Edge Podcast - A Beginner’s Guide to Monad’s Parallelized EVM - Revelo Intel

Edge Podcast – A Beginner’s Guide to Monad’s Parallelized EVM

In this episode of Edge Podcast, which took place on February 14, 2024, DeFi Dad, Nomatic, and Keone discuss Monad, its features, parallel EVMs, and more. Read our notes below to learn more.

Background

  • DeFi Dad (Host) – investor at 4RC Capital
  • Nomatic (Host) –  investor at 4RC Capital
  • Keone Hon (Guest) – Co-Founder of Monad
  • Monad – a layer-1 that provides parallel execution to the EVM

Background and founding Monad

  • Keone says that he has worked in crypto for several years, including at Jump Trading and on Solana DeFi projects. He joined Jump Trading in 2021, focused on Futures Trading, and helped scale up the team. He saw the need for performant EVM solutions and left Jump to start Monad Labs with two others.
  • He adds that he left Jump Trading at the beginning of 2022 along with two other individuals. He started Monad Labs as an effort to make the EVM more performant. There is a significant need for scalable EVM solutions due to high transaction fees.

Why we need Monad

  • Keone says that Monad aims to make the EVM extremely performant. Ethereum currently processes about a million transactions per day, while Monad aims to increase this throughput to a billion transactions per day. The focus on daily transactions rather than transactions per second helps envision successful applications with mass user adoption. Applications with millions of daily active users would require significantly higher throughput than what is currently available.
  • He adds that Monad takes a different direction compared to other layer-2 solutions. It rebuilds the execution stack from scratch in C++, introducing architectural improvements for enhanced performance. Unlike fractal scaling approaches like rollups, which rely on separate environments inheriting Ethereum security, Monad goes deep into rebuilding the execution stack.
  • Keone says that Monad implements optimistic parallel execution where multiple transactions run independently in parallel. This process generates pending results that track inputs and outputs (state changes) of each transaction. These pending results are then committed one by one in the original order of transactions, ensuring consistency with serial execution.
  • He adds that parallel execution allows squeezing more work into the same amount of time, making computation more dense. By increasing computational density, a single unit of blockchain can serve more users effectively.

Monad’s parallel execution

  • Keone says that to prepare for success, it is crucial to compare with successful Web2 applications and assess if the current infrastructure stack can handle similar demands. If the answer is no, as in Monad’s case, a new solution needs to be built to meet those requirements.
  • He adds that parallel execution and increased computational density are key goals for Monad. Early on, parallel execution was implemented to run transactions independently and generate pending results. Rescheduling conflicts are not a major concern due to the relatively small computational workload per transaction.
  • Keone says that in Monad, optimistic parallel execution runs multiple transactions independently in parallel. This generates pending results that track the inputs and outputs (state changes) of each transaction. Committing these pending results ensures consistency with serial execution.
  • He adds that rescheduling is inexpensive due to the small computational workload per typical transaction (e.g., addition or multiplication). Concerns about constant rescheduling due to interrelated transactions are mitigated by efficient handling of conflicts.
  • Keone says that computing a hash is relatively inexpensive in terms of CPU time. Reading values from state, such as going to SSD, can be expensive. Re-execution is cheaper because state dependencies are usually in memory (cache). Parallel execution allows running multiple transactions simultaneously, generating pending results, and committing them in the original order. In case of conflicts, rescheduling and re-executing the work is possible.

Benefits of parallel execution for users

  • Keone says that parallel EVM is a new layer-1 that is fully compatible with existing Ethereum smart contracts. Monad offers over 10,000 transactions per second throughput, 1-second block times, and single-slot finality. Monad provides a fast blockchain with high throughput while maintaining backward compatibility with Ethereum smart contracts and tooling like Etherscan or MetaMask.

Pipelining

  • Keone says that pipelining refers to utilizing system resources more efficiently by creating stages of work that can be completed in parallel. An example of pipelining in everyday life is doing multiple loads of laundry simultaneously instead of completing each load sequentially. In computer systems, pipelining involves creating multiple stages of work to improve efficiency. In Monad, pipelining separates consensus (nodes coming to agreement on transaction ordering) from execution (processing transactions), allowing parallel processing for improved performance.
  • He adds that in most blockchains, execution and consensus are interleaved, with consensus taking up most of the time. Ethereum, for example, allocates only a small fraction of block time (around 100 milliseconds) to execution. Monad moves execution out of the hot path of consensus into a separate pipeline, allowing nodes to agree on transaction ordering first before executing them. By separating consensus and execution, Monad improves performance by allocating more time for actual transaction processing.
  • Keone says that in Monad, nodes first come to agreement on the official ordering of transactions (consensus). Once consensus is reached, two things can happen in parallel: starting consensus on the next block and executing the agreed-upon set of transactions. Separating these processes allows for efficient pipelining and better utilization of system resources.

