►
From YouTube: HyperDrive Deep Dive
Description
As the HyperDrive upgrade approaches in less than two weeks, to further explain the upgrade, Filecoin CryptoEcon Engineer, ZX, will be providing a deep dive to highlight what this upgrade enables and how to tactically take advantage of the new change in different chain utilization scenarios. Post event, lotus and CE devs will answer miner specific questions via Slack.
A
Great
hi
everyone,
my
name
is
cx.
I
work
on
crypto
economics
and
ecosystem
for
powerpoint
and
today
I'm
very
excited
to
talk
to
you
about
hyperdrive,
deep
dive
and
the
various
incentive
consideration
around
the
upgrade
and
hybrid
drive
is
a
pretty
big
milestone
on
the
falcon
network
since
the
network
launch.
So
I'm
very
excited
to
to
just
go
over
lots
of
this,
like
lots
of
this
consideration
without
further
ado,
let's
dive
right
in
just
some
disclaimer
first,
this
is
a
research
talk
on
crypto
economics
and
incentives.
A
Right,
I
think
a
big
part
of
incentive
is
that
different
agents
have
different
private
belief.
They
have
their
preferences,
their
expectation,
so
that
will
influence
their
behavior
and
their
perception
of
the
system
right
and
then
because
there
are
so
many
agents
in
this
network.
The
dynamics
can
be
complex
in
this
in
this
kind
of
setting
and
of
course
numbers
can
be
wrong.
A
So
please
do
your
own
research
and
make
your
own
decision,
but
if
you
see
anything
too
off
feel
free
to
reach
out
to
us
happy
to
discuss
on
like
slack
github
and
so
on.
So
with
that
in
mind,
let's
just
recap
on
some
of
the
concepts
about
like
gas
hyperdrive
and
what
other
different
like
moving
parts
that
this
upgrade
would
impact.
So
to
start
off,
let's
just
remind
ourselves:
what
is
gas
right
get
this
gas
is
a
measure
of
competition
and
short
resources
consumed
by
messages.
A
It's
a
pretty
common
concept
within
blockchain
today,
and
then
there
are
a
few
other
secondary
concepts
that
are
relevant
to
gas.
So
one
is
like
gas
limit.
Gas
limit
is
relevant
for
both
messages
and
a
block.
That's
basically
there's
a
limit
on
the
amount
of
gas,
a
message
execution
can
consume
right,
and
then
you
also
have
the
gas
limit
for
a
block
which
is
like
how
many,
how
much,
how
much
gas
this
block
can
consume
right
and
then
you
have
the
gas
usage,
which
is
the
amount
of
gas.
A
So
with
that,
and
what
there's
another
concept
called
the
base
these?
What
is
the
base
fee
base
b
was
introduced
in
eip1559,
congratulations
to
ethereum
community
for
launching
this
on
a
robson
test
and
earlier
in
the
week
we'll
see
how
that
goes
and
feel
free
to
experiment
and-
and
similarly,
we
used
a
very
similar
mechanism
on
file
point
after
some
evaluation
early
on,
like
there
are
many
good
properties
that
we
want
to
include
from
efp1559,
but
on
the
high
level.
A
It's
a
poster
price
mechanism
that
improves
the
fee
estimation
ux
without
prior
to
this
mechanism.
Message
center
need
to
kind
of,
evaluate
and
guess
right
how?
How
much
is
the
message
work
for
them
and
trying
to
bid
bid
for
this
bit
for
the
block
space
before
the
space
in
the
block,
so
base
b
is
basically
telling
here
so
pierce
the
price
right,
take
it
or
live
if
it's
too
high
for
you,
okay,
walk
away
right.
A
You
can
come
back
later
if
it's,
if
otherwise,
you
can
take
the
you,
can
take
the
poster
price
and
then
submit
your
message
on
chain
that
simplifies
the
whole
ux
and
basically
really
is
a
result
of
relative
supply
and
demand
of
the
network
capacity
and-
and
it
can
be
spiky
at
times,
because
the
demand
for
using
a
network
is
spiky,
but
it
grows
exponentially
to
try
to
find
the
equilibrium
between
supply
and
demand
fairly
quickly.
