►
From YouTube: Keynote: Exploring the Frontiers of Cloud Native - Bryan Che, Chief Strategy Officer, Huawei
Description
Don’t miss out! Join us at our upcoming hybrid event: KubeCon + CloudNativeCon North America 2022 from October 24-28 in Detroit (and online!). 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.
Keynote: Exploring the Frontiers of Cloud Native - Bryan Che, Chief Strategy Officer, Huawei
Cloud native technologies and adoption have certainly come a long way since the early days of small-cluster Kubernetes and containers. But, there are still so many exciting frontiers for cloud native to explore. At Huawei, we have been working to push cloud native deployments into the depths of even outer space, to wider scales than ever before, and to higher-level workloads such as AI and machine learning. And, we have been working together with others to broaden the cloud native community.
A
Good
morning,
so
those
of
you
who've
been
around
the
community
for
a
while
will
remember
the
early
days
of
cncf,
and
there
was
a
lot
of
discussion
around
what
was
the
right
scheduling
approach
and
which
scheduler
was
going
to
overall
win.
On
the
one
hand,
of
course,
we
had
kubernetes
here
with
cncf,
advocating
a
very
strong
cloud
native
approach,
but
we
also
had
this
world
coming
from
the
distributed
computing
side.
Some
of
you
remember
a
lot
of
the
you
know.
Hpc
schedulers
people
were
using.
A
A
A
This
has
been
really
a
big
focus
for
us
in
terms
of
how
do
we
start
to
bring
these
projects
and
these
different
areas
together,
and
a
lot
of
the
things
I
want
to
explore
with
you
today
are
some
of
the
new
open
source
developments
that
we've
been
driving
to
go
towards
this
new
area
of
what
we
start
to
call
distributed.
Cloud-Native
computing
and
there's
three
different
areas
that
I
want
to
start
to
focus
on.
In
terms
of
where
we
see
the
expansion
of
where
cloud
native
technologies
are
going.
A
The
problem,
though,
is
that
kubernetes
depends
on
state
sync
and
when
you
have
a
lot
of
different
vehicles
moving
around,
if
you
enter
a
tunnel,
if
you
go
into
an
area
with
bad
network
connection,
what
are
you
going
to
do?
Are
you
going
to
evict
your
node
or
your
pod?
Doesn't
work
so
well
when
their
vehicle
is
driving
around
the
road,
so
the
way
that
we've
been
handling?
A
That
traditionally,
is
that
we
default
and
fall
back
to
things
like
a
data
sync
and
that
works
okay,
but
it's
not
really
cloud
native,
and
so
what
we've
been
focused
on
is
how
project
called
kubej,
which
allows
us
to
bring
kubernetes
out
to
the
edge.
We
can
do
full
state
sync
and
then
we
can
also
start
to
bring
in
lots
and
lots
of
vehicles.
A
Today
today,
we
already
have
over
200
000
vehicles
a
year
being
manufactured
with
cube
edge
installed,
bringing
kubernetes
into
these
different
environments,
and
we
can
manage
a
cluster
of
100
000
vehicles
with
just
one
cluster.
So
it's
a
really
really
big
advancement
in
terms
of
how
do
we
push
kubernetes
into
these
areas
that
look?
Nothing
like
a
traditional
cloud
data
center.
A
Now,
a
lot
of
the
big
workloads
that
are
running
in
cloud
data
centers
today
are
all
around
machine
learning,
training
and
the
problem
is
that
you
deal
with
very,
very
sophisticated
scheduling,
algorithms,
that
don't
work
so
well
on
cloud
native,
because
you've
got
to
understand
things
like
fair
share.
You've
got
to
understand
resource
scheduling
over
time.
You've
got
to
have
you
know
very,
very
dedicated
resource
affinity
and
so
on,
and
this
is
very
challenging
when
you
take
a
look
at
bringing
that
into
cloud
native.
A
So
one
of
the
things
that
we've
introduced
is
a
new
cncf
project.
You
know
from
a
couple
years
ago
called
volcano
and
this
adds
a
batch
system
onto
kubernetes.
So
we
can
start
to
bring
these
two
worlds
together.
We
can
start
to
give
you
extreme
skill
and
a
lot
of
these
batch
primitives,
but
in
a
cloud
native
environment,
and
so
we've
seen
massive
improvements
in
terms
of
how
do
we
start
to
bring
these
very
sophisticated
workloads
into
kubernetes
with
volcano
huge
advancements
in
throughput,
scalability
and
resource
efficiency?
A
The
third
area
that
we've
been
focused
on
is
how
do
we
start
to
do
multi-cluster
and
bringing
kubernetes
across
clouds
so
that
it
becomes
really
ubiquitous?
Everyone
wants
to
do
multi-cloud
today,
but
the
problem
is
that
when
you
take
a
look
at
going
to
different
clouds,
everyone's
kubernetes
is
a
little
bit
different
from
the
other.
So
what
a
lot
of
people
do
is
they'll
take.
Perhaps
you
know
one
kubernetes
distribution
and
install
them
from
cloud
to
cloud
to
cloud.
This
still
has
a
couple
limitations,
though
now
first
you're
tying
yourself
into
one
particular
vendor.
A
The
second
is
that
you're
still
dealing
with
the
constraints
of
scheduling
into
a
single
cluster
boundary.
You
don't
really
go
cross-cluster
cross-cloud
in
a
single
pool,
and
so
this
is
why
we're
excited
about
this
new
kmr
project,
which
is
the
first
cncf
project
in
terms
of
how
do
we
deal
with
cross-cluster
cross-cloud
scheduling
it's
in
cncf,
so
it's
open
for
everyone,
and
it's
also
very
much
focused
on
how
do
we
start
to
combine
these
different
clusters
into
one
pool
that
works
across
different
environments?
A
So
a
new
approach
that
uses
standard
kubernetes
apis
to
tie
everything
together
and
then
the
final
thing,
just
just
as
I
wrap
up
we're
very
excited,
also
to
expand
the
community
of
who's
involved
with
cloud
native.
One
of
the
things
that
we've
been
focused
on,
especially
in
china,
is
that
there's
a
lot
of
end
users
who
are
still
not
cloud
native
and
how
do
we
bring
them
into
the
community?
A
So
we've
been
partnering
with
the
cncf
to
build
out
what
we
call
a
cloud
native
elite
club,
and
this
is
on
boarding
a
lot
of
new
end
users
into
the
overall
cncf
community
to
become
adopters
of
all
the
technologies.
We're
all
talking
about
here.
One
of
the
things
we're
really
excited
is
we've.
Also
been
focusing
a
lot
about
how
do
we
bring
more
women
into
technology
through
this
initiative,
and
so
we're
very
excited
about
the
growth
that
we're
seeing
in
where
cloud
native
is
going
as
well
as
what
the
community's
been
doing.