►
Description
Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon Europe in Amsterdam, The Netherlands from April 17-21, 2023. 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: From Silicon to Serverless: A Full Stack Journey - Cathy Zhang, Senior Principal Engineer, Intel
Speakers: Cathy Zhang
The cloud native journey starts with silicon and ends with delivered applications. In this talk, Cathy Zhang will walk us through everything from cores to cloud.
A
A
A
Intel
operates
at
virtually
every
layer
of
the
cloud
technology
stack.
We
have
started
to
take
a
software
first
approach.
An
open
source
Community
plays
a
key
role
in
that
a
work
in
the
open
source
and
the
close
collaboration
we
have
been
had
with
the
Upstream
developers
are
crucial
to
our
mission
and
our
success.
It
is
our
goal
to
make
sure
that
the
open
source
Community
is
benefiting
as
much
from
us
as
we
are
from
them.
A
One
great
thing
about
cloud
is
that
it
hides
the
complexity
of
the
hardware
and
infrastructure
from
the
end
users.
You
may
think
that
the
low
layers
of
the
cloud
technology
stack
do
not
matter
anymore,
but
after
joining
Intel
for
over
one
year,
I
have
realized
how
important
the
low
layer
work
is
to
the
cloud.
I
have
had
the
opportunity
to
really
dig
in
and
see
how
efficiency,
acceleration
and
isolation
are.
The
Silicon
layer
can
help
with
cost
performance
and
security
of
the
cloud.
A
Just
to
give
you
one
example
about
how
Hardware
topology
aware
scheduling
helps
with
performance
and
cost.
You
may
know
that
the
CPU
and
memory
on
a
physical
server
are
grouped
into
numer
nodes.
Obviously,
internumal
node
communication
will
be
slower
than
in
the
numer
node
communication.
Therefore,
if
the
cloud
scheduler
can
take
the
low
layer,
Hardware
topology
into
consideration,
your
workloads
will
be
your
workloads.
Performance
will
be
enhanced.
The
CPU
and
memory
on
the
server
will
be
used
more
efficiently,
which
results
in
lower
Cloud
costs.
A
A
We
need
confidential
Computing
Technologies
like
sgx
and
TDX,
to
build
a
trusted
execution
environment
around
the
data
when
the
data
is
being
is
being
processed,
Hardware,
topology,
aware
scheduling
and
confidential
Computing
Technologies
are
just
two
of
the
areas
we
have
been
working
with
the
Upstream
Community
to
take
the
cloud,
cost
performance
and
security
to
the
next
level
to
take
the
crowd
to
the
next
level.
There's
still
a
lot
of
innovation
and
work
to
do.
The
industry
is
moving
from
bare
metal
to
virtual
machine
to
container
and
then
to
serverless
function.
A
Serverless
enables
developer
velocity
velocity
by
offloading
from
the
developer
side
to
the
cloud
platform
side,
not
just
how,
but
also
when
to
run
a
workload.
Serverless
offers
many
benefits,
but
it
also
brings
some
challenges.
Two
of
the
major
challenges
are
code
start
latency
and
the
auto
scaling
speed
in
response
to
a
burst
of
traffic.
A
To
solve
these
challenges,
we
have
been
working
on
a
new
snapshot
based
container
image
to
start
from
a
workload
which
greatly
reduces
the
cold
start
latency.
We
are
also
developing
a
new,
auto
scaling
approach
to
increase
the
scaling
speed
when
there
is
a
burst
of
traffic.
This
is
just
one
example
of
the
Innovative
work
to
solve
the
pain
points
in
the
cloud.
A
Intel
has
contributed
to
many
open
source
projects.
This
slide
just
shows
a
few
of
the
projects
we
have
been
contributing
to.
As
you
can
see
our
contribution
to
the
cloud
spans
all
the
software
stack.
If
you
are
curious,
you
can
learn
more
about
our
work
at
Intel
boost
in
the
solutions
showcase.
You
may
be
surprised
at
what
you
will
see
and
we
would
love
to
get
your
feedback
there's
a
lot
more
I
would
like
to
cover,
but
I
can
only
say
so
much
in
five
minutes.