Original article: “Foresight Ventures: The Road to LP Productization of Perpetual DEX”
By Kylo @Foresight Ventures
- Uniswap’s AMM mechanism and Compound’s liquidity mining are the ancestors of LP productization
- DeFi Native products should be oriented towards both LPs and interactive users
- Order Book Systems as a Web 2.5 Narrative Shouldn’t Be Too Confined to Web 3
- The subsequent development of Perpetual Protocol was hindered due to the problem of LP productization
- The most successful part of P2Pool Perpetual DEX lies in the productization of LP
- P2Pool Perpetual DEX LP will spawn a large number of derivative products
The main body of the narrative in this article is the productization of LP, that is, the process in which the liquidity of the DeFi protocol becomes a standardized product. In people’s initial cognition, a good product should serve users as a whole. Take DyDx and 0x protocol as an example, the two built market makers into the product, and packaged it into a trading platform through good UI design, providing Trader with smooth transaction services and capturing transaction fees, which is A typical single-user-oriented product thinking. The emergence of Uniswap’s AMM model and Compound’s liquidity mining should be the ancestors of LP productization: the AMM model enables LP and Trader to interact; Compound’s liquidity mining stimulates users’ deposits and borrowings. The product itself can also be lent out as liquidity, the identity boundary between the borrower (interactor) and the depositor (LP) is blurred, and the LP provided by it can obtain income without loss.
The above two scenarios actually reflect the characteristics of LP productization:
- LP can obtain standardized benefits under the premise of controllable risks
- After providing liquidity, corresponding LP notes will be issued to lay the foundation for subsequent LP leverage
The productization model of LP has been proven to be a feasible solution in the spot AMM DEX and lending agreement. In the current development process of Perpetual DEX, the idea of LP productization is also gradually highlighted. The mechanism of Perpetual DEX has undergone three major changes from 2017 to now: from DyDx’s order book mechanism to Perpetual Protocol’s vAMM, and now to the current P2Pool trading model. These changes in transaction patterns actually reflect the aesthetic and morphological changes of Perpetual DEX on the chain: DeFi products have gradually changed from a single product form for interactive users to a dual product form for LPs and interactive users. This model actually answers the question of what is a good DeFi product, namely:
- The silkiness of the interactor experience, that is, launching interactive products
- The divisibility of LP risks and the simplicity of market making, that is, the launch of LP products
For now, the order book system and the P2Pool model have the opportunity to realize LP productization, while the vAMM model cannot be realized temporarily. The reason why vAMM cannot be realized is that there is currently no effective LP management method, that is, it is impossible to effectively divide the risks and benefits of LP; the order book system is assisted by some on-chain financing agreements and Advisor Protocol, which can realize the democracy of LP’s income rights However, the management rights of LP are still limited; GMX’s GLP pool and gTrade’s gDAI pool in the P2Pool model are typical LP products. This means that for GMX and gTrade, Perpetual Trading is the trading product they launched for Trader, while GLP and gDAI are income-generating products for ordinary retail investors. The two products complement each other to create a logical and self-consistent system.
In the following text structure, we will interpret the mechanism of the three types of Perpetual DEX and explain the possibility and implementation method of realizing LP productization. Finally, we explain why LP productization plays an important role in the development of DeFi from the perspective of the historical view of DeFi development.
1. Orderbook transaction mode
The order book system is currently the most widely used and most robust trading model. Inherited from TradFI, it provides Perpetual’s price discovery, liquidity provision and trading functions. At present, the mainstream Perp trading products, such as WOO Network, Orderly Network, DyDx, etc., all adopt the order book transaction mode. Since most of the trading models in traditional trading markets are order books, the migration cost of Trader’s interactive order book products in the blockchain is very low. In addition, for market makers, there are already a series of risk control tools for the order book system. Under the condition of controllable risks, market makers can obtain stable handling fees and customer loss income. The Orderbook transaction model is already a very mature transaction model in the TradFi industry. However, due to the significant differences between the characteristics of the blockchain and TradFi, the order book system may face unacceptable conditions in the blockchain system, including:
- Due to the limitation of TPS on the chain, the process of transaction matching is realized off the chain, and only the settlement of the final transaction will be uploaded to the chain
- Although the liquidity provided by market makers is very efficient, it cannot be leveraged
- Liquidity is monopolized by market makers, and users on the chain cannot share the profits of market makers
In fact, the above-mentioned so-called acclimatization problem does not affect the fundamentals of the order book model. It is only due to the collision of two different ideologies between TradFi and DeFi Native that the above-mentioned problem was raised. Therefore, regarding the future development of order book DEX, Web 3 is internally divided into two perspectives, one is the perspective of Web 3 Native, and the other is the perspective of pragmatism.
