►
From YouTube: Centaurus Monthly TSC Meeting 02/22/2022
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
C
C
Yeah
yeah
money:
could
you
have
a
mega
mega
paw
in
the
zoom,
so
that
that
I
think
we
have
three
dsm
members
at
the
meeting
now,
so
we
can
vote
and
then
yeah
after
that,
then
we
can
get.
We
may
get
the
result
from
other
numbers.
C
Because
so
far
we
only
get
one
nomination
from
featuring
for
deepak.
So
deepak
is
the
only
candidate.
D
What
about
shawnee
you're,
not
gonna,
you're,
not
gonna!
You
don't
have
any
interest
to
continue.
C
It's
already
been
one
year
and
honestly,
it's
a
little
difficult
for
me
to
to
operate
effectively,
because
now
I
don't
have,
I
don't
have
enough
time
and
also
I
don't
have
access
to
another
kind
of
the
project
planning
stuff.
So
it's
a
little
difficult
for
me
to
continue
operator
effectively.
I
and
I
think
that
deepak
is
a
perfect
candidate
for
the
ksc
chair
because
he
spent
a
lot
of
time
and
contributed
not
yeah.
E
E
C
A
A
So
I
will
be
great
to
have
you
you
know
as
even
though
I
know
you,
you
have
a
lot
of
commitment
in
your
company,
but
it'll
be
great
to
have
you
and
you
as
part
of
our
meeting
as
a
psc
member,
so
yeah
I
mean
I'm
excited
to.
Hopefully
we
do
a
lot
of
work
actually,
so
this
kind
of
change
of
direction
so
the
way
centaurus
started
out
and
now
the
direction
we
are
going
in
it's
a
lot
of
you
know
challenging.
G
D
F
B
A
G
C
Okay,
I
think
now
we
probably
can
move
to
the
second
agenda
item:
hey
I'm
still
there.
Yes,
yes,
so
I
think
we
can
get
started.
Data
management.
H
Okay,
yeah
sure
thanks
yeah
thanks
for
this
slot
to
speak
I'll
just
share
my
screen.
H
H
And
some
of
the
industry
trend
and
then
what
we
are
thinking
at
soda
foundation,
a
glimpse
of
it
I'll
go
through
a
little
faster,
because
I
have
around
30
slides
some
I'll
skip
and
if
you
have
any
questions
or
online
or
offline,
you
can
discuss.
H
Okay,
so
this
is
self
explanatory,
because
everything
is
connected
anywhere
edge
on
premise
and
cloud
and
everything
things
are
connected,
so
things
are
connected
and
the
technologies
are
data
driven,
so
data
management
is
very
critical,
so
basically
what
we
need
as
a
user,
we
want
something
like
unified
global
because
things
are
distributed,
but
still
we
need
a
global
view
of
things.
What
is
happening,
whether
the
deployment
is
in
edge
kind
of
scenario
or
in
on-premise
it's
nothing
but
on-premise
or
cloud.
H
You
want
full
control
and
management
on
your
data
and,
as
you
know,
in
terms
of
storage
file,
block
and
different
object
or
file
formats
or
the
storage
formats
is
a
legacy.
So
the
dispute
or
the
debate
is
still
on
that.
Whether
file
is
better
block
is
better
object.
Is
better
things
like
that,
though?
H
In
the
cloud
native
scenario,
object
is
predominantly
taking
over
our
object
and
file
combinations,
but
still
we
need
to
live
with
this
situation
and
then
there
are
influences
and
impact
based
on
different
platforms,
say,
for
example,
open
stack,
kind
of
platforms
or
virtualization
based
platforms
or
containerized
place
platforms
or
mixed
kind
of
things
also
affect
how
you
manage
your
data
and
different
applications
and
use
cases.
H
So
there
are
a
lot
of
surveys
recently
happened.
Even
soda
did
a
data
and
storage
survey
last
year,
and
also
there
are
some
industry
reports.
I
pin
the
links
in
the
notes
of
the
slides,
so
predominantly
what
you
see
is
that,
though
a
major
chunk
of
on-premise
or
enterprise
kind
of
situation
is
still
there.
H
H
Containers
that
is
predominantly
in
our
report,
our
survey
also
this
is
this-
has
come
out
very
obviously,
and
similarly,
the
industry
report
also
says
that
application
deployments
are
moving
towards
container
significantly
and
out
of
this,
the
movement
to
our
momentum
towards
container
deployments.
