►
From YouTube: Beyond the API Gateway - Kong's Product Vision & Strategy: VP, Product, Reza Shafii, Kong Summit 22
Description
Beyond the API Gateway- Kong's Product Vision and Strategy
API Management as it was once known is dead. Kong's Vice President of Product Reza Shafii talks through the evolutions of monoliths to microservices and beyond, the birth of Kong, as well as the growth and future of Kong in his Kong Summit 2022 keynote.
Learn about the Kong suite of products, how businesses can better supporting their dev ops functions, catch a recap of a live demo of Kong Insomnia and more.
https://www.linkedin.com/in/rezashafii/
https://twitter.com/rezaloo
C
Right
yeah,
we
did
good
morning.
Everyone
hope
you
enjoyed
yesterday
welcome
to
day
two
we're
very
happy
to
see
all
of
you
here
today.
This
morning
we
have
many
interesting
sessions
in
store
for
you
today
and
we're
finishing
off
day
two
with.
What's
personally
my
favorite
part,
the
awards
ceremony,
so
please
make
sure
to
attend
this
afternoon.
At
the
end.
B
And
I
don't
want
to
delay
more
the
thing
that
you
all
collected
here
to
learn.
What's
going
to
be
a
future
in
Kong
and
for
this
I
would
like
to
invite
on
stage
the
man
the
myth,
the
legend
vice
president
of
the
product
that
called
Reza.
A
Kong
began
as
a
concept
in
2009
when
a
couple
of
friends
were
in
the
garage
and
started
building,
but
today
over
10
trillion
API
calls
per
month
the
most
popular
API
Gateway
on
GitHub,
with
more
than
67
000
Stars,
300
million
Kong,
Gateway
downloads
and
Counting.
We
couldn't
have
achieved
this
without
you.
You
contributed
thousands
of
lines
of
code
each
year,
you've
championed
Kong
across
the
world
and
you've
connected
with
each
other
across
more
than
80
meetup
groups
on
every
major
continent.
D
D
D
D
D
D
D
The
actual
architecture
of
the
applications
themselves
have
fundamentally
changed.
There
is
no
such
thing
as
monoliths
for
new
applications
anymore,
whereas
back
then
it
was
all
monoliths.
We
hadn't
heard
of
the
concept
of
microservices
and
importantly,
the
way
the
changes
to
these
applications
are
applied
day
to
day
has
changed.
When
we
were
in
a
system
where
everything
was
done
through
a
ticket-based
model,
changes
were
done.
Maybe
once
a
week,
not
every
piece
of
the
application
can
change
literally
dozens
of
times
a
week
culminating
to
hundreds
of
changes
potentially
a
week,
so
everything
has
changed.
D
D
This
startup
also
started
with
m,
has
something
to
do
with
monkeys
my
shape
and
when
they
looked
around,
they
couldn't
find
the
right
type
of
API
gate.
They
wanted
something
that
worked
well
with
containerization.
They
wanted
and
they
hear
a
Gateway
that
was
extremely
fast.
They
wanted
an
API
Gateway.
That
was
lightweight.
They
wanted
an
API
Gateway
that
had
a
fully
programmatic
interface,
so
young
and
foolish
you
stay
where
what
do
they
do?
D
D
D
They
brought
all
the
power
of
Kong
gateway
to
kubernetes
through
the
Kong
Ingress
controller,
and
now
at
Congress
controller
is
one
of
the
most
powerful
Ingress
controllers
out
there,
one
of
the
most
popular
Ingress
controllers
out
there
when
they
saw
that
the
mesh
space
was
too
complex.
It
wasn't
addressing
the
actual
needs
of
our
customers,
which
was
a
journey
from
virtualized
world
to
kubernetes
world,
so
needed
to
address
both
virtualized
and
kubernetes
environments.
D
They
created
humor
the
open
source
project
and
called
mesh
the
product
on
top
of
it
and
last
but
not
least
when
they
saw
that
the
dev
tools
that
API
developers
were
using
for
both
creating
apis
and
consuming
apis
were
too
heavyweight
or
too
bloated.
We're
getting
in
the
way
of
the
Developers
we
acquired
insomnia
three
years
ago,
the
most
popular
open
source,
API,
Dev
tool
and
we've
been
investing
in
it
ever
since
what
brings
This
Modern,
Cloud
native
API,
Tech
stack
together,
is
can't
connect.
D
We
introduced
how
connect
a
year
ago
and
can't
connect,
does
three
things
first
today
it
makes
the
actual
run
time
running
and
provisioning
of
con
gateways
much
simpler.
It
makes
their
day
two
operation
of
coin
gateways
much
simpler
over
time.
It's
going
to
do
that
to
all
these
other
peripheral
nervous
system
components
you
see
at
the
top.
D
Second,
con
Gateway
enables
platform
Builders
platform
Specialists
to
provide
a
self-service
interface
to
the
actual
application
teams,
while
providing
essential
governance
view
a
Federated,
Central
governance,
View.
Third,
it
provides
applications
like
portal,
like
analytics,
like
a
service
catalog
on
top
of
all
of
these
peripheral
systems
as
a
service.
D
D
D
D
The
dev
teams
want
to
be
focused
on
developing
business
Logic
on
innovating,
on
the
business
logic,
so
they
can
deliver
strategic
business
outcomes
for
your
organization,
and
apis
is
at
the
core
of
that.
We
all
know:
we've
heard
all
of
the
different
meetings
and
all
we've
talked
to
each
other
on
this.
We
understand
that.
D
So
how
do
we
get
those
deaf
teams
to
focus
on
those
strategic
business
outcomes
well
by
abstracting
the
runtime
layer
by
letting
them
deal
with
the
work
with
the
tools
they're
comfortable,
with
get
as
a
Version
Control
System,
whatever
ID
they
want
to
use?
You
know.
Saju
here
loves
emacs,
some
of
you
like
VI,
that's,
okay
and
insomnia
enables
those
developers
to
have
specialized
capabilities
around
consumption
of
those
apis
and
around
creation
of
those
apis
and
also
enables
them
to
then
generate
configuration
that
then
gets
mapped
to
the
right
runtime.
D
E
So
today,
I'm
going
to
wear
a
hat,
wear
a
hat
of
a
API
developer
where
I
work
for
Kang
Airway
I
wanted
to
develop
an
API
endpoints
that
receive
user
feedback
about
individual
flight
I'm
going
to
apply
API
Ops.
