►
From YouTube: Values of IBM Z 20180927
Description
Ever wonder why IBM Z mainframes have been such an attractive and successful platform for modern enterprises? In this talk, Iris will discuss IBM Z’s strengths in performance, scalability, security, isolation and RAS.
A
So
I'm
waiting,
okay,
so
up
yeah
but
also
know,
don't
know
me:
I'm,
Irish,
Bell
and
I'm,
a
technical
advocate
for
runtime
Swansea
and
in
my
role,
I
typically
talk
to
customers
on
these
are
dedicated
and
have
big
businesses
on
Z
and
trying
to
advocate
at
one
time
talking
about
Java,
node
and
Swift
today
on
Z
and
whether
they
want
to
migrate,
extend
modernize
their
application
with
those
one
time
and
now
I'm,
switching
heart
kind
of
and
I
do
the
opposite.
I'll
talk
to
the
runtime
steam
about
the
values
of
V.
A
A
A
That
explains
what
is
the
mainframe,
but
it's
a
I
like
it,
because
it's
very
unformatted
I
took
it
from
another
marketing
slide
for
Z
and
and
let's
see
what
it
is,
it's
basically
an
integrated,
highly
scalable
computer
system
allowing
many
different
types
of
workloads
running
simultaneously,
showing
same
information
as
needed
with
protection
very
large
amounts
of
information,
many
users
with
security
without
user,
experiencing
any
failure
in
service.
So
what
does
it
mean
all
together?
Basically
we're
talking
about
a
huge
server
that
can
do
the
scalability.
It
can
handle
many
types
of
workloads
to
many
users.
A
So
some
people
think
that
a
set
aside,
the
mainframe
I'll
dead-
you
pick
somebody
the
street
and
you
heard
about
Anthony
was
totally
sir
really
do
you
still
remember
and
the
truth
is
we
really
do
and
it's
a
life
and
it's
a
life
more
than
you
think
cook.
It's
probably
any
one
of
you
are
touching
Jersey
every
day
in
your
life,
whether
you're
going
to
an
ATM
you're,
doing
credit
card
you're
doing
online
shopping.
A
Many
many
aspects
of
this
is
a
booking.
For
example,
Hotel
flight
typically
is
v
application,
so
we
can
see
on
the
left
here
with
the
most
industry
so
top
of
the
world
by
worldwide
bang
top
insurances
top
retailers
are
running
Z.
So
it's
really
and
we
can
see
a
line
here.
So
it's
airline
tip
goes
to
the
Medicare
environment.
Even
sometimes
you
don't
know.
A
For
example,
you
could
do
good
for
distillation
in
your
tea
application
runs
on
Z,
you
didn't
know
it
right,
and
so,
but
what
we
can
see
here
is
that
it's
a
lot
of
business,
critical
applications
and
on
the
right
we
can
see
the
volume.
So
we
are
talking
about
huge
data
and
for
those
businesses
29
billions
of
ATM
transaction
annually.
A
lot
seven
point:
seven
trillion
credit
card
payments,
$1
that's
more
than
the
Canadian
GDP
altogether
annual,
probably
five
times.
A
So
it's
a
lot
of
volume
and
also
businesses
kicks
transaction
kicks
it's
middleware,
one
Engels
is
doing
online
transactions,
mostly
in
businesses,
but
also
others,
industry,
1.3
million
transaction
every
second,
globally,
much
much
more
than
the
Google
searches
that
you
can
do
that
is
participating
globally
per
second,
so
the
volume
and
magnitude
of
the
business,
it's
huge
and
that's
what
we
need
to
know
and
in
fact
they
say,
80
percent
of
the
corporate
data
worldwide
is
a
result
of
a
region.
It's
on
thee,
so
it's
off
everywhere.
A
So
that's
a
a
perception
that
was
what
is
wrong
if
you
would
think
that
it's
that
it's
completely
not
that
and
then
some
outline
for
the
talk,
we'll
talk
about
performance.
We
are
all
related
to
this
here
and
is
known
to
be
very
performant.
We'll
see
why
we'll
talk
about
the
cost
benefits
and
the
three
key
factors
that
we
I
mentioned
before
the
quality
of
service
flexibility,
scalability
and
security?
