Compiled & edited by TechFlow
Guests: Keone Hon, Co-founder & CEO of Monad Labs; Jay Jog, Co-founder of Sei Labs; Vijay Chetty, CEO of Eclipse Labs
Moderator: Laura Shin, Writer and Host of Unchained
Podcast source: Unchained
原标题：How to Build the Fastest Onchain Experience With Monad, Sei, and Eclipse
Air Date: August 28, 2024
Summary of key points
In this episode, Keone Hon of Monad Labs, Jay Jog of Sei Labs, and Vijay Chetty of Eclipse Labs share their different approaches to scalability and performance on high-throughput blockchains. They discuss the technical advantages of parallelizing the Ethereum Virtual Machine (EVM), the strategic decisions behind blockchain architecture, and the innovations driving the next generation of high-speed chains.
Monad
How Monad got started and its original mission
Keone introduced the origin and mission of Monad. Monad is an Ethereum Virtual Machine-compatible blockchain that uses a design called pipeline architecture, which can achieve a high throughput of 10,000 transactions per second and a block generation time of one second. Monad was founded in early 2022 by three co-founders, Eunice Girda, James Hon Saker, and Keone. Keone and James worked at Jump Trading for about eight years, focusing on building high-performance trading systems, and participated in Solana DeFi related work in Jump's crypto team.
They realized there was a pressing need in the market for higher performance EVM execution, so they decided to leave Jump and start Monad with Eunice Girda.
Keone pointed out that the core goal of Monad is to get the maximum performance from minimal hardware, which requires building new software from scratch and introducing new architectural improvements. The design concept of Monad is to stack four major improvements on top of each other, each of which has a multiplier effect on the overall efficiency of the system, thereby achieving excellent performance.
Features that make Monad a high-throughput blockchain
Keone details several key features of Monad that enable high throughput.
First, Monad uses a customized state database that can natively store Ethereum Merkle tree data on SSD, allowing for more efficient state access.
The second improvement is optimistic parallel execution, which allows multiple transactions to run in parallel while ensuring that the end result is consistent with that of serial execution.
The third item is asynchronous execution, which creates independent “swim lanes” between consensus and execution, allowing execution to take full advantage of the entire block time, rather than just a fraction of it.
Finally, Monad BFT is a performant consensus mechanism capable of maintaining synchronization among hundreds of globally distributed nodes.
Regarding the composability of parallel processing, Keone explained that Monad's blocks are still linear, and the transactions within the blocks are also defined in a linear manner, so the total order between transactions is maintained. Parallel processing is performed in the background and does not affect the final result of transaction execution. The guarantee of this optimistic parallel execution is that although some processing is performed in parallel, the final submission is still carried out in the original transaction order, ensuring that each submission is correct and can be re-executed if any input changes.
In addition, Keone also mentioned that the customized state database is designed to solve the problem of blockchain expansion. As the state continues to grow, the cost and latency of accessing the state will also increase. The database structure used by existing Ethereum clients is complex, resulting in a large number of interactions when accessing a state in the Merkle tree. Monad's customized database (called Mona DB) is specifically optimized for storing Ethereum Merkle tree data. Although building a database from scratch is a complex process, it is an important step to improve execution efficiency.
Why Monad chose to create a new blockchain instead of L2
Keone explained why Monad chose to be a standalone blockchain rather than L2.
First of all, decentralization is their primary consideration. They believe that decentralized block production and nodes participating in consensus are very important, which is essential for censorship resistance and decentralization of overall network control. Most existing L2 solutions still rely on centralized sorters, which is contrary to their decentralized philosophy.
Secondly, performance and cost are also key factors. In order to build a truly high-performance system, every layer of the technology stack must be optimized, including execution components, consensus components (keeping nodes synchronized), and other considerations such as data availability. Keone emphasized that Monad's goal is to comprehensively address these issues and build an efficient single system to achieve maximum performance.
