►
From YouTube: Ceph Days NYC 2023: Ceph Crossing the Chasm
Description
Presented by: Vincent Hsu | IBM
The new generation of hybrid cloud provides a common platform across all your cloud, on-premises, and edge environments. That means you can skill once, build once, and manage from a single pane of glass. That also implies platform needs to support diverse workloads and different level of maturity in management skills. In this presentation, we will cover the open source projects and proposals to enhance Ceph's consumability and manageability to enable Ceph in more environments.
https://ceph.io/en/community/events/2023/ceph-days-nyc/
A
All
right
good
morning,
everyone,
my
name,
is
Vincent
shu
I'm,
an
IBM
fellows
and
CTO
for
storage.
I
am
a
new
in
Seth.
This
is
the
first
time
I
attend
the
Sev
event.
So,
first
of
all,
I'd
like
to
thank
the
Bloomberg
team
to
host
this
event,
I
just
I
Rush
here
from
South
France,
so
yeah
where
there's
different
wow,
it's
a
it's
a
business
meeting
anyway,
so
anyway,
so
I'm
glad
to
have
this
opportunity
to
to
be
invited
to
give
this
talk
as
I
say
that
I've
been
in.
A
A
You
know
some
people,
like
you
know
Potter
is
their
passion,
dance
and
hikings
storage.
Is
my
passions
actually
just
just
try
to
check
the
temperature
here
so
raise
your
hand
if
stories
is
your
passions?
A
Okay,
good,
so
I
tried
to
make
sure
that
IBM
team
who
didn't
raise
their
hands
Okay.
So
before
we
talk
about
this
crossing,
the
chasms
I
do
want
to
address
the
the
elephant
in
the
room.
So
in
case
you
missed
the
news
that
January
1st,
that
we
are
very
happy
to
you
know.
Welcome
Sev
team
joined
IBM
story,
divisions.
Okay,
we
thought
it's
a
very
natural
things.
A
I
I
never
expect
this
right.
So
on
the
internet.
There
are
a
lot
of
a
lot
of
comments
right
from
you
know,
ranging
on
all
kinds
of
rumors
from
you
know,
what's
the
IBM?
What
what
do
you?
You
know?
Those
are
three
letter
agencies
do
with
the
people.
I
was
like
it
is
the
first
time
I
IBM
being
considered
as
part
of
the
three-letter
agencies,
so
I
want
to
make
sure
that
I
did
not
modify
this
this
this
page.
A
This
page
is
official
communication,
page
that
on
in
January,
we
communicate
with
the
safe
team
odf
team
right.
So
we
talk
about
a
customer
we'll
be
able
to
remain
100,
open
source
and
Upstream
first
and
in
development.
Okay,
I
want
to
measure
that
you
know
everybody
be
a
piece
here
that
IBM
continued
to
be
the
open
source,
open
source
first
strategies.
So
you
know
IBM
teams,
they're,
you
know,
working
on
the
active.
The
community
leadership
roles
will
continue
to
be
a
role
here
and
you
know
IBM.
A
Now,
instead
of
redhead
IBM
joined
the
governor
sport
as
a
premium
member,
so
everybody
I
hope
that
you
know
you
got
this
and
I
apologize.
I
have
to
repeat
this
one
because
everywhere
I
go
people
talk
about
this.
You
know
the
three
letter
agencies
deal
with
the
open
source
Technologies,
but
I
don't
want
to
give
you
the
false
impression
that
oh
IBM
just
started
looking
at
yourself
from
six
weeks
ago.
A
That's
not
true
I
want
to
by
the
way
you
don't
have
to
read
through
all
this
stuff
I'm,
just
telling
you
that
this
is
the
list
of
the
project
we've
been
working
on
with
Seth
for
the
last.
You
know
three
years
we're
talking
about,
like
you
know,
lip
rdb's.
You
know
native
encryptions
and
then,
like
you,
know,
fabric
that
Jonas
just
presented.
So
there
are
a
lot
of
activities
going
on
right,
I
mean
in
in
safe
communities
that
you
know
IBM
has
been
looking
at
this
for
a
long
time.
A
This
is
not
like
you
know:
hey
December,
1st,
we
just
during
the
New
Year's
party.
We
decided,
you
know
why
not
bring
the
cell
phone
team
over.
We
have
been
thinking
about
this.
This
is
really
a
very
important
disruptions
that
we
are
seeing
in
the
market
and
we
need
the
new
tools.
We
need
the
new
tools.
