Andy is a guest at the Human Mine Computing School to dismantle Filecoin's economic model and ecological construction

The Filecoin Space Race is in full swing, the mainnet is about to be launched, and the economic model coefficients are updated one after another. On September 16, the 4th issue of "Computing Power School" hosted by Anmai Cloud & XnMatrix and co-organized by Renren Mine IPFS Community & Lieyun Finance invited Andy, the founder of the Filecoin Beijing community, to be a guest in the live broadcast room to dismantle the Filecoin economic model and ecological construction, and help IPFS miners understand the FIL value ecology.

Guest profile: Andy, founder of Filecoin Beijing community, CEO of Open Source Mining Pool, co-author of the second part of "Blockchain+", former investment director of Node Capital, former head of project incubation of Huobi Labs. Filecoin Beijing is a non-profit community organization jointly initiated by early IPFS code contributors, investors and enthusiasts in Beijing, aiming to gather various people interested in Filecoin in Beijing and build the IPFS ecosystem.

RRMine/Renren Mine IPFS community has compiled the content record. For the convenience of reading, some deletions and modifications have been made without changing the original intention of the sharer. For more details, please check the live broadcast replay.

Andy:

Hello everyone, I am Andy, the initiator of Filecoin Beijing community, and also the CEO of the open source mining pool. I am very grateful for the invitation from XnMatrix, Renren Mine and Lieyun Finance. Today I would like to share with you the latest economic model of Filecoin.

Filecoin, a great social experiment

The economic model is a highly abstract mathematical model of the real complex world. It is hoped that by designing an accurate, complex and specific economic model, it can guide or simulate and predict some future economic development laws and trends. However, in the field of blockchain, it is very difficult to use a relatively complete economic model to build a real borderless business empire.

At present, the only successful ones are Bitcoin and Ethereum, which use some simple algorithms to implement a relatively simple economic model in some relatively simple fields. However, if blockchain wants to be implemented, it is definitely necessary to design a more complex economic model and a more complete set of code structures.

The current blockchain implementation is heading towards a deep water area and no man's land. The one that has gone the deepest and furthest in this field is the most popular Filecoin. It is a great social experiment, because few people can go so far in the implementation of economic models. Although it has not yet been launched, I think it is still worth looking forward to. And even if Filecoin fails, it will provide a lot of useful nutrients for later generations to learn from, so Filecoin's economic model is still very worthy of everyone's attention, whether from the perspective of investment, idealism or social practice, it is worth studying by people who work in blockchain, or even those who study economics.

The economic model makes the network run quickly

In a nutshell, the official hopes that the economic model can serve a healthy ecosystem, rather than being like some previous altcoins or Ponzi schemes, where everyone gets a piece of the pie and then ends up with nothing. It should be said that the Filecoin project has attracted nearly 50 to 60 billion yuan in funding so far, which is not much compared to traditional industries, but it is already a very big giant in the blockchain field. The official hopes that Filecoin can really be used, and it can really use storage to do effective proof of work and reduce unnecessary waste of electricity.

But in the early stages, it is indeed very similar to a POW Bitcoin mining, because it now requires very high computing power, and the time limit for storage and retrieval will have to wait until the main network for further observation.

First of all, the official has designed a benchmark casting model. The casting model does not simply give you rewards based on the packaging speed of the hardware. Instead, the official designed a baseline. If the baseline is not reached, even if you have mined the reward, it will not be released to you immediately. Baseline can be understood as a KPI. What does it mean specifically? For example, in the current space race, the computing power of the entire network must reach 200p before it can release 1.5 million global rewards. After the mainnet is launched, it will also use this model, which means encouraging everyone to collaborate, and then let the total computing power and storage capacity of the network reach a certain scale, so that the network can run quickly.

