►
From YouTube: Understanding the new Kubernetes Gateway API vs Ingress
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
It's
great
to
have
everyone
on
this
webinar,
so
most
likely
we're
going
to
spend
more
time
talking
about
code
rather
than
talking
about
slides
and
something,
even
though
I
do
have
some
slides
as
a
conversation
started
today,
I'm
going
to
be
talking
about
a
gateway
again,
it's
a
new
thing.
We
just
released
a
initial
support
for
gateway,
api
and
cloning
grass
controller,
and
if
you
will
stick
around
until
the
end,
there's
one
one
extra
thing
that
you
probably
also
will
be
excited.
A
If
you
are
today
with
us
to
learn
about
the
gateway
api,
essentially,
today
we're
gonna
spending
more
time.
I
would
like
to
spend
more
time
answering
questions
rather
than
I
am
a
technical
person,
so
you
can
have
any
technical
questions
if
I
don't
know
because
sometimes
I
don't
know
many
things.
If
I
don't
know
something,
I
know
people
who
can
help
you
and
I'll
help
to
navigate
these
questions
afterwards.
A
Ingress
many
of
you
probably
already
somehow
familiar
if
you're
running
any
type
of
workloads
inside
your
kubernetes
cluster,
and
you
need
to
provide
external
access
to
this
to
this
service.
You
usually
may
be
using
ingress
or
maybe
using
load
balancer,
and
that's
why.
I
would
like
to
see
your
responses
in
the
poll
so
and
after
that
we're
going
to
be
talking
about
the
gateway
api.
So
what
is
this
gateway
api?
What
kind
of
problems
first
of
all
it
will
solve
for
for
kubernetes
and
for
users
of
kubernetes?
A
And
after
that,
the
rest
of
this
session?
We
will
spend
doing
demos
I
like
to
go
through
the
slides
as
soon
as
possible
as
quickly
as
possible,
because
I
know
many
of
you
are
here
eager
to
learn
some
technical
details
all
right.
So
in
terms
of
networking,
just
real,
quick,
we
deploy
our
applications
inside
kubernetes
kubernetes
cluster
might
consist
of
a
number
of
physical
nodes.
A
Those
nodes
can
be
virtual
machines
or
some
bare
metal
things
and
inside
your
applique
inside
this
commerce,
cluster
compressors
will
be
responsible
for
deploying
your
application
inside
the
pods.
These
pods
can
be
distributed
across
multiple
nodes,
based
on
the
replication
factor
and
so
far,
and
so
on
and
usually
in
order
to
get
access
inside
kubernetes
cluster,
you
define
a
service
with
either
the
type
of
cluster,
ip
or
node
port
cluster
ip
allow,
it's
kind
of
like
default
settings,
allows
you
to
get
some
ip
address.
That
would
be
addressable
inside
cobras
cluster.
A
In
order
to
get
access
to
outside
world.
To
you
know,
provide
the
application
access
to
the
to
to
your
applications,
we're
using
a
load
balancer
service,
so
a
load
balancer
service
usually
will
have
an
external
ip
address
external
to
kubernetes
cluster,
and
after
that,
this
load
balancer
will
be
somehow
either
about
to
the
service.
You
know
you
can
create
the
service
with
type
of
load,
balancer
and
you'll
get
external
id
address.
A
Now
the
as
you
grow
as
you
deploy
more
applications,
it
quickly
becomes
kind
of
like
a
challenge
to
figure
out
how
to
get
access
provide
access
to
those
those
applications
to
those
pods
from
outside
of
commerce
cluster.
A
If
you
have
many
services,
like
you,
either
kind
of
like
deploy
a
load
balancer
per
each
kind
of
like
application
functionality,
but
it
very
quickly
become
a
very
kind
of
wasteful
the
pattern,
because
in
this
case,
like
every
time,
you
need
to
get
external
ip
address
and
that's
why
we
kind
of
wants
to
centralize
it
through
some
sort
of
proxy.
So
this
proxy
is
essentially
what
ingress
does
it's
a
object
that
will
be
performing
routing?
A
Some
of
the
requests
that
come
into
this
system
will
go
into
now
this
ingress
and
based
on
certain
either
path
or
either
other
rules
like
maybe
headers
or
some
some
other
stuff.
It
will
be
routed
to
a
particular
surface.
A
The
ingress
object
that
looks
like
this.
With
the
with
this,
we
have
a
couple
things
here.
This
is
the
kind
called
ingress.
This
is
something
that
many
of
you
already
did
this,
so
this
ingress
will
be
served
by
some
sort
of
like
ingress
controller.
It
can
be
engine
x
and
gears
controller
or
be
called
congress
controller
or
some
something
else
and
based
on
certain
rules
either
it's
going
to
be
like
a
header
like
I
said,
host
or
particular
path.
A
We
will
route
this
request
to
particular
service
in
this
in
this
case
bill.
So
every
time
we'll
go
to
example.com
bills.
