A preliminary study on the Filecoin mining pool allocation model and revenue calculation

Filecoin mainnet has been online for nearly a month and a half, and the popularity continues to rise. There is a strong demand for computing power growth, which can be seen from the gas baseFee last week. The baseFee of 1.0~2.0 nanoFil basically declares that the computing power growth capacity of the Filecoin network has reached the limit of its network support. Perhaps after v1.2.0 is launched, the situation will be slightly relieved, but don't expect the gas fee to be reduced. According to the current market enthusiasm, what you may see is that the daily computing power growth rate has increased a bit, and the gas fee will still remain high.

The gas fee continues to be high and sideways, which requires miners and mining pools to automatically adjust. There is essentially no other way. Because the blockchain always has a TPS limit, and for the security of the network, TPS should maintain a reasonable upper limit. This part is not the focus of today's discussion.

Today we are going to talk about another issue: Since Filecoin has just been launched, although everyone can understand the mining pool operation model, the theoretical preparation for some of the revenue calculations and profit distribution is insufficient. Investors and mining pool operators need an adaptation period, and the market needs an education period and understanding process. At the same time, since Filecoin's economic model is very complex, it brings some difficulties to everyone's learning and understanding. But the good news is that we have seen that many investors are very interested in this area and hope to learn more. This article makes some preliminary explorations, hoping to throw out bricks and jade, arouse more discussions, and finally lay a theoretical foundation and form a consensus in the industry.

Bitcoin, Ethereum and other mining pools have been running for many years, and their revenue calculation and distribution models are relatively mature. These have some reference value for the Filecoin network. Here we take Bitcoin as an example. There are many distribution models for Bitcoin mining pools, all of which have very clear definitions.

Roughly speaking, the available modes are:

• PROP (proportional)

• FPPS (Full Pay Per Share)

• SMPPS (Shared Maximum Pay Per Share)

• ESMPPS (Equalized Shared Maximum Pay Per Share)

• CPPSRB (Capped Pay Per Share with Recent Backpay)

• PPS (Pay Per Share)

• PPLNS (Pay Per Last N Share) and lastly

• PPS+ (Pay Per Share Plus)

There seem to be many, but there are actually two main types that are widely adopted:

1. PPS (Pay Per Share) : It means to distribute benefits according to shares. The shares mentioned here are not the shares you invest in a mining pool, but the share of computing power you occupy in the entire network. If the entire network produces a block: 6.25 BTC, if you account for one millionth of the total computing power of the network, then you will get one millionth of this block (of course, the mining pool will charge a service fee). This distribution model has the following characteristics:

1. No matter what the income of the mining pool you participate in is, your income will not be affected and is deterministic.

2. The mining pool will bear a greater risk, because even if the mining pool has problems, it still needs to allocate funds to customers.

3. Because mining pools bear greater risks, the proportion of service fees will be higher than other distribution methods.

4. This method will promote the mining pool to provide good services, improve the stability and performance of the mining pool, luck value, etc., because the extra money earned belongs to the mining pool.

   
   

2. PPLNS (Pay Per Last N Share) : It is to distribute according to the actual income of the mining pool and the proportion of shares owned by investors at the distribution stage. For example, it is distributed once a day and settled at 0:00 every day, but you bought it at 12:00 noon, then the share you bought can only be counted as half in the first distribution. No problem. This distribution method has the following characteristics:

1. Your income is directly related to the operation of the mining pool. If the mining pool is not running well today and there is a power outage, then sorry, the distribution will be less. But if the mining pool is lucky today, then your income will increase accordingly.

2. The risk of mining pools is relatively low, and there will be no over-allocation;

3. Since the mining pool has a relatively small wind direction, the service fee can be relatively lower than PPS

4. When customers choose a mining pool, the luck value is very important. To see the profitability of a mining pool, you can observe the luck value over a period of time as a reference.

The current status of the operation and distribution of each Filecoin mining pool is relatively simple. Basically, the actual income of the mining pool is distributed according to the investment share. This model is similar to the PPLNS model. At the same time, due to the complexity of the Filecoin network, the current mining pool has basically not formed a model of purchasing and participating in the distribution at any time. Basically, it takes some time for the corresponding mining pool to start operation and distribution after purchase, so there is no problem of Last N Share, which is N Share. Therefore, it can be abbreviated as PPNS. The difference between this distribution method and Bitcoin's PPLNS is:

1. The user's mining pool share basically does not change, and continuous calculation is relatively easy

2. Mining pool income is a key factor. Due to the mortgage mechanism and lock-up issues, the calculation of mining pool income is much more complicated than Bitcoin. For more information on income calculation, please see below.

3. Due to the lock-up issue, the mining pool income and the immediate distribution are two different things, and a large part of the income will be distributed in a delayed manner; please refer to the discussion below for this part

One thing that needs to be explained here is that there is no direct relationship between the user's mining pool share and the computing power share. This is because the computing power of the Filecoin network is gradually increasing. Judging from the current market situation, the total computing power sold and the computing power to be invested in the entire network is an unknown number, and it continues to grow. Basically, it is not a problem to reach a total of 5EiB within a year. Because each mining pool will allocate according to the computing power sold, users can roughly estimate the total computing power sold and their own computing power share in the entire network. In addition, I don’t know how much computing power the entire network has sold. If anyone has a better algorithm, please leave a message.

