
is
where
they
are.
They
are
really
super
powerful
machines,
and
one
of
the
things
I
came
to
the
realization
of
in
these
past
few
years
is
that
when
I
got
into
Linux,
so
I
started
playing
around
with
Linux
like
2001
got
my
first
systems
engineering
job
in
2006.
A
So
like
this,
at
this
time,
Google
had
already
come
into
the
market
and
said
everyone.
It's
great.
You
can
get
rid
of
all
your
old
big
machines
and
just
use
a
whole
bunch
of
x86
hardware,
they're
cheap
and
they
break
all
the
time,
but
they're
really
cheap.
So
we
can
just
string
them
all
together,
and
so
this
is
what
I
grew
up
in
as
a
Systems
Administrator
and
that's
kind
of
what
I
thought.
A
The
whole
world
was,
but
now
I'm
looking
at
mainframes
and
things
and
I'm
like
actually
there's
really
good
Hardware
out
there,
that
a
lot
of
people
are
just
ignoring
because
Google
said
it's
better,
just
to
use
a
lot
of
cheap
stuff
and
one
of
the
things
that
we've
also
been
finding
recently.
Is
it's
no
longer
cheap
stuff.
A
So
some
of
the
big
things
about
the
mainframe,
though,
is
like
there's
this,
like
enterprise-grade
storage,
is
totally
a
thing
and
not
to
take
away
from
the
great
projects
that
are
doing
distributed.
Storage
like
Vitesse,
which
just
graduated
there
are
places
for
this
enterprise,
like
for
the
for
these
distributed
systems,
for
storage
to
be
and
I,
think
it's
awesome
that
this
technology
is
out
there.
But
it's
not
the
answer
for
everyone
for
one.
It
takes
administrative
burden.
A
You
have
to
have
people
who
know
how
to
run
these
distributed
systems,
and
then
you
have
to
make
sure
it
doesn't
break,
because
losing
data
is
bad,
but
mainframes
I
mean
like
when
I,
when
I
look
at
some
of
the
discussions
on
the
Ceph
mailing
list.
It's
kind
of
funny,
because
I
now
that
I'm
getting
familiar
with
the
enterprise
grade,
storage,
stuff
I,
will
see
a
pape
like
a
white
paper
from
like
1983,
and
it's
like
the
same
discussion
as
they're
having
on
the
set
mailing
list
and
I'm
like
you
guys.
A
This
was
solved
already
you're,
not
gonna,
solve
it
over
beers
tonight.
So
like
Enterprise,
great
storage
is
a
real
thing,
and
so,
if
you
want
to
just
don't
like
drop
in
a
giant
box,
that
does
your
storage
and
you
don't
have
to
worry
about
it
like
a
mainframe
can
attach
to
one
of
those
I
mentioned.
The
processors
are
really
fast.
We
like
to
say,
there's
five
nines
of
uptime,
but
during
the
Z
15
release
there
was
like
independently
verified
numbers.
A
They
were
saying
seven
nines,
but
since
that's
not
really
a
thing,
but
like
these
machines,
don't
go
down.
There's
this
video
on
YouTube,
you
can
look
at,
you
can
just
look
up,
earthquake
mainframe
and
they
take
a
mainframe
and
they
lift
it
up
on
the
big
thing
and
they
shake
it
and
see
what
falls
out
so
they're
like
tested
to
be
like
up
to
like
8.0
earthquakes,
so
it
like
Fukushima
when
the
earthquake
and
the
tsunami
hit.
A
A
A
Other
things
so
like
one
of
the
other
things
I've
seen
in
these
infrastructure
projects
where
they're
going
to
modernize
is
networking
is
a
huge
pain
like
you.
Don't
really
think
about
it.
So
much
when
you're
doing
your
plan,
but
like
the
connection
between
VMs
and
the
mainframe
they're
already
configured
super
fast.
So
you
still
need
to
worry
about
like
container
networking,
but
you
don't
necessarily
have
to
work
worried
about
like
the
VM
communication
or
anything
that's
on
the
mainframe
itself.
A
So
one
of
my
favorite
things
is
that
when
we
ported
golang
to
this
architecture,
we
actually
made
it
so
golang
can
access
the
hardware
crypto.
So
if
you're
using
crypto
and
golang
on
a
mainframe
on
linux,
it's
crazy
fast.
So
that's
really
cool
and
the
sales
guys
love
talking
about
the
hsm
like
I,
don't
know
HSM,
it's
never
been
impressive
to
me,
but
apparently
it
is
the
industry's
best.