From the perspective of Web3 Native, the order book data should be on-chain; (Currently, millisecond-level public chains such as Sui Network claim that there is an opportunity to realize the complete on-chain of the order book, but it can only be realized in the test network environment; DyDx V4 The version also stated that it will use a mempool-like model to realize the decentralization of the order book.) The liquidity of the order book system should be leveraged in a certain way.
But from a pragmatic point of view, the full-chain order book system still faces two major problems:
- Professional traders have a need for private transactions, and the existence of a full-chain order book may destroy the environment for private transactions
- There is a certain contradiction between decentralization and high performance, and all data in the order book is actually a compromise of the centralization of the blockchain.Compared with centralization at the level of matching, the centralization of the entire blockchain level may be even more unacceptable
In addition, the democratization of LP income and the leverage of LP also face the problems of inefficiency and high cost. First of all, the democratization of LP income needs to rely on Advisor Portocol and some on-chain fund agreements, that is, order book market makers raise funds on the chain and use the funds raised to participate in the market-making of the order book system, but the management of the LP is still controlled by the Market makers monopolize themselves. Secondly, the leverage of LP relies on the democratization of funding sources. After raising funds on the chain, market makers will generally issue a note to the user who provided the funds, representing the user’s ownership of part of the market maker’s funds. Since the note has the ownership of the market maker’s funds and some income rights, it also has the fundamentals to participate in other DeFi ecology. In the whole process, the order book needs to pay three levels of additional costs in order to realize the so-called LP productization:
- Financing costs
- The integration cost of issuing LP notes to participate in other DeFi ecology
- The cost of maintaining LP exit liquidity
Too high a cost means a bad business model. Therefore, the model of LP productization may be realized on the chain, but it may not be the best choice.
So from a practical point of view, what will be the future development trend of the order book? As an imported product of TradFi, the order book itself should be positioned at the stage of Web 2.5. This means that it should take advantage of both TradFi and DeFi to serve Web 3 users together: the advantage of TradFi lies in the market maker system and mature order book model; the advantage of DeFi lies in the transparency brought about by the on-chain settlement of assets and Asset property of token. Reasonable use of Token’s asset attributes to stimulate liquidity and cooperate with marketing is the best gift DeFi brings to TradFi.
2. vAMM transaction mode
The vAMM mechanism was first proposed by Perpetual Protocol, and currently there are two versions, V1 and V2. The V1 version is a trading method that can provide unlimited liquidity, and the transaction does not require long and short counterparties. There is no real liquidity of ETH-USDC in the V1 version, but we can understand it as an AMM virtual pool, and the k value is manually adjusted by the official. When Trader conducts a transaction, assuming that it is long on ETH 5x leverage, Perpetual Protocol will generate 5x virtual USDC, and exchange the 5x USDC for equivalent 5x ETH in the virtual pool, and the 5x ETH is equivalent to Trader’s 5x leverage is long ETH position. Since there is no real liquidity in the V1 version, the above swap process did not actually happen, and we can simply understand it as a bookkeeping process.
The core of the whole mechanism is the setting of k value, which is directly related to the price impact of each transaction on the contract. If the value of k is set too high, the price of the contract will not be sensitive to the amount of open positions, which will affect the Trader experience; if the value of k is set too low, the amount of open positions will have a large impact on the contract price, which will directly cause losses for Trader. The reasonable choice of k value lies in the real-time adjustment according to the liquidity of CEX. When the number of pools managed by Perpetual Protocol is small, manually adjusting the k value is a reasonable choice; but when the number of pools continues to increase, manual adjustment may be difficult to keep up with market changes. Therefore, considering the market-oriented pricing of the k value, Perpetual Protocol proposed the V2 version.
Perpetual Protocol V2 utilizes Uniswap V3 to construct actual “virtual liquidity”. “Actual” means that there are real LPs in the V2 version, and “virtual liquidity” means that the liquidity in Perpetual Protocol V2 is not the common AMM LP liquidity, but a virtual generated non-real value AMM LP.
