Lightning Network: Routing Cost Economics

Foreword: Lightning Network is still in the initial stage of development and needs to solve many problems. One of the controversial areas is the incentive problem with routing nodes. Lightning network not only guarantees the user's low-cost payment experience, but also realizes sufficient income of the node. Otherwise, it is difficult to attract the profit-oriented node to enter. How to find the balance between the profit rate of locked funds and the low cost of users is lightning network. Need to be solved.

Abstract: BitMEX studies the market dynamics of lightning network routing costs, and also studies the financial incentives for providing mobility network node operators with liquidity. We believe that there is a correlation and balance between the cost of lightning network routing and the return on investment of the nodes that provide mobility to the channel. This is the main challenge of Lightning Networks, and the computer science aspect of routing problems is not a major challenge. Our conclusion is that if the lightning network expands, at least in theory, in the broader financial markets, such as interest rate changes and investor sentiment will have an impact on the lightning network cost market. However, regardless of the current economic situation, in the long run, competition will be a key driver of price. A low barrier to entry into the market may mean that the balance is beneficial to users and low costs, and is not conducive to the return on investment of liquidity providers.

(Lightning crossed the night sky of Singapore) (Pexels)

Overview

I have written a lightning network before, and today, as the lightning network moves from theory to landing, it is necessary to take another look. This paper analyzes lightning networks from the perspective of finance and investment, especially the incentives for cost and lightning network providers. This article does not cover technical aspects.

Routing problem

Comments on Lightning Networks often see routing as a major issue, often as a "computer science problem that has not yet been resolved." In general, we don't really agree with this, and we don't think computer science about routing is a major challenge. Finding the path between channels for payment may be relatively clear and similar to other P2P networks, such as Bitcoin.

However, we believe that its main challenge is the interaction and balance between financial and economic liquidity provision and payment routing. Lightning network node operators need to be motivated by routing costs to provide sufficient liquidity to achieve a smooth payment. Mobility needs to be specifically assigned to channels with requirements, and identifying them is challenging, especially when new businesses enter the network.

There is a cost-balanced challenge to ensure that the network is low enough for the user, but on the other hand, it must be ensured that the cost is high enough to motivate the liquidity provider. The severity of this problem and the rate of market liquidation may depend on economic conditions.

Lightning network cost market dynamics

For bitcoin chain transactions, the user (or their wallet) assigns a fee for each transaction at the time of payment, and the miners select a higher transaction fee transaction to try to produce the block in order to maximize the fee income. In contrast, Lightning Networks currently seems to be the opposite path, with the node operator setting the fee, and then the user chooses the path to pay and selects the channel to minimize the cost.

That is to say, the lightning network is initially set by the node provider, not the user. Therefore, the lightning network can provide a superior cost structure. Because the provider provides specialized services, it is more appropriate to compete by the interest rate between the providers, instead of the ordinary users to compete for the cost, the priority is simplicity.

In lightning networks, node operators must specify two types of routing fees: basic fees, rates.

  • Basic fee: The fixed fee charged for each transaction through the route payment, expressed in one thousandth. For example, the base fee per transaction is 1,000, which means that the base fee for each transaction is 1 Cong.
  • Rate: refers to a certain percentage of the value of the payment. This is expressed in terms of one millionth of the transfer. For example, the rate of 1,000 is 1,000/1,000,000, that is, 0.1% is its fee. Once the transfer is successful, the routing channel will charge a transfer fee of 0.1%. Equivalent to 10bps.

Investment capital

In order to provide liquidity for routing payments and thereby earn fee income, Lightning Network node operators need to lock capital (bitcoin) in the payment channel.

Two types of channel capacity

  • Entry capacity (capacity to enter the node)

Inbound liquidity refers to funds in the node's payment channel and can be used to receive payments. These funds are owned by other participants within the Lightning Network. If the payment channel is closed, these funds will not be returned to the node operator.

There are two ways to create an in-node balance: 1. When another network participant opens a payment channel with that node 2. When the node operator pays through the existing channel

  • Outbound capacity (outflow node capacity)

Outbound liquidity is the payment of funds in the node's payment channel, which can be used for the outflow of funds. These funds are owned by the node operators and partly by their investment capital. The node operator may consider the opportunity cost of other investments while considering the total outbound balance. If the payment channel is closed, these funds will be returned to the node operator.

There are three ways to create an outbound balance: 1. When the node operator opens a payment channel with other network nodes; 2. When the node operator receives payment through the existing channel; 3. When the node routes the payment and receives the fee.

