Data on the work in the Lightning Network from the operator of the fourth largest node

Report on the experience of managing your own lightning nodes from River Financial, which has been participating in the Lightning Network since 2019.

River Financial - financial and technologicala company that assists in the accumulation and use of bitcoin. We offer brokerage services, cloud mining, automated BTC uniform purchase (dollar value averaging tactic) without commissions, and deposit and withdrawal transactions via the Lightning Network.

We have been participating in the Lightning Network since 2019.By 2022, our node has become the fourth in the network in terms of BTC capacity. We now have a team of four full-time developers who are busy maintaining and updating not only our projects in the Lightning Network, but also the Lightning infrastructure of the Chivo wallet for the Republic of El Salvador.

From the experience of supporting and managing lightning nodeshigh capacity and with hundreds of channels, we learned a lot about the Lightning Network (LN). In this report, we have collected the knowledge gained about various aspects of working in LN and grouped them into the following categories:

Lightning Network Topology →
- There is no optimal profile →
- Capacitance is not the main thing, but it is a signal →
- Number of channels as a signal →
- The cost of opening channels →
- Number of channels and their quality →
- Commissions and Liquidity →
Routing in the Lightning Network →
- When lightning transactions are made →
- The discrepancy between activity and profitability →
- Why lightning payments may fail →
Yield in LN →
LN infrastructure →
- Lightning infrastructure for business →
Conclusion →

Topology Lightning Network

The Lightning Network is made up of nodes managed byvarious users and companies. These nodes, communicating with each other through two-way channels, form a huge interconnected network covering the entire globe.

Connecting «highways» this networkare routing nodes that, for a commission, send other people's payments through their channels. Anyone can create such a node, but for its successful and profitable operation it will require much more time and effort, which we will discuss later. Non-routing nodes with multiple connections (links) are the blue dots on the edges of the trunks; they belong to users who send and receive payments primarily for personal reasons.

As with Bitcoin, people may use Lightning for a variety of reasons. The goals of running your own node vary, and five types of profiles can be distinguished among node operators.

Profile type	goal
cost	Reduced commission costs for the user or company
Business	Enabling customers to purchase goods and services that benefit from instant global settlements and/or microtransactions.
Profitable	Earning BTC returns with very low counterparty risk by routing payments to other users
Trader	Allows traders to participate in arbitrage without using stablecoins

Each of these goals influences which tasksface the node operator, what are his incentives and therefore on the configuration of his node and channels. Based on these goals, you can answer questions about the required capacity, the number of channels, connections, commissions and rates in different ways, as well as whether outgoing or incoming financial traffic is a priority for the node owner. Any node can correspond to several categories at the same time.

There is no optimal profile

Some may consider «optimal»nodes with the highest return on capacity, but depending on the goals and type of profile, there may be other indicators of success that are more important to the node owner. An exchange, for example, may make far more profit from trades on its platform than from charging a user a fee for routing transactions in Lightning, and therefore may not optimize its nodes for the greatest profitability, while an individual interested in extracting return on invested capital will most likely try to do this.

We look at the publicly available node metrics (capacity, channels, fees, and liquidity) and what we've learned about them in relation to their goals.

Capacity is not the main thing, but it is a signal

As of 2022, our River Financial 1 node is the fourth largest Lightning node in terms of capacity, closely followed by our second node.

Nodes with maximum capacity - this is not necessarythe best nodes to connect to. If the node has a lot of capital, that's great, but it still needs to be used constructively. High capacity alone does not guarantee that a node will route payments with the highest success rate and the most competitive fees. For example, a high-capacity node might be owned by an amateur who accumulated a lot of bitcoin early on but put relatively little effort into managing liquidity and channels.

High node capacity really increasesthe likelihood that we are talking about a business node. Unlike users, companies have much stronger incentives to use their funds efficiently, which is less likely to happen when allocating capital to a channel with a private user without a track record. On top of these opportunity costs, there are also security costs associated with storing capital outside of a cold wallet, as occurs when using Lightning.

However, for private node operators it is easieraccess most of the network by having a link with the business node. The easiest way to create a connection with a Business Node is to open a channel with it and buy something from it with a fraction of the funds.