Kubernetes
is
one
of
the
top
candidates
as
an
application
orchestration
platform,
and
today
I
mean
most
of
the
technologies
when
it
comes,
and
we
we
hype
about
a
particular
technology.
H
Most
of
the
cases
it
will
be
notional
in
the
initial
phase,
but
today,
container
storage
deployment
is
not
really
notional,
it
is
real
use
case.
The
deployments
are
happening.
Production
kind
of
production.
Great
deployments
are
happening
based
on
container,
so
things
are
moving.
I
think
earlier
we
started
hearing
about
storage
as
a
service
when
more
and
more
cloud
vendors
came
or
on-premise
kind
of
private
cloud
and
public
cloud.
H
So
the
summary
is
that
its
container
centric
storage
solutions
are
more
prominent
and
hybrid
cloud.
Data
management,
especially
for
disaster
recovery
or
data
protection
kind
of
scenario,
is
becoming
hot
and
well.
Demand
is
very
high
on
that
and,
of
course,
momentum
towards
container
storage,
because
the
container
application
deployments
are
high,
so
hence
the
container
storage.
H
So
most
of
the
solutions.
If
you
see
there
are
logical
solutions
to
provide
end-to-end
management,
so
I
have
given
some
examples
and
we
will
see
some
more
details
on
the
further
slides.
H
H
And
many
other
products
coming
in
this
way
that
end-to-end
data
management
solutions
and
earlier
most
of
the
storage
vendors
were
confined
to
the
storage
boxes,
but
not
today,
so
they
are
all
coming
out
of
storage
box.
H
At
the
same
time,
they
are
increasing
the
capability
of
the
storage
boxes,
but
they
come
up
come
out
with
solutions
on
top
of
their
storage
boxes
and
container
deployment,
ready
cloud
native
ready,
those
kind
of
stuff
and
multi-cloud
application
aware
this
is
another
key
area
where
most
of
the
people
are
working
application,
aware:
data
management
because
there
are
different
use
cases,
big
data
or
ai,
on
different
use
cases.
The
data
management
needs
to
be
more
tuned
to
the
application
or
the
workload
and
cross
vendor
product
ecosystem.
This
is
one
of
the
key
areas.
H
Most
of
the
vendors
I
mean,
I
think,
in
the
storage
history
itself.
I
think
this
momentum
is
for
the
first
time
that
they
are
coming
out
to
build
a
better
ecosystem
with
partners
and
open
source
projects.
H
This
is
not
about
storage,
but
on
the
container
landscape.
This
is
this.
I
have
taken
from
the
cncf
project
landscape.
So
if
you
see
the
container
landscape
is
growing
exponentially.
So
you
time
to
time
the
number
of
projects
number
of
new
solutions
or
use
cases
coming
is
quite
high
and
when
it
comes
to
container
data
management.
Okay,
these
are
some
terms
I'll
just
skip,
because
there
are
different
terms.
H
We
confuse
with
cloud
native
storage,
container
storage
and
the
container
storage
interface
csi,
so
so,
basically,
cloud
native
storage
is
kind
of
software
defined
storage
that
apa,
driven
mainly
on
api
driven
container
storage,
is
contained
as
centric
storages
and
container
storages
type
of
cloud
native
storage,
but
not
the
other
way
and
csi
is
just
a
specification.
So
this
is
a
one
of
the
I
mean
term.
We
use
understand
it
in
differently
in
many
cases.
So
csa
is
just
a
specification
for
storage
interconnect.
H
Connection
with
the
content
deployments
across
not
only
for
kubernetes,
but
whichever
is
compliant
message
or
cloud
foundry.
All
these
are
combined
with
csi,
where
the
storage
connection
interface
specification
so
based
on
the
specifications,
vendors
will
prepare
the
csi
plugins,
which
can,
which
is
the
real,
the
driver
which
can
be
deployed
to
connect
their
vendor
storage.
H
So
this
is
a
high
level
of
view
of
cloud
native
storage,
like
I
mean
typical,
any
other
application,
but
the
lower
half
is
slightly
different,
so
business
application,
then
application
platforms
and
then
the
container
orchestration.
That
is
where
kubernetes
and
all
other
projects
will
come
into
play,
then
on
container
and
in
container
itself
in
the
runtime
storage
and
the
the
processor
runtime
and
the
network
will
come
into
play.
So
this
is
where
we
are
interested
in
the
container
storage
and
the
storage
boxes
connected
through
csi,
especially
in
the
container
part.