Graphical description of the entry and exit capacity of the channel (ie graphical description of the capacity of the incoming and outgoing nodes)

Source: Bitcoin Lightning Wallet, note: the orange balance is the inbound capacity (inflow node), and the blue balance is the outbound capacity (outbound node)

Lightning network fee market operation

 

Being a successful routing node is more difficult than people think. At the time of this writing, there are 7615 public lightning network nodes based on 1ml statistics. However, of these nodes, there may be only a few hundred that work well, and these nodes provide liquidity through node management, rebalance channels, and set appropriate fees.

The node operator may need to:

  • At the same time adjust the rate and base fee, monitor the impact of adjustments, and recalibrate to achieve the optimal income fee setting.
  • Analyze the network and look for lightning network nodes that have high payment requirements but are difficult to connect, such as new merchants
  • Analyze the cost network, not only for the entire network, but also for your high-demand, low-capacity routing
  • Continuously monitor and rebalance channels to ensure adequate two-way mobility
  • Perform a customized backup solution for the latest channel state to protect funds when a node computer crashes

Currently, no automation system can perform the above functions. If this does not change, you may need to set up a professional business to provide liquidity for the lightning network. However, just like liquidity, the challenge of overcoming these technical issues does not mean that payments can become difficult or expensive.

These technical challenges may simply be adjusting the equilibrium market rate. The harder it is to overcome these problems, the higher the potential return on investment for channel operators and the greater the motivation to solve them. This will be a need to drive the success of Lightning Networks, not the challenges faced by node operators.

In order for the lightning network cost market to work, the node operator may need to adjust the cost based on the competition, which can be based on algorithm or manual processing, in order to achieve the maximum fee income. To mimic what might eventually become standard practice, the BitMEX study attempts to modify the rate on the node within three months, as shown in the next section.

Rate experiment

This experiment attempts to assess the status of the charging market. Although the lightning network is currently in its infancy. We set up lightning network nodes and periodically change rates, trying to determine what rates would maximize revenue revenue, just as network operators eventually hope to do as the network expands.

The basic analysis we performed from a node (not strictly scientific) is illustrated in the scatter plot below. It seems to indicate that the expense rate currently has an impact on the cost of revenue for lightning network nodes. When the expense rate is increased from 0 to approximately 0.1 bps, the daily fee income will increase rapidly. Once the cost increases beyond this ratio, the average daily fee income seems to gradually decrease.

Therefore, based on this experiment, it seems that the maximum rate of income is around 0.1bps, which is definitely very low compared to other payment systems. However, this is only a one-hop fee and the payment may be more hopping. At the same time, the lightning network cost market is almost non-existent. In fact, only a few lightning network nodes are trying to maximize the economic benefits through cost adjustment.

Once the network is expanded, other participants try to maximize revenue, and the cost market conditions will be completely different. Therefore, this work can only be seen as an illustrative experiment, rather than a special disclosure of the lightning network cost market.

Lightning network node daily fee income Vs rate

(Source: BitMEX Research) (Lightning Network Expense Revenue Data Chart – Notes and Additional Notes)

*Day data time period: from December 31, 2018 to March 24, 2019

*Data comes from a lightning network node

*The basic cost for the entire period is 0

*The return on investment data does not include transaction costs on the Bitcoin chain. When the impact is included, the optimal cost rate bar chart shows a negative return on investment.

*Data includes weekdays and weekends. Generally speaking, the lightning network's network traffic during the weekend is significantly lower than the working day.

* The rate is changed daily at 21:00 UTC. The rate decreases daily, then rises to the top of the expense range after a few days of decline, and begins the next rate down cycle. The reason is that some wallets (such as mobile wallets) do not always ask for fees when they try to route payments through nodes, so many payments will fail when the expense rate increases. For example, when opening a channel from a mobile wallet to a node, then increasing the rate and attempting to pay immediately, the payment often fails because the wallet attempts to pay at a low cost. In our view, in order for the lightning network cost market to work, the node operator may need to change the fee on a regular basis, so the wallet may need to query the expense rate more frequently.

* Manually process channels to rebalance, once every two weeks. It takes about 30 minutes each time

*Lightning network node runs on LND, software is updated every two weeks

* Use automatic operation to open approximately 30% of the channel (by value), and the remaining 70% for manual operation

* The return on investment is calculated by taking the channel outflow capacity of the daily network, calculating the annualized rate of return on investment based on the daily fee income, and then calculating a simple average of all dates based on the rate within a specific range.

*Data is based only on one node and a specific channel set, and the experience of other node operators may be very different

* We tried to use the public node for this experiment. However, the fee income was too sporadic, and some network participants paid far more than the cost of advertising, which made the data less reliable.

*Unfortunately, we need to use a logarithmic scale for both axes. Considering that we are not sure what rate is appropriate, or even what magnitude is appropriate, we have tried a wide range of expense rates, from 0.0001% to 0.5%, so it is appropriate to use a logarithmic scale. At the same time, daily fee income is highly volatile, ranging from 0 to more than 3,000. Therefore, it is best to use a logarithmic scale to indicate. As the network evolves and becomes more reliable, cost-market intelligence is improved, and linear scales may be more appropriate.