The
traffic
will
go
inside
this
this
service,
so
underlying
inside
inside
this
ingress.
There's
some
sort
of
like
a
proxy
server
sits
there,
but
you
don't
see
it
in
ingress.
So
that's
why
you
will
be
relying
on
ingress
controller
that
will
be
serving
a
particular
ingress.
A
I
will
show
you
how
you
can
define
this
and
there's
like
a
standard
ways
and
there's
some
of
the
legacy
ways
how
you
can
do
this
now
with
this,
this
proxy
server
actually
can
be,
can
be,
can
do
some
some
other
things
apart
from
being
just
a
simple
router
as
a
proxy
server,
you
will
be
able
to
do
some
of
the
advanced
things
like
defining
a
some
of
the
security
concepts.
Maybe
you
are
doing
ssl
termination
doing
some
sort
of
like
certificate
management.
A
Maybe
integration
with
some
other
components
like
assert
manager,
external
dns
and
so
far
and
so
on,
but
also
there
should
be
some
of
the
traffic
management
and
traffic
balances.
So
here's
the
problem
with
this,
so
those
stuff
usually
not
defined
in
not
usually
it
is
not
defined
in
ingress
in
the
standard
kubernetes
ingress.
This
stuff
is
defined
only
based
on
the
implementation,
so
you
need
to
go
into
particular
implementation
and
figure
it
out
how
you
define
all
these
rules
in
terms
of
enabling
plugins
and
in
terms
of
enabling
different
traffic
rules.
A
I've
been
doing
some
of
the
presentations
around
how
the
qualcomm
grass
controller
can
be
used
with
ingress,
so
basically
extending
ingress
using
the
standard
annotations.
However,
this
is
the
this
is
the
like
challenge
for
for
many
vendors
and
the
initiative
group
figure
out
the
some
of
the
limitations
of
ingress,
another
very
popular
limitation,
another
limitation
that
many
of
you
aware
that
ingress
is
solely
designed
to
be
serving
like
l7
traffic
http
traffic.
A
However,
some
of
the
services
that
you
will
be
serving
through
from
kubernetes
might
not
be
only
l7,
so
it
can
be
like
tcp
service,
udp
service.
So
in
this
case
again,
you
need
to
rely
on
some
implementations
that
your
you
know,
gateway
controller
like
or
like
a
kubernetes
controller
will
provide.
So
this
is
where
we're
going
into
this
is
where
we're
going
into
gateway
api.
A
So
I
will
bring
this
slide
one
more
time,
so
when
we
essentially
define
the
ingress
when
we
define
this
ingress,
the
communication
is
going
through
the
proxy,
but
proxy
is
not
outlined
in
any
ways
and
you
need
to
either
rely
on
whatever
vendor
will
provide
you
with
some
of
the
things.
So
there's
no
way
how
you
can
say
like
what
exactly
we're
running
there
so
with
gateway
api.
This
information
would
be
a
little
bit
more
explicit,
so
this
information
would
be
a
little
bit
more
outlined.
A
So
with
this
gateway
api,
we
start
getting
some
of
the
new
objects.
So,
first
of
all,
we
will
this
particular
scheme
of
like
objects,
so
those
things
that
I
will
be
showing
you
today
right
now
like
this
one
gateway,
class
gateway
route
and
particular
service.
You
already
have
a
service,
that's
that's
already
given,
however,
as
you
can
see,
those
new
objects
are
very
much
bind
to
kind
of
like
a
role
model
so
and
based
on
this
role
model,
we
can
have
a
more
granular.
A
I
don't
know
security
to
distinguish
different
roles
around
different
objects
and
provide
different
are
back
rules,
so
a
gateway
class.
This
is
something
new
that
comes
into
gateway
api
and
we're
going
to
see
how
it
works
in
the
real
world
so
get
to
a
class,
we'll
be
defining
what
kind
of
gateway
we'll
be
using.
So
it's
it's.
It's
a
it's
exactly
what
it
says.
A
class,
if
you
familiar
with
any
type
of
object-oriented
programming,
so
class,
is
essentially
a
template.
It's
something
that
will
be
defining
your
actual
instance.
A
Where
is
the
gateway,
it's
actual
instance
or
implementation?
If
you
also
came
from
a
programming
world,
so
essentially
with
this
with
gateway
that
cluster
operator
can
configure
to
you,
if
different
instances
of
this
get
gateway
can
be,
can
be
shared
across
multiple
developers
and
as
a
developer
here,
you're
just
defining
a
route
to
move
the
traffic
inside
your
application,
one
of
the
limitations
that
we
have
the
ingress
that
the
basically
management
of
the
traffic
was
handled
or
different
rules.
A
Rather,
you
need
to
either
rely
on
some
sort
of
extensions
or
you
need
to
rely
on
functionality
and
available
ingress,
but
in
ingress
it
was
limited
by
I
don't
know
path
and
host,
but
with
the
gateway
you
will
be
able
to
define
this
routing
thing
also
based
on
headers,
like
some
arbitrary
headers
I'll
show
you
in
a
demo
in
a
bit
and
after
that,
based
on
different
rules.
