►
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.
Keynote: The Next Big Things for Istio - Idit Levine, Founder + CEO, Solo.io
After 6 years, most open source projects are focusing more on stability than innovation. But as more companies deploy Istio at scale, the need for new innovation is accelerating. In particular, the data plane, proxy modularity, security, and observability are all areas with rich new innovations that will make it easier to operate, lower cost, and improve application performance. This keynote will highlight why Istio is the future of service mesh.
A
B
All
right
so
I'm,
sorry,
it's
Nick
said,
and
he
did
Levine
I'm,
the
father
and
the
CEO
of
solo.
So
there's
a
company
who
is
focusing
on
the
service
mesh
and
projects
like
psyllium
sdo,
we're
doing
Envy
for
a
long
long
time.
So
I'm
not
going
to
focus
a
lot
about
the
company
itself,
I'm
not
going
to
focus
about
what
I
think
will
be
the
future
of
STL.
So
in
general
it
really
really
simple
where
we
are
playing
as
a
company.
B
I
think
that
in
the
beginning
you
know
we're
all
here,
because
Cuban
has
become
very,
very
meaningful
for
us.
It's
what
in
Solo
we
like
to
call
Cloud
Nanny
1.0,
and
what
does
it
mean
is?
Basically
you
know
we
adopt
microservices
we
adapt
kubernetes.
We
were
pretty
excited
about
that.
But,
honestly,
that's
also
when
the
problem
started
and
to
me
personally
I.
Don't
think
that
when
we
adopting
this
amazing
technology
that
giving
us
a
lot
a
lot
of
benefit
and
velocity
to
make
a
lot
of
sense
to
keep
the
same.
B
For
instance,
Legacy
API,
Gateway,
they're
running
active,
active
Cassandra
behind
the
scene
or
I
think
that
when
you're
doing
this,
it
makes
sense
to
kind
of
like
fit
to
the
ecosystem
of
kubernetes
and
do
everything
the
clarity
and
eventually
consistent
so
githubs
can
play
a
big
big
play
here
and
everything
that
related
to
observability.
Now
you
have
100
hundreds
and
thousands
of
micro,
Services
everywhere
replica
of
them.
You
don't
really
know
even
where
the
request
is
Flowing.
So
again,
how
do
you
understand
what's
going
on
in
your
infrastructure?
Are
they
making
sure?
B
Because
now
everything
is
on
the
network,
that
you
will
trust
it
right?
Zero
trust,
net,
a
security,
it's
a
big
big
player
here
and
I
think
this
is
what
most
of
us
here
multi-class
there
again.
We
have
a
lot
of
customers.
They
are
running,
you
know
in
multi-cluster,
environment,
multi-cloud,
environment,
right,
so
basically
kind
of
like
fell,
fell
over
and
so
on,
and
I
think
that
there
is
some
interesting
technology
that
we
can
kind
of
like
Leverage
like
serverless
and
graphql.
That
is
very
interesting.
B
So
what
solo
is
focusing
is
what
we
call
Cloud
native
2.0
right,
which
is
basically
all
this
part
above
that
make
you
successful
in
Cloud
native
one.
So
what
do
we
do?
So?
We
have
three
pillars
to
our
platform.
We
have
a
platform
called
glue
and
there
is
three
pillar
to
it.
The
first
one
is
the
glue
Gateway
really
really
nicely
aligned
with
the
Gateway
Market
and
basically,
what
we
did.
We
took
the
sto
in
grass
and
we
enhance
it
by
actually
all
plugins
and
Android
filter
to
actually
make
it
full-blown
API
Gateway.
B
So
again,
it's
SEO
native,
the
only
one
that
exists
in
the
market,
so
we
are
basically
taking
an
upstream
sto
and
we're
building
the
API
Gateway
on
top
of
it.
So
that's
the
first
one
again
very
nicely:
aligned
with
Gateway
the
other
one
is
Google
match
and
glue.
Mesh
is
basically
rst
or
distro
and
besides
the
fact
that
again,
we're
taking
Upstream
estia
we're
adding
a
lot
a
lot
of
extensibility.
So
if
this
is
extensibility
of
is
to
use
like
certification,
for
instance,
if
this
is
extensibility
around
multi-clusters,
and
how
do
you
do
multi-cluster?
B
How
do
you
fail
over?
How
do
you
write
base?
Look
a
localization
right.
It
will
start
with
the
close
one
if
it's
not
there,
you're
failing
over
to
the
other
one
and
so
on.
So
there's
a
lot
of
kind
of
like
a
enhancement
that
we
did
to
the
regular
app
streamstr.
