Opinion: Filecoin mining is far more complicated than Bitcoin

1. Filecoin mining is much more complicated than Bitcoin and requires professional maintenance

2. Mining income only accounts for a part of Filecoin miners’ income. Miners can only get the most benefits if they find real customers.

3. Be careful when buying Filecoin mining machines. Mining pools will occupy the majority of the market share in the early stages of the network.

Although there are thousands of digital currencies in existence, the dominant consensus mechanism is still PoW (Proof of Work), which is proof of workload. This method is simple and fair, and users can participate in the calculation and contribute to the network regardless of their computing power. This is the most mature and widely used mechanism, and it is also a way for the majority of users to participate. However, one of the most criticized issues with this method is energy consumption. A large amount of electricity is used to calculate "meaningless" hash values. Although this consumption ensures a trust system, it has no intuitive and practical use and is considered a waste.

In order to reduce waste, the PoS (Proof of Stake) consensus mechanism has been extensively studied since the Ethereum was proposed, but it has not yet reached the practical stage. This includes security issues and fairness issues. Another similar consensus mechanism, DPoS (Delegated Proof of Stake), is adopted by EOS and similar projects. This method uses a small number of consensus nodes to verify and generate blocks, which is actually semi-centralized. At the same time, the majority of users are completely excluded from the consensus mechanism.

The highly publicized PoC (Proof of Capacity) last year set off a wave of hard disk mining. Although PoC avoids the large amount of power consumption by the CPU, it introduces hard disk resources for mortgage. Most projects are meaningless consumption of hard disk resources, and the resources are still not really utilized.

Filecoin uses a completely different consensus mechanism, which uses the actual usage of the hard disk for election and mining . This is also a PoS (Proof of Storage, not Proof of Stake). In the Filecoin project, PoRep (Proof of Replication) and PoSt (Proof of Spacetime) are used to implement it. Its election consensus mechanism is EC (Expected Consensus). This consensus mechanism uses VRF (Verifiable Random Function) to secretly elect the leader of each round to obtain block generation qualifications and rewards. This election method to reach consensus can effectively reduce energy consumption and resist forks. Currently, Snow White[1], Algorand[2], etc. also use the VRF mechanism to achieve consensus. On the other hand, users are encouraged to use hard disks to store truly meaningful data to obtain greater benefits, and resources are truly utilized.

For a Filecoin miner, the income mainly comes from the following aspects:

1. Block rewards and transaction fees: fixed release according to a certain algorithm (constant)

2. Storage fee: the fee paid by storage users to miners, which is released linearly according to the storage period (the total amount increases with the storage amount)

3. Retrieval fee: the fee paid by users to miners for accessing content, generally paid to retrieval miners. If retrieval miners need to obtain content from storage miners, they will also pay storage miners (not yet implemented, but will definitely be implemented in the main network)

It can be expected that in the early stage of the network, among these three parts of rewards, block reward fees will account for the majority of the total miner income for a period of time. However, as the network develops, the proportion of storage fees and retrieval fees received by miners will gradually increase. After a certain scale, these two parts of valuable income will be higher than block rewards. See the figure below for a specific illustration:

This is an inevitable trend, and the main reason is that the distribution of block rewards is fixed, while the actual value of Filecoin lies in the storage and retrieval of data. When the network grows, the block reward of the entire network remains unchanged, so the reward per unit computing power will inevitably become smaller. However, the income from storage and retrieval increases with the increase of computing power (specifically the amount of valid data stored).

1. Find users who are really willing to pay to store valid data;

Miners can not only obtain computing power, but also obtain storage fees from users

1. If no valid data is found, fill it with junk data

Miners only get computing power, but since they fill in the data themselves, they need to pay for storage, and this fee will be released linearly to the miners.

Note the similarities between 1. and 2. Miners can both obtain computing power, but the difference is very obvious. Not only can self-filling and self-mining not only fail to obtain storage benefits, but also need to produce coins for storage, which is equivalent to mortgaging a fee to the network and gradually recovering it. (This method of self-filling and self-mining may exist in large quantities in the early stage of the network, but will gradually disappear as the network develops.)

