ZkOmnichain Deserves More Attention Than ZkRollup
By Frederick Kang, founder of Fox Tech and Way Network and chairman of Danyang Investment
If Ethereum itself can maintain a 70% + market share for years, then Omnichain itself is a fake concept. However, Ethereum can't maintain a 70% market share forever. The chart below shows the share trend of Ethereum TVL from August 2020 to February 2023. Ethereum's latest market share is 59.24%. The Chain layer infrastructure of Web3 maintains the pattern of "one super play, many strong players" for a long time, and Ethereum itself is constantly upgraded. However, with the traffic absorption of Layer2 of Ethereum like zkRollup and opRollup to Layer1, and the strong growth of challengers like BNB Chain and Polygon, With the rapid rise of new public chains such as Atpos and Sui, and with the Cosmos system fast growing, we will see Ethereum itself below 50% market share for a long time after 2025.
Almost all applications after DeFi Summer 2020 are implemented in a multi-chain state by deploying each of the original chains individually. This simple and crude approach not only brings a lot of trouble to developers in deployment and operation, but also leads to two major problems: the liquidity is severely divided in each chain, and the price difference of the same token on two chains is huge.
In addition to the above original approach, there are three main paths for the applications to achieveomnichain coverage: The first path is also a common path of the first generation of cross-chain products, that is, the Application interoperates with the start chain, and the target chain through an intermediate consensus protocol. Application is usually not deployed on this intermediate chain. The typical representation is Cosmos and Polkadot, which are their own systems. Multichain, Celer, Wormhole, Thorchain, Axelar; The second path is to interoperate with the remaining chains through Optimistic mechanisms. A typical example of this type of path subdivision is Optimistic Verification & Fraud Proofs, such as Synapse, Nomad and Hyperlane. Optimistic Oracle & None Proofs, such as LayerZero and Across; The third path is that the applications interoperate with other chains through an Omnichain interoperability protocol based on zero-knowledge proof. The typical representative is Electron, Polymer, Succinct, which all come out from Cosmos system. Overearlity, the focusing on the NFT across chains, as well as the Way Network dedicated to providing universal solutions.
Finally, there is a path where the Application deploys to a new smart contract public chain through which it interoperates with other chains, typically ZetaChain and Map Protocol. This is a heavy development and operational path. First of all, we need to solve the problem of book alignment when EVM public chains cross heterogeneous chains, which is the same as the first generation of products. Secondly, due to the overlap of functions with existing smart contract public chains, they must compete directly with them for application resources. But no matter which path you take, you are shared the same target to turn single-chain applications into omnichain applications.
In the first-generation intermediate chain (consensus protocol) mode, if there are too many verification nodes, it is slow, and if there are too few verification nodes, it is unsafe. Moreover, node operation is extremely heavy, which makes it difficult to assume the major role of Web3 Network Fabric. In this way, we can only focus on the non-intermediate chain (consensus protocol) model, focusing on the second-generation product model opOmnichain and the third generation product model zkOmnichain.
Omnichain interoperability protocol Layer0, like Ethereum expansion scheme Layer2, forms OP and ZK solutions. Let's first look at the Optimistic Oracle submodel in the OP scenario, where LayerZero uses a Chainlink oracle, Chainlink has its own Cross-Chain Interoperability Protocol (CCIP), which makes them both partners and competitors. Across uses UMA as the oracle. As you know, UMA and Across are two projects belonging to the same team.
Such schemes cannot realize decentralization in the true sense and do not have Shared Security. Shared Securit means that a particular Token or Application running on a given infrastructure is not free to choose its security model, but must comply with whatever security requirements the infrastructure provides. All Layer1 and Layer2 have Shared Security, and Layer0 should also have Shared Security, rather than Isolated Security, because only in this way can unify the security provided for end users. And no matter what application they are using, no matter what the internal security policies of the application are. That's what the word "infrastructure" means.
In addition, the above Optimistic Oracle has strict trust assumptions, and the end user must trust that Oracle will not be randomly validated. At this point, Oracle is A Third Trusted Party in this mode. Users must trust that Oracle and Relayer are not conspiring to steal their assets.