Okay,
so,
let's
start
with
performance,
don't
get
scared
here.
A
I
showed
the
latest
five
releases
of
movies
over
the
last
decades,
and
maybe
we
should
start
by
saying
that
every
two
or
three
years
there's
a
new
release
of
the
hardware,
so
it's
life
with
the
people
I
can
invest
in
it
a
lot
and
becoming
better
and
better
with
every
release.
That's
coming
up.
The
latest
one
is
that
14
already
a
year
more
than
a
year
now
available
out
there
and
I
showed
underneath
each
one.
Some
of
the
features
that
came
out
with
those
releases
I
didn't
list
all
of
them.
A
There
are
many
others,
but
I
want
to
touch
on
the
ones
that
we
relate
to,
and
one
thing
to
notice
is
that
some
of
the
innovation
leading
innovation
technology
in
the
industry
came
out
from
these,
for
example,
some
suction
memory
in
ac-12
two
releases
ago.
The
later
that
13
came
out
is
published.
Dc
I
think
java
application
having
better
response
time
to
strict
SLA.
So
that's
something
completely
new
and
I.
A
Maybe
even
underneath
that
so
same
D,
for
example,
is
one
example
that
we
exploit
in
our
date
and
it's
transparent
to
the
application.
Smt
is
another
example:
well
we
don't
exploit
it,
but
it
feels
mainly
for
Java
applications
or
others,
but
Java
takes
a
big
benefit
of
it,
and
if
we
look
at
at
the
bottom,
we
can
see
some
other
architectural
measures
so
clock
speed.
Everybody
knows
he
has
very
big,
very
large,
fast
clock
speed.
So
it's
around
5
gigahertz.
Now
as
a
platform
which,
to
this
clock
did
I
know
it's
not
it.
A
It's
not
the
only
thing
that
matters,
but
it
does
matter
it
matters,
especially
because
we
have
instructions
which
can
run
single
cycle.
So
you
got
the
speed
out
of
that
I
listed
here.
The
cache
size
I'm
not
going
to
go
over
all
of
them,
but
there
is
an
increase
in
cache
size
all
throughout
the
l1,
l2
and
l3,
and
the
reason
why
is
it
important?
Because
we
are
dealing
with
more
and
more
data
and
we
want
to
make
sure
that
we
access
them
and
handle
this
data
very
fast.
A
So
that's
one
thing
that
we
can
do
with
that.
Now:
let's
look
a
little
bit
higher
level.
We
look
at
the
system,
level
performance
and
it's
a
quadric
lateral
view
of
their
releases
that
they
show
before.
But
now
we
look
at
them
from
how
much
memory,
how
many
cores
you
can
see
what
the
bandwidth
PCI
is
basically
like
meet
IBM
X
and
we
see
the
growth
through
all
dimensions
and
we
can
see
that
14,
for
example,
can
fit
170
course
that
can
be
used
by
customers
with
extremely
large
server.
You
can
think
of
an
X.
A
It's
a
much
much
bigger
need.
Many
access
to
accommodate
same
amount
score
memo
is
the
same
band
the
same,
so
we
grow
in
all
direction,
and
why
is
that
important?
Because
today
was
the
businesses
grow
with
the
mobile?
With
Internet
of
Things,
they
want
to
do
analytics.
They
want
to
do
all
in
one
place
so
for
consolidation,
so
this
gives
them
the
ability
to
do
that
and
that's
why
they
like
it.
Another
way
to
look
at
performance
is
comparing
throughput.
We
all
know
that
right.
A
So
here's
a
sample
point
I,
didn't
taste
all
of
them.
There
are
many
out
there.
I
did
pick
three,
but
the
composition
is
between
Linux
1
and
PI
2
versus
equivalent
x86.
Maybe
I'll
start
with
what
is
Linux
1
Linux
1
is
a
line
of
servers
that
is
built
on
top
of
mainframe
hardware,
but
designed
to
run
only
Linux
world
workloads
so
same
how's
about
running
Linux
and
/.
2
means
that
they
it's
based
on
z14
hardware,
so
the
latest
one
and
we
compare
for
example,
why
we
compare
Acme
air.
