►
Description
Combining Chocolate and Peanut Butter at the Linux Foundation - Dan Middleton, Intel
Hyperledger and the Confidential Compute Consortium are two Linux Foundation projects providing unique enterprise computing models. At the intersection of these two projects is Avalon. Avalon makes the trusted execution capabilities of CCC usable in the distributed ledger frameworks provided by Hyperledger. Beyond deployment in HL frameworks Avalon also provides ease of composition in other B2B and distributed applications.
In this talk we will discuss how these open source efforts intersect and some of the impacts of that intersection emerging in cloud service provider capabilities. Technology leaders will benefit from learning about Avalon’s building blocks and lessons learned from some of the first commercial adoptions. Developers will learn about the application design model and cloud friendly container patterns for rapid prototyping. We will tie these topics and audiences together by walking through at least one end to end usage. This usage will demonstrate how to process confidential information between parties both attached to and disconnected from Fabric and Besu networks.
----
Hyperledger – Open Source Blockchain Technologies
Hyperledger is an open source community focused on developing a suite of stable frameworks, tools and libraries for enterprise-grade blockchain deployments.
It serves as a neutral home for various distributed ledger frameworks including Hyperledger Besu, Fabric, Sawtooth, Indy, as well as tools like Hyperledger Avalon, Cactus and libraries like Hyperledger Aries, Ursa. Learn more about Hyperledger projects: https://www.hyperledger.org/use
Case Studies: https://www.hyperledger.org/learn/case-studies
Training & Certification: https://www.hyperledger.org/learn/training
Tutorials: https://www.hyperledger.org/use/tutorials
Webinars: https://www.hyperledger.org/learn/webinars
Events: https://www.hyperledger.org/events
Vendor Directory: https://www.hyperledger.org/use/vendor-directory
Subscribe to the Hyperledger Newsletter: https://www.hyperledger.org/newsletter
Follow-us on Twitter @Hyperledger
Learn about Hyperledger Membership: https://www.hyperledger.org/about/join
#Hyperledger #HyperledgerGlobalForum2020 #Blockchain
B
A
But
if
you
want
to
understand
how
to
get
going
an
open
source,
maybe
you
haven't
made
contributions
before
or
you
are
actually
a
maintainer
on
a
project,
and
you
want
some
tips
about
how
you
encourage
community
development.
My
friends,
Tracy
and
swetha
will
be
giving
a
good
talk
tomorrow
and
I
encourage
you
to
attend
that
now.
What
we're
gonna
focus
on
here
are
delicious
open
source
flavors
and
bringing
them
together
and
I
want
you
guys
to
understand
where
these
ingredients
come
from.
A
A
So
if
you
are
allergic
to
peanut
butter,
you
will
be
safe
here.
If
you
are
metaphorically
allergic
to
metaphoric,
peanut
butter
I
think
I
either
really
want
to
hang
out
with
you
and
understand
you
better
or
stay
far
away
from
you,
we'll
figure
that
out.
So,
let's
jump
into
the
three
of
peanut
butter.
When
you
start
a
story,
you
got
to
figure
out
where
the
beginning
is
and
I
thought,
let's
go
to
a
safe
period.
A
A
Good
and
that's,
of
course,
seven
years
after
Bitcoin
that
was
clear
from
that
day
right.
That
was
clear,
okay,
but
of
course,
hyper
ledger
didn't
just
sort
of
spring
out
of
nothing.
Companies
like
Intel
have
been
doing
research
into
distributed
systems
for
ages.
Ibm
had
been
working
in
this
space,
r3
delivering
products
already
digital
asset
same
thing,
so
all
of
these
companies
had
ingredients
that
were
interesting
individually
and
we
all
came
together
to
say,
hey.
A
A
So
we
don't
just
create
projects,
though,
at
the
same
time
that
these
projects
are
growing,
the
community
around
them
grows,
and
so
this
is
just
some
of
the
member
companies
of
hyper
ledger
and
then,
when
you
take
into
account
a
lot
of
the
companies
that
contribute
code,
don't
actually
pay
for
memberships.
We
have
hundreds
of
companies
that
are
around
this
ecosystem.
This
has
been
the
fastest
growing
open
source
project,
the
linux
foundation,
and
it's
only
getting
bigger
and
growing
faster.
A
We
heard
about
Walmart
joining
this
morning,
so
we
continue
to
get
more
growth
in
this
area,
but
you
probably
already
know
a
lot
of
this
background
because
you're
here
at
this
hyper
ledger
conference.