There are two main considerations for designing pledge

In fact, many miners are more concerned about the issue of staking. There are two points to consider when designing staking. In economics, it is called the principal-agent theory. There are two main roles in the principal-agent theory, one is called the owner, and the other is called the manager. When designing this model, the owner will consider two aspects of the manager. The first is how to get him to participate, and the second is how to maximize the benefits.

Therefore, it can be seen that the Filecoin team has studied economic theory. First of all, the pledge must be small enough, so that small miners do not have to exit the network and can still be profitable. However, the pledge must also be large. If it is too small, the security of the network will be threatened. Many large miners may enter the network to make money, which will eventually cause a great impact on the network and even cause the network to suffer a 51% attack.

From the perspective of miners, the authorities are asking you to spend money, but it is not targeting you alone, but it means that if you want the long-term development of the network, you must design this thing. Recently, everyone is quite angry about Meituan Takeaway and Ele.me Takeaway systems, but in fact, when designing the economic model, this model also makes miners very angry. In order to make the network realize long-term, the authorities must consider the trade-offs of various interests, and it must also consider the rights and interests of early investors, including considering the realization of their own idealism. After all, this network is not designed for miners. If it is designed purely for miners, the Filecoin project will eventually become a mining coin, and its value will be much smaller. It is precisely because the Protocol Labs has such an original intention and wants to make the network better, so more people, including many people outside the circle, come in to help the ecosystem develop and grow.

Consensus Staking and Storage Staking

In the pledge model, part of the pledge is called storage pledge, and part of the pledge is called consensus pledge. The design model of storage pledge is mainly based on future sector income, and what kind of punishment should I receive if the storage fails, data is lost, or the service with the customer is terminated. Therefore, the storage pledge mainly considers two points. One is your future sector income. If I can only get 10 coins for 1 sector, and you ask me to pledge 100 coins, this is definitely unreasonable, so your pledge limit will not exceed 10 coins.

On the other hand, if the system deducts my coins based on my sector errors, or based on my dishonesty, or based on my breach of contract, for example, it will deduct 5 or 8 coins at most, and it is definitely unreasonable for you to pledge 1 coin. So the initial storage pledge is based on this consideration.

There is also a consensus pledge, the main goal of which is to maintain the security of the network. For example, if the entire storage is 200p, you can contribute a proportion of pledge according to the percentage of your own computing power in the total network computing power. At present, the official coefficient is 30%, which means that 30% of the total circulation of the network is pledged as part of the network security.

In addition, the official mentioned the issue of sector quality. Different sectors or different storage have different qualities. How to evaluate quality? In fact, it mainly depends on the real user storage, because in the current miner competition, basically all miners are self-produced and self-sold, that is, I feed myself data, but from the perspective of the network, it does not increase the real value of the network. Everyone brushes data, just like Didi's early orders. There may be Didi drivers who drive the client and run for a day without doing anything, and finally get a lot of rewards.

Validator mechanism verifies the authenticity of node data

In fact, the quality of sectors is generally low, but how to verify the authenticity of the data is another dimension of the problem, which may also be an ultimate philosophical question. In fact, I will not discuss this issue too much here. In short, the official hopes that everyone will store real data, and it will calculate a coefficient based on the weight. Storing real data can get more rewards. If you store ordinary data or feed yourself data, you will get less rewards.

Then the second question will arise, that is, how do we verify that the data is real? The official has designed a validator mechanism, and many people are paying attention to the latest progress of the mechanism. However, at present, the official has not fully developed the verification node mechanism. The official has only put forward such a concept and idea. Of course, it will take a relatively long time to implement it. It is expected that it will have a clearer version after the main network is launched.

In the previous versions of AMA, the official proposed some ideas for verification nodes, such as it hopes to be distributed, and some reputable institutions can participate in the network. It wants to use the reputation of some traditional institutions to endorse the network and let them choose which data is real data. Of course, the official assumed such a role in the early days. In the Filecoin discover project, the official has an operation to sell hard disks. Selling hard disks is not particularly expensive. It is 265U, which is equivalent to about 1,800 yuan. It is not expensive, but the official did make everyone feel that there are some centralized elements in it through this behavior in the early days. That is to say, the official you decide what data is valid data, and this right must be decentralized. If you decide the rules of the game alone, the game will become a very centralized game.

