After nearly 3 years of operation, the disadvantages brought about by the complexity of the Filecoin protocol have gradually emerged, especially after the launch of FEVM, the development of the ecology and the innovation of applications are constrained by the complex core protocol. As an important part of the Web3.0 world, Filecoin's development has been hindered, and it has also affected the entire ecology. In order to speed up the development of the Filecoin network and at the same time reduce the resistance to ecological innovation, it is imperative to simplify the Filecoin core protocol.
**Filecoin has ushered in a turning point in its current development. This turning point is the launch of Filecoin's virtual machine FEVM, and the next launch of the native FVM under development. **The launch of the Filecoin virtual machine has brought rich imaginations to everyone, because Filecoin is not only a blockchain network for storing data, but also can perform calculations and become a public chain. Since then, blockchain-based applications are no longer poor in content, but rich in content.
**After the launch of FEVM, some contracts appeared on the Filecoin network, which is not bad, but it is still far from everyone's expectations. It is a bit lukewarm and cannot boil. There are many historical reasons for this. The most important reason is that when Filecoin was originally designed, because there was no support from virtual machines and contracts, it was necessary to realize the basic functions of the Filecoin data storage market and play a role in stimulating data storage. All functions can only be realized in core layer. This makes the core layer of Filecoin very complicated. **
We must first understand that under the circumstances at that time, such a design was inevitable and necessary, otherwise Filecoin would have problems in satisfying the vision of data storage. At the same time, these designs did play a very good role in the early days. It played a role in fueling the early development of Filecoin.
However, when Filecoin starts to build an ecology and support applications, the complex dependencies in these designs and the limitations brought about by simple and fixed definitions hinder the development of applications. These complexities are reflected in:
Define the data market at the core layer, lack of changes and flexibility, making the realization and improvement of the market to meet user needs very slow;
The agreement defines the transaction method and terms template, which may not be able to meet the real user transaction needs;
In order to incentivize the storage of real data, block rewards are used to reward certified data storage, which is Fil+. However, adopting the same value determination (10 times computing power) for all authentication data actually creates a market where bad coins drive out good coins;
Initial Pledge Collateral, Fault Slashing, and Termination Fee are equal to data-related concepts defined in the Layer 1 protocol layer, and many economic data lack consistency, such as mortgage termination fees, which may vary greatly at different heights, affecting the entire system. Financial calculations bring great complexity and risk of uncertainty.
The block consensus relies on QAP, which brings some uncertainty to the security of the protocol. Although this is not obvious in the actual network, in theory, in some special cases, it may bring harm to the security of the protocol.
Filecoin's consensus mechanism is complex, relying on market modules, authentication modules, etc. On the one hand, this makes the evolution of the consensus itself difficult. On the other hand, because the economic model of the market varies greatly in different periods, this leads to changes in the market that cause consensus unfair.
**Because of these complexities, Filecoin’s ecological development has felt resistance in many aspects, **this is also the reason why ecological applications are struggling, here are a few examples:
**The complexity of the Filecoin economic model and the uncertainty of risks have led to the continuous obstruction of the development of the Filecoin lending market. **This issue existed before FEVM went live. Even for a centralized lending platform, because it is very difficult to control risks, the mortgage owner is usually required to obtain the authority to terminate the sector before lending, and due to the complexity of the calculation, the lending interest rate is usually very high. Up to now, no lending platform that can utilize ordinary token holders has developed, and those that can operate must rely on the support of PL and FF; similarly, decentralized lending platforms will also suffer from the complexity of calculations and risks. Uncontrollable also becomes very complicated;
**Similarly because of the complexity of financial logic and the risk cost caused by excessive penalties, even though Filecoin has an annual income of more than 40% from the perspective of PoS, the increase in computing power is still not enough. ** There is a saying that storage providers cannot borrow coins. Is this a problem of not being able to borrow coins? In a free market, as long as it is profitable, someone will naturally do it. The important reason is that the risk cost is too high, so that the input-output ratio is already in a reasonable range. It is not in line with business logic to increase any more;
Because the consensus relies on the built-in market, it is difficult to develop flexible market strategies, or the user-defined market needs to rely on the core-defined market, which makes the market behavior costly and hinders development;
**Fil+ uses block rewards as an incentive to solidify the incentive model and make the incentive method inflexible, so that other data service platforms can only be competitive if they provide applications through Fil+. **This is in the development process, and Fil+ can certainly provide support, but Filecoin has a tendency for DC to completely replace CC. Just imagine, when DC basically completely replaces CC, the incentive of Fil+ will lose its practical significance and just increase the complexity.