A
If
you
mess
with
the
H
at
the
harder
security
module
like
able
to
zero
itself
out
and
not
let
you
do
anything
and
of
course
they
run
Linux
and
they've
been
running
Linux
for
over
twenty
years.
This
is
one
of
the
things
I
didn't
really
know
about.
So
in
1998
there
was
a
community
effort
that
came
out
and
they
were
playing
with
Linux
on
the
mainframe.
Now
1998
is
like
I
will
install
Linux
on
everything.
A
Ibm
had
already
been
working
about
this
and
working
with
this
internally
because
they
were
already
on
the
Linux
train,
like
IBM's,
been
contributing
to
Linux
since
the
mid-90s,
so
they
were
like
oh
I,
guess
we
have
to
tell
everyone
about
this,
because
now
the
community
is
like
scooping
us
so
in
December
of
1999.
So
almost
exactly
20
years
ago,
IBM
came
out
with
the
first
patches
for
the
Linux
kernel.
A
I
think
was
like
a
2
to
kernel
so
a
long
time
ago
and
then
in
October
of
2000,
the
first
version
of
SUSE,
a
Enterprise
Linux
came
out
and
it
was
for
the
mainframe
and
it
was
called
SUSE.
A
Enterprise
Linux
was
because
it
was
for
the
mainframe.
Now
there
was
like
lots
of
rebranding,
shuffling
and
stuff,
and
eventually
they
wanted
to
use
Enterprise
Linux
for
all
of
their
things.
But
that
means
they
had.
That
means
that
that
Souza
Enterprise
Linux
is
mainframe
native
and
x86
is
the
port.
A
It's
not
really
true,
but
that's
the
story
I'm
telling
so
the
x86
version
of
Susannah
Prez
Linux
came
out
in
2001,
so
I
say
you
have
to
recompile
software,
and
that
has
been
done.
There's
a
so.
This
is
just
like
our
glamour
slide
of
like
a
ton
of
the
projects
that
have
been
ported
to
s
390x
or
is
the
architecture.
So
when
we
talk
about
running
Linux
on
it,
you're
running
Red,
Hat,
Susy
or
a
bun
to
some
of
the
community
projects
have
access
to
mainframes
to
build
it.
A
A
So
anyway,
if
you
looked
in
the
list,
kubernetes
is
in
there,
and
this
is
just
a
screenshot
you
can.
You
can't
read
this
of
the
the
download
page
for
the
cou
Bernays
project
and
there
are
s.
390X
builds.
So
if
you
aren't
familiar
with
what
s
390x
is,
it
is
a
mainframe
and
kubernetes
builds
for
it.
They
also
build
for
power
and
arm.
So
you
can
actually
just
go
and
download
the
client
binaries
and
the
server
binaries
directly
from
them,
but
like
realistically
like
who
actually
does
this?
A
Who
is
running
like
some
mainframe,
so
suzay
has
an
investment
in
it.
There
was
this
interesting
article
that
was
published
in
March
about
running
the
latest
version
of
kubernetes
on
sleds
12,
so
sousei
12,
which
is
came
out
a
few
years
ago,
but
running
Eligius
latest
version
of
kubernetes,
because
suzay
12
is
old
and
the
newest
version
of
kubernetes
did
not
come
with
it,
and
so
the
interesting
thing
about
this
article
was
they
were
like.
A
A
So
this
is
a
blog
post
that
was
announced
by
IBM.
Like
my,
he
was
last
month
where
they
were
talking
about
how
IBM
has
pretty
much
gotten
rid
of
a
lot
of
their
container
efforts
that
were
doing
internally
and
switched
entirely
to
kubernetes,
and
this
includes
Z,
so
OpenShift
is
being
ported
to
Z
they're,
integrating
it
into
their
plan
for
like
IBM
cloud
packs,
which
make
like
getting
a
lot
of
the
IBM
software.
Much
easier
on.
Openshift
Ubuntu
has
their
cdk
the
canonical
distribution
of
kubernetes,
and
so
there's
this
Ubuntu
on
big
iron
blog.
A
If
you're
interested
in
mainframes
and
Ubuntu-
and
you
haven't
seen
this
blog,
it
is
awesome.
