257 total views
This article enumerates the technical nature of the two cross-chain application forms of “BTC anchored assets” and “cross-layer fast track”, and analyzes them based on their core technical characteristics and project examples.
Original title: “Paka Labs 4D Report (2/4) | Connecting isolated islands to mainland: BTC anchored assets and Ethereum cross-layer fast channel”
Written by: MIDDLE.X, Paka Labs Researcher Review: Shawn Lin, 1PAR Research founder, PAKA Research Collaborator
This article is the second part of the whole article, about 19000 words, the recommended reading time is 45 minutes; the whole article has four parts, which will be published one after another.
In the first article , we divided the cross-chain technology into five categories: atomic swap, witness, light node side chain, relay chain, and shared validator. From the atomicity of cross-chain transaction, cross-chain message verification, and asset custody The four dimensions of multi-chain adaptation were disassembled and elaborated respectively, and a cognitive framework and panoramic view of cross-chain technology were established.
Starting from this article, we will give examples of the application forms of cross-chain technology. This article will focus on the two application forms of “BTC anchored assets” and “Ethereum cross-layer fast asset bridge”, combined with project examples for in-depth analysis.
BTC anchored assets
BTC is the cornerstone of the blockchain world. Because of its special position in the blockchain field and the profound ideas contained in its architecture, BTC has gathered many firm believers. Until now, the market value, liquidity value, and user base of BTC in the encryption ecosystem have an absolute dominant position. Therefore, the nascent public chain often has great motivation to introduce BTC to inject strong vitality into its own ecology. As a result, a large number of BTC anchored assets have been produced.
The basic principles of BTC anchored assets are:
Lock BTC on the BTC chain and cast anchor assets on the target chain
Destroy anchored assets on the target chain and release BTC on the BTC chain
Since the BTC chain does not have Turing completeness, the link of locking BTC on the BTC chain needs to be completed through an escrow account, which is managed by a witness; in addition, the light node contract of the target chain cannot be constructed on the BTC chain. When the Burn transaction occurs, the witness must manually complete the Unlock link.
The target chain is Turing complete in general, so in the Lock-Mint link, there can be different choices. You can choose to deploy a BTC light node contract on the target chain. When the user passes the Lock transaction to the light node contract, the light node contract verifies and executes the Mint action, or you can choose to still rely on the witness to verify the Lock transaction and trigger the Mint. Generally speaking, the project side tends to choose the former, but if the target chain is Ethereum, which is expensive for Gas, the project side is more inclined to choose the latter.
The key point of the design of BTC anchored assets is the witness mechanism. Different projects have different designs on the witness mechanism, which are generally divided into two categories: trust type (no mortgage required) and non-trust type (need mortgage) .
Trusted BTC anchored assets
Witnesses of trusted BTC-anchored assets do not require collateral deposits. Such projects are often simple in design and low in cost, so they occupy a huge market share. They rely on long-term accumulation of off-chain goodwill and public custodial addresses to accept supervision. The trust of the community. It should be noted that the witness of trusted BTC-anchored assets is not necessarily a single subject, but may also be an alliance of multiple subjects.
The earliest BTC anchored asset was issued on the side chain of BTC. The RSK developed and launched by Rootstock Labs in January 2018 is the first side chain of BTC, followed closely by the BTC side chain Liquid Network developed by Blockstream in September 2018. Both created BTC anchored assets on the side chain, namely sBTC (SmartBTC) and LBTC (Liquid BTC). Later, when Rootstock updated its white paper in 2019, it renamed sBTC to RBTC (RSK BTC).
RSK is an EVM-compatible smart contract platform written in Solidoty language, designed to give BTC programmability. RSK uses DECOR+ + (a unique variant of the Satoshi Nakamoto consensus) to encourage BTC miners to concurrently serve as RSK nodes through the “combined mining” mechanism, so that RSK achieves a high level of security.
RBTC is anchored to BTC at a ratio of 1:1. The minting and redemption process of RBTC relies on the witness set called PowPeg Alliance.
- The casting process of RBTC is as follows:
The user transfers BTC to a multi-signature address controlled by the RSK PowPeg alliance. The BTC arriving at this address is locked. The SPV certificate of the transferred transaction is sent by the PoWPeg alliance to the bridge contract on the RSK side chain. Once After the bridge contract obtains this proof, it will send the same amount of RBTC to the RSK sidechain address specified by the user. This process is called Peg-in.
The Peg-In process is non-trusted and does not need to rely on witnesses. Although the PoWPeg alliance is responsible for delivering the SPV certificate to the bridge contract. But the process can actually be carried out by anyone. When the PoWPeg alliance is not delivering in time, the user can also deliver it by himself.
- The redemption process of BTC is as follows:
The user sends RBTC to the bridging contract address on the RSK blockchain. Since the BTC chain cannot verify transactions on the RSK chain, we need the RSK PowPeg alliance to assist in signing the redemption transaction on the BTC chain. When there are 15 alliance members, there are 11 When the signature is completed, the same amount of BTC can be transferred to the user from the multi-signature address. This process is called Peg-Out.
The Peg-Out process is trustworthy, and we need to believe that the PowPeg Alliance will not conspire to sign wrong transactions and steal custody assets.
The PoWPeg Alliance is composed of 15 well-known, reputable blockchain companies with high security standards and node maintenance technical capabilities. The existing alliance members are distributed in diversified geographic units and jurisdictions, and they manage the addition and withdrawal of members through 11-of-15 signatures. In addition, RSK Labs has formulated a detailed charter of alliance members, including security policies, backup requirements, etc.
In order to ensure safety and avoid the loss caused by the possible reorganization of the blockchain, RSK adopts a very conservative finality principle. The Peg-In process takes about 14 hours (100 BTC blocks), and the Peg-Out process takes about 33 hours (4000 RSK blocks). In view of the long waiting time required for Peg-In and Peg-Out operations, many institutional users (such as exchanges) will directly cast ready-made RBTC, and ordinary users can obtain them directly.
Liquid and RSK use roughly the same design, which is also an M-of-N alliance multi-signature asset custody solution, and also uses an asymmetric solution of “untrusted Peg-In + trusted Peg_out”.
The difference is that RSK focuses more on giving BTC programmability, while Liquid focuses more on providing a fast transaction network for BTC. Under this guiding ideology, Liquid was designed as a consortium chain, and the consortium responsible for asset custody is actually also a node of the Liqiud Network. In addition, Liquid Network supports private transactions. Through Liduid Network transactions, transaction amounts and asset types can be hidden.
Both RSK and Liquid were once king-level projects and were highly anticipated. However, due to the rise of Ethereum and its DeFi ecosystem, the brilliance of RBTC and LBTC was overshadowed by WBTC.
