►
From YouTube: Welcome + Opening Remarks - Ricardo Rocha, CERN
Description
Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon Europe 2023 in Amsterdam, The Netherlands from April 17-21. Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Welcome + Opening Remarks - Ricardo Rocha, CERN
A
All
right,
I
think
we
can
start
so
welcome
everyone
to
the
second
batch
on
HPC
today,
thanks
for
for
coming,
we'll
have
a
pack
day,
so
I'll
just
do
a
quick
introduction
and
then
we
we
go
deep
into
the
directly
into
the
talks.
The
first
thing
is
to
thank
our
dominant
sponsor
the
Oracle
Cloud
infrastructure.
So
thanks
a
lot,
then
just
a
quick,
quick
reminder.
So
this
is
the
second
time
we
organize
this
HPC
and
batch
day.
A
The
goal
of
the
first
one
was
to
to
have
an
event
that
would
kind
of
bridge
the
traditional
HPC
world
that
has
its
own
conferences
and
its
own
communities
and
then
the
cloud
native
world
and
to
have
any
there
was
an
event
missing
that
would
kind
of
bridge
bridge
the
Two
Worlds.
A
So
we
had
the
first
one
back
in
Valencia
coupon
earlier
this
year
it
was
quite
successful,
so
we
had
in
person
we
had
40
over
40
people
and
almost
200
on
the
live
stream
and
I
just
put
here
some
some
stats
that
we
got
from
the
event
that
are
also
quite
interesting
to
have
an
idea
of
what
we
are
facing.
A
So
if
you
look
at
the
jump
jump
function,
you
can
see
that
half
of
the
people
are
either
developers
or
or
kind
of
devops
is
admin
people,
but
there's
then
the
rest
kind
of
splits
on
different
roles
in
the
organizations.
So
I
guess
people
are
feeling
what's
possible
in
the
using
Cloud
native
for
hpcn
batch
and
the
event
was
very
successful.
A
So
we
decided
to
continue
and
this
time
instead
of
a
half
day
that
we
had
in
in
Valencia.
We
have
a
full
day
here
in
in
Detroit,
so
welcome
you're
in
the
right
place.
We
start
now
we
go
until
five.
I
remember
remind
everyone
that
we
have
the
networking
reception
at
5,
PM,
5
PM
as
well.
So
don't
miss
that
and
then
for
the
agenda.
We
have
one
keynote,
then
we'll
have
eight
full
sessions
and
we'll
have
four
lightning
talks.
A
This
was
some
of
the
feedback
we
got
from
the
first
event
as
well
that
people
wanted
more
of
lightning
talks
as
well
and
thanks
everyone
for
for
sending
the
submissions.
It
was
interesting
to
to
go
through
them
and
do
the
selection.
A
A
Some
sort
of
high
performance
or
high
through
workloads,
running
batch
workloads
as
well,
and
doing
some
sort
of
AIML
kind
of
workloads
as
well,
and
the
specificity
of
this
event
is
that
we
basically
all
are
interested
or
already
running,
Cloud
native
based
infrastructure,
meeting
kubernetes
and
all
the
tools
around
it,
and
we
are
all
looking
for
the
best
way
to
to
to
build
on
the
on
the
tooling,
from
coming
from
the
cloud
native
Community
to
to
kind
of
simplify
the
deployments
we
have
for
HPC
and
batch.
A
There
are
a
couple
of
groups
that
we
meet
so
if
you're
new
to
this
community
there's
a
couple
of
places
that
you
can
look
at
the
first
one
I
would
highlight,
is
the
kind
of
less
newly
but
still
newly
formed
the
batch
working
group
in
in
kubernetes,
and
that's
where.
Basically,
the
focus
is
on
expanding
the
support
in
kubernetes
to
to
to
batch
and
HPC,
and
this
is
things
like
job
API
and
the
scheduler
there's.
Also,
the
cncf
batch
system
initiative
proposal:
it's
not
a
proposal
anymore.
A
So
if
you
go
to
the
to
the
link
there
in
the
QR
code,
you
will
actually
see
that
it's
a
new
newly
formed
batch
working
group
inside
the
cncf-
and
this
is
kind
of
complementary
to
the
to
the
kubernetes
batch
working
group
in
the
sense
that
it
Aggregates
the
efforts
from
different
projects
in
the
cncf
that
are
not
necessarily
the
core
kubernetes,
but
that
have
been
developable
over
the
last
few
years,
extending
the
functionality
to
support
HPC
and
batch.
A
And
then
the
final
one
I
will
highlight
is
the
research
User
Group.
This
is
a
a
group
that
is
focused
on
Research
workloads
that
was
formed
as
an
idea
back
in
kubecon
Barcelona,
and
it
has
been
running
for
a
few
years
and
there
you
will
see
a
batch
as
a
topic
HPC
as
well,
but
you
will
see
other
things
like
how
do
I
run
notebooks
properly
using
kubernetes
or
how
do
I
run
engineering
applications
and
optimizations.
So
there
there's
kind
of
a
more
broad
scope,
but
still
still
in
context
with
the
topic
today.
A
So
the
challenges
that
we
all
face-
the
kubernetes
batch
API,
the
initial
design-
was
focusing
on
particular
workloads.
There
are
ideas
on
how
to
enhance
and
extend
this
functionality
to
support
the
things
we
need:
The
Primitives
for
advanced,
queuing
and
fair
share.
So
this
is
not
something
that
exists
necessarily
in
the
kubernetes.
A
Scheduler
tools
have
been
extending
that
to
support
this
functionality,
go
scheduling
and
similar,
but
there's
an
interest
to
to
have
this
built
into
the
core
scheduler
as
well,
then
a
topic
that
is
maybe
not
in
any
of
the
groups
before,
but
it's
very
close
to
to
this
kind
of
problem
as
well,
which
is
support
for
multi-cluster
deployments
and
hybrid
deployments.
So
you
might
have
a
large
on-premises
infrastructure,
but
you
want
to
burst
out
to
public
cloud
in
some
cases
or
access
to
accelerators
things
like
this
rootless
containers.
A
If
you
run
an
HPC
Center,
this
is
kind
of
a
big
thing.
So
this
is
also
on
topic
and
then
all
sorts
of
optimizations
of
the
the
workload
submission,
but
also
of
the
workload
execution
on
the
nodes,
and
then
I
mentioned
before
specialized
hardware.
And
then
the
overall
idea
is
that
to
try
to
reduce
or
avoid
somehow
fragmentation
of
this
functionality
across
multiple
tools
and
where
it
makes
sense
to
kind
of
consolidate
in
a
in
a
single
place.
A
So
for
today,
again,
if
you
look
at
the
agenda,
you
will
see
quite
a
lot
of
focus
on
batch
workloads
for
for
this
kind
of
activities
I
mentioned.
The
other
main
goal
is
to
bring
together
Community
experts,
and
this
is
the
people
working
on
the
tools,
but
also
the
end
users
that
have
feedback
on
what
they
actually
need.
A
The
topics
on
the
talks
we'll
have
programming
paradigms
in
batch
and
HPC,
covering
scheduling
and
optimization
hybrid
deployments
I
mentioned
earlier,
and
then
user
reports
and
a
lot
of
discussions
so
we'll
be
packed,
and
hopefully
a
lot
of
discussion
as
well.
So
thank
you
again
for
for
sponsoring
to
for
Oracle,
Cloud
infrastructure
and
that's
all
I
have,
and
we
can
go
straight
to
the
our
first
talk
from
from
Oracle.