►
Description
Don’t miss out! Join us at our upcoming events: EnvoyCon Virtual on October 15 and KubeCon + CloudNativeCon North America 2020 Virtual from November 17-20. Learn more at https://kubecon.io. The conferences feature presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Intro + Deep Dive - Provider IBM Cloud - Sahdev Zala & Khalid Ahmed, IBM
In this session, the Provider IBM Cloud leads will give an overview of this Cloud Provider subproject, its activities and discuss the recent developments. Everyone --active contributors, new contributors, and conference attendees -- with interest in the Provider IBM Cloud are welcome to attend.
https://sched.co/ZevW
A
Hi,
my
name
is
sadhu
zala.
I
am
an
open
source
developer
at
ibm
and
I
contribute
to
kubernetes.
I
am
a
co-chair
of
this
project
called
provider
ibm
cloud
which
is
a
sub-project
of
the
kubernetes
cloud
provider,
sub
special
interest
group
and
I'm
also
one
of
the
maintainers
for
the
hd
project.
Khalid.
Would
you
like
to
introduce
yourself
please.
A
All
right,
thank
you
khalid
all
right,
so
welcome
to
this
video.
We
have
a
few
items
in
the
agenda.
We'll
provide
an
overview
of
the
sig
cloud
provider
followed
by
you
know
this
sub
project
for
ibm
clouds,
we'll
go
through
the
structure
activities
and
then
we'll
talk
about
cluster
api
provider,
ibm
cloud,
which
is
a
github
repository
owned
by
this
project.
A
I'll
talk
about
ibm
cloud
provider
and
what's
how
you
know
what
roadmap
we
have
for
the
future
there
and
then
we'll
talk
about
the
hybrid
cloud
ecosystem
around
ibm
cloud
all
right,
so
the
cloud
port
or
spatial
interest
group
is
one
of
the
sig
in
kubernetes
and
it
does
a
lot
of
things.
Actually,
it
owns
the
code
around
cloud
for
interface
in
you
know
all
the
related
work.
A
As
you
may
already
know,
there
are
specific
control
groups
that
comes
you
know,
with
communities
to
enable
kubernetes
functionalities.
For
you
know
when
you
deploy
on
a
particular
cloud
and
those
control
loops
includes
many
different
things:
different
interfaces
to
enable
the
load
balancer,
you
know
for
the
nodes
to
find
out
the
ips
zoning
information.
A
But
that
repository
that
has
the
code
for
this
cloud
loops
interface,
it's
it's
it's
owned
by
cloud
provider.
There's
a
url
here!
You
can
learn
more
there.
The
seek
also
owns
all
the
sub
projects.
They
were
formally
known
by
individual
cloud
provider,
sigs
like
c
aws
or
c
kazoo
or
gcp
sig,
ibm
cloud
and
and
similarly
seek
openstack
c
vmware.
A
A
A
A
A
A
As
for
activities
are
concerned,
you
know
segments
every
month
with
the
slack
channels
with
emails
mailing
lists.
You
know
we
don't
see
a
need
to
meet
every
week,
kind
of
things
which
we
used
to
have
in
the
past,
but
now
the
segments
every
month
last
wednesday,
2
p.m.
Eastern
time,
if
you're
in
different
time
zone,
if
you
can
you
know
if
you
could
not
attend
those.
A
As
far
as
you
know,
upstream
contributions
are
concerned.
We
have
a
github
repositories
called
cluster
api
for
driving
cloud.
You
you
may
only
know
about
the
cluster
api
project,
it's
a
cubans
project
and
the
ibm
cloud
api
provider.
It's
basically
an
implementation
of
this
project
and
it
it
provides.
You
know,
optional
functionalities,
on
top
of
the
core
kubernetes
around
managing
the
life
cycle
of
kubernetes
cluster
khalid
will
talk
more
on
that
and
then
we
are
working
on
on
support
for
out
of
free,
ibm
cloud
provider.