A
This
particular
thing
will
hit
a
particular
route.
This
this
actually
opens
up
like
a
very
interesting
concept
that
you
can
be.
It
can
be
done
through
the
gateway
apis.
It's
totally
fine,
still
using
the
ingress.
So,
as
far
as
I
can
tell,
the
gateway
api
will
supersede
ingress
only
from
a
features
perspective,
but
ingress
will
not
go
anywhere.
So
there's
no
conversation
around
duplicating
this
and
having
the
something
that
allows
you
to
configure
this
more
or
less
application,
specific
trout,
routing
logic,
it's
actually
very
nice
for
for
many
developers.
A
So
that's
that's
my
slides
just
I
don't
want
to
talk
about
slides
anymore,
so,
let's
jump
right
in
into
the
code.
Before
I
jump
right
in,
I
will
see
there's
a
lot
of
chat
is
going
on
in
in
in
a
chat.
So
the
first
question
supports
alb.
Is
it
application
load
balancer
that
is
in
aws?
A
I
need
a.
I
need
the
clarifications
for
this.
So
short
answer
is:
yes,
we
do
support
load
balancers
in
aws
into
more.
You
know
more
examples
how
kong
is
better
than
traffic.
I
love
this
type
of
question.
So
essentially
I
don't
believe
something
is
better
or
something
is
is
is
worse.
A
There
are
plenty
of
different
card
types
in
different
card
types.
It's
you
know
based
on
preference,
personal
preference
based
on
certain
characteristics.
You
know
you
can
buy
the
fast
sport
card
to
get
you
from.
You
know
also
from
point
a
to
point
b
bottom
line.
The
car
will
get
you
any
car
will
get
you
from
point
a
to
point
b,
but
you
can
get
this
with.
A
I
don't
know
honda,
civic,
there's
nothing
wrong
with
honda
civic,
it's
a
great
car
but
like
if
you
want
to
get
from
point
a
to
b
with
style,
adding
some
additional
features.
You
know
you
can
get
the
more
advanced
car
so
think
about
this
like
this.
So
if
some
of
the
tools
or
some
of
the
cars,
some
of
the
vehicles
will
get
you
from
point
a
to
point
b
and
you're
enjoying
the
ride.
A
That's
that's!
That's!
That's
all
the
best,
but
my
I'm
not
the
car
salesman
here.
So
I'm
here
not
selling
that
and
this
car
the
cone
would
be
better
than
traffic
or
in
nginx
or
contour
or
some
some
other.
A
A
A
All
right,
where
does
the
name
gateway
api
come
from,
especially
a
reasoning
for
the
word
api
for
the
name?
That's
easy,
so
this
very
easy
question
so
api.
In
this
particular
case
it
refers
to
a
kubernetes
api,
so
I
show
the
slide.
Let
me
see
if
I
still
have
it
a
little
quick,
essentially
when
I
was
showing
you
a
this
guy,
so
kubernetes
itself
as
a
system
as
a
technology,
it's
a
it
be
based
on
api
server.
Kubernetes
exposes
different
apis
to
operating
or
to
interacting
different
components.
A
Every
time
when
you
fire
up
coup
control
command
line
tool,
you
have
you
interacting
with
kubernetes
api.
So
for
many
people
gateway
api
might
sound
a
little
bit
kind
of
like
api
gateway
like
how
they
how
they
relate
it.
So
short
answer
is
naming
is
hard
in
in
this
world.
In
it
we
know
that
naming
and
the
caching
validations
is
probably
the
hardest
problems,
but
picking
up
good
name,
it's
also
a
problem
so
before
it
was
called
gateway.
Api.
So
api
again
refers
to
kubernetes
api
gateway
api
for
kubernetes.
A
Before
that
this
specification
was
called.
Actually
it
was
called
a
service
api,
but
it
was
creating
even
more
confusion.
So
that's
why
a
special
interest
group
and
the
working
group
decide
to
rename
it
so
so,
in
this
particular
case,
a
gateway
brings
more
into
the
table
from
perspective
of
a
piece
of
functionality
that
will
be
available
that
as
a
name,
it
will
be
showing
that
it
will
provide
more
features
than
than
simple,
simple
ingress.
A
So
I
hope
it
makes
sense.
But
again
we
can
talk
about
these
names
and
you
know
anytime,
okay,
cool,
so
I
will
get
back
to
the
questions
right
after
I
will
do
this,
so
I
have
a
kubernetes
cluster
up
and
running.
I
already
have
something
that's
installed
there.
So,
first
of
all,
I
will
start
with
the
quan
ingress
controller
and
I
installed
quank
ingress
controller
using
health
chart.
So,
if
I'll
do
something
like
helm,
install
and
I'll
just
do
this
one.
A
This
is
how
I
usually
install
using
help
and
if
I'll,
do
less
so
in
all
limb
spaces
in
the
in
quang,
we'll
see
that
quank
is
up
and
running
interesting
thing
here
that
I
need
to.