Number of channels as a signal

The number of open channels can be one morea sign of useful nodes. Although this does not guarantee that the node will work well, paying commissions or other expenses for opening many channels suggests that the owner has a certain interest in this.

In terms of the number of open channels, our (River Financial) node does not come close to the top 10. We are in the fourth ten.

Relatively few channelsour conscious policy. We have no ambition to be the most connected node in the network. We are more focused on ensuring the high reliability of each of our connections. We will return to this topic when we talk about payment routing.

The cost of opening channels

To open each channel you need to paybitcoin transaction fee. In 2022, the average Bitcoin transaction fee was 7728 Satoshi, or about $1.30 at the exchange rate at time of writing. Since a transaction to open a lightning channel takes up slightly more space in the block, the transaction fee for such a transaction is usually slightly higher than average, unless the node operator chooses to wait longer for it to complete. In thousands of on-chain transactions for our lightning channels, the average fee was 3905 satoshi, or ~$0.64 at current prices.

If the purpose of opening a node is to getprofits, then opening tens or hundreds of channels will quickly lead to an increase in costs that will be difficult to recoup if it turns out that open channels are inactive. In the future, if bitcoin transaction fees rise, creating new successful routing nodes may become more costly.

Number of channels and their quality

Every node operator wants good peers.This is a big challenge for carriers who want to earn revenue from routing other people's payments, and for users who want to save on transaction fees when using the Lightning Network. Node operators need to convince other operators to allocate capital to create a channel with them in order to have incoming liquidity. If the node is new, then there is still no idea of the potential benefit from blocking capital in the channel with it.

As a result, marketplaces for buying andsales of incoming liquidity for LN channels. This is not ideal because buying inbound liquidity does not always result in active routing, so the process may need to be repeated multiple times. These marketplaces provide people with a choice, which is a welcome development.

Node operators seeking to obtainprofitability, often close outgoing channels that have been inactive for a long time and redistribute the capital blocked in them. Since 2022, we have begun to manage our channels much more actively - so much so that today we open and close dozens of channels daily.

We have no doubt that over time suchactions, such as setting the size of commissions for opening channels or closing inactive channels, will be automated, which will significantly increase the convenience of managing lightning nodes. But even so, we expect that the creation of a successful routing node will require a significant time investment.

Commissions and Liquidity

The concepts of incoming and outgoing liquidity can be confusing, so let's look at them here:

Incoming liquidityis the number of bitcoins that can bereceive a node, or value, located on the other side of the channels connected to it. The node has no control over the fees charged for incoming liquidity and does not make money on incoming transactions.

According to Lightning developer Joost Jager, the lack of inbound routing fees could be a potential limiting factor for LN efficiency.

Example: Alice uses her node to open channelswith Bob and Carol with a capacity of 1 BTC each. Incoming liquidity is 0 BTC until she sends 0.5 BTC to Bob and 0.2 BTC to Carol. The incoming liquidity is now 0.7 BTC, which is the sum of the value on the other side of each channel.

Outgoing liquidityis the number of bitcoins a node cansend, that is, the value located on the side of this node in the channels connected to it. The node controls the fees charged for outgoing liquidity.

Example: using the previous example, Alice starts withoutgoing liquidity of 2 BTC. After she sent 0.5 BTC to Bob and 0.2 BTC to Carol, she has 1.3 BTC of outgoing liquidity, of which 0.5 BTC is in the channel with Bob and 0.8 BTC is in the channel with Carol.

And let's also take a quick look at commissions,because the Lightning Network has two types of fees: the base one, which is a flat fee charged for each transaction, and the fee rate, which depends on the size of the transaction. There is an LN zero base fee initiative that we support as it improves the efficiency of multi-route payments to increase the likelihood of success.

Below we look at liquidity fee rates for the top 10 nodes by channel capacity.

As we have already mentioned, receiving incomingliquidity is the hardest part for many LN participants. Looking at Bitfinex nodes, for example, their peers charge a relatively higher fee to send BTC to Bitfinex, while sending BTC from a Bitfinex node costs similar to other nodes.