A
I'll
be
talking
about
that
right
after
I'm,
I
talk
manager
offerings,
we
have
so
iks
ibm
cloud.
Kubernetes
service
is
one
of
the
managed
offering
to
create
cumulus
clusters,
so
iks
implements
the
ibm
cloud
provider
to
provide
the
functionalities
when
you,
as
an
user,
create
a
cubase
cluster
using
iks
and
as
as
you
may
know,
there
is
a
conformance
program
in
communities.
A
So
you
know
this:
the
iks
is
a
certified
cuban
is
offering
you
can
read
more
about
iks
on
on
this
url
provided
right
here
in
this
slide,
rocks
red
hat
openshift
I
mean
clouds
is
another
managed
offering
openshift.
As
you
may
know,
already
know
it's
a
product
that
is
built
on
top
of
kubernetes
and
provides
many
additional
features
there.
A
It
is
also
a
certified
kubernetes
offering
and
you
can
learn
more
about
openshift
on
the
ibm.com
all
right,
so
the
ibm
cloud
provider
there
there
has
been
some
discussions
going
on.
So,
if
you're,
following
the
discussion
discussions
in
the
community
around
cloud
provider,
seek
you
probably
know
that
in
all
those
control
loops
I
mentioned
earlier
the
code
where
part
of
the
core
kubernetes
so
they're
part
of
the
cube
controller
manager
and
their
architecture
are
called
kcm.
A
But
the
community
realized
that
you
know.
That's
that's
that
that
that
is
something
needs
to
be
decoupled
from
the
core
communities
so
that
the
kubernetes
and
cloud
provider
can
evolve
independently.
So
that
code
is,
you
know
the
process
still
going
on.
Some
of
the
providers
already
moved
the
code
outside
it
called
out
of
free
codes
versus
entry,
but
ibm
cloud
provider.
We
already
have
support
for
ccm.
That's
what
we
use
now.
A
We
implement
the
first
three
one,
the
load
balancer.
We
have
two
offerings
for
the
classic
cloud
infrastructure,
the
nlb
version
1.0
iptable
based
and
the
2.0,
which
is
ipvs
based
instances.
Why
is
instances
v2?
Is
it's
a
new
new
interface
with
which
we'll
be
working
to
implement
and
and
and
you
know
we
implement
the
zones
clusters
and
routes?
We
have
no
need
to
implement
them
right
now
for
the
routes
we
use,
what
calico
provides
for
the
future.
A
You
know,
as
I
mentioned
there
was
you
know
in
the
initial
days
of
communities,
early
versions,
the
code
that
that
enables
the
cloud
provider
functionality
is
the
control
loops.
They
are
part
of
the
code
kubernetes.
So
at
one
point
the
community
decided
to
stop
taking
any
new
code
there
from
any.
You
know
new
cloud
provider
and
that's
where
you
know
we
started
exporting
it.
So
we
don't
have
code
there.
A
What
we
have
is
it's
it's
based
on
ccm,
so
they
are
pretty
much
ready
for
the
out
of
three
I
out
of
three
implementation,
so
we
will
be
open
sourcing
it.
You
know
as
we
following
what's
going
on
there
and
you
know
we'll
be
working
on
that,
but
on
that
line
along
with
that,
we'll
work
on
improving
the
documentations
and
you
know,
align
the
bills,
taste
and
releases
process
with
the
committee.
A
B
Sure,
thank
you.
Thank
you.
So
I'll
talk
about
a
few
projects
that
we're
working
on
in
the
areas
of
hybrid
cloud,
and
by
hybrid
we
mean
not
only
the
ability
to
run
on
premise,
as
well
as
various
public
clouds,
but
also
hybrid,
in
the
sense
that
applications
and
workloads
are.
Although
the
cloud
native
workloads
run
in
kubernetes,
there
are
other
were.
