►
Description
Autonomous Risk Management of DeFi - Vipin Bharathan, dlt.nyc
DeFi has seen rapid development. With this rapid development comes reports of spectacular heists, scams and hacks. Rapid development and great returns come with risks. Without risk, there is no reward. The broader financial system has been dealing with this phenomenon for centuries, mostly through the practice of Risk Management.
This talk answers the question "Can such Risk management practices be adopted to DeFi?" Our answer is yes. We propose a framework and suggestions for a solution fusing familiar risk management proposals.
These include ways to deal with the vast amount of data, Value at Risk Methods (including governance risk), trading risk methods integrated and extended to the DeFi world through automated risk monitoring and mitigation. The surfacing of these risks and integration into trading systems on the edge (aka wallets) with a transparent scoring mechanism will result in wider acceptance of DeFi.
Drawing parallels with structures in Finance; the hacks that have happened, ideas for infrastructure, automation of scoring, division between the edge and the core and integration of such techniques into the trading and risk management of Web3:Yield Farms, DEXes, NTFs ...
Sessions Details & Slides (if available): https://sched.co/1AySO
Access All Sessions: https://events.linuxfoundation.org/hyperledger-global-forum/program/schedule/
----
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
Discord: https://discord.gg/rVCn4j7UFd
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 #Blockchain
License
Creative Commons Attribution license (reuse allowed)
A
So
I
got
a
call
from
somebody
about
two
days
ago
saying
you
have
to
present
this
talk
because
they
first
did
not
take
it
and
someone
canceled.
They
wanted
somebody
a
sacrificial
goat
at
the
last
minute.
So
that's
that's
why
I'm
here
and
I
threw
together
this
presentation
in
maybe
the
last
couple
of
hours
but
I'm
very
familiar
with
the
topic,
so
there
may
not
be
everything
that
I
wanted
to
say
in
here.
So
please
excuse
me.
A
So
the
themes
are
very
simple:
what
are
you
know:
institutions,
the
evolution
of
the
institution,
the
D5
stack,
risk
and
management
of
risk,
the
default
stack,
reworked
implementation
and,
of
course,
references
and
Q,
a
h.
Of
course
you
can
do
you
can
ask
questions
anytime
you
want,
but
for
your
questions
to
be
recorded,
you
have
to
have
a
mic
or
you
have
to
tell
me
so
that
I
can
repeat
the
question
all
right,
so
the
D5
stack
you
know
is.
A
Basically,
this
their
thesis
I
mean
the
new
institutional
economics
thesis
is
that
institutions
are
the
basis
of
society.
Institution
is
just
a
law
like
some
kind
of
a
law,
the
rule
of
law,
property
rights
being
the
primary
one
which
the
Western
societies
have
adopted,
and
he
Douglas
North
and
others
think
that
that
is
the
main
reason
for
the
rapid
economic
development
of
Western
societies
and
obviously
the
IMF
has
taken
that
and
said.
Oh,
you
know
governance
matters,
so
they
want
every
society
to
be
reflect.
A
There
are
institutions
evolved
through
organizations,
organizations,
meaning
more
organized
groups
of
people
or
actual
structures
that
created
this
sort
of
takeoff
point,
but
the
main
thesis
I
want
to
say
is
that
it
is
not
an
end
point.
Meaning
Western
Society
in
this
particular
form
is
not
an
endpoint
and
that's
what
Douglas
North
was
saying
that
it's
an
end
point.
A
B
A
You
know
the
rights
enforcement
of
Rights
and
enforcement
is
usually
not
just
a
law
standing
by
itself,
but
it's
some
form
of
government
that
has
to
do
the
enforcement
with
the
standing
army
or
some
other
way
of
enforcing
that
law.
So
the
thing
that
the
D5
stack
defies
lack
today
is
a
structure
similar
to
that
it
does
not
have
to
be
a
centralized
government,
but
it
has
to
be
some
kind
of
a
something
that
is
monitoring
and
enforcing
the
law
of
property,
which
means,
if
I
own
a
token.
A
A
You
know
this
is
proposed
by
some
guys
a
long
long
before,
but
it
is
a
very
familiar
stack
for
us
right.
There
is
a
base,
immutable
orders,
blockchain
assets,
tokens
of
all
kinds
and
what
distinguishes
Defy
is
logic
logically,
so
this
is
all
reworked
by
me.