H
H
For
your
docker
features
and
things
like
that,
which
is
portable
and,
on
the
right
hand,
side,
security,
compliance
and
key
manager.
These
are
some
key
areas,
even
in
the
container
data
management
or
container
deployment
area
that
security
compliance
is
one
of
the
areas
where
it
needs
more
maturity.
H
Okay
now
comes
to
cloud
native
storages.
These
are
some
of
the
projects
which
is
in
cncf
landscape
about
cloud
native
storages.
I
just
circled
some
of
them
just
to
show
that
some
of
these
projects
are
part
of
soda
project
as
well.
Zenko,
open
ebs
and
linstor,
incidentally,
soda
is
also
part
of
this
storage
landscape.
H
Key
platforms
from
different
vendors,
so
I
think
you
don't
need
to
read.
This
is
just
to
depict
that
most
of
the
vendors
like
netapp,
pure
storage,
hp,
dell,
emc
hitachi,
vmware
rancher.
I
mean
something
like
rustic:
it
is
an
open
source
project,
even
my
data,
so
most
of
these
vendors,
so
they
are
focusing
on
various
use
cases
predominantly
on
data
protection.
H
So
it
can
be
backup,
restore
replication,
failover
and
those
kind
of
things
so
most
of
the
solutions
they
are
coming
up
with
that
and
also
most
of
them,
are
something
like
a
software
as
a
service
and
logically
connecting
like
I
mentioned
earlier.
They
build
components
from
third
party
or
their
own
components,
to
build
this
kind
of
a
solution
and
focusing
on
even
the
container
a
deployment
part
and
even
in
the
container
deployment.
Their
first
focus
is
on
the
provisioning
and
data
protection.
A
Okay,
just
go
back
so
the
question:
where
did
you
get
this
list
from?
Because
the
reason
I'm
asking
that
question
is
because
this
whole
data
protection
and
disaster
as
a
service
and
all
that
the
leader
in
the
space
is,
you
know,
companies
like
rubrik
and
cohesity
and
all
those
kind
of
those
guys
they
don't
show
up
in
the
list,
though,
let
me
all
right.
H
A
Okay,
you
know,
I
think,
the
whole
database.
The
management
term
itself
is
confusing,
though,
because
so
from
your
perspective,
data
management
is,
if
you're,
a
storage
company
your
data
management,
so
that
falls
under
data
management,
but
there's
another.
So
it
looks
like
yeah
go
ahead.
Yeah,
you
wanted
to
say
something.
Sorry.
A
I
think
one
of
the
key
thing
I
wanted
to
kind
of
highlight
that
there
was
another
I'll
post
that
article,
a
very
nice
article,
published
recently
on
new
stack.
This
whole
data
management
can
is
evolving
behind
where
beyond
kind
of
silos,
and
all
that
this
whole
theorem,
you
know
especially
companies
like
rubric
and
cohesity,
and
all
that
people
and
all
the
companies
are
outsourcing
their
data,
so
they
are
essentially,
you
know,
like
a
holistic
picture
of.
You
know
the
data
management,
so
the
is
that
something
you
have
you.
A
H
Are
they
are
no
they're?
They
are
not.
They
are
not.
They
are
not
so
this
okay,
so
data
management-
probably
it
is
very
generic-
and
I
know
probably
at
a
different
level.
They
see
it
different,
but
conceptually
conceptually,
if
you
see,
for
example,
rubric
or
cohesity
or
anything
you
talk
on
on
top,
even
the
big
data
processing
side,
there
are
applications
and
platforms
on
top
wherein
they
also
do
the
same
data
management
more
from
the
application
perspective,
so
even
the
metadata
management
and
also
the
the
real
data
management.
H
In
the
I
mean
the
data
plan
and
the
control
plane
part
here,
we
are
more
on
one
level
down
that
because
most
of
these
features
traditionally,
where
on
application
level
utilizing
some
of
the
features
from
the
lower
I
mean,
maybe
the
storage
or
the
infrastructure
part,
but
right
now
what
we
are
seeing
is
that
even
the
storage
vendor
coming
up
and
building
solutions
to
make
their
solution,
even
the
data
protection
or.
H
H
A
They
managed
it
so
cloud
public
cloud,
agnostic.
B
H
Okay,
so
these
are
some
of
the
I
mean
views
about
so
like
I
mentioned
data
pasta,
if
you
see
cohesity
also,
it
is
something
like
it
will
provide
application,
aware
public
cloud
data
management
across
hybrid
cloud
kind
of
scenarios.