B
So
this
product
is
available
it
the
product
for
that
is
called
the
red
hat,
advanced
cost
management
for
kubernetes,
and
you
can
go
and
download
and
transmit
in
your
environment,
and
it
gives
a
base
management
plane
for
kubernetes
to
deploy
out
and
manage
multiple
kubernetes
environment.
So
if
we
can
go
to
the
next
slide,
one
aspect
of
that
is
the
policy
management,
and
this
is
where
both
red
hat
and
ibm
are
encouraging
an
ecosystem
of
contributions
around
how
to
manage
and
develop
policies
in
a
open
but
kubernetes
native
model.
B
B
I
want
to
talk
about
another
related
project.
This
is
about
manage
iq
again.
This
is
a
totally
open
source
project.
The
history
of
that
was
that
it
was
based
on
an
acquisition
that
red
hat
made
and
and
it
led
to
a
product
from
from
red
hat
called
cloud
forms,
but
as
part
of
the
acquisition
and
a
relationship
with
with
red
hat.
B
This
manage
iq
project
is
being
kind
of
handled
by
ibm
now,
and
it
is
part
of
an
offering
from
from
ibm
I'll
talk
a
bit
about
that
later
right,
but
one
of
the
new
things
about
this
project.
So
what
what
manage
iq
does
is
allows
you
to
manage
traditional
virtual
environments,
so
things
like
vmware
or
kbm,
based
solutions,
kubernetes
and
red
hat,
and
so
on,
manage
container
native
virtualization.
B
This
allows
you
to
bring
in
traditional
workloads
with
traditional
virtualization
into
a
kubernetes
based
management
model.
The
traditional
managed
iq
deployment
was
on
bms.
The
latest
iteration
of
that
is
a
modified
version
of
manage
iq.
Where
the
manage
iq,
control,
plane
itself
can
run
on
kubernetes
and
we've.
You
know
made
manage
iq
into
a
set
of
microservices
that
can
be
deployed
within
a
kubernetes
environment,
and
it
supports
a
variety
of
you
know:
providers
that
I
mentioned
right.
It
allows
you
to
connect
to
a
variety
of
cloud
providers
as
well
as
on-premise.
B
B
Okay,
so
this
is
another
ecosystem
project
that
helps
you
know
the
hybrid
management
aspect
of
the
thing,
so
the
open
cluster
management
deals
with
kubernetes
and
and
openshift
the
this
project
deals
with
traditional
vm
based
workloads
and
ias
workloads
within
within
your
your
cloud
environment.
The
next
chart,
so
assad
have
alluded
to
the
cluster
api
project,
and
this
is
about
the
lifecycle
management
of
individual
clusters,
and
this
is
ibm,
contributes
to
that.
B
We
have
a
ibm
cloud
plugin
that
that
allows
you
to
declare
how
you
want
a
cluster
to
be
provisioned
and
from
a
bootstrap
node.
You
can
provision
multiple
clusters
and
it
interacts
with
the
infrastructure
providers
to
provision
the
appropriate
infrastructure
resources
and
met
and
connect
the
the
all
the
kubernetes
control
plane
elements,
the
masters,
the
xcd
and
the
worker
nodes
and
proxy
et
cetera
and
allows
you
to
you
know:
do
things
like
scaling
up
and
scaling
down
number
of
worker
loads
and
so
on
and
there's.
B
B
Okay,
so
hybrid
applications
are
are
about
workloads,
so
there's
a
github
project
called
the
hybrid
app
io
and
there's
a
link
to
the
documentation
and
what
this
project
does
is
to
extend
the
notion
of
applications.
It's
based
on
the
sig
app
model
that
that
is
in
the
kubernetes
ecosystem,
but
it
extends
it
to
being
able
to
manage
multi-component
applications
that
span
both
kubernetes
resources,
as
well
as
traditional
vm
resources
and
in
the
future.
Looking
at
you
know
other
types
of
resources
from
a
cloud
infrastructure.