LP needs to deposit USDC when providing liquidity in the V2 version. Taking ETH-USDC LP as an example, if LP wants to use 1000 USDC to provide liquidity, the system will generate up to 10000 vUSDC for it. The LP can exchange the 10,000 vUSDC into a vUSDC-vETH LP trading pair equivalent to 10,000 vUSDC according to the current vETH-vUSDC exchange rate, and deposit the LP into Uniswap V3 after setting a reasonable price range to provide liquidity for Perpetual V2. (Of course, the LP can also retain part of vUSDC, for example, only provide 8000 vUSDC equivalent liquidity, and the remaining 2000 vUSDC can be used for additional transactions.) When the liquidity of vETH – vUSDC is sufficient, the k value of LP will naturally It has been confirmed by marketization.
The price of vETH in the vETH-vUSDC LP pool calculated by the vAMM model is the pricing of Perpetual Protocol V2 for ETH perp. In order to anchor the price of perp with the spot price, Perpetual Protocol introduces a funding rate to balance long and short.
In fact, we can make a more concrete description of the abstract concept of the actual “virtual liquidity” of Perpetual Protocol V2. The vUSDC generated by LP when depositing into USDC actually exists in the form of ERC-20, and the constructed vETH-vUSDC LP also exists in Uniswap V3. The liquidity provided by LP is essentially providing vETH-vUSDC liquidity. The perp transaction conducted by Trader is essentially converted into a vETH-vUSDC spot transaction. The price of vETH is the current pricing of the perp. Since the pricing of vETH is independent of that of ETH, in order to make the two prices as anchored as possible, the tool of funding rate is naturally introduced into the Perpetual Protocol.
From the above mechanism, it can be seen that the boundary between LP and Trader is very unclear, mainly reflected in three points:
- After the LP provides liquidity, the unused vUSDC can be opened in the trading account
- When the long and short sides are unbalanced, LP needs to bet against Trader as the counterparty
- When LP provides liquidity, it also provides leveraged liquidity
The second and third points are the main risks faced by LP, and this part of the risk is caused by the mechanism of vAMM itself, and there is no way to hedge it in other ways. The risk of LP as the counterparty to bet against Trader is that some of its LP positions automatically become trading positions. Since LP automatically chooses to provide leveraged liquidity when making the market, this means that this part of the trading position is also a leveraged position, LP needs to pay the funding rate, and has the possibility of being liquidated.
The Perpetual Protocol project party is also aware of the above problems, and reduces the risks faced by LP as much as possible through the subsidy and Insurance Fund mechanism. The main way of subsidy is that when LP faces impermanent losses and losses caused by gambling with Trader, Perpetual Protocol will subsidize LP with a certain amount of $Perp; Fund first acts as the counterparty to bet against Trader, which reduces the risk that LP may face in directly participating in the market.
These two remedies seem to be effective, but they are actually transfer payments in disguise. The $Perp subsidy is mainly to eliminate the losses faced by LP through the bubble in the secondary market, which is essentially unsustainable; the source of funds for the Insurance Fund is part of the liquidation fee and 20% of the Taker opening and closing fees, which essentially come from the agreement fee and LP income. This is equivalent to forcibly distributing part of the income that should be obtained by LP to the Insurance Fund, which reduces the overall income of LP and correspondingly reduces the overall risk.
In general, the vAMM model of Perpetual Protocol seems to be a crypto native perp pricing model at present. This mechanism cleverly takes advantage of the trading mechanism of Uniswap V3, but also inherits the disadvantage of Uniswap V3’s impermanent loss. The price range set by LP in Uniswap V3 is equivalent to the price range that LP is willing to use as a counterparty to Trader. In this price range, if Trader’s buyers and sellers are not very balanced, then LP will play a game with Trader as the counterparty. In the same way, in Perpetual Protocol V2, if there is no balance between the long and short parties, LP will play a game with Trader as an opponent and bear various profits and losses. In CEX, the game between market makers and Traders already has a variety of mature risk management strategies, such as limiting Delta positions. However, since the trading mode of Perpetual V2 is automatic and mandatory, LP has no way to manage its own position risk reasonably. When the price fluctuates sharply, LP will be exposed to greater risks due to impermanent losses and the existence of Trader PnL.