So
let's
talk
about
something
you
might
know
less
about
this
new
sort
of
chocolate
thing.
So
where
does
chocolate
come
from?
Let's
go
back
in
time
again
to
a
strange
and
mysterious
time
known
as
octo.
A
We
have
had
companies
working
in
this
space
for
a
number
of
years
until
has
actually
been
delivering
silicon
in
this
space
for
several
years.
A
lot
of
there's
there's
a
lot
of
similar
members
from
the
initial
creation
of
hyper
ledger
are
participating
here
and
one
of
the
first
things
that
we
can
start
to
compare
about
the
flavors
of
these
two
organizations
as
you'll
see
some
different
names
up
there.
When
you
see
Google,
that's
actually
Google
Cloud
Microsoft
is
a
sure
and
then
you'll
see
other
names
like
Alibaba
Huawei
Baidu.
A
A
Also,
a
little
difference
is
we're.
Hyper
ledger
started
with
one
project
and
then
we
started
to
proliferate
a
lot
of
similar
projects
with
the
intent
that
they
diverge,
maybe
cross
pollinate,
test
out
different
ideas
and
now
we're
at
a
point
with
hyper
ledger
where
we
actually
probably
want
to
start
seeing
some
convergence
where
the
CCC
is
starting
out
is
some
very
similar
projects,
with
the
explicit
intent
that
these
things
converge
and
what
all
these
these
three
projects
up
here
and
they
don't
have
a
handy
little
green
house
logo.
A
Yet,
but
with
these
three
projects
do
these
are
all
SDKs
over
confidential
compute
capabilities
and
you
probably
are
trying
to
decide-
or
maybe
not
whether
or
not
you
care
whether
or
not
these
things
converged,
because
we
haven't
talked
about
what
the
capabilities
are
yet.
So,
let's
figure
out
what
kind
of
capabilities
we're
talking
about
with
confidential
computing.
What
is
the
scope
of
this
new
open
source
organization
if
you
want
to
come
up
to
speed
on
a
broader
space
quickly?
The
UN
recently
released
this
handbook
on
privacy
preserving
computation
techniques.
A
This
covers
a
wider
area
than
what
confidential
computing
entails.
It
covers
everything
from
homomorphic
encryption
to
the
trusted
execution
environments
that
we're
going
to
be
focused
on
I
found
it
a
surprisingly
easy
and
quick
read,
so
you
might
enjoy
that
on
your
flight
home.
What
confidential
computing
focuses
on
specifically,
though,
is
protecting
data
in
use
using
hardware
based
trusted
execution
environments
and
data
in
use
is
kind
of
a
new
idea,
so
everybody's
probably
familiar
with
protecting
data
at
rest.
This
means,
when
you
write
something
to
disk
you
encrypt
it.
A
Everybody
should
also
be
very
familiar
with
protecting
data
in
transit
or
data
in
flight.
You
really
don't
ever
connect
to
any
system
on
the
internet
anymore
in
in
clear-text,
with
with
a
vanilla
connection,
you've
got
HTTPS.
You've
got
a
TLS
connection,
you're,
probably
actually
VPN
into
your
company
at
some
point
today
and
then
making
a
TLS
connection
through
that
tunnel.
A
Double
encrypting
your
traffic,
but
when
we
actually
go
to
use
this
data
that
we've
encrypted
in
every
other
part
of
its
life
cycle,
we're
suddenly
exposing
it
and
in
the
clear-
and
that
is
a
problem
or
can
be
a
problem.
You
think
about
the
whole
stack
that
your
application
runs
on,
so
this
is,
of
course,
not
to
scale,
but
you
might
have
a
VM
m
on
there.
A
You
want
to
minimize
your
tcp
if
you
can
compromise
the
operating
system-
and
we
know
that
that's
not
a
terribly
difficult
thing
for
hackers
to
do.
Then
they
can
get
at
the
secrets
that
your
application
is
using.
You
might
also
have
compromised
applications
that
can
pivot
into
the
operating
system.
There's
a
whole
lot
of
ways
that
you
can
get
at
that.
A
So
the
idea
with
a
trusted
execution,
environment
or
an
Enclave,
which
is
a
little
bit
less
of
a
mouthful
for
me
to
say,
is
if
you
can
take
that
critical,
that
security
critical
part
of
your
application
and
take
it
somewhere,
where
the
operating
system
can't
get
at
it
or
more
generally,
you
can
minimize
the
TCB.