So
there's
notion
of
network
transaction
fee,
which
is
for
every
single
message.
A
It
would
be
the
amount
of
gas
used
multiplied
by
the
base
fee
right.
Basically,
it's
the
measure
of
okay
at
this
moment.
What
is
the
relative
demand
supply
by
using
a
network
and
gas
usage
is
for
each
message
or
now
as
a
whole,
how
much
gas
has
been
consumed
and
then
there's
a
network
transaction
fee
that
is
burned
to
align
everyone's
interest
if
the
network,
and
also
just
to
touch
point
on
why
this
network
transaction
fee
is
burned.
There
are
many
incentive
considerations.
A
A
There
are
used
for
verifying
messages,
so
this
is
the
cost
that
you
showed
by
the
whole
network,
so
hence
it
makes
sense
for
the
message
centers
to
pay
the
network,
and
while
the
many
for
a
more
direct
way
is
to
through
this
kind
of
network
transaction
alignment,
number
two
not
burning,
also
create
some
incentive
issue,
which
is
like
senders
and
block
producing.
Miner
can
always
arrange
our
band
payment
right
this.
This
is
documented
in
the
paper
as
well,
and
the
alignment
issue
number
three
we
also
falcon
is
a
utility
token
right.
A
It's
meant
to
be
used.
We
want
to
align
the
token
supply
with
the
network
utility,
that's
manifested
by
this
demand
for
using
the
network
and
some
historical
charts.
So
far,
the
network
has
has
consumed
more
than
26
million
of
dollar
coin
in
our
transaction
fee,
which
is
pretty
remarkable,
and
the
average
daily
no
transaction
fee
is
about
75
to
90
000
file
coins
every
single
day.
So
this
is
like
a
this
is
a
this.
A
Is
a
sign
of
a
very
popular,
highly
demanded,
heavily
used
network,
which
is
pretty
rare,
even
which
is
pretty
rare
in
the
crypto
world
today
and
then,
just
earlier
in
april,
the
network
has
a
pretty
big
milestone,
which
is
the
crossing
of
the
baseline
for
the
very
first
time.
Let's
just
recap
on
what
is
the
baseline
baseline
is
another
innovative
component
within
the
filecoin
economics,
where
it
sets
a
kpi
for
the
network
right.
A
The
network
maintain
kind
of
like
is
proportional
to
network
utility
as
measured
by
this
kpi,
so
the
network
so
far
has
been
growing
very,
very
quickly.
It's
faster
than
this
kpi
that
is
set
for
the
network,
which
starts
from
2.8
x
sub,
I
at
genesis,
and
it
doubles
every
year
so
and
there's
some
other
incentive
consideration
that
comes
out
of
this.
A
So
when
we
are
below
the
baseline,
your
reward
is
sort
of
proportional
to
the
storage
and
it's
increasing,
because
the
blood
reward,
the
total
pool
of
borders
is
increasing
when
you're
under
the
baseline,
as
the
network
is
catching
up
to
its
kpi.
But
after
the
baseline,
the
network
then
follows
the
six
year
half-life
exponential
decay.
That's
for
the
total
network
reward
and
that
shifts
the
kind
of
the
incentive
and
the
dynamic
amount
miners
as
well.
What
that
means
more
concretely
is.
A
We
are
now
in
this,
like
bitcoin
or
ethereum
style
block
reward
model
where
there's
a
fixed
blob
reward
right
like
there's
a
fixed
floor
that
is
going
down
and
the
miners
are
competing
for
a
greater
share
of
the
pie
right
they
are
in
this
cooperative
competition
relationship
is
cooperative
because
there's
a
shared
branding
for
falcoid.
There
is
this:
everyone
is
a
stakeholder
within
the
network
as
a
token
holders.
A
We
want
the
network
to
succeed
and
there's
lots
of
room
to
collaborate
with
the
best
practices
and
so
on
by
the
same
time,
we're
competing
for
this
common
pool
of
block
reward
right
and
and
if
minus
adds
more
than
the
network
average
right
they
get
ahead.
