►
From YouTube: Chainlink by Johann Eid
Description
Berlin Ethereum Meetup 2019/06/18
Interoperability Edition
A
All
right,
hello,
everyone
thank
you
all
for
coming,
so
I'm
you
and
I'm
a
product
manager
and
developer
evangelist
at
channelling.
So
today,
I
would
like
to
discuss
the
problem
that
you
are
solving
at
channeling,
so
the
smart
contract
connectivity
problem
first
off.
Why
does
my
contract?
Why
does
my
contracts
matter?
Why
do
we
like
small
contracts?
Why
is
there
super
practical
and
super
well
revolutionary?
First,
my
contracts
have
some
really
key
properties.
They
are
deterministic
they're,
reliable.
There
are
tamper-proof.
Whatever
input
you
have
into
a
smart
contract,
you
will
know
the
output.
A
Now
that's
a
big
change
from
the
digital
agreements
that
we
have
today.
So
in
today's
word,
for
a
derivatives
contract
for
an
insurance
contract
for
trade
finance,
it
can
always
be
delayed
right.
So
let's
say
a
payment
needs
to
be
done
once
package
is
delivered
well.
For
some
reason,
this
payment
could
be
delayed
by
a
third
party
which
would
act
maliciously
or
just
be
unreliable
right.
So
basically,
digital
agreements
of
today
are
quite
unsure,
they're
unreliable
and
if
you
have
a
choice
to
basically
have
our
world
economy
turn
into
something
that
is
better.
A
That
is
more
secure
and
more
reliable.
Then
you
should
probably
do
it
right,
so
smart
contracts
are
great,
but
why
are
we
not
using
them
today?
Why,
today,
don't
you
have
all
these
enterprises,
the
businesses
using
smart
contracts?
Well,
currently,
smart
contracts
cannot
communicate
with
the
outside
world.
It
means
that
the
real
world
data,
so
market
fields,
IOT
data
for
trade
finance
for
insurance.
All
of
this
stuff
is
not
accessible
on
smart
contracts,
so
basically
they
are
siloed
and
that's.
Where
is
the
interoperability
expect
comes?
A
If
you
don't
have
access
to
real
world
data,
you
have
access
to
nothing.
Basically,
you
have
access
to
the
data.
That's
on
the
blockchain.
This
data
is
not
going
to
change
over
you're,
not
going
to
make
a
lot
of
stuff
with
it
right.
So
let's
go
over
the
history
of
smart
contract
to
basically
see
where,
where
we
came
from
and
where
we
want
to
go
so
at
first,
when
Bitcoin
came,
we
had
multi-sig,
so
basically
I
think
if
you
had
multiple
signatures,
you
allow
the
transaction
to
happen.
A
So
here
you
get
a
couple
of
input
which
were
the
signatures
and
the
output,
which
would
always
execute,
was
allowing
the
transaction
to
happen.
Then
we
got
protocol
opcode,
smart
contract.
So
basically
you
add
a
smart
contract.
You
want
to
embed
it
into
the
protocol.
You
had
to
contact
protocol
developers
for
them
to
kind
of
implement
your
smile
contract.
So
that's
a
very
bad
approach
because
first
it's
very
lengthy.
A
It
takes
time
to
implement
and
it's
very
unsecure
because
you
have
to
modify
the
protocol
control
to
the
protocol
code
each
time
you
don't
want
to
start
touching
protocol
codes
every
time
you
want
to
implement
something
right,
it's
very
unsecure.
So
then
we
got
etherium
saviour,
basically,
which
created
the
10x
improvement
on
which
read
before
now.
You
are
able
to
estimate
contracts
which
could
be
rolled
up
in
a
few
well
hours
a
few
days
and
without
needing
any
protocol
change
right.
So
when
this
happened,
we
saw
the
wave
of
tokens
emerging.
A
So
basically
people
had
the
freedom
to
start
coding
up
tokens,
releasing
them
on
the
blockchain
and
that's
what
happened
like
you.
If
you
give
developers
tools,
they'll,
start
building
and
they'd
start
creating
stuff
that
you
wouldn't
think
of
before.
So
that's
what
happened.
We
had
a
lot
of
tokens
being
created
and
that's
fine.
That's
great.
A
However,
it
didn't
change
the
road
right
like
nothing
that
disrupted
already
it
kind
of
created
a
movement,
but
if
you
want
to
take
it
further,
if
you
want
to
kind
of
disrupt
to
kind
of
create
something
better,
more
transparent,
more
secure
for
the
enterprises,
the
businesses,
then
we
need
to
go
a
step
further.
What
is
a
step
further?