WBTC (Wrapped BTC)
WBTC is the abbreviation of Wrapped BTC, which was jointly launched by BitGo and Kyber Network in January 2019. WBTC is currently the largest BTC anchored asset with the largest amount of minted BTC, with more than 2 million BTC locked up, accounting for more than 80% of the share of BTC anchored assets, and it is the irrefutable king. Now WBTC has two formats: WBTC on Etheruem and WBTC on Tron. WBTC manages the process of asset exchange through a set of witnesses composed of WBTC DAO members.
WBTC split the role of witnesses into Custodian and Merchant. The exchanger is an intermediary role between the user and the custodian, which isolates the user from the custodian, the user interacts with the exchanger, and the exchanger and the custodian interact.
Picture from WBTC white paper
In theory, there is more than one exchanger and custodian. Each exchanger and each custodian manages an independently controlled escrow address. Regardless of whether it is an exchange or a custodian, they must be members of WBTC DAO. The addition and withdrawal of WBTC DAO members are managed by existing members through M-of-N multi-signature contracts, but not Certainly all WBTC DAO members will become exchangers or custodians.
- The process of an exchanger applying for minting WBTC is as follows:
The exchanger initiates a minting request for X WBTC and sends X BTC to the custodian’s custodial address. The custodian receives the BTC and waits for the confirmation of 6 BTC blocks (about 1 hour) before creating a WBTC minting transaction. The exchanger distributes X WBTC.
- The process for the exchanger to apply for redemption of BTC is as follows:
The exchanger initiates a Burn transaction and destroys X WBTC (the transaction is a multi-signature transaction and requires both the exchanger and the custodian to sign to formally complete it). After the custodian waits for 25 Ethereum blocks (about 6 minutes) to confirm, Release X BTC to the address of the exchanger, and finally the custodian signs the Burn transaction, making the destruction officially completed.
Users do not have the right to apply for minting and redemption, but can only buy or sell WBTC from the exchanger. The custodian also has no right to directly serve the end users, but can only serve the exchanger. Users can freely choose trusted exchangers. As for how exchangers choose custodians, the mechanism is not clear. Currently, WBTC has only one custodian, BitGo.
Users can exchange BTC and exchange providers for WBTC through atomic transactions based on hash time locks, or exchange WBTC back to BTC in the same way.
Through the separation of the roles of the custodian and the exchanger, the user experience has been greatly improved. If users and the custodian are allowed to directly interact, users will have to endure the slow process of Mint and Burn, but the exchanger can reserve WBTC and BTC in advance. , So that the user’s exchange can be completed quickly.
Although WBTC is relatively centralized and requires users to KYC, it is the most successful BTC-anchored asset. Paka Labs believes that there are three important reasons that can inspire us:
First: learn from previous experience and encapsulate the complexity between the exchanger and the custodian, giving users a simple and fast experience; second: WBTC attaches great importance to the business development of the exchange, and the exchange has actually become a distributor of WBTC , WBTC has successfully constructed a chain of commercial interests; third: WBTC leverages the potential of Ethereum and preemptively expands its application in DeFi, which in turn promotes the casting demand of WBTC.
BTCB is the abbreviation of Bitcoin BEP2. It is the BTC anchored asset issued by Binance Smart Chain, which has hosted about 100,000 BTC. HBTC is the BTC anchored asset issued by Huobi, which has hosted about 40,000 BTC. The two are the BTC anchored assets of Top2 and Top4 respectively. (Top3 is RenBTC)
Both are issued by centralized institutions relying on their large user groups and long-term business reputation. They have a good user experience and their technical routes are extremely centralized. The issuing institution directly acts as a witness and is responsible for asset custody. And verification of cross-chain transactions.
XBTC is a BTC anchored asset launched by the cross-chain project ChainX in May 2018. Similar to RBTC/LBTC, ChainX itself was also developed as a side chain of BTC, and adopted an asymmetric design of “untrusted Peg-In + trusted Peg-Out”.
ChainX’s advantage is based on the development of the Substrate framework. In the future, it will have the opportunity to become Polkadot’s parachain and bring XBTC to the Polkadot ecosystem. ChainX uses WASM technology to implement BTC light nodes for the first time on the chain, and supports SPV verification of BTC transactions, enabling untrusted security in the XBTC casting link.
In the BTC redemption link, multiple “trusts” complete the destruction-unlocking process through a majority vote on the chain. There will be 15 trusts in total, and they manage the escrow address through multi-signature. The first batch of trusts will be selected from among the validators of the ChainX testnet, and will be replaced by the “Concession System” in the future. There are two escrow addresses for ChainX, one is a cold wallet address, and the other is a hot wallet address. The cold and hot are separated to improve security, the signature records are open and transparent, and they are subject to community supervision.
ChainX is not satisfied with the current trustworthy design. ChainX claims that its asset custody solution will continue to evolve. The above is version 1.0, and there will be asset custody solutions of 2.0, 3.0 and 4.0 in the future, step by step towards decentralization.
Version 2.0: The XCLAIM mechanism is adopted to enable a distributed hosting solution.
Version 3.0: Change the account control technology from multi-signature to private key fragmentation.
Version 4.0: Let users control a special private key shard of the escrow account, which has a veto right. Through this design, the security of asset custody is improved, and the witness is free from over-collateralization.
Untrusted BTC anchored assets
Although trusted BTC-anchored assets have advantages in cost and experience, the industry is more enthusiastic about the exploration of non-trusted solutions, and many unique projects have emerged. These projects have made innovations in the access mechanism, grouping mechanism, and mortgage mechanism of witnesses based on the concept of eliminating non-trust.
tBTC: early explorers
tBTC is a BTC-mapped asset issued on the Ethereum platform. It is an open source project supported by multiple teams such as Keep Network. tBTC was issued in May 2020 and is the first untrusted BTC-anchored asset project on Ethereum.
The signer is responsible for managing the custody account of tBTC. The signer is free to enter, but it needs to use 150% of the value of ETH for over-collateralization. Since ETH is used as an overcollateralization, it will involve the oracle price feed and mortgage debt liquidation issues. tBTC directly refers to the MakerDAO mechanism and feeds the price through the MakerDAO oracle machine.
The signatories of tBTC are randomly divided into groups of 3 people, and each group manages a escrow address through 3-of-3 multi-signature. Each escrow address only escrows 1 BTC.
It can be said that tBTC is a very geeky open system, but its design is highly complex and its ease of use is extremely anti-human. First, the user can only mint 1 tBTC at a time. If you want to mint more than one tBTC, you need to apply for multiple times, and the user cannot mint less than 1 tBTC. Secondly, it is very troublesome to replace the signer of a multi-signature account, so tBTC does not Design a flexible replacement mechanism for signers. In order to allow signers to withdraw, the tBTC system requires users to redeem BTC every 6 months.