So
it's
just
the
CPU
working
with
your
process
and
the
critical
portion
of
your
process
of
your
application.
A
What
that
would
look
like,
then
anything
on
your
system,
including
something
that's
trying
to
snoop
the
buses
in
between
system
memory
and
the
CPU
is
everything
that
you
put
inside
this
specialized
area
is
encrypted,
so
even
if
the
operating
system
gets
compromised,
it
doesn't
have
any
special
privilege
that
lets
it.
Look
inside
that
space
that
area
that
says
cache
there.
If
you
were
to
look
in
to
a
little
lower
level
of
documentation,
you'd
call
that
the
Enclave
page
cache
and
the
Enclave
gate
page
cache
is
this
special
area
that
is
automatically
encrypted
any
time.
A
So
encryption
is
cool
and
we've
never
before
had
this
way
to
encrypt
the
data,
while
we're
using
it
protect
the
process
while
we're
using
it
but
really
I
think
the
cool
thing
you
know
everybody
say
that
the
encryption
wasn't
cool
but
I.
Think
really.
The
cool
thing
is
the
integrity
and
authenticity
capabilities
that
are
added
for
these
new
trusted
execution
or
confidential
computing
environments,
and
they
illustrate
that
a
little
bit
more
I
had
a
lot
I
had
a
lot
of
ideas
on
this
slide
actually
started
thinning
it
down.
A
A
So
what
you
get
that
is
unique
within
Enclave
is
when
you
deploy
this
application.
This
is
in
a
way
similar
to
a
Bitcoin
transaction.
In
that
you
have
a
signature
over
the
contents
of
the
code.
This
is
not
just
data
integrity.
This
is
code,
integrity
and
code,
authenticity,
so
you've
deployed
an
application
and
when
the
CPU
goes
to
load
the
application
into
that
special
Enclave
page
cache,
it's
going
to
do
a
hash
of
that
memory.
A
A
It
fails
that
application
doesn't
get
to
load,
but
if
it
does
get
to
load,
what
the
CPU
will
do
is,
after
it's
checked
that
that
that
signature
matches
it
will
generate
for
you,
a
report
that
says
here
is
the
cryptographic
identity
of
the
application
that
you
are
running
and
it
passes
your
signature
check
and,
what's
more
here,
are
these
security
properties
of
the
system
that
we
call
the
TCB
level?
Now
your
application,
your
remote
application
in
an
automated
way,
can
make
security
policy
decision.
A
Is
this
a
system
that
I
want
to
go
actually
deploy
secrets
into
for
a
lot
of
our
chain,
related
transaction
block
chain
related
applications
that
might
be
deploying
a
secret
but
also
importantly,
is
not
to
just
continue
to
think
about
protecting
the
data.
But
you
now
have
an
opportunity
to
rely
on
the
execution.
That's
provided.
So
if
we
need
to
put
some
private
inputs
into
a
computation,
we're
gonna
get
a
result
and
you
can
get
and
add
a
station
to
that
result.
A
So
you
know
that
the
result
was
created
with
code
that
you
or
your
consortium
have
already
vetted,
and
it's
got
this
cryptographic
signature
around
it,
and
the
last
point
with
this
is
that
there's
api's
over
this
and
other
capabilities
like
sealing
that
we
won't
be
talking
about
and
those
building
blocks.
Let
you
create
new
things
that
people
haven't
been
able
to
create
before,
and
so
if
we
come
back
and
I
put
too
many
builds
in
the
slide.
A
If
we
come
back
to
what
the
confidential
compute
consortium
is
going
to
be
working
on,
they
have
those
SDK
applications
which
are
on
the
right
side
of
the
slide
or
those
SDKs.
Rather
so
those
are
these,
these
new
api's
that
let
you
build
new
security
properties,
new
behaviors
into
especially
distributed
applications,
and
let
you
create
new
things,
so,
what's
the
rest
of
their
green
house
going
to
start
to
build
out
with,
because
that's
just
the
beginning,
for
what
we
do
at
the
ccc?
Well,
probably,
the
next
thing
that
we'll
see
is
a
higher
level
abstraction.
A
This
is
sort
of
a
natural
flow
of
things.
