Exploring a New Concept in the Industry—Parallel EVM, Overview of Four Projects

Former Polygon co-founder JD once stated on social media that he anticipates every L2 in 2024 will rebrand themselves and label themselves as “Parallel EVM”.

Paradigm's CTO Georgios also believes that 2024 will be the “Year of Parallel EVM” and mentioned that exploration and design of related technologies are also underway internally at Paradigm.

The Definition and Implementation of Parallel EVM

The Parallel EVM (Ethereum Virtual Machine) is a concept aimed at improving the performance and efficiency of the existing EVM. Currently, the EVM has a crucial feature in its design to maintain network consistency and security:

Transactions are executed sequentially. Sequential execution ensures that transactions and smart contracts are executed in a deterministic order, making it easier to manage and predict the state of the blockchain. This design choice prioritizes security and reduces potential complexities and vulnerabilities associated with parallel execution. However, under high loads, it may lead to network congestion and delays.

Parallel EVM allows different independent transactions or smart contracts to execute simultaneously, greatly improving the processing speed and system throughput of the EVM.

General methods for parallel EVM processing include:

• Partitioning or sharding: Transactions are partitioned or grouped, allowing them to execute in parallel. This means that different transactions can be executed simultaneously on different processing units instead of one after another. Additionally, Solana's SVM adopts a similar processing logic.

• Optimization algorithms: Development of new scheduling algorithms and optimization techniques to effectively manage and execute parallel tasks while maintaining transaction correctness and order.

• Security and consistency guarantees: Implementation of complex synchronization mechanisms and consistency models to ensure the security and data consistency of the entire system even in parallel processing scenarios.

Monad

Monad is a high-performance EVM blockchain designed to significantly enhance scalability and transaction speed through parallel execution. It optimizes four key areas to become a high-performance blockchain.

Monad BFT

Monad BFT is the consensus protocol utilized by the Monad blockchain to achieve transaction ordering consistency. It allows many nodes in the network to securely agree on the order of transactions, even in the presence of dishonest or offline nodes. Monad BFT operates on a leader-based two-phase commit approach, collecting threshold signatures from 2f+1 validators, which are aggregated into a single signature using pair-based cryptographic techniques. This enhances efficiency, scalability, and reduces storage space compared to individual signatures.

Separate Execution

Monad separates consensus and execution to improve performance. The consensus layer involves nodes agreeing on the order of transactions but does not execute transactions. The execution layer consists of nodes independently executing sorted transactions to update the state. By separating consensus and execution, Monad achieves faster consensus through larger transaction batches. With single slot finality achieved in 1 second, execution latency is less than 1 second, ensuring consistency while achieving high transaction throughput on a single shard. This architecture decouples consensus and execution to optimize transaction ordering and processing.

Parallel Execution

Monad allows for parallel execution of transactions within a block to enhance efficiency. It employs an Optimistic approach, initiating the execution of new transactions before the completion of the previous step. To handle incorrect outcomes, Monad tracks input/output and re-executes inconsistent transactions. A static code analyzer predicts dependencies, avoiding invalid parallelism and reverting to a simple mode when uncertain. This parallel execution increases throughput while reducing the likelihood of transaction failures.

Monad DB

Monad executes asynchronous I/O (Input/Output operations) to achieve parallel transaction execution, unlike traditional systems where one must wait for I/O results before processing the next transaction. With asynchronous I/O, the CPU can handle other transactions without waiting for specific I/O results. Monad DB leverages advanced Linux kernel features to facilitate efficient asynchronous disk operations, eliminating the restrictions of synchronous I/O. Traditional Ethereum databases lack support for asynchronous I/O, but Monad DB is designed to fully utilize it. This asynchronous approach significantly improves transaction processing efficiency while maintaining Ethereum compatibility.

Sei

Sei is an open-source Layer 1 blockchain designed specifically for transaction optimization, aiming to provide advanced infrastructure for various transaction applications, including DeFi, NFT markets, and game DEXs.

In the latest V2 version (expected to be implemented in the first half of 2024), Sei prioritizes parallel EVM. Sei V2 represents a significant upgrade to the Sei network, aiming to become the first fully parallel EVM. This upgrade will provide the following features for Sei:

Backward Compatible EVM Smart Contracts

Developers will be able to directly port smart contracts that have been audited on other EVM-compatible chains to Sei without any code changes. Sei nodes will import Geth (the Go implementation of the Ethereum Virtual Machine), which will be used to process EVM transactions. Any updates resulting from this (including state updates or calls to non-EVM related contracts) will be facilitated through a special interface created by Sei for the EVM. This is a significant upgrade as it simplifies the process of migrating existing smart contracts from other blockchains such as Ethereum to Sei.

Optimistic Parallelization

Like Monad, Sei V2 will also utilize Optimistic parallelism. This allows the blockchain to execute transactions simultaneously without developers needing to define any dependencies. When conflicts occur, the blockchain will track the storage portions each transaction touches and rerun these transactions sequentially. This process will recursively continue until all unresolved conflicts are resolved.

Sei DB

Sei DB is a novel dual-component data structure designed to optimize storage and prevent blockchain bloat. It separates state storage from state commitment, deviating from the traditional single IAVL tree design. Compared to a single tree, this dual structure reduces latency and disk usage while enhancing multi-threaded read/write performance.

Neon EVM

Neon EVM is a fully compatible EVM (Ethereum Virtual Machine) solution built on the Solana blockchain. It operates as smart contracts within Solana, allowing developers to deploy Ethereum DApps without any reconfiguration.

Neon EVM enables DApps built with Solidity and Vyper to leverage Solana's network advantages: low fees, high transaction speeds, and parallel transaction execution capabilities. This means developers can create Solana-compatible contracts using familiar Ethereum ecosystem tools.

Neon EVM ensures key Ethereum DApp tools are compatible with Solana, including Vyper, Solidity, and MetaMask. This solution allows any Ethereum application to run on Solana with minimal reconfiguration, including Uniswap, SushiSwap, 0x, and MakerDAO.

Eclipse

Eclipse is the next-generation Optimistic Layer 2 solution on Ethereum, supported by the Solana Virtual Machine (SVM). By integrating Ethereum's settlement functionality with SVM's high performance and parallel execution, the Eclipse mainnet offers a blend of speed, scalability, and security. It settles transactions on Ethereum and uses ETH as the

gas token

, ensuring compatibility and security while enhancing transaction throughput.

Eclipse aims to address Ethereum's scalability challenges, leveraging Celestia for data availability and RISC Zero for zero-knowledge proofs, aiming to achieve greater transaction capacity without sacrificing security and decentralization. The Eclipse team will closely monitor Ethereum's EIP-4844 upgrade and consider migrating to Ethereum's DAO in the future, further enhancing Eclipse's security.

SVM and its Sealevel runtime support parallel transaction execution. Similar to Sei V2 and Monad, transactions that do not involve overlapping states can be executed in parallel instead of sequentially.