B
Okay
and
let's
switch
to
the
next
chart-
and
this
is
from
a
a
product
perspective
or
an
offering
perspective.
This
is
what
we
bring
together:
a
number
of
the
open
source
projects,
the
hybrid
app,
the
open
cluster
management
or
the
red
hat
advanced
cluster
management,
as
well
as
manage
iq
to
give
a
open
standards-based
management
plan
for
hybrid
applications.
Right,
it's
built
on
kubernetes,
it's
you
know
trying
to
avoid
any
lock-in.
B
It's
built
around
a
number
of
open
source
components
and
it
enables
an
ecosystem
because
it
is
based
on
kubernetes
and
kubernetes
is
an
open
standard
in
in
the
community.
It
allows
other
partners
and
other
projects
to
plug
in
in
the
areas
of
operational
automation,
application
management,
security
and
governance
and
visibility,
and
so
on.
Okay,
I
won't
go
through
all
the
other.
You
know
components
that
that
make
up
this
product,
but
it
leverages
a
lot
and
contributes
to
a
lot
of
those
ecosystem
projects
that
you
see
in
there.
B
B
So
this
is
a
project
that
tries
to
marry
ai
with
scheduling
capabilities.
It
uses
a
learning
based
algorithm
reinforcement,
learning
algorithm
to
allow
you
to
detect
certain
problems
that
are
kind
of
hard
to
do.
If
you
do
a
traditional
algorithmic
model,
things
like
pods,
always
failing
on
a
certain
node
that
they
can
be.
That
can
happen
for
a
number
of
reasons.
For
example,
the
the
pvc
is
not
bound
the
nasa
access
on
a
node
has
been
revoked
and
the
device
plug-ins
for
gpus
are
not
configured.
Probably.
B
B
What
we
did
is
a
a
scheduler
extender
that
uses
a
reinforcement,
learning
algorithm
to
adaptively,
learn
where
a
particular
type
of
workload
is
best
placed
based
on
the
variety
of
conditions
and
based
on
learning
from
different
failure,
modes
and
different
cases
of
bad
performance
and
so
on.
So
the
scheduler
extender
basically
extends
the
kubernetes
scheduler
to
take
into
account
the
results
of
the
ai
algorithm
to
direct
the
workloads
to
a
particular
node
that
best
fits
that
that
set
of
conditions
that
that
might
cause
problems.
In
terms
of
that.
B
B
So
this
last
project
again
the
github
location.
Is
there
it's
again
around
scheduling
but
about
multi-cluster
how
how
to
dispatch
workloads
from
a
central
location
to
multiple
kubernetes
clusters,
and
this
addresses
some
of
the
challenges
that
that
people
have
especially
around
managing
batch
type
of
workloads,
complex
workloads,
so
there's
a
notion
of
a
app
wrapper
which
can
wrap
a
variety
of
kubernetes
resources
that
gets
scheduled
or
placed
into
this
dispatcher
running
in
a
dispatcher
cluster
and
from
there
they
can
be
deployed
out
to
different
execution
clusters.
B
This
looks
at
you
know,
available
capacities
for
clusters,
as
well
as
energy
awareness
of
the
the
the
clusters,
energy
consumption
and
so
on,
and
tries
to
make
a
intelligent
decision
about
where
to
run
these
workloads.
Okay,
so
that's
it
from
my
side.
I
think
that
that's
the
overall
talk,
I
don't
know
if
there's
any
other
charts,
but
okay,
any
closing
comments.
Out
of
so.
A
A
Shout
out
through
slack
channel
through
mailing
list,
you
know
various
ways.
If
you
have
any
question
about
this
presentation
about
the
project-
and
you
know
you
would
like
to
have
your
participations
in
the
c
club
folder
in
the
provided,
ibm
cloud
project,
so
look
look
forward
to
work
with
you
and
with
that
khalid
thanks
again,
thanks
everyone.