You
know
these
names
were
not
in
the
original
paper,
but
we
are
familiar
with
this
and
the
logic
layer
is
the
most
important
layer
for
D5,
which
we
already
know.
A
So,
let's
back
up
a
little
bit,
we
have
now
risk
in
D5,
which
is
liquidity,
risk
transformation
of
maturity,
risk,
which
means
that
you
know
we're
trying
to
Leverage
using
one
asset
and
exchanging
it
for
another
asset.
The
various
exploits,
so
the
main
way
that
we
can
manage
the
risk
in
this
is
using
ideas
from
C5.
A
Scoring
the
logic
layer
or
scoring
the
asset
layer,
the
other
one
is
circuit
breakers.
You
might
be
familiar
with
the
idea
of
circuit
breakers
from
Equity
markets,
which
is
also
due
to
program.
You
know,
program
trading,
which
is
a
kind
of
autonomous
execution
of
Trades
without
human
intervention,
but
it
has
cost
a
lot
of
problems,
including
Rapid
Falls,
in
the
price
of
certain
equities
or
a
turmoil
or
volatility
in
the
market,
which
is
not
a
natural
volatility.
It
is
cost
purely
by
the
autonomous
programs
acting
on
their
own.
A
After
the
after
one
of
the
crashes,
people
proposed
that
we
should
have
some
kind
of
a
circuit
breaker.
That
means
if,
if
the
price
of
an
equity
Falls
at
a
certain
rate
in
you
know
like
20
percent
in
an
hour
or
in
three
minutes
or
whatever,
whatever
the
rate
is
that
security
will
no
longer
be
traded,
there
is
an
automatic
pause
that
goes
on.
For
you
know,
depending
on
the
exchange,
it
would
be
for
20
minutes,
30
minutes
and
so
on.
A
So
that
is
the
idea
of
circuit
breakers,
the
other
one,
which
is
a
scoring
stuff,
we're
already
familiar
with
it.
From
normally
it
is
scoring
of
individuals,
but
you
can
also
score
any
organization
or
an
asset
using
certain
conditions
using
certain
criteria.
Sorry
and
when
we
say
Dynamic
scoring,
that
means
that
score
constantly
changes
and
who,
how
would
this
work,
is
if
you're,
a
Trader
in
that
market?
A
First
of
all,
you
have
two
types
of
Trades:
one
is
the
ones
that
you
hold
positions
in
and
the
other
is
stuff
moving
up
and
down
all
the
time,
and
you
decide
either
to
participate
in
the
market
or
sell
off
in
the
market.
So
it's
for
individual
decision
making
and
also
you
know,
this
kind
of
method
can
be
used
for
scoring
of
any
asset,
including
decentralized
assets.
A
A
A
So
if
there's
a
run
on
on
usdt,
where
people
are
rapidly,
you
know
trying
to
get
dollars
back
from
the
usdt
market,
then
they
have
to
sell
off
the
corporate
bonds
rapidly,
which
of
course
will
cause
prices
of
corporate
bonds
to
fall.
So
now
you
have
a
cross
system
Contagion
the
liquidity
and
maturity
risk
is
you
know.
We
know
that
if
buyers
disappear
from
the
market,
that
prices
will
fall
and
if
sellers
disappear,
the
you
know,
prices
will
will
rise
if
there's
more
demand
and
Supply
Supply.
You
know
this.
A
A
A
A
A
Decentralized
autonomous
organizations
cannot
be
censored
right,
but
there
is
a
thought
today
that
we
have
to
extend
the
token
standards
to
include
to
include
facilities
to
control
the
token
meaning,
like
ttf,
proposed
that
it
should
be.
There
should
be
an
interface
which
passes
the
token
and
then
resumes
the
token,
but
obviously
not
everybody
can
call
this.
A
A
So
this
is
probably
one
of
the
last
slides
and
I
I.
Don't
know
how
I'm
doing
for
time,
but
I
have
about
15
minutes,
but
you
know
I
I
want
to
leave
enough
time
for
questions,
because
this
is
not
just
some.
This
is
this
is
all
material
that
is
absolutely
new
in
some
sense
it
so
to
build
such
a
system.
A
What
do
you
need?
A
system
that
consists
of
tunability?
That
means,
like
I,
will
say,
Okay
20
down.
If
the
token
goes
down
by
20
percent,
then
we
pause
or
I
can
tune
it
to
say
it
is
if
it
goes
down
by
19.5
percent
or
you
know
some
other
percentage.
