With continued improvements in scalability, many new high-performance blockchains have been formed with a focus on parallel processing to scale. Solana emerged as the largest and most prominent of these high-performance blockchains, pioneering this approach, which has gone on to drive significant developer adoption and usage. Following in its footsteps, several other high-performance blockchains, including Sui, Monad and MegaETH, have come into focus. In this piece, we will look at these chains to understand how they each approach parallel processing, how they differ, and the future of high-performance blockchains.
Given that many of these new high-performance chains are built using the EVM (Ethereum Virtual Machine), it’s helpful to recap Ethereum and its approach to scalability. Ethereum uses discrete block-building in which blocks are constructed sequentially. Blocks are produced on average every 12 seconds. The current block gas limit on Ethereum is 30m. Ethereum currently achieves around 13 TPS on the mainnet, a 21x improvement from 3 years ago. The recorded TPS peak was 62, with a theoretical maximum of 119 TPS. The throughput of Ethereum combined with its associated rollups achieves closer to 334 TPS across over 105 rollups.
Ethereum’s eventual goal is to achieve 100k TPS across the combined mainnet and its associated rollups as part of the surge roadmap. To achieve this, Ethereum needs to implement PeerDAS (Peer Data Availability Sampling), an upcoming upgrade that aims to enhance data availability scaling and allow validators to verify access to data without downloading entire blocks.
Vitalik Blog - Possible futures of the Ethereum protocol, Part 2: The Surge
Ethereum is built using the EVM (Ethereum Virtual Machine), a Turing-complete virtual machine that operates as the runtime environment for smart contracts on Ethereum. It has a few high-level programming languages, the most well-known being Solidity. The EVM is one of the most established virtual machines in the crypto space, leading in terms of developer mindshare, ecosystem, tooling, and user base.
While very similar to Ethereum, Solana has some distinct differences in how transactions are processed, which change the user experience for end users and block producers. Solana utilises continuous block building in which blocks are assembled in parallel and streamed dynamically over an allocated time slot. Slot times on Solana last for an average of 400 milliseconds and each leader is assigned four consecutive slots lasting around 1.6 seconds, after which it rotates to the next selected leader.
Unlike other blockchains, Solana does not have a mempool and instead operates on a leader schedule in which the network forwards all transactions to the predetermined lead validator for each epoch. The current TPS for Solana is around 708, excluding vote transactions. The max recorded TPS for Solana was 7,229 TPS, and its theoretical maximum is 65k TPS.
Helius - Solana How it works overview
The Jump Crypto team is building their new validator client, Firedancer, from the ground up, optimising it for performance based on their experience in the low-latency trading space. The Jump Crypto team claims their new unreleased Solana Firedancer client has achieved 1m TPS in their test environment. While it’s unclear how this will perform in the wild, it is believed it could bring an order of magnitude improvement to the chain in terms of performance.
Solana uses the SVM (Solana Virtual Machine), the execution environment that processes transactions and smart contracts/programs on Solana. Solana’s main languages for writing smart contracts are Rust, C and C++, which are compiled into BPF bytecode for the SVM to be executed by validators on the network.