ZK Co-processor: A New Paradigm for Blockchain Applications

Co-processors in the field of computing are responsible for handling complex tasks for the CPU. For example, in 2013, Apple launched the M7 motion co-processor to enhance the motion sensitivity of smart devices, and in 2007, Nvidia introduced the GPU co-processor to handle tasks such as graphic rendering. Co-processors can offload complex and performance-intensive code, allowing the CPU to handle more flexible parts.

There are two issues on the Ethereum blockchain that constrain the development of applications:

1. High Gas fees limit the scope of on-chain application development. Most contract codes revolve around asset operations, and complex operations require a large amount of Gas, hindering widespread adoption of applications and users.

2. Smart contracts can only access data from the most recent 256 blocks. Future upgrades will result in full nodes no longer storing past block data. Data loss makes it difficult for data-driven innovative applications to emerge, limiting the appearance of "mass adoption" products.

These issues stem from the fact that the original design intention of the Ethereum Blockchain was not to handle large-scale computational and data-intensive tasks. To accommodate these applications, the concept of co-processors needs to be introduced. The Ethereum chain acts as the CPU, while the co-processors are similar to GPUs, handling computational and data-intensive tasks. By integrating ZK technology, it can ensure that co-processors perform trustworthy calculations and data usage off-chain.

The application scope of ZK co-processors is extensive, covering social, gaming, DeFi, risk control systems, oracles, data storage, large model training, and more. Theoretically, all functions that Web2 applications can achieve can also be realized by ZK co-processors, utilizing Ethereum as the settlement layer to ensure application security.

Currently, the definitions of ZK co-processors in the industry vary. Terms like ZK-Query, ZK-Oracle, and ZKM all belong to this category, which can assist in querying complete on-chain data, off-chain trusted data, and off-chain computation results. From a certain perspective, Layer 2 can also be seen as an Ethereum co-processor.

Co-processor Project Overview

Currently, well-known co-processing projects are mainly divided into three categories: on-chain data indexing, oracles, and ZKML. The General-ZKM project encompasses all three scenarios. Different projects employ different off-chain virtual machines, such as Delphinus focusing on zkWASM and Risc Zero focusing on Risc-V architecture.

Co-processor Technology Architecture

The following analyzes several typical general-purpose ZK co-processor projects, discussing the similarities and differences in their technology and mechanism design.

Risc Zero

The ZK co-processor of Risc Zero is named Bonsai, which is a set of blockchain-independent zero-knowledge proof components. Based on the Risc-V instruction set architecture, it supports multiple languages such as Rust, C++, Solidity, and Go. Its main functions include:

1. Universal zkVM, capable of running any virtual machine in a zero-knowledge/verifiable environment.
2. ZK proof generation system that can be directly integrated into smart contracts or chains.
3. General rollup, distributing the computations proven on Bonsai to the blockchain.

The key components of Bonsai include the Prover Network, Request Pool, Rollup Engine, Image Hub, State Store, and Proving Marketplace.

Lagrange

Lagrange aims to build co-processors and verifiable databases that contain Blockchain historical data to facilitate the development of trustless applications. Its core features include:

1. Verifiable Database: Index chain on smart contract storage, reconstructing Blockchain storage, state, and Block.
2. Computation based on MapReduce principles: using data separation with multiple instances for parallel computation, and finally integrating the results.

Lagrange's database design covers contract storage data, EOA state data, and Block data. Its ZKMR virtual machine uses Map and Reduce steps for computation and proof.

Succinct

Succinct Network is committed to integrating programmable facts into all aspects of Blockchain development. Its co-processor supports multiple programming languages, including Solidity and a specialized zero-knowledge domain language. Succinct's off-chain ZKVM is called SP(Succinct Processor), supporting Rust and other LLVM languages.

Comparative Analysis

When comparing general-purpose ZK co-processors, the following factors should be considered:

1. Data Indexing/Synchronization Capability
2. The adopted ZK technology ( SNARKs vs STARKs )
3. Does it support recursion?
4. Proof System Efficiency
5. Ecological Cooperation Status
6. Financing and VC Support

Overall, the technical paths of various projects are converging, such as using wrappers from STARKs to SNARKs, supporting recursion, building prover networks, and cloud computing markets. In the case of similar technologies, team strength and the ecological resources of the VC behind them may become key competitive factors.

The difference between co-processors and Layer 2

Unlike user-oriented Layer2, co-processors are application-oriented. They can serve as acceleration components or modular components for the following scenarios:

1. As an off-chain virtual machine component of ZK Layer2
2. Provide off-chain computing power for public chain applications
3. Oracle for public chain applications to obtain verifiable data from other chains
4. Act as a cross-chain bridge for message passing

The co-processor brings the potential for real-time synchronized data across the entire Blockchain and high-performance, low-cost trusted computing, which can be used to reconstruct various middleware for Blockchain.

Challenges Facing Co-Processors

1. The entry barrier for development is high, requiring mastery of specific languages and tools.
2. The industry is in its early stages, performance standards are complex, and the landscape is not yet clear.
3. The hardware and other infrastructure are not yet fully ready.
4. The technical paths are similar, making it difficult to achieve a significant technological lead.

Summary and Outlook

ZK technology has strong versatility and is expected to reconstruct multiple key aspects of the blockchain ecosystem. General-purpose ZK co-processors are one of the important tools for the implementation of ZK technology, with application boundaries covering almost all dapp scenarios.

The large-scale adoption of ZK co-processors requires meeting two key criteria: a fully chain real-time provable database and low-cost off-chain computation. As these goals are gradually achieved, the software development paradigm is expected to undergo a fundamental transformation. The commercialization of ZK computing chips will be an important prerequisite for the large-scale implementation of ZK co-processors.

Although there is a relative lack of innovation in the current cycle, this is precisely the key window period for building the next generation of "mass adoption" technologies and applications. It is expected that in the next round of cycles, the ZK industry chain will achieve commercial implementation. Now is the time to focus on those core technologies that can truly support Web3 to carry billions of on-chain user interactions.