Each node solves its own problems withliquidity. A user who uses the Lightning Network primarily for spending needs to replenish outgoing liquidity frequently. And a business that primarily accepts payments needs to replenish incoming liquidity, otherwise each client will have to open a channel with it directly.

Routing nodes want to have as much as possibleless capacity in inactive channels. You can build a formula to determine the percentage of actively used capacity and hence the efficiency of a routing node. Only the node operator can calculate this indicator, since information about node activity is not public. Example:

</p>

Over time, as more data becomes available, the datafrom many operators can be aggregated to analyze the range of expected activity across LN channels. Below we will make a first timid attempt in this direction.

Routing in Lightning Network

Number of routed by uslightning transactions is growing this year by ~5% per month, or 20% per month, if we take data only for the third quarter. It should be noted that this is not an indicator of the growth of the Lightning Network as a whole. As we have already said, we did not aim to maximize the connectivity of our nodes or the number of routable transactions. We tried to focus rather on the high reliability of payment routing.

What has changed significantly as a result isthe number of bitcoins routed by our lightning nodes. The relatively high success rate makes it more likely that other peers will route payments through our lightning nodes.

We cannot analyze the final recipientsthese transactions and how they proceeded. Thanks to the use of onion routing, or encapsulated encryption layers, a node cannot determine its position in the payment unless it is the final recipient. Its software can “peel back just one layer” to find instructions on which channel to route the payment through next.

When lightning transactions are made

What can be analyzed, for example, istimestamps of routed transactions. If we put the timestamps (in UTC) of 115,648 transactions routed by our lightning nodes in September 2022 into a weekly heat map, we get the following picture.

It can be immediately noted that on our routabletraffic is significantly impacted by activity in US time zones, as average sleep hours in America are marked by a significant decrease in routing activity. This is not surprising given that we have many lightning channels in the US and are the Lightning service provider for the El Salvadoran wallet Chivo.

If we adjust the heatmap for the hourlyUTC-6 zone, then we get the map view on the right. Here you can clearly see the night hours according to American time and, except for Sunday, the activity in the network is distributed fairly evenly.

If we could get data from other large routing nodes, we could both see the differences and get a better understanding of network activity in general.

The discrepancy between activity and profitability

Most popular days in terms of quantitytransactions are not always the most profitable in terms of commissions earned. The charts below, broken down by day of the week for August and September 2022, show quite different results (after adjusting the data for the number of each of the days of the week in those months). The second half of the week is usually relatively more profitable.

Why lightning payments may not go through

Today there is no guarantee of successlightning payments. There are many technical reasons why these payments can fail, such as nodes not being able to find a successful route due to the necessary channels being unavailable at a given time, or taking too long to find a route. A lot of effort has gone into the Lightning developer community to solve these problems.

User funds in case of payment failureare not lost, however this creates a poor user experience. To mitigate this, we, as previously mentioned, try to prioritize the quality of our channels. And so far, the results of work in this direction are promising: in September, the percentage of successful payments was 98.7% with thousands of daily transactions, the average transaction size was 230 thousand satoshi (~$40), and the median was 24.4 thousand satoshi (~$4) .

This is much better than the earliestthe public data we were able to find (which is from the early days of Lightning in 2018) where $5 transactions didn't reach their destination ~48% of the time.

We can examine the September data tounderstand why payments are not going through. Of the 1,467 payments we failed to route this month, 866 failures were due to timeouts and 558 due to no available route at that time. Other, less common causes were server outages for nodes (1 case), incorrect payment details provided (14 cases), or insufficient balance when liquidity ran out between the node's route calculation and payment sending (28 cases).

We expect that the percentage of outstanding payments will decrease over time due to the increased reliability of the Lightning Network due to the many efforts of developers.

Yield in LN

Since March 2022, by actively managing commission rates and expanding the number of peers, we have started earning millions of Satoshi from transaction fees. Growth was incredibly fast.

In September 2022, the average capacity of ourlightning nodes was ~420 BTC and we earned 40 million Satoshi in routing fees. This corresponds to a yield of 0.095%, which gives ~1.15% per annum. The rate is relatively low, but growing rapidly.