As the first decentralized BTC-anchored asset on Ethereum without the assumption of trust, tBTC is of groundbreaking significance. But due to its poor user experience, tBTC did not succeed.
renBTC (Ren Protocol ): MPC network
renBTC is also a BTC anchored asset issued on Ethereum in May 2020.
The witnesses on RenBTC are called Dark Nodes. Numerous dark nodes form an MPC (Secure Multi-Party Computing) network through the BFT consensus mechanism. Ren Protocol abstracts the cognition of the network into a Ren virtual machine (RenVM).
The dark nodes are divided into groups of 100. Every 100 dark nodes form a RenVM partition, which is responsible for managing a private key sharding account and hosting a part of BTC, but the grouping is not fixed, and the cards will be shuffled once a day. Dark nodes can join freely or withdraw at any time, and do not need to withdraw when the assets under custody are redeemed like tBTC. The dark node is “dark” because it is hidden under RenVM. As an abstract whole, RenVM assumes the function of witnesses and interacts with users, but in essence, the dark node performs related operations. Since RenVM is a BFT consensus mechanism, we can consider that 100 dark nodes in each partition only need 67 or more signatures (2/3) to operate on custody assets.
Dark nodes need to pledge 100,000 REN tokens in Ren Protocol’s Ethereum contract in order to work, and they cannot be redeemed until they exit the dark node. If the dark node tries to steal assets, the pledged REN will be slashed. The number of REN pledged by dark nodes will be adjusted with the number of BTCs hosted in the network to maintain a safe over-collateralization rate. This adjustment is done through governance and does not use oracles.
It is worth mentioning that renBTC and Acala have established a cooperative relationship. Ren Protocol supports the casting of renBTC on Acala. Acala is a Polkadot ecological DeFi platform. Its test network Karura has passed an auction and obtained Kusama’s parachain plug-in. groove. In the future, renBTC will likely flow into the Polkadot ecosystem through Acala.
Picture from Ren Protocol blog
In addition to supporting users to cast RenBTC on Ethereum and RenBTC on Acala, Ren Protocol also supports users to switch between the two. Assuming that the user wants to convert RenBTC on Ethereum to RenBTC on Acala, the user does not need to redeem RenBTC on Ethereum as BTC and then mint RenBTC on Acala, but can directly convert it directly through the Burn-Mint logic . This design saves operation steps and conversion costs, effectively improving user experience.
renBTC is one of the mapping assets supported by Ren Protocol. Ren Protocol also supports renZEC and renBCH, with the same mechanism as renBTC.
renBTC introduction document:
eBTC (DeCus) overlapping grouping mechanism
eBTC is a BTC anchored asset issued by DeCus on Ethereum. The signer in eBTC is called Keeper, and Keeper can freely access the network without permission. When users need to mint eBTC or redeem BTC, a Keeper Group composed of 3 Keepers allocated by the system will handle it. The process is as follows:
%% picture from Decus Docs
Every 3 Keepers is responsible for forming a group to manage a 2/3 threshold of hosting addresses. At the current stage, Keepers manage escrow accounts through multi-signature. The first batch of Keepers will be generated through auction activities. Keepers cannot withdraw freely. It will then iterate to accounts controlled by private key sharding technology, and Keepers will be able to enter and exit freely.
The biggest feature of eBTC is its unique “overlapping grouping” mechanism. Decus claims: Through this mechanism, only a mortgage rate of less than 50% is required to achieve security equivalent to 100% of the mortgage rate.
Suppose there are 6 Keepers, divided into two groups of 3 persons responsible for the management of 50% of the assets. These two groups are the ABC group and the CDE group respectively. Assuming there are 2 malicious Keepers, then these 2 malicious Keepers are divided into one The probability of the group is 40%, and the proportion of the custody assets that can be threatened is 50% ;
Assuming that 6 Keepers are divided into multiple three-person groups by overlapping grouping, they can be divided into ABC, ABD, BCE…20 groups, and each group is responsible for managing 5% of assets. If there are two malicious Keepers, the probability of them being assigned to the same group is 100%. There are 4 groups in which the two are in the same group, and the percentage of managed assets that can be threatened is 20%.
We found that the “hazard probability*threat asset ratio” of the two malicious Keepers is constant and has not been changed. The essence of overlapping grouping is to increase the hazard probability while reducing the threat asset ratio. However, this change is extraordinary. Keeper only needs to pledge 20% of the deposit to ensure the security of the system. Because of the existence of the deposit, even if the probability of harm is high, the malicious Keeper dare not act rashly.
The above is an example of 2 malicious Keepers among 6 Keepers. Decus plans to recruit 121 Keepers, which will generate C(121,3)=287980 Keeper Groups. Assuming that the number of malicious Keepers is m, then they can destroy the groups The number is (121-m)*C(m,2)+C(m,3). When m is 10, the number of groups that can be destroyed is 5115, accounting for about 1.78% of the total number of groups, which means The proportion of assets that can be threatened is only 1.78%. For other m-values, we can also calculate them in turn and compare them with the conventional grouping scheme:
%% Comparison of security mortgage rate between overlapping grouping scheme and conventional grouping scheme
It can be seen from the values in the table that the overlapping grouping scheme always has a lower security mortgage rate than the conventional grouping scheme. But we also found that Decus claims a safe mortgage rate of less than 50%, which is hypothetical : only if the proportion of malicious keepers does not exceed half, the safe mortgage ratio does not need to exceed 50%.
InterBTC independent control address matrix
In 2016, Polkadot proposed in the white paper to establish an interoperability relationship with BTC. In January 2020, the Web Foundation commissioned Interlay to design and develop a BTC-anchored asset on Polkadot based on XCLAIM. Therefore, PolkaBTC developed by Interlay can be considered as having Polka’s official background. Later, PolkaBTC was renamed InterBTC.
Interlay uses Rust language and Substrate framework to develop a BTC-Parachain, and plans to enter the Polkadot ecosystem as a parallel link in the future. In October 2021, Kintsugi, Interlay’s pioneering network, has taken the Kusama slot and connected to the Kusama network.
Similar to ChainX, BTC light nodes are deployed on BTC-Parachain, enabling BTC-Parachain to have the ability to verify transactions on the BTC chain. In the InterBTC system, escrow accounts are a number of independently controlled BTC accounts. These accounts are called Vaults, and the witnesses are called Vault administrators. Any entity can become a Vault administrator by staking DOT, and the amount of staking DOT is proportional to the maximum amount of BTC that can be escrowed. The user transfers BTC to the Vault. After BTC-Parachain is verified, the user will be issued InterBTC. When the user needs to redeem BTC, BTC-Parachian destroys InterBTC, and the Vault administrator returns the BTC to the user. Although each Vault is an independently controlled account, a 150% over-collateralization can protect these Vault managers from doing evil.
