The Competition of Parallel EVMs: Analyzing Monad, MegaETH, and Pharos

Recently, three heavyweight parallel EVM projects launched their testnets one after another. Monad made its debut on February 19, followed by MegaETH and Pharos, which went live with their testnets on March 21 and 24, respectively. This series of actions seems to indicate that, following the surge in artificial intelligence, the focus of Web3 technology is once again returning to the parallel EVM, which is the most anticipated field in early 2024.

EVM, as the core component of Ethereum, is responsible for executing smart contracts and processing transactions. However, its sequential execution model can easily lead to network congestion and latency issues under high load. Parallel EVM technology significantly improves network throughput by allowing multiple tasks to run simultaneously, thereby enhancing the overall performance and scalability of the blockchain. In fact, parallel EVM refers not only to parallel execution but also includes comprehensive upgrades from consensus mechanisms, transaction processing, pipeline optimization, storage systems to hardware acceleration, aiming to significantly reduce transaction processing time and effectively address the performance bottlenecks faced by traditional blockchains.

Monad: Pursuing a Balance Between High Performance and Decentralization

Monad is a high-performance EVM-compatible Layer 1 blockchain developed by Monad Labs. The project aims to enhance system scalability while maintaining decentralization, addressing the throughput issues of existing EVM-compatible chains.

The core advantage of Monad lies in its ability to process 10,000 transactions per second, with a block generation time of just 1 second. This performance improvement is mainly attributed to several innovations in the following areas:

1. MonadBFT: This is a high-performance consensus mechanism based on an improved HotStuff. It adopts a two-phase BFT algorithm, with linear communication overhead under normal circumstances and quadratic communication overhead in timeout cases. Additionally, MonadBFT employs a hybrid signature scheme and the RaptorCast protocol, effectively enhancing consensus efficiency and network bandwidth utilization.

2. Asynchronous Execution: By separating consensus from execution, Monad significantly increases execution throughput. This design allows the execution process to take up the entire block time, rather than just a small portion.

3. Parallel Execution: Monad adopts an optimistic execution method that allows subsequent transactions to begin execution before the prior transactions are completed. By tracking the inputs during the execution process and comparing them with the outputs of previous transactions, Monad can re-execute transactions when necessary to ensure the correctness of the results.

4. MonadDB: This is a custom KV database designed for storing verified blockchain data. MonadDB natively implements the Merkle Patricia Trie data structure on disk and in memory, and employs technologies such as asynchronous I/O and concurrency control, significantly improving data access efficiency.

MegaETH: A Layer 2 solution focused on real-time performance

MegaETH is a high-performance Layer2 blockchain developed by MegaLabs, distinguished by its pursuit of extreme real-time performance, providing ultra-low latency and high scalability for applications that require instant responsiveness.

MegaETH claims to achieve 100k TPS and approximately 10ms block times, maintaining millisecond-level response speeds even under high load. This exceptional performance is primarily attributed to the following technical features:

1. Node specialization: Nodes in MegaETH are divided into different roles, each with its own responsibilities. The sorter is responsible for transaction ordering and execution, the prover performs stateless validation, and the full node is responsible for updating local state and validating block validity.

2. Targeted Optimization: In response to various challenges faced by traditional EVM blockchains, MegaETH has implemented a series of targeted measures. For example, designing an efficient new state Trie to address the delay in state data retrieval, and using a JIT compiler to improve interpreter efficiency.

3. Mini Blocks: MegaETH performs a pre-confirmation every 10 milliseconds, known as Mini Blocks. This design significantly reduces the time it takes for transactions to propagate to the rest of the network while providing a more efficient data retrieval method for light clients.

Pharos: Full-stack Parallelized EVM Compatible Layer1

Pharos is positioned as a high-performance EVM-compatible Layer 1 blockchain, dedicated to providing optimal solutions for the RWA and payment ecosystem. The project claims to handle 50,000 transactions per second, consuming 2 billion units of gas per second.

Pharos proposed the "Degree of Parallelization ( DP )" framework, which divides the parallelization capability of blockchain into six levels (DP0-DP5). Pharos itself adopts a DP5 full-stack parallel architecture, undergoing comprehensive upgrades from consensus to hardware acceleration:

1. Scalable consensus protocol: Adopts a high-throughput, low-latency BFT consensus protocol.

2. Dual virtual machines execute in parallel: EVM and WASM execute in parallel using advanced compilation technology.

3. Full Lifecycle Asynchronous Pipeline: Achieving parallel and asynchronous processing of the entire transaction lifecycle and between blocks.

4. High-performance storage: Utilizing certified data structure (ADS), providing high throughput, low latency I/O, and cost-effective state storage.

5. Modular Special Processing Network (SPN): Seamlessly integrates new technologies and supports diverse application scenarios.

Summary

EVM has become a presence in the Web3 world similar to JavaScript in Web2, with the most developers and the largest DApp ecosystem. However, Ethereum's scalability issues have severely restricted the further development of EVM, making parallel EVM one of the most important technical directions.

Monad seeks to balance scalability and decentralization through its parallel execution model, providing developers with 10,000 TPS throughput while maintaining EVM compatibility. Its independent consensus mechanism offers autonomy but also means giving up the security guarantees of Ethereum.

MegaETH performs exceptionally well in terms of latency and throughput, with an ultra-low latency of 10 milliseconds and a throughput of 100,000 TPS, making it particularly suitable for applications that require near-instantaneous responses. However, its centralized ordering design may raise concerns regarding decentralization.

Pharos demonstrates performance comparable to Monad and MegaETH, while focusing on institutional clients and regulatory requirements for RWA-Fi with its "ant gene", making it likely to meet the future market demand for compliant and efficient blockchain infrastructure.

Although MegaETH and Pharos show superior performance in public data, considering the large financing obtained by Monad and the potential technological breakthroughs it may bring, there is actually no absolute leader among these three projects. Developers need to weigh the priorities of performance, level of decentralization, and specialization when making a choice.