Filecoin’s crypto-economic design

Original title: "Filecoin's Cryptoeconomic Design"

Original source: Blue Fox Notes

The Filecoin mainnet will be launched at block 148,888. This means that Filecoin, which has been waiting for more than 3 years, is about to arrive.

What is Filecoin?

When talking about Filecoin, we must first mention IPFS. IPFS is a peer-to-peer distributed file system. Its ultimate goal is to replace HTTP. Its core is based on content addressing, rather than traditional domain name addressing. This will change the way people get information. Under HTTP, people look for content and get services through domain names. Under the IPFS architecture, people directly look for content instead of getting information through domain names.

So, what does this have to do with Filecoin? Filecoin is the incentive layer of IPFS, it is a decentralized storage market. To realize the vision of IPFS, a decentralized storage market like Filecoin is needed.

In addition, from the perspective of web3.0, decentralized storage is also very important. Filecoin's decentralized peer-to-peer storage market allows people to store data quickly and cheaply, thereby reducing dependence on centralized servers.

From a mechanism point of view, in simple terms, Filecoin can be seen as a decentralized storage market, where storage users pay miners for services. Storage miners run software and hardware to help users store files, and are therefore paid, and the payment is made in Filecoin's native token FIL. Filecoin's blockchain records transactions of FIL tokens, as well as storage miners' storage proofs. For more information about IPFS and Filecoin, please refer to Blue Fox Notes' previous articles "Filecoin's Value Capture", "Filecoin's Ultimate Guide: Digging Deep into the Filecoin White Paper", and "Why is IPFS Interesting?"

Filecoin enters a fertile market

Filecoin will be active in at least two markets, and because of its possibilities, more new things will be derived in the future. These two markets are cloud storage and CDN services. In 2019, the output value of the cloud storage industry exceeded 46 billion US dollars, and the output value of CDN services (content distribution network) also exceeded 12 billion US dollars. These two markets are still growing, because with the continuous popularization and upgrading of hardware such as mobile phones, people generate data faster and faster, and more and more video and image data.

This is a very fertile market. Currently, most of the market share is occupied by centralized companies, especially giants. This is also consistent with the dominance of centralized services. However, with the development of dApp, decentralized applications are more willing to use decentralized infrastructure, and this situation will gradually change, although it will take a long time.

From a model perspective, if we simply compare Filecoin to Airbnb in the decentralized storage space, it allows anyone to become a data storage service provider, and anyone can also obtain decentralized storage services. Unlike other cloud storage and CDN services, Filecoin provides services in a decentralized manner, which means that it does not rely on trusted third-party organizations to provide services. This also means that it requires verifiable storage, permissionless participation, etc.

Therefore, Filecoin needs a blockchain and cryptographic verification to verify whether the user's data is being stored, that is, storage proof. At the same time, it needs to pay and subsidize through native tokens to incentivize miners to provide storage services to more people. Permissionless participation means that anyone can become a storage service provider on the Filecoin network as long as they have certain hardware equipment and are connected to the Internet.

In addition, since the participants are distributed, it may also develop into a more efficient CDN service, giving popular content the opportunity to be distributed near users. Therefore, Filecoin must not only enter the territory of AWS, but also the territory of CloudFlare (CDN service). Files closer to users will be faster, especially those popular content will have an advantage, and even institutions can participate in mining and provide data retrieval services for nearby users.

In short, Filecoin is a decentralized service network with a storage and retrieval market, and on this basis, various decentralized new applications will be born. These new applications will flourish in the next few years and will become very interesting.

Sustainable Design of Filecoin Token Economy

The design of token mechanism in crypto economy is very important. If it is not well designed, it may cause instability in the market. For example, some current DEXs have very good fundamentals, both in terms of trading volume and liquidity, but the token price has not been able to support itself. One of the important reasons is that the design of token economic mechanism is not well thought out. It only considers the issue of liquidity provider incentives, but does not consider the more important token value capture, ecosystem builder measurement and overall long-term sustainability from a global perspective. These DEXs can actually integrate the interests of liquidity providers with the value capture of tokens. From this perspective, the current token economic mechanism design will undergo more iterations.