Users need to pay DOT as a handling fee for casting and redemption, and the fee will be given to the Vaults manager as an incentive. If the Vault administrator attempts to steal the custodial BTC, the pledged DOT will be confiscated and compensation will be given to the users who suffered losses. The existence of the over-collateralization mechanism requires the system to configure the oracle to feed the price, but since InterBTC has not yet been launched, the supplier of the oracle has not been determined yet.
In addition to managing the custodial account, the Vaults administrator also assumes the functions of Relayer, and is responsible for transmitting the block header of the BTC chain to BTC-Parachain.
The InterBTC system allows users to choose to build their own Vault to host their own BTC. This function is necessary for large liquidity providers who need to convert large amounts of BTC into InterBTC.
%% Picture from Interlay Docs
Although it is not mentioned in the introduction document, according to our understanding, Vaults administrators should not be able to withdraw at any time, and need to wait until the custodial BTC is redeemed before they can withdraw. However, we can make a guess: what happens if you support the Vault manager to sell your own Vault?
Summary on BTC anchored assets
Above we have introduced 8 representative BTC anchored assets. Trusted anchored assets have economic efficiency advantages and are more successful overall. Untrusted anchored assets are more in line with the spirit of the blockchain, but they are still being explored. Medium: tBTC made the earliest attempt; then RenBTC solved the problem of the witness’s exit mechanism by using a private key sharding mechanism, and used regularly refreshed groups to improve security; eBTC changed the grouping method when mortgage was inevitable. To reduce the security mortgage rate, InterBTC focuses on increasing the distribution of witnesses and providing users with more choices.
On the whole, there is still a lot of room for innovation in untrusted BTC-anchored assets, and what kind of scheme is the best practice will still be verified by time.
The following table is a feature comparison table for the BTC anchored assets mentioned in 5.1.1-5.1.2:
%% BTC anchored assets comparison table
Click here to copy the link to view the larger image.
Ethereum cross-layer fast asset bridge
If BTC is the most successful digital asset, then Ethereum is the most successful smart contract platform. Regardless of asset scale, user scale, or ecological scale, Ethereum is undoubtedly the most attractive public chain for Dapp deployment. The Solidity language and EVM environment have also become the most used development tools for blockchain developers. However, the ecological capacity of Ethereum seems to have reached a bottleneck, and slow transactions and high gas fees have become the core limiting factors hindering its continued development.
Although the sharding scheme of Ethereum 2.0 is already planned, it will take time to actually land, so various expansion schemes have been proposed, including side chains, Lightning Network, Plasma, Truebit, state channels, and Rollup. Among them, although the side chain can solve the expansion problem, its security is independently responsible and cannot inherit the security of Ethereum, so the side chain is considered a new Layer 1, and the rest of the expansion plan is considered Layer 2.
For the sake of brevity, the following text will abbreviate Layer1 and Layer2 as L1 and L2, respectively.
The implementation of L2 has evolved, and finally Rollup has become the main expansion technology. The Rollup solution has the best overall performance in terms of L1 level security, data availability, scalability, and user experience. Most of the expansion networks in the Ethereum ecosystem have chosen to use the Rollup solution to build. The L2 mentioned later will refer specifically to the Rollup Layer 2 network.
The basic route of Rollup is to submit all the state transition information of the second-layer network to L1, but at the same time, by providing validity or fraud proof, L1 can realize “lazy verification” (just verify the fraud proof or validity proof. It is equivalent to verifying all state transitions submitted by L2), thereby saving L1’s computing resources.
According to the form of proof submitted by the Rollup network to L1, the Rollup network is divided into two categories. One is Optimistic Rollup (hereinafter referred to as Op Rollup) that generates fraud proofs through the challenge period, and the other is the use of zero-knowledge proof technology to generate validity proofs. Zk Rollup.
Why do we need a cross-layer fast asset bridge
Due to the challenge period, Op Rollup has a shortcoming, that is, the user’s withdrawal cycle from L2 to L1 is long, which takes about 7 days. ZK Rollup has a relatively large calculation complexity for generating zero-knowledge proofs. , There is also a waiting time of about 1 hour. However, Op Rollup is currently more widely used, because Op Rollup is better compatible with EVM and is easy to migrate L1 Dapp. ZK Rollup is more difficult to be compatible with EVM and requires a lot of technical research work, which has not yet been implemented. In addition, even if ZK Rollup is successfully compatible with EVM in the future, the withdrawal time of about 1 hour is unacceptable for impatient users. What users expect is: minutes, even seconds!
In addition to fast withdrawals, there is a typical demand, that is, fast transfers between L2 and other L2. The conventional way is to first withdraw money from L2-1 to L1, and then deposit from L1 to target L2-2, but this is slow and uneconomical.
In view of the above requirements, many cross-layer fast asset bridges that provide L2→L1 fast withdrawal and L2⇋L2 fast transfer have been developed.
The technical nature of the cross-layer fast asset bridge
The fast asset bridge itself did not speed up the original asset circulation channel, but set up a new one to construct a new asset circulation channel. In the following text, we will refer to them as “original channel” and “fast channel” respectively .
In the express channel, a new role has been added to provide users with liquidity advances. The user pays the liquidity provider on the source ledger, and the liquidity provider pays the user’s target address on the target ledger. Then, the liquidity provider rebalances its assets through the original channel to achieve liquidity return.
We found that the cross-layer fast asset bridge, in addition to supporting L2→L1 fast withdrawal and L2⇋L2 fast transfer, can often also support fast transfer between L2 and other EVM compatible chains, because EVM compatible chains are often established with Ethereum In order to bridge the relationship (including the side chain of Ethereum), there is a multi-hop original channel between L2 and L2.
Some documents attribute the liquidity advance model of the cross-layer fast asset bridge to a new cross-chain solution, and call it “liquidity swap.” Pay attention to its core-trust mechanism. According to its trust mechanism, all cross-layer fast asset bridges can basically fall into two frameworks, one is the atomic transaction mode and the other is the witness mode . We will introduce several typical projects separately:
Atomic transaction type cross-layer fast asset bridge
cBridge is a cross-layer fast asset bridge built by the Ethereum L2 layer expansion platform celer.network.
cBridge adopts a hash time lock scheme, and makes the relay node (Rely Node) the public counterparty. The completion process of a transaction is as follows:
- Step1: The user initiates a TransferOut transaction on the source ledger to transfer money to the relay node, and the transaction sets a hash time lock;
- Step2: The relay node initiates a TransferIn transaction on the target ledger, transfers money to the user, and sets the same hash time lock;
- Step3: The user confirms the TransferOut transaction on the source ledger, and the original image of the hash lock is disclosed;
- Step4: The relay node confirms the TransferIn transaction on the target ledger, and the cross-layer transaction is completed.