In
order
to
this
thing
to
work,
I
need
to
enable
this
secret
features
that
not
that
secret,
but
we
call
it
feature
gates,
and
we
need
to
tell
that
ingress
controller
need
to
start,
and
you
know
in
the
form
of
gateway
through
so
the
way
how
it
looks
like.
A
So,
if
I
go
into
my
kubernetes
cluster
and
if
I
yeah
my
internet
is
still
struggling.
If
I
go
to
quank-
and
I
will
see
this-
my
quan
congrats
controller
and
inside
my
con
congress,
if
I
do
something
like
gateway.
A
So
I
see
that
when
I
run
this
configuration
was
found
for
a
gated
feature.
So
let's
do
okay,
and
I
will
just
do
this.
One
and
I'll
just
do
gate,
and
in
this
case
it
will
start
this
ingress
controller.
With
support
of
this
now
I
mentioned,
I
mentioned
a
couple
things
that
this
gateway
api
introduces
two
or
three
or
four
few
new
apis.
A
One
of
this
new
api
would
be
called
a
gateway
class
that
allows
me
to
define
my
quang,
as
as
a
gateway
provider
in
in
the
concept
of
gateway
api.
This
thing,
this
the
gateway
class,
will
be
something
that
infrastructure
provider
will
define.
So
this
is
something
like
system
administrator
will
configure
and
after
that
we
define
this
class.
So
in
this
case,
if
I
go
in
my
convenience,
cluster.
A
And
I
say
not
this
one:
I
will
just
go
namespace
and
I'll
go
in
clone
namespace
and
after
that
I'll
show
gateway
class.
A
So
it
says,
like
there's
one
gateway
class
right
now
for
congress
controller
now
so
the
next
thing
that
I
will
define
a
gateway
itself
so
gateway
will
the
object
that
I
will
be
reference
and
link
with
my
routes,
so
the
request
will
comes
in
into
this
application.
A
A
Why
I
did
this
so
we
need
to
access
this
somehow
and
in
order
to
access
this
somehow
I
need
to
go
to
services
and
define
that
this
one
service
with
load
balancing.
So
this
is
the
external
ap
address.
So
like
everyone
right
now,
if
you
have
a
like
any
tools
that
curl
or
insomnia,
for
example,
we
can
hit
this
and
I'll
go
ahead,
say
new
request,
my
request:
just
do
one
create,
so
I
can
go
ahead
and
just
copy
this
guy
and
hit
this
url.
A
So
this
is
the
traditional
response
that
I'm
getting
from
traditional
response
that
I'm
getting
from
conch
means
that
the
quank
is
up
and
running.
There's
nothing
is
configured,
so
I
see
that
clones
is
working.
Also,
I
configured
my
I
configured.
My
server
would
be
available
on
dns
name.
So
if
I'll
just
do
something
like
dig
short.
A
I
will
be
able
to
hit
this,
so
this
is
our
guy,
so
we're
gonna
be
doing
with
this
proxy
me
yeah.
It
works
still
still
works
so
far,
so
good
now.
So
next
thing
is
that
since
I
already
configured
these
two
guys,
I
need
to
next
have
some
sort
of
deployment
and,
if
I
can
say
in
my
namespace
default
I'll,
have
a
two
two
services:
one
service
will
be
responsible
for
providing
me
back
to
future
quotes.
A
Another
service
will
be
responsible
for
providing
me
with
chuck
norris
quotes,
so
those
deployments
are
pretty
simple.
There's
nothing
really
special
about
this.
There's
a
there's,
a
one
image
that
again
you
can
grab
it
and
and
use
it
yourself,
and
if
you
provide
different
environment
variable,
it
will
start
either
back
to
future
or
chuck
norris
facts
service.
So,
with
today,
routing
examples
I'll
try
to
configure
configure
access
to
both
the
services.
Now
remember
those
services
in
our
default
namespace,
all
good.
A
So
next
thing
that's
how
we
can
configure
some
sort
of
you
know
the
access
to
this
one
we're
going
to
start
with.
First,
the
very
basic
example
we're
going
to
be
using
http
route.
This
is
a
another
new
type
of
api
that
comes
as
a
part
of
the
gateway
api,
so
you
learn
about
gateway
class.
You
learn
about
gateway,
and
now
it's
a
http
route.
A
The
way
how
we
can
like
link
those
together
map.
Those
together
is
by
defining
this
parent
reference.
I
will
bring
this
slide
a
little
quick,
so
you
have
a
idea.
What
are
we
talking
about
slide
with
this
relationship
again?
So
this
is
what
we
defined
as
a
class.
We
use
a
quantity
as
implementation
of
this
gateway.
Now
we're
defining
this
http
route
and
this
http
route
will
be
mapped.
You
see
this
many
to
many
map
to
gateway
and
we
will
see
how
this
will
help
us,
so
I'm
go
ahead
and
just
do
apply.