We
need
the
tool
to
disrupt
the
industry,
the
storage
industry
and
that's
what
we
talk
about
crossing
the
chasm
and
that's
why
we
are
you
know
we,
you
know
the
evaluation.
A
Would
you
say
that
this
is
the
right
move
for
both
company
for
Redhead,
Seth
and
IBM?
So
that's
why
we
welcome
them
in
the
teams.
So
I
know
again.
Please
don't
ask
me
the
two
details.
I
was
sitting
there
listening
to
you
guys,
asking
Yana's
question
I
said:
please
don't
ask
me:
let
level
questions
in
Seth,
I'm
new
guys
in
Seth
I
in
in
the
beginning,
actually,
I
always
spell
stuff.
As
a
capital
c
e
p
h
and
my
team
told
me
that
that
stands
for
Center
For
Education
of
Public
Health.
A
That's
not
Seth!
Stop,
spelling
yourself
like
that.
So
I,
you
know
I
I
change
that
so
all
right.
So
since
I've
been
in
the
storage
business
for
30
years,
let
me
just
share
with
you
right.
So
you
know
this
is
sort
of
brief
history.
I
gotta
lead
to
the
disruptions
right
so
in
early
days.
Well,
since
early
days,
the
1990
around
that
time
right
the
world
of
the
story.
When
you
talk
about
storage,
it's
giant
things.
A
Okay,
I've
been
working
on
the
my
first
assignments
working
on
the
IBM
3390
hard
drive,
it's
like
your
washer
and
size
and
that
things
like
whopping,
four
gigabyte
of
capacities
right
and
that
thing
is
with
tons.
And
then
it's
like
boy.
It's
like
you
seriously!
You
had
you
know
when
we
go
to
the
lab,
we
have
to
sign
some
special.
A
You
know
badge
to
go
into
that.
I
have
to
see
those
things
and
everything
is
a
proprietary
design,
proprietary
Hardware,
proprietary
operating
systems.
We
have
it
all
I
mean
in
my
project
the
first
project
we
have
about
two
mineralized
assembly
codes.
Let
me
tell
you
the
first
code:
I
wrote
is
control
the
magnetic
head
to
be
able
to
skip
from
one
one
sector
to
the
other
sectors,
so
we
can
read
things
more
efficiently,
so
you
know.
A
Obviously
that
is
not
very
you
know
very
agile
if
you
will
so
the
Innovation
is
very,
very
slow.
Now
we
come
to
the
server
base
right
now.
Everybody's
thinking
about
the
server
is
fast
enough.
The
memory
is
much
cheaper.
Let's
just
use
the
generic,
you
know
commodity
server.
If
you
will
that
you
know
it's
a
gen,
you
know
sort
of
commodity
hdds,
so
the
Innovation
you
can
see
the
Innovation
store
to
move
to
the
software.
The
hardware
become
become
I'm,
not
saying
the
hardware
is
not
important.
A
A
Then
you
know
the
industry
talk
about
in
most
of
the
data
centers
at
the
time
have
three
three
stakeholders:
if
you
will
compute
Network
and
storage
and
every
time
you
want
to
do
some
things,
you
have
to
talk
to
like
three
guys:
I,
don't
I
mean
I,
mean
I,
really
feel
bad
I
feel
old.
To
talk
to
you
know
so
many
young
people
here,
I
don't
know.
A
Do
you
remember,
there's
a
day
that
when
you
go
ask
for
capacity
for
your
applications,
you
you
need
to
figure
out
forms
and
it
takes
like
a
best
six
weeks
because
they
have
to
make
sure
their
computer
network
storage
guy
all
agree
on
how
much
capacity.
That
is
why
a
lot
of
people
just
ask
a
lot
of
storage
in
the
beginnings.
You
know
they,
the
VMware
volume.
Some
of
them
are
a
couple
terabyte
because
they
they
related
to
ask
ahead
of
time
and
instead
of
come
back
to
ask
more
and
so
the
world
comes
to.
A
Like
okay,
we
have
something
called
converge
system
basically
prescribed.
Compute,
Network
and
storage.
Is
the
package
right
so
but
later
on,
the
people
figure
out.
Even
that
is
too
complicated.
How
about
we
have
substances
called
hyper
converge.
You
know
we
basically
truly
extract
the
storage
software
storage
into
story
function
into
a
software
running
storage
as
an
almost
like
applications
right
in
hyper-converged
infrastructures.
This
is
really
commodity.
Server,
direct
attached
storage,
then
2015
come
around.