We found that the above process is slightly different from a typical hash- based time lock . A typical hash time lock transaction should be confirmed by the relay node TranferOut on the source chain, and the user should confirm TransferIn on the target chain.
The purpose of cBridge’s design is to improve the user experience and avoid the need for users to switch wallets during cross-layer transactions. In the optimized transaction process, all operations of the user are completed in the source ledger, and there is no need to switch to the target ledger wallet to perform any operations.
This adjustment brings about a small problem: if the relay node does not perform Step4 operation, and waits until TranferIn expires after the timeout, it will obtain the assets of the user TransferOut for free. In fact, after the completion of Step3, the original hash has been disclosed. Anyone can complete the operation of Step4. cBridge recommends that users pay attention to the status of the TransferIn transaction in time. If the relay node has not completed Step4 for a long time, the user is required Go to the target ledger to confirm the TransferIn transaction.
After cBridge is actually running, the relay node will faithfully perform the Confirm operation in most cases. The transaction amount of TransferIn will be slightly less than TransferOut, and the difference is the commission charged by the relay node. If the relay node does not confirm the transaction for many times, it may lose its qualification as a relay node.
In the cBridge2.0 update in September 2021, the relay node role has been merged into Celer’s State Guardian Network (SGN) validator, and there will be no separate relay node role in the future.
In addition to supporting Ethereum and its second-layer network Arbitrum, cBridge currently supports the rapid transfer of assets between BSC, Fantom, Avalanche, OKExChain, Polygon and other blockchain networks that are compatible with Ethereum EVM.
NXTP is released by the Connext team. Its full name is Noncustodial Xchain Transfer Protocol, which is translated into non-custodial cross-chain transfer protocol in Chinese. The protocol uses an atomic transaction mechanism similar to a hash time lock, but does not rely on the original image of the hash. Instead, it is based on a smart contract that directly sets the trigger condition of a transaction to provide a signature for another transaction.
When users conduct cross-layer transactions through NXTP, the transaction will go through three stages:
- Step1: Router bidding
Users broadcast transaction requirements to the network. Router secretly bids with the promised transaction completion time and handling fee amount. The user selects an advantageous Router (selection of bids) and enters the next stage;
- Step2: Prepare
After the user’s bid selection is completed, the asset to be transferred is locked to the transaction management contract on the source ledger (the transaction contains bid selection information). After the successful bidder monitors the transaction, it locks the user’s due to the transaction management contract of the target ledger. The number of assets (the number of assets the user deserves is related to the total liquidity distribution in the system, which will be described later);
- Step3: Fulfill
The user provides a signature to obtain the assets locked by the Router on the target ledger. The Router uses the signature information disclosed by the user to unlock the assets locked by the user on the source ledger. In order to prevent users from switching wallets, other non-successful routers will provide relay services. Users do not need to go to the target chain to provide signatures to unlock assets. Instead, they can send the signature information to the relay router, and the relay router will send it to the target chain. Unlock it on the chain. Relay Router will also charge a small fee.
