Cold Thinking on Application Chains: What is the Best Form of Uniswap?

Original Title: 《Will the Era of Application Chains Really Arrive?》

Author: Degen Researcher 100y.eth

Compiled by: flowie, ChainCatcher

Abstract

• The recent release of the Cosmos 2.0 white paper has brought widespread attention to application chains.
• Recently, the opinion piece by Dan Elitzer titled 《The Inevitability of UNIchain》 discusses the popular topic that "Uniswap will ultimately shift to the form of an application chain," but I do not agree with this viewpoint. The following text will elaborate on my reasons.
• I believe that blockchain applications can take various forms, each with distinct advantages and disadvantages. In the future, the blockchain ecosystem will no longer be dominated by a single service, but rather a world where various applications present themselves in their optimal forms.

I. Introduction

Recently, with the release of the Cosmos 2.0 white paper, interest in application chains has surged. Typically, many applications are developed and run on general networks like Ethereum and Solana. However, with the launch of Cosmos, which advocates for the "blockchain internet," the concept of application-specific chains (App-Specific Chains) has emerged.

DApp 、 Application Chain 、App Rollup

Application chains are blockchain networks specifically designed for certain functions and applications (such as games and DeFi). Therefore, application chains leverage the highly flexible characteristics of blockchain to optimize the desired applications. For example, consensus algorithms, the number of validators, block time, block size, virtual machine types, and many other features can be tailored specifically for the application. Additionally, applications implemented on application chains will be able to fully utilize the resources of the blockchain network (bandwidth, transaction throughput, storage, etc.), which is an advantage that application chains have over DApps.

In addition to application chains, networks in the form of AppRollups have also emerged, with notable examples being networks like Immutable X or Sorare that use StarkEx solutions. The purpose of AppRollups is almost the same as that of application chains, but the biggest difference lies in whether they form a complete blockchain network by building their own set of validators (application chains) and whether they rely on other blockchains (ETH, SOL, etc.) to ensure network security through fraud proofs or validity proofs.

So, is an application chain a better choice than a DApp? Recently, application chains have received significant attention, and many articles supporting application chains have emerged. While the article from AllianceDAO presents a relatively neutral view on the pros and cons of application chains, Dan Elitzer from Nascent advocates for application chains, stating that the decentralized exchange Uniswap, built on the Ethereum network, will ultimately have to transition to the form of an application chain.

Before expressing my viewpoint, let’s analyze Dan Elitzer's opinion piece titled 《The Inevitability of UNIchain》.

II. Rebuttal to the Article "The Inevitability of UNIchain"

1. Article Summary

(The following content represents the views of Dan Elitzer from Nascent, summarized in the first person for convenience.)

Application chains are unique blockchains focused on specific protocols/applications, with their own validators/sequenceers. Uniswap is the largest DApp on the Ethereum network and will ultimately gain many advantages by transitioning to an application chain or AppRollup. Uniswap achieved approximately $700 billion in trading volume over the past year, and what fees did users incur during this active trading:

1. Swap Transaction Fees ------ Traders typically pay a fee of 0.3% when swapping tokens (Uniswap v2), sometimes paying different fees for each pool (Uniswap v3). All these fees are paid to liquidity providers. Last year, Swap transaction fees amounted to $1.2 billion, which is not the largest expenditure.

• Transaction Gas Fees ------ Fees paid to the underlying layer (Ethereum) on which Uniswap (DApp) is based. The transaction fees generated by Uniswap over the past year were $1.63 billion, some of which will be burned by EIP-1559, while the remainder is generated by miners (previously PoW), validators, and stakers (currently PoS).
• MEV ------ Over 60% of Uniswap's trading volume comes from MEV bots. It is currently unclear how much MEV occurred in Uniswap last year, but it is estimated to be $1.76 billion, which is the capital for operators, validators, ETH stakers, etc.

Source: @Danning.Sui

Currently, UNI holders only have governance rights, such as deciding whether to enable the fee switch, allowing a portion of the fees paid by LPs to be allocated to Uniswap's treasury. Although swap fees have little value accumulation in the UNI token, the value of transaction fees and MEV has not accumulated in the UNI token either.

If Uniswap (DApp) transitions to UniChain (application chain), the value generated by the protocol can be better accumulated in the token. First, it can pay UNI tokens to staked validators and stakers without having to pay transaction fees through the Ethereum network. Additionally, since application chains can help an application fully utilize blockchain resources, this is a significant advantage over DApps.

Of course, if Uniswap leaves the Ethereum ecosystem in the form of an application chain, it will face questions regarding composability (the extent to which it can interact with other protocols). However, we are already living in a multi-chain era, and most economic activities will not occur on a single chain. In other words, to address this issue, improved trustless or trust-minimized forms of cross-chain bridges are expected to emerge in the future, with enhancements in user experience and security.

2. Rebuttal to the Article

As mentioned above, with the growth of DApps, the economic benefits of switching to application chain forms seem considerable. However, I believe the above viewpoint overlooks some aspects.