**Filecoin's proof of storage (PoRep+PoSt) occupies a large amount of network resources, including bandwidth, storage, and computing, which limits the bandwidth that the application ecosystem can have. ** In order to protect the proof of SP's computing power, Gas Lane is still being discussed, but this cannot fundamentally solve the problem. The solution to the problem is to use subnets or shards. The current way of Layer 1 Paul Vientiane makes the interaction between the subnet and the mainnet very complicated, which is not conducive to long-term development.
The above-mentioned problems are all problems in development. This will appear in the development of every technology, and the development of blockchain is also the same. You only need to look at the development process of Ethereum to understand. The same is true for Filecoin. We must find a path to gradually realize the vision of Filecoin through gradual changes and optimizations. At present, the most urgent thing is to remove the protocol layer barriers of ecological development. That is to simplify Layer 1 and build a better architecture.
This architecture is necessarily layered. Any complex system must go through certain abstraction, design a reasonable structure, layer and try to modularize. This brings many benefits. Layering will make the implementation of each part easier, and different layers can evolve independently without being affected by other layers; modularization can also make ecological participation easier, and there can be different implementation methods. Similar to the current Internet, Ethernet is on the second layer, regardless of whether the third layer is an IP network or other similar networks. Theoretically, an IP network can also be grafted onto other Layer 2 networks.
For Filecoin, my suggestion is to simplify the core layer as much as possible and do a few basic things well, such as:
**Computing power consensus: **Currently based on storage-based EC algorithm
**Virtual machine: **Support Turing complete computing, and support storage
**Storage proof: **Storage proof in units of Sector, including PoRep and PoSt. Each Sector of equal size can be regarded as a homogeneous space, that is, Fungible Space. And it can prove the data (Piece) stored in it, what the data is, the core layer does not need to care.
This is enough, because Filecoin is a proof-of-storage network, just like Bitcoin is a distributed ledger used to store Token ownership and transfer records. The bottom layer of Filecoin is the same, except that the management of Fil is similar to BTC, and it also includes the storage person (SpId: SectorId) used to store data Piece, and the life cycle of Piece, including birth, migration and termination.
From this point of view, the part of proof of storage can also be migrated out, which can also be tried in the future. How to do it is to be studied.
Then there is a question, how to realize this migration under the current network status? In fact, some things are already being done. A lot of FIPs and discussions are going in this direction, it's just that the overall direction is still clear and vague. Also give a few examples:
FIP-0033: Explicit premium for FIL+ verified deals Hope to simplify the way of consensus and reward distribution to decouple Market and Power;
**FIP-0034:**Fix pre-commit deposit independent of sector content Simplify the calculation method of pre-commit deposit and remove obstacles for the establishment of the user market;
FIP-0041: Forward Compatibility for PreCommit and ReplicaUpdate is compatible with the future user market and implements UnSealedSectorCID;
**FIP-0045:**De-couple verified registry from markets is still preparing for the user market;
**Discussion #442:**Decoupling Storage and Deal with independent POW and POS mechanisms Simplify consensus and prepare for more flexible incentive methods;
**Discussion #719:**Preparing deal collateral for flexible data onboarding
**Discussion #712:**Capped Penalty for Sector Termination and Fault Slashing Reduce financial risks and activate the lending market;
。。。
Based on the above proposals and discussions, at this stage, we need a clear guideline and an achievable path. And this path must be a gradual change. Every change needs to be compatible with the current market implementation and needs. It is not easy, but if it is realized, it will be of great benefit to future development.
One visible direction is that we try to move up the implementation of some existing Built-in Actors, these actors can include: Market, Datacap, VerifReg, etc., but if you want to move up these Actors, you must be Built-in Actors are decoupled from them, that is, computations cannot be done based on their state. Then concepts such as Deal, DC, and QAP also need to be moved up to form two levels. Layer 1 will be easy. Since in the current implementation of Filecoin, many parts, including consensus and storage, are related to the Market, and the implementation of Fil+ has invaded into multiple Actors, these need to go through a lot of discussions, reach a consensus, and find an achievable path of.
