I guess we can kick it off
with our first speaker. And we have
Jameson Lopp from CASA talking
about Enterprise Lightning.
Applause
Alright.
Pray that the technology gods are kind to us
today. I'm Jameson
Lopp. I'm CTO of CASA.
And I'm actually here
today to talk about
why I believe enterprise
adoption of
lightning network technology is incredibly
important for the health
and optimal use
of the network. So,
while I have not been dealing with
rolling out enterprise
level lightning technology myself,
this is something that I've actually been thinking about
for almost four years now.
So, why
enterprise lightning? There are a few
of course
answers to that, like lower fees,
instant withdrawals.
These are obvious features of the
lightning network, which is why almost anybody
would want to use it over on chain
transactions. But
when we start to look at the economics
of what is in play,
think about the way
that your regular day-to-day
transactions work. And in fact,
if you think back
to the last transaction that you
have made personally,
was that with
a business or with an individual?
If your last transaction was with
a business, could you just raise your hand?
Looks like
a lot. So, if your last transaction that
you made in the real world
was with another individual, could
you raise your hand?
So, it seems like probably
the majority of
recent transactions were
with businesses. I think that you find
if you go back through your own
accounting ledger and
the payments that you've made, most of them of course
are going to be with companies with
enterprises that are providing services.
So, if we
make the assumption that the economics
on a mainstream adoption
lightning network are going to be
similar to what we're seeing today in the real
world, I think it makes a lot of sense
that we want to get as many enterprises
onboarded as possible.
And I'll go into a lot more detail about why I think
that is important. Also, another
of course answer is
we can get some improved privacy,
especially with various services that
are doing AML KYC.
I think that a lot of people would really prefer
to have a peace of mind that
they could make a deposit onto
an exchange or make a withdrawal from a
service and not have to worry about
the flow of that value
over the time
in the future being
tracked via on-chain analysis
or other methods.
The main point that I'm going to be
talking about in detail is that
the architecture of
the lightning network itself
is predicated upon
the idea that the net
flow of money in an
economy is close to zero.
It's that you have value
flowing in one direction, but eventually
it'll be flowing all around
and around, and you're not just moving
all of the value from one party
to another and it just sits there.
Any healthy economy is going to continuously have
value that is flowing
around and around through
various enterprises and individuals.
Just a sort of side
note, I don't think that
fee income is going to be much
of a driver of adoption,
but rather the savings
that enterprises are going to get
by not having to receive
tons of UTXOs and consolidate
them and pay lots of on-chain fees
to actually consume
those UTXOs. This was
a study that BitMEX research
did that showed
that maybe you could eke out about
a 1% annual
return on investment if you're really
good at routing, but that's also
very early days and I think
that over time, as lightning network
becomes greater and greater adopted, there's
going to be more and more competition for
routing of value through the network,
so the fees are probably going to go down.
Now, four years
ago, when I first started
researching this lightning network thing
because I wanted to write an article about
it, I asked Delalu, I was
like, hey man, what do you
have to say in response to all
these people who are saying, well
the economics of the lightning network
are going to result in an extreme level
of centralization, so
we're just going to end up with a few big hubs
and everybody else is going to be a spoke
off of those hubs.
And he responded to me
and he was like, well yeah man, the people
that are saying this, they actually got a really good
point because there's going to be a lot of centralization, there's
going to be some concentrated areas of
economic liquidity where you have bigger players,
but then also you're going to have some smaller players, and
then you're going to have a lot of individuals, and so what we really
think is going to happen is that there's going to be
a scale-free network topology.
And I was like, a scale-free
network topology? I've never
heard of that, but that's a very interesting
theory, it makes sense to me,
and I guess we'll just see how it plays
out. And hopefully
you can see, it's probably pretty faint,
but if you actually go
to any of these lightning
network explorers today,
it appears to me
that the topology that is
forming on the lightning network is in
fact scale-free, that while
there is some centrality
in certain areas
of the lightning network, that
it's fairly well distributed, and
that because more players
can join the network, leave the
network at will, that I am
not particularly concerned
about a hub-and-spoke
model actually
happening, even if we have greater enterprise
adoption. So I've
got a few video simulations
here. This first one, on the left
we've got A, B, C, and D. These are
people who are buying coffee every day. We've got
E, F, G, and H. They're selling coffee
every day. And this simulation
is showing a very simplistic network
where the value is basically moving
from the payers to the enterprises.
And thanks to
AJ Towns for creating this simulation
a few years ago. But what you end
up seeing is that all of the value
moves across the network, and then
eventually it just stops.