Why Keone thinks Monad provides the best developer experience and why he doesn’t like the title of “Ethereum Killer”
Keone mentioned that the reason why Monad provides the best experience for developers is mainly reflected in two aspects: performance and compatibility. First, Monad is fully compatible with EVM (Ethereum Virtual Machine), which means that developers can continue to use the tools and libraries they are familiar with, such as the Solidity programming language. This compatibility enables developers to reuse existing components or ensure backward compatibility with Ethereum standards when building new applications.
Second, Monads provide significant performance improvements. Through asynchronous execution and an efficient consensus mechanism, developers can deploy applications in a Monad environment to enjoy higher transaction processing capabilities. This is especially important for applications looking to achieve mainstream adoption, as they need to process large volumes of transactions.
What do you think of the title of “Ethereum Killer”?
Keone expressed a different view. He believes that this description is not accurate because the Monad project focuses on some under-researched directions in the field of Ethereum expansion. Keone emphasized that the team hopes to enhance Ethereum's capabilities by exploring different technical pillars, rather than replacing it. Their goal is to contribute to Ethereum's expansion roadmap by proposing improvement proposals (EIPs) and collaborating with other researchers.
Keone further pointed out that Monad can be seen as a pioneer environment for certain Ethereum improvements, which may drive some improvements into Ethereum or other environments over time. Therefore, Monad's goal is to enhance Ethereum's capabilities rather than destroy it.
Monad's Large Venture Capital Financing and How It Uses the Funds
Monad recently completed the largest financing event in 2024, raising a total of $225 million. Keone said that the funds will be used in many aspects, with the main goal of expanding the team and recruiting more talents who can drive the Monad project and the entire industry forward. They not only hope to recruit professionals in the encryption field, but also plan to attract some junior engineers with rich experience in other fields and introduce them to the encryption industry.
Keone stressed that the team is grateful for the support of investors and believes that the funding provides them with ample resources to realize their vision.
Keone also mentioned that although Monad has not yet been officially launched, it has attracted the attention of many venture investors and applications. For example, Monad applications such as Kintsu and Kuru have recently received financing. This attention reflects the capabilities of the teams in the Monad ecosystem and investors' expectations for features such as high throughput, low fees, and full EVM compatibility. In addition, investors are also confident in other improvements that the Monad team continues to promote, which will help expand and enhance the capabilities of EVM in the long term.
Monad's strong community
Keone pointed out that Monad has a strong community, mainly due to the active contributions of community members and the efforts within the team. He believes that the growth of the community is ultimately driven by members who feel they can contribute. The role of the Monad team is to create a friendly environment where people feel welcome, while managing spam and creating a good atmosphere.
Keone expressed gratitude to the engaged individuals in the community who set the bar high, create great memes and artwork, and organize events. For example, Monad has also launched the Monad Running Club, a Strava group where members log their runs with the goal of improving fitness, and the Mon Lingo program, which is aimed at learning a new language together.
These community events showcase the fun of the Monad community, emphasizing friendship and relationship building rather than just focusing on financial gain or “farming” behavior. This kind of social network building is critical to a successful blockchain community.
What’s next for Monad and whether it will launch a token soon
Keone said that Monad's testnet and mainnet are about to be launched, and the team is working hard. Although there is no exact release date, the progress is very smooth.
Regarding the Monad token, Keone was unable to provide specific information at this time, despite the leak of the token name and subsequent withdrawal. He said that he could not comment on it for the time being.
Be
What is Sei, and the role of the GameStop incident in its creation
Sei is a fully functional, first-level blockchain that was initially inspired by the GameStop incident that Jay Jog and his co-founder Jeff Feng experienced in 2021. Jay was the head of engineering at Robinhood at the time, and he experienced the internal chaos of the incident firsthand. When buying of GameStop and other stocks was suspended, Robinhood faced tremendous public pressure, which made him very uncomfortable because as an insider, they did not have full context and had to bear reputational risks.