A
For
now,
with
known
we
compare
a
way
ycs
big
with
MongoDB
and
in
all
of
them
we
so
like
two
to
two-and-a-half
times
better
throughput
per
core,
so
that's
very
impressive
by
itself
and
other
proof
point
showing
the
same
thing.
So
it's
not
like
I
pick
the
best,
it's
kind
of
consistent,
and
so
we
saw
the
performance
input
in
all
kinds
of
aspects,
and
we
know
it's
performance.
A
But
let's
see
what
else
why
customers
really
like
it
in
other
aspects
and
before
we
do
that
I
do
want
to
talk
about
cost,
because
I
I
think
many
people
have
all
kinds
of
ideas
of
about
the
cost
of
V
and
perception,
and
so
let's
talk
about
it
and
when
enterprises
are
doing
cost
analysis
and
choosing
the
right
platform,
they
need
to
take
into
account
many
factors.
So
one
way
to
look
at
it
and
say
only
tch-tch
is
a
total
cost
acquisition.
So
it's
only
hardware
and
software,
but
that's
not
enough.
A
And
let's
see
what
came
up
so
as
I
said,
the
case
study
is
like
looking
at
12
instances,
so
they
need
three
that
they
need:
24,
general
VMs,
but
they're
not
important
and
what's
important,
so
that
is
equal
comparison
and
they
break
down
the
cost
into
three
parts
infrastructure
which
is
basically
the
cost
of
hardware
in
memory
or
where
hypervisor
storage,
low
thing
middleware,
which
is
a
software
and
licensing
and
labor
on
top.
So
what
do
we
see
on
the
right?
We
have
Linux
1
&
/.
A
If
we
look
only
on
infrastructure,
yes,
it
is
expensive
much
more
than
if
you
look
connects
on-premise,
so
you
can
hardly
see
the
infrastructure
here,
it's
low,
but
it's
very,
very
small
compared
to
the
see
how
the
cost
for
infrastructure
on
public
cloud
Lee.
It's
also
a
big
jump,
but
that's
probably
because
it's
the
M
propriety
and
also
the
ability
to
grow
and
agility.
Probably
that
accumulate
for
the
price
and
in
terms
of
middleware.
We
see
that
on
X
it's
much
more
expensive
than
the
licensing
on
Linux
one.
A
We
are
looking
at
less
cost
to
do
the
same
work
on
Linux
one.
So
that
might
be
a
reason.
Labor
is
kind
of
the
same,
so
yes,
infrastructure
is
high.
But,
overall,
when
we
looked
at
three
years
calculation,
we
see
the
solution
on
these
actually
the
cheapest.
So
by
far,
it's
not
like
a
little
bit
by
far
it's
like
46%,
less
than
on-premise,
with
X
and
57%
lesson
on
the
cloud
solution.
So
that's
something
to
take
into
account
too
right
when
we
do
the
calculation.
A
Another
very
compelling
number
to
remember
about
cost
68
for
six
rules.
So
this
came
out
from
customers
right,
which
says
not
one
but
many
if
with
the
meant
out
of
all
the
workloads
that
they
have
68%
one
from
V,
but
the
cost
out
of
all
the
Ticos
only
6%
goes
overseas,
so
that's
really
impressive
and
it's
a
big
motivator
for
the
Z.
So
what
we
take
out
of
here?
A
So
now,
let's
remember
the
three
key
point
that
I
mentioned
at
the
beginning
about
what
is
a
mainframe:
let's
see
how
they
play
and
what
customers
like
that.
So
we'll
start
with
flexibility
and
scalability
and
chosen
has
a
beautiful
picture
very
colorful
right.
So
what
I
want
to
to
per
say
that
on
V
you
can
have
two
levels
of
virtualization.
A
So
if
we
looked
on
this
picture,
the
FIR
Umbro
is
the
cps
that
calls
themselves.
On
top
of
it.
You
have
the
first
level
of
virtualization,
which
is
called
logical
partition.
So
remember
we
have
tons,
of
course,
and
now
we
we
have
the
option
to
a
partitioning
and
virtualize
it.
So
there
is
a
lot
of
control
and
comparable
configurable
m
with
the
El
Paso.
Each
one
of
alpha
is
basically
kind
of
a
full
system
by
itself,
and
you
can.