The
reset
just
talked
about
to
create
a
design
specs
in
insomnia
and
generate
a
declared
account
Gateway
configuration
file
through
single
command
line
experience
and
finally,
deploy
a
real
Gateway
instance
into
connect.
Runtime
group,
let
me
get
started.
F
E
E
D
E
So,
first
and
foremost,
insomnia
validates
against
the
industry,
recognized
open,
API
standard.
It
helps
API
developer
myself
in
this
case
to
identify
the
gaps
in
the
API
design
specs
and
follow
the
best
practice.
So
one
more
feedback
I
got
here
is
that
I
need
to
add
the
response
code
for
my
get
endpoints.
So
let
me
do
that
quickly.
D
E
But
insomnia
goes
beyond.
We
will
be
releasing
an
Innovative
feature
that
will
actually
allow
the
customer
to
generate
a
custom
view
the
best
to
tailor
to
their
own
business
context
and
business
needs
now
before
I
show
you
the
rule.
Let
me
finish
my
design,
so
I
want
to
create
another
endpoint
post
endpoint.
E
Now,
as
we
discussed
earlier,
we
want
to
develop
a
specific
rule
that
only
tailored
to
my
own
business
needs
for
the
demo
purpose.
I
create
a
very
simple
rule
that
called
every
single
tag
needs
to
have
a
description,
make
sure
I
set
a
function
as
choose
it.
So
if
I
go
back
to
the
design
specs
and
let's
see
what
happens
if
I
remove
the
description.
E
D
D
D
E
E
D
D
D
One
of
my
favorite
anecdotes
over
the
last
year
was
a
call
ahead
with
one
of
our
big
banking
customers,
and
they
had
just
done
a
full
stress
test
on
their
environment,
a
full
performance
load,
stress,
test
on
their
environment
and,
as
I
got
into
the
zoom
I
saw
there
was
a
bit
of
gloomy
faces
and
asked
what's
going
on,
they
said.
Well,
you
know
what
we
just
did
this
stress
test
and
everything
fell
apart.
D
D
Of
course,
the
coin
Ingress
control
has
a
new
version
that
now
supports
con
Gateway
3.0
we've
also
introduced
a
new
kubernetes
operator
that
works
with
calling
Gateway
3.0
that
automates
day
two
operations
for
current
Gateway,
such
as
upgrading
the
folks
who
were
here
yesterday
at
the
tail
end
for
the
closing
keynote
saw
that
in
action,
Kong
Gateway
3.0
introduces
really
exciting
capabilities
and
amplifies
the
automation.
There's
a
resiliency
and
the
traffic
management
capabilities.
I
was
just
talking
about
so
with
the
kubernetes
operator.
D
That's
one
thing
day:
two
operations
were
just
at
the
beginning
of
that,
with
backing
things
up
with
being
able
to
upgrade.
Backups
are
coming
up
next,
we
also
are
supporting
now
the
kubernetes
Gateway
apis
in
our
Ingress
controller.
This
is
a
big
deal
because
it
makes
all
of
the
Gateway
operations
on
kubernetes
be
standards,
and
we
spearheaded
the
effort
to
work
with
the
community
to
come
up
with
those
standards
as
part
of
a
special
interest
group.
D
D
Well,
the
capabilities
of
the
con
Gateway
plug-in
management
have
further
been
enhanced.
We
have
plug-in
ordering
now,
so
you
can
move
around
those
plugins
and
change
the
order
in
which
they
happen
and
execute.
We
have
non-native
support
for
websockets.
Your
plugins
can
actually
examine
the
websocket
frames
and
manage
them,
and
we
have
a
new
routing
domain
domain
specific
language.
D
What
that
does
is
not
only
gives
you
more
powerful
capabilities
around
specifying
the
routing
rules
in
terms
of
how
requests
come
in
where
they
should
be
directed.
It's
so
flexible
and
cool
that
the
engineering
team
internally
called
it
air
traffic
control,
but
also
it
improves
the
performance
capabilities
of
the
con
Gateway
when
running
in
hybrid
mode.
D
D
D
D
D
So
what
you
see
in
this
slide
on
those
circles
is
the
10
practices
that
cyber
Security
Experts
say.
We
need
to
follow
in
order
to
make
sure
our
apis
are
secure,
and
the
con
Gateway
enables
you
to
do
all
of
this.
The
3.0
version
of
the
con
Gateway
further
amplifies
all
those
capabilities
around
authorization.
Our
open
policy
agent
support
our
encryption
capabilities,
our
role-based
access
control,
all
of
that
was
around
before
3.0
with
3.0.
Our
secret
manager
now
enables
you
to
manage
the
tokens
and
any
other
property
of
your
API.
D
That
must
be
sensitive
with
hashical
Vault,
with
gcp
secret
manager
and
with
the
AWS
secret
manager.
Couple
that,
with
the
new
fips
140-2
compliance,
the
security
capabilities
of
The
Conch
3.0
are
even
further
there
to
help.
You
make
sure
what
happened
to
Optus
doesn't
happen
to
you
all
right.
So
we've
heard
about
Khan
Gateway.
We
heard
about
API
Ops,
we
heard
that
API
management
is
dead
and
a
new
stack
and
a
new
process
has
taken
over.
G
D
So
Davis
I
remember
when
we
started
our
discussions
about
two
years
ago.
It
was
yes.
G
G
G
We
had
very
less
Insight
in
the
platform
we
had
a
hard
time
building
capabilities
on
top
of
the
platform
and
we're
really
struggling
with
it
right,
because
the
platform
is
the
foundation
that
helps
the
engineer,
build
products
to
meet
the
customer
in
the
store
in
the
app
everywhere
around
2019
20.
That's
where
it
all
starts
happening.
We
started
looking
into
Kong,
for
example,
helps
us
power
all
the
capabilities
that
we
needed
for
the
engineers
and
that's
when
it
took
off.
G
We're
running
around
100
plus
teams,
150
we
have
more
than
300
apis
and
like
10,
000
million
API
requests
going
through
and
with
the
request
in
Ikea.
If
you
think
about
it,
you
do
you
don't
do
a
lot
of
requests
until
you
go
to
a
checkout
right,
so
you
find
a
product.