So
that's
what
I
call
tuning
I
mean
that's
a
simple
example
of
a
of
a
of
a
token
that
is
controlled
by
an
exterior
system,
not
exterior,
but
you
know
now
you
can.
A
You
can
try
to
enlarge
the
system
by
saying
okay,
how
can
I
make
it
decentralized?
How
can
I
make
the
analysis
system
analytic
system
itself
decentralized?
These
are
some
of
the
ideas
that
we
are
playing
around
with
you
know.
I
have
been
talking
to
some
people
about
these
implementing.
You
know
building
something
like
this.
The
other
is
upgrades,
which
is
how
do
you
upgrade
either
the
analytic
layer
or
the
smart
contract
itself?
A
There
are
ways
to
do
the
SMART
contract
upgrades
it's
a
very
rigorous
process,
you
know,
and
so
on
then
adopting
techniques
from
stock
markets
which
we
discussed
just
now.
There
are
other
techniques
that
we
can
talk
about.
That
is,
you
know
either
to
throttle
the
system.
That
means,
if,
if
the
system
rate
is
observed
to
be
let's
say
20
percent
higher
than
normal,
then
you
can
say
bring
down
the
rate
so
that
it's
it's
you
know,
let's
say
only
10
percent
higher
you
can.
A
You
can
do
a
lot
of
different
things
with
this
and
it
has
to
be
in
the
governance
that
means
most
tokens
today,
most
Dao's
operate
with
the
white
paper.
Maybe
we
have
to
State
these
governance
ideas
in
the
white
paper
itself
so
that
it
controls
the
the
way
you
know
the
arm
works
and
people
are
not
hit
with
unexpected
effects.
A
The
fourth
idea
is,
of
course,
do
not
give
all
the
control
to
programs
but
have
some
way
of
manual
intervention,
and
these
are
my
references.
You'll
you'll
have
you'll
get
the
you
know
the
PowerPoint
presentation
from
the
site,
and
now,
of
course,
it's
q.
A
if
you
guys
have
any
questions,
I'd
be
glad
to
answer
it
and
he's
going
to
hand
out
the
mice
the
mic.
C
So
my
question
is:
don't
we
already
have
this
in
the
economies
that
we
know
since
years,
and
it's
called
like
governments,
for
example:
yes
and
I'm,
not
criticizing
what
you
are
saying
or
what
you're
suggesting,
on
the
contrary,
I'm
seeing
the
and
I'm
critical
towards
the
how
defy
works
and
what
is
being
claimed
there,
because,
generally
there
is
and
again
that's
my
personal
opinion,
like
a
lot
of
claims.
Put
there
like.
C
Oh,
we
don't
trust
our
governments,
we
don't
trust
who
controls
the
economies
and
so
on,
and
therefore
we
try
to
bypass
all
that
through
D5,
but
at
the
end
of
the
day,
if
you
that
I
think
the
reality
is
that,
if
you
don't
trust
your
governments
or
what
you
name,
the
institutions
that
govern
your
everyday
life.
Probably
you
have
more
problems
that
you
can
solve
with
the
blockchain
and
it's
it's
more
like
a
theoretical
and
how
to
say
it
like
a
philosophic
point
to
discuss
rather
than
technical.
C
A
Bit
yes,
definitely
most
D5
is
actually
centralized,
because
anybody
some
some
wallet
controls
the
smart
contract.
I
can
put
another
smart
contract
in
there
and
that
can
steal
all
the
all
the
all
the
stuff.
So
I
agree
with
you
that
there
you
know
there
exists
already
government,
but
there
also
exists.
A
The
Dao
and
people
are
trading
on
it
daily
with
billions
of
dollars.
So
we
have
this
two
situations:
I'm
trying
to
say
how
do
we
bridge
the
gap?
How
do
we
bring
them
together?
What
can
we
do
that
is
similar
to
a
decentralized
way
of
thinking
that
will
bring?
You
know
that
will
bring
some
kind
of
an
institution,
a
rule
of
law
to
the
Dao.
That's
that's!
A
A
Programs,
and
only
if,
like
two
out
of
three
agree,
then
you
will
do
the
pause.
I
mean
that's
one
way
of
doing
it,
because
human
beings
cannot
intervene
in
the
time
that
it
takes
for
these
to
go
south.
It
always
happens
very
rapidly.
How
we
are
not.
