The Lightning Network is a layer 2 payment network built on top of Bitcoin. It is denominated in Bitcoin, and disputes are settled entirely by the Bitcoin network. Like Bitcoin, it is decentralized. Ultimately, the utility, speed and cost of the network will be defined by the sum of how all of the participants behave. In this post, I will explore at least 6 different user types I predict will exist on the Lightning Network and my expectation of how they will behave.
Users on the Lightning network can be roughly categorized by: The size, number, distribution and placement of their channels, whether they route payments and whether or not they are online as close to 24/7 as possible.
DISCLAIMER: These are unsourced, unfounded ramblings to be released to the community to be debated and discussed. There are likely to be novel scenarios and complicating factors that could render some or all of this inaccurate. This is a sort of crystal ball reading of what the Lightning Network could look like with a given population having above double digit Lightning Network usage among its citizens.
Example individual: Average Joe, who buys coffee from Starblocks.
Expected average channel size: small (<$100)
Expected relative channel number: small (1-4)
Expected relative channel distribution: Mostly local, or 50/50 if onboarded via an Exchange.
Channel connection points: to an Exchange, Merchant, Enthusiast, Employer, or Watchtower.
Online 24/7: no
Payment routing: no
Multiple nodes: no
The first, and simplest user of the Lightning Network will likely be the most prolific: the Consumer. This user is most similar to a user making normal Bitcoin purchases. The Consumer might be onboarded by any other user who routes payments, who can onboard the Consumer by opening a channel to the Consumer rather than sending or receiving Bitcoin directly.
Thus, the Consumer is given access to the entire network via an Exchange or Merchant, who can charge routing fees for subsequent transactions. The Consumer might also be onboarded by an Enthusiast friend, or an Employer who can pay the Consumer hourly, thus settling payroll by the day rather than 2 weeks or a month at a time. The Consumer has no incentive to be make many separate channels, so long as their existing channels are usable for their payments.
The Consumer will likely not hold many funds in channels, as they cannot be online 24/7 to police them and punish individuals that try to cheat them. However the Consumer can work with a Watchtower who is online 24/7 to punish cheaters.
The Consumer's optimal channel connection strategy is to have a channel funded mostly on their end to a well-connected node, such as an Exchange, or a Merchant. The Consumer will then make micropayments out of this channel until it is depleted, or the other side wishes to settle their counterparty risk.
Consumers will likely have most of the funds in their channels on their side, with payments deducting small amounts until they or their channel point closes the channel due to counterparty risk. The exception is channels with Exchanges, which will be covered below.
It will probably not be common for Consumers to open channels between one another, except for cases of multiple, continuous payments in one direction (example: A Consumer opens a channel to another Consumer with the amount equaling 12 consecutive grass cuttings. If a Consumer has many of these situations, they should not be considered a Consumer, but rather an Enthusiast or a Merchant.)
Example individual: Overstock, or Fred who offers computer repair services in his spare time.
Expected average channel size: medium (<$10,000)
Expected relative channel number: medium (hundreds, perhaps more or less)
Expected relative channel distribution: Mostly local balances with important nodes, mostly remote balances with Consumers.
Channel connection points: to Exchanges, Merchants, Enthusiasts, Employers, and Consumers
Online 24/7: probably
Payment routing: maybe
Multiple nodes: no
The next use case is the Merchant. This is a business that accepts Lightning-enabled payments, such as a coffee shop or internet cafe. Merchants will likely make up the largest raw percentage of nodes, as the barrier to entry is very low. Merchant is also a misnomer, because Consumers and Enthusiasts can elevate themselves to being a Merchant simply by running their node 24/7 and selling goods or services, such as a lemonade stand, or a one person lawn care service.
The principal difference between a Merchant and an Enthusiast is whether or not the Merchant wishes to dedicate time to managing their node. Most Merchants will wish for their node to run automatically, opening and closing channels according to settings they want such as total amount of funds to be locked up in channels, and total acceptable counterparty risk.
Merchants have little incentive to minimize fees as their overall rate of transactions will be relatively low, and will prioritize usability above all else. Thus, they will likely have a couple to a few large channels with the most well-connected nodes, Exchanges, larger Merchants and Enthusiasts, and their smaller channels will largely be repeat Consumers. Merchants will want their channel distributions to be mostly on the remote end with both well-connected nodes (Enthusiasts, Exchanges and other well-connected or better-connected Merchants), and with Consumers, until the Consumers have exhausted the funds in their channel.
Merchants, like Consumers, can work with a Watchtower to secure their funds and punish cheaters. Smaller Merchants, perhaps Consumers too well-connected to be considered a Consumer, will want to be online as often as possible ("synchronizing" once every couple of days, depending on channel lock time, could be more than sufficient.) This means many Merchants could accept payments via their smartphone, provided they require channel timelocks to be sufficiently long.