A
So
what
it
says
like
if
I
will
hit
say
this
server
no
routes,
but
if
I
add
header,
say
host
and
that's
going
to
be
back
back
to
future
proxy
me
and
now
what
we
see
here,
it's
the
response
that
came
from
back
to
future
service.
As
you
can
see
here,
I
have
my
application.
Add
some
extra
headers
that
are
reports
back
where
this
application,
where
this
response
came
from.
A
So
in
this
particular
case,
this
response
came
from
back
to
you
back
to
future
code
service
that
is
backed
by
this
is
my
service
and
this
service,
backed
by
this
deployment.
A
So
if
I
go
here-
and
this
is
my
back
to
future
service
now
with
this-
if
I
will
go
and
just
do
say,
let's
see
what
kind
of
hp
routes
we
have
so
that's
http
route
we
created.
A
Let
me
do
describe
http
route,
connect
it
to
kong
and
it
will
yeah
that's
the
part
of
specification,
that's
the
the
the
the
parent
reference
so
which
gateway
will
be
serving
traffic
from
from
this
from
this
particular
thing
and
we're
not
gonna
use
any
path
prefixes.
Here
it's
going
to
be
our
next
example.
So
next
example
is
using
path
so
now
like.
If
I
will
hit
this
particular
url.
A
I
will
be
let
me
put
this
here,
but
now,
if
I
will
just
do
back
to
future
I'll,
get
the
404
from
kong.
So
there's
you
know,
nothing
is,
is
there
so
what
I
will
do
is
go
ahead
and
just
do
apply.
A
In
and
now
we
see
here
in
my
http
rouse,
there's
another
http
routes
that
will
be
using
path.
So
if
I
will
go
ahead
here
and
do
this,
let
me
remove
the
header,
because
in
this
case
it
will
be
interpret
and
path
in
header
with
this,
and
I
don't
have
a
rule
for
that.
So
now
I
you
can
see
here,
I'm
still
getting
the
response
from
this
back
to
future.
My
response
still
goes
through
the
quant,
and
I
have
some
payload
is,
as
my
you
know,
output
here
now.
A
Next
thing
is
that
we
want
to
do
something
a
little
bit
more
interesting,
so
in
this
particular
case
we're
going
to
be
using
the
header-based
routing,
so
we're
going
to
be
using
a
header
to
define
what
kind
of
service
we
want
to
hit.
So
in
this
particular
case,
if
I
hit
this
cloneproxy.meservice
with
particular
header,
in
this
case
it's
going
to
be
arbitrary
header.
In
this
case
I
will
be
using
quote
header
with
the
roundhouse
kick
value.
In
this
case
the
traffic
will
go
to
chuck
norris
code
service.
A
So
let's
try
to
do
this
and
in
this
case
I
will
go
ahead
and
just
do
apply
and
let's
get
back,
we
see
my
new
route
check
header
here.
So
if
I'll
just
do
quote
and.
A
A
So
that's
that's
what
it
gives
you,
so
you
can
do
a
different
sorts
of
routes
and,
more
importantly,
the
similar
way
you
can
configure
you
can
configure
like
the
same
communication
to
to
to
your
services,
now
important
thing
to
note
that
if
you
want
to
see,
if
you
want
to
play
around
this
with
yourself,
so
you
can
go
to
this-
that's
my
repository
called
demo
scene
and
you
can
find
gateway
api
with
a
brief
explanation.
What
you
need
to
do
and
how
you
can
do
this.
A
That's
very
simple,
I'll!
Just
quick
pause
before
we
jump
to
another
part
which
is
it's
exciting.
It
was
not
advertised
and
for
those
of
you
who
stuck
around
for
for
a
while,
you
will.
Hopefully
you
will
enjoy
this
okay.
So
let
me
see
some
of
the
questions.
Can
it
be
used
in
the
multi-cluster
deployment?
A
So
that's
a
rather
very
complex
questions.
So
technically,
nothing
prevents
you
from
using
this
in
the
multi-cluster
environment,
however,
like
how
you
will
be
deploying
multiple
multi-cluster-
and
this
is
where
we're
going
to
be
talking
some
of
like
some
something
like
this
in
a
in
a
quick
second,
usually
bringing
stuff
like
a
service
mesh
that
allows
you
to
do
like
a
multi-cluster
communication,
and
some
of
the
things
that
were
recently
announced
in
in
the
service
mesh
world
is
gateway.
Class
and
gateway
are
cluster
scope,
resources
or
namescape
namespace
scope.
A
So
essentially
you
let
me
bring
the.
Let
me
bring
this
slide,
so
you
defined
the
classes
and
and
gateways
based
on
the
cluster
and,
more
importantly,
that's
the
cluster
operator
role
to
define
this
for
you
so
you're
as
a
infrastructure
provider
like
quang,
we
can
provide
you
with
the
gateway
class
and
gateway
implementation
and
the
its
responsibility
of
cluster
administrator
and
the
infrastructure
provider
to
to
configure
this.