You
know
we
are
not
cyborgs
not
yet
anyway,
and
we
have
to
have
delegation
to
certain
things
that
we
can
trust,
but
who
will
watch
the
Watchers
I
mean
this.
A
B
A
B
Thank
you
so,
when
you're
talking
about
the
management
of
risk,
so
if
I
understand
correctly
you're
trying
to
superimpose
the
concept
of
Adele
to
basically
do
risk
management,
but
with
that
was
in
production
like
the
way
that
we're
so
when
you
deploy
a
smart
contract
right,
a
lot
of
the
code
that
has
been
deployed
out
there,
you
cannot
fundamentally
change.
You
can
change
the
state,
though,
through
like
subsequent
actions
and
when
you're
saying
mitigate
risk.
B
So,
like
my
primary
concern,
is
that
any
centralized
system
is
always
going
to
be
strictly
faster
than
a
decentralized
system
right
and
you,
your
stipulation,
is
correct
that
humans
cannot
intervene
in
a
Time
necessary
for
a
malicious
attack
to
happen.
So
my
primary
question
is:
is
it
necessarily
like
true
that,
if
we're
imposing
these
like
analytic
mechanisms
like
we're
operating
at
the
speed
of
code
here,
how
can
we
necessarily
at
scale
prevent
you
know
like
basically
a
plus
one
attack,
so
an
exploit
happens?
B
How
can
humans
like
in
a
Dell
necessarily
intervene
because
it's
like
some
attacks
are
too
large
in
which
you
know
an
exchange
gets
hacked
right
and
you
could
have
an
emergency
fund
to
you
know
just
reverse
that
to
the
parties
affected,
but
I
think
for
like
a
truly
scalable
solution.
I
think
there
needs
to
be
the
implementation
of,
for
example,
formal,
like
formal
verification
or
things
that
could
really
operate
at
the
edge.
So
we
could
basically
verify
that
what
do
we
determine
and
how
can
we
arbitrate
on
an
autonomous
way?
B
A
Well,
you
made
several
statements
which
I
have
problems
with
right
away
right.
The
first
thing
you
said
is:
smart
contracts
cannot
be
changed,
they
can
be.
Second
thing
is
the
first
Dao
attack
was
handled
by
the
fact
that
the
Dao
did
not
allow
immediate
withdrawal
of
funds,
but
had
to
be
segregated
into
a
separate
area
before
and
had
to
wait
for
48
hours
or
whatever,
before
the
person
could
withdraw
it.
A
So
vitalik
and
others
created
a
a
fork
where
they
actually
prescribed
that
wallet
address
and
that
wall
could
never
be
spent
inside
ethereum
after
that
pork.
So
you
know
all
of
these
things
have
happened
before,
but
what
I'm
proposing
is
you
know
some
way
to.
First
of
all,
there's
no
system
without
risk.
A
A
It
means
that
we
have
a
scoring
mechanism
and
then
we
see
where
are
the
false
positives,
where
the
false
negatives
and
we
keep
improving
that
system.
So
in
terms
of
that,
the
scoring
mechanism
would
be
a
edge
solution.
That
means
U
as
an
investor
in
multiple
tokens:
Can
Say,
oh
multiple
systems,
ethereum
based
Solana
based,
you
know,
whatever
else
based-
and
you
say,
okay,
what
is
my
risk
across
the
Spectrum
and
how
would
I
manage
that
risk?
So,
along
with
risk,
there
is
risk
management
that
is
not
perfect.
A
I
mean
we
had
the
crash
of
2008
2009.
There
were
plenty
of
risk
systems
in
operation
then,
but
they
have
changed
the
practice
of
risk
after
2008-2009,
mainly
by
slowing
down
things,
mainly
by
changing
things.
All
I'm
talking
about
is
you
know,
reality
based
approach,
not
idealism,
you
know
not
something
that
is
I
mean.
B
A
A
Is
that
the
time
yeah
yeah
so
I
mean
if
you
have
any
comments
or
questions,
I'm,
always
willing
and
able
to
answer
them
or
hopefully
we
if
I
cannot,
then
you
know
we
can
explore
it
together
and
anyway,
all
of
this
is
sort
of
spun
out
of
my
mind
in
a
way
in
very
short
period
of
time
and
I'm
writing
a
paper
for
UCL
University,
College
London,
which
is
on
this
topic,
and
you
know
it's
it's
a
work
in
progress.
Let
me
put
it
that
way.