After experiencing the GameStop incident, Jay and Jeff decided to create a decentralized trading platform aimed at avoiding similar situations. At first, they wanted to build an on-chain decentralized exchange (DEX), but found that there was no suitable ecosystem to support an on-chain order book-based exchange at the time. Therefore, they decided to build it as an independent chain, which eventually evolved into Sei, a blockchain platform focused on efficient trading and decentralized finance.
Why Jay thinks the EVM developer ecosystem is so strong
Jay believes that the EVM (Ethereum Virtual Machine) developer ecosystem is strong for the following reasons:
High developer ratio: Currently, about 80% to 90% of developers in the crypto space are EVM developers. This means that the EVM ecosystem has a large developer base.
Technical and ideological stickiness: EVM is not just a technology stack, but an ecosystem. A rich tooling and developer culture has been formed around EVM, giving developers a stronger sense of belonging and stickiness in this environment. Many developers are opposed to switching to other execution environments, partly because of technical complexity, and moving to a new environment may lead to errors in the code, thus affecting the security of the entire project.
Performance limitations: Jay pointed out that the biggest limitation of EVM is insufficient throughput, which leads to poor user experience. For example, Ethereum's transaction processing capacity is low, and users need to pay high transaction fees during peak hours, which makes it difficult for ordinary users to operate on the chain.
Potential for parallel processing: Jay mentioned that the EVM currently processes transactions in a single thread, while modern computers are multi-core and can handle multiple workflows simultaneously. Their core insight is that by parallelizing the EVM, modern hardware can be more efficiently utilized, thereby increasing transaction processing capabilities. This is also the reason why they chose to build Sei, making it the first parallelized EVM.
Combining the strengths of EVM: By combining the strengths of EVM with the features of high-performance blockchains such as Solana, Sei is able to provide higher performance and user experience while maintaining the EVM developer ecosystem.
Why Sei switched from Cosmos to EVM, leading to the launch of its v2
Jay explained Sei’s transition from the original Cosmos to EVM and the reasons for launching V2, mainly including the following points:
Initial choice: In the initial stage of Sei, the team chose to use Cosmos SDK to build application-specific chains because Cosmos SDK is a well-tested framework that makes it easier to start projects. This enabled them to quickly build infrastructure.
Transition to a general-purpose chain: As the project developed, the Sei team realized that they needed a more general solution, and the Cosmos SDK was still a good starting point. However, they found some inefficiencies in their use, so they made several optimizations.
Optimization of consensus mechanism: Initially, Sei used the more basic Tendermint consensus mechanism. In order to improve performance, they optimized the block propagation and processing methods, and finally developed the "twin turbo consensus", making Sei one of the fastest blockchains currently.
Performance advantage: With these improvements, Sei achieves a final confirmation time of 400 milliseconds, which is faster than the confirmation time of other chains such as Solana. This high performance makes Sei stand out from the competition.
Decision to switch to EVM: With a deeper understanding of market demand and the developer ecosystem, Sei decided to launch V2 and switch to the EVM environment. This shift will not only attract a large number of EVM developers, but also take advantage of the rich tools and resources in the EVM ecosystem to improve user experience and development efficiency.
What makes Sei “the fastest chain, even faster than Solana”
Jay explained how Sei became the “fastest chain” and detailed the characteristics of twin turbo consensus, mainly including the following points:
The mechanism of dual-turbo consensus:
In the traditional Tendermint consensus mechanism, processing a block requires two rounds of voting: pre-vote and pre-commit, and then the block can be processed. This process is time-consuming.
Sei’s innovation is that while these two rounds of voting are taking place, blocks are processed in parallel. This means that during voting, blocks can be processed simultaneously, saving time. For example, if the voting process takes 300 milliseconds, Sei can use this time to process blocks, improving overall performance.
Block propagation optimization:
In traditional Tendermint, when a block is submitted, the entire block must be sent over the network, even if every validator already has most of the transactions in their memory pool.
Sei's solution is to send only the transaction hash (about 32 bytes per hash) instead of the entire transaction. This greatly reduces the amount of data transmitted over the network, allowing validators to reconstruct blocks locally, thereby improving performance.
