Compared to zkRollup, the zkOmnichain in the Layer0 track is more worth paying attention to

Author: Kang Shuaiyue, Founder of Fox Tech and Way Network, Chairman of Danyang Investment

Preface: zkRollup belongs to the Layer2 track, aiming to solve the congestion problem of Ethereum; zkOmnichain belongs to the Layer0 track, aiming to build the entire Web3 network. However, due to the significant attention on the Ethereum ecosystem, most focus is on zkRollup, neglecting zkOmnichain.

If Ethereum's market share can maintain above 70% for many years, then Omnichain itself is a pseudo-concept. However, Ethereum cannot maintain a 70% market share forever. The following chart shows the trend of Ethereum's TVL share from August 2020 to February 2023. Ethereum's latest market share is 59.24%. Although the chain-layer infrastructure of Web3 will likely maintain a "one strong, many strong" pattern for a long time, and Ethereum itself continues to upgrade its technology, with Layer2 solutions like zkRollup and opRollup absorbing traffic from Layer1, and challengers like BNB Chain and Polygon growing rapidly, along with the swift rise of new public chains like Atpos and Sui, and the rapid spread of the Cosmos ecosystem, we will see Ethereum's market share remain below 50% for an extended period after 2025.

Figure 1: Ethereum TVL share from August 2020 to February 2023, data source: defillama

Almost all applications after the DeFi summer of 2020 were implemented through the most primitive method of deploying each chain individually to achieve multi-chain status. This simple and crude approach not only causes significant trouble for developers in deployment and operation but also leads to two major issues: severe fragmentation of liquidity across chains and significant price differences for the same token between chains.

The path for applications to achieve full-chain coverage, besides the aforementioned primitive method, includes three main paths: The first path, which is a common path for first-generation cross-chain products, involves applications inter-operating with the starting chain and target chain via an intermediary consensus protocol. Applications typically do not deploy on this intermediary chain, with typical representatives being the self-contained ecosystems of Cosmos, Polkadot, as well as Multichain, Celer, Wormhole, Thorchain, and Axelar; The second path is to achieve interoperability with other chains through an Optimistic mechanism. Typical representatives of this path include Synapse, Nomad, and Hyperlane, which use Optimistic Verification & Fraud Proofs, as well as LayerZero and Across, which use Optimistic Oracle & No Proofs; The third path is for applications to inter-operate with other chains through the zero-knowledge proof-based full-chain interoperability protocol zkOmnichain, with typical representatives being Electron, Polymer, and Succinct from the Cosmos ecosystem, Overearlity focusing on NFT cross-chain, and Way Network dedicated to providing general solutions.

Figure 2: Comparison of historical cross-chain protocols

Lastly, there is a path where applications are deployed on a brand new smart contract public chain and inter-operate with other chains through this chain, with typical representatives being ZetaChain and Map Protocol. This is a path that is heavy in both development and operation. First, it needs to solve the ledger alignment problem faced by first-generation products when cross-chain with EVM public chains. Secondly, due to overlapping functionalities with existing smart contract public chains, it inevitably needs to compete directly for application resources. Regardless of the path taken, they all aim to transform single-chain applications into full-chain applications.

The first-generation intermediary chain (consensus protocol) model suffers from slow performance with too many verification nodes and insecurity with too few, and node operation is particularly heavy, making it difficult to take on the responsibility of the Web3 Network Fabric. Thus, we can only focus on non-intermediary chain (consensus protocol) models, specifically examining the second-generation product model opOmnichain and the third-generation product model zkOmnichain.

Figure 3: Comparison of OP and ZK full-chain interoperability protocols

The full-chain interoperability protocol Layer0 track, like Ethereum's scaling solution Layer2 track, has formed two major categories of solutions: OP and ZK. First, let's look at the Optimistic Oracle sub-model within the OP solution, where LayerZero uses Chainlink oracles, and the Chainlink project also has its own cross-chain interoperability protocol (Cross-Chain Interoperability Protocol, CCIP), with both projects cooperating and competing. Across uses UMA as its oracle, and UMA and Across are two projects from the same team.

Such solutions cannot achieve true decentralization and lack Shared Security. Shared Security means that specific tokens or applications running on a given infrastructure cannot freely choose their security model but must comply with any security requirements provided by the infrastructure. All Layer1 and Layer2 have Shared Security, and Layer0 should also have Shared Security, rather than Isolated Security, because only in this way can unified underlying security guarantees be provided to end users, regardless of which application they use or what the internal security policies of the application are. The term "infrastructure" implies this foundational aspect.