Here, for a miner, you need to ask yourself:

It should be noted that in the formal network, the system may provide two ways to reach a transaction: 1) The user reaches a transaction directly with a miner; 2) The user reaches a transaction with the miner through the network.

During the operation of the system, both methods will be adopted by users. When the network continues to run for a long enough time, there will inevitably be some large miners or high-quality large mining pools. They have more advantages in terms of price, storage guarantee, access speed, etc. Most users will inevitably embrace these more secure mining pools. At the same time, large mining pools also have more ability and resources to develop customers.

So, a basic requirement for maximizing storage costs is that you must be stable enough, large enough, and have sufficient resources. In other words, it is best to participate in a large mining pool.

Furthermore, how to maximize the retrieval fee is not such a simple question. According to the current analysis of cloud storage business revenue, the retrieval fee income is far greater than the storage fee. In the early days of the Filecoin network, when there was not much content, the proportion of retrieval fees was relatively small. But once the network develops to a certain scale, the retrieval fee will account for the largest part of the miners' income in the entire network. This is when the network becomes a truly decentralized storage network. Since the retrieval is performed off-chain, it can be expected that some super retrieval miners will appear, who will not only crawl data in the entire Filecoin network, but also store a large amount of hot data to directly respond to user requests and provide a better user experience. The retrieval node will also provide a bridge between the IPFS network and the Filecoin network, using the IPFS network to directly provide retrieval services.

Maximizing retrieval fees, for mining pools, can easily develop retrieval mining machines and retrieval software, and directly use their own mining pools to provide high-quality retrieval services. For general miners, if they can get orders for hot data, I believe it will also be a big boost to retrieval income.

What are the situations that can lead to losses? Actually there are many, listed below:

1. Hard drive damage: It is bound to happen. The probability varies depending on the quality of the hard drive and the length of time it has been used. If you buy a second-hand hard drive, the probability is likely to be very high.

2. Network outage: Less likely, but still more likely in a home environment

3. Software failure: inevitable, a matter of probability

4. Mining machine hardware failure, etc.

So how to reduce the loss? In a word, redundant backup and quick handling of failures. Hard disks can use RAID to build high availability, and node failures can be redundancy through RAID-like or EC (Eraser Coding) technology between nodes. But these are not easy for individual users to do.

So how about individual users purchasing mining machines and joining mining pools? This should be an idea, but there are still many problems involved. For example, how does the mining pool share the profits with the mining machine users? Does it only consider the block reward? Or does it take storage and retrieval into account? How does the mining pool guide the mining machines? Are the mining machines treated equally? Or are there differences based on the different configurations and performances of the mining machines?

It should be noted that one of the main reasons why Bitcoin miners can easily join mining pools is that mining pools can assign computing tasks to miners, and at the same time, the computing power of miners can be calculated and recorded at all times, which can be very fair. However, Filecoin miners are different. The computing power comes from the amount of stored data. If the mining pool only considers block rewards and divides them according to computing power, then miners themselves need to consider how to improve computing power, which is not easy in itself.

It can be expected that the participation model of Filecoin mining will be more based on cloud mining pools, where mining pools can provide users with professional traffic diversion, computing power improvement, redundant backup, system maintenance, and even mortgage guarantees. This allows users to truly feel at ease. However, such mining pools require professional design. Considering that the Filecoin test network requires an application to participate, it is estimated that Proto Labs has already considered these aspects. The Filecoin network is decentralized, but each miner node needs to be professional. Because this is a real service network, and also a network where the fittest survive, any unprofessional participant may be eliminated.

refer to:

[1] Snow White: Provably Secure Proofs of Stake. Iddo Bentov, Rafael Pass, Elaine Shi

[2] ALGORAND AGREEMENT Super Fast and Partition Resilient Byzantine Agreement. Jing Chen, Sergey Gorbunov, Silvio Micali, Georgios Vlachos

[3] Filecoin & IPFS Integration: https://github.com/filecoin-project/specs/issues/144

[4] Filecoin Proofs

Source: WeChat official account Hu Feitong

Author: Hu Feitong