Introduction of parallelization:
In addition to Dual Turbo Consensus, Sei also introduces parallelization technology to further improve the performance of the execution layer. This parallel processing capability allows Sei to utilize computing resources more efficiently.
Market feedback and transformation:
When Sei launched V1, it only supported Cosmos and Rust smart contracts, which led to developers being reluctant to learn new programming languages and a relatively small number of people using it.
In communication with developers, Sei received consistent feedback that they wanted to support EVM. Therefore, the team decided to turn to EVM and make further optimizations and improvements on this basis.
Successful launch of V2:
Sei announced V2 in November last year and officially launched it in May this year. V2 combines EVM and Cosmos smart contracts, allowing developers to use different smart contract languages on the same chain.
With the launch of V2, Sei's ecosystem has developed rapidly, attracting the attention of a large number of new projects and investors, and TVL (total value locked) has grown significantly.
How Sei DB works and why Jay believes that monolithic architecture has many advantages over modular architecture
How SeiDB works
State storage and data writing:
With transaction parallelization and higher throughput, the amount of data written on the Sei blockchain has increased significantly, which has led to state load issues. State load is one of the main causes of blockchain performance degradation.
To solve this problem, Sei introduced CDB, a new database solution with two main ideas at its core:
Memory-mapped IVL tree: In V1, the entire tree structure was stored on disk, but in V2, Sei split it into multiple files and removed a lot of metadata, resulting in a reduction of about 60% of stored data. This makes it easier to run a full node and can scale more quickly.
Asynchronous writes to disk: State is created in memory rather than immediately written to disk, resulting in faster writes, which Jay mentioned were 287x faster when committing blocks. This improvement significantly improves performance, reduces latency and increases throughput.
Advantages of Monolithic Architecture
Complexity Management:
Jay believes that monolithic architecture has obvious advantages in managing complexity compared to modular architecture. In modular architecture, the dependencies between multiple microservices will increase the complexity of the system. For example, the separation of the execution layer, settlement layer, and data availability layer may lead to communication problems between them, increasing the fragility of the system.
Performance optimization:
In a monolithic architecture, all components operate within the same system, eliminating communication delays between different modules, thereby improving overall performance. Jay emphasized that top performance is always better in a monolithic chain than in a modular system because modular architectures need to handle additional communication overhead.
Fault Tolerance:
Failures in individual parts of a modular architecture can affect the finality and availability of the entire system, whereas in a monolithic architecture, all functions are centralized in one system, reducing the scope of impact when a failure occurs.
Jay emphasized that although modular architecture may be more appropriate in some cases (such as requiring dedicated block space), for applications pursuing top performance, monolithic architecture provides a better solution. Sei DB is designed to support high performance requirements by optimizing state storage and increasing write speed while reducing system complexity.
Eclipse
How Eclipse works by combining Ethereum, Solana, and Cosmos
Eclipse's Vision and Goals
High-throughput L2 solution: Eclipse aims to build a high-throughput Ethereum Layer 2 (L2) solution that aims to improve efficiency by optimizing the execution environment. Vijay emphasized that Eclipse chose to use the Solana Virtual Machine (SVM) because it provides higher parallel processing capabilities and throughput than the Ethereum Virtual Machine (EVM).
Modular architecture
Modular design: Eclipse adopts a modular design concept, using different blockchain technologies to meet different needs:
Execution layer: Eclipse uses the Solana Virtual Machine (SVM) because it provides higher throughput and has a large Rust developer community. This allows Web 2 developers to easily transition to Web 3 and build applications using SVM's high performance.
Settlement and Consensus Layer: Eclipse chose Ethereum as the settlement and consensus layer to take advantage of Ethereum's large user base and diverse assets. This combination allows users to enjoy the superior user experience of Solana applications while still being able to access Ethereum's assets and ecosystem.
Data Availability Layer: Eclipse selected Celestia as the data availability solution because it is a leader in modular data availability and provides the best block data publishing cost.