Moreover, the aforementioned Optimistic Oracle has strict trust assumptions, requiring end users to trust that the oracle will not validate randomly. In this case, the oracle acts as a Third Trusted Party in this model. Users must trust that the oracle and relayer will not collude to steal user assets. Additionally, such solutions do not produce any Fraud Proofs or Validity Proofs, nor do they have on-chain verification Proofs. Various drawbacks and issues have been pointed out by teams like L2BEAT, Nomad, and Way Network, and will not be discussed further here.

Next, we look at the Optimistic Verification sub-model within the OP solution, which is more rigorous and cautious than the aforementioned Optimistic Oracle and produces Fraud Proofs. Messages are optimistically signed on the source chain, and a network of off-chain validators is responsible for submitting fraud proofs to the target chain within the enforced optimistic window period (Timeout Period). During this window period, fraudulent messages may be challenged. If any errors are noticed, validators can check and reject the messages and prohibit any fraudulent transactions. Compared to the first-generation products that rely on the honest majority assumption of external verification networks, Optimistic Verification relies on a single honest validator assumption. Optimistic verification only requires one honest validator initiating a change to safeguard the system's security. However, the tradeoff is the delay introduced by this window period. Fortunately, this latency is not as long as that of opRollup; instead of waiting 7 days, it is a matter of tens of minutes, for example, Nomad sets a 30-minute Timeout Period.

Finally, we look at the most important zkOmnichain, which belongs to Zero-Knowledge Verification and has better robustness. zkOmnichain, like zkRollup, also produces Validity Proofs. A crucial step is calling the Sum-Check Protocol module, where proofs are made off-chain (Sum) and verified on-chain (Check). Through this rigorous mathematical and cryptographic verification, completeness and soundness are achieved. As for succinctness, Groth16 can be used to further compress Proof Size. For instance, the original Proof Size generated by Way Network's zkSpark is about 100KB, but after Groth16 compression, it can be reduced to 130B.

Figure 4: Way Network zkPark Sum-Check Protocol

With the emergence of zkOmnichain, a "strong cryptography" full-chain interoperability protocol, we can finally achieve the highest standard for infrastructure products, which is the removal of any third-party trusted entities for trustlessness and the removal of any central privileged entities for decentralization. With the advent of zkOmnichain, we finally see the possibility of transforming from "chains" to "networks," and the feasibility of building a Layer0 infrastructure that can support large-scale full-chain application deployments across various smart contract public chains.

What changes will full-chain applications based on zkOmnichain bring? First, it eliminates the intermediary chain from the first-generation cross-chain products, reducing the costly operational costs of verification nodes; secondly, it removes the Intermediate Token and Wrapped/Pegged Token derivative tokens, allowing users to complete Native Token cross-chain transactions without needing to trust these tokens. Finally, it removes the third-party trusted entity oracle, which helps improve the resilience of the full-chain network.

Full-chain applications based on zkOmnichain do not rely on intermediary chains, nor do they depend on oracles, and they do not have intermediary tokens or derivative tokens. They achieve decentralization and trustlessness through rigorous mathematics and cryptography for validity proofs, verified on-chain, enabling peer-to-peer full-chain communication similar to the Bitcoin network.

From Singlechain to Multichain to Omnichain, this is the basic trend of the Web3 wave. In the future, full-chain applications will no longer be sparse as they are now but will be ubiquitous. Applications based on zkOmnichain will have higher cross-chain security, lower communication costs, and faster communication speeds, making them more likely to succeed.

Figure 5: Full-chain applications based on zkOmnichain

Recently, numerous full-chain interoperability protocols aimed at connecting various blockchains into a true Web3 have emerged. In the context of Web2, Cosmos is akin to building a Local Area Network (LAN), while Way Network is dedicated to constructing a Wide Area Network (WAN) or even the Internet. The standard for judging the quality of such products lies not in the number of chains the protocol can connect to or other factors, but in whether the access method embodies Satoshi's consensus, namely Decentralized and Trustless, because this is "1"; everything else is "0." Without "1," no amount of "0" is useful.

Conclusion

It is foreseeable that full-chain applications will gradually replace single-chain/multi-chain applications as the mainstream deployment model for Web3 applications in the next three years, and the zkOmnichain Interoperability Protocol is the key technological means and infrastructure to achieve this process. Soon, we will see full-chain AMM applications, full-chain lending applications, full-chain SocialFi, full-chain NFTs, and more emerging like mushrooms after rain, all based on zkOmnichain.