It's
super
technical,
it's
written
by
a
guy
at
MIT,
canonical
who
used
to
work
on
mainframes
and
then
came
to
canonical
and
now
works
on
mainframes
again,
but
he
wrote
posts
tons
of
really
technical
articles,
so
he's
talking
about
using
Canonical's
deployment
technology
to
install
kubernetes
on
the
mainframe,
and
he
sent
me
over
this
slide,
and
so
you
see
how
you
install
it.
A
So
you've
got
your
L
par
running
on
your
mainframe
so
and
then
you
install
ubuntu
server
on
the
L
par.
So
that's
the
logical
partition.
That's
you
know
the
space
they
have
to
install
stuff
on
instead
of
using
docker,
canonical
and
Ubuntu
is
very
invested
in
lxd,
so
they're
using
lxd
as
their
container
environment.
And
then
you
see
all
the
familiar
services
from
kubernetes.
You've
got
your
s
CD
and
all
your
workers
and
and
everything.
But
this
is
pretty
much
what
it
looks
like,
and
this
is
specifically
for
Z.
A
Of
course
the
x86
version
looks
very
similar,
but
it's
not
running
on
an
L
par
there's.
Also,
a
few
consulting
services
that
are
doing
kubernetes
on
the
mainframe
I
want
to
say
that
I
don't
actually
know
how
to
say
the
name
of
this
company,
because
I
only
talked
to
them
on
email
and
like
mailing
lists
and
stuff
yeah
see.
A
Oh
all
right
yeah,
but
the
great
thing
about
these
guys
that
they've
been
they've
they've,
been
working
with
open,
shipped
origin
and,
like
they've,
been
offering
consulting
services
for
that
and
also
being
super
active
in
the
community.
So
they
give
away
a
bunch
of
their
like
PDF
guides
on
how
to
install
open
shift
origin
on
Z,
and
they
are
also
really
active
on
the
there's.
A
there's,
a
Linux
mailing
list
for
us
390
and
they're
super
active
or
anytime.
A
Anyone
acces
asks
about
kubernetes
and
when
I
was
preparing
for
this
talk,
one
of
the
one
of
like
the
research
things
I
did
was
reach
out
to
the
mailing
list
and
say:
hey.
You
know,
who's
who's,
running
kubernetes
and
tell
me
your
stories
and
they
reached
out.
They
were
like
here's
all
our
PDFs
and
they're
good
for
good
folks.
I
also
discovered
ICU
IT
services
out
of
the
Netherlands
and
the
representative,
who
got
back
to
me
from
this
company.
They
were.
A
They
were
saying
that
they're
they're
helping
our
customers
integrate
with
traditional
z/os
environments
and
new
cloud
environments,
and
that's
what
they're
doing
with
kubernetes
so
they're
taking
kubernetes
and
the
entire
ecosystem
of
software
that
they
are
compiling
for
Z
or
have
already
been
compiled
and
using
that
to
help
their
customers,
modernize
but
they're,
still
using
z/os
and
and
other
other
private
environments
and
the
reason
they
all
want
to
use.
This
is
the
same
reason.
We
all
want
to
use
kubernetes
right.
They
want
to
do
contain.
A
They
want
to
use
containers,
they
want
to
use
micro
services.
They
want
that
orchestration,
because
once
you
get
more
than
like
a
dozen
containers,
it's
a
giant
mess
and
they
want
access
to
the
big
ecosystem
of
services
and
partners
in
there
and,
as
we
heard
from
the
one
consulting
service
like
they
want
to
integrate
with
the
traditional
z/os
environments.
A
So
they
want
to
run
their
containerized
workloads
alongside
all
of
their
programs
that
are
written
in
assembler
and
COBOL
and
like
everything
that
actually
runs
their
business
and
actually
runs
really
well
and
they
want
it
close
to
their
data.
So
they
run
kubernetes
on
the
mainframe,
that's
sitting
next
to
the
other
mainframe
that's
running
z/os
or
on
the
same
one,
because
you
can
mix
like
Linux
and
z/os
on
the
same
mainframe
and
then
attach
to
the
storage.
So
they
have
really
fast
access
to
any
of
the
data
that
they
want
to
get.
A
So,
in
the
one
example
like
they
want
to
be
accessing
db2
on
z/os
or
Oracle
running
on
Linux,
on
Z
and
having
their
micro
services
on
the
machine.
The
same
machines,
a
big
deal,
they're
also
really
interested
in
leveraging
the
hardware
crypto
we
heard
in
the
keynote
yesterday
that,
like
Intel's,
totally
on
board
with
doing
end-to-end,
Hardware
crypto
like
x86,
is
definitely
getting
there,
but
mainframes
been
doing
it
for
quite
some
time.