You
create
new
low
level
building
blocks,
and
then
you
start
to
find
higher
abstractions
around
them
and
since
we
all
probably
do
a
lot
of
containerized
applications,
we're
used
to
the
idea
of
wrapping
our
application
and
all
the
dependencies
that
it
needs.
So
we
can
easily
push
it
out
to
some
other
environment
cloud
or
what
have
you?
A
That
is
a
whole
topic
in
and
of
itself,
and
we're
not
going
to
get
into
detail
about
that.
But
I
did
want
to
give
you
a
pointer
if
you
want
to
check
out
an
open
source
project
in
that
space,
you
can
search
for
graphing,
which
is
hosted
at
the
oscar
lab
github
group,
and
there
are
other
projects
in
that
space.
A
So
that's
just
to
get
you
started
in
there
and
that'll
give
you
an
opportunity
to
see
what
that
next
layer
of
abstraction
is
and
then
moving
further
out
from
there
we're
going
to
come
back
to
what
the
CCC
might
be
doing
at
a
later
date,
a
little
later
in
this
presentation,
but
let's
go
ahead
and
pack
that
up
in
a
greenhouse
just
so
it
looks
a
little
bit
more
like
what
we're
you
seeing
with
hyper
ledger,
and
these
are
sort
of
predictive
things.
This
isn't
exact.
A
A
So
if
we
come
back
to
our
metaphor
here,
we
now
understand
this
chocolaty,
confidentially,
goodness
and
you're
already
familiar
with.
Maybe
you
like
the
the
creamy
blockchains
or
the
crunchy
blockchains
you
get
to
pick
either
it's
at
hyper,
ledger
and
sort
of
the
next
evolution
of
what's
happening
at
hyper,
ledger
and
generally
in
the
industry.
That
dave
treat
mentioned
in
his
keynote
is
that
these
systems,
a
lot
of
them,
aren't
even
necessarily
technically
blockchains
I,
think
that
was
a
good
insight
that
our
three
had
initially
with
Corte.
A
You
can
provide
a
lot
of
the
functionality
that
people
want
without
fitting
a
very
narrow
description
of
what
that
technology
could
be
so
I
think
more
generally,
these
are
multi-party
transaction
systems.
So
now
what
do
we
finally
get
when
we
mix
together
the
chocolate
and
the
peanut
butter
here?
Well,
of
course,
we
get
some
delicious
private
multi-party
transactions-
delicious
okay.
A
So
how
would
we
reach
out
and
and
taste
this
this
yummy
new
capability,
just
gonna
reach
out
and
grab
it?
Of
course
not.
We
can't
touch
anything
with
our
hands
they're
covered
with
disgusting
germs.
So
what
we
really
want
here
is
the
power
of
the
fork.
It
will
do
just
a
little
bit
more
history
here
to
see
why
this
is
interesting,
so
I'm
going
to
grab
a
few
of
the
projects
out
of
hyper
ledger.
A
We've
got
fabric
sawtooth
and
basic
so
fabric
in
a
pass
22.0
now
sawtooth
was
where
my
company
started
a
lot
of
its
contributions
in
this
space
and
and
Mick
who
created
sawtooth
as
its
initial
author
inventor.
He
got
to
see
the
community
come
around
and
mature
this
code
to
the
point
that
it
reached
a
production
released.
It
went
on
in
one
oh
and
Mick
said
well,
hey
I'm,
familiar
with
this
code.
I
can
use
this
I've
got
a
new
idea.
A
I
want
to
use
sawtooth,
but
you've
added
a
whole
lot
of
production
bells
and
whistles
that
I
don't
really
need
so
I
actually
want
to
kind
of
strip
some
of
that
off,
because
I
really
just
want
to
use
the
blockchain
portion
of
sawtooth
for
a
narrow
part
of
my
new
idea.
Submit'
grabs
a
fork
plucks
out
a
chunk
of
sawtooth
and
now
you've
got
a
private
data
objects
lab.
So
this
takes
place
in
that
little
lab
corner
of
the
hyper
ledger.
Greenhouse
and
Mick
is
working
away.
A
Making
trusted
execution
work
with
private
data
objects
he's
having
a
good
time.
It
looks
delicious
the
fabric
guys
look
over
and
they
say
that
looks
tasty.
Can
we
have
some
and
what
is
Mick
say
he
says
no.
This
is
mine,
go
away.
No
Mick
doesn't
say
that
makes
nice,
but
Mick
couldn't
stop
them.