It's
basically
allowing
smart
contracts
to
have
access
to
real
world
data,
so
real
world
data.
A
What
are
those
well
market
prices
insurance
like
basically
everything
as
it
can
be
in
a
digital
agreements
as
of
today,
we
need
to
have
it
on
the
blockchain.
We
need
all
the
important
inputs
to
be
fed
into
smart
contracts
to
trigger
reliable
contracts
which
can
basically
allow
you
to
have
services
which
you
had
before
in
digital
agreements,
but
which
were
unreliable
right.
So,
basically,
can
you
explain
what
I
was
saying
here?
A
We
have
basically
data
providers
which
would
be
market
prices
which
would
be
IOT
data
and
basically
those
will
allow
you
to
as
smart
contracts
for,
for
instance,
if
you
want
to
track
the
package
delivery
of
value
of
a
package
right.
So
you
have
multiple
API
endpoints,
which
are
fitting
you,
which
are
tracking
a
certain
package.
When
this
package
gets
delivered,
you
will
have
access
to
a
payment
right,
so
you
want
to
do
a
payment,
let's
say
in
Bitcoin.
So
basically.
B
A
Case
you
need
to
have
access
to
the
price
of
Bitcoin,
so
you
meet
multiple
data
fiends
to
provide
you
with
this
price.
So
that's
what
I'm
saying.
Basically
it's
having
inputs,
which
are
which
will
basically
allow
you
to
trigger
a
smile
contract
which
will
do
a
payment
which
will
basically
generate
an
output
and
in
this
case,
unions
for
insurance
contract
I,
was
discussing
the
tracking
of
the
delivery
of
the
package
delivery
and
the
price
of
Bitcoin,
for
instance.
A
So
you
have
multiple
use
cases
with
Shannon
curves
inputs,
the
outputs
and
being
able
to
communicate
across
block
chains.
Basically,
so
now,
all
of
this
is
fine.
We
won't
secure
input
here
on
secure
outputs
and
now
how
do
we
achieve
it?
Well,
first,
let's
talk
about
how
not
to
achieve
it.
Let's
say
currently
if
you
were
to
use
centralized
Oracle
to
trigger
a
smart
contract.
So
let's
say
you
have
one
record
feeding
data
into
a
highly
secure,
highly
decentralized
computation
system
which
is
built
on
a
theory
on
mass
market
right
now.
A
This
system
doesn't
make
much
sense
right
because
you
have
the
middle
piece,
the
smart
contract,
which
is
very
secure,
and
then
you
have
the
input
which
relies
on
one
centralized
Oracle,
which
means
that
if
your
input
gets
corrupted,
it
gets
hacked
or
is
just
malicious.
Then
your
contract
will
be
triggered
for
no
reason
whatsoever.
A
So
you
basically
don't
want
to
have
an
architecture
where
the
middle
part
is
very
secure
or
where
the
inputs
and
the
outputs
are
insecure
or
centralized.
So
in
this
case,
we
strongly
advocate
against
such
a
system,
which
shouldn't
make
sense.
You
want
end-to-end
reliability.
Your
system
is
as
secure
as
the
weakest
part
of
your
system
right.
So
how
do
you
have
an
to
end?
Reliability
now,
a
challenge.
We
have
multiple
approaches
to
it
right.
A
Our
main
approach
that
you
are
using
right
now
is
decentralization
so
very
similar
to
a
theory
on
to
Bitcoin,
where
you
have
multiple
actors.
Each
of
these
actors
has
a
version
of
the
state
of
the
network,
so
basically
of
the
block
that
should
be
appended
to
the
blockchain.
Well
here
each
Oracle
has
a
version
of
the
data,
so
it
should
be
fed
into
the
smart
contract
and
probabilistically.
A
We
can
assume
that
if
most
Oracle's
have
the
same
version
of
the
data,
then
this
data
is
probably
the
reliable.
It's
probably
the
right
data
right,
very
similar
to
consensus
in
Bitcoin
or
in
theory
on
so
we
are
using
the
centralization
to
find
out
what
the
right
data
point
is.
What
the
source
of
truth
that
you'll
be
using
in
your
contract
is
so
that
was
the
example.
I
was
talking
about
previously,
but
I
forgot
that
I
just
slide
here
so
too
bad.
A
Well,
just
to
go
over
it,
you
have
a
delivery
package
API
which
is
easy
post,
which
will
feed
data
into
your
smart
contracts.
You
have
multiple
Oracle's
to
do
this,
basically
having
multiple
or
Acuras
filling
use,
the
same
data
source
is
important
because
if
you
had
only
one,
if
it
was
taken
down
or
malicious,
it
could
probably
die
right
so
having
multiple
ones
is
an
insurance
for
security
and
for
reliability.