A
It
has
dedicated
memory,
but
other
resources
can
be
either
shared
or
not
and
do
some
load,
balancing
and
cetera.
If
one
helper
comes
down,
nothing
happens
to
the
other
apart.
So
when
was
the
first
level
and
this
important
say
that
L
Pass
has
EA
L
5,
plus
certificate
evaluation
assurance
level.
What
does
it
mean?
It's
a
standard
with
measure
security,
5
being
very
high,
like
other
smart
cards,
may
be
other
OSS
and
v.
A
Arms
are
typically
less
with
EA
l
4
and
what
what
it
means,
basically,
that
with
el5
the
isolation
of
work
is
equivalent
to
put
into
this
work
on
completely
different
systems.
So
you
can
guarantee
the
privacy
and
security
of
the
applications
at
one
on
air
part,
not
to
have
anything
to
do
with
something
that
I'm
from
a
differential
power.
So
that's
very
important
because
that
gives
us
a
village
with
capability
to
run
many
types
of
application
in
parallel
with
many
users
right
without
compromising
the
security
into
multiplication,
simple
medical
stuff.
A
A
As
I
said,
there
is
another
level
of
virtualization.
On
top
of
the
del
per,
you
can
put
a
either
the
legacy
that
be
I'm,
okay,
VM
and
one
different
OSS,
and
you
can
scale
very
large.
You
can
get
to
thousands
of
virtual
Linux
guests,
for
example,
the
13
I
remember
said
8,000,
probably
the
14
is
more
two
million
dollar
containers.
Is
that
14?
A
That's
enormous
number
so
when
and
you
can
scale
you
can
scale
or
isn't
a
little
scale
vertically
can
virtualize
a
lot
and
this
picture
kind
of
shows
where
all
kinds
of
configurations
you
can
have
so
there's
about
this
side
of
scalability.
But
there
is
another
aspect
of
it,
so
whether
there
is
a
lot
of
course,
but
people
buy
them
and
they
can
actually
activate
only
some
of
them
and
keeping
the
rest
in
active
integer
is
peak
in
demand.
So,
if
you
think
about
the
cell
cycle,
it's
not
flat
right,
the
demand
is
not
flat.
A
There
are
jumps
and
Peaks
and
seasonal,
or,
for
example,
cyber
monday
singles
day
in
China.
So
those
means
extra
capacity
now
to
handle
all
the
demands
that
comes
in,
and
then
there
is
a
concept
of
on
demand
so
automatically
without
an
interruption
for
the
application
it
kicks
in
the
inactive
course.
Basically,
moreover,
the
some
calls
that
are
spare
and
if
there
is
a
failure
automatically
again
it's
all
a
matter
of
without
any
destruction
for
the
application
level,
it
will
kick
out
on
the
course
so
that
kind
of
a
resiliency
for
trailers.
A
So
let's
look
a
little
bit
what
x86
looks
like
so
we
know
that
from
Z
we
can
fit
a
lot
of
stuff
and
you
can
do
the
same
for
X
with,
but
he
will
need
many
servers.
Probably
you
will
have
an
eye
provider.
You'll
have
several
guests
on
top
of
them
and
they
will
talk
with
network
doable.
Yes,
but
it's
a
very
rigid
solution,
so
you
have
to
define
ahead
of
time
number
of
course,
number
of
I/o
memory.
Everything
not
like
we
talked
before.
A
There
is
no
concept
of
on
demand,
so
much
no
partition,
definitely
not
the
security
of
isolation.
So
if
you
want
isolation,
put
it
on
another
machine
right,
it's
less,
there
is
no
spare
course.
So,
that's
reliable!
You
can
pull
them
and
you
manage
them,
but
it's
not
so
out
of
the
box.
I
convey
utilization.
If
you
want
to
plan
for
peak
demand,
you
need
to
reduce
your
utilization
right
to
have
capacities
are
talking
about
utilization.
V
is
planned
to
run
with
90%
utilization
and
up
on
each
course
which
is
allotted.
A
X
is
not
designed
like
that.
Not
to
talk
about
network
that
there
is
risk
from
meeting
things
on
the
network
on
V.