Filecoin's token economic design serves the participants in the decentralized storage market. It aims to provide participants with a more reasonable token incentive mechanism, mitigate their short-term behavior (which is inevitable to some extent), and continue to contribute to the development of the network while ensuring sufficient attractiveness.

(Filecoin participants, Filecoin.io)

In Filecoin's token economy, storage users pay FIL tokens to store data, and storage miners are paid FIL by providing storage space. Storage proofs are verified through the Filecoin network. Storage pricing is defined by available storage capacity and storage period. Retrieval is a similar market, where users pay retrieval miners FIL tokens to obtain copies of the data they provide.

The economic prosperity of Filecoin requires the active participation of users and miners, and requires more large-scale use cases, just like an economy, if transactions are frequent and vigorous, then it will flourish. Ultimately, the value of FIL depends on the size of its storage and retrieval market. When the demand for storage and retrieval on the Filecoin network increases, the demand for FIL will increase. Because the miners' staking needs and the users' demand for purchasing storage and retrieval services are inseparable from FIL tokens.

Therefore, the speed of Filecoin's token generation must match its value creation speed, otherwise it will be detrimental to overall development. Filecoin needs to form an economy that encourages real transaction demand and can still operate without FIL subsidies.

To understand Filecoin’s token economy, we must first understand the main participants and rules of participation in its market.

Filecoin storage service provider

Unlike the centralized storage market, the decentralized storage market needs to coordinate the transaction parties to ensure reliable service provision without the need for third-party trust. To this end, Filecoin has done a lot of design, including committed capacity, sectors, miner pledges, etc.

* Incentivize miners to meet real storage services

Any market is made up of transactions, and so is the Filecoin network. However, Filecoin currently cannot distinguish between orders from real users and "washing" orders from miners themselves. To this end, Filecoin has designed a "committed capacity" architecture and introduced the concept of "verified users", which will gradually make "internal transactions" economically uneconomical.

(Storage services are provided on Filecoin, Filecoin.io)

In the early days of the network launch, storage services provided on Filecoin have the opportunity to receive rewards even if they do not store any content. The significance of this is to leave free space for the system. This is the design of Filecoin: "committed capacity".

Storage miners not only provide storage services, but also ensure consensus on the chain. Sector is the most basic unit of storage on Filecoin. It has a standard size and a clear service period. When a sector capacity has only part of the transaction order, the rest will be regarded as "committed capacity", and sectors without transaction orders will also be regarded as "committed capacity sectors". Miners can also get rewards by proving their "committed storage capacity" to the Filecoin network. Of course, if they can find the real storage demand, they can earn higher returns.

In order to prevent miners from ignoring the real storage needs and only providing "committed capacity", Filecoin also introduces the concept of "verified users". "Verified users" can be non-profit organizations, academic institutions, basic projects on the blockchain, etc., which have real needs for data storage.

"Verified users" are verified through a distributed network of verifiers. After verification, sectors that trade with verified users will receive more storage computing power and more block rewards. In other words, storing data of "verified users" will receive higher block rewards. Such incentives will encourage miners to find real data demanders to use their storage space, thereby improving the competitiveness of the entire system.

*Filecoin sectors

Filecoin is a trading market for storage, which involves the concept of sectors. As mentioned above, sectors are the most basic unit of storage on Filecoin. They have a standard size and a clear service period. Therefore, the storage proof of miners is mainly related to sectors, involving the size and usage period of sectors.

However, different sectors on Filecoin also have different weights. In addition to the most basic sector size and sector commitment service cycle, it also involves the weight of transaction orders, transaction order quality multiples, sector quality multiples, and weighted byte computing power. For example, the weight of verified user transaction orders in a sector is greater than the weight of ordinary transaction orders; different transaction order categories have different transaction order quality multiples; weighted byte computing power will consider sector quality multiples on the basis of raw byte computing power, etc. By distinguishing these weights, the quantity and quality of services can be balanced, thereby achieving a fairer distribution of tokens.