They
get
a
bigger
share
of
the
pilot
reward
increase.
If
money
adds
the
same,
the
rewards
stay
the
same.
If
it
has
less
than
our
average,
the
reward
would
decrease.
So
here's
like
another
just
some
historical
data
from
like
a
week
ago.
A
The
red
line
is
the
the
network
growth
over
this
week
right
and
then
like
the
this
is
the
y-axis
is
the
percentage
minor
above
that
growth
right,
so
we
have
about
20
of
the
miners
are
growing
faster
than
the
network,
so
their
share
of
the
reward
will
increase.
The
remaining
80
are
kind
of
like
lagging
behind
so
like.
If,
given
that
this,
this
is
a
dynamic.
This
is
the
overall
macro
macro
environment
that
we
are
in.
A
Miners
are
strongly
incentivized
to
increase
the
quality
adjusted
power,
either
through
falcon
plus
deals
or
through
studying,
more
committed
capacity
sectors.
Another
thing
to
just
call
out
here
the
collateral
requirement
is
proportional
to
this:
the
expected
reward
and
the
storage
right.
So,
as
the
network
increases
in
size,
the
product
per
unit
storage
reward
will
go
down,
but
but
the
per
unit
collateral
will
also
go
down,
so
these
two
go
hand
in
hand
and
then,
like
the
the
network,
keeps
track
of
the
rate
of
growth.
A
So
just
to
recap
again:
what
is
quality?
Adjust
the
power
right
quality,
just
the
power
of
a
sector?
Is
the
sector
quality
multiplier,
multiplied
by
the
raw
by
power
and
what
is
the
quality
multiplier?
It's
the
spacetime
weighted
average
of
the
the
quality
multiplier
of
the
content
in
a
sector
in
a
sector.
Then
what
are
the
different
contents?
There
are
three
main
kinds:
you
have
committed
capacity
of
the
quality
multiplier
of
one.
A
Then
you
have
regular
deals
multiplier
of
one
and
then
you
have
falcon
plus
client
deal
with
multiplier
of
10.,
and
here
are
some
examples
you
can
think
of
each
sector
as
a
container,
then
you
can
put
different
kind
of
content
within
the
sector.
So,
for
example,
let's
say
half
of
the
sector
space
time
is
occupied
by
by
falcon
plus
deals
right.
Then
we
get
this
like
5.5
of
sector
quality,
the
sector
quality,
just
the
power
will
be
5.5
multiplied
by
the
raw
bio,
which
is
32
gigabyte.
A
In
this
case
right,
if,
let's
say
a
quarter
is
occupied
by
falcon
plus
steels,
then
you
you
get
a
different
sector
quality,
and
how
do
you
get?
How
do
you
get
more
falcon,
plus
deals
right
like
this
is,
I
think,
powerpoint
plus
has
made
tremendous
progress.
I
encourage
everyone
to
participate.
It's
a
pretty
groundbreaking
thing
in
all
of
f3,
where
we
add
this
layer
of
social
consensus
on
top
of
the
pure
resource
driven
network.
A
So
far,
there
have
been
two
two
more
than
two
petabytes
of
falcon
plus
data
allocation
of
notaries
and
miners
for
all
over
the
world
doing
vd
for
the
network.
I
just
learned
that
I
think
just
earlier
just
like
earlier
on.
There's
this,
the
new
filecoin
plus
notary
applications
have
been
approved,
so
we
have
another
extra.
I
think
more
than
five
petabyte
of
file
completed
allocation.
The
home
network
should
work
together
to
find
more
clients
right,
there's
also
a
falcon
plus
for
large
data
sets
to
to
just
let
them
try.
A
Firecoin
showcase,
stylecoin
and
and
and
store
to
earn
right,
potentially
right
there
lots
of
different
room
for
experimentation
just
to
get
people
make
use
of
our
coin.
To
really
demonstrate
all
the
different
features
that
our
miners
can
provide.
Right,
there's
also
store
and
retrieve.
I
encourage,
like
clients,
can
just
get
some
good
allocation,
try
to
find
a
miner
and
given
that
it's
two
pretty
early
stages,
lots
of
room
lots
of
business
opportunities
that
will
emerge
on
the
falcon
network.