So
here,
if
you
see
the
netapp
astra,
is
a
solution
from
the
storage
vendor,
wherein
they
try
to
use
their
lower
layer
to
connect
their
boxes
through
the
container
interfaces.
H
Especially
the
plugin
is
called
trident
and
then
bring
your
own
kubernetes.
You
can
have
optimized
kubernetes,
whether
it
can
be
vanilla,
communities
or
public
cloud.
Maybe
it
is
from
aws
or
gcp
or
enterprise,
private
cloud
or
hybrid
cloud,
and
then
you
put
your
application
on
top
of
it,
so
they
are
also
trying
to
have
even
the
astra
service
actually
provides
some
kind
of
a
management
layer
on
top.
H
So
similarly,
even
pso
is
trying
hp.
I
mean
some
of
the
project
products
which
I
mentioned
in
the
earlier
slide.
They
all
have
something
similar
and
the
direction
is
in
this
line
that
they
want
to
provide
end-to-end
application,
aware
data
management
and
lifecycle
management
data,
lifecycle
management.
H
Okay,
so,
and
at
soda
what
we
are
thinking,
I
think
some
one
one
or
two
points
I
think
even
deepak
mentioned
when
you
ask
the
question.
H
So
basically,
it's
kind
of
a
xylose
like
you
have
different
platforms
with
different
storages
and
if
you
buy
a
storage
storage
will
have
its
own
management
plane
through
which
you
need
to
connect
the
arrays
to
get
the
information
or
manage
your
say.
Data
protection
will
be
supported
in
storage
outside
storage
things
like
that,
and
there
are
different
kinds
of
first,
the
object
or
file
or
block,
and
some
of
the
deployments
will
be
in
vm.
H
So
on
top
of
that,
you
have
public
public
cloud,
private
cloud
edge,
distributed
data
centers,
and
things
like
that.
So
you
will.
You
already
have
such
kind
of
situations
where
the
capacity
performance,
the
overall
data
management
is
becoming
hard
and
most
of
the
cases
you
may
have
to
get
locked
in
with
some
of
the
vendor
solutions,
wherein
you
will
have
limited
third-party
storage
support
and
their
solutions
will
be
optimized
for
their
storage
and
etc.
H
So
this
is
in
a
changing
mode
that
more
and
more
third
party
support
is
coming
more
and
more
unified
solutions
are
coming.
That
is
where,
in
soda,
what
we
envision
is
that
a
unified
data
framework
which
support
edge
core
and
cloud
and
also
support
heterogeneous
storages
behind
whether
it
is
cloud
or
optimized.
A
H
H
We
are
moving
in
that
direction
to
step
by
step
that
how
we
can
get
there
with
an
open
solution.
Today.
Today
we
have
a
kind
of
good
solutions
on
on-premise
and
cloud
edge.
We
have
just
started,
and
this
year
we
are
trying
to
focus
on
certain
very
specific
things,
because,
as
a
foundation,
it's
just
started
in
2020,
so
whatever
we
can
focus
step
by
step
incrementally,
along
with
the
partners,
we
wanted
to
our
progress
and
build
some
solutions
which
can
be
really
deployed
in
the
production
scenario.
H
That's
where
we
are
moving.
I
will
just
touch
upon
that
in
the
subsequent
slides.
So
overall,
this
is
a
view
of
soda
open
data
framework.
We
call
it
as
open
data
framework,
so
we
want
to
provide
end-to-end
whether
it
is
edge
core
or
cloud
manage
all
the
data
in
between
and
any
platform
any
application
and
any
storage
and
our
key
propositions
are
open
source
and
connect
data
silos.
I
mean
we
want
to
eliminate
the
data
cellulose
like
the
work
also
mentioned
this,
and
it
is
extensible.
H
It
is
like,
like
the
solutions
or
the
modules
can
be
plugged
in,
and
we
would
like
to
move
towards
some
kind
of
api
standardization
and
build
an
ecosystem
with
hardware,
software
and
solution
and
service
vendors
and
also
in
future.
We
want
to
have
some
kind
of
certification
based
on
this
standardized
ecosystem,
so
this
is
overall
stack.
You
can
see
on
the
right
hand
side
this
is
we
just
show
one
of
the
use
cases,
a
typical
use
cases
like
a
connected
car
platform
wherein
you
have
edge,
you
have
on-premise,
you
have
cloud.
H
Okay,
so
going
forward,
we
are
trying
to
consolidate
some
of
our
projects
and
landscape
in
soda
foundation,
so
these
are
some
of
the
key
components
of
our
landscape
today,
so
you
have
an
sds
controller.