ZK Fraud Proofs
Fraud Proof Layer: Eclipse plans to use zero-knowledge (ZK) fraud proofs to create a trust-minimized L2 solution. Vijay believes that all optimistic approaches will eventually move to ZK fraud proofs, so Eclipse hopes to be at the forefront in this area.
Solving industry problems
Solving the Fragmentation Problem: Vijay mentioned that there is a fragmentation problem in the current Ethereum ecosystem, which makes coordination between different RFQs and liquidity providers complicated. Eclipse aims to provide a general L2 solution that can support the needs of most applications, thereby reducing the coordination problems caused by this fragmentation.
How Eclipse handles the complexity of its modular architecture
In-house development and testing
Team effort: The Eclipse team does a lot of internal development and testing to ensure efficient communication between the various layers. The close cooperation with Celestia has significantly improved the efficiency of data publishing.
Optimization of data publishing
Data volume and metadata: Vijay mentioned that Eclipse recently set a record for the largest amount of data released in a single day, and the reason behind this achievement is the optimization of data structure. He pointed out that in addition to the transaction list, the compressed Solana block also contains a lot of other metadata. Therefore, reducing the amount of metadata and improving the efficiency of data release have become important optimization directions.
Continuous learning and optimization
Dynamic Adjustment: The Eclipse team is constantly learning and adjusting to cope with the growing volume of activity. The goal is to reduce the ratio of published data to activity as activity increases, thereby improving overall efficiency.
Ecosystem collaboration
Collective testing and collaboration: Vijay stressed that “collective battle testing” between Eclipse and Celestia and other ecosystem partners is an important part of the optimization process. This collaboration not only helps improve their own solutions, but also promotes the progress of the entire ecosystem.
What are the ways to trade in SOL on Eclipse?
Bridging mechanism
Bridge required: SOL transactions on Eclipse require the use of a bridge mechanism. Eclipse will operate a dedicated bridge for transferring Ether (ETH) to Eclipse.
Asset Bridging Cooperation
Partners: Eclipse is also working with other bridge service providers, such as Hyperlane, to facilitate bridging from other L1 and L2 assets. This means that users can more easily move a variety of assets to the Eclipse platform.
Wrapped SOL Tokens
Wrapped SOL: On Eclipse, SOL will exist as a wrapped token. This means that during the bridge process, users will convert the original SOL into a wrapped version in order to trade on the Eclipse network.
Vijay's reaction to how Eclipse Labs responded to the allegations made by its founder and former CEO Neel Somani
Focus on product and team: Vijay emphasized that his main focus is the future of the company, the product and its development roadmap. He is committed to ensuring that the team is working in a unified direction, and 99% of his energy is focused on building the product and team.
No comment on personal actions: Vijay said he will not comment on Neel Somani’s personal actions or allegations as he believes the veracity of the allegations is not for him to judge. However, he supports those who speak out and their right to express themselves.
Company and product priority: Vijay stressed that for him, the most important thing is to focus on the company and the product that is being built. He believes that the controversy surrounding Neel is mainly a personal matter, Neel has left the company, and the team will continue to focus on their work.
How Eclipse attracts developers
Attracting Solana Developers
Easily port applications: Eclipse provides an easy way for existing Solana developers to migrate their applications to Ethereum to reach Ethereum users and assets. This is a growth opportunity for protocols like Mango, Soul Land, etc. to attract more users and assets while bringing the Solana application experience to their users.
Attracting Ethereum developers
Opportunities for cross-chain development: Vijay noted that there has been little communication between Ethereum and Solana developers, mainly due to the differences between the two languages (such as Solidity and Rust). However, this situation has improved recently, and developers have begun to explore how to use Rust on Ethereum, and tools such as Rep are also supporting this cross-chain development.