First is user experience. Uniswap has become the largest DApp because it is located on the Ethereum network, which has the largest ecosystem. It is unreasonable to think that high transaction fees and MEV fees on the Ethereum network justify switching DApps to application chains to avoid high costs.

For example, in the real world, while rental costs in the suburbs are indeed much lower than in prime locations, the corresponding foot traffic and sales are also lower. Furthermore, if Uniswap were to move to an application chain, Uniswap users would have to move Ethereum assets through cross-chain bridges, which would negatively impact user experience.

Second is the issue of security. When Uniswap transitions to an application chain, it needs to obtain a set of special validators, and the security level of Uniswap, which originally benefited from the Ethereum network's security, will be significantly reduced.

Uniswap's high TVL is because it is the DEX on the safest Ethereum network. If the security level of the application chain supporting Uniswap decreases, users with substantial funds may not be able to use the Uniswap application chain.

However, if there is a desire to move forward in the form of an application chain but concerns about security exist, AppRollup can address this issue. Since AppRollup can have a degree of freedom while relying on Ethereum's security, I believe that if Uniswap were to transition to an application chain, adopting AppRollup would be a better solution. Of course, if it takes the form of AppRollup, it will have to submit state routes and transaction data while relying on Ethereum's security, treating gas fees as a form of taxation.

Third is composability. The author of the above article believes that composability will naturally resolve with the development of cross-chain bridge technology and user experience, but I think it is still too early for that. Composability refers to the extent to which one smart contract can read and write the state of another smart contract. For example, if a DApp can automate liquidity positions on Uniswap v3, then these two DApps are said to be composable. There are two types of composability: synchronous composability and asynchronous composability.

The former means that interactions occur within a given limited time, and DApps on the same network are usually synchronously composable. The latter means that interactions occur at unknown times, and DApps on different networks typically exhibit this characteristic. For an A' DApp on chain A to interact with a B' DApp on chain B, messages must be sent through a cross-chain bridge, which can be described as asynchronously composable.

Uniswap is the largest DEX on the Ethereum network, interacting with many DApps such as 1inch and Gelato, and many ERC-20 tokens on the Ethereum network have liquidity pools on Uniswap. If Uniswap switches from a DApp to an application chain, its composability with various DApps in the Ethereum network will shift from synchronous to asynchronous, which will greatly improve the user experience when interacting with other DApps.

III. What is the Best Form for Uniswap?

Using Uniswap as an example, we have analyzed Dan Elitzer's supportive views on application chains and my criticisms of application chains. So, what type of Uniswap is best? Before exploring this question, let’s understand the various forms of applications and their characteristics.

1. DApp (Ethereum)

Advantages ------ Dominant DApp ecosystem, synchronous composability, largest community, highest security level, low development difficulty.

Disadvantages ------ High transaction fees, slow speed, low application flexibility, no value accumulation for MEV.

2. DApp (Ethereum Rollup)

Advantages --- Low transaction fees, fast speed, synchronous composability (among DApps in the rollup), high security level, trust-minimized cross-chain bridges, low development difficulty.

Disadvantages ------ Low application flexibility, insufficient message bridging, no value accumulation for MEV.

3. Cosmos Application Chain

Advantages --- Low transaction fees, fast speed, high application flexibility, trust-minimized cross-chain bridges (relying on Cosmos's security when using cross-chain security), advanced message bridging between chains (e.g., IBC, ICA).

Disadvantages --- Requires a validator set and needs to build its own security, asynchronous composability, (must pay security fees to the ATOM ecosystem when using cross-chain security), high development difficulty.

4. Polkadot Parachain

Advantages ------ High security level, low transaction fees, fast speed, high application flexibility, trust-minimized cross-chain, MEV relies on value accumulation in the Polkadot relay chain.

Disadvantages ------ Insufficient ecosystem, security fees must be paid to the Polkadot relay chain, asynchronous composability, high development difficulty.

5. Avalanche Subnet

Advantages ------ Low transaction fees, fast speed, high application flexibility, relies on the security of some Avalanche validators, accumulates value for MEV.

Disadvantages ------ Insufficient token bridging and messaging levels, asynchronous composability, security fees must be paid to validators, high development difficulty.

6. P olygon Supernets

Advantages ------ Low transaction fees, fast speed, relies on the security of some Polygon validators, high application flexibility, no need to pay native tokens to validators, accumulates value for MEV.

Disadvantages ------ Insufficient message bridging levels, asynchronous composability, high development difficulty.

7. App Rollup (e.g., Stark X)

Advantages --- Low transaction fees, fast speed, high security level (relying on Ethereum), high application flexibility, trust-minimized cross-chain bridges, accumulates value for MEV.

Disadvantages ------ High gas fees relying on Ethereum security (will be cheap after EIP-4844), poor message bridging, asynchronous composability, high development difficulty.

8. What is the Best Form for Uniswap?

So far, we have examined the pros and cons of various options. Which type of Uniswap is the best?

First, aside from swap transaction fees, there are also gas transaction fees and MEV, and currently, Uniswap accumulates very little value in the protocol. To address this issue, the DApp (Ethereum rollup) form can be preferred (Uniswap has already launched on Optimism and Arbitrum).