You
put
it
in
the
cart,
go
to
the
checkout
and
you're
done,
find.
D
G
So
if
we're
gonna
meet
new
markets
and
new
business
opportunities,
so
let's
say
we
have
a
start
with
the
store
on
ikea.com.
We
have
the
app
what's
next
right,
that's
about
having
apis,
not
but
not
just
apis.
They
need
to
be
reusable
right
exactly
as
our
flat
back
concept.
We
create
same
products
from
the
same
and
for
the
future.
We
see
that
a
lot
focus
on
developer
experience,
because
that's
the
key
to
help
us
scale.
D
D
D
Sometimes
you
need
to
do
a
quick
data
filtering
on
one
of
the
backend
apis.
Well,
you
can't
go
put
that
in
the
backlog
of
the
application
team
because
it
might
take
longer
than
you
need
a
presentation
layer.
The
uis
might
have
different
needs.
They
need
to
go
at
a
faster
pace
and
again
you
might
need
to
make
minor
modifications
to
the
backend
apis
in
a
rapid
way.
D
But
yet
it
is
true
that
one
should
not
put
data
transformations
in
the
Gateway.
We
believe
that
so
when
we
looked
at
what
we
could
do
to
help,
what
we
realized
was
Kong.
The
runtime
provides
a
couple
of
capabilities
around
Dynamic
configuration
automation,
just
even
the
plug-in
model,
and
the
fact
that
everything
runs
can
run
at
cognitive
speed.
That
makes
it
ideal
for
actually
driving
transformations,
but
at
the
same
time
Khan
doesn't
have
to
act
as
the
Gateway.
Only
crank
can
act
as
a
data
transformation
engine,
so
you
can
actually
separate
the
concerns.
D
D
Kayak
is
cloud
native
data
Transformations
at
play,
Let
Me
Explain
what
it
is
and
why
we
think
it's
Cloud
native
so
think
most
people
here
know
about
the
max
rasm
X
wasn't
announced
last
year.
We
just
made
it
Tech
preview
this
year
and
what
was
MX
does
it
enables
you
to
build
traffic
management
logic
or
con
Gateway,
executable
logic
in
any
of
the
supported
languages,
JavaScript,
golang,
rust,
C,
plus,
plus,
and
so
expands
a
set
of
developers?
You
can
Target
for
creating
Kong
customizable
capabilities,
but
does
so
at
cognitive
speed.
Why?
D
Because
that
code
gets
compiled
into
Azam
and
runs
natively
on
the
cone
Gateway?
This
is
going
to
be
really
powerful
stuff
for
con
Gateway
in
general,
but
when
we
saw
that
we
thought
about
that
data
transformation
issue,
we
said
okay
well,
what?
If
we
could
marry
the
two
and
that's
what
kayak
does
with
kayak?
We
created
a
new
DSL
called
k-script
is
made
to
make
data
Transformations
very
easy
to
create
to
build,
in
fact,
there's
a
demo
later
today
at
four
o'clock.
D
D
So
kayak
enables
you
to
create
those
Transformations
and
run
among
runtimes
and
calling
native
speed,
and
this
way
you'll
be
able
to
have
con
gateways
doing
what
they're
supposed
to
do
pure
gateways
and
con
kayaks
that
are
running
your
Transformations
on
the
site.
We're
very
excited
about
this.
It
should
be
coming
out
in
Tech
preview
over
time,
I'm
hoping
that
at
next
year's
Summit
I'll
be
able
to
announce
much
more
progress
on
this
and
get
make
sure
it's
in
your
hands
for
deployment.
So,
looking
forward
to
that.
D
D
Couple
of
hands
I
like
to
swim
once
in
a
while
myself
and
I
have
to
go
to
a
shared
swimming
pool
olympic
sized
swimming
pool
ideally,
would
be
good
and
I
get
my
own
streamline
if
it's
not
busy
and
see
I
have
to
tell
you
I'm,
always
a
little
bit
uncomfortable
because
I'm,
not
a
good
swimmer
and
so
I'm
worried
about
the
other
shimmers
in
their
swim.
Lanes,
and
then
you
know
if
there
are
kids
in
the
pool
gets
a
fee.
Sometimes
you
know
so.
I
would
way
rather
have
my
own
swimming
pool.
D
D
So
how
do
we
have
our
cake?
I?
Need
it
too?
Well
with
comconnect.
We
believe
you
can
do
that.
Remember
I
said
can't
connect
us
three
things.
Our
connect
enables
you
to
run
your
data
planes
in
your
own
environment,
on
any
Cloud
that
is
supported
today
on
your
own
data
centers
in
any
Cloud
adjust
the
data
planes,
which
is
have
the
lowest
amount
of
operational
complexity.
D
The
second
thing
that
it
does
is
that
it
allows
you
to
create
the
swim
lengths,
but
in
a
way
that
looks
like
they're
just
dedicated
swimming
pools,
that's
what
we
call
runtime
groups.
Each
runtime
group
is
a
dedicated
con
control
plane
to
Any
Given
tenant
that
you
might
have
for
your
platform
and
that
keeps
configuration
scoping
completely
isolated.
That
gives
you
the
ability
to
do
our
back
on
the
whole
thing
so
that
the
right
people
only
see
the
right
stuff.
D
So
that
gives
you
that
governance
model-
and
the
third
thing
remember,
is
that
enables
you
to
create
a
central
catalog
of
all
your
apis
through
something
we
call
service
Hub.
It
enables
you
to
take
those
services
and
productize
it
to
make
sure
they're
searchable
and
discoverable,
and
it's
got
full
analytics
operational
today,
but
business
analytics
coming
up
and
I'll
talk
more
about
that
very
soon
and
one
last
thing:
every
single
runtime
group
is
like
a
control
plane,
remember
so,
if
you're
doing
API
Ops
using
deck,
for
example,
or
if
you're
using
the
admin
API.
D
D
So
without
further
ado,
I
want
to
bring
on
stage
Frank
van
falkenberg,
who
is
a
business
integration
architect
at
Rabobank.
We
started
working
with
rubble
Bank
six
months
ago,
eight
months
ago,
and
the
progress
they've
had
in
enabling
connects
and
in
actually
implementing
a
platform
program
with
it.