Picture from Connext document
The Router in NXTP and the relay node in cBridge assume the same role. The difference is that the former customizes its service price and bids on users, while the service price of the latter is uniformly stipulated by the agreement and adjusted through governance.
It is worth mentioning that there is a virtual AMM mechanism for NXTP’s liquidity provision, which means that the user pays 1 USDC on the source ledger and obtains it on the target ledger, which may not be (1-r) USDC (set the handling fee as r), and it may be (0.99-r) or (1.02-r). The specific value depends on the ratio of total liquidity on the source ledger to the target ledger. The purpose of this design is to add a negative feedback mechanism to promote Router to balance the liquidity of different ledgers according to demand .
As of the post, NXTP is still under audit. After NXTP is released, it will first implement support for Ethereum, Optimism, Arbitrum One, BSC, xDAI, Polygon, Fantom Opera.
StarkWare is a zero-knowledge proof research and development organization, and the developer of the ZK Rollup Layer 2 network StarkNet. StarkEx is an extensible tool set developed by StarkWare for StarkNet, including StarkEx Bridge.
StarkEx’s current service model is L2 as a Service. It supports other projects to use StarkNet technology to build its own independent L2 network. Currently StarkEx’s customers include Immutable X, DyDx, DeversiFi, all of which have built their own L2 network with the support of StarkNet .因此，StarkEx Bridge 首先要解决的只是StarkEx 生态内的L2 网络的跨层交易问题，然后才会逐步扩展为适配所有L2 的解决方案。
StarkEx Bridge 采用了类似Connext 的条件交易机制，通过该机制实现无信任的跨层原子交易。StarkEx Bridge 上有专门的LiquidProvider (LP) 作为公共交易对手，提供流动性。其过程如下：
▸ StarkEx L2→L1
- Step1 : Alice 在L2 向LP 发起1 ETH + 手续费的条件转账T(X)，条件是T(Y)：LP 在L1 向Alice 付款1 ETH，在T(Y) 生效前，T(X) 状态为[invalid]；
- Step2：LP 签署T(Y)，在L1 上向Alice 付款，Alice 立即可在L1 使用该资金，LP 拿到T(Y) 的生效证明；
- Step3：LP 使用T(Y) 的生效证明，更新T(X) 状态，T(X)[invalid] 转化为T(X)[Valid] ；
- Step4 ：LP 通知L2 节点打包T(X)[Valid] 到零知识证明批次中，该批次抵达L1 并被验证时，LP 正式拿到Alice 的付款。
%% 图片源于StarkWare 博客
▸ StarkEx L2→ StarkEx L2
- Step1：Alice 在L2-1 中向LP 发起1 ETH + 手续费的条件转账_T(X) ，触发条件设为_T(Y) : LP 将1 ETH 转到Alice 的L2-2 账户。在_T(Y)_生效之前，_T(X)_为[invalid] 状态；
- Step2：LP 签署_T(Y)_，在L2-2 上向Alice 付款1 ETH，该交易立即生效，Alice 立即可在L2-2 使用该资金；
- Step3：_T(Y)_被L2-2 节点打包到零知识证明批次中，提交给L1 并被L1 验证，LP 拿到_T(Y)_在L1 的生效证明；
- Step4：LP 使用_T(Y)的生效证明，更新_T(X)_状态，_T(X) [invalid]_转化为_T(X)[Valid] ；
- Step5：LP 通知L2-1 节点打包_T(X)[Valid]_到零知识证明批次中，该批次抵达L1 并被验证时，LP 正式拿到了Alice 的付款。
注意：上述描述对条件交易过程进行了简化，事实上，如果要创建以T(Y) 作为触发条件的T(X) ，T(Y) 需要先被创建，只是处于未签署状态，或者可以称为invalid 状态。可以将创建T(Y)[Invalid] ，理解为一笔Lock 操作。另外，StarkEx L2 作为非独立的状态机，一切交易的生效与否都以被L1 验证为准。
我们发现，与Connext 不同， StarkEx Bridge 采用了一种非对称的原子交易设计，用户可以立即拿到资产，但LP 却需要等待一段时间，等到LP 对用户的付款信息通过原始通道抵达，才能解锁用户的付款。这个等待时间不会太长，大约1 小时左右。这个等待时间就是LP 的资金占压成本。
StarkWare 还提供了StarkNet L2⇄侧链的快速交易通道，流程与StarkNet L2→ StarkNet L2 大体相似。
StarkWare 认为，ZK Rollup 相比Op Rollup，有一个重要的优势：Zk Rollup 的跨层快速通道，对于LP 而言，具有更优的资金效率，而Op Rollup 的跨层快速通道，LP 的资金占压达7 天之久，资金效率更低，这会转化为昂贵的流动性手续费。
StarkEx Bridge 介绍：
我们发现，原子交易型的跨层快速资产桥，设计的关键点在于如何避免要求用户切换钱包操作，cBridge、NXTP、StarkEx Bridge 采用了不同的设计。
cBridge 选择将原子交易的执行顺序进行微调，NXTP 则选择让中继Router 代替用户去目标账本解锁资产。
StarkEx Bridge 则进行了更彻底的改变：cBridge、NXTP 的原子交易都是让用户的转账触发流动性提供商的转账，而StarkEx 则交换了两者，让LP 的转账触发用户的转账，并让触发过程走原始通道，这样一来用户可以先拿到钱，而且不用手动去操作为LP 付款的事情。
Hop Exchange，也称Hop Protocol，由Authereum 钱包团队打造，其创始人是以太坊编程语言Solidity 的开发者之一Chris Whinfrey。Authereum 团队在开发该钱包的时候发现了当时以太坊对于扩容的紧迫性，所以将精力转移至了L2 相关的设施中。
Hop Exchange 设计了一个媒介通证：hToken，例如hWETH，hDAI，hUSDC，并在各个支持的L2 上部署了hToken:Token 的AMM 兑换池。
Hop Exchange 中见证人角色被称为Bonder ，中文可译为连接者。Bonder 是实现快速提款和快速交易的流动性垫付者，也是在Layer 之间传递消息的中间人。Bonder 要在L1 上抵押原生Token，以获得在L2 上铸造对应的hToken 的额度。（此处注意，只是获得额度，并不是L1 抵押原生Token ，立即在L2 生成hToken ）。
- 依照当前系统设置，设AMM 流动池手续费0.3%
- 依照当前系统设置，设Bonder 提供的垫付服务手续费为0.2%
- 假设AMM 流动池是绝对平衡的，1 USDC 始终兑换1 hUSDC