This
is
is
like
it
comprises
of
three
projects:
api
controller
and
talk
dock,
where
the
storage
can
be
connected.
H
This
is
primarily
on
premise,
so,
irrespective
of
your
storages
file
and
block,
this
can
provide
you
an
unified
api
to
connect
with
the
different
storage
and
you
can
use
your
provisioning
data
management,
like
snapshot,
file,
share
and
etc,
and
now
the
container
management
is
the
new
project.
The
green
things
are
the
incubator
new
project,
which
we
are
planning
in
2020
to
one
for
container
data
management.
Primarily,
we
will
be
focusing
on
backup
and
restore
to
start
with
and
then
observability.
H
Why?
Observability,
because
we
already
have
a
project
called
delphin,
which
is
for
storage
monitoring,
which
can
support
heterogeneous
storage,
monitoring
across
different
storage
vendors
and
provide
a
single
dashboard
to
provide
all
the
infrastructure
information
like
resources,
performance
alerts.
H
So
this
is
one
key
project
which
people
are
also
more
interested,
because
this
is
one
of
the
tough
ask
tasks
and
also
this
is
not
in
the
critical
path,
even
for
their
current
deployment.
It
is
not
in
the
critical
path
on
the
monitoring
side.
H
They
can
actually
deploy
and
get
all
the
insights
with
a
common
dashboard.
So
this
delphin
we
have
already
ported
to.
We
have
experimented
with
kubernetes.
It
is
on
enterprise
side
currently,
but
we
have
tested
in
kubernetes
and
it's
working
fine.
So
we
wanted
to
migrate
this
as
a
native
observer,
observability
storage,
observability
framework
for
a
container
environment,
which
is
a
key
gap
today,
because
in
csi
it
is
only
providing
the
volume
level
matrix.
H
So
the
storage
specific
information
is
missing.
Still
people
are
using
the
vendor
managed
platforms,
so
this
could
be
a
good
replacement
and
then
the
multi-cloud
data
management.
We
have
a
project
called
strato,
so
this
is
a
multi-cloud
wherein
we
support,
I
think,
six
or
seven
cloud
vendors.
H
So
the
user
can
have
a
transparent
api
interface
where
you
can
have
the
the
object,
provisioning
data,
migration
for
on-premise
or
cloud-
and
we
are
also
considering
we
are
working
with.
H
One
of
the
I
mean
top
partners
of
soda
where
they
are
very
much
interested
in
a
data
lake
kind
of
project.
They
have
very
specific
requirement
on
object
management
across
their
data
center
and
the
multiple
cloud
vendors.
So
this
is
one
of
the
other
projects
we
are
actively
discussing
to
be
launched
soon.
H
So
container
data
management
and
data
lake
will
be
our
new
additions
to
this
year
and
focus,
and
you
see
a
lot
of-
I
mean
blue
boxes
here.
These
are
some
of
the
echo
project
soda
echo
project,
wherein
the
external
project
join
as
a
partnering
project
with
soda
landscape,
wherein
we
try
to
work
for
collaborative
solutions
with
those
projects
like
lean
store,
is
a
block
storage
solution.
H
Acura
edge
is
an
edge
kind
of
platform,
it's
originated
by
huawei
and
it
is
a
cncf
project
and
zenko
is
an
object.
Data
store
and
egg
is
a
massive
object.
Storage
is
a
optimized,
ceph
kind
of
a
solution.
A
cortex
is
from
seagate
and
that
was
from
intel.
It
is
an
nvm
object.
Storage
and
open
ebs
is
a
container
attached
storage
from
my
data,
and
now
it
is
data
core.
H
C
A
A
G
A
Their
seven
billion
dollar
valuation-
these
guys,
so
industry,
has
realized
that
you
know
this.
That's
not
the
way
it's
gonna
go
basically
with
data,
so
data
management
is
gonna,
be
cloud
agnostic
companies
like
direct.
A
H
Yeah,
so
one
one
thing
is
that
we
are
going
step
by
step.
So
one
thing
is
that
we
first
started
with
something
like
an
end
users,
because
we
wanted
to
get
some
of
the
real
use
cases
so
kp
and
vodafone,
or
that
kind
of
an
end
users,
toyota,
even
huawei
huawei,
is
kind
of
a
vendor,
plus
storage
or
storage
vendor.
H
Yes,
we
did
not
get
big
success
except
seagate
and
some
companies
because,
as
you
know,
storage
organizations
are
generally
not
very
open
for
this
kind
of
solution.