A
Just
to
recap,
on
this
corporate
competition
we
are
now
past
the
baseline.
We
follow
the
traditional
proof
of
work,
kind
of
minty
model,
there's
a
fixed
tool
of
blood
reward
and
that's
divided
by
the
two
of
miners.
Everyone
is
strongly
incentivized
to
increase
the
share
of
their
offset
of
the
reward,
and,
just
to
recap,
on
the
stages
of
economy.
We
are
kind
of
moving
past
stage
one.
We
have
built
lots
of
capacity.
A
So
far,
the
storage
has
been
pretty
reliable,
a
very
robust
and
thriving
dynamic
ecosystem
with
secure
proof
and
and
us
baseline
maintenance
picks
up
right,
like
I
think,
hyperdrive
is
a
classic
example
of
like
reducing
the
cost
of
mining
and
using
the
network,
making
the
proof
more
efficient
and
also
making
the
chain
more
scalable
right.
I
think
going
forward
we're
going
to
we're
trying
to
try
to
make
as
a
community
maybe
triple
m
more
efficient,
and
I
think
all
miners
are
now
strongly
incentivized
improve
the
quality
of
service.
A
So
with
that,
I
think
now,
let's
that's
the
overall.
On
the
macro
side,
let's
now
guide
you
to
some
of
the
gas
changes
that
have
to
have
that
concept.
So,
on
the
high
level
hybrid
drive,
unlocks
10
to
25
x,
increase
in
narrow
capacity,
sometimes
people
call
the
tps
or
like
the
throughput.
This
is
huge
right
and
this
is
done
through
reducing
the
unit
storage,
gas
usage
of
some
of
the
most
popular
messages
remind
remember.
A
We
talked
about
the
gas
limit
and
the
gas
usage
when
you
reduce
the
gas
usage
of
certain
messages.
You
are
improving.
The
throughput
and
gas
limit
is
pretty
much
set
by
some
other
physical
limit,
but
this
is.
This
is
all
on
how
this
is
all
good,
but
there
are
other
there's
some
incentive
consideration
that
we
must
take
into
account.
A
Oops,
so
I
think
the
order
gets
off
here.
So
let's
talk
about
this
slide
first,
so
there
are
a
few
incentive
issues.
The
first
one
is
a
minor
sorry
correct.
So
we
talked
about
here's,
an
illustration
of
the
reduction
in
gas
usage
per
unit
of
storage
right
like
so.
This
is
this
is
the
centric.
This
is
a
gas
usage
for
adding
one
unit
of
sector
onto
the
chain
right,
so
without
aggregation,
adding
a
single
method,
it
will
cost
you
about
60
million
gas
units
right
and
with
aggregation.
A
This
costs
get
amortized
over
the
sectors
that
you're
aggregating.
So
as
we
can
see,
this
unit
storage
gas
usage
kind
of
decreases
pretty
significantly
as
you
aggregate
right.
So
this
this
motivates
us
to
talk
about
some
of
the
incentive
problem
with
with
hyperdrive
as
it
was.
So
one
is
minor
failures
right
like
so
as
we've
seen
from
the
previous
slide
right,
like
this
cost
saving
decreases.
A
This
cost
saving
is
pretty
significant
as
the
amount
of
aggregation
increases
right,
so
that
creates
an
additional
advantage
for
economies
of
scale
right
so
like
smaller
miners
or
like
clients
and
so
on
and
so
forth.
Might
not,
it
might
not
be
fair
for
miners
who
can't
aggregate
we
can't
upgrade
as
much
and
so
on
and
so
forth.
So
we
want
to
make
sure
the
gas
cost
is
proportional
to
the
storage,
to
make
it
fair,
just
like
your
reward
is
proportional
to
the
storage
as
well
right.
A
The
second
point
here
is
as
we
make
as
we
create
this
guy
saving
for
particular
kind
of
particular
messages.
We
really
want
to
share
this
cost
reduction
with
other
messages
as
well.
This
include,
for
example,
like
deals.