If
you
wanted
to,
because
it's
open
source
anybody's,
welcome
to
go
grab
a
fork
a
bit,
you
can
run
a
copy
directly
or
you
can
start
modifying
it,
which
is
usually
the
reason
that
you
make
a
fork.
A
So
our
fabric
colleagues
go
grab
a
fork
of
private
data
objects
and
now
we've
got
a
new
lab
fabric,
a
private
chain
code
and
they're
working
away
at
that,
and
while
all
this
is
going
on,
there's
other
parts
of
the
blockchain
ecosystem
that
you're
probably
familiar
with.
There's
the
enterprise
of
hiriam
alliance
and
they're
working
away,
trying
to
create
enterprise
capabilities
around
the
etherium
code
base
and
they
look
over
and
they
say,
hey
that
looks
delicious.
Can
we
have
some
and
once
again,
nick
tells
him
to
go
hit.
A
So
initially,
Avalon
is
called
something
else,
but
as
it
matures
as
it
gains
some
community
around
it,
it
gets
promoted
to
be
a
top-level
project.
It
gets
a
new
name.
It
becomes
known
as
Avalon,
so
fabric
private
chain
code.
Meanwhile,
that
also
starts
to
mature,
and
that
starts
to
get
integrated
into
fabric.
Now,
when
you
come
to
a
conference
like
this,
you
might
be
thinking
hey
it's
open
source.
All
of
this
is
public
information.
A
A
Avalon,
like
I,
said
promoted
to
a
top-level
project
and
that's
going
to
be
a
lot
visible
in
and
of
itself.
So
there's
a
lot
to
talk
about
with
fabric
private
chain
code,
which
is
why,
your
afternoon,
tomorrow,
you
can
go
to
another
room
at
4:15
and
Mick
and
Jeb
are
going
to
be
talking
about
fabric
private
chain
code
in
great
detail.
You'll
learn
a
lot
at
that
and
then,
if
you
want
to
see
what
that
looks
like
in
real
life,
you
can
see
Bruno
and
Marcus
run
an
actual
demo
of
it.
A
So
I
think
that'd
be
a
good
way
for
you
to
spend
your
afternoon
tomorrow,
and
you
can
see
in
a
kind
of
detail.
That's
going
to
give
you
the
higher
level
picture
that
you
need
in
order
to
figure
out.
Is
that
something
that
you
want
to
be
working
with?
You
know,
particularly
if
you're
already
working
with
fabric
now
Avalon
has
maybe
a
little
broader
dance,
because
Avalon
is
blockchain
neutral.
It
can
sit
as
a
layer
two
which
is
what's
the
EEA
wanted
out
of
that
capability,
why
they
created
that
spec
to
begin
with.
A
But,
interestingly,
it
means
that
it's
not
just
a
layer
two
for
aetherium,
it's
something
that
you
can
work
with
with
fabric,
so
Avalon
is
something
that's
really
applicable
across
hyper
ledger
and
outside
of
hyper
ledger
and
so
again
on
Thursday.
We
have
a
whole
workshop
with
Avalon
and
something
that's
always
helpful
for
me
is
if
I
can
get
hands-on
with
something.
A
So
if
I
see
an
integration
point
in
the
future,
I
can
take
advantage
of
it,
and
I
also
threw
up
this
just
interesting
fact
that
it
does
have
a
very
broad
appeal
and
that's
why
we
had
more
sponsors
for
this
project
than
any
other
project
that
we've
seen
at
hyper
ledger.
Yet
so,
let's
go
ahead,
maybe
and
dive
into
some
of
the
details
of
Avalon.
A
So
it
can
be
a
little
hard
to
work
through
an
architecture
diagram
in
a
setting
like
this.
So,
let's
start
by
focusing
on
just
the
orange
pieces,
if
this
was
a
normal
blockchain
application
that
you're
familiar
with,
maybe
something
like
Bitcoin
down
in
the
lower
right
corner
here,
where
you've
got
the
bright
orange
box.
That
is
your
trans
actor.
That's
your
Bitcoin
wallet!
A
This
is
somebody
that
wants
to
initiate
some
sort
of
private
transaction
generally,
it's
any
sort
of
workload
that
requires
private
inputs,
where
you
don't
want
to
publicly
share
those,
but
you
want
to
be
able
to
get
results
that
multiple
parties
can
do.
Something
with
could
be
financial
could
be
something
else.
A
It
might
go
into
something
that's
more
broadly
from
that
UN
report
about
different
cryptographic
methods,
that's
all
built
into
the
spec
there.