的情况下，阐述Hop Exchange 的系统设计。
当Alice 需要从L2 快速提款1000 USDC 到L1 时，需要经历以下过程：
- Step1：Alice 通过L2 上的AMM 兑换池，将1000 USDC 兑换为997 个hUSDC；
- Step2：Alice 通过L2 的HopBridgeContract (下文简称HBC)，在L2 上销毁997 hUSDC，设该交易为X；
- Step3：Bonder 监听到交易X，在L1 上从自己的抵押金里垫付995 USDC 给Alice ，Alice 在L1 上拿到995 USDC，对于Alice 而言，交易已完成；
- Step4：交易X 被提交到L1，通过欺诈证明或零知识证明，被L1 验证；
- Step5：L1 的HBC 获取到交易X 已被完成的信息，向Bonder 归还997 USDC。
当Alice 需要将L2-1 上的1000 USDC ，转移到L2-2 上时，需要经过以下过程：
- Step1：Alice 使用L2-1 上的AMM 兑换池，将1000 USDC 兑换成997 hUSDC；
- Step2：Alice 调用L2-1 上的HBC，销毁997 hUSDC，设该交易为X；
- Step3：Bonder 监听到交易X，使用自己的hUSDC 铸造额度，调用L2-2 上的HBC，为Alice 铸造995 hUSDC，Alice 将995hUSDC，在AMM 池中兑换为992 USDC，此时，对Alice 而言，交易已完成；
- Step4：交易X 被传回L1 ，经过欺诈证明或零知识证明，被L1 验证；
- Step5：Bonder 监听到被验证的交易X，同步给L2-2 上的节点；（为了加快速度，Hop Exchange 建议Bonder 自己运行L2 节点）。
- Step6：L2-2 节点验证后，L2-2 上的HBC 为Bonder 铸造997 hUSDC。
如果每一笔兑换的流动性归还步骤，都需要走一遍L1 的话，会耗费很多Gas，所以Hop Exchange 实际上是批量处理流动性归还的。系统会收集一段时间内的垫付，再把流动性归还的需求打包为一个Transfer Root 处理，代价是Bonder 的回款周期会延长一些。
Bonder 在提供垫付服务时，事实上充当了见证人的角色，因为垫付是提前发生的，早于目标账本自身对来自源账本的交易有效性验证。因此为了完成垫付，源账本上的交易会首先由Bonder 进行监听和验证，并传递给目标账本上的HBC。倘若恶意的Bonder 向目标账本传递虚假交易，则可能造成不正确的垫付。
Bonder 在L1 上的抵押金不光是其在L2 铸造hToken 的额度，也是其忠实履行职责的保证金。Hop Exchange 设定了专门的挑战者（Challenge Watcher），一旦发现Bonder 的欺诈行为，Bonder 的抵押金将被没收，变成给挑战者的赏金。
Hop Exchange 会在所有支持的L2 上部署AMM 兑换池，任何人都可以往兑换池里注入hToken:Token 的流动性，以赚取兑换手续费。由于hToken 和原生Token 的价格属性是基本一致的，Hop Exchange 在AMM 中选择使用类似于Curve 的Stablecoin 价格曲线，减小滑点。尽管如此，hToken 和原生Token 可能还是会存在微弱的价差，这当中会有套利者的空间。套利者并不是Hop Exchange 当中的正式系统角色，任何人都可以充当套利者。套利者的存在将使得hUSDC 和USDC 的价格基本保持平衡。
与StarkEx Bridge 相似，Hop Exchange 也设计了一个不对称的系统，让用户可以立即获得资金，但让流动性提供者，也就是Bonder 的，等待通过原始通道拿到回款。
Hop Exchange 界面截图
hToken+AMM 的设计是相对复杂的， 但Hop Exchange 的前端界面上努力做到简洁，让用户几乎可以不用感知hToken 和AMM 兑换过程的存在。
hToken + AMM 的引入，事实上降低了整体的资金效率，因为AMM 中的流动性是需要额外占压资金的。Hop Protocol 之所以采取这样的设计，是出于技术性的考量，这样做可以避免去维护一个L2 上的资产合约清单，详细的解释，可以参考这篇文章。
Hop Exchange 目前已支持以太坊主网与Arbitrum、Optimism、xDai Chain、Polygon 的跨层兑换。
Hyphen 是由Biconomy 推出的跨层快速资产桥。Hyphen 目前的开发还处于早期阶段，仅支持Ethereum 和Polygon。
Hyphen 将LiquidtiyPoolManager 合约（后文简称LPM 合约）部署在所有受支持的账本上。LPM 合约中存储了所有的流动性。任何人都可以向LPM 合约中存款来提供流动性，提供流动性的人我们称之为LP ，LP 将获得流动性费用。
用户需要跨层转账时，只需在源账本将资产存入LPM 合约，Hyphen 网络的执行节点（Executor Node）会监听存入事件，并转发给目标链上的LPM 合约，目标链上的LPM 合约接收到之后，就会释放资产到用户提供的目标链地址。执行节点并不需要提供流动性，所有流动性都在LPM 合约中，并通过AMM 机制调节兑换价格比。但这个AMM 机制有点特别，一个交易对的两种资产不在一条链上，AMM 想要计算出价格，还需要执行节点来提供另一种种资产的储备量。
Hyphen 的特色功能是：用户在源账本和目标账本上需要支付的Gas，都会由Hyphen 网络统一代付，并折算为用户要兑换的资产类型向用户收取。用户在目标账本收到的金额，将会是源账本上的金额-源账本Gas 费-目标账本Gas 费-给执行节点的手续费-给LP 的流动性费用。这样做的好处是，用户可以在没有ETH 的情况下进行兑换，而且可以清晰的看到兑换的总成本。
我们发现Hyphen 的设计极其简洁，通过Hyphen 进行跨账本转账也非常迅速，可以在几秒内完成。但我们细心留意，也可以发现，Hyphen 的设计，包含了对执行节点的信任假设，是一种相对中心化的方案。
Degate 的目标是创造一个更强大的Uniswap，打造一个功能丰富程度可以比肩中心化交易所的去中心化交易所。Degate 将通过各个模块，分别支持AMM 交易、订单薄交易、杠杆交易、跨层交易等功能。Degate Bridge 是Degate 用以实现跨层快速交易的模块。
Degate Bridge 界面截图
Degate Home DAO 作为单个主体，承担见证人的角色，并以其锁定的资产作为担保。此外，Degate Bridge 的流动性是放在链下的，由Degate Home DAO 直接在源账本上接收用户资产，并在目标账本上向用户输出资产。Degate 设置了虚拟AMM 机制，采用类似Curve 的Stablecoin 自动做市曲线。
DeGate 在白皮书中表示，当以太坊生态出现服务于L2→L1 消息传递的成熟预言机服务后，DeGate Bridge 将依托该预言机服务，实现去中心化的跨层兑换服务。
Degate 的开发还在早期阶段，仅支持了以太坊和Arbitrum 之间的跨层交易。
Optimism DAI Bridge ( MakerDAO )
为了扩大DAI 的使用，MakerDAO 正在逐步推动在L2 部署DAI 的合约，目前已经在Optimism 上部署，与此同步推出的是名为Optimism DAI Bridge 的桥接器，该桥接器将支持Optimism 上的DAI （称为oDAI）到Ethereum DAI 的快速提款。
Optimism DAI Bridge 本质上是依赖于一个中心化的预言机（Oracle）在L1 和L2 之间传递消息，来实现快速提款的。
当用户需要把oDAI ，提现为L1 上的DAI 的时候，会经历这样的过程。
- Step1：用户在L2 上通过DAI Bridge 合约，销毁oDAI；
- Step2：Maker Oracle 将销毁信息从L2 传递给L1 的DAI Bridge 合约，L1 的DAI Bridge 合约为用户铸造fDAI；
- Step3：用户拿到fDAI 之后，可以选择在7 天（挑战期）后到MakerDAO 的财政库中兑换DAI，也可以选择拿fDAI 作抵押，从Maker 财政库借出DAI (借出数量＜抵押数量)，当抵押的fDAI 过了挑战期，债务将被自动结算。
%% 注意：没有fDAI:DAI 的AMM 兑换池，因为考虑到在不同时间的操作中获得的fDAI 挑战期到期时间不同，fDAI 被设计为了NFT。
%% 图片源于MakerDAO 博客
Optimism DAI Bridge 尽管只支持DAI 这一种资产的跨层兑换，但是以DAI 为媒介，也可以实现其他资产的跨层兑换，不过，那样做在Gas 费上不占优势。
Optimism DAI Bridge 介绍资料