So
we
hope
that
we
can
get
some
of
them
on
board
and
then
we
also
work
with
integrators
and
probably
these
kind
of
projects
also,
we
will
reach
out
to
to
to
have
them
part
of
our
ecosystem.
H
So
this
is
again
whatever
I
explained
so
I'll
just
skip
this
part,
okay,
so
on
the
container
data
management.
I'll,
just
brief,
some
ideas
on
that.
So
what
we
plan
to
do
is
that
we
will
be
focusing
more
on
kubernetes
kind
of
solution
wherein
we
have
two
major
areas
we
are
focusing.
One
is
about
the
csi
part,
the
other
one
is
on
the
data
management
extensions.
H
Interface,
which
is
provided
in
the
container
environment,
to
connect
to
storages-
and
if
you
see
today,
every
vendor
will
have
their
own
csi
plugin
and
the
csa
plugin
can
support
a
series
of
series
of
models
from
the
same
vendor
and
also
there
are
some
open
source
solutions
which
support
multiple
vendors.
H
Our
idea
is
that
it
is
a
heterogeneous
csi
plugin,
wherein
you
can
connect
your
vanilla,
csi
plugin
to
this,
so
that
you
don't
need
to
have
manage
multiple
plugins
that
will
be
obstructed
by
the
soda
csa
plugin
to
to
make
it
easy
to
manage
port
it,
because
we
want
to
support
a
heterogeneous
and
also
whatever
the
metadata
management
of
across
the
storage.
Vendors
can
be
very
easy.
H
H
Some
discussions
in
the
community
and
at
the
same
time
hobby
also
was
interested
because
we
wanted
some
user
some
user.
So
we
told
that
yeah
we
are
very
much
interested,
so
we
are
already
integrated
with
the
product
on
this
project
and
similarly
on
the
data
protection.
Also,
we
are
getting
some
interest
from
the
community
like
ntt,
softbank
and
different
partners
of
soda,
wherein
they
are
interested
in
the
data
protection
aspect
of
it,
whether
in
the
container
environment.
H
So
we
thought
this
is
predominantly
going
up.
So
this
is
one
good
area
to
start
with,
and
basically
we
are
also
in
the
process
of
getting
the
real
users
and
the
deployment,
because
that
is
very
important
for
the
success
of
the
project
and
earlier
we
used
to
focus
on
multiple
projects
and
across
the
stack.
H
So
we
are
trying
to
reduce
that
kind
of
a
wider
focus
and
go
into
a
specific
solution,
kind
of
focus
even
with
the
echo
project
solutions
so
wherein
we
can
give
a
specific
solution
for
a
particular
challenge,
so,
for
example,
data
backup
and
restore,
or
only
just
the
storage
monitoring
across
the
storage.
Vendors
unify
your
heterogeneous
storages
and
things
like
that.
So
that
is
where
we
are
heading
to
in
the
container
space.
H
This
is
the
csi
part
which
I
have
already
explained,
and
these
are
some
of
the
current
features
at
a
poc
level.
I
won't
say
that
it
is
a
ga
or
that
kind
of
a
quality,
but
there
are
some
initial
pre-releases
on
these
features,
which
is
available
in
our
solar
releases,
which
can
be
tried
out.
It
can
be
trialled
in
kubernetes
environment.
H
Even
we
have
the
delve
in
the
monitoring
which
is,
I
have
not
mentioned.
I
think
the
heterogeneous
storage
monitoring
can
work
with
kubernetes
as
it
is.
The
delphin
can
be
deployed
in
kubernetes
in
that
way.
In
the
latest
release-
and
next,
like
I
mentioned,
we
will
be
focusing
on
data
protection,
observability
and
the
csa
enhancement.
Even
the
csa
community
is
working
towards
more
features
on
the
data
protection,
because
csi
doesn't
support
any
of
the
key
data
protection
kind
of
features.
H
Across
the
storage,
so
they're
also
working
to
improve
the
csa
spec,
so
we
will
be
working
with
them
as
well.
So
this
is
about
some
demo
link
and
another
one
is
that
the
soda
data
lake
kind
of
project?
So
this
is
a
very
high
level
kind
of
a
view
about
a
typical
data
like
kind
of
solution,
mainly
on
data
metadata
management
and
kind
of
managing
the
data.
H
So
the
current
focus
is
to
start
with,
because
this
is
a
big
scope,
very
specific
scope.