Besides
the
fact
that
we're,
of
course,
making
sure
that
all
the
CVS
is
for
the
you
know,
n
minus
4,
and
that
you
are
successful
in
production,
the
last
but
not
least,
is
basically
Google.
B
Network
and
glue
network
is
How
We,
believe
we
can
take
the
lower
level
and
basically
enhance
the
service
mesh.
So
this
is
stuff,
that's
related
to
the
cni.
Specifically,
we
are
going
to
drive
any
CNN
that
you
have,
but
we
are
personally.
If
you
will
ask
one
from
us,
we
will
give
you
cilia
them
exactly
like
you're
doing
with
sto
or
what
we're
doing
with
SEO
Upstream
psyllium,
with
enhancement,
making
sure
that
it's
good
in
production
and
so
on,
and
besides
that
we're
leveraging
a
lot
of
vpf
to
enhance
it.
B
So
everything
that
related
to
layer,
seven
of
layer,
you
know
layer,
four
observability,
but
also
layer,
7
observability,
acceleration
of
your
network.
How
can
you
know
we
can
redirect
traffic
to
when
we
will
talk
about
ambient
and
so
on,
so
that
International
is
the
three
pillars
again
nicely
in
line
for
a
market
that
you
know
Gateway,
mesh
and
networking
right
in
cni.
B
B
B
So
you're
on
you
know
all
you're
all
in
the
space
of
service
range
and
therefore
you
know
that
the
the
kind
of
like
model
that
most
of
the
people
are
using
today
is
basically
the
sidecar,
which
means
that
if
a
request
is
flying,
it
will
go
to
the
first
side
car.
If
we
go
to
the
second
side
car
and
it
will
go
to
the
destination.
B
So
this
is
great,
but
there
is
some
problem
with
this
model,
so
the
first
problem
that
kind
of
like,
for
instance,
interesting
or
no
problem,
it's
a
it's
making
it
harder
to
operate
so
the
first
one
is
everything
that
related
to
operational
problem.
For
instance
right.
Let's
say
that
you
want
to
upgrade
your
microservices
right,
it's
coming
up,
but
what?
If
the
sidecar
is
not
coming
immediately
up
and
it's
happening
right,
you
can
you,
you
can't
100
control
that.
B
So
the
problem
is
that
if
your
application
is
something
like
SQL,
for
instance,
there's
a
good
chance
that
it
will
crash
right
so
a
problem,
so
the
request
will
go
and
you
will
lose
it.
The
second
one,
an
example
and
I'm
just
putting
some
example
that
we
see
from
our
customers
right.
Let's
say
that
you
want
to
run
some
job.
This
is
great.
The
job
is
done,
but
guess
what
your
sidecar
is
still
there.
So
now
it's
not
going
to
be
cleaned.
B
So,
theoretically,
you
might
find
in
a
lot
of
kind
of
like
sidecar
that
doesn't
have
even
application
attached
to
it
in
your
infrastructure,
a
little
bit
of
waste
in
terms
of
performance
right
I
mean
if
you're
going
from
one
microservices
to
the
other.
You
know
it
would
be
faster
than
going
through
the
side
car.
It
is
what
it
is
right
and
then
I
say
not
least,
is
honestly
there's
a
lot
of
proxy
out
there
right.
So
that's
pretty
expensive.
B
Okay,
so
a
year
and
a
half
ago
still
decided
after
working
with
a
lot
a
lot
of
customers
to
try
to
attack
that
problem.
We
created
an
internal
project
and
we
start
working
on
how
to
create
the
future
of
sto
like
eight
months
or
so
after.
We
discovered
that
Google
doing
the
same
thing
internally,
so
we
decided
to
partner
and
we
worked
on
it
under
NDA
us
and
Google,
and
a
month
ago
we
announced
mbn.
B
So
what
is
mbn
and
how
it's
making
our
life
easier.
So
it's
very
simple.
We
are
very,
very
careful
and
again
there
will
be
a
very
deep
dive
talk
here,
so
we'll
go
fast,
but
basically
we're
making
sure
that
we
separate
between
layer,
four,
who
is
usually
honestly,
a
very
simple
operation
to
layer,
seven
right
so
we're
putting
a
layer
for
proxy
on
the
Node.
It's
very
simple
right
now
is
Envoy
we're
working
internally
in
the
sdo
community
externally
and
that's
the
community
in
order
to
make
it
a
rust
implementation.
B
B
Once
it's
equipped,
it
will
go
to
the
other
side
and
basically
we'll
go
to
the
destination.
That's
not
that
different.
Besides
the
fact
that
we
got
rid
from
a
lot
of
the
of
the
of
the
sidecar,
but
if
what
you
want
is
layer,
7,
that's
where
complex
stuff
happens,
layers
4
is
honestly
very,
very
simple,
operation
encryption,
maybe
some
you
know,
there's
not
going
to
be
any
problem
there,
but
on
layer,
7,
that's
where
we're
doing
stuff
that
are
a
little
bit
more
complicated.