Leveraging
the
runtime
groups
and
teams
concept
has
just
been
phenomenal.
H
Thanks
guys,
integration
is
one
of
the
three
largest
banks
in
the
Netherlands,
but
based
on
our
philosophies,
we
have
been
approaching
API
management
so
much
different
than
most
other
large
companies.
We
have
some
600
devops
teams,
which
we
call
squats.
Every
squad
has
their
own
distinct
responsibility,
typically
one
or
more
business
applications,
and
every
squat
is
ideally
front
to
back
responsible
for
these
applications
with
the
main
goal,
to
make
them
autonomous
and
as
little
dependent
on
other
squats
as
possible.
In
total.
H
We're
talking
about
some
2400
apis
at
the
moment
and
it's
growing
rubberbank
has
a
multi-cloud
strategy
where
every
squat
or
even
every
application
gets
their
own
separated
environment
in
the
cloud
and
consumption
of
apis,
very
importantly,
is
done
over
public
networks
no
longer
over
private
networks.
Now
making
scores
respondable
for
building
and
maintaining
their
own
apis
is
one
thing.
H
Making
them
responsible
for
maintaining
an
API
platform
is
another
thing
that
requires
specialist
knowledge,
and
that
is
why
Arab
Bank
offers
the
squats
a
central
API
management
platform,
a
shared
platform
which
is
compelling
and
easy
to
use
for
squats,
because
it
saves
them
the
trouble
of
having
to
manage
an
API
platform
themselves.
But,
most
importantly,
it
is
a
self-service
platform
and
we've
been
doing
API
management.
This
way
for
the
last
four
to
five
years
already,
but
our
current
platform
is
end
of
life
and
no
longer
living
up
to
the
requirements
we
have.
H
So
we
need
to
replace
that
platform
and
a
challenge
with
selecting
a
replacement
platform
has
been
that
not
a
single
API
management
vendor
fully
supported
our
single
platform
cell
service
operating
model,
at
least
not
out
of
the
box.
So
why
did
we
eventually
choose
com
as
the
basis
for
a
new
API
management
platform?
Couple
of
reasons?
First,
the
primary
focus
of
Congress
API
management,
which
is
important
for
us.
H
Secondly,
we
align
on
Visions
Rabobank
wants
a
SAS
based
management
plan
with
distributed
gateways
in
the
various
environments
we
have
following
the
application
workloads
both
believe
in
lightweight
Integrations
logic
should
be
in
applications,
multi
and
apis,
and
both
Guang
and
Rabobank
believe
in
automating
the
crap
out
of
everything.
Thirdly,
Kong
delivers
lightweight
run
times,
which
facilitates
our
operating
model,
where
every
team,
or
even
every
application,
gets
their
own
isolate
Gateway.
H
And
fourthly,
but
most
importantly,
we
have
found
going
to
be
very
willing
to
think
along
with
us
and
even
involve
product
management
to
facilitate
the
requirements
we
have
on
our
API
management
platform.
We
will
be
using
connect
as
the
central
management
plane
from
Kung,
and
that
is
where
we
Define
the
600
squats
in
separate
runtime
groups.
Now
this
concept
of
using
isolated,
lightweight
gateways
per
squat
or
even
per
application
supports
a
security
requirement.
Rabobank
has
to
segregate
workloads
from
different
squads
or
applications,
and
it
does
it
very
nicely.
H
And
finally,
we
have
a
central
integration
catalog
in
which
we
publish
all
the
reusable
services,
not
only
apis
on
the
central
API
management
platform,
but
also,
for
example,
Eventing
topics
on
our
Kafka
platform.
At
this
moment,
our
team
are
very
busy
implementing
this
new
API
management
platform
and
they
are
doing
this
in
close
cooperation
with
com,
so
in
next
year's
gun
summits,
I
expect
to
be
here
physically
with
my
colleagues
to
present
you
a
success
story
of
how
we've
implemented
exactly
this
platform.
That
I've
been
talking
about
back
to
you,
Raisa.
D
D
D
What
Frank
was
talking
about
is
something
we
call
Federated
governance
at
scale
and
something
we're
seeing
very
often
out
there
right
platform
teams
want
to
enable
developer
teams
to
go
fast.
They
want
them
to
be
able
to
innovate
that
they
want
them
to
do
such
so
with
some
Guardians.
They
want
them
to
do
it
in
a
way
that
builds
consumable
apis.
They
want
them
in
a
way
that
reuses
existing
API
assets
they
have,
and
they
want
to
do
it
in
a
way.
I
So
when
I
log
into
connect
for
the
first
time,
I
will
land
on
runtime
manager,
runtime
manager
is
a
great
way
to
logically
separate
out
environments,
business
units
Etc.
These
are
the
swim
lanes
that
you
referred
earlier.
During
your
talk
for
my
particular
organization,
I
have
chosen
to
use
environments
as
my
separation
units.
That's
why
you
will
see
production
and
staging
on
my
runtime
manager,
I'm
using
a
blue
green
deployment
model,
which
is
why
we
have
production
blue
production,
green,
staging
blue
and
staging
green.
Let's
take
a
look
at
production
blue
here.
I
D
D
I
So
within
my
Google
kubernetes
engine
I
can
see
that
all
of
my
runtime
groups
actually
map
to
kubernetes
clusters,
and
this
is
the
same
replication
of
what
I
see
inside
Kong
manager-
oh
sorry,
inside
runtime
manager.
So
when
I
go
into
one
such
cluster
or
my
production
blue.
To
continue
my
example,
you
will
see
that
all
of
my
runtime
instances
actually
map
to
parts.
I
What
is
interesting
here
is
that
not
just
can
we
deploy
this
on
GCB
we
can
actually
Deploy
on
any
cloud
provider
of
your
choice.
We
have
AWS
Azure,
you
name
it.
We
can
also
deploy
it
on
any
infrastructure
or
any
environment
that
you
want,
including
on-prem,
which
is
what
makes
Kong
the
most
flexible
deployment
options.
That's
out
there
I.
D
I
D
I
Is
a
scary
idea
for
sure,
which
is
why
connect
provides
you
a
very
fine-grained
permission
systems?