Moreover, such schemes would not produce any Fraud Proofs or Validity Proofs Proofs, nor could there be any on-chain verification proofs. All kinds of disadvantages and problems, L2BEAT, Nomad, Way Network, and other teams have written articles to point out.
Next, let's look at the Optimistic Verification sub-model in OP scheme, which is more rigorous and prudent than the Optimistic Oracle above, and generate Fraud Proofs. Messages are optimistically signed on the source chain, and a network of validators below the chain is responsible for submitting proof of fraud to the target chain during the enforced Optimistic Window (Timeout Period). During this window, fraudulent messages may be challenged. If any errors are noticed, the Validator can examine and reject the message and prohibit any fraudulent transactions. Optimistic Verification relies on the single honest verifier hypothesis, while the external verification network of first-generation products relies on the honest majority hypothesis. Optimistic validation only requires an honest Validator to issue a Challenge to secure the system. But the Tradeoff is precisely the delay that comes with this window. Fortunately, this Latency is not as long as opRollup, and instead of waiting seven days, just wait tens of minutes, as Nomad, for example, has a 30-minute Timeout Period.
Let's turn finally to the most important one, zkOmnichain. This model is Zero-Knowledge Verification, so it has better Robustness. zkOmnichain, just like zkRollup, will also have Validity Proofs. One of the important links is to call the sum-Check Protocol module, generate proof (Sum) off chain, and do verification (Check) on chain. Through this very rigorous mathematical and cryptographic verification, to achieve Completeness and Soundness. As for Succinctness, Groth16 can be used to further compress the Proof Size. For example, the original Proof Size produced by Way Network's zkSpark is about 100KB, but it can be reduced to 130B after being compressed by Groth16.
It is with zkOmnichain, a "strong cryptography"omnichain interoperability protocol, that we can finally achieve the highest standard of infrastructure products, that is, Trustless, without any third trusted party, and the Decentralized, none central privilege body. With zkOmnichain, we finally see the possibility of moving from a "chain" to a "network" and the possibility of building an infrastructure called Layer0 that can support the deployment of large-scale omnichain applications to various smart contract public chains.
What changes will an omnichain application based on zkOmnichain bring? First, the Intermediate Chain in the first generation of cross-chain products was removed, reducing the expensive operation cost of verification nodes. Second, Intermediate Token and Wrapped/Pegged/derivative tokens are removed, so users can complete Native Token cross-chain transactions without trusting these tokens. Finally, the third trusted party Oracle Oracle is removed, which helps improve the resilience of the whole chain network. The zkOmnichain application does not rely on intermediate chains, nor does it rely on oracle, nor does it have intermediate tokens or derivative tokens. It uses rigorous mathematics and cryptography to prove the validity and on-chain verification to realize the goals of Decentralized and Trustless, and realize peer-to-peer omnichain communication just like Bitcoin network.
From Singlechain to Multichain to Omnichain, this is the basic trend of the Web3 wave, where omnichain applications will not be as sparse as they are now, but ubiquitous.
zkOmnichain-based applications have higher cross-chain security, lower communication costs, and faster communication speeds, so they are more likely to succeed.
In the past few years, there has been a proliferation of omnichain interoperability protocols aimed at connecting the various blockchains into a veritable Web3. Compared with the context of Web2, Cosmos is equivalent to the construction of a Local Area Network (LAN), while Way Network is dedicated to the construction of a Wide Area Network (WAN) and even the Internet. The criterion to judge the quality of such products lies not in the number of chains or other things that can be accessed by this protocol, but in whether the access methods are Decentralized and Trustless, i.e., the consensus of Satoshi Nakamoto, because this is "1" and others are "0", without 1, 0 is useless.
Conclusion: The zkOmnichain Interoperability Protocol is a key technical tool and infrastructure for Web3 applications, which will gradually replace single-chain/multi-chain applications as the dominant deployment mode in the next three years. Soon, we'll see zkOmnichain-based AMM applications, zkOmnichain-based Lending applications, zkOmnichain-based SocialFi, zkOmnichain-based NFT, and so on.