You
do
not
want
to
share
that.
B
It
will
go
to
the
layer
four
and
then
we'll
go
to
the
service.
So
that's
very,
very
high
level
and
again
there
is
a
talk
on
it.
So
I'm
not
going
to
put
it
too
much,
but
what
we
did
here
is
basically
reduce
the
cost
drastically,
which
is
important
to
some
people
right
by
actually
going
into
one
proxy.
We
in
Solo
created
a
lot
of
Vlog,
and
one
of
those
blog
was
about
what
it's
doing
to
your
wallet.
So
we
see
a
multiple,
reducing
in
three
multiply
by
three.
B
Four
proxy
is
relatively
fast
with
one
a
layer,
7
proxy,
which
is
usually
where
it's
taking
more
time,
so
we
believe
it
will
be
faster
and
the
most
important
thing
is
the
operational
and
to
me
honestly,
this
is
the
most
important
and
I
think
that
they
will
talk
about.
You
know
a
lot
of
the
customers
who
today
is
actually
adopting
service
mesh.
They
sometimes
only
want
layer
7,
but
they
still
need
to
build
to
buy
to
basically
buy
into
this
like
sidecar
model,
and
it's
a
little
bit
too
expensive
and
too
complex.
B
It's
very
simple:
you
can
apply
a
mesh.
You
can
put
whatever
policy
you
want,
you
can
remove
the
mesh,
the
application
doesn't
even
know,
and
this
is
why
the
name
ambient
it's
really
really
transparent
to
the
application
and
I
think
this
is
free.
So
what
did
we
do
right?
Very
important,
zero
trust
right
I
mean
we
did
not.
We
can't
separatize
the
security,
that's
really
really
important,
and
so
did
that
solo
and
Google
did
a
Blog
on
this
on.
Why
we
believe
that
it's
actually
better,
but
we
are
reducing
the
cost.
B
B
That
I
think
I
really
recommend
you
to
to
go.
So
you
know
we
as
solo
put
it
in
our
product
already.
So
if
you
are
interested
in
consuming
it,
this
will
be
something
that
that
we
would
love
to
help
you
with,
and
this
is
what
basically
is
solo
doing
so
I
will
summarize,
if
I
said,
what's
special
about
solo,
because
I
think
that
there's
something
very,
very
unique
that
is
overlooked
this
ecosystem
we
are
here,
you
will
see
a
lot
of
projects.
B
We
are
contributing
to
sto
to
handle
and,
honestly
to
anything,
we
need
right
to
make
our
stuff
successful.
We
have
two
Toc
member
of
the
five
that
estio
has
two
from
Google
to
from
solo
and
one
for
my
VM,
and
we
are
going
to
pushing
the
boundary
with
mbn,
but
not
only
with
mbn.
If
you
will
look
at
solo,
you
will
see
that
we're
all
about
pushing
the
boundaries
and
innovation.
B
If
you
will
look
at
the
work
that
we
did
with
Washim,
you
know
most
of
the
stuff
that
people
will
talk
today
about
all
this
orchestration
of
service
mesh
is
something
that
we
talked
about
it
in
2018.
We
built
it.
We
made
it
wrong,
we
fixed
it,
so
we
right
now
feel
very,
very
comfortable
about
this
solution
that
we
have.
B
You
know
NBN,
we
talked
about
it,
ebpf
we're
a
big
believer
in
it
and
we
built
Bumblebee,
who
is
the
docker
Docker
experience
from
running
ebpf
models
and
the
another
point
which
is
extremely
important
is
that
we
have
a
lot
of
customers
and
I
think
this
is
very,
very
unique
in
our
infrastructure
in
our
ecosystem.
We
see
this,
we
run
it.
We
see
all
this
environment,
we
solve
those
problems,
so
we
can
help
anybody
else
and
well
about
education.
So
yesterday
we
had
an
event.
B
B
Took
me
not
a
lot
of
time
to
understand
that
I
need
to
put
my
ego
inside
and
basically
bring
everybody
to
the
front
of
this
stage
and
that's
exactly
what
I'm
doing
Sola
is
not
me.
So
it's
not
Christian
solo
is
not
lean,
not
Nick,
not
everybody!
Solo
is
everyone
working
in
Solo
and
you
don't
see
the
200
people
walking.
There
is
a
plus
player
and
there
are
a
lot
of
them
are
here.
So
please,
you
know,
learn
from
them,
talk
to
them
and
let
us
know
how
we
can
help.
Thank
you
so
much.