You
can
take
advantage
of
our
out
of
the
box
predefined
teams
to
give
access
to
your
users,
or
you
could
create
your
own
custom
teams
that
have
a
granular
set
of
permissions
that
is
based
on
Services
runtime
groups,
or
even
regions
like
North,
America
or
Europe.
So
that
way
you
can
control
who
has
access
to
what
great.
I
I
A
very
good
question
now,
even
though
P99
latency
means
that
99
of
my
traffic
is
seeing
less
latency
than
this,
but
497
is
definitely
high.
So,
let's
figure
out
if
we
can
identify
where
this
latency
is
actually
coming
from
I'm
going
to
go
into
my
plugins
I'm,
going
to
give
you
a
sneak
peek
into
my
plugin
hub.
I
Have
an
open,
Telemetry
plugin,
which
is
a
3.0
compatible
plugin.
The
latest
Gateway
that
came
out
right,
I'm,
going
to
enable
that
plugin
I
have
actually
configured
my
plugin
to
work
with
datadog
so
that
I
can
figure
out.
Why
and
where
we
are
seeing
this
latency.
I
F
I
The
time
taken
by
the
con
Gateway
itself
in
the
entire
in
the
entire
transaction,
that's.
I
Is
correct
so
when
I
come
back
here
all
of
these
instances
or
all
of
my
run
times
actually
roll
up
to
my
service,
Hub
and
I
can
see
that
in
the
service
Hub
itself,
but
before
we
go
there.
I
want
to
take
you
to
another
thing.
That's
on
my
screen.
You
can
see
here.
I
have
North
America
and
Europe.
I
Now
what
this
means
is
connect
is
designed
to
support
data
residency
requirements
for
various
regions,
for
my
example
I'm
only
using
North,
America
and
Europe,
but
what
this
means
is
any
data,
any
services
or
runtime
groups
that
I
Define
in
North
America
are
specific
to
North
America
and
anything
that
I
have
in
Europe
will
be
specific
to
Europe.
Yes,.
I
D
D
So
connect
has
come
a
long
way
in
the
last
year.
We're
able
to
deploy
I
think
it
was
300
updates
a
month
to
connect
at
this
point,
and
new
features
have
been
coming
in
it's
one
of
the
things
I
love
about
connect.
Is
that
we're
able
to
elevate
the
capabilities
real
time
with
all
of
you
and
give
you
that
benefit
no
more
waiting
for
the
next
release.
It's
just
going
to
be
there
and
we
have
much
more
capabilities
right
around
the
corner.
D
We're
going
to
have
on
the
portal
side
we're
creating
API
abilities
for
the
portal
itself
on
the
analytics
side,
we're
creating
the
ability
to
have
business
metrics,
as
I
mentioned,
you'll,
be
able
to
query
the
number
of
services
that
might
have
not
the
right
security
posture.
From
your
perspective,
we're
also
creating
custom
dashboards
for
you
to
be
able
to
create,
on
the
platform
side
we're
further,
enabling
things
with
things
like
audit
trails
and
there's
going
to
be
a
capability
to
enable
self-service
for
you
to
update
your
own
custom
plugins.
D
D
B
Thank
you
Raza,
and
thank
you
for
this
insightful
keynote
when,
when
the
keynote
was
on,
we
were
talking
in
the
great
room
with
Dali
and
Daly
told
me
that
she
is
actually
a
professional
swimmer.
She.
B
Now,
let's
get
serious,
let's
get
to
back
to
business!
So
next
speaker,
it's
gonna
be
sharing
some
of
the
real
world
examples
from
financial
industry
and
the
banking
sector.
It's
going
to
be
very
eye-opening.
F
Good
morning,
good
morning,
hello,
I
think
the
challenge
is
follow
that
how
do
we
follow
follow
something
like
that
as
we
bring
you
right
down,
you
know
you've
you've
seen
the
level
of
detail.
You've
seen
the
Fantastic
things
that
are
happening
with
with
Kong,
but
we
want
to
talk
about
why?
F
Why
HSBC?
Why
have
they
invited
a
large
global
bank
to
talk
about
Kong
about
these
capabilities
and
why
it
makes
such
a
difference
to
us?
So,
first
of
all,
we
won't
take
you
down
to
that
level
of
detail.
We
want
to
try
and
change
the
way
that
we
talk
about
our
technology
choices
and
the
way
that
we
talk
about
our
Partnerships
and
to
do
that,
we
need
to
take
you
on
a
bit
of
a
journey.
F
We
need
to
take
you
on
a
transformation
Journey
not
of
just
of
apis
but
of
people
of
culture
of
Technology,
so
who
is
HSBC
so
are
our
global
banking
organization
up
there
in
the
tier
one
level
of
banks?
That
means,
if
something
fails
in
the
global
banking
system,
somebody
goes
to
jail.
I
never
realized
that
until
I
joined
a
bank,
but
but
if
you
fail
somebody's
responsible
for
that
failure
and
the
and
the
penalty
is
quite
severe,
you
talk
about
resilience
yeah.
F
If
somebody,
if
you
know
something
in
Kong,
fails
which
it
doesn't
by
the
way
you
know
if
something
in
Kong
fails
or
something
in
any
other
vendor
product
fails
and
that
results
in
a
failure
in
our
service.
It's
had
quite
serious
implications.
Yeah
somebody
could
go
to
jail
if
our
service
is
not
resilient,
somebody
could
go
to
jail
if
our
apis
are
not
resilient.
If
our
perimeter
is
not
resilient.
F
F
That's
why
we're
here
we're
here
to
talk
about
the
complexity
of
our
organization
and
the
cultural
changes
so
enough
from
me.
First
of
all,
thank
many
thanks
to
Reza
many
thanks
for
calling
for
inviting
us.
We
wanted
to
try
and
do
a
little
bit
on
the
fly
from
what
we
saw
from
the
keynote
this
morning.
We're
a
Gateway
we're
a
2.x,
Gateway
Enterprise
Gateway
customer
we're
in
year,
one
of
our
multiple
year,
Journey
with
Kong,
and
a
partnership
that
I'm
so
pleased
about.
F
J
J
F
So
we
talked
about,
we
talked
about
the
yeah.
Why
we
are
here-
and
you
know
why
why
have
Khan
chose
an
HSBC
to
talk
today.