So
once
your
package
got
delivered,
something
will
be
triggered
inside
the
smart
contract
which
says
okay
package
got
delivered.
I
should
make
your
payments.
A
A
So
it's
basically
how
we
envision
such
a
simple
use
case
right,
but
you
have
so
many
out
there
like,
let's
say
for
insurance,
you
could
have
you
know
the
new
Tesla
cars
have
a
black
box
in
the
car
right,
so
you
could
have
IOT
data
feed
data
to
your
my
contract
whenever
this,
like
er
gets
into
an
accident
or
something
similar,
the
black
box
will
signal
it
to
the
smart
contracts
and
you'll
be
able
to
reimburse
automatically
using
an
insurance
contract.
The
person
who
who
got
into
the
accident
right.
A
So
that's
another
example,
and
you
have
so
many
out
there,
like.
Basically,
every
digital
agreements
can
be
transformed
into
a
smart
contract
so
yeah.
Basically,
our
line
of
body
of
work,
a
chain
link,
is
creating
multiple
chain
links,
multiple
Oracle
and
data
providers,
which
can
allow
you
to
have
all
the
functionalities
that
basically
currently
iOS
developers
or
Android
developers
have
whenever
you
create
Weber,
was
created.
For
instance,
you
had
access
to
Google
API
rights
with
access
to
SMS
API
to
payment
API.
Well,
we
want
to
have
the
same
for
smart
phones.
A
Why
developers
we
want
to
give
you
guys
the
tools
to
have
access
to
whatever
API
you
need
to
build
in
complex
systems
which
can
one
day
disrupt
and
revolutionize
small
like
digital
agreements.
Basically
it's
about
making
smart
contract
the
main
form
of
their
agreements
in
today's
society,
and
for
this
you
need
the
right
tools
to
do
it
right.
So
that's
basically
our
vision.
Now
on
our
security,
we
have
multiple
approaches
right,
but
the
more
secure
is
the.
A
A
So
chaining
recently
is
working
with
Isis
three,
basically
Isis
redevelop,
Town
Crier
Town
Crier
is
basically
an
Oracle
which
is
using
the
trusted
education
environments
to
feed
data
into
smart
contracts.
So
these
have
some
very
interesting,
proper
properties.
First,
one
is
I,
have
a
very
small
attack
surface.
So
whenever
you're
using
the
TE,
it
cannot
be
accessed
by
the
OS
or
the
application.
So
there
is
a
very
small
attack
surface
on
this
kind
of
hardware,
which
is
trusted
execution
environment,
and
how
does
it
work
exactly?
So?
A
What
is
the
T
and
how
do
we
use
it?
Like
don't
cry?
For
instance,
let's
say
you,
you
want
to
feed
data
into
a
smart
contract,
so
you'll
have
a
ton
per
contract
on
the
blockchain.
It
will
get
a
request
from
a
users
contract
to
provide
data.
It
will
ask
the
new
operator,
which
is
running
a
trusted
execution
environment,
to
provide
this
data,
and
is
this
not
operator
like
so
te
e
is
composed
of
an
enclave
and
an
enclave
is
a
part
of
the
trust
execution
environment
which
is
basically
which
guarantees
you
confidentiality
and
integrity.
A
So
the
one
can
see
what
no
problem,
what
the
not
operator
is
running
and
even
the
not
operator,
cannot
see
what
code
is
running
right.
So
basically
it
allows
you
to
be
fully
confidential
in
what
you're
doing
and
this
property
of
confidentiality
and
integrity
allows
to
have
many
use
cases.
The
first
one
is
basically
optional
computation.
Where
you
can
send
encrypted
code
to
the
trusted
execution
environment.
It
will
decrypt
it,
it
will
compute
it
and
then
it
will
send
it
back.
A
So
that's
one
of
the
properties,
another
one
is
being
able
to
manage
credentials
so
now
that
we
know
that
trusted
execution
environments
allow
you
to
be
fully
confidential,
free
private,
even
the
not
operator,
cannot
access
it.
It
means
that,
basically,
you
can
send
credentials
to
it
and
no
one
will
be
able
to
see
what
these
credentials
are.
So
that's
how
you
want
to
do
a
payment
using
a
banking,
API
or
PayPal.
A
You
know,
whatever,
like
you'll,
be
able
to
send
these
credentials
to
the
trusted
execution
environment
for
it
to
be
processed,
and
it's
the
same
for
private
keys.
I.