You get through about 30
different sets of
iterations before it blows up.
Now this simulation is doing
the same thing, but you may notice at the
bottom we've got this new entity,
this ExchangeX.
ExchangeX is not
highly
connected to many other nodes. It doesn't
even have a ton of liquidity.
But what it is doing is
providing the ability
for payers and
payees to make out
of band payments to the Exchange
in order to rebalance their channel.
So if A, B, C, or D
finds that all of the value on their
channel is on the far end and they can no
longer push any value to buy coffee,
they can actually send
out of band some fiat to the Exchange
and the Exchange will then
basically rebalance their
channel for them by pushing value over the
Lightning Network. And then vice versa
with the enterprises
E, F, G, and H. Is that when
all of the value ends up on their side
of the channel and they can't receive any more payments,
they just push some value over the Lightning
Network to the Exchange and it can pay them
out of band through some other payment
network. So
the main point to all of this
is that as soon as we
have some halfway decent
liquidity provider, you know,
enterprise level exchange that
is integrated with Lightning Network,
we can now
make the use of
these Lightning channels much more long
lived. With the previous simulation
you can only get through about 30 payments
from each consumer before they
ran out of the ability to
push more value and they would have to close the
channel and reopen it or something like that.
With this
very simple simulation
it basically becomes
unlimited.
The video itself goes for the
equivalent of three years and we just
stopped rendering it after that point because there's no point
continuing. But
this is the main point that
I'm trying to make is that having enterprises,
especially exchanges, is going
to make the use of
Lightning Network and
the reuse of long lived channels
much more feasible.
So let's go to the
next slide now. This is
from Chainalysis a few years ago
where they were looking at general activity
of addresses on the network,
basically showing that
about 150 million
of recently active addresses could
be ascribed to various
enterprises, whereas about
25 million appeared
to be maybe individual wallets.
So this is just giving you some context of
general activity on the network.
This is more recent if you're
looking at payments per day
which is basically TX
outputs per day. There's a
pretty strong case to be made
that somewhere around half of
all of the transaction outputs that are created
are in fact coming
from exchanges.
This is just from looking at the transactions
that have many outputs on them
which are
almost necessarily
either a batch payment
from an exchange or possibly
a coin join, though there aren't
a ton of coin joins happening in
comparison to exchange activity and also
not all exchanges are batching
their outputs. So it probably evens
out pretty well.
Once again, the point being that if
50% of the on-chain activity
is coming from exchanges
then getting these providers
on the Lightning Network gives us
a potentially large optimization
for reducing on-chain
block space.
This is looking at a
specific transaction that was a
Binance withdrawal. This just happened yesterday.
I pulled it up on oxt.me.
It's very easy for anyone
to go to some of these
wallet analysis
services and actually look at the transactions
that are known to be
coming from exchanges.
This particular transaction is fairly
standard. It had 40-something
outputs. Six of those
outputs were all going to BitMEX.
A handful of them
were going to Coinbase.
Another handful of them were going to HitBTC.
The main point here being
that when exchanges are doing
these large batch payments
they don't necessarily know
that maybe a lot of the outputs
in a single transaction are going
to the same place because they're just
receiving a pseudonymous Bitcoin
address and they're just batching them all together
and sending them. The reason
why the Lightning Network is so
interesting is that it can
automatically deduplicate this.
You just give a Lightning invoice
and it gets routed automatically
if there's a lot of activity
going back and forth. The Lightning
Network as a protocol
is automatically deduplicating
a lot of these back
and forth transfers. This is another
bit of analysis from BitFury
which is probably hard to read but essentially
this is from 2018
showing that there are a number
of different exchanges.
This was done at a country aggregation level
but basically showing that billions
and billions of dollars
in on-chain activity
are only going from one exchange
to another exchange.
It's just back and forth, back and forth.
And to actually hit on
that a little bit more, this is another chain analysis
diagram from a few years ago.
All of the blue
edge activity are exchanges.
Yellow are some merchant
providers and the upper left
quadrant are some darknet markets.
But once again the main thing
that I want you to take away from this
is just look at all of the activity
that's going from blue to blue, exchange
to exchange.
Traders on one exchange
may be arbitraging with another exchange.
They may just be rebalancing their risk profile.
Who knows
what the traders are actually doing
but there's just so much evidence
that shows that a ton
of our transaction activity
is between exchanges.
So imagine how much
of this redundancy you could
deduplicate if instead
the value was flowing over
something like Lightning Network.