You
know,
because
we're
certainly
not
going
to
talk
about
the
technical
level
of
detail
that
some
of
you
want
to
hear
and
you've
heard
some
of
that
this
morning
and
you'll
hear
some
more
today
again,
we
want
to
refresh
you
on
the
the
culture
and
Technology
Journeys
that
we
have
to
go
on
I.
Think
to
to
illustrate
that
I
think
Jen
you're
going
to
talk
about
our
backstory
yeah
yeah.
J
J
J
There
is
a
CEO
there's,
a
CIO,
and
all
of
those
people
have
decision-making
rights
over
what
happens
in
their
Tech
organization,
because
they
are
the
risk
holders
they're
the
people
that
go
to
jail
if
it
goes
wrong,
and
so
for
somebody
like
Gareth-
and
someone
like
me
who
sit
in
global
roles,
trying
to
affect
change
is
really
freaking
hard,
really
hard
and
that's
all
because
of
our
complexity
and
just
how
big
and
unwieldy
we
are.
Thank.
F
You
I
think
to
illustrate
that
we
are
paid
by
one
of
the
same
boxes
on
this
screen.
You
know
they
pay
our
wages,
but
this
is
probably
physically
as
close
as
Jen
and
I
ever
get
to
each
other.
Ever
we
live
in
the
same
country.
We
don't
live
in
the
same
city
yep,
but
nothing
brings
us
together.
Our
technology
strategies
take
us
further
apart
and
further
apart.
I
made
decisions
over
here,
no
idea
what
Jen's
doing
over
there,
probably.
F
Nice
segue
into
common
goals
and
outcomes
behind
every
technology
strategy
and
business
strategy
is
a
set
of
values
and
principles.
I
won't
drains
them
up,
but
from
our
chief
executive
officer
from
our
chairman
right
down
to
our
Branch
staff
right
down
to
our
call
center
operators
to
us
as
VPS
driving
technology
across
our
organization,
we
try
and
pivot
around
people
cultural
values,
yeah.
You
can
see
one
on
there,
which
says
you
know
we
succeed
together.
F
F
That's
something
that
we're
very
good
at
doing
is
owning
a
problem
taking
it
through
to
the
end,
but
if
that
problem
is
owned
by
Jen,
she
probably
won't
ask
me
what
my
problems
are
and
I
won't
ask
Jim
what
her
problems
are
we,
but
we
get
it
done.
I
think
is
the
most
important
value
that
we
have
in
our
organization.
F
F
So
we
have
40
000
people
working
in
technology,
40
000
people.
What
are
they
doing?
They
can't
they're,
not
coding
they're,
not
developing
they're,
not
creating.
What
are
those
40
000
people
doing
so
you
can
just
see
again
I'm
sorry
to
labor
the
point
some
of
the
challenges
we
have,
so
it
helps
to
have
a
values
driven
organizational
purpose.
It
also
helps
to
have
a
technology
strategy.
Everyone
here
has
talked
about
technology
strategies
we
heard
yesterday
we
heard
from
Ikea
this
morning
we
heard
from
Rabobank
as
well.
We've
all
got
our
own
technology
strategies.
F
J
J
We
have
a
strategy
to
digitize
the
organization.
Some
of
you
will
be
working
at
startups
where
you
started
as
a
digital
organization.
Some
of
you
are
transforming
your
organizations
we're
transforming.
We
have
to
transform
at
scale.
You
saw
how
complex
it
was.
We
could
keep
talking
about
that
I.
Think
we've
probably
made
the
point.
Thank.
F
You
I'll
keep
making
that
point.
Yeah
yeah.
J
So,
for
us
this
is
all
around
creating
and
delivering
quickly
easily
getting
things
in
front
of
customers,
because
as
a
as
an
organization
who
are
all
about
servicing
customers,
whether
that
is
somebody
taking
some
money
out
of
an
ATM
or
whether
that
is
a
very
high
net
worth
individual
or
a
complex
corporate
customer,
we're
all
about
getting
Solutions
in
front
of
customers
the
longer
it
takes
us
frankly,
the
less
profit
we
make
from
them.
J
We
talk
about
partnering
with
technology,
innovators.
Historically,
organizations
like
ours
have
worked
with
the
ibms
of
the
world.
The
Oracles
of
the
world
we've
not
really
worked
with
the
smaller
organizations.
Our
internal
processes
make
it
difficult
to
do
that
sometimes,
but
but
our
teams,
particularly
so
we
we
sit
really
in
the
the
eye
of
the
storm
of
this
we're
working
with
smaller,
faster
leaner
organizations
to
deliver
this
quicker.
J
The
bank
as
a
whole,
though,
has
recognized
that
that's
how
we
need
to
partner
to
move
quicker.
We
need
to
think
differently
and
whether
that's
from
a
technology
perspective
or
a
cultural
perspective
or
just
an
execution
perspective,
we
have
to
be
different.
Those
40
000
people
in
Tech
all
need
to
slightly
change
the
way
they
work
to
deliver
quicker,
resiliency
and
security.
J
It
kind
of
goes
without
saying
that
we
need
to
be
a
resilient
organization,
anything
that
happens
within
our
services.
If
something
goes
down,
you
know
our
mobile
apps
do
have
periods
where
they're
down
we
don't
want
them
to
be,
but
we
do
is
an
issue
for
us.
It's
an
issue
for
us,
because
our
customers
can't
get
access
to
their
money,
but
it's
also
an
issue
because
The
Regulators
are
watching
every
time
we're
on
escalation.
Calls
and
I
was
on
one
earlier
on
in
breakfast.
J
J
So
we
need
our
solutions
to
be
resilient
and
secure,
like
you
wouldn't
believe,
and
then
lastly,
speed
and
automation
at
scale
when
you
have
40
odd
countries,
when
you
have
thousands
of
apis,
when
you
have
millions
and
millions
of
customers
and
40
000
technicians,
you
have
to
get
things
out
quickly
right
now,
if
I
want
one
of
my
teams
to
build
and
deploy
a
new
API,
it
takes
them
weeks.
J
F
Thanks
Jen
so
digital
strategy,
but
how
does
that
impact
me
selfishly
I'm,
Commercial
Banking?
This
is
my
business
yeah
and
you
can
see
some
some
figures
on
there.
We
flow
a
lot
of
money
around
the
organization
or
around
the
planet.
We
fund
a
lot
of
things.