The V2 model of Perpetual Protocol is actually a P2Pool trading model, and LPs will also bet against Trader under certain conditions. However, the risk tolerance of LPs under the vAMM model is far lower than that of LPs under the P2Pool model. The main reason is that LPs under the vAMM model bear their own risks, while LPs under the P2Pool model share all risks, personal risks The bearing capacity is far lower than that of the collective. For P2Pool, the losses caused by extreme market conditions will be equally distributed to all LPs; but for vAMM mode, some LP positions may be directly liquidated. Because LPs in vAMM obtain market-making benefits and bear risks are closely related to LP’s own risk management capabilities, and there is currently a lack of risk management tools for Uni V3, which leads to higher market-making risks on Perpetual V2, while The income is relatively low, and there is no way for LP to participate in the liquidity provision of Perpetual Protocol in a standard form. The subsidy provided to LP is also a disguised transfer payment, which cannot solve practical problems. Therefore, for Perpetual Protocol, it is difficult to present its LP in a productized form, and its subsequent development will be restricted.
From the data point of view, the TVL data of Perpetual Protocol is not good-looking, and the root of the problem lies in the productization of LP.
3. P2Pool transaction mode
For Perpetual Protocol V2, the vAMM design is indeed an innovation in the transaction paradigm of the blockchain industry, which can realize Perputual’s price discovery function on the chain. But the cost of the price discovery function is that LP takes too much risk, which is difficult to hedge, which means that the risk is uncontrollable under the condition of certain income. With Trader users having an inconsistent experience with LPs, it will be difficult for the product to grow into a general protocol. GMX’s P2Pool mode is on the stage of history under the above background. The P2Pool Perpetual Trading mode did not start in GMX but has flourished in GMX. Its main advantage is that the trading module can be modularized and can be split into LP, trading users, price feeding system, settlement system and other parts. Each part can operate independently, and the risk of LP can also be split.
Modular design should be one of the most important product ideas brought to the blockchain industry by the ETH 2.0 upgrade. GMX-style product design is philosophically consistent with ETH 2.0. In the future planning of ETH 2.0, PBS (proposer-builder separation) is one of the important steps in the upgrade of Ethereum. Its main function is to separate the two steps of block construction and block proposal. All blocks including the beacon chain block need to be constructed within a slot time, which requires high computing power. However, for public chains, high computation and decentralization are contradictory, and the existence of high computation will seriously weaken the decentralization characteristics of the blockchain. The existence of PBS is to combine the high amount of calculation and decentralization. The specific implementation method is to transfer the high amount of calculation to off-chain, and only return the calculation result after the calculation is completed.
The implementation logic of the above PBS can answer the pricing question of Perpetual Trading on the chain. The pricing of derivatives requires a high amount of calculation and effectiveness. If the calculations related to the price discovery of derivatives are forced to be placed on the chain, the delay and error of the calculation results due to insufficient calculation will eventually be jointly borne by Trader and LP. However, if the calculations related to the pricing of derivatives are placed off-chain, and the results are transmitted to the chain through the oracle machine after the calculation, Trader and LP will not face losses caused by the price discovery mechanism.
Therefore, the design of P2Pool Perpetual DEX is actually just to hand over the complex derivatives pricing problem to CEX for quotation, and then use the CFD model to conduct transactions on the chain, and the final result is settled on the chain. To complete the settlement, a fund pool is required to provide liquidity, and the fund pool is actually the Pool in P2Pool.
At present, the classic projects in P2Pool Perpetual DEX include GMX and gTrade, but the basic mechanisms of the two are quite different. On the whole, GMX has a more risk appetite and a more aggressive style than gTrade; in addition, due to the complexity of the composition and risk complexity of GLP, there are more derivatives related to GLP than gTrade.
3.1 GMX trading mechanism
The occurrence of a transaction on GMX requires multiple steps:
- GMX obtains various contract price data from CEX through oracles
- Calculate the contract price for opening a position on GMX based on the price feed and the weight of each CEX
- Lend a part of the assets from the GLP pool according to the opening leverage for the final asset settlement
- According to the duration of the position, pay borrow fees and corresponding opening and closing fees
There is no slippage in the contract price of GMX through calling CEX, which is an advantage of GMX to attract Trader. But often the advantages of a product are also its disadvantages. Since slippage is the cost of liquidity, no slippage means that a third party pays for this part of the cost. In the GMX scenario, the GLP pool finally bears this part of the loss caused by slippage. (GMX’s AVAX mainnet once had arbitrage behavior due to insufficient AVAX liquidity. At present, quantitative funds often conduct arbitrage activities between GMX and CEX based on the liquidity of CEX.) Therefore, there is no slippage The trading model essentially subsidizes Traders through the GLP pool, thereby attracting Traders to trade. However, the arbitrage behavior of slippage itself has not become a serious problem for GMX. The main reason is that the revenue sharing of the GLP pool is relatively high, which conceals the losses caused by arbitrage activities.