Now Balaji Srinivasan
gave a talk back in I think 2013
where he was talking about network effects
and how we grow the network
from an economic standpoint.
I think that this is a pretty accurate chart
though it is probably missing
the exchange component to it.
Because as we're well aware
merchants and users aren't
only interacting with each other at this point
because we don't have a
fully circular
100% contained crypto
economy. I think if we ever want to
get to the point where we have a 100%
circular crypto economy
we need a stepping stone.
And that stepping stone is getting these
exchanges onboarded so that
while we're not using Bitcoin
as a unit of account, the exchanges
can still convert things to
fiat and basically be able to pay
their suppliers in
other currencies until we hopefully
eventually one day get to the point
where merchants can pay their
suppliers in crypto directly and those
suppliers can pay their employees
and other enterprise
companies that they deal with also
in crypto. So what do
enterprises actually need?
I wasn't quite sure
how well timed
this talk would be but it actually seems
to be right on the money because
if you look at your schedule you'll find
several other talks that are happening
at this conference that are specifically
focusing on some of the
needs that are going to have
to be met in order for enterprises to be
comfortable with using
lightning network technology.
So what are they going to need? Well they're
going to need really high
reliable services.
In many cases these enterprises
may be dealing with large sums
of money or in the cases
of exchanges they may be dealing with money
that is technically not even theirs.
So the risk factor,
the level of risk that
they're going to be willing to take is going to be a lot lower.
So we're going to need the ability
to run large
numbers of nodes redundantly
that can be failed over without
losing money.
Exchanges and high
volume services are going to need
much better tools to help them
manage the liquidity on these nodes
and hopefully eventually also
plug into their exchanges
and other liquidity providers so that they can
rebalance channels without even having
to think about it.
Honestly from a
standpoint of even having to deal with
channel balancing I don't think that
lightning network will ever be able to get
to mainstream level adoption
until we can
really abstract away the need
to even be worrying about that.
From a scalability standpoint
of course a lot of
these companies are going to need
to be able to add and
remove nodes as
needed so that
the load balancing
is not overwhelming
a single node causing it to slow down and crash.
You're also going to need
a much better history of
being able to query
what has happened on these nodes from
a personal note
like something that
I ran into a lot when I was at BitGo
and realized that one of the reasons
BitGo was able to amass a large number
of enterprise customers was
because the
standard software like Bitcoin Core
was not really designed
to be enterprise level
scalable software. That's fine.
It works well for what
it's trying to do but
what a lot of companies ended up
running into is they would say
we want to add Bitcoin support or we want to create
a Bitcoin based company
and they would basically
throw this problem at the developers and the
developers would spool up Bitcoin Core and start
making RPC calls to add
wallets and add accounts
and keep track of things and then they would
find once they got to like a hundred
thousand users and accounts
that Bitcoin Core would slow down to a crawl
and it would not scale very well.
This was one of the things that
BitGo as an enterprise
was able to basically
break out a lot of the different pieces
of functionality that
were commonly being used from
Bitcoin Core, break them out into a
scalable architecture
and then essentially sell access
to that to other companies so that
they could have highly performant
reliable and robust queries
that they could make to a
service to get data and to
interact with the network.
From a security standpoint
these are hot wallets
so we're going to need much better
level security than just keeping
private keys in memory on some
virtual machine.
Preferably having actual
hardware security modules that are managing
the hot keys and then having other
systems to help manage the keys that don't
necessarily have to be hot. I know Chris Stewart
is going to be talking about
some of the ways that you can approach
the private key security
with Lightning Nodes later on today
and from a business perspective
additional policy enforcement that
could perhaps also be done
at the HSM level but basically
some additional level
of assurance so that
you are fairly
confident that someone won't be able
to just make a single API call
and drain all of your Lightning wallets.
And of course boring
old accounting tools, being able
to put together history on
a per user basis and figure
out what different users balances
within your system are going to be
very important.
I think that covers most of it
but in general
this is something that I think is
going to be ongoing for the foreseeable
future and I'm hopeful
that we can get some people thinking about
these problems, building some tools
hopefully even open source tools
that can be collaboratively
built on rather than
us ending up with a system
where every enterprise has to
solve all of these problems independently
and ends up duplicating
a lot of the work across different companies
in this space.
So I hope I convinced you that
having enterprise adoption
and especially exchange adoption is
important for a Lightning network so
if you're not building these tools yourself
feel free to
badger the support cues
at your local exchange and ask
when Lightning. Thank you.