Public
storage
fields,
falcon
is
a
utility
network,
is
a
useful
network.
We
want
to
make
this
cheap
on
firepoint
and
also
with
this
request.
A
So
we
really
want
to
like
kind
of
like
even
out
not
create
any
unbalance
and
the
incentive
everybody
share.
We
kind
of
like
share
the
we
share
this
like
risk
with
the
network.
We
share
this
successful
network
as
well.
We
mitigate
this
uneven
impact
to
different
participants
with
with
this
reduction
in
gas
usage.
A
So
here
are
the
gas
changes,
so
there
are
three
concepts
that
we
introduced.
One
is
called
batch
gas
charge,
so
the
network
fee
should
be
proportional
to
to
the
amount
of
charge
that
you're
adding.
So
basically
you
have
the
single
proof.
Gas
usage
right.
The
gas
charge
for
this,
like
proof
coming
messages,
is
really
based
on
the
number
of
proof
that
you
are
batching,
so
so
it's
proportional
to
the
amount
of
trade
that
you're
adding.
A
But
given
that,
if,
let's
say
oil
is
equal,
the
supply
of
network
capacity
increases
the
base
field
should
go
down
right.
So
this
net
is
still
a
very
big
saving
for
all
miners
right
and
then
we
introduce
the
dash
balancer
and
the
batch
discount.
These
two
basically
work
together
to
create
some
kind
of
balancing
dynamic
and
to
share
this
cost
reduction
with
other
messages,
to
kind
of
like
hit
the
basic
low
and
also
have
like
different
to
make
the
cost
of
making
deals
cheaper.
For
example,
we
will
go
over
that
in
a
moment.
A
So
introducing
gas
charge
right,
so
this
is
this.
This
chart
here
on
the
x-axis.
We
have
like
the
daily
network
growth
for
the
network
right
on
the
y-axis.
It's
the
gas
unit
in
terms
of
gas
usage
right
on
log
scale,
so
the
orange
dots
here
this
are
like
what
would
how
much
gas
would
we
consume
on
a
daily
basis
if
the
network
is
on
boarding
at
a
certain
rate
right
and
then
on
the
gray
dot
is
basically
the
gas
usage.
A
If
actually
everybody's
aggregating,
to
the
the
max,
how
much
gas
would
we
consume
right,
but
clearly
it's
not
proportional
right
after
when
gas
charge
introduced.
Now
this
this
is
the
this
green
dot
right.
The
the
gas
charge
that's
unit
spent
is
proportional
to
the
amount
of
rate
that
you're
adding
just
like.
Let's
see
if
there's
no
aggregation,
but
the
beauty
of
this
is
it
does
not.
Interf
interfere
with
the
gas
usage,
so
the
gas
charge
is
just
an
additional
charge.
A
It
doesn't
change
the
base
fee
right,
so
basically
it
can
be
way
low,
so
that
makes
the
cost
savings
still
very
significant,
but
that
makes
sure
that
there's
no
incentive
misalignment
on
the
on
the
micro
level,
and
then
we
have
the
batch
balance
sheet
and
batch
discount
right.
It
creates
this
balancing
dynamic
for
base
b,
as
mentioned
earlier,
like
hyperdrive,
introduces
10
to
25x
increase
in
supply
right,
but
we
it's
unclear
how
the
system
will
adjust.
It's
unclear
how
the
demand
will
pick
up
right.
A
So
this
creates
this
like
balancing
dynamic
to
incentivize
aggregation
when
the
network
is
approaching
its
capacity
to
free
up,
mulching
bandwidth
and
decorate
its
self-regulating
force
and
the
net.
The
other
effect
here
is
to
share
the
cost
reduction
of
other
messages.
So
as
a
rough
estimate
right,
these
numbers
could
be
wrong
right
with
the
parameter
that
we
recommend.
A
We
hope
the
bc
would
hover
around
like
a
0.15
nano
field,
compare
a
significant
drop
on
where
it
was
today
right
and
note
that
there's
no
note
that,
if
the
let's
say
the
demand
is
very,
very
strong,
there's
no
guarantee
that
the
basically
would
stay
in
the
level
right.