Now
they
will
be,
can
be
accessed.
A
A
Those
are
different
namespaces,
but
they
can
share
same
gateway
that
defined
for
them.
They
can
bounce
these
resources
to
this
gateway
and
this
gateway
can
be
shared.
So
I
hope
I
answered
this
question
what's
missing
in
gateway
api
today
that
exists
in
the
ingers
controller.
So
it's
it's
very
good
question
anonymous,
attendee,
anonymous
name,
attendee
is
the
last
name,
so
essentially
nothing
much
so
the
the
the
right
now
the
problem
with
gito
api
and
the
wider
adoption
is
still
right
now
in
alpha.
A
If
you
go
and
see
this,
so
all
the
api,
all
api
versions,
let
me
remove
this
zoom
table.
You
see
here.
It
is
right
now
in
alpha
two,
so
it
is
not.
This
is
not
something
that's
like
very
stable.
Another
problem
is
that
this
crds.
A
A
That
is
somehow
you
can
find
in
in
in
this
website.
You
can
see:
there's
implementations,
there's
actually
quite
a
few
of
open
source
projects
and
the
community
members
who
working
on
this
we're
working
on
this
right
now
so
like
in
the
technical
sense,
if
we
will,
comparing
the
ingress
and
gateway
gateway
already,
has
a
quite
quite
the
features
that
allows
you
to
use
this
like
much
easier
than
than
ingress.
A
For
example,
the
the
header
based
this
like
a
header
based
routing,
which
is
something
that
you
can
do,
but
with
particular
implementation
like
you,
can
do
this
with
quant,
if
called
your
ingress
controller,
but
there's
nothing
in
ingress
that
allows
you
to
do
that:
you're,
extending
the
basic
functionality
of
ingress
and
instead
of
instantiating
header,
writing
plugin.
How
I
can
do
that
with
new
kind
of
kubernetes.
Now
I
can
do
this
with
new
kind
of
brands.
Is
that
right,
yeah,
that's
correct!
A
So
previously,
extension
of
ingress
happened
on
based
on
certain
functionality
that
vendor
will
provide
you
again
nginx,
quality
and
stuff
like
that,
and
you
need
to
have
external
plugins
to
do
this
configuration
with
the
with
gateway
api.
You
not
doing
this
anymore,
so
it's
going
to
be
part
of
the
standard.
A
How
I
can
use
gateway
api
if
I
have
a
multiple
cone
gateway
running
in
the
same
cluster,
is
kick
have
much
name
space
and
filter
in
and
grasp
very
good
question.
I
promise
you.
It
was
not
seated
very
good
question.
I
know
thanks
so
much
for
for
asking
this
here's
another
thing,
interesting
thing
that
will
set
up
answer
to
this
question,
so
you
can
find
this
very
interesting
thing.
When
we
define
the
gateway,
there
is
a
special
annotation
says,
gateway,
unmanaged,
true
or
false
right
now.
A
This
particular
object
expects
you
to
have
a
calling
deployed,
but
wouldn't
it
be
cool
if
you
can
spin
up
the
actual
gateway
actual
proxy
on
demand,
based
on
your
particular
request,
so
you
define
your
cluster
administrator,
define
you
a
role
in
your,
not
the
role,
the
gateway
class,
and
after
that
you
can,
as
you
need.
You
can
spin
up
like
multiple
different
proxies
for
particular
different
use
cases,
so
that's
where
the
gateway
unmanaged
false
will
come
into
play
where
there's
something
that
we're
working
right.
A
Now,
it's
right
now
in
the
in
the
poc
state
and
if
you're
interested
in
this
type
of
js,
I
recommend
you
to
the
follow
development
in
quant
congress,
controller
repository,
so
there's
a
pc
state
that
allows
to
kind
of
like
spin
up
the
gateways
as
as
you
please,
this
is
some
something
that
we're
working
on
a
gateway
operator
that
allows
to
start
different
instances
with
different,
maybe
even
different
configs
for
for
this
particular
use
case.
A
So
answer
to
your
question,
so
you
can
run
multiple
gateways
within
one
cluster
and
multiple
gateways
with
the
different
gateway
semantics.
Now
the
I
want
to
show
you
something
cool
before
we
before
before
we
go
just
the
last
one
is
very
quick:
does
gateway
support,
mtls?
Well,
it
does
support
a
tls.
You
can
expose
your
you
can
expose
your
you
can
provide
gps
for
hp
route.
It's
it's
there,
but
mtls
between
gateway
in
your
service.
A
I
think
it
still
it's
a
part
of
implementation
details.
So
kong
does
support
this.
You
can
do
this
with
quank,
like
mtls,
between
quank
and
underline
service.
You
can
do
that.
Okay.
Now,
as
I
already
said,
and
if
I
didn't
mention
this
more
explicit
gateway
api,
it's
a
standard
for
api,
so
meaning
that
similar
functionality
will
be
available
on
the
different
providers
and
there's
something
that
we're
working
on
in
in
say
another
side
of
the
clunk
here.