View Original
The content is for reference only, not a solicitation or offer. No investment, tax, or legal advice provided. See Disclaimer for more risks disclosure.
It is imperative to simplify the core protocol of Filecoin
After nearly 3 years of operation, the disadvantages brought about by the complexity of the Filecoin protocol have gradually emerged, especially after the launch of FEVM, the development of the ecology and the innovation of applications are constrained by the complex core protocol. As an important part of the Web3.0 world, Filecoin's development has been hindered, and it has also affected the entire ecology. In order to speed up the development of the Filecoin network and at the same time reduce the resistance to ecological innovation, it is imperative to simplify the Filecoin core protocol.
**Filecoin has ushered in a turning point in its current development. This turning point is the launch of Filecoin's virtual machine FEVM, and the next launch of the native FVM under development. **The launch of the Filecoin virtual machine has brought rich imaginations to everyone, because Filecoin is not only a blockchain network for storing data, but also can perform calculations and become a public chain. Since then, blockchain-based applications are no longer poor in content, but rich in content.
**After the launch of FEVM, some contracts appeared on the Filecoin network, which is not bad, but it is still far from everyone's expectations. It is a bit lukewarm and cannot boil. There are many historical reasons for this. The most important reason is that when Filecoin was originally designed, because there was no support from virtual machines and contracts, it was necessary to realize the basic functions of the Filecoin data storage market and play a role in stimulating data storage. All functions can only be realized in core layer. This makes the core layer of Filecoin very complicated. **
We must first understand that under the circumstances at that time, such a design was inevitable and necessary, otherwise Filecoin would have problems in satisfying the vision of data storage. At the same time, these designs did play a very good role in the early days. It played a role in fueling the early development of Filecoin.
However, when Filecoin starts to build an ecology and support applications, the complex dependencies in these designs and the limitations brought about by simple and fixed definitions hinder the development of applications. These complexities are reflected in:
**Because of these complexities, Filecoin’s ecological development has felt resistance in many aspects, **this is also the reason why ecological applications are struggling, here are a few examples:
The above-mentioned problems are all problems in development. This will appear in the development of every technology, and the development of blockchain is also the same. You only need to look at the development process of Ethereum to understand. The same is true for Filecoin. We must find a path to gradually realize the vision of Filecoin through gradual changes and optimizations. At present, the most urgent thing is to remove the protocol layer barriers of ecological development. That is to simplify Layer 1 and build a better architecture.
This architecture is necessarily layered. Any complex system must go through certain abstraction, design a reasonable structure, layer and try to modularize. This brings many benefits. Layering will make the implementation of each part easier, and different layers can evolve independently without being affected by other layers; modularization can also make ecological participation easier, and there can be different implementation methods. Similar to the current Internet, Ethernet is on the second layer, regardless of whether the third layer is an IP network or other similar networks. Theoretically, an IP network can also be grafted onto other Layer 2 networks.
For Filecoin, my suggestion is to simplify the core layer as much as possible and do a few basic things well, such as:
This is enough, because Filecoin is a proof-of-storage network, just like Bitcoin is a distributed ledger used to store Token ownership and transfer records. The bottom layer of Filecoin is the same, except that the management of Fil is similar to BTC, and it also includes the storage person (SpId: SectorId) used to store data Piece, and the life cycle of Piece, including birth, migration and termination.
From this point of view, the part of proof of storage can also be migrated out, which can also be tried in the future. How to do it is to be studied.
Then there is a question, how to realize this migration under the current network status? In fact, some things are already being done. A lot of FIPs and discussions are going in this direction, it's just that the overall direction is still clear and vague. Also give a few examples:
Based on the above proposals and discussions, at this stage, we need a clear guideline and an achievable path. And this path must be a gradual change. Every change needs to be compatible with the current market implementation and needs. It is not easy, but if it is realized, it will be of great benefit to future development.
One visible direction is that we try to move up the implementation of some existing Built-in Actors, these actors can include: Market, Datacap, VerifReg, etc., but if you want to move up these Actors, you must be Built-in Actors are decoupled from them, that is, computations cannot be done based on their state. Then concepts such as Deal, DC, and QAP also need to be moved up to form two levels. Layer 1 will be easy. Since in the current implementation of Filecoin, many parts, including consensus and storage, are related to the Market, and the implementation of Fil+ has invaded into multiple Actors, these need to go through a lot of discussions, reach a consensus, and find an achievable path of.