So
it's
got
some
flexibility.
It's
got
some
longevity
to
it
to
explore
different
paths
but
I
think
even
better
than
working
through
more
details
of
that
diagram.
Immediately,
we
could
do
another
good
open-source
thing
which
is
I
can
fork
slides
from
another
company.
So,
instead
of
just
fork
encode
we
can
grab
slides
and
actually
sometimes
they're
offered.
A
These
slides
when
we
asked
around
amongst
the
Avalon
community
who's
got
a
use
case
that
we
can
share
at
Global,
Forum
and
what
Santander
said
is
well.
We
don't
want
you
to
share
this
other
thing
that
we're
working
on,
because
we
think
that
that
has
some
good
proprietary
value
to
us
and
also
this
other
thing,
and
this
other
thing
that
we've
been
talking
about
discreetly,
but
we
have
a
whole
lot
of
things
that
we
think
are
pretty
public,
and
this
is
one
of
them.
A
So
this
is
going
to
be
an
interesting
thing
to
walk
through,
but
I
said
I'm
not
real
great,
with
walking
through
other
people's
slides
I
make
silly
metaphors
like
chocolate
and
peanut
butter.
I,
don't
understand
banking
I'm,
not
a
banker,
and
so
they
said
we'll
make
it
easy
for
you.
You
don't
have
to
understand
banking.
You
don't
understand
that
banks
exist,
though
I
nodded
politely
and
that
generally,
these
banks
compete
I
get
that
and
sometimes
they
have
to
cooperate
and
they
have
to
cooperate
usually
around
legal
issues.
A
Things
like
AML
kyc,
this
sort
of
thing
they
can
benefit
where
they
have
to
benefit,
they
have
to
participate,
and
they
also
tell
me
that
banks
sometimes
have
customers
that
have
lots
and
lots
of
money.
I
told
them.
I
did
not
understand
this,
and
maybe
if
they
could
give
me
lots
and
lots
of
money,
then
maybe
I
would
they
said
they
would
not
help
me
that
way.
A
But
I
got
the
idea
that
they
have
account
information
that
they
need
to
protect,
not
just
for
the
benefit
of
their
customers,
but
also
for
their
own
selfish
interests
that
they
don't
want
to
share
customer
lists.
They
don't
want
to
know
where
they
don't
want
the
other
banks
to
know
their
clientele
necessarily
but
again,
there's
this
potential
problem
that,
if
there's
something
fishy
going
on,
they
are
responsible.
A
So
if
they
had
a
global
view
of
these
three
accounts,
for
example,
they
might
see
that
the
same
million
euros
is
going
in
a
circle
between
the
same
three
accounts
and
that
triggers
some
sort
of
AML
problem.
They
said
we
didn't
have
a
way
to
solve
this
problem.
If
we
adopted
something
like
a
Bitcoin
model
where
everything
is
transparent,
sure
we
could
solve
this
problem,
but
it
would
be
at
the
loss
of
any
confidentiality
for
our
customers,
and
this
is
the
sort
of
the
general
problem
that
we've
had
with
blockchain
since
their
inception.
A
Great,
we
all
want
this
full
transparency
and
oh
yeah.
We
also
have
necessary
confidentiality
in
order
to
operate
a
business.
They
said
we
have
no
way
to
solve
this
problem
until,
of
course,
oh
the
magic
peanut
butter
cup
shows
up
and,
of
course,
they
think
of
this
as
the
central
place
or
not
the
central
place.
But
this
is
their
mechanism
where
they
can
now
provide
simultaneously
the
confidentiality
and
a
way
of
solving
the
the
main
problems
that
would
request
transparency
and
they
don't
think
of
it
as
a
peanut
butter
cup.
A
They
think
of
it
as
something
a
little
bit
more
Banky
as
a
vault.
But
what
goes
into
that?
Vault
is
initially
this
code.
This
signed
code
that
the
banks
have
had
an
opportunity
to
review.
Everybody
understands
the
signature
of
that
code.
Everybody
trusts
that
code,
so
that
code
gets
loaded
into
something
like
an
enclave
page
cache.
It
gets
an
attestation
which
is
absolutely
critical
for
this
to
work,
because
you
probably
wouldn't
accept
this
model.
A
If
that
code
was
just
loaded
into
some
encrypted
memory,
you
want
some
cryptographic
token
that
says
that
code
is
actually
running
in
an
environment,
that
I
trust.