The PPLNS / PPNS model is more complicated than the PPS model. For mining pools, they have to bear the pressure of mining pool efficiency. If the efficiency is low, users will not be satisfied. This user sentiment has already appeared. And it is possible that users will not be very satisfied no matter what. Why? Because Filecoin's economic model is very complicated and there are many things to calculate. Users often do not perform very detailed calculations, so they may always feel that something is wrong.

Then the PPS model is much simpler, because PPS refers to the average income of the entire network. As long as the formula is out, everyone will calculate the same, and it can be done in the browser. If everyone has no objection, just follow the contract.

But why is no one using the PPS model? Or is it because Filecoin is too complex to be mature at this stage? One of the prerequisites of PPS is that the purchase of computing power takes effect immediately, which is rarely done by Filecoin mining pool operators at this stage. However, due to the simplicity and directness of the model, it is expected that the adoption of this model will soon be considered by mining pool service providers.

Whether it is the PPS model or the PPNS model, there is a revenue issue. Filecoin's revenue is more complicated than Bitcoin's. On the one hand, it is because the deserved revenue and the available revenue are different. On the other hand, miners always have a lot of consumption, especially when the baseFee is uncertain, the consumption may be different.

The PPS income is the average of the income of the entire network, so the calculation is relatively simple. Its 24-hour income can refer to the following formula:

Total network revenue = 24-hour FIL output - 24-hour FIL destruction

So,

The income of a single T is equal to: the income of the entire network / the total computing power of the entire network (T)

Note: The amount of FIL destroyed must be calculated here, because judging from the current network operation status, when the baseFee is high, the amount of destruction may account for a relatively large proportion, such as close to 30%.

Due to the characteristics of blockchain, all transactions are transparent and replayable, so the revenue and operation status of each mining pool cannot be concealed. However, it is still quite troublesome to calculate. Here is a simple method to calculate according to a single miner address. A miner has three roles:

• Worker

• Miner

• Controller (often requires an address to do WindowedPoSt alone)

Considering these three as a whole, this is the basic unit of a miner (pool). You only need to calculate the fund flow relationship between these three addresses.

remember:

• Balance(h) is the sum of the balances of these three addresses at height h

• TransOut(h1_h2) is the sum of all transfers from the three addresses between heights h1 and h2 (transfers result in a decrease in balance, which should be counted as income). Transfers between these three addresses are not counted.

• TransIn(h1_h2) is the sum of all transfers to the three addresses between heights h1 and h2 (transfers increase the balance and should be counted as loans)

Then the income of this miner (pool) between height h1 and height h2 is:

Profit(h1_h2) = Balance(h2) + TransOut(h1_h2)

- Balance(h1) - TransIn(h1_h2)

Tip: Whether the Owner address should be included in the calculation of the mining pool's revenue is actually irrelevant. If it is included, the Owner is in the internal address of the mining pool and is included in the balance calculation, and the rules for transfers in and out will change accordingly. There is no problem with this. For example, if the Owner is included, the transfer from the Owner to the internal Worker is not considered a transfer in, and the Owner's withdrawal of coins is not considered a transfer out.

Filecoin user income or mining pool income is not equal to available income. Part of it may be used for mortgage (if the computing power continues to grow), and another part may be locked and released later. Some of the available income is also released from the previous lock-up, which makes the problem very complicated.

Due to the existence of the mortgage mechanism, the growth of computing power requires lock-up. However, the amount of lock-up cannot be queried through messages. All of this is calculated internally. At the same time, block rewards cannot be queried through messages, and there is no relevant data in the block header, which brings a lot of trouble to bookkeeping. Fortunately, some current browsers have the function of real-time recording. The specific block reward amount can be queried through the browser (if the browser calculation is correct), and it can also be compared through different browsers.

At the same time, the amount of locked positions (including mortgage and reward locked positions) can be directly queried through the chain status. Combined with the profit calculation method mentioned above, we have a simple formula:

Total release on the day = total income on the day - total lock-up increment on the day

The daily allocable will come from the daily release. However, another aspect must be considered. If the mining pool wants to continue to increase its computing power, it needs to leave enough funds for the mining pool to increase its computing power, so not all the allocable parts can be allocated. Of course, it is also necessary to leave a margin part required to maintain computing power and maintain the stability of the mining pool. When the mining pool continues to operate at a certain rhythm, this part of the demand is generally relatively constant. Of course, another very important influencing factor is the gas fee mentioned at the beginning of the article.

In short, the calculation and distribution of Filecoin's revenue is relatively complicated, but there are simple ways to deal with it. Both mining pool service providers and customers should actively look for simple, effective and easy-to-understand ways to do it. I hope this article can inspire some investors, and welcome more discussions and better theoretical summaries to promote the healthy development of the ecosystem.