Title: "Why Cross-Chain Bridges Have Become a Battleground?"

Author: Chloe @IOSG Ventures

Why Cross-Chain Bridges Have Become a Battleground

The summer of Layer 2 has arrived. Major Rollup scaling solutions, including Zk Rollup and Optimistic Rollup, are being deployed and expanded at a rapid pace. Many leading DeFi projects are also choosing to deploy on Layer 2. One of the leaders in the Layer 2 race, Arbitrum, has seen its TVL reach an astonishing nearly $3 billion in a short period, maintaining a strong growth momentum. From the TVL of various cross-chain bridges, we can see that assets on Ethereum have been massively transferred to sidechains and Rollups.

Image source: https://dune.xyz/eliasimos/Bridge-Away-(from-Ethereum)

However, despite the high demand for asset transfers, the usability of native cross-chain bridges for Rollups and sidechains is not so smooth—transfers across Layer 1 and Layer 2 can take up to 40 minutes or even more than 7 days, and gas fees are quite high. Additionally, sidechains and Rollups are isolated from each other, making it even more challenging for users to transfer assets across "Rollup - sidechain." This situation has triggered and accelerated the rise of third-party cross-chain bridges.

This article will focus on three leading cross-chain solutions based on Ethereum: Hop, Celer, and Connext, comparing their underlying architectures and respective advantages and disadvantages.

cBridge - Celer

Celer is a Layer 2 platform that operates across multiple chains and Layer 2, with a broader vision. cBridge is the cross-chain bridge launched by Celer, designed to support different Layer 1 blockchains, such as Ethereum and Polkadot, various Rollup solutions like Optimistic Rollup and Zk Rollup, as well as sidechains like Polygon.

Unlike Hop's architecture, Celer achieves asset transfers by building state channels and conducting atomic swaps on these channels. The security of the entire process is ensured by SGN. SGN is a dedicated PoS chain that monitors Layer 1 transactions related to Layer 2 states and relays information back from Layer 2 to Layer 1.

Cbridge's user interface is quite similar to Hop Bridge

An interesting observation is that the fees charged by CBridge can vary significantly depending on the timing of the operation (as shown, the time difference between the upper and lower images is one day), and choosing the right time for transfers can save you a lot in fees. :)

Although the user interfaces are similar, the underlying operations of Celer are entirely different, as there is no burning and minting of synthetic tokens.

1. 100 USDT is locked in a secure contract on Arbitrum (default lock time is 12 hours).

2. A state channel is created, and Cbridges generates a key to access the contract, sharing the encrypted hash of this key with the relay node.

3. The relay node creates a secure contract address on Optimism and deposits the same amount of USDT into it.

4. The user redeems USDT on Optimism using the key and shares the key with the relay node through a hash lock.

5. The relay node uses this key to unlock USDT on Arbitrum, completing the transaction.

Here, the user's counterparty is the "relay node," which has rich liquidity across multiple chains. When a user transfers USDT from Arbitrum to Optimism, their assets are first sent to the relay node on Arbitrum. Since the relay node also has liquidity on Optimism, it will further send an equivalent amount of USDT to the user's destination on Optimism and charge a portion of the fees.

In addition to the common features of atomic swaps, other advantages and disadvantages of Cbridge include:

Advantages:

1. External verification systems like SGN are universal and easy to build. Therefore, Celer can quickly deploy across various sidechains and Rollups.

2. Celer supports the most chains and tokens, providing users with the greatest selection.

Disadvantages:

1. Gateways can be more decentralized: Celer's gateway for filtering registered nodes can be further decentralized, for example, by automatically determining whitelists based on liquidity, historical bridging success rates, fees, etc., to improve efficiency.

2. User experience still has room for improvement: If the relay node suddenly goes offline before the two-step HTLC transfer is completed, users will have to wait a long time to receive their returned tokens.

Hop Bridge - Hop Protocol

Hop Bridge is a universal token bridge across Rollups and Layer 1. Hop Bridge supports Ethereum Rollup solutions like Arbitrum and Optimism, sidechains like BSC and Polygon, as well as transfers between the Ethereum mainnet. The underlying architecture of Hop is relatively novel, borrowing the burning-minting synthetic token model and adding two AMMs at both ends of the source and target chains.

User interface of Hop Bridge:

In the image above, we can see that Hop charges a portion of the fees; 100 USDT on Arbitrum can be exchanged for 98.9 USDT on Optimism, with the price difference accounting for the rewards paid to LPs for providing liquidity. Additionally, there is an extra fee of 2.25 USDT, which includes 1 USDT paid to the bonder and 1.25 USDT for transaction costs. Transaction costs will fluctuate with Layer 1 gas fees. Ultimately, the user will receive 96.7 USDT on Optimism, which is much cheaper and faster than using the native bridge.