以上，我们介绍了四个采用见证人机制的跨层快速资产桥项目。在Hop Exchange、Hyphen、Degate Bridge、 Optimism DAI Bridge 中，见证人分别是Bonder、执行节点、Degate Home DAO 和Maker Oracle。其中，Bonder 的信誉来自于抵押，执行节点、Degate Home DAO 的信誉来自于链下商誉，包含了信任假设，Maker Oracle 则比较特别，如果Maker Oracle 行为不当，承受损失的是DAI 财政库，也就是MKR 的持有者，相当于MKR 持有者替Maker Oracle 作了担保抵押。
在流动性垫付职能方面，Bonder 和Degate Home DAO 兼任了流动性提供者的功能，执行节点和Maker Oracle 则只负责传递跨层信息，Hyphen 的流动性提供者是LP，Optimsm DAI Bridge 的流动性提供者是DAI 财政库。
FundMovR 和DataMovR，都出自同一个项目—— MovR Network，二者是MovR Network 的两个独立模块。聚合和互操作是跨层快速通道进阶演化的两个重要方向，MovR Network 准确的抓住了这两个方向，可以看出对于赛道的理解非常深刻。
跨层资产桥已经足够多了，FundMovR 选择做一个聚合器，将各资产桥聚合起来，并为用户推荐最优路径。由于各跨层资产桥都只支持同质资产的兑换（例如，USDT 只能兑换USDT，不能兑换其他资产），FundMovR 不光聚合了跨层资产桥，还聚合了各账本上的Dex ，以便用户可以通过FundMovR 直接完成异质资产的兑换。
假设Alice 在Arbitrum 上有DAI，但她换成Optimism 上的ETH。她可以使用多种路径来实现：
- 路径一：通过1inch on Arbitrum，把DAI 换成ETH ，然后通过Hop Exchange，把ETH 从Arbitrum 换到Optimism ；
- 路径二：通过Paraswap on Arbitrum，把DAI 换成ETH ，然后通过Connext 将ETH 从Arbitrum 换到Optimism；
- 路径三：通过Hyphen，把DAI 从Arbitrum 换到Optimism，然后通过Uniswap on Optimism，把DAI 换成ETH ；
- 路径四：通过cBridge，把DAI 从Arbitrum 换到Optimism ，然后通过Sushiswap on Optimism， 把DAI 换成ETH。
假设兑换资金量很大，那么含有AMM 机制的Hyphen、Hop 会不可取，因为滑点会比较大，cBridge 会比较合适；
假设兑换资金量很小，Gas 将成为主要成本，为了节约Gas，Hop Exchange 可能是更好的选择；
假设Alice 对速度的要求是最优先的，那么采用中心化方式的Hyphen 和Degate Bridge 会更有优势；
至于DEX 兑换环节，是在源账本Arbitrum 上完成，还是在目标账本Optimism 完成，取决于哪边流动性更大，滑点更低。
如果没有FundMovR，上述权衡和考量，将由Alice 自己完成。但通过FundMovR ，系统会自动找到所有可用的路线，并分别以下面三条标准进行排序：
- 最低Gas 费用
假设Alice 想将100 DAI 从Optimism 转移到Arbitrum，而Bob 想将50 DAI 从Arbitrum 转移到Optimism。FundMovR 将相互清算DAI，并将剩余的50 个DAI 从Optimism 转移到Arbitrum。
图片源于MovR Network 文档
对于跨层部署的Dapp，桥接功能是其必要组成部分。Dapp 开发人员希望构建内置的桥接器，为用户提供无缝的用户体验。FundMovR 通过SDK 和API 使Dapp 开发人员可以轻松集成FundMovR，以实现内置的效果，例如，像Aave、Instadapp 等Dapp 可以轻松地允许用户从不同L2 迁移用户抵押资金。
DataMovR 将允许任意形式的状态转换跨层传递，这意味着DataMovR 将支持任意形式的跨层互操作。DataMovR 有两个独立的组件，分别是负责消息传递的MMF(MovR Messaging Framework)，和负责先于L1，快速验证L2 状态的一个见证人网络，该网络被称为Finality Gadget （终结器）。
与FundMovR 相同，DataMovR 也将被开发为一个可内置于Dapp 的模块。导入该模块的Dapp 将在免于自己开发的前提下实现跨层互操作。例如：
- Yearn Finance 可以在诸多的L1，L2，乃至侧链上，寻找最优的收益率，并在此基础上采取行动；
- 可以允许用户在Arbitrum 上的Aave 销毁aToken，并在Optimism 上的Aave 上提取质押；
- Uniswap 可以支持LP 跨层快速挪动流动性，以实现各层价格的均衡。
DataMovR SDK 将帮助Dapp 在各Layer 上部署的合约可以相互通信，以实现以上用例。
更进一步，DataMovR 还有野心实现跨Dapp 的跨层互操作，例如让用户可将Aave on Optimism 当中的抵押品挪动到Uniswap on Arbitrum 中提供流动性，或是让用户将Compound on Arbitrum 的抵押品挪动到Aave on Polygon。这点的实现，有赖于足够多的Dapp 内置了DataMobR SDK。
关于DataMovR 的更多技术细节，MovR Network 暂时没有披露更多，Paka Labs 将持续关注。
2021 年以来，业界对以太坊扩容的努力方向发生转向，人们放弃了对改进以太坊本身的期待，而是将目光投向了L2 和EVM 兼容链，相关项目喷发式出现：
Op Rollup 日趋成熟，主流DeFi 项目陆续开始向之迁移；拥有巨大用户体量加持的HECO，BSC 被构建为了以太坊的高性能侧链；EAR、Fantom、Avalanche、Substrate 等以底层创新著名的区块链也快速地加入了EVM 版图；进展最慢的兼容EVM 的Zk Rollup 也呼之欲出。
一个以Ethereum 为中心的，众多L2、EVM 兼容链组成的泛EVM 多链网络已逐渐形成。尽管EVM 的执行效率低于WASM，但是在巨大的惯性力量和成熟的DeFi 代码库的加持下，EVM 依旧凝聚起来了最大的势能。
在这样一个泛EVM 多链网络中，跨层快速资产桥是其重要基础设施，也是其不可或缺的一部分。快速通道将承担大多数的资产交易，原始通道只负责流动性的结算，释放以太坊的压力。除此之外，我们更加期待，快速资产桥演变为快速互操作桥，让整个泛EVM 生态的DeFi 连为一体，产生乐高效应。
当然，我们也得认识到，泛EVM 网络或许还不是以太坊生态演化的终局，还有很多其他方向的尝试，例如Celer 开发的Layer2.Finance 和StarkWare 开发的Caspin 正在尝试在不割裂L1 流动性的情况下进行原地扩容，Polygon 则正在开发以自身为中枢的EVM 分片网络，ETH2.0 虽然要等待很久，但终究会来，到时候，以太坊生态会是什么样，犹未可知，让我们拭目以待。
本文列举了两大跨链应用形态，分别是「BTC 锚定资产」和「跨层快速通道」。我们分析了两者的技术本质，并根据其核心技术特征，进行了分类，在分类的基础上，结合项目举例进行了深入分析。至此，我们在对跨链技术有全景式框架认知的技术上，对「BTC 锚定资产」和「跨层快速通道」两大应用形态进行了纵深探究。在后续的篇目中，我们将对更多的应用形态进行举例分析。
- BTC 锚定资产
- cBridge/Connext/StarEx Bridge
- Hop Exchange/Hyphen/Degate Bridge
本文的案例选取仅从技术方案的代表性角度考量，不含推广目的，亦不作为投资建议。如果您发现文中所写的内容与实际的情况不符，或者对本文观点存在不同的理解，欢迎关注PAKA Labs 公众号并留言，非常期待与您交流。
感谢Nic Lin @imtoken 在本文写作过程中提供的帮助。
Blockcast.cc does not endorse any content or product on this page. While we aim at providing you all important information that we could obtain, readers should do their own research before taking any actions related to the company and carry full responsibility for their decisions, nor can this article be considered as investment advice or recommendations. Every investment and trading move involves risk, you should conduct your own research when making a decision.