In this case, it can enjoy Ethereum's security level and have fast scalability while maintaining synchronous composability with DApps in the rollup. In other words, even if transaction fees do not accumulate in token value, they will become cheaper, and low fees are a motivation to attract more users. However, even if a DApp (Ethereum rollup) is chosen, it still cannot accumulate value for MEV, and in terms of synchronous composability, only DApps in the rollup can achieve this, making it difficult to fully enjoy the benefits brought by Ethereum's vast ecosystem and MEV.

Secondly, the form of AppRollup (Ethereum) can be considered. While enjoying Ethereum's security level and fast scalability, it can achieve value accumulation for transaction fees and MEV. However, as mentioned above, it may significantly impact user experience and composability, and whether it is necessary to accumulate transaction fees and value for MEV is also a question.

• The Cosmos application chain and Polkadot parachain can be considered for cross-chain security. Since Uniswap needs to handle massive liquidity, it should be built on a robust network, thus excluding the relatively less secure Avalanche subnets and Polygon supernets.

The Cosmos application chain and Polkadot parachain have relatively high security, and it is possible to accumulate transaction fees and value for MEV. Furthermore, although they have asynchronous composability, the cross-chain messaging and asset transfer capabilities, such as IBC and XCMP, are already quite advanced, providing a relatively high user experience for bridges, even if they have not reached the level of synchronous composability.

However, their biggest drawback is the disparity in ecosystems. Uniswap has the most interactions with Ethereum DApps due to its presence in the Ethereum ecosystem, and migrating to the Cosmos or Polkadot ecosystem would almost mean giving up the economies of scale established in Ethereum, which would be fatal.

Uniswap's success is attributed to its position on the Ethereum network. Without the largest ecosystem, community, and the synchronous composability of numerous DApps, there would be no Uniswap as we know it today. To enjoy these benefits, if the UNI token does not accumulate transaction fees and value for MEV, I believe it should be abandoned.

As Dan Elitzer stated, if you leave the Ethereum ecosystem in the form of an application chain to accumulate value, the question is whether Uniswap will still be the best DEX after leaving. Therefore, I believe that Uniswap maintaining its current form as an Ethereum DApp or Ethereum rollup is the best option, and if it appears in the form of an application chain, I think AppRollup is the appropriate form.

9. Nevertheless, application chains are still useful

Uniswap may be better off as an Ethereum DApp, but this is just the case for Uniswap; the optimal form for each application may vary.

1. Blockchain Games ------ If it is a game that does not require much interaction with other ecosystems or DApps, there is no need to worry about composability. Additionally, Cosmos application chains, Avalanche subnets, Polygon Supernets, and AppRollups may be good choices, as more centralized operations can be tolerated. However, if it is a game that requires extensive interaction with other DApps (such as Sandbox), I believe the DApp form on a general blockchain is appropriate.
2. High-Performance DeFi ------ Running order book DeFi protocols on general blockchains is quite troublesome. Of course, there are attempts to achieve this in highly scalable networks, such as those powered by Solara, Aurora, and zkVM, but if the network's computational resources are fully utilized for Order Book DeFi protocols, the user experience will be much better. A representative example is dYdX, which has transitioned from AppRollup to application chain, and Injective Protocol has done the same. For DeFi, there is also a blockchain that lies between general blockchains and application chains, with Sei Network being a representative example. Sei Network is not an application chain targeting a single application but a network specifically for DeFi DApps. Sei Network has built-in features unique to DeFi, such as order matching engines, MEV prevention, and parallel processing, to provide a better experience.
3. Cross-Chain Bridges ------ In the case of transferring tokens or messages between heterogeneous blockchains, the situation of application chains can be considered. In the case of bridges like Wormhole and Multichain, there may be issues such as single points of failure due to the verification of bridging transactions by third-party centralized entities. However, if a blockchain is built to provide bridging services, a more transparent and decentralized bridging service can be constructed, as nodes will verify bridging transactions. A typical example is Axela Bridge or Tor chain. Of course, aside from the application chain approach, there are safer methods, such as using light clients, so application chains are not the only option for cross-chain bridges.

Conclusion

I do not agree that the future of blockchain will be reorganized around application chains, as the technical requirements for each service differ, leading to many different types of services that can leverage blockchain.

There is no such thing as a free lunch. As Jon Charbonneau from Delphi Digital commented, if flying car technology becomes feasible, and the time from prime locations to suburbs becomes very fast, the best restaurants in prime locations will only do business in the suburbs.

In other words, for Uniswap to leave the Ethereum ecosystem in the form of an application chain, both the technology and user experience for cross-chain token and message transfers must be significantly improved, which seems difficult to resolve in the near term.

However, regardless of the chosen method, each has its pros and cons, and in the future, there will be several services suitable for specific applications.

References

https://medium.com/nascent-xyz/the-inevitability-of-unichain-bc600c92c5c4

https://medium.com/alliancedao/the-appchain-universe-the-risks-and-opportunities-9a22530e2a0c

https://www.alexbeckett.xyz/composability-in-a-rollup-ecosystem/