That's
the
code
that
I
trust,
and
so
that's
what
that
attestation
gives
you
once
each
of
these
banks
gets
an
attestation
that
there's
an
endpoint
with
the
right
code
that
they
can
trust.
A
Then
they
can
set
up
a
secure
connection
directly
into
that
memory:
space,
not
a
connection
to
an
operating
system
to
a
network
driver
but
directly
into
that's
Enclave
and
deliver
just
the
financial
data
they're
responsible
for
delivering
and
then
there's
an
output.
That
comes
with
that.
Maybe
that
output
goes
to
an
auditor,
or
maybe
it
goes
back
to
the
bank
that
initiated
the
investigation.
A
But
there
again,
you
trust
the
attestation
that
the
code
that
I
trusted
created
an
output-
it
probably
sanitized
some
data,
so
it
doesn't
necessarily
give
need
to
give
it
count,
values
for
data
that
turned
out
to
be
irrelevant
to
that
investigation,
but
you
have
whatever
evidence,
is
required
and
part
of
that
evidence
is
the
attestation
from
your
Enclave
and
I
say:
hey
great
slides
and
they
said
no,
not
just
slides.
We
have
written
this.
This
runs
on
Avalon,
they
said.
Would
you
like
to
run
the
code
and
I
said?
A
I
am
NOT
a
confident
enough
speaker
to
toggle
out
of
PowerPoint
and
try
to
run
an
application
during
a
presentation,
but
some
screenshots
would
be
nice,
so
something
there
has
not
just
sort
of
the
the
back
end.
Part
of
this
that
works
with
Avalon,
but
also
they've
got
a
whole
front
end
for
it,
and
this
is
actually
not
super
new.
A
We've
got
all
these
banking
participants
and
relying
on
trusted
code.
That's
in
that
vault
putting
their
secrets
into
that
vault.
So
if
we
bring
that
back
to
the
architecture
picture
to
understand
that
a
little
bit
more
you're
trans
actor
again
is
in
the
lower
right
and
the
secured
area
for
again
the
code
and
the
data
is
embodied
by
that
top
box.
A
So
let's
dig
into
that
just
a
little
bit
more
and
something
that
is
interesting.
That
I
alluded
to
earlier
is
that
you
don't
necessarily
even
need
to
use
a
blockchain,
so
blockchain
might
be
a
convenient
way
to
advertise
the
availability
of
trusted
workers
to
have
a
good
place
to
submit
and
to
submit
work
and
to
retrieve
receipts
of
completed
work.
All
that
would
go
in
encrypted
form
onto
a
blockchain,
but
you
might
not
have
a
big
enough
network
that
mandates
a
blockchain.
A
You
might
have
a
use
case
that
doesn't
mandate
a
blockchain,
and
so
you
can
go
directly
into
a
different
listener,
which
is
the
the
bottom
box
there
either
way
about
half
the
job
that
Avalon
is
doing
is
orchestration.
This
is
orchestration
of
trust
and
execution
environments.
So
if
we
think
back
to
what.
A
A
What
that
leaves
us
with
is
a
whole
new
way
of
doing
execution
with
multiple
parties
for
private
transactions,
whether
it's
in
a
blockchain
environment
or
not
a
blockchain
environment
and
you've
got
a
good
opportunity
while
you're
here,
you
can
always
go
out
online
and
read
document.
It's
open
source,
it's
always
going
to
be
there.
We've
got
chat
channels,
we've
got
mail
lists,
so
once
you
leave
here,
you're,
not
you
know,
blocked
in
any
way
from
engaging
to
find
out
more
with
with
the
actual
people
developing
it.
A
But
your
opportunity
here
at
Global
Forum,
is
to
get
it
sort
of
straight
from
the
horse's
mouth.
So
I
encourage
you
to
go
see
my
fellow
horses
here
later
today
and
I
think
we
might
have
one
or
two
minutes
for
quick
questions
and
then
I
also
for
anybody
who
seriously
doesn't
have
a
peanut,
allergy
or
anything
I've
got
some
peanut
butter
cups
to
pay
you
for
your
nice
attention
during
this
presentation.
A
Great
question,
so
there
is
for
people
who
aren't
familiar
with
this.