A
So
how
many
of
you
actually
know
that
we
are
working
on
service
mesh
which
is
kuma?
It's
it's
open,
source,
cncf
project
that
give
me
a
second
I'll,
bring
the
I'll
bring
the
example
of
this.
That's
something
that
we're
working
on.
That's
a
cncf
project.
That's
called
the
kuma
and
in
kuma
as
a
service,
mesh
service
mesh
was
designed
to
running
your
application
inside
your
your
in
your
data
center
and
providing
this
intra
application
connectivity
features.
A
However,
kuma
was
supporting
quang
for
a
while
as
a
external
gateway,
sometimes
many
users
they
don't
want
or
don't
need
to
use
quant
or
they
want
to
use
like
some
simple
features
like
routing
that
I
just
showed
you
without
you
know
doing
anything
like
super
fancy
like
enabling
some
plugins,
or
even
this
example,
with
these
mtls,
like
the
kuma,
allows
to
do
mtls
super
quickly
and
if
gateway
would
be
part
of
the
service
mesh.
You
can
enable
this
very
quickly.
So
now
the
the
the
kuma
also
supports
kubernetes
gateway
api.
A
So
let
me
quickly
spin
up
a
spin-up
akuma.
You
can
find
example
for
kuma
as
well
in
in.
A
A
Kuma
also
has
a
feature
flag
that
disable
this
by
by
default.
Now
we
will
enable
and
we'll
see
how
we
can
do
the
same
stuff
with
different
implementation.
That's
kind
of
like
one
more
thing
that
I
wanted
to
show
you,
so
it
will
grab
the
helm
chart
for
kuma
and
it
will
spin
up
a
let's
see.
A
Will
be
part
of
my
thing,
so
in
this
case
I
need
to
create
a
namespace
that
allows
me
to
do
that.
Allows
me
to
do
injection
of
okay.
So,
let's
see
if
it's,
if
it's
still
a
thing,
if
I'll
just
go
to
them
space
and
I'll,
just
do
a
comma
demo,
okay
and
I'll
just
do
describe,
and
it
still
allows
me
to
inject
sidecar
containers
so
like
inside
this
namespace,
I
will
deploy
my
applications
and
my
applications
will
become
part
of
this
service
mesh.
So
what
I
will
do
there
here.
A
A
Fresh
okay,
so
now
I
have
a
two
services
that
runs
inside
my
application.
One
service
is
back
to
future
service
and
another
one
chuck
norris
service
now.
So
this
is
how
you
can
you
know,
run
this
stuff
in
your
in
your
service
mesh
and
now
you
can
enable
like
mtls
between
the
services
and
and
all
these
things
like
super
super
cool
super
fast.
A
But
now
what
I
wanted
to
do
is
to
provide
a
gateway
as
right
now,
there's
no
gateway
data
for
proxies
data,
plane
proxies
that
will
allow
me
to
get
access
to
this
application.
So
for
this
I
need
to
do
is
to
create
a
gateway
for
kuma.
Now,
as
you
can
see
here,
I
also
define
in
gateway
class,
but
in
this
case
it's
going
to
be
a
gateway,
scum,
io,
slash
controller.
A
Compare
files
yeah,
so
you
can
see
they
are
very
much
alike
that
you're
going
to
be
using
the
same
thing.
The
the
differences
I
prefer
api
versions
should
be
go
up.
Api
version
should
so
this
is
going
to
be
using
nameco
and
it's
going
to
be
named
kuma
in
this
case.
Api
version
also
go
up
gateway,
annotations,
it's
not
using
any
annotations
here.
In
this
particular
case,
it's
going
to
be
named.
Kuma
we'll
put
this
here
for
namespace
demo,
and
now
it
also
will
create
a
gateway
for
kuma.
A
A
A
A
I
will
update,
I
will
update
my
dns,
so
we
can
access
to
this
host
in
a
more
fancy
way.
I
prefer
this.
A
So
there's
nothing
here,
but
gateway
is
up
and
running.
As
you
can
see,
his
response
came
in
from
the
from
the
gateway
now.
So
what
I
will
do
here,
I'm
not
gonna.
Do
this
one
by
one.
I
will
just
go
ahead
and
show
you
the
same
thing
that
we
did
here.
For
example,
four,
four
for
kuma
where's,
the
kuma
for
header-based
routing.
I
will
just
go
ahead
and
apply
this
with
good
control
apply.
A
Should
I
do
to
do
http
route
name
just
do
namespace
namespace
kuma
demo.
So
in
this
case,
if
I
will
hit
conc
me
dot
sh
with
this
particular
header,
I
should
be
able
to
get
access
to
my
chuck
noise.
A
I
see
that
it's
actually
comes
from
the
service
that
is
chuck
norris
and
bound
to
my
kuma
demo
nam
space
and
obviously
response
we
have,
and
the
cool
thing
is
that
I
also
see
this
in
my
ui,
because
this
gateway
is
also
part
of
service
mesh.