GMX’s Trader needs to borrow assets and pay borrow fees when trading. And this part of borrow fees is one of the sources of income for GLP. But the purpose of borrowing assets at this time is different from the principle of leverage platforms such as Gearbox. GMX lends funds to prepare sufficient spot for each position for final settlement; while Gearbox is similar to margin trading, and the assets lent represent the actual leverage it acquires.
Another source of GLP’s income is Trader PnL, which is the loss of Trader when trading. Since there is no funding rate in GMX to balance long and short positions, GLP holders are faced with naked short positions, that is, LPs will directly enter the market to bet against Trader. But this problem may not be a problem from the perspective of long-termism. Trader’s winning rate is generally less than 50%. fluctuation.
The community has experienced heated discussions on whether the GMX model can continue. The argument that GMX is unsustainable is mainly reflected in the imbalance of GMX positions due to the lack of funding rates. In the case of a unilateral market in the bull market, the GLP pool may face a huge risk of naked short positions, which may result in large losses. But in fact the above question is similar to answering the question of how CEX avoids naked short positions in a bull market. The currencies supported in the GLP pool are only a few blue-chip currencies with excellent liquidity, so there will be no unilateral market similar to Luna. Moreover, there is sufficient liquidity in major CEXs. GMX officials can hedge risks through CEXs or other on-chain DEXs or directly implement position limits. On the whole, there are still many risk aversion methods available to GMX during the bull market.
In fact, GMX officials are also aware of possible flaws and risks in its model design. It mentioned several possible future improvements in the X4 version, but it is still a compromise in essence:
- Introduce new assets to the GLP pool, but the underlying assets that can be opened are still limited to the constituent assets of GLP
- Introduce funding rate to balance the problem of excessive naked short positions under extreme market conditions, which means that Trader may need to pay borrow fee and funding fee at the same time when trading on GMX
- When extreme market conditions occur, adjust the ratio of non-stable coins to stable coins in the GLP pool to increase the capacity that can be opened
The above-mentioned improvement plans have not yet been finalized and are still in the discussion stage. Discussions related to GMX are likely to continue.
So to sum up, the characteristics of GMX are:
- There is no slippage in the transaction, and this part of the cost is borne by GLP
- No funding rate, so GLP may face a large naked short position
- Trader needs to pay borrow fee to GLP
- Limited by GLP, only some currencies can be traded on GMX, which lacks scalability
From the model design of GLP, it can be observed that most of the LPs attracted by GMX have a high risk appetite. Whether it is the no-slip trading mode or the existence of naked short positions, it means that GLP holders take more risks and obtain risk returns. On the contrary, the mode design of gTrade is more conservative and friendly to LP.
3.2 Gains Network (gTrade) trading mechanism
gTrade (GNS) was originally a Perp Trading product on Polygon, but it was migrated to Arbitrum. Its overall design is very different from GMX. The specific difference lies in the following points:
- When feeding the price through the oracle machine, the slippage and opening fee of the contract opening will be calculated according to the liquidity of CEX and the size of the position
- The mechanism introduces a funding rate to balance long and short positions and reduce the risk of the gDAI pool
- Only DAI stablecoin is stored in the settlement pool (gDAI)
- Trader does not need to pay the borrow fee to open a position, but needs to pay the extension fee
- A single address has a unilateral position limit, which can be increased by holding NFT
- Support transactions of multiple currencies, bulk commodities, and foreign exchange, with strong scalability
The settlement pool inside gTrade is only DAI, and its fund size is far smaller than the daily trading volume of gTrade. In the case that the agreement does not introduce slippage, arbitrageurs can easily cause losses in the settlement pool through slippage arbitrage; the introduction of the long-short funding rate is also to ensure that there will be no more naked short positions in the settlement pool and reduce the settlement pool. Risk; the purpose of the rollover fee is to reduce the time risk and try to avoid the one-time huge loss of the settlement pool caused by the long-term holding of profitable positions; the unilateral position limit is also to avoid the huge loss of the settlement pool. Therefore, it can be seen from the mechanism design of gTrader that its style is risk-averse compared to GMX. There are various risk control mechanisms in the mechanism design to ensure the safety of funds in the gDAI pool.