Some Merchants will take initiative to route payments, but most will accept default settings for fee rates, which will likely respond to market forces automatically, and perhaps not in the most efficient way. As a result, they will likely make money routing payments, but not enough to be worth mentioning.
Example individual: Gemini, Bitstamp, a power LocalBitcoins user
Expected average channel size: large to very large (>$10,000 or >$1,000,000)
Expected relative channel number: large to very large (multiple hundreds to thousands)
Expected relative channel distribution: 50/50 to all participants
Channel connection points: to Exchanges, Merchants, Enthusiasts, Employers, and Consumers
Online 24/7: yes
Payment routing: yes
Multiple nodes: yes
The Exchange is the next user in the Lightning Network ecosystem. The Exchanges will have the highest liquidity on the Network, and will be central hubs for only the very largest transactions. The definition of a large transaction is also to be determined, since at some point, the benefits of micropayments on LN outweigh the risk of counterparties locking up your funds.
Exchanges will have a person or persons actively managing their nodes on the Network. They will also be the first example of a user with more than one node on the Network. They are also the first example of a user who has a financial incentive to have efficient channels. These users will monitor channel distributions, likely using specialized software, to reduce risk, find optimal channels, offload funds from "hot" nodes to cold wallets, and very closely manage transaction fees. Exchanges will likely make the most from transaction fees, as they can front the most liquidity and monitor the Network to find optimal channels.
Exchanges will open direct channels between each other with high amounts of liquidity to enable high-frequency trading between exchanges. Since most on-chain transactions are from one exchange to another (or so I've read), this will singlehandedly push on-chain transaction fees down immensely, though the total effect remains to be seen.
Exchanges will have backup nodes, disaster recovery plans and so forth. They are likely to be the most reliable nodes on the network, and as a result, they will be common onboarding points for every other user to the Network. This further increases the Exchange's take-home from operating a routing node.
Exchanges may also offer services no other node can offer, mostly around translating legacy fiat into Bitcoin onto the Lightning Network since they already have infrastructure in place to comply with KMC/AML. One example of such a service is using LN to back and trade stocks on a stock market, or a Merchant or Employer paying an exchange in fiat or Bitcoin to allow paying their customers on Lightning without requiring the Merchant or Employer to actually write or use software that does this.
- A point to note: No user on LN has to comply with KYC/AML since at no point is a node, routing or otherwise, a custodian of someone else's funds. Exchanges could offer services where they willingly act as custodians of funds, though, which triggers KYC/AML. I am not a lawyer, and most governments are still struggling to wrap their heads around Bitcoin, so it is a large unknown that only Exchanges may feel comfortable finding novel Lightning-enabled services.
Example individual: ACME, Inc., Walter's Widget Co.
Expected average channel size: medium to large to very large (<$10,000, >$10,000 or >$1,000,000)
Expected relative channel number: large (multiple hundreds to thousands)
Expected relative channel distribution: Mostly local balances
Channel connection points: to Exchanges, Merchants, Enthusiasts, Employers, Consumers and Watchtowers
Online 24/7: yes
Payment routing: maybe
Multiple nodes: maybe
The Employer is the next user type. Employers typically make large outgoing payments (through direct channels or via an Exchange) such as for rent, utilities, supplies (perhaps to a Merchant) and paying employees (Consumers and Enthusiasts).
The Employer is very similar to the Merchant and the Exchange, except the Employer will make large purchases on the Network, either streamed in small amounts over time or in large chunks. This means Employers (if they themselves do not accept Lightning payments for their goods/services) will likely have a net outflow of funds. If they do accept Lightning payments, they will operate in a very similar manner to Exchanges, only with the addition of making regular payments through their channels.
Employers can work with Watchtowers to bridge the gap between how much time they intend to dedicate to managing their Lightning services and how much liquidity they are comfortable locking up on the Network. Some may not need their services and will be self-managed, and others may not want the headache of managing their software even though they have the resources to.
Employers may route payments, depending again on their level of interest in managing their node.
The next two users, Watchtowers and Enthusiasts are not well demarcated. Consider them as sort of an optional flag to be enabled, or a different flavor of an already existing user type.