Is
that
how
we
can
manage
object
data
across
the
cloud
with
the
service
provider
data
center,
along
with
the
client,
the
customer
data
center?
H
So
there
are
some
specific
use
cases
we
are
getting
from
our
end
users,
and
we
are
trying
to
create
that
in
2022
to
support
that
the
workload
the
customer
data
center
application
workload
can
seamlessly
manage
the
object,
data
management
across
multiple
cloud
with
the
access
management
provided
by
the
service
provider
service
provider,
data
centers.
So
this
is
one
of
the
I
mean
deployment
use
case
today.
H
H
H
We
don't
we,
don't
we
don't
we
don't
catch
data,
we
okay
right
now.
Okay,
so
till
now,
till
now,
most
of
the
cases
our
data
management
limited
to
control
plane,
but
going
forward
even
right
now
the
object
management
which
we
are
talking
about
is
predominantly
on
the
control
plane,
but
the
kind
of
requirements.
What
we
are
getting,
especially
on
the
data
lake
side.
G
H
Request
is
that
you
also
should
help
in
the
data
plane.
So
probably
we
will
move
and
support
in
the
data
plane
area
also
because
data
lake
is
more
on
that
direction.
So.
E
Do
you
because
it
do
you
manage
the
soda
to
sort
of
manage
the
relationship
between
the
data
that
from
different
cloud
or
for
different
clusters,
so
in
the
okay
for
example,
in
this
case,
you
may
have
the
data
actually
related
in
one
cluster
to
another
or
to
your
ws.
Do
you
manage
that
relationship
so
that.
H
Okay,
so
what
we
are
right
now
doing
is
that
say,
for
example,
you
have
aws
and
gcp.
So
suppose
I
am
a
user.
I
have
some
data
in
aws.
H
You
want
to
move
the
data
to
say
gcp
some
data
to
gcp,
so
you
want
to
do
the
data
data
migration
from
one
cloud
to
another
which
we
facilitate.
We
don't
get
into
the
data,
but
we
can
migrate
the
data
we
can
do
life
cycle
management
say,
for
example,
in
aws
itself.
I
mean
not
about
multiple
cloud,
even
one
cloud
itself.
If
you
want
to
do
life
cycle
management
from
the
standard,
it
has
to
go
to
glacier
or
maybe
archival
kind
of
thing.
H
So
so
this
this
level
at
the
moment
we
manage
at
this
level
but,
as
you
know,
s3,
when
you
start
working
with
the
object,
s3
kind
of
case,
you
are
not
only
in
control
plane.
You.
G
H
Say
that
you're
purely
on
control
plane
so
going
forward,
probably
the
kind
of
requirements
what
we
get
we
we
may
get
into
the
data
plane
kind
of
processing
as
well.
But
at
the
moment
I
would
say
that
predominantly
we
work
in
control
plane.
We
don't
get
into
the
data,
but
we
are
outside
and
manage
the
data.
A
H
H
That's
that's
correct!
That's
correct,
okay!
So
because
right
now
right
now
we
support
file,
block
and
object,
even
in
the
multi
cloud.
Suppose,
if
aws
and
gcp
support
file
block
file
storage
and
if
it
is
same
format,
we
can
do
especially
in
block
case.
The
problem
is
made
predominantly
on
the
block
case,
because
object
is
more
or
less
s3
compatible
and
the
file
is
also
kind
of
manageable.
But
in
the
case
of
block
is
where
you
will
face
problems,
because
the
formats
are
different.
H
Okay,
so
right
now,
right
now
we
have
two
levels
of
security.
One
is
that,
because
we
sit
I'm
not
talking
about
the
data
lake,
I'm
talking
about
the
multi-cloud,
we
actually
use
the
back
end
security
specific
to
the
cloud.
That's
one
thing
and
our
apis
are
authentication
based
right
now
we
are
using
keystone
based
authentication
for
the
api
level
and
the
cloud
level.
Security
is
purely
based
on
the
cloud.
So
we
support
some
kind
of
acl
sse
kind
of
support.
C
I
mean
for
the
because
you
move
data
but,
however,
the
the
metadata
part
like
these
tags
and
the
sal
associated
with
these
data
blocks,
where
they
also
be
kind
of
replicated
to
the
new
data
data
storage.
H
G
A
H
Provisioning
data
management,
everything
so
you
have
a
single
endpoint.
So
if
you
create
something
outside,
you
don't
have
an
access.
A
S3
and
I
go
through
your
api
behind
the
scene-
you
you
know
use
as
three
api
to
deploy
that.