So, when it comes to the latter part, everyone can look forward to how the validator nodes will be opened after the mainnet is launched. The validator nodes will become an object of competition among all mining machine manufacturers in two or three months, because at present, the competition is still in a relatively low-level dimension of computing power. If the official wants to really turn the network into a storage network that is comparable to Alibaba Cloud, in fact, in the future, when everyone looks for real data, it may bring a more important work and competition dimension.

The final parameters for reward unlocking are still to be determined

Regarding the estimation of the daily block reward, it was actually easier to estimate before the Baseline was announced. The coins released by the entire network every day were based on the number of sectors you packaged, and then divided by the number of packaged sectors in the entire network to get your share. But now there is a Baseline, which means that it is not necessarily that 100% of the coins that you should theoretically release will be released to you. The official parameter before was 20+180. That is, the coins I mined must first be locked for 20 days, and then released after 180 days. But now you can actually see in slack that miners and the official team are also in a game stage. Miners have to fight for their own rights and interests. In order to allow more people to participate, the protocol lab will also make some appropriate compromises. In the latest economic model, the 20-day lock-up period has been reduced. In addition, the failure fee, inspection fee and penalty of the miner's sector have been reduced.

Another question is how to calculate the minimum storage pledge? Storage pledge is mainly for the providers of storage orders. When we use the network, the pledge may be provided by two parties. One party is my miner, as the party providing the service, and the other party is the provider of my order, which is my user. In fact, users also need to deposit and provide some pledges, which means that if there is a breach of contract, both parties may breach the contract.

It also cites an incentive idea, that is, first of all, the pledge should provide the minimum service guarantee for the storage user's order. In addition, it also depends on how much resources the storage order actually consumes. Just like Ethereum, it will have a gas fee, so on the Filecoin network, it also requires users to provide a storage pledge, but there are still some tricks in storage pledge, which will be discussed later. The next step is blockchain reward unlocking. In fact, there is no very clear final parameter for reward unlocking at present, but as mentioned earlier, 20+180, including the Baseline, these three factors determine how your rewards are unlocked and released.

Of course, many people will consider the issue of initial pledged coins. In fact, the official also has many considerations for pledged coins. First of all, miners must pre-stake, otherwise the network will have a lot of computing power coming in. For example, the largest computing power may be traditional cloud service companies, such as Alibaba Cloud. If their computing power surges in, it will not be a good thing for current miners. So we don’t want to say that the less the pledge, the better. If the less the pledge, the better, the network will soon collapse.

Pre-pledge can be obtained from many places, such as exchanges, space races, and the official will set up its own mortgage lending platform after going online. Everyone can pledge digital assets to borrow the currency, and then to pledge. There may also be many third-party lending platforms.

Many customers will ask, what if the pledged coin I bought falls? In fact, it is the same as everyone's DEFI mining, right? You use Ethereum to pledge, and then a coin comes out, and then what if your Ethereum falls? In fact, the official has not made any promises about this, because the official cannot completely control the trend of the coin price. The coin price should be said to be a market game behavior, and the official will not bear the loss of the pledged coin.

Penalty Mechanism

Penalty may be an issue that miners care about. Penalty is actually a KPI that the network puts forward to miners. You must ensure that you are online 24 hours a day, 7 days a week, and you must be able to achieve "6 nines" in line with Alibaba Cloud. Then you must be able to quickly recover within 1 to 2 hours. If you can't do it, you will face corresponding penalties. For example, if your sector is wrong, you may have to pay a sector failure fee. Because the current test network uses robots to detect faults, there will be an opportunity every day for robots to detect your error sectors. If you can recover within one day, your computing power can also be restored, but if you conceal it after being detected, you may be punished for breach of trust.