For comparison, here are the statistics @cold_sats,A Lightning enthusiast running a top 100 node with a total capacity of ~19.8 BTC. He reveals earnings and most of his strategies, which is commendable and makes his data useful for comparison. In March, he estimated his return for 2022 at 1.44% per annum, but if you take the data for September, this figure will be 0.55% per month, or 6.8% per annum. This is significantly higher than our figures, with the proviso that we have focused our efforts primarily on creating infrastructure.

There is also a plateau at whichincreasing capacity may no longer provide a linear increase in profit, depending on the current network activity and its size. Data from many lightning node operators would allow us to visualize this plateau and its evolution over time.

While a return of a few percent sounds great, it needs to be accompanied by a caveat about the need for more effort.

LN infrastructure

Launch a lightning node to send and receivepayments on LN are relatively simple. However, building a successful routing node from scratch, or integrating such a node with business systems, is much more labor intensive and has a steep learning curve. Each of these tasks is an independent project for several months, the implementation of which has to be done by trial and error. There's a great guide from @cold_sats, a Lightning enthusiast and a major node operator, that gives you an idea of all the tasks involved in setting up a lightning node from scratch.

To manage a large lightning node, you needat least a few hours a week. The main tasks are channel and liquidity management. Some time savings can be achieved through automation, but so far such solutions are custom-made and are not available in standard and widely available software. We expect that with further updates to Lightning, node management will continue to be simplified.

But even after such updates, we expect thatthe average user will not be able to run a highly profitable routing node. This is not a problem for the smooth operation of the Lightning Network itself, as the number of profitable nodes is a self-balancing mechanism. If there are too many operators trying to manage profitable nodes, then some of them will be weeded out over time and increase the profitability of the rest. If there are too few profitable nodes, then people will learn how to manage them more efficiently. Likewise, most content creators on platforms like YouTube or TikTok don't earn large sums, but the experience for both content creators and content consumers remains positive.

Likewise, we do not expect the average businessboth inside and outside of the bitcoin industry will want to hire people to set up and maintain the Lightning infrastructure. There are many more considerations when running Lightning nodes as a business.

Lightning infrastructure for business

As a provider of Lightning infrastructure services to other companies, we take on much broader responsibilities than are required of ordinary node operators.

Like any operator, we need to updatesoftware of our nodes. However, when our business partners rely on the system for their payments, the system should never go offline. And this means that we (and any other business that wants to run its own Lightning infrastructure) must create and maintain multiple nodes for unexpected system failures, software updates and maintenance. The company also needs alert systems to replenish incoming liquidity so that users can use Lightning to pay without interruption.

There are also many accountingconsiderations and necessary integrations with business systems for accurate reporting. And then there's the critical issue of security, since Lightning channel capacity is stored in a live system, no more secure than a «hot» wallet.

Ensuring that the system is always up and running is oneaspect, but when it doesn't work, the user needs to be properly informed about what happened and what action he can take. Sometimes lightning payments may still not go through or get stuck. due to unexpected liquidity depletion. As we already mentioned, we manage to process payments with a 98.7% success rate, but this also implies a 1.3% chance that users may encounter some kind of error and may need to try again. This is a high failure rate compared to other consumer payment systems, but Lightning is a relatively new network and we expect our successful payment processing rate to increase in the coming years.

Conclusion

Manage your own Lightning infrastructurestill challenging and interesting. And with the advent of Taro, we envision a bright future for the Lightning Network, as it has the potential to completely change the idea of what is possible with Bitcoin.

Despite this, we came to the conclusion that forFor many companies, it does not make sense to allocate resources to launch and manage lightning nodes on their own. There are many unique challenges that businesses need to address in order to use Lightning at significant scale, especially those related to reporting, security, and maintenance. As a result, many companies may find it beneficial to outsource such responsibilities.

As for River Financial, our mission with Lightning is simple: to be the best peer on the web.

BitNews disclaim responsibility for anyinvestment recommendations that may be contained in this article. All the opinions expressed express exclusively the personal opinions of the author and the respondents. Any actions related to investments and trading on crypto markets involve the risk of losing the invested funds. Based on the data provided, you make investment decisions in a balanced, responsible manner and at your own risk.