We
don't
build
buildings,
we
don't
sponsor
or
we
don't
build
airports,
but
you'll
see
our
logos
everywhere.
F
Commercial
Bank,
it's
very
important
that
we
operate
at
a
certain
Pace.
It's
very
operate
very
important
to
operate
at
a
certain
speed
to
not
miss
opportunities.
One
of
those
opportunities
is
around
digitization
of
our
services.
What
I've
just
brought
up
on
the
screen
is
a
challenge
that
we
go
on
to
try
and
get
a
mobile
app
live
on.
A
cloud
native
platform
within
two
years
sounds
like
a
walk
in
the
park
or
a
cakewalk
for
some
people
out
there,
but
we've
never
launched
a
new
product.
F
We've
never
launched
a
new
mobile
Channel
like
this
in
15
years.
How
do
we
do
it
in
two
years
on
a
technology
stack?
That's
not
proven.
That's
not
understood,
that's
not
known,
but
we
like
to
take
those
challenges
on.
Let's
look
at
one
of
our
values.
We
get
things
done,
but
actually
our
normal
operating
pace
is
like
this.
F
This
is
how
we
work,
and
we
all
know
how
long
it
takes
to
turn
an
oil
tanker
or
a
cargo
tanker.
But
when
I
looked
at
this
graphic,
when
we
did
the
rehearsal
yesterday,
I
thought
actually
this
message
or
this
Picture
Tells
A
lot
of
different
stories.
It
tells
you,
doesn't
it
there's
a
cargo
tanker
or
an
oil
tanker,
that's
incredibly
difficult
to
turn
around
or
move
or
change
course,
but
actually,
what
is
on
that
cargo
tanker
containers?
F
Yeah
I,
didn't
use
this
because
we're
going
to
talk
about
Docker
by
the
way
containers
on
that
cargo
tanker.
That
is
our
neural
network
yeah,
because
we
are
funding
the
contents
of
those
containers.
We
are
funding
the
people
building
it
in
one
part
of
the
world.
We
are
funding
the
people
requiring
it
on
another
part
of
the
world,
we're
funding
the
shipping
company
making
the
containers
we're
funding
the
shipping
companies
agent
shipping,
the
container
they
are
our
neural
network.
F
There
are
apis
talking
across
our
organization
and
if
we
cannot
get
those
apis
talking,
if
we
cannot
get
those
containers
talking
and
what
is
our
cargo
tanker
end
up,
it
ends
up
on
its
side
blocking
a
canal
somewhere
yeah.
We
don't
want
that
because
we
know
what
happens
in
the
global
supply
chain
when
one
container
ship
goes
wrong,
one
container
ship.
J
Is
it
this
is,
who
I
mean
good
point
and
we
exist
to
give
people
access
to
their
money
to
Investments.
We
also
exist
to
make
sure
that
the
owners
of
the
companies
who
bank
with
Gareth
is
part
of
the
organization
have
somewhere
to
keep
their
loose
change.
J
Now.
Wpb
is
big
and
it's
unwieldy,
we're
not
as
necessarily
slow
as
some
of
the
other
parts
of
the
organization.
What
we
are
is
chaotic,
so
we
have
many
many
thousands
of
people
working
in
our
our
part
of
the
organization,
38
million
customers.
That's
quite
a
lot
of
customers
spread
across
the
group,
most
of
them
in
Europe
and
in
Asia,
but
we
have
presence
all
over
the
world
and
we
have
people
building
the
same
solutions
for
those
customers
across
the
world.
J
If
we
asked
any
of
our
teams
today
who
is
building-
and
we
we
do
this
all
the
time
who
is
building
in
a
New,
originations
Journey
for
checking
accounts,
there
will
be
at
least
six
teams
doing
the
same
thing.
At
least
six
teams,
that's
the
number
I
plucked
out
for
my
head.
It's
probably
like
where
we
have
a
big
challenge.
Is
we
have
to
get
our
teams
access
to
something?
J
That's
already
been
built
that
they
can
reuse
we're
about
economy
of
scale,
because
that's
the
way
that
we
get
new
things
in
front
of
our
customers,
not
the
same
thing.
Re-Engineered,
not
the
same,
stop
check
process.
Do
people
still
use
checks
they
do
here.
Don't
they
we've
been
re-engineering,
stop
tech
for
a
long
time,
but
we
don't
want
to
re-engineer
the
same.
J
J
Every
now
and
again
we
do
funky
stuff.
This
is
this:
is
a
payments
application
in
Hong,
Kong
called
pay
me,
which
some
of
you
may
have
come
across
pay
me
was
started
a
few
years
ago
in
the
same
vein
as
kinetic
was
in
wholesale.
It
was,
it
was
a
startup
mentality.
It
sat
separately
to
the
bank,
you
don't
have
to
be
a
HSBC
customer
to
use
it.
J
It's
been
extremely
successful.
It
has
its
challenges,
we're
trying
to
bring
it
a
little
bit
back
into
the
bank
now
so
that
we
can
support
it,
because
we
have
the
same
resiliency
challenges,
the
same
security
challenges
with
with
our
our
sort
of
offshoots
as
we're
doing
with
the
main
part
of
the
bank,
but
we
need
to
move
fast
as
well
from
time
to
time
and
there's
all
sorts
of
things
that
happen
within
my
world
that
mean
that
this
stuff
is
really
important.
F
Thanks
Jen
thanks
again,
let's
Refresh
on
a
digital
strategy,
and
what
do
we
do
and
how
do
we
make
that
different
yeah?
The
first
thing
is,
you
know
we
look
at
fast,
easy,
digital,
agile,
and
we
see
the
blockers.
You
know
resource
hungry
resources.
What
do
we
do
about
that?
Yeah?
We
need
some
tool
out
there
to
help
us
discover.
We
need
some
set
of
tools
out
there
to
help
us
reuse.
My
business
Jen's
business
is
all
about
reuse,
reducing
cost,
taking
our
bottom
line,
making
it
more
productive
to
reinvest
into
other
things.
J
At
the
moment,
our
API
developers
are
very
Niche
they're
in
teams
across
the
world.
That's
all
they
do
apis
apis
apis
they're,
not
skills
that
are
readily
available
in
the
organization
they're,
not
embedded
within
our
agile
teams.
We
have
to
transform
our
Workforce.