A
So
like
this
end-to-end
encryption
thing
is
a
really
big
deal
for
us
and
the
mainframes
actually
do
it
really
well
and
one
of
the
things
that
actually
fell
into
our
laps,
it
seems
like
is
blockchain.
So
I
gave
a
history
of
Linux
Anzhi
talk
back
in
July
for
a
Linux
users
group
in
Philadelphia
and
when
I
gave
him
my
little
talk.
Abstract
I
was
like
here's.
A
Another
thing
that
people
tend
to
like
it
for
is,
if
they're
going
for
some
sort
of
hybrid
cloud
approach,
so
they
want
to
run
the
same,
say
docker
container
across
like
a
bunch
of
different
environments,
and
so
you
know
when
the
next
like
meltdown
or
Spectre
hits
and
their
x86
machines
take
a
performance
dive.
They
can
offload
some
of
their
workload
to
a
different
architecture.
This
is
like
the
same
story
if
you're
talking
about
arm
or
power
or
any
of
these
other
architectures,
it's
becoming
increasingly
important
to
diversify
your
infrastructure.
A
If
you
actually
need
a
lot
of
resiliency
in
here
in
your
system
and
like
I
love
I,
also
like
love
geeking
out
over
alternate
architectures
I
have
like
a
spark,
an
ultra
spark
in
my
office.
That's
not
running
at
the
moment.
I
have
to
install
BSD
or
something
on
it,
but
I've
played
with
it
a
lot
in
the
past,
and
my
husband
has
a
whole
bunch
of
like
MIPS
boxes
in
the
garage
and
it's
a
big
crazy
place
in
my
house.
A
A
But
you
know
x86
has
been
a
huge
heart
target
because
of
these
specific
issues
that
have
been
coming
up
these
past
few
years,
so
I
kind
of
want
to
wrap
up
and
leave
some
time
for
questions
that
hopefully
I
can
answer.
I
just
started
in
April
be
gentle,
but
some
of
my
suggestions.
So
you
know
coming
back
to
our
story.
Like
you
know,
the
new
DevOps
team
came
in,
they
couldn't
complete
their
project
cuz.
They
still
will
have
the
mainframe
they're
very
sad,
but
this
is
not
actually
a
sad
result.
A
You
are
leveraging
the
strengths
of
kubernetes
and
the
strengths
of
your
mainframe
and
the
point
is
not
to
like
be
cool
and
modern
and
modernize
everything.
The
point
is
to
deliver
the
business
need
to
your
customer
and
if
a
mainframe
is
the
thing
that
is
delivering
that
need
and
handling
your
data,
and
it's
doing
really
well
and
it
has
been
for
30
years.
There's
actually
no
reason
to
change
that.
So
one
of
the
things
you
should
do
is
when
you
start
this.
A
Modernization
project
and
you're
diving
into
kubernetes
is
get
the
mainframe
team
out
of
hiding.
Sometimes
they
won't
like
it
because
they
feel
like
they're
being
replaced
or
they
just
are
set
in
their
ways.
They
haven't
had
to
modernize.
They've
been
doing
the
same
thing
for
a
long
time,
but
including
them
in
the
process
is
really
really
important
and
so
play
to
those
strengths
and
integrate
it
into
your
plans.
So
consider
the
mainframe
like
what
is
it
good
for
you
know
what?
What
would
what
talent
are
we
investing
in
in
our
organization?
A
Do
we
want
to
run
a
distributed
storage
operation,
or
do
we
just
want
to
stick
it
on
a
big
box
that
has
like
six
petabytes
and
so
work
that
into
your
plan
and
there's
also
a
project?
There's
the
open
mainframe
project,
part
of
the
Linux
Foundation
they've,
been
putting
out
some
really
interesting
software
around
z/os.
A
So
there's
a
project
called
Zoe
zo
w/e,
which
puts
like
a
consolidated
API
across
a
bunch
of
mainframe
services
and
so
you're
able
to
run
like
there's
like
a
CLI
client
that
you
can
run
things
against,
and
so
you
can
run
your
Jenkins
jobs
against
things
in
the
mainframe,
which
is
pretty
cool,
because
then
you
can
just
integrate
it
into
a
CI
CD
pipeline
and
you
don't
even
need
to
change
anything.
I
mean
you
have
to
run
Zoe
but,
like
you,
don't
have
to
change
things
on
the
mainframe
site
itself.