Basically,
it's
still
in
the
hands
of
the
community
and
and
with
this
level
of
basic
narrow,
can
grow
at
a
much
much
faster
rate
that
it
was
not
possible
before
I
think,
like
more
than
500
petabytes
and
more
and
at
that
level.
A
Basically,
the
point
storage
still
should
cost
about
much.
You
should
call
cost
a
much
smaller
number
compared
to
what
it
was
today
and
the
amortized
proof
commit
cost
will
also
go
down
pretty
significantly
compared
to
where
it
was
today
as
well
and
just
going
back
to
the
previous
slide.
To
just
mention
one
more
point,
just
to
note
that
all
these
thoughts
right,
like
the
blue
line,
is
the
net
the
gas
limit
of
the
network.
A
So
in
theory,
you
can't
really
grow
way
much
higher
than
this
blue
line
without
incurring
a
pretty
high
pretty
high
base
fee.
But
I
just
want
to
call
as
well
right.
This
chart
is
not
static.
It's
also
dynamic
in
the
sense
that
engineers
and
developer
make
improvement
to
the
protocol
all
the
time
right
before
hyperdrive.
The
network.
A
Great,
so
I'm
not
I'm
going
to
just
talk
now,
I'm
going
to
run
through
some
some
scenario
to
just
to
illustrate
this
balancing
dynamic
that
the
balance
that
the
balancer
and
the
discount
creates.
So
we
have
the
so
at
the
network
base
fee
or
0.01
nano
field
right,
so
the
unit
economics
is
actually
in
favor
of
adding
single
proof.
This
is
mean
the
network
is
very
underutilized.
It
doesn't
really
it's
better.
It's
it's!
A
Okay,
if
you
don't
aggregate
right
so,
like
smaller
miner
can
actually
take
advantage
of
this
to
onboard
storage
more
cheaply,
but
as
the
network
demand
picks
up
right
like
as
the
base
fee
increases,
so
let's
say
0.1
nano
fuel
or
even
like
0.15.
The
unit
network
fee
for
a
single
proof
goes
up
right.
This
is
where
the
amortization
of
advantage
starts
to
kick
starts
to
kick
in
right.
You
see
this
descending
gradient.
A
People
are
incentivized
to
onboard
to
aggregate
onboard
onboard
storage
through
batch,
commit
and
then
in
the
hypothetical
scenario,
when,
basically
it's
pretty
is
going
up
to
one
to
nanofil,
which
is
considered
low.
Historically,
right
miners
are
then
strongly
incentivized
to
aggregate
that
will
take
advantage
of
the
cost
saving
and
that
will
free
up
marching
capacity,
so
that
will
create
this
balancing
dynamic
for
the
basic.
A
So
overall,
so
here
is
the
summary
of
some
key
takeaway.
I
think,
like
mining
will
become
cheaper.
The
unit
storage
cost
of
mining
reduces
very
significantly
as
we
can
see
from
from
the
charts
faster
network
growth.
The
network
can
now
grow
at
a
much
faster
rate,
as
miners
cannot
capitalize
on
the
gas
cost,
saving
and
also
cheaper
deals
right,
because
we
have
this
mechanism
share.
A
This
cost
reduction
with
with
other
operations,
most
importantly
deals
and
the
whole
network
benefits,
because
miners
who
benefit
more
also
pay
more
to
the
network
to
to
be
aligned
with
the
network's
incentive.
And
lastly,
we
are
now
in
this
cooperative
competition,
landscape
and
miners
are
incentivized
to
grow
their
quality
adjusted
power,
whether
through
more
ceiling
or
through
more
falcon
plus
deals
to
outpace
the
network.
With
that
I'm
super
high
for
hyperdrive.
A
I
hope
you
are
high
level
hyperdrive
as
well,
and
now
we
can
move
on
to
some
q,
a
so
in
terms
of
batch
commit
messages.
If
you
see
miners
who
maliciously
choose
not
to
include
the
message
on
chain
what
would
happen
so
I
think,
like
there
are
multiple
miners
where,
like
it's,
it
depends
if,
let's
say
all
miners
decide
to
not
include
that
could
be
an
issue
and
there's
also
with
with
a
hyperdrive.