One
of
the
reasons
that
you
see
a
lot
of
the
the
cloud
parties
participating
in
the
CCC
is
that
when
you've
got
these
workloads
that
that
are
on
the
same
physical
platform,
I
ideally
should
know
that
they're
on
the
same
physical
platform,
but
you
might
have
different
customers
using
the
same
hardware
and
the
CSP
is
more
and
more
want
to
provide
some
better
isolation
than
just
what
a
VM
em
can
provide
and
so
providing
encryption
around.
A
Each
of
those
containers
is
important
for
security
they're.
What
you
get
with
the
t's
over
and
above
that
is
the
Aten
and
the
minimization
of
your
TCB.
So
if
you've
encrypted
the
entire
vmm,
if
you've
encrypted
the
entire
VM,
that
might
include
an
operating
system
in
it
and
you're
still
going
to
be
susceptible
to.
If
something
penetrates
into
that
into
that
space,
it
can
still
pivot
around
once
it's
inside
the
operating
system,
so
Intel
doesn't
have
any
plans
to
get
rid
of
these
finer
grained
capabilities
as
the
the
ecosystem
moves
to
coarser
grained
ones.
B
B
A
So
there's
two
different
ways
to
do
it.
One
is
the
sort
of
the
most
asynchronous
way
to
do.
It
is
that
when
the
Enclave
is
created
that
Enclave
self
generates
key,
that
that
key
only
exists
for
the
runtime
of
that
Enclave.
If
you
killed
the
Enclave,
the
key
goes
away.
It
creates
a
new
one.
So
when
that
Enclave
advertises
itself,
maybe
it's
advertising
itself
in
one
of
those
registries
on
the
blockchain.
It
advertises
its
public
key.
A
So
the
asynchronous
way
is
you
go
you're
gonna
encrypt
your
data
or
your
workload
using
that
public
key,
so
the
data
comes
over
encrypted
into
the
Enclave
and
the
Enclave
decrypts
it.
The
synchronous
way
to
do
that
is
that
within
the
Enclave
it's
got
a
small
ssl
stack
in
there
and
that
can
directly
create
a
TLS
connection
in
between
two
endpoints.
A
So
there's
a
couple
ways
to
do
that
for
the
good
question:
yeah
yeah,
and
so
some
of
those
container
technologies
that
I
alluded
to
is
sort
of
the
next
wave
of
what
happens
at
the
CCC.
What
those,
let
you
do
is
you
can
describe
in
the
container
manifest
I
need
a
trusted
port
and
then
that
will
map
directly
into
the
Enclave
that
sits
over
that
container.
B
A
So
a
lot
of
that
is
gonna,
be
policy-based.
I
see
I
get
back
to
the
could
have
stopped
at
the
first
architecture.
Version
first
slide
there,
but
so
the
things
that
end
up
getting
customized
here.
Maybe
this
isn't
actually
the
best
slide
for
this,
but
you
want
your
it's
always
difficult
to
talk
about
distributed,
applications
to
talk
about
which
end
point
you're
referring
to,
but
as
a
relying
party
to
something
that's
happening
in
an
enclave.
A
I
want
to
set
my
own
policy
for
what
I
trust
so
I,
don't
necessarily
just
trust
that
I
got
an
attestation
from
an
Intel
CPU.
That's
meaningless,
I
need
to
know
a
priori
what
the
identity
of
the
Enclave
is
and
I
want
to
know
what
the
signature
is
that
I
trust.
So
when
you
deploy
that
code,
that
code
has
something
on
it
called
an
mr
Enclave.
A
No,
not
mr
enclave
misses
ignorant.
Yes,
what
is
that
and
more
signer?
That's
the
right
way
to
say
that
Mr
signer!
So
these
you,
you
have
a
an
identity
that
you've
chosen
as
the
person
that
publishes
that
code.
That's
where
one
of
those
identities
comes
from,
so
I
trust.
This
code
I'm
the
person
that
signs
it
when
I
get
an
attestation
back
from
the
hardware.
A
It
includes
that
signer
value
that
signer
identity,
so
that
is
something
that's
specified
down
at
the
the
TE
SDK,
but
when
you
work
with
Avalon,
there's
going
to
be
higher
level
abstraction.
So
here's
here's
my
policy
for
the
things
that
I
trust
and
you
might
want
to
do
higher
level
things
over
and
above
that,
okay.
Well,
thanks
for
the
good
questions,
I
think
I,
better
wind
things
down
here
for
the
next
session,
which,
by
the
way,
is
also
a
very
cool
session
on
security
topics.