You
can
think
about
it
like
a
full
data
center
in
one
machine
of
the
consolidation
matter
in
terms
of
security
and
performance
and
management,
of
course,
so
that
that's
how
we
can
reach
to
having
so
many
application
running
on
same
system,
many
many
users
without
compromising
the
security
and
and
there
the
business
applications.
Users
I
use
easy,
have
a
demand
for
such
a
big
capacity
like
I.
A
Just
yesterday,
wasn't
call
with
the
customer
that
took
for
200
CPS
for
their
application,
so
they
do
need
those
capacity.
It's
not
just
there
for
nice
to
have
so.
We
talked
a
little
bit
about
resiliency
of
fellow
call
but
different
levels
of
resiliency
here,
right,
here's
another
level,
so
this
picture
is
taken
from
Japan
2011
after
the
earthquake
and
it's
a
data
center
and
we
see
several
Z
systems,
some
of
them
laying
down
life
but
very
interesting
application,
continued
to
run
and.
A
It's
true
that
so
that's
the
power
of
these
you,
the
continuous
availability
and
serviceability
quality
of
service,
means
a
lot
to
our
customers
and
there
are
probably
most
concern
topics
that
they
want
to
Kern.
They
trust
me
for
this:
a
continuous
availability
and
quality
of
service,
so
talking
about
reliability
and
responsiveness.
Maybe
I
should
ask
I
should've
asked
maybe
before
anybody
knows
what
the
C
stands
for
in
IBM
V
and
it
yes.
A
Yes,
yes
exactly
well,
it's
pretty
obvious
on
the
slide,
but
that's
exactly
it.
So,
if
you're
down
a
down
time-
and
if
we
look
here
at
the
graph,
this
is
a
comparison
of
several
hard
work
done
by
ithc
in
2015,
looking
at
downtime
of
more
than
four
hours,
and
we
see
that
IBM
TN,
Linux
wanna,
with
zero
presented
as
very
low
and
actually
measures
between
failures
are
counted
in
decades
versus
commodity
platforms
with
each
month,
and
there
is
another
concept
of
them
5/9.
A
If
anybody
heard
about
it,
so
applications
are
designed
to
run
on
three
with
a
service
level
of
99%
99.999%
availability.
Just
to
give
you
what
it
means,
it
means
that
they
can
go
down
unplanned
downtime
less
than
five
minutes
per
year,
so
that's
equivalent.
But
it
is
that's
a
very
nice
and
customers
do
like
that
features
and
they
want
actually
to
increase
product
availability
and
here's
an
example
ICBC
it's
today,
the
largest
bank
in
the
world.
A
So
if
one
datacenter
completely
goes
down,
everything
will
automatically
move
to
the
other
one
and
continuous
to
run
without
any
interruption
in
servicing
application.
And
this
technology
of
continuous
availability
is
something
very
important
for
customers.
Could
you
think
about
business?
You
think
about
all
those
big
enterprises
with
every
downtime
counts
for
them
cause
it'll
either
lost
money,
lost
opportunity,
lost
customers
because
customer
will
not
trust
them
anymore,
and
so
they
do
care
a
lot
about
this
availability
and
the
solution
comes
from
the
level
of
data
center
having
two
servers
one
next
to
each
other.
A
It
can
go
regional
so,
for
example,
a
bank
sitting
in
terms
of
might
have
another
data
center
in
Mississauga,
just
in
case
for
disaster
recovery.
Balancing
Frances
facing
it
can
go
to
further
distance
like
we
saw
in
Shanghai
and
Beijing
with
ICBC,
and
it
can
go
up
to
multi
sites
up
to
32
servers
in
a
cluster
which
can
all
back
up
and
guarantee
the
full,
continuous
availability
of
applications
and
I
know.
Customers
are
using
that
features
and
building
the
full
map
of
clustering.
A
A
So
let's
talk
about
security
and
securities
and
other
things.
That
is
a
very
easy
say
famous
for
and
for
many
many
years
it
was
live
and
it
still
the
most
secure
platform
on
earth
with
the
slogan
and
that's
because
many
many
years
of
innovation
in
this
field
and
actually
the
hardware
and
software
are
combined
to
produce
the
best
security
possible.