After the mainnet is launched, there may not be enough resources to send robots to each miner. It mainly depends on everyone's self-consciousness. If you are not conscious, the network will find out and punish you. If a sector error lasts for more than 14 days, the miner must pay a sector termination fee, which means that I admit that I have breached the contract. If I cannot find the data after more than 14 days, the official believes that the miner has lost the data. However, the new version of the economic model just released by the official only mentions three items, which reduces the punishment for miners and the lock-up. The lock-up for the first 20 days is cancelled, but this model is still not a final model.

I just mentioned the issue of storage pledge. Storage pledge actually refers to whether users will be compensated if their data is lost. This question is actually a difficult one to answer. For example, if I am a miner and feed myself data, I am also a storage user. If my data is lost, can I get compensation? This is actually a difficult question to answer.

At present, the government will not set compensation, because the government cannot distinguish whether the storage user is a miner. If the miner is a storage user, there is a loophole for stealing the wool. That is, if I store data for myself and then delete my data, and then I get compensation, will my compensation be higher than the cost of the pledge? It's just like the wool stealing of new energy vehicles in 2014. Many gasoline cars then add a battery. It's like an electric car. The cost may be only 50,000 yuan, but the government will give me a subsidy of 70,000 yuan. So if there is such an unreasonable situation, someone will definitely steal the wool.

The economic model ensures that miners will not do evil

But at present, it is impossible to verify the difference between storage users and miners at the code level. However, the official also mentioned another point, that is, although miners do not provide you with compensation, because of the various KPI requirements for miners, you can understand it as a heavy tax, and miners will have a strong incentive to ensure the stability of their storage, that is, if miners lose the data, they will be punished in various ways.

Its pledge may be deducted, and it may even lose its mining qualification, so miners have a strong motivation to maintain the network. We can say with certainty that miners will not take the initiative to delete user data. Therefore, how to make the network more stable and fix the vulnerabilities more completely is a more serious issue that the authorities need to consider.

Another question is where the burned Filecoin goes. Just like Ethereum, the gas fee will be sent to an empty address and cannot be transferred back. If you want to transfer it back, you can only do it through a hard fork. In fact, from this perspective, the circulating disk will become smaller, which means that the entire network is a deflationary model. The supply of the entire network is decreasing, but its value is actually increasing.

The current network requires miners to stay online 24/7, including during the test network period, including Ammann Cloud or open source mining pools, or other miners. During the competition, miners are in a very tired state. Basically, everyone can only sleep for three or four hours a day, and everyone is in a shift, so there will always be someone online. If one of you is absent from work, or if one of you is offline and causes user data to be lost, you may face a more serious punishment. Because one of the things the official wants to do is decentralized cloud storage, in fact, the benchmark is still Alibaba Cloud, and Alibaba Cloud should be very reliable.

Why does the government still want to do this? In fact, it is not about competing on reliability and price, but about considering this from the perspective of a new underlying network structure, or from the perspective of security and fairness. Alibaba Cloud now has a reliability of 99.9999%. At present, the official KPI for miners is 80%. If you cannot achieve 80% reliability, you will not be eligible to receive space race rewards. This requirement will be higher after the main network is launched.

When will the final parameters be released?

The system will continue to improve. It can be seen that it has been three years since the white paper was released in 2017, and the parameters have been changing in these three years. Since the first version of the test network on December 11 last year, four or five versions have been tested so far, and each version will have major changes.

Even in the middle of a version, for example, since the large miner test started on August 25, the official has released 7 versions. You have to upgrade and deploy each version, and then find new problems. It can be said that miners are now exhausted. If you don’t have enough accumulation and a deep understanding of the code, you are in a state of firefighting every day. It is actually difficult to do this well. The system is constantly improving, and new problems will be discovered during the improvement process. At the beginning, the official may not have expected that so many Chinese miners would participate in this test, so it was a surprise to the official. After December 11, the main network that was originally expected to be launched in April was postponed to June, and then postponed to August. Now it should be October. Whether it will be 100% available in October, in fact, no one dares to guarantee it, because when you draw a bigger circle, you will encounter more problems. How to solve more problems is actually an open ecosystem.