You
know,
maybe,
if
I
don't
take
this
day
delivery.
This
is
just
a
roughly
right.
The
cloud
story
right.
A
A
Now.
The
question
is:
what's
the
next
next
disruption
here,
but
before
we
talk
about
this,
there's
another
things
that
I
don't
know
that
you
guys
realize
this
or
not
that
in
the
storage
world,
nothing
goes
away
right,
everything's
still
here
everything
they
found.
Oh
probably
not
this
one,
but
everything
is
still
here.
You
know
usually
what
I
always
talk
to
customers
that
when
I
go
to
walk
into
customer
Data,
Center
I
was
joking
about
their
like
it's
soji
museums.
They
have
every
different
type
of
storage
on
different
generation.
A
A
This
architecture
divide,
you
know
it
really
really
slowed
down.
The
Innovation
people
spend
so
much
time
to
manage
their
storage,
the
diversity
of
storage,
because
people
believe
that
you
need
fiber
channel
storage
because
it's
faster
okay.
You
need
particular
storage
too.
If
storage
is
like
the
personality
sort
of
dictating
the
you
know,
the
the
we
we
we
put
such
a
hard
personality
on
particular
Technologies
right.
So
we
it's
become
you
know
in
in
in
in
the
in
the
most
data
center.
I
see,
has
a
lot
of
you
know,
story
silos
right.
A
So
you
know.
Sometimes,
when
you
talk
to
storage
guide,
there
is
a
in
general.
There
is
a
Mainframe
guys
and
there's
the
distributor
guys
the
distributor
guide,
there's
the
block
guy
file
guys
and
object
guys
and
the
file
guy
they're,
like
you
know
this.
You
know
basically
there's
a
decision
Tree
on
who
you're
going
to
talk
to
right
and
finally,
they
finally
talk
to
well,
that's
my
area
and
that's
what
you
know
every
time.
But
but
data
is
data.
A
You
know
I,
like
man
you're
talking
about
the
data
is
data
and
in
order
for
people
to
share
data
among
in
their
data
center,
what
do
they
do?
They?
Usually
just
you
know.
Okay,
fine
I
mean
with
all
of
this
okay
organization.
Chart
I
need
data
from
this.
This
one
to
that
one.
Okay,
I
had
to
copy
the
data
over.
So
one
of
the
customer
I
had
you
know,
I
support
in
the
lower
Manhattan.
A
He
told
me
that
in
their
company
they
believe
that
the
same
context
same
content
of
data
probably
get
replicated
about
30
times.
Okay
I
was
shocked.
I
was
like
oh,
my
God.
This
is
great
to
be
in
a
storage
business.
You
know,
how
can
we
help
you
to
replicate
more
well,
that's
not
I
was
just
joking.
That
is
not
that's,
not
good,
because
you
know
the
the
growth
of
the
data
will
immediately
pushing
this
one
and
we
definitely
don't
want
this
kind
of
stupid
practice
to
cause
more
climate
change.
A
So
you
know
so
that's
the
problem
right
everybody
talking
about
this.
You
know
making
different
workloads
and
everybody's
sort
of
building
the
storage
museums.
Okay,
if
it
was
just
and
idea
is,
can
I
have
a
consistent
storage
layers
to
host
all
my
data
can
I
do
this?
Can
I
do
this
to
simplify
this,
because
people
really
look
I
storage,
my
passion,
I
love
to
talk
about
storage,
I'd
like
to
tingle
with
storage,
but
not
everybody
want
to
do
that.
Right,
I
mean
there
are
people
that
just
want
to
do
their
day.
A
Jobs
like
you
know
doing
stuff
that
they
really
want
to
focus
on
The
Innovation.
So
we
want
to
make
this
storage
as
easy
as
possible
to
use.
So
this
is
where
we
want
to
do
that
in
order
to
get
out
this.
You
know
storage,
Museum
business,
that
we
really
need
as
something
that
we
believe
that
is
programmables
and
use
unified
software
software-defined
storage
for
hybrid
Cloud.
Today,
if
you
talk
about
the
edge
storage-
oh
Edge
storage,
you
know
has
to
be
this
particular
things.
A
If
you
talk
about
a
cloud
data
storage,
you
have
to
be
a
particular
things,
and
if
you
talk
about
a
core
data
center
again,
you
talk
about
file,
block
objects
and
go
to
and
then
file
what
kind
of
file
you
have
NFS
team.
You
have
a
sift
teams,
an
object.