There
is
a
longer
window
for
you
to
submit
your
messages
so
over
that
period.
A
A
And
there's
another
question
about
the
more
specific
equations
around:
how
do
we
compute
pre-commit
and
proof
commit?
I
believe
that
is
the
that's
for
the
pre-commit
deposit
and
then
for
proof
commit
that
would
be.
That
would
be
the
collateral
so
for
pre-commit
is
expected
20
days
of
reward.
I
think
the
the
calculations
are
on
chain
right.
A
Next
question:
this
is
very
hard
to
translate.
Guys
next
question
is
so
there
there
are
some.
There
were
some
like
basically
spiking
attack
right.
How
do
we
really
solve
this
right,
like
previously
their
minor
burning
or
accidentally
burning
some
foul
coin
to
despite
the
base
heat?
So
here's
my
answer
to
that
attack
right?
Let's,
let's
not,
it's
not
really
an
attack,
it's
a
protection
mechanism
right
when
somebody
is
spamming.
The
network,
with
many
messages
right
like
basically
go
up
right,
that
is
deterring
them
from
like
that.
B
A
B
Yeah,
so
you
know
the
way
that
I
think
about
this
is
that
by
creating
this
lane,
for
all
of
the
aggregated
proof
commits
spiking.
The
base
fan.
Getting
it
really
high
now
is,
is
usually
a
huge
fraction
of
the
of
the
messages
that
contributed
to
that
were
that
they
contributed
to
the
normal
gas
cost,
where
the
standard,
pre
and
proof
commit
messages.
B
Now
that,
as
soon
as
the
base
fee
rises
above
a
certain
level,
most
of
those
would
move
into
aggregation
and
would
go
into
the
the
aggregate
channel,
and
so
they
would
no
longer
contribute
to
that.
So
that
means
that
the
expenditures
of
somebody
trying
to
mount
this
attack
would
be
like
would
have
to
be
way
higher.
So
any
party
that
is
trying
to
kind
of
increase
the
base
fee
and
so
on
will
will
have
to
pay
dramatically
more
in
order
to
be
able
to
do
that.
A
A
B
Yeah
so
and
all
blockchain
networks
have
these
kinds
of
attacks.
Whenever
you
have
a
permissionless
setting
that
enables
anyone
to
use
a
fixed
resource
of
a
network,
then
that
means
that
all
of
these
are
open
to
an
irrational
attack
from
parties
that
just
want
to
spam,
and
so
this
is
why
you
have
mechanisms
that
that
are
that
are
market
oriented
that
increase
the
fee
structure
according
to
the
usage,
so
that,
if
that
you
know
spam
keeps
going
on,
then
the
the
price
of
that
keeps
increases.
B
Is
very
useful
for
this
kind
of
thing.
Any
of
these
kinds
of
attacks
will
will
run
up
the
base
on
the
attacker
themselves
and
they
will
burn
a
large
amount
of
large
amount
of
money
doing
this
now.
The
benefit
of
aggregation
is
that
during
that
period,
when
that's
happening,
you
can
still
do
all
kinds
of
pre-input
improvements
in
the
aggregation
layer
and
move
a
huge
fraction
of
the
gas
gas
cost
into.
B
A
Yep-
and
then
junior
also
mentioned
just
reading
out
for
the
audience
out
there
answering
to
jackie's
question
earlier
about
like
about
the
irrational
attack
through
gas.
Rather,
if
the
whole
network
is
a
factor,
most
minor
can
decide
to
not
include
the
messages
and
then
like,
and
it's
also
easy
to
identify
in
the
event
of
such
attacks.
B
Yeah
and
that
might
also
help
for
her.
There
might
also
be
a
good
solution
for
spam
style
messages.
Those
miners
can
also,
of
course,
identify
those
and
try
to
not
include
them.
There's
another
good
semi
here
around
saying:
hey
shifting
to
our
creation
today
requires
restarting
the
lotus
miner,
so
it's
not
really
easy
to
do.