3.3 LP products under the P2Pool mechanism
Returning to the main narrative of LP productization, GMX’s GLP pool and gTrade’s gDAI pool are typical LP products. This means that for GMX and gTrade, Perpetual Trading is the trading product they launched for Trader, while GLP and gDAI are income-generating products for ordinary retail investors. The two products complement each other to create a logical and self-consistent system. GLP has higher risks and higher returns but does not guarantee capital; gDAI has lower risks, but gTrade has set up a series of buffer mechanisms for it to prevent gDAI users from incurring losses as much as possible.
GLP and gDAI are already relatively complete LP products, but for GLP, because GMX does not use the funding rate to balance long and short positions, GLP pool often has large naked short positions, which is equivalent to GLP Pool being forced to act as an opponent Disc and Trader game. In addition, the composition of GLP includes BTC, ETH, Link, Uni and stable coin. The value of GLP will fluctuate with the fluctuation of its spot value, which means that the currency price is also an important factor affecting the profit and loss of GLP. Therefore, on the whole, the risk of GLP can be simply split into two independent risks: Delta risk and Trader PnL risk.
The detachability of risks means that there is the possibility of derivative development of GLP. At present, there are already a large number of protocols trying to deal with the Delta risk faced by Hedge GLP, such as GMD Protocol, Umami Finance and Rage Trade. However, Trader PnL risk may be difficult to hedge from the perspective of derivatives. The feasible solution is to improve the internal mechanism of GMX or try to hedge in other DEX or CEX.
4. Why is LP productization important for the development of DeFi?
After sorting out the entire development history of DeFi, we can get a simple three-tier architecture:
- The first layer is the core currency, that is, ETH, BTC and other blue-chip core assets
- The second layer is DEX, lending and stablecoin
- The third layer is the so-called DeFi 2.0 protocol
The connection logic of this three-tier architecture is also very simple. The connection logic between the first layer and the second layer lies in the leverage of the core currency; the connection logic between the second layer and the third layer lies in the composability between DeFi protocols. Adding leverage to the core currency is a stage that all public chains must go through when building their own DeFi ecosystem, so that an endogenous DeFi ecosystem can be built based on leverage. It is worth noting that the connection between the second layer and the third layer, that is, the composability between protocols, needs to rely on the real benefits generated by the second layer protocol. In other words, the composability between protocols actually reflects the flow of real benefits between different protocols.
At present, there are not many scenarios that can generate real income on the chain, including:
- Financial service fees, such as transaction fees, lending fees, etc.
- bribery fee
- Staking Yield
These scenarios with real benefits are the basis for the development of DeFi-related products. Judging from the proven narrative of the last DeFi cycle, the core engine that drives the development of the initial DeFi protocol can be summarized as the following figure:
The income agreement here includes agreements such as Yearn Finance, and the most important function it performs is the productization of LP. In the early days of DeFi development, there were not many users who knew how to fill liquidity in the DeFi protocol, and liquidity mining was still a relatively new narrative. The emergence of Yearn FInance provides a scenario for LP to obtain income. It strategies the way to obtain income in DeFi, so that LP can passively obtain income according to the strategy of Yearn Finance without knowing the specific operation.
Although it seems that the strategy given by Yearn Finance at that time was too simple, it greatly promoted the development of DeFI. After the commercialization of LP, the increase in TVL of the DeFi protocol has entered the fast lane, and the increase in TVL has also increased the experience of interactive users for DeFi products. In addition, the productization of LP also makes it possible for LP to be leveraged, making it the basis for the development and integration of subsequent derivative products and the construction of DeFile lego.
Therefore, for Perpetual DEX, the road to LP productization is actually just replicating the logic of the initial DeFi protocol development. So what are your impressions? An underappreciated narrative actually existed in the previous story, but this story was forgotten in the corner after changing the cover. When people suddenly remember that the story can be told in another way, it takes on a new look, but if you go back to its roots, the core of this story has always existed in the long river of time that has passed.