(The impact of transaction type on sector quality, Filecoin.io)

In the design of Filecoin, sectors have a life cycle, which involves the packaging of sectors, life cycle settings, and related failure costs.

(Lifecycle of a sector, Filecoin.io)

When Filecoin encapsulates a sector, it generates an identification code for the sector, which is implemented through a proof of replication (PoRep). Once the encapsulation is completed, the storage miner will generate a proof and run SNARK to compress the proof. The compressed result is submitted to the Filecoin chain as a proof of storage commitment.

The transaction order is submitted to the chain. Storage miners begin to generate proof of spacetime (PoSt) in sectors. With the storage proof of spacetime, there is an opportunity to obtain block rewards and earn storage fees. In addition to size, the current commitment period of sectors is as short as 6 months and as long as 18 months. The minimum sector life cycle is set to 6 months in order to effectively utilize the bandwidth of the chain, while the longest sector life cycle is subject to the security of the proof structure. The security of Filecoin's proof of replication (PoRep) will decrease as the sector life cycle increases.

Finally, let’s talk about the sector failure fee processing. If a sector is in a faulty state, a fee needs to be paid every day. In order to encourage miners to handle faults as quickly as possible, the amount of the fee is slightly higher than the expected income of the sector. If a sector still has faults for two consecutive weeks or the sector is terminated before the sector expires, the sector needs to pay a termination fee, refund the user’s remaining order fees, and be deleted from the chain. In addition, if the miner does not report a fault, but an unreported fault is found on the chain, a one-time fee is required. In the latest design, Filecoin has lowered the sector fault detection fee and sector termination fee to reduce the operation and maintenance risks of miners.

Termination fees and failure fees are designed to encourage miners to improve the quality of their services.

*To become a Filecoin miner, you need to stake FIL tokens

In terms of the security of the PoW mechanism, miners purchase dedicated ASIC mining machines, which are difficult to resell after purchase. If they are used for attacks, this part of the cost will be difficult to recover; PoS uses token staking as a security defense. If attackers purchase a large number of tokens for an attack, the token price will be pushed up, thereby raising the threshold for attacks.

In order to obtain security, Filecoin needs to rely on token staking in addition to the hardware resource threshold (reusable commercial hardware for mining) for higher security defense. Attackers must run hardware at the same time and also need to stake a large number of tokens.

Currently, Filecoin miners need to stake tokens to participate in mining, including storage stake, consensus stake, and block reward stake.

Storage pledge means that miners need to pledge a certain amount of FIL tokens for each sector. Its main purpose is to ensure the quality of storage services of the network in order to deal with sector failures, fines, etc. Filecoin deals with the bilateral market problem between users and miners. If miners can easily abandon storage commitments, it will cause data loss and service quality degradation for storage users. Therefore, pledge is required to prevent miners from acting arbitrarily. However, if the pledge cost is too high, it will not be conducive to more miners to enter. Therefore, in the design of Filecoin, locked block rewards or unlocked rewards can also be used as collateral.

The consensus stake is determined by the sector's weighted byte power and the FIL supply in circulation on the network. When the network reaches or exceeds the baseline (described below), approximately 30% of the FIL supply in circulation is locked in the initial consensus stake.

Finally, the protocol also requires storage providers to have a minimum pledge to guarantee the minimum storage. If the transaction order is terminated early, this part of the pledge will be punished. After the transaction order of the sector is successful, the pledge will be returned to the storage provider.

Filecoin storage users

Implementing decentralized storage on Filecoin is attractive to storage users. It is relatively easy for storage users to obtain Filecoin storage services.

First, the storage user needs to obtain FIL tokens, and then be verified to become a verifiable user. Of course, this step can be skipped. Then start looking for miners and initiate transactions with the corresponding miners. The two parties reach an agreement to verify the storage for the user through the Filecoin chain and pay the miner, and the user obtains the storage file from the miner.

