►
Description
London OpenShift Commons Gathering 2019 Lightning Talk NGINX
A
There
you
go
go
for
it.
Thank
you
very
much,
nice
to
see
everybody,
so
my
name
is
David
Luc
solution,
architect
within
internets,
I'm
sure
you've.
All
heard
of
nginx
we've
had
a
very
aggressive
in
a
good
way,
marketing
campaign
in
the
last
two
or
three
years,
so
I'm,
hoping
that
nearly
everyone
here
has
at
least
heard
of
us.
A
We
come
from
a
background
of
originally
a
web
server
in
the
open-source
community
way
back
ten
eleven
years
ago
and
since
then
that
evolved
to
be
use
cases
where
the
actual
request
flows
through
nginx
onto
somewhere
else,
rather
than
ending
it
nginx.
So
the
kind
of
web
server
thing
is
still
widely
deployed,
in
probably
a
million
plus
websites
over
the
world.
But
in
recent
years
we've
moved
more
into
the
space
where
actually
proxying
those
requests
through.
So
that's
given
rise
to
a
number
of
really
good
use
cases.
Api
gateway
is
a
huge
one
for
us.
A
Now
that's
a
very
popular
popular
with
customers,
we've
added
a
control
plane
layer
which
also
backs
that
up
so
yeah,
also
physical
load,
balancer
replacement,
some
of
the
big
guys.
I
won't
mention
their
names,
big
red
boxes
in
racks
that
do
application
delivery
control.
We
can
offer
a
software-only
alternative,
we're
fully
agnostic
to
any
cloud
provider,
any
technology
provider
we
play
nicely
with
with
the
Red
Hat
openshift.
We
also
play
night
nicely
with
bare
metal,
kubernetes
or
any
of
the
other
different
cloud
based
implementations.
A
So
that's
a
little
bit
of
background
on
us,
but
a
key
piece
of
value
that
we
provide
in
this
particular
environment,
which
we're
we're
to
talk
about
today
is,
of
course
the
router,
and
you
can
see
that
little
green
hexagons
in
the
middle
that
says
n+
engine
x+
is
the
commercial
version
of
our
products,
and
the
open
stores
are
still
alive
and
well
they're,
still
very
strong.
What
nginx
plus
gives
you
is
as
well
as
the
class-leading
technical
support
piece.
A
It
also
gives
you
a
number
of
enterprise,
great
features
that
enable
you
to
actually
deploy
nginx
in
an
enterprise
environment.
There
are
things
like
service
discovery,
high
availability
and
yeah,
a
number
of
other
features
over
and
above
that
to
give
you
yeah
what
enterprises
are
commonly
asking
for
over
the
top
of
of
the
open-source
version.
So
why
is
the
open
shift?
Route
are
important,
and
why
is
the
engine
X
plus
version
of
that
important?
It's
a
very,
very
simple
concept.
It's
your
one
way
in
to
any
open
shift
cluster.
A
So
if
you
think
about
that
from
a
practical
perspective,
let's
say
you've
got
like
in
this
example
of
four
node
cluster.
Okay,
that's
not
all
just
in
fact,
four
pods
there
doesn't
show
the
number
of
nodes,
but
let's
say
you've
got
I've
got
some
customers
who
say
to
me
we're
going
to
scale
to
a
50
node
cluster
in
the
next
12
months.
So
suddenly
the
router
is
now
handling
all
the
traffic.
Is
your
single
way
into
that
cluster,
which
is
going
to
and
we're
going
to
see
clusters
scale
in
the
next
few
years?
A
Two
very
large
volumes
of
nodes
which
affects
every
company.
That's
just
presented
and
everybody
in
this
room
as
we
see
that
start
to
happen
in
the
real
world,
so
performance
is
not
high.
Performance
is
something
which
everyone
mentions
and
you
hear
it
hundred
times
a
day.
I
even
saw
a
system
in
screw
fix
direct
for
a
toilet
that
was
high
performance.
So
it's
it's
become
marketing
speak,
but
what
does
it
actually
mean
for
us?
It's
actually
about
you've
got
this
open
shift.
A
Rooter,
it's
handling
thousands
of
SSL
terminations,
it's
doing
path
based
routing
host
based
routing,
it's
doing
service
discovery.
You
might
have
that
pod.
There
in
app
3
suddenly
decides
to
grow
to
8
pods
with
an
auto
scaling
event.
So
you've
got
a
lot
of
things
going
on
and
you've
got
massive
traffic
volumes
and
complex
operations
on
each
of
those
requests
coming
into
the
router
and
then
going
out
to
wherever
it's
destined
inside
your
ever-growing
cluster
according
to
your
business
need
and
scale.
So
it
is
performance.
A
It
was
originally
written
in
the
C
language
10
years
ago
that
the
first
version
came
out
simply
in
response
to
what
was
then
called
the
C
10
K
problem,
which
is
how
can
you
have
10000
concurrent
actions
on
a
single
instance
of
a
process
running
inside
Linux
and
actually
get
it
to
work
and
not
consume
a
million
threads
or
yeah
all
of
the
CPU
load
or
flood
the
network
ports
with
unnecessary
traffic?
So
that
was
the
core
philosophy
of
nginx
all
those
years
ago,
and
it
continues
to
this
day
it's
still
written
in
C.
A
So
that's
that's
the
cool
reason
and
that's
the
one
reason,
even
if
you're
not
considering
the
various
enterprise
great
features
here,
I
mean
we
do
have
a
feature
comparison
matrix
and
it
shows
you.
You
know
the
things
that
we
can
do
versus
the
default
H,
a
proxy
implementation
and
also
we
have
our
open
source
implementation.
So,
but
this
aside,
this
doesn't
actually
contain
I
had
the
one
piece
which
I
found
important,
which
is
yeah
the
the
performance
and
yeah.
A
This
is
the
one
area
in
the
whole
of
the
open
shift
ecosystem
where
high
performance
is
not
just
a
nice
buzzword.
It's
absolutely
critical,
so
that's!
Why
we're
here
today
and
that's
why
we've
got
we're
focusing
on
this
specific
piece
of
the
ecosystem.
You
can,
of
course,
use
nginx
in
in
other
areas.
You
can
use
it
as
a
sidecar
proxy.
Put
it
inside
every
pod
to
do
east-west,
SSL
mutual
TLS.
You
can
also
use
it
as
a
web
server
in
the
in
the
classic
deployment
that
it
was
originally
designed
for
all
those
years
ago.
A
You
can
also
even
use
it
outside
the
cluster
as
a
physical
application,
delivery
controller
replacement,
so
we're
right
through
the
ecosystem,
and
you
can
use
nginx
plus
in
all
of
those
scenarios,
but
the
one
specific
piece
of
value
just
for
OpenShift
is,
of
course
the
rotor
so
that
that's
the
bit
I
want
to
focus
on.
Today
we
have
yeah
we're
over
in
I,
can't
remember
which
room
it
is,
but
it's
the
one
on
the
left
when
you
come
in
through
the
front
door.