From: thepipeline_xyz
The development of high-performance blockchain networks like Monad involves distinct stages and significant engineering efforts, often tackling problems previously deemed impossible [00:01:22].
Core Engineering Philosophy
The engineering approach emphasizes deeply understanding and optimizing for modern hardware capabilities, such as impressive SSD drives, PCI-E4, and PCI-E5 bandwidth [00:01:31]. To achieve maximum performance, it is crucial to avoid relying on general-purpose, “out-of-the-box” software like standard database or file system software, which often has algorithmic overhead and is optimized for general use cases rather than specialized high-performance needs [00:02:00].
Instead, developers must:
- Specialize and build from scratch [00:02:30].
- Build everything very low-level, avoiding reliance on the operating system to perform tasks [00:03:00].
- Understand hardware behavior, even down to how different SSD drives perform in specific ways [00:03:13].
- Optimize the EVM stack by understanding how people use the EVM, including usage patterns of ERC-20 contracts and Uniswap [00:04:27]. This involves analyzing transaction patterns, state access, scheduling, and autocorrelation (e.g., an account used recently is likely to be used again soon) [00:04:50].
- Arrange data structures to leverage analytical insights into usage patterns [00:05:41].
This process is a continuous cycle of improvement, often requiring days or weeks of engineering time to resolve specific performance issues [00:04:01].
Development Stages
Devnet (Development Network)
The Devnet is an internal deployment phase [00:06:42]. During this stage:
- The software can be “a little rough around the edges” [00:06:51].
- Manual configuration of systems is acceptable [00:06:55].
- Internal communication between development and deployment teams handles issues [00:07:38].
- For Monad, the Devnet currently achieves 10,000 transactions per second (TPS) internally with full 100% EVM compatibility, sub-cent gas fees, and hardware requirements similar to Ethereum [00:00:57].
Testnet (Test Network)
Moving from Devnet to Testnet requires significant polish and preparation [00:07:11]. The goal is to make the software a product that external users can download, run, and use independently without needing developer support [00:07:15]. This involves making the setup and usage process much smoother [00:07:33].
Throughout both stages, a significant portion of the team continuously works on performance optimization, aiming to make the network even faster [00:07:51].
Why Build a New Layer One (L1)?
Despite the prevalence of Layer 2 (L2) solutions, Monad chose to launch as a new L1 due to several factors related to scaling blockchain ecosystems and challenges in blockchain system design:
- Bandwidth Constraints: Existing L2 data availability solutions do not offer the necessary bandwidth. A new L1 like Monad can target higher connection speeds (e.g., 100 megabits) to fully utilize hardware potential [00:08:47].
- Complexity: Building a decentralized L2 involves significant complexity and trade-offs in design [00:09:29]. By building all components of Monad internally and modularly, much of this external complexity is avoided [00:09:49].
- Control and Performance: Building everything from scratch provides full control over the system, ensuring the necessary performance is achieved [00:10:40].
- User Experience (UX): An L1 can offer a simpler and better user experience by avoiding the complexities inherent in multi-chain interactions (e.g., one chain for data, another for execution) [00:10:55].
The decision to build a new L1 was driven by a competitive advantage stemming from experience in high-frequency trading (HFT), which involves building extremely low-latency systems [00:11:47]. This expertise allows for building something faster than existing L1s and L2s, many of which had made trade-offs or copied components [00:12:24].
Challenges in Development
Developing a high-performance blockchain like Monad presents significant challenges in blockchain system design:
- Database Management: A major technical challenge is handling the database. Unlike HFT systems that operate purely in memory, blockchains must store tens to hundreds of gigabytes of state data on disk [00:13:12]. This requires learning new technologies, extensive experimentation, and benchmarking existing open-source databases (like RocksDB, LevelDB, LMDB) to understand their limitations [00:14:09].
- Talent Acquisition: A non-technical challenge is finding and hiring skilled people in a competitive employment market, competing with big tech companies and the finance sector [00:14:40].
Future Outlook
The ultimate goal of these development efforts is to enable the creation of “cool apps” that leverage the network’s high performance [00:16:43]. This will provide validation for the hard work put in and foster collaboration with application teams to improve the user experience [00:16:57].
The enhanced performance unlocks new use cases and application possibilities that were previously unfeasible on lower-performing chains [00:18:26]. This not only allows for scaling blockchain ecosystems to more users (e.g., supporting 10,000 users on an order book instead of 1,000) but also increases the complexity and richness of applications per user [00:19:11]. The anticipation for these new applications marks the next exciting stage of blockchain adoption [00:19:39].