You
have
a
different
team,
you
have
a
object
for
the
primary
data
usage,
a
backup
data.
As
an
archive.
You
basically
have
all
these
decision
trees.
A
The
question
is:
can
they
build
from
the
same
substrate
and
we
are
able
to
configure
them
just
like
I'm,
sorry
that
yeah
okay?
This
is
not
I'm,
so
I'm,
so
glad
this
is
not
lunch
time
yet
so
this
is
like
you
know,
can
we
use
the
same
pasta
and
you
just
put
a
different
sauce
and
create
a
different
kind
of
plate,
different
kind
of
meals,
so
that
sort
of
relies
on
this
programmable,
unified
software
defined
storage?
This
software
device
storage
needs
to
be
programmable.
A
You
need
to
be
unified,
you
need
to
be
able
to
need
to
be
API
driven.
So
this
is
where
we
believe
that
you
know
in
the
in
in
going
the
next
disruptions.
This
is
this
I
forgot,
you
know,
I
forgot
one
more
one
more
one
more
things
I
want
to
put
on
it's
programmer,
unified
and
it
needs
to
be
intelligence,
and
then
we're
going
to
talk
about
a
little
bit
more
on.
What
do
you
mean
by
that?
A
So
today
you
know
you
look
at
the
driving
car
right
so
by
the
way,
I
30
years
ago,
I
know
how
to
drive,
stick
shift.
I,
don't
think
I
can
do
that
anymore,
but
I
realized
that
in
Europe,
but
most
of
people
still
drive
stick
shift
right
from
that
experience
to
automatic
to
the
a
lot
of
the
car
assist
to
True
autonomic
driven
technology.
We
do
think
that
in
these
storage
layers,
the
Next
Generation
need
to
have
the
autonomous
capability
start
be
able
to
do
self.
A
You
know
self,
you
know
monitoring,
self-adjusting
and
it
all
based
on
very
important
work
that
Telemetry
works.
That
later
on,
we'll
talk
a
little
bit
more
okay.
So
this
is
our
aspirations.
Okay!
A
No,
we
we
like
to
see
that
Seth
become
the
best
of
reprogrammable
SDS
for
hybrid
cloud
data
managements.
We
want
to
be
a
harmful
of
data.
I
know
this
is
not
going
to
be
happen
overnight.
I
know
and
part
of
that
is
not
so
much
technology,
it's
the
it's,
the
it's
the
business
practice
or
the
just
the
fear
of
unknown
that,
but
I
think
that
we
we
have
this
good.
You
know
technology
that
we
believe
that
can
extend
us
for
to
host
all
data.
A
Be
able
to
extend
from
age
to
core
and
cloud
has
a
unify
interface.
We
can
support
diverse
workloads
and
autonomous
managements.
Okay,
this
is
not
just
automatic
management.
We
need
to
be
able
to
apply
the
new
technology.
How
do
we
make
this
management
a
lot
easier
again?
I
want
to
I
mean,
obviously
in
the
call
here,
I
mean
if
the
fact
you
are
here
you
probably
must
talking
about.
You
must
be
very
familiar
with
elasticity
of
the
self
and
how
robust
it
is.
But
I
think
that
I
want
to
talk
about
the
simplicities.
A
We
need
to
focus
on.
How
do
we
make
this
thing
simple
right?
We
in
order
for
us
to
take
on
this
aspiration,
this
gold,
that
things
need
to
be
able
to
much
easier,
to
manage
much
easier
to
upgrade
much
easier
to
you,
know
the
scales
and
be
able
to
respond
to
the
the
new
workloads
and
new
requirement
autonomously.
A
A
This
is
the
this
is
where
we
you
know
we
like
to
see
that
our
plant
contribution
to
the
community
go
in
the
next
few
years
and
I
think
that
Josh
already
talked
many
of
talk
about
many
of
them,
but
in
high
level,
I
think
break
up
into
three
categories.
One
is
focus
on
the
consumability
automations.
Okay.
This
is
where
how
we're
gonna?
You
know
the
Telemetry
informations
and
then
we're
going
to
apply
some
technology,
so
AI
Ops
to
to
apply
the
AI
technology
to
managing
this
this
clusters.
A
So
we
want
to
drive
the
Innovation
on
acceleration
of
time
to
values.
This
is
not
just
not
not
time
to
Market
time
the
values
right
when
people
start
to
us
getting
safe
in
your
data
center.
How
long
does
it
take
for
them
to
set
up
their
clusters?
How
fast
can
they
apply
the
new
functions,
new
fixes
and
be
able
to
enjoy
the
benefit
again?