A
So
it's
not
different
features,
but
they
all
work
together,
integrated
solution
and
when
we
talked
about
when
we
talk
about
hardware,
for
example,
and
that's
something
available
for
many
years,
there
is,
for
example,
a
core
processor
for
every
core.
That's
doing
accelerating
the
encryption
functions
called
super
keV.
In
addition,
we
have
crypto
cards
that
can
extend
external
cards
that
can
do
the
crypto,
Security
and
offload
those
in
work.
A
What's
interesting
that
the
security
rating
for
them,
it's
tips,
140
level.
So
for
those
who
know
what
it
is,
if
not
I'll
try
to
explain
it's
the
highest
level
of
security
possible,
which
means
that
it
has,
for
example,
tamper
resistant.
So
even
if
somebody
physically
going
and
trying
to
manipulate
this
card,
it
will
detect
it
it
automatically
erase
everything
and
stop
operating
so
that
the
highest
level
of
security
could
have
four
cards
like
that.
A
We
talked
about
logical
partitions
before
with
higher
security
for
isolation,
so
those
bill
is
built
up
from
done
up
and
on
this
verse,
14
on
today,
words-
and
we
talked
about
data
law,
data
becomes
a
very
important
factor,
and
it's
actually
one
of
the
top
valuable
resources
in
companies.
You
can
think
about
data.
What
do
these
companies
hold
in
their
data?
A
list
of
patients,
for
example,
medical
records,
accounting
information,
all
very
critical
data
for
data.
We
can
another
level
that
needs
a
completed
encryption
and
that's
what
the
reporting
time
with
pervasive
encryption.
A
Big
thing:
meaning
data
is
encrypted
everywhere.
All
the
time
I
need
to
feature,
and
you
can
turn
it
on
customer-
can
turn
it
on
without
again
any
disruption
in
application.
It
automatically
will
kick
up
and
this
goals
again
full-stack
solution,
so
data
at
rest.
It's
not
just
when
we
move
data
that
we
need
to
encrypt
it,
but
now
we
encrypt
it.
We
encrypted
filesystem.
A
There
is
an
option
to
encrypt
file
systems.
The
data
set,
for
example,
everything
with
storing
this
data,
and
we
talked
about
clustering,
so
all
the
information
that
they
exchange
to
control
this
clustering
and
recovery
and
continuous
availability,
whether
it's
locking
or
management
or
this
data
is
also
encrypted,
and
there
are
many
other
features.
I
don't
want
to
I
didn't
put
them
all,
because
there
are
many
others
really
many
other
software
solutions.
I
just
want
to
give
you
some
feeling
of
how
much
innovation
and
investment
is
done
in
security.
Other
point
is
secure
service
containers.
A
This
is
for
Linux
one.
So
this
M
it's
a
solution
to
take
your
installation,
always
whatever
it
is
that
you
want
to
run
application,
put
it
in
a
container
and
and
you're
probably
no
container,
but
this
is
a
special
container
because
it's
a
very
security
security.
Extreme
first
of
all,
everything
is
encrypted,
but
there
is
no
memory
access
to
it.
You
cannot
ssh
to
it.
A
A
So
what
do
I
wanted
to
take
without
about
many
things?
We
talked
about
performance,
we
talked
about
cost,
we
talked
about
quality
of
service,
scalability
security
right,
and
we
thought
how
great
these
are
also
suspect.
So
maybe
one
thing
is
to
take
this.
There
is
no
one
thing
that
shine
from
thee:
it's
the
accumulation
of
the
other
things
that
makes
this
so
powerful
and
that's
what
customers
are
looking
and
that's
what
they
like
about
thee
and
they
that's
why
they
choose
thee
for
their
solution.
A
A
Right,
it's
a
hypervisor,
it's
another
level!
Oh,
so
the
question
is
what
that
VM
is
in
the
picture,
so
it
and
we
can
one
that
always
on
top
of
it
and
we
can
run
Linux
on
top
of
it.
It's
another
level
of
hypervisor
and
separation
legacy
kind
of
yes.
Yes,
it's
all
about!
You
can
have
all
this
in
one
machine
here
so
that's
where
ages
ago,
and
then
they
want
to
run
Linux
and
they
want
to
able
to
scale
those
as
well.
So
that
was
the
solution.
Then,
before
Linux,
one
good
for.