Think
a
very
nice
property
which
which
you
should
be
able
to
appreciate,
is
basically
being
able
to
use
private
keys,
which
can
be
managed
by
not
operators
running
T's,
and
since
no
one
can
see
this
private
key,
it
can
just
be
sent
and
it
can
allow
us
to
sign
transactions
on
other
block
chains
right
using
su
nice
property
of
confidentiality
that
trust
the
execution
environment
offers
so
yeah.
A
Let's
not
go
too
much
into
this
yeah,
basically
I
think
that's
mostly
what
I
wanted
to
discuss.
So
just
quick
recap
I
think
it's
important
to
understand
that
inputs
and
outputs
like
what
we're
trying
to
do
a
chain
link,
is
really
securing
and
connecting
the
real
world
data
to
smart
contract
to
the
blockchain
world.
So
you
can
never
be
too
secure
to
do
this.
That's
why
we
want.
A
In
depth
approach,
we
use
decentralization,
we
use
transit
execution
environments
and
we
are
going
to
use
many
other
solutions,
so
we
are
using.
We
aren't
thinking
about
many
other
ones,
so
you
want
to
be
able
to
connect
all
the
real
world
data
into
smart
contracts,
for
you
guys
to
be
able
to
develop,
compelling
applications
which
can
one
day
disrupt
today's
digital
agreements.
A
So,
basically,
our
mission
is
to
give
you
guys
the
tools
to
build
stuff
that
currently
can
be
built
with
applications
such
as
iOS
and
Android
being
able
to
build
these
applications
on
the
blockchain
and
leverage
the
smart
contracts
which
have
such
nice
properties
such
as
not
relying
on
third
parties
being
deterministic
and
so
many
hazards.
So
it's
basically
building
tools
which
can
allow
to
build
more
secure,
more
transparent
and
more
efficient
systems
for
developers.
A
That's
about
it.
We've
been
on
block
also,
we've
been
on
the
main
net
for
I
think
two
weeks,
so
we
are
currently
ramping
up
more
and
more
on
our
integration
side.
So
I
don't
know
if
you
guys
saw
the
Google
announcement
recently,
where
Google
will
be
using
chain
link
to
provide
bigquery
data
into
a
theorem.
So
basically
chaining
is
used
as
a
middle.
Where
to
do
this
and
we
have
partnerships
with
Swift.
B
A
The
real
words
we
want
to
kind
of
take
crypto
out
of
the
intranet
where
it
is
currently
based.
Basically,
it's
very
siloed,
you
know
and
it
doesn't
communicate
with
the
real
world
and
take
it
to
some
things
that
can
really.
You
know,
change
the
world.
Basically
so
yeah!
That's
it
thank
you
and
if
you
have
any
questions,
happy.
C
A
Code
yeah,
it
has,
it
has
a
lot
of
promises.
It
also
has
some
veneer
abilities.
Currently
many
companies
are
working
on
it.
Intel
SGX
arms
also
is
using
its
own
implementation.
So
you
can
go
to
tamper
your
website
and
see
basically
in
production
use
cases,
but
it's
I
think
too
early
to
have
contracts
rely
with
billions
of
dollars
on
it.
So
it
has
a
lot
of
promises.
A
It
could
really
revolutionize
a
lot
of
stuff
you
are
doing
currently
like,
but
still
takes
more
time,
like
you
I
think
it's
the
second
layer
of
the
defensive
depth
approach.
You
know
like
the
first
one
that
you
have
to
remember
is
decentralization.
The
second
one
which
is
very
nice
to
have
is
G
if
it
works.
It's
huge
yeah,
just.
B
A
So
sorry,
chain
link
is
not
its
own
chain,
so
basically
chaining
deploys
smart
contract.
Are
you
theorem?
Currently
we
are
doing
a
theorem
which
we
are
blocks
an
agnostic
right,
so
we
want
to
be
on
every
big
block
chain
where's.
My
contacts
are
developed,
so
there
is
my
contractor
theorem.
Is
this
my
contract,
which
will
receive
a
query
from
a
user
contractor?
A
The
user
contract
wants
to
get
some
piece
of
data
market
prices,
IOT
data,
whichever
you
know,
so
it
queries
a
chain
in
contract
which
will
carry
an
Oracle
node,
which
is
obtained,
and
this
Oracle
node
will
query
data
source
will
fetch
data
from
an
API
endpoint
right
or
multiple
and
ports?
So
that's
the
way
it
is
and
so
yeah?
No,
it's
not
it's
own
chain.
However.
We
have
not
operators
to
run
this
note.
Basically,
this
Oracle
note
we
are
talking
about.