I'm,
a
storage,
guys
I
love
to
Tinker
with
things,
but
people
want
to
get
job
done
and
they
want
to
go
home
with
their
families.
So
this
is
very
important.
A
That
I
always
ask
my
teams
that
when
I
apply
a
new
fix,
how
long
does
it
take
for
a
customer
to
enjoy
the
functions
can
start
to
receive
the
benefit
of
this
okay,
so
time
to
value
is
extremely
important
in
this
measure
measurement.
The
second
one
is
a
cloud
and
data
management
right,
so
you
just
unified
hybrid
cloud
and
focus
on
the
future
of
data
management.
The
world
is
evolving.
The
data
is
become
everywhere
and
have
so
much
data.
How
do
you
provide
the
best
data
management
capability?
It's
not.
A
You
know
by
the
way,
I
love,
storage,
but
a
lot
of
people.
Don't
even
care
about
story.
They
just
want
to
get
their
data
right.
So
how
do
we
play
the
role,
the
integration
with
data
management?
This
you
know,
let
me
just
they
I,
don't
know
how
many
people
heard
about
this
phrase
called
lake
house
right.
A
There
is
a
you
know,
1990
again,
there's
something
called
Data
Warehouse
since
then
you
know
you
base
a
warehouse
of
most
of
these
destructive
data
databases,
Oracle
db2,
IBM,
db2
teradata
sets
Etc
and
come
along
with
something
called
the
Big
Data
Big
Data,
mostly
for
the
unstructured
data
right.
So
the
problem
is
people
like
the
predictability
and
efficiency
of
the
structural
data
data
warehouse,
but
I,
really
like
the
elasticities
of
scalability
of
the
unstructured
data
like
big
data
and
that's
called
the
you
know:
data
Lake.
A
Unfortunately,
most
of
the
data
lake
become
data
junkyards,
because
you
know
the
people
just
dump
things
there
right,
so
the
lake
become
polluted.
So
this
concept
called
Data
lighthouse,
it's
kind
of
funny.
It's
it
take
half
worth
on
the
first
phrase
and
six
have
another
set
half
word
from
the
other.
The
other
phrase
is
lake
house.
It
really
the
combination
of
the
best.
It's
basically
asking
for
the
elasticity
of
the
of
the
data
Lake
but
predictability
and
Enterprise
capabilities
of
the
data
warehouse,
and
how
do
we
support
that?
A
How
do
we
participate
in
that
kind
of
big
major
movement?
How
do
we
have
a
enable
our
customers?
There
are
a
lot
of
things
we
can.
We
need
to
do
to
storage
and
to
be
smart
to
support
that.
So
that's
the
second
area
that
we're
working
on
the
third
one
to
resilience
in
cyber
security
again
I
think
that
unless
you
live
in
the
cave,
which
I
don't
think
any
one
of
them,
you
know
for
the
last
every
other
week.
A
Every
month
you
must
heard
about
something
somebody
get
hacked
with
the
cyber
cyber
intrusion
and
things
like
this
so
and
we
need
to
revisit
these
questions
again.
The
industry
start
to
floating
a
new
term
called
cyber
storage.
Basically,
the
storage
can
assist
on
detecting
the
ransomware
intrusions.
Okay.
When
can
you
know
when
people
start
to
updating
data?
Can
we
buy
our
looking
at
data
patterns?
Look
at
the
assets
pattern.
Can
we
start
detecting
that?
There's
a
there's
a
there's,
a
cyber
incidence,
the
potential
cyber
instance,
and
can
we
stop
the
blast
radius
right?
A
Can
we
reduce
the
blast
radius
so,
for
example,
if
we
start
to
realize
that
well,
maybe
somebody
is
trying
to
encrypt
the
data
right
unauthorized
encryptions.
Can
we
stop
that
before
we
encrypt
the
entire
databases?
Can
we
maybe
just
stop
and
just
encrypt
in
a
very
small
fraction
of
data,
and
if
these
things
happen,
what
is
the
role
for
us
to
participate
in
recover
fast,
recover
fasters?
So
how
long
you
think
that
most
a
company
recover
from
cyber
incidence?
A
Three
months:
yeah
yeah,
well,
never
wow,
I,
don't
know,
I
mean
I,
think
many
of
them
recover
okay,
but
it
takes
a
long
time.
In
fact,
most
of
you
know
we're
talking
about
in
the
past.
We
asked
people
like
that
and
really
feel
like
I'm
showing
my
age
here.