(Storage services available on Filecoin, Filecoin.io)

During the transaction process, transaction information is transmitted through the Gossipsub network, and transactions can be matched by a third party to speed up the transaction. Once the storage demand and supply order are matched, the transaction order can be reached. Storage requirements include capacity, price, term, miner's pledge amount, etc. Because the state and bandwidth of the chain are limited assets, in order to ensure that providers provide users with higher quality services, the protocol will set a minimum provider transaction pledge. If the total amount of network transaction orders reaches the network baseline, the sum of transaction pledges will target 5% of the circulating supply of Filecoin.

Transaction payments are locked when the transaction order is on-chain, and payments are distributed to miners in batches during each payment cycle. If a miner terminates before the order expires, the miner needs to pay the sector termination fee, and the provider's transaction stake will also be punished and destroyed. The remaining transaction payment will also be returned to the user.

Users not only have storage needs, but also query needs for specific data. If a user raises a query demand for specific data, the retrieval miner node uses Gossipsub to propagate these queries in the miner network. The retrieval miner who owns the content returns a retrieval proposal containing price information, and the user selects the retrieval miner and pays the fee. The income from retrieval depends on price, popularity, local advantage, latency, etc.

Network Baseline Adjusts Token Distribution

The total amount of Filecoin's token FIL is 2 billion. In the Genesis block allocation, 10% of it is allocated to early investors (7.5% sold, 2.5% retained), and 20% is allocated to the project party (15% allocated to Protocol Labs, 5% allocated to the Filecoin Foundation).

The remaining 70% is allocated to Filecoin's storage miners, retrieval miners, and maintenance miners, etc., to subsidize miners to provide data storage services, maintain blockchains, distribute data, run contracts, etc. Among them, storage miners have the highest reward ratio, receiving 55% of the token allocation, because they are not only responsible for providing storage services, but also for maintaining core functions. This part of the reward is issued through block rewards to reward the maintenance of blockchains, running contracts and storage services. Retrieval miners, maintenance miners, etc. can get 15% of the token incentives.

(Filecoin’s token distribution, Filecoin.io)

Filecoin is different from Bitcoin's halving mechanism every four years, and also different from DeFi's previous short-term rapid halving mechanism. In order to achieve sustainable development, Filecoin introduced a hybrid index casting mechanism. The casting mechanism of the hybrid index involves the concept of the network baseline.

If the early rewards are too high, storage miners will be forced to package storage mining as soon as possible, invest a lot of hardware, and quit after obtaining the early high rewards, which will lead to losses for storage users and affect its long-term sustainability. In this case, users will not be willing to come to the Filecoin network for storage, and miners will not be willing to provide services for a long time.

Therefore, token release based entirely on time cannot work on the Filecoin network. It must take into account the storage computing power on the network, the basic utility and value of the network, and generate corresponding block rewards based on this.

This is why Filecoin introduces the concept of a network baseline. If the network reaches the baseline, its token release follows an exponential decay model. If the network baseline is not reached, a portion of the block rewards will be delayed. Filecoin's hybrid exponential casting mechanism, part of the reward comes from exponential decay casting, and part comes from the benchmark casting of the network baseline. The reward for each time period is the sum of the two rewards.

30% of them are distributed through simple casting, and 70% are distributed through benchmark casting. On the one hand, the most basic incentives can be provided, while the baseline attempts to match the basic business scale. Filecoin's network baseline starts at 1EiB (less than 0.01% of global storage) and grows at a rate of 200% per year (the annual growth rate of world storage is 40%). When the storage provided by the network is about 1-10% of global storage, the community can decide to slow down the growth rate. However, according to the latest announcement, Filecoin has increased the initial value of its baseline and reduced the rate to achieve more lasting economic incentives.

If the community hopes to release more token shares of the benchmark minted, it can make improvements in replication proof, consensus algorithm, transaction order market, etc. Make the replication proof algorithm more performant, with less on-chain data, faster verification time, lower hardware cost, longer sector lifecycle, and sector upgrades without repackaging; make the consensus algorithm more scalable, provide greater throughput, and process more messages in a shorter time; sector transaction orders last longer, etc.