For example, there was no Base fee before. No one thought that the Base fee could be raised, and then many miners on the network would be eliminated? In addition, there were many open and secret fights between miners on the network, and no one thought it would be so strict at the beginning. Therefore, the current main network parameters are still in a state of "to be continued", and the parameters are constantly updated. The official has thousands of calculations and measurements in front of the results of each version of the parameters. Going back to the past, based on the situation at that time, the current choice is still the best choice that could be made at that time. The current situation is the best that the official can do.

How to balance the relationship between low price and profit?

Filecoin encourages everyone to participate. Users want low-cost storage, and miners actually pursue profits. So is low-cost storage and miners' pursuit of profits a contradiction? In fact, this should be a very profound question that Marx talked about in Capital. There will always be such a game in capitalism, but this also promotes the development of the free market. If miners are allowed to do charity, capitalism will not exist and society will not develop. For business, making money is the greatest charity. If business does not make money, no one will do business, and everyone will be unemployed. This is the most extreme situation. In the end, this society will regress to feudal society. This is actually a very profound philosophical problem.

The miners raised this question, that is, users want to get low-cost storage. The storage cost on Alibaba Cloud is already low enough. Do I need to store it on you? On the one hand, the miners want to pursue more profits, such as making a private cloud, public cloud or hybrid cloud, which can also make money, right? Why should I put it on Filecoin, right? In fact, this thing is the same as Didi. At the beginning, Didi had 10 billion subsidies to subsidize drivers. In the end, drivers formed a habit through subsidies, and finally users also formed such a habit. Then it is equivalent to Didi spending 10 billion to trade the market. In fact, now Filecoin is also cultivating the market through Filecoin tokens. It may be said that it will cost a lot, but this road is still worth exploring.

In economics, there is a concept of consumer surplus, which means that I am willing to pay a price for this thing. At the same time, there will be a producer surplus, which means that my miner will quote a price. In fact, there is a difference between the price quoted by the miner and the price the user is willing to pay. The difference is actually the so-called welfare. If there is no transaction, market welfare will not exist. Even if you want to pay a very high price, no one will provide you with this service.

Miners may want to pursue more producer surplus, which is the so-called profit. In the storage user stage, they want to pursue more consumer surplus, that is, I want more room, I can fight for my own welfare more. It should be said that once the economic model is established, it is a borderless country. In a borderless country, just like the real world now, there will be many differences and interests.

There will be constant bargaining between interested parties, just like the bargaining between trade unions and capitalists, but the final profit will depend on the price of the currency, which is another issue for everyone to consider. How miners can make the ecology more valuable is actually a high-dimensional issue. If we are limited to internal circulation, it is difficult to increase the value of the ecology, so the official considerations may not have been too much to weigh the miners and users, but to consider how to make the network run, develop rapidly, and make its ecological value stronger. Then, if your ecological value becomes more powerful, it means that money becomes more valuable, and at this time, workers will have fewer disputes to consider.

How will Filecoin develop in the future? I think it still depends on the perfection of the technology and the maturity of the application. At present, we are also in contact with many traditional companies, including some listed companies. They also have high expectations for the Filecoin network. Some of them may even be willing to try to store some real data. If the Filecoin technology is perfect enough, I think its implementation may be faster than expected. If the implementation on the B-side can run, it may be faster on the C-side. If more and more ordinary users store some personal files, the network will actually develop very quickly.

The above is mainly my interpretation of the latest economic model. After the new economic model comes out, I will have a deeper interpretation. I hope to have the opportunity to continue to share it with you. Okay, thank you.