So
you
know
long
long
time
ago
the
storage
industry
doesn't
even
do
the
disaster
recover
or
hadr.
A
The
turning
point
is
the
World
Trade
Centers,
you
know
bombing.
So
at
that
time
overnight,
everybody
in
everybody
you
know
I
was
at
the
time
doing
the
my
Mainframe
storage,
every
customer
asking
for
Disaster
Recovery
plan
and
that
basically
they
can
they
can.
They
can
respond
to
the
data
center
failure
stuff,
but
now
the
next
one
is
not
so
much
data
center
fail.
It's
The
Logical
fail,
it's
the
it's
the
Cyber
intrusions
and
how
do
we
protect
our
customers
most
important
assets
that
we
are
able
to
detect
it
earlier
and
we
can
recover
fast
right.
A
You
know
somebody
the
gentleman's
in
the
back
says
three
months:
three
months
recover.
Can
you
imagine
some
company
shut
down
for
three
months
right?
That's
that's
Unthinkable
scenarios
that
many
companies
will
be
just
out
of
business
if
they
shut
down
three
months.
So
this
the
third
area
we're
talking
about
the
the
Dr
Solutions
and
then
the
you
know,
such
as
the
rdb
live
migrations,
automated
Dr
solution
and
encryption
support,
and
things
like
that.
So
well,
okay,
I'm!
Just
going
to
give
you
an
example
what
I'm
talking
about
here?
A
So
this
is,
you
know
when
we
every
time
the
good
thing
about
Steph
is
every
time
I'm
asking
the
team
that
do
we
have
this
one?
The
answer
is
yes,
we
do
right
and
okay,
that's
great.
So,
for
example,
that
do
we
have
the
Dr
solution.
Yes,
we
do
okay,
but
the
thing
is
that
we
have
the
data
data
replication
from
point
A
to
point
B.
A
This
is
the
project
we
are
working
in
the
open
source
area
that
you
know
in
the
you
know
the
there's,
a
PV
and
PVC
right
in
the
in
the
in
the
event
of
the
failovers
that
you
need.
We
need
to
be
able
to
make
sure
that
all
the
metadata
are
covered,
so
you
can
resume
the
application
on
the
the
talker
site
and
this
automation.
This
is
what
we're
talking
about
this
automation.
A
Sure
if
customers
know
ahead
of
time,
they
can
do
it
themselves
right,
but
just
just
it's
just
saying
that
yeah
we
have
the
ER
solution.
We
can
copy
the
data,
but
customer
without
a
metadata.
They
cannot
restart
the
application
anyway.
So
this
is
what
I'm
talking
about
be
able
to
make
this
solution
as
a
consumables.
Okay.
A
This
is
where
I'm
talking
about
the
time
to
Value
things,
so
so
I
am
firmly
believe
that
you
know
that
the
storage
industry
is
about
to
have
a
new
disruptions
with
all
the
big
data,
all
the
large
amount
of
data,
there's
trinomial
data
and
all
these
AIS
and
all
this
cyber
Securities-
and
you
know
people
need
to
have
a
real,
a
new
storage
substrate.
If
you
will
to
support
those
Innovations
and
I'm
here
to
Champion
this
idea,
I'm
hoping
that
the
the
community
will
help
us
to
get
there.
Okay,
that's
all
I
have
questions.
A
Do
not
ask
that
kind
of
level
detail
questions
I
already
told
you
I
I,
just
I
even
spell
Steph
incorrectly,
when
will
IBM
natively
support
cell
phone
AIX
Seth
on
AIX?
Okay,
that
is
a
tough
questions.
Okay
who
wants
it?
Okay,
one
person
wants
it.
Okay,
yeah
I'm,
surprised,
you
didn't
ask
me:
when's
the
IBM
gonna
support
native
on
Z,
the
Z
OS,
so
I
have
to
get
back
to
you
on
that.
So
we
have
a
product
management
team.
Uday,
raise
your
hand
here.
Okay,
please
ask
him
the
hard
question
to
him.
A
He
controlled
the
the
program,
the
black
items,
any
other
questions.
Okay,
I'll,
be
around
all
day.
If
there's
any
questions
that
I
can
help
with.
Please
let
me
know
and
I
hope
that
again,
this
is
presentation
is
not
about
IBM's,
but
it
is
about
IBM's
visions
and
towards
the
chef
and
I
hope
that
I
collaborate
with
the
the
community
that,
let's
make
this
happen
all
right.
Thank
you.