Solving Ethereum’s scalability issues
Decentralized trading experience: Many Ethereum developers move their DeFi trading processes off-chain or to specific application chains due to scalability limitations. Vijay believes that this leads to a fragmented experience and makes it difficult to achieve liquidity unification and transparency. Eclipse provides a solution that enables developers to deploy institutional-grade central limit order books (CLOBs) and seamless request for quotes (RFQ) systems while keeping these operations completely on-chain.
Providing high-quality DeFi application experience
Institutional DeFi Applications: Eclipse aims to provide users with a high-quality DeFi application experience while maintaining full on-chain operations, which is contrary to the current trend of many developers moving intent and RFQ systems off-chain. In this way, Eclipse can help developers achieve better user experience and higher transparency.
What areas Vijay expects to flourish at Eclipse
DeFi
Institutional-level DeFi experience: Vijay is very much looking forward to seeing some institutional-level DeFi experiences during the hackathon, especially the central limit order book (CLOB) and request for quote (RFQ) systems. He emphasized that these systems need to have a good user experience so that ordinary users can easily access them.
Meme Coin Field
Fair Launch and Community Ownership: Vijay noted that there has been a lot of activity on Solana around Meme Coin, and he looks forward to seeing some innovation related to fair launch and community ownership. These concepts can be applied to venture-backed crypto companies to promote community participation and development.
Consumer Experience
High-throughput consumer applications: Vijay emphasized that the high throughput provided by Eclipse provides a huge opportunity to build first-class consumer experiences. He mentioned that mainstream Web 2 social applications have a large number of transactions and interactions per second, so being able to use Eclipse's capabilities to support these applications is a very interesting design space. He mentioned some successful examples such as Friend Tech, Forecaster, and Lens, which have demonstrated potential in this area.
Infrastructure
Decentralized physical infrastructure: In terms of infrastructure, Vijay mentioned the potential of decentralized sensor networks and decentralized artificial intelligence. He believes that although there is some hype in this area, there are still many interesting possibilities in the design space of large-scale sensors and deep networks.
What’s next for Eclipse and when the mainnet might launch?
Vijay mentioned that Eclipse has launched the mainnet for developers, but has not yet actively promoted the use of front-end tools for ordinary users. Their focus is to help developers and infrastructure partners deploy applications to the mainnet environment.
The public mainnet is expected to be launched around September, when a variety of interfaces and gamification experiences will be available to ordinary users.
Post-launch development plan
Withdrawal function: Before the public mainnet is launched, users will be allowed to withdraw, which is a basic requirement.
Phased development: After the mainnet launch, Eclipse will transition to the first phase of Rollup in the next few months, which will include the launch of features such as permissionless fraud proofs, trust-minimized bridges, and forced inclusion.
Gradually open source: Eclipse plans to gradually open source parts of its technology stack in the future. The currently released parts adopt the Apache 2.0 license, which allows academic use and reuse.
Decentralization and Governance
Decentralized future: Vijay emphasized that decentralization and governance are part of the Eclipse roadmap and the Eclipse Foundation is working on a timeline.
Competitiveness and Innovation
High-throughput opportunity: Eclipse aims to provide significantly higher throughput and plans to be the first to implement a standalone validator client developed by the Jump team for Solana.
Hardware acceleration: Vijay said that Eclipse will also conduct research on hardware acceleration, with the goal of achieving a throughput of hundreds of thousands of transactions per second to support the scale of Web 2 and Web 3 applications.
Suitable application types
Key Application Areas: Vijay specifically mentioned that the types of applications that Eclipse is best suited for include on-chain central limit order books and RFQ systems, mainstream consumer applications that can reach millions of users, and deep applications such as large sensor networks.
Summarize
In summary:
Monad focuses on optimizing EVM performance to provide an efficient Ethereum experience.
Sei improves DeFi transaction efficiency through a decentralized trading platform combined with parallelized EVM.
Eclipse combines multiple blockchain technologies to provide a high-throughput Ethereum L2 solution through a modular architecture and ZK fraud proofs.
These three projects have each proposed innovative solutions to different challenges of blockchain performance and decentralized applications, promoting the diversified development of blockchain technology.