G
H
Exactly
exactly
basically
right
now,
right
now,
the
flow
is
something
like
you
start,
creating
the
back
end.
What
back
end
you
want,
because
we
support
many
back
ends,
so
it
can
be
aws
as
or
gcp,
and
things
like
that.
This
is
one
way
you
register
the
back
end
when
you
register
the
back
end,
you
give
the
access,
keys
and
etc,
etc.
H
Then,
once
you
register
the
back
end,
you
can
start
from
provisioning
the
bucket
creation
data,
upload,
download
all
the
cred
operations
and
data
lifecycle
management.
Now
we
are
also
adding
some
more
intelligence
that
abstracting
the
back
end.
Also
something
like
because
this
we
got
one
of
the
partners
we
they
had
a
requirement
saying
that
they
want
storage
service
plans.
H
In
the
sense
they
want
to
say
to
the
customer
saying
that
you
are
in
gold
plan
or
silver
plan,
but
gold
means
in
behind
gold
will
support,
say
aws
and
azure
only
in
particular
region
and
then
silver
means
it
will
only
support,
say,
ibm
or
some
other
cloud.
Now
the
user
will
see
that
okay,
my
data
is
in
gold
or
silver,
something
like
that.
So
these
kind
of
features
intelligence.
A
H
That
will
be
fantastic
because
we
always
explore
the
possibilities
of
creating
some
combined
solution.
That's
where
we.
A
A
A
Could
be
a
very
good
thing
to
integrate
that,
so
that's
something
we
should
talk
more
actually
in
the
next
meeting.
Also,
we
should
brainstorm.
This
is
very
good.
H
Right
right
sure
sure
we'll
do
that.
Okay,
I
think
this
is
my
last
two
slides,
okay,
so
soda
edge.
We
have
been
talking
about
and
last
year
we
did
some
trial
with
kibe
edge.
It
was
a
kind
of
I
mean
in
line
with
our
thought
process,
because
we
support
csi
and
qbi
just
given
it
as
a
native
edge
platform,
so
we
could
deploy
a
soda
csi
with
cuba
edge
and
maybe
this
year
we
will
work
more
closely
with
them
to
try
to
have
some
kind
of
solutions
if
possible
on
edge.
H
But
this
is
not
a
real
focus
on
2022.,
as
qubit
is
an
echo
project
in
soda
foundation.
Probably
they
may
be
contributing.
H
H
So
we
think
that
a
unified,
open
source
solution
has
its
own
place,
especially
in
the
container
ecosystem,
because
container,
if
you
see
most
of
the
projects,
are
coming
from
the
open
landscape,
especially
at
the
platform
level
application
level,
there
can
be
proprietary
solutions,
but
platform
level,
more
or
less.
It
is
open
source
and
then
container
storage
as
a
service
and
cross
cluster.
This
is
one
of
the
pain
points
in
kubernetes
as
well
like.
A
H
I
I
hope
you
have,
you
have
seen
karma
also
right
from
hawaii.
A
H
Think
kevin
is
on
yeah.
Kevin
is
also
discussing
with
us
that
they
are
interested
in
the
storage
piece,
whether
we
can
support.
A
A
This
is
exactly
what
we're
saying
we're
saying:
we're
gonna,
take
your
your
storage
broker,
odf,
which
you
call
odf
and
then
integrate
that
with
our
federation
there,
which
is
much
beyond
karma,
and
all
that.
H
A
Yes,
yes,
we
should,
I
think,
next
meeting
we
should
have
brainstorming
session
in
next
meeting
as
well.
I
think
this
is
something
we
should
definitely
work
on
and
we
can
so
who's.
So
from
active
membership.
I
mean
who's
the
would
that
be
you
basically.
So
let's
say
we
don't
wanna.
Do
this
initiate
this
project.
You
know
so
yeah
odf
integration,
the
centaurus.
H
What
we
can
do
is
that
there
are,
we
can
discuss
further,
but
there
could
be
some
steps
like
maybe
centaurus
can
consider
to
join
us
a
echo
project
in
soda
sure,
and
then
then
then
what
we
can
do
is
that
usually
what
we
do
is
that
for
specific
project
collaboration
project
we
discuss
with
that
community
say,
for
example,.
G
H
G
A
I
much
good
thanks
thanks
for
everybody,
I
think
we
pretty
much
done
to
everybody
else
too.
I
know
mangani
has
to
rush
to
the
meeting
as
well.
Thanks
everybody.