This
is
not
just
about
a
technology
transformation.
This
is
not
just
about
implementing
a
new
platform.
This
is
about
transformation
of
an
entire
Workforce.
J
F
You
and
we
talked
about
resilience
yeah
we
talked
about
somebody
actually
does
go
to
jail
here
and
there's
a
big
fine,
but
I
think
you
know
I
think
somebody
else
mentioned
you
know
one
one:
does
the
fine
really
Warren
implementing
if
we're
making
so
much
money?
What's
the
point
we'll
just
take
a
fine,
so
our
regulator's
got
a
bit
serious
and
said:
okay,
if
a
Fine's
not
enough
about
going
to
jail
instead
and
spending
time
behind
bars.
So
that's
why
we
need
resilient,
secure,
robust
apis.
F
J
A
J
Few
years
is
the
vendor
that
we're
on
still
going
to
be
the
right
decision.
Decisions
that
we
took
three
four
five
years
ago
are
really
hurting
us
now
really
hurting
us
now,
and
we
can't
do
that
again.
We
can't
make
the
same
decisions
and
the
same
mistakes
and
tie
ourselves
into
one
vendor,
because
we'll
never
be
in
a
situation
where
we
can
keep
moving
faster
and
the
business
don't
want
to
keep
funding
technology
transformation
they
want
to
keep
funding
new
business.
F
Here
anyway,
I
feel
like
we're,
educating
everybody
on
global
problems
of
the
world
and
how
banks
operate.
That's
not
a
pull
on
people
to
come
and
join
HSBC.
J
F
Okay,
we've
managed
to
get
almost
all
the
way
without
even
talking
about
our
technology.
That's
how
good
we
are
HSBC.
Yeah
we've
talked
about
the
problems
we
haven't
talked
about
Solutions.
So
how
do
we
bring
Kong
into
that
conversation?
That's
why
we're
here?
That's
why
we're
here
to
talk
about
and
and
reflecting
the
great
products
and
services
that
Khan
is
building
and
developing.
F
So
where
does
that
fit
in
our
ecosystem?
It
fits
in
our
ecosystem,
where
we
try
and
identify
the
single
unit
of
measure
yeah.
We
said
we
have
between
12
and
15
000
of
these
things
somewhere
in
the
bank,
lots
of
them
all
doing
the
same
thing
spread
across
100
or
so
thousand
containers
on
every
flavor
of
container
platform.
You
could
possibly
imagine
so.
How
do
we
get
better
at
that?
First
of
all,
let's
try
and
rationalize
down
on
that
container
platform.
F
F
I,
take
that
any
day
of
the
week,
because
it
really
doesn't
cover
the
cracks
of
Elsewhere
across
our
organization,
because
we
without
a
development
framework
without
landing
on
a
runtime
for
our
microservices,
it's
pointless
because
we
just
proliferate
these
things
wherever
we
want,
that's
where
for
us,
Kong
comes
in,
and
there
are
three
Kong
gorillas
on
this
go
on
the
screen
for
a
very
clear
reason.
As
all
of
the
great
stuff
we've
talked
about
how
Kong
is
trying
to
bring
us
together,
we're
still
deciding
to
build
three
different
types
of
Kong
in
our
organization.
F
F
So
it's
that
good.
We
want
three
of
them
right,
we're
choosing
the
Enterprise
Gateway
product.
For
now
we
are
really
really
impressed
with
what
we
see
from
connect
yeah.
That
might
actually
give
us
one
I
didn't
think
that
number
existed
in
an
organization,
but
that
might
actually
give
us
one
one
day
yep,
but
you
can
see
now.
F
I
talk
about
Federated
I,
really
talk
about
Federated
Jen's
got
one
I've
got
one,
and
actually
in
my
one,
I've
got
three
so
there's
four
and
then
probably
Jen's
got
more
and
our
other
parts
of
the
bank
have
got
even
more.
So
you
can
just
see
how
this
gets
worse
and
it
gets
further
compounded
when
we
talk
about
things
that
we
have
to
do
for
our
customers,
a
truly
Global
multi-cloud
strategy,
not
two
Cloud
partners
for
public
cloud
and
our
on-premise.
That's
the
anthos
stack.
F
So
you
can
see
alone,
even
with
a
single
unit
of
measure,
the
combinations
of
the
three
implementations
of
Kong,
the
combination
of
the
five
hosting
Regional
choices
and
flavors.
We
have
I
think
I
worked
out.
There
was
something
like
120
choices
that
we
have
from
developers
to
try
and
host
a
workload,
look
after
a
workload
and
manage
and
contain
that
workload,
but
that's
there
for
a
reason.
I
think
you
were
going
to
talk
about
why
we
have
so
many
Cloud
partners.
J
J
The
other
reason
is
we're
a
highly
regulated
environment,
our
Regulators
Force
us
to
have
an
exit
pass
for
all
of
our
services.
We
have
a
concept
of
a
stressed
exit,
which
some
of
you
will
be
familiar
with.
If
you
work
in
that
industry,
where
we
have
to
take
our
critical
applications
and
move
them
within
three
months,
that's
not
a
small
piece
of
work.
J
So,
whilst
we've
got
multiple
Cloud
Partners,
while
we
have
multiple
instances
of
Kong,
what
we
need
to
do
is
to
make
sure
that
if
I
need
to
move
my
services
and
my
applications
onto
gareth's
Kong-
which
hopefully
there's
a
little
bit
of
room
for
me-
to
do
that-
but
I'm
over
here.
Sorry,
when
you're
over
there
that
it
looks
and
feels
as
close
to
my
implementation
of
Congress
possible
so
that
the
apis
that
I'm
running
will
work
with
Gareth
and
vice
versa,
because
when
we
need
to
do
this,
we
need
to
do
it
quickly.
J
J
F
F
C
Well,
we
will
be
back
here
in
about
10
minutes
for
the
next
session,
which
will
be
from
Ikea
here
in
the
main
stage
on
track.
Two,
you
will
have
introduction
to
the
con
Gateway
Operator
by
Victor
on
track
three.
We
have
managing
Kuma
tokens
with
hashicorp
Vault
and
on
track
four,
you
have
webassembly
in
Kong,
using
wasm
X,
so
enjoy
a
small
break
and
we'll
see
you
in
a
bit.