Example individual: Lightning Security Services, Inc., Anti-fraud Dept at Gemini
Expected average channel size: medium ($100-$10,000)
Expected relative channel number: small (dozens to hundreds)
Expected relative channel distribution: Mostly remote balances
Channel connection points: to Exchanges, Merchants, Enthusiasts, Employers, Consumers and Watchtowers
Online 24/7: yes
Payment routing: maybe
Multiple nodes: yes, probably multiple Internet connections
The Watchtower is probably the most interesting Lightning Network user. This is a user who watches the Lightning Network for fraud, and scoops up the funds of fraudsters, while also providing services to any other user type to recover their funds in case of fraud. The Watchtower has a financial incentive to be well-connected, highly available, highly technically capable and specialized. Watchtowers share common ground with Enthusiasts (below) and Merchants in that they can route payments and accept payments of their own, but a Watchtower can monitor the network with 0 channels, if they wish, though they will likely have open channels to take advantage of Lightning's low transaction fees anyway.
The Watchtower will actively interrogate the Network, connecting to every node possible and ensuring they are the best informed user as to the state of the Network. They may publish data on the Network as a public service, or provide it for a fee.
Any user can negotiate with a Watchtower to watch their funds, either outside of the Network or in a decentralized manner in a configuration as yet to be determined, and provide the Watchtower with the most up-to-date state of their transactions. Then, if the Watchtower sees someone trying to broadcast a previous state to the Bitcoin blockchain, they can sweep the funds themselves and return them to their owner, if they can sweep them in time.
The main difference between Watchtowers and enthusiasts will likely be resources available at their disposal, and goals for their usage of the network. A Watchtower will primarily be looking to "snipe" funds, and may route them as a side-benefit.
Watchtowers may also be one of the few user types to utilize "hidden" channels, so their routes don't show up themselves on network graphs. Since they are well informed, routing Watchtowers can manually create guaranteed fee rates and uptime that Enthusiasts or Exchanges may not want to negotiate. They can also create channels themselves that exploit network inefficiencies using the data they collect, should the decide to take on the additional responsibility of routing payments.
Example individual: Satoshi, Blockstream
Expected average channel size: medium ($100-$10,000)
Expected relative channel number: medium to large (hundreds to thousands)
Expected relative channel distribution: Either mostly local or mostly remote, with not a whole lot of 50/50
Channel connection points: to Exchanges, Merchants, Enthusiasts, Employers, Consumers
Online 24/7: yes
Payment routing: yes
Multiple nodes: maybe
Enthusiasts are the last user type. As I write this, there are 450 Lightning nodes on the main Bitcoin network, and these are virtually entirely Enthusiast nodes. Like I explained above, like a Watchtower an Enthusiast can also be any other type of user, however the primary difference between Enthusiasts and other user types is their motivation.
Enthusiasts may have any number of reasons for running a Lightning node. Primary among them is likely to have a very small (very, very small) stream of additional income by locking up liquidity in channels on their node. Enthusiasts may lock up a few hundred dollars or potentially tens of thousands of dollars in their channels, especially if they were an early adopter of Bitcoin. Enthusiasts are largely programmers, "true believers" and ideologists who understand that by participating, they make the network more valuable.
Enthusiasts may, like Watchtowers who choose to route payments, look for inefficiencies in the network and open channels to exploit them, using community tools or tools they develop themselves. Enthusiasts will directly counter the centralization of Exchanges. Should an Exchange start abusing their position, an Enthusiast with sufficient liquidity will gladly offer it up and bypass the Exchange's dominance in exchange for the fees one could collect on such a large channel, just slightly below whatever the Exchange was charging.
They might also keep the network honest by performing Watchtower-like services, but what they do with the funds will likely be altruistic, such as returning them for free (with proof of ownership and mediation) or by donating funds they snag to Bitcoin-related causes, further disincentivizing cheating.
Enthusiasts may also onboard their friends, family, employers who are not Lightning-aware and give them exposure or offer their services in a direct and ad-hoc manner to "bridge the gap" between fiat or normal Bitcoin transactions and the Lightning Network. Developers of Lightning Network implementations and secondary software enriching the primary implementations are also Enthusiasts. Some Enthusiasts may not even run a Mainnet node, and instead operate solely on Testnet troubleshooting, developing and teaching others.
Enthusiasts will also stress-test the network in every way possible. White-hat and Black-that hackers in the space will both be considered Enthusiasts, and possibly Watchtowers, though Watchtowers will need to maintain a reputation as trustworthy.
In summary, each of these users will have different yet distinct use cases and objectives in the Lightning Network, and it remains to be seen if my predictions of dominance hold true especially over a longer timeframe. I expect the network won't look like what I'm describing for at least a year or two from now, and only once all three current implementations, (Eclair, LND and c-lightning) are mainnet-ready.
I wanted to go into more detail regarding strategies each user will take, such as automatic management vs manual, and how groups of users will coordinate but unfortunately this ended up way longer than I planned for.
I appreciate any and all feedback and input.