When a user makes a transfer, Hop Bridge operates as follows:

1. Through the AMM on Arbitrum (provided by LPs with token-htoken pairs), Hop swaps USDT for hUSDT.

2. The hUSDT on Arbitrum is burned, and the bonder will collateralize assets to mint new hUSDT on Optimism.

3. Through the AMM on Optimism, Hop swaps hUSDT back to USDT.

4. The user receives USDT on Optimism.

In this process, there are two important roles: Bonder and LP. Bonder is a whitelist of nodes that need to collateralize assets to mint htokens and validate transactions. LP provides liquidity for the AMM on the chain (token and htoken pairs). Both will charge a small fee.

The advantages and disadvantages of Hop Bridge can be summarized as follows:

Advantages:

1. Strong scalability, as Hop implements a synthetic version compatible with the underlying assets, which works best on EVM-compatible chains.

2. Faster than atomic swaps, as the entire process involves only one asset transfer.

3. Unlike Celer and Connext, the roles of bonder and LP in Hop are separate, and the protocol does not require LPs to run nodes, thus lowering the threshold for becoming an LP.

Disadvantages:

1. Centralized bonders; currently, bonders are added to the whitelist by Hop Bridge's smart contract, and their number is very limited. If a bonder is unavailable during a transaction, although there is no risk of funds, the system's broadcasting speed will be much slower.

2. Limited types of supported Layer 2; currently, non-Rollup Layer 2 solutions are not compatible. Additionally, Hop Protocol does not support other Layer 1s besides Ethereum.

3. Relatively lower security; the security of Hop Bridge depends on the lowest security Rollup it supports. This poses systemic risks to bonders, LPs, and users.

4. Low capital efficiency; a single transaction requires locking a large amount of liquidity in two AMMs.

xPollinate - Connext

Connext is another Ethereum Layer 2 interoperability protocol that extends its scope from asset transfers to cross-chain data calls. In theory, Connext can support all EVM-compatible chains.

Similar to Celer, Connext builds liquidity pools on supported chains and utilizes state channels to achieve atomic swaps. However, there is a major difference between the two: the transmission of cross-chain information. Celer deploys SGN to verify cross-chain data, which consists of multiple external validators. In contrast, Connext deploys nxtp to achieve internal verification of cross-chain information.

celer SGN network

Connext nxtp

Currently, Connext does not support Optimism, and its user interface is very similar to the previous ones.

(Although Connext appears to be the cheapest among the three, it is difficult to compare due to the different source and destination chains.)

1. The user's transfer request is broadcast to the network, and various routers bid to select the final router through an auction process.

2. The user's transaction is submitted to the transaction management contract, which locks the assets the user wants to transfer on Polygon. The router passes this transaction to the transaction management contract on Fantom and locks a portion of the auction fees (less than the transfer fee).

3. The Relayer (usually another router that detects the transaction preparation event) receives the user's information and submits it to the transaction management contract to unlock the user's assets on Polygon and send them to the Polygon wallet address, confirming that the router has locked the funds. The router submits the same signed information to unlock the original amount on Fantom and completes the transaction.

This process seems more complex than Celer. It involves two roles: one is the "router," which serves a similar function to Celer's "relay node," and the other is the "relayer," which performs internal verification of cross-chain information.

Advantages:

1. Users can directly view how much liquidity is available in the pool on the webpage.

2. Connext only requires routers to lock auction fees instead of all liquidity, resulting in higher capital efficiency.

3. Connext is internally verified, thus minimizing trust (no new security assumptions are needed).

The page supports viewing liquidity

Disadvantages:

1. The underlying architecture of Connext is not as universal as Celer's PoS system, making it harder to deploy.

2. Connext's liquidity is relatively limited.

Conclusion

Although each of the three projects has its strengths and weaknesses, there is currently no perfect architectural solution. However, we can glimpse the underlying logic from the planning of the projects. Hop has a clear market positioning as a universal token bridge for Ethereum. Although Hop still has the potential to integrate with other Layer 1s, this is not its focus. Instead, scalability on Ethereum is their greatest pursuit. Celer and Connext have more ambitious goals. State channels expand the scope from asset transfers on Ethereum to cross-Layer 1 data calls. Celer quickly deployed and penetrated the market by building a relatively low-security but easy-to-generalize PoS system, which is why Celer currently supports the most chains and tokens. In contrast, Connext focuses on security, with universality being a secondary concern.

It is very difficult to determine which of the three projects is the best, but it is certain that the cross-chain bridge sector has a bright future. Firstly, with the rapid development of Layer 2, the on-chain ecosystem is becoming increasingly rich, and the demand for cross-chain solutions will grow with the development of Layer 2. Therefore, cross-chain bridges will have a vast market. Secondly, we believe that cross-chain technology has gradually matured, and existing technologies can solve most problems. What project teams need to do is continue to refine their products, improve details, and enhance user experience. For example, Celer is about to release the Cbridge 2.0 version, which will address many existing shortcomings.

Due to the length of this article, many cross-chain solutions have not been mentioned. We have attached a table at the end of the article for reference. In summary, various cross-chain solutions are developing towards security, low cost, high capital efficiency, and resistance to censorship. In the future, cross-chain bridges will undoubtedly be an indispensable part of the blockchain landscape.