Selling Monad to EVM devs

  • Keone says that other developers are generally supportive and excited about Monad’s vision. The value proposition of lower fees and higher throughput is an easy sell. People are eager to onboard Monad when it goes live.
  • He adds that lower fees and higher throughput are highly desired by everyone. A commercial from the 2000s featuring complaining turtles who found their internet speed too fast is reminiscent of people wanting lower fees. Building on Monad opens up possibilities for more expansive designs without having to heavily consider gas optimization.
  • Keone says that building on Monad leads to significantly lower fees and more plentiful transactions, enhancing user experience. The improved user experience makes it easier to acquire more users. Building on Monad allows developers to focus on more expansive designs without being burdened by gas optimization concerns.

Decentralization vs high-performance

  • Keone says that the number of nodes participating in consensus, hardware requirements, and the composition of the validator set are key considerations for decentralization. Monad aims to strike a balance between professional validator companies and community-driven validators. High throughput chains need to carefully consider decentralization tradeoffs.
  • He adds that achieving performance gains through choices that centralize the network contradicts the goal of decentralization. Building a performant layer one blockchain with high decentralization requires hundreds or thousands of globally distributed nodes. Keeping nodes close together can be seen as an attempt to centralize the network.
  • Keone says that Monad’s hardware requirements are similar to Ethereum’s, with 32GB RAM being sufficient. Compared to other blockchains like Solana, which require 256GB RAM, Monad offers performance gains through algorithmic improvements rather than expensive hardware.

MonadBFT consensus

  • Keone says that MonadBFT is a high-performance consensus mechanism built from scratch in Rust. A blockchain consists of two major components: execution and consensus. MonadBFT focuses on the consensus part. Execution system was designed to enable parallel execution and a performant state database underneath it.
  • He adds that consensus ensures that globally distributed nodes agree on the official work, such as the list of transactions. MonadBFT is derived from the Hot Stuff consensus mechanism, which emphasizes linear communication. In comparison to Tendermint, another popular consensus mechanism, Monad BFT has lower communication complexity and allows for a larger network size.
  • Keone says that Hot Stuff’s linear communication protocol reduces the number of messages sent in the network. The leader for a particular slot is chosen through a leader election process and creates a block proposal. Other nodes evaluate the block proposal’s validity and send their votes back to the next leader in a fan-in approach.

Fees on Monad and $MONAD token

  • Keone says that the execution time and cost of a transaction on Monad vary depending on its complexity. A Uniswap V2 transaction with 150,000 gas costs around $5-50 on Ethereum Mainnet. On Monad, the same transaction would cost less than a cent due to higher throughput and more block space.

New possibilities to build on Monad

  • Keone says that high throughput chains like Monad open up new possibilities compared to Web2 or low throughput chains. A fully onchain limit order book is one such possibility enabled by high throughput chains. It allows for the self-custody of assets and full composability for other applications to build upon.
  • He adds that fully onchain limit order books have been limited in the EVM space due to high update costs. Market makers would need to pay significant amounts for frequent updates on existing chains. Monad’s lower costs per update (fraction of a cent) make fully onchain limit order books more feasible.

Monad community building

  • Keone says that community is a vital part of the success of projects like Monad and crypto overall. Projects like Bitcoin and Ethereum have gained massive community support due to shared ideals. Building a strong community is crucial for widespread adoption and achieving decentralization goals.
  • He adds that open-source technology ensures that the best tech will be adopted over time. Individuals who care about decentralization play a key role in using crypto services and spreading awareness among their friends. Community involvement is essential for achieving crypto’s mission.
  • Keone says that building a supportive community is crucial in the crypto industry. Over time, individuals may become more focused on technical details and lose sight of the significance of community. However, it is eventually recognized that open-source technology adoption and the difficulty of creating new communities make community-building essential.

Closing and ETA for Monad Mainnet

  • Keone says that Monad’s mainnet is expected to go live in 2024, with a public testnet launching in the second half of this year.

Check Out These Important Links

Show Information

  • Medium: Youtube
  • Show: The Edge Podcast
  • Show Title: A Beginner’s Guide to Monad’s Parallelized EVM
  • Show Date: February 14, 2023