Fee income and return on investment

In addition to the daily fee income, you can also consider the annualized investment income associated with running a lightning node and various rates. This is calculated by annualizing the daily income and dividing the number by the liquidity of the daily node outflow.

The highest annualized investment yield in this experiment was 2.75%, while the highest cost column investment return was almost 1%. This seems to be a reasonable rate of return because it is theoretically a relatively low-risk investment, at least once the ability to back up the lightning path in real time. Existing Bitcoin investors may be attracted by such rates of return and provide liquidity to Lightning Networks, or dollar investors will buy Bitcoin, use leverage to hedge bitcoin price exposures, and try to earn Bitcoin Lightning Network fee income.

Calculate the annualized investment income of lightning network nodes according to the cost column chart

(Source: BitMEX Research)

Of course, current liquidity providers in lightning networks are less likely to be driven by return on investment. Current node providers are more like amateurs, and most node providers have lost money in view of the chain costs required to open and rebalance lightning network channels. Although hobby-based liquidity can sustain the network for a while, in order to achieve the massive scale that lightning network expects, more investors need to attract potential investment returns.

Lightning network costs and economic conditions

In the current low-yield environment, 1% of investment income seems attractive, but lightning network to attract suitable commercial liquidity providers may have difficulties from the beginning. Investors in this area are looking for high-risk and high-return investments, which seem to be exactly the opposite of what Lightning Networks offers because it provides liquidity providers with relatively low-risk and low-return investment returns. Therefore, a new type of investor may be needed to suit this portrait.

If the lightning network reaches a large scale, it is likely to become a highly liquid investment product with a stable low risk return, which is sensitive to economic conditions.

Consider the following scenario:

  • Federal Reserve benchmark interest rate is 1.0%
  • Lightning network node operators typically earn 1.5% of annualized investment income on their outbound balance (node ​​outflow)
  • Federal Reserve Board Open Market Committee raised interest rates from 1% to 3% due to strong economic conditions and inflationary pressures
  • Because its return on investment is more attractive, Lightning Network node operators withdraw capital from the lightning network and buy government bonds.
  • Due to the lower mobility in the lightning network, users are forced to pay higher fees for routing payments, and lightning networks become more expensive

However, if the lightning network is large enough to be able to use the above logic, then the lightning network will be a huge success anyway.

Risk free return

In a sense, if the lightning network is mature, one can consider the return on investment running the lightning network node as the risk-free rate of bitcoin, or at least the rate of return without credit risk. In traditional finance, this is usually the interest rate that investors receive by holding government bonds. The government has a legal obligation to pay the principal and coupons, and it is also the holder of new money to pay the bonds. These risks are almost zero.

In theory, all other investment projects or loans in the economy should have a higher rate of return than this risk-free rate. The same is true for Bitcoin, where the return rate of a lightning network node liquidity provider can be seen as the benchmark interest rate within the Bitcoin ecosystem.

In the future, if most of the technical challenges involved in running nodes are overcome and there are competitive cost setting algorithms, then the risk-free interest rate of Lightning Networks can ultimately be determined as follows:

  • Conditions for a broader financial market – higher interest rates may mean higher risk-free rates for lightning networks
  • Demand for lightning network transactions – more demand or higher capital flows, should increase the risk-free rate of Lightning Networks

in conclusion

Whether professional hedge funds and venture capital investors will have the same enthusiasm for becoming a lightning network liquidity provider, as they did in PoS's “equity-as-a-service” business model in mid-2018, remains to be seen. While the return on investment provided by Lightning Network Mobility providers does not seem compelling, we do see the potential benefits of this business model as the network is in the formative stage.

In our view, lightning-only networks can easily scale to many times the size of transactions on the current bitcoin chain, purely based on a hobby-based liquidity provider, without encountering any economic expense market cycles or problems.

However, if Lightning Networks wants to reach the scale that many supporters expect, it will need liquidity support for profit-hungry investors who seek to maximize risk-adjusted investment returns. If this happens, unfortunately, as the investment environment changes, the network may experience significant changes in the cost market conditions.

However, it is relatively easy to build a node that provides liquidity and attempts to earn income by weakening peers. In the operational channel of the operating node, where the balance between the degree of liquidity provision and the return on investment is finally balanced, we obviously do not know. However, if we have to guess, based on the architecture and design of the lightning network, we would say that the system is designed to be somewhat user-friendly and low-cost, rather than designed for liquidity providers.

The author of this article is BitMEX Research, translated by the "Blue Fox Notes" community "Sien".