It's
part
of
the
of
the
of
the
puma,
so
you
can
see
here.
There's
data
plane
proxies
one
is
the
standard,
that's
my
application
that
runs
inside
here
and
my
gateway.
So
now
you
can
have
this
as
a
part
of
your
service
mesh
and
as
well.
A
So,
let's
see
if
the
host
also
works
here
with
with
here,
I
will
just
do
name
space-
yes
correctly
name
space
here
and
when
I
will
go
ahead
and
update
this
route.
A
And
response
waits
is
nothing
to
do
with
it:
okay,
marty,
okay
and
response
from
the
kuma
demo
and
the
grumps
from
this
back
to
future
service.
All
right,
so
that's
that's
it
for
now.
So,
let's
see,
if
there's
some
any
any
other
questions
now,
what
you
just
saw
is
that
inside
one
cluster
I
have
a
two.
A
Gateways
one
gateway
based
on
kong,
another
one
based
on
kuma
with
quank.
You
can
have
all
things
that
available
in
kong
with
kuma.
You
can
have
automatic
integration
with
the
service
mesh.
You
can
use
quang
as
gateway
for
your
kuma
and
but
you
cannot
use
yeah,
so
you
can
use
clone
gas
gateway
for
kuma.
You
can
use
gateway
for
kuma
using
gateway
api.
So
there's
plenty
of
things
that
you
can
do.
There's
like
a
variety
of
the
things.
Just
all
you
can.
A
It
would
be
nice
to
use
for
provisioning
data
plane,
but
still
have
control
plane
to
manage,
find
great
control
in
the
plug-ins.
So
I
think
the
there's
some
of
the
confusion
and
the
naming
here.
So
I
I
don't
quite
understand
the
question:
quan
kubernetes
gateway,
api
and
chrome
gateway
api.
So,
okay,
so
probably
the
question
is:
can
I
use
a
gateway,
api,
compress
gateway
api,
but
also
using
some
of
the
cooling
functionality
like
a
plugins?
A
A
For
example,
if
I
go
in
my
calling
plugin,
so
I
don't
have
any
like
a
crds,
so
you
still
can
apply.
You
still
can
apply
all
these
quantum
crds
like,
for
example,
the
cluster
plugin,
plugins
and
tcp
ingress,
and
all
these
kind
of
things
you
will
be
able
to
apply
in
the
same
way
you
should
put
the
annotation
inside
your
gateway,
or
rather
http
route.
A
So
if
you
go
to
kick
and
say
here,
so
you
will
be
able
to
put
somewhere
kind
of
like
annotation
same
way
as
you
do
like
hq
dot,
com,
slash,
plugins,
and
just
do
your
you
know
rate
limiting
yeah
this
type
of
stuff.
A
So
you
should
be
able
to
do
this
as
well,
but
some
of
the
functionality
that
available
in
quant
extending
functionality
needs
to
be
handled
explicitly
using
either.
You
know
plug-in
and
hp
ingress
and
all
those
kind
of
stuff
yeah,
but
quank
based
tcp
ingress
and
gdp
and
grass.
All
these
things
will
be
eventually
available
as
a
part
of
gateway
api.
All
right,
I
just
have
like
a
two
more
slides
that
I
want
to
people
pay
attention.
A
We
have
a
still
good
present
right
from
from
perspective
of
like
80
people
joined
61
people
stick
around.
I
appreciate
that
that
was
great.
I
also
appreciate
if
we
will
go
to
this
nice,
a
link
called
youtube.comconking.
A
This
is
the
youtube
channel
that
I
creating
a
lot
of
content.
Recently.
There
is
a
lot
of
work
around
the
tutorials
for
clunking
dress,
controller
using
helm
without
helm,
just
like
doing
things
from
scratch:
different
tutorials
for
different
tools.
Just
go
there.
You
know
click
subscribe,
enable
notification,
you'll
see
some
of
the
some
of
the
live
streams
that
we
do.
We
have
insomnia
team
doing
like
a
weekly
live
streams
that
they
can
break
down
the
process,
how
they
develop
insomnia
in
open
source.
A
So
I
would
really
appreciate
if
you
would
go
there,
plus
very
soon,
that
this
page
will
be
live
for
the
consignment
join
our
community.
We
have
a
great
community
forum.
Ask
your
questions
in
stack
overflow.
Most
importantly,
most
important.
I
don't
know
why.
I
didn't
mention
this
in
the
very
beginning.
Follow
me
on
twitter.
A
Follow
me
on
twitter,
it's
a
it's!
It's
it's!
My
twitter
was
conveniently
placed
on
every
slide
here.
Just
go.
Follow
me
on
twitter
dm
me.
If
you
have
any
questions
and
go
follow
for
all
announcements
and
the
konk
inc
in
twitter
as
well
and
with
this
my
name
is
victor
gamov
and,
as
always,
have
a
nice
day.