From: thepipeline_xyz
Monad, with its fully bytecode EVM compatibility and underlying architectural changes, is designed to enable new types of applications that require high performance [00:25:21]. The core of its innovation lies in its approach to parallel execution and its custom-built database, which together achieve significantly higher throughput [00:25:34].
The Problem of State Access in Blockchains
The original vision for Monad was to create a more performant version of Ethereum’s execution and consensus layers to build a superior engine for decentralized smart contracts [00:28:51]. Upon profiling existing Ethereum clients and other blockchain clients, it became evident that the single biggest bottleneck to performance was “State access” [00:29:13].
Existing Ethereum clients, such as Go Ethereum (Geth), rely on inefficient commodity key-value stores like LevelDB for storing Merkle tree data [00:29:39]. While Merkle roots are crucial for commitments to global state and composability [00:29:55], the storage method requires many lookups just to retrieve one piece of state, making it extremely inefficient [00:31:30]. This inefficiency means that even if a hundred virtual machines (VMs) were running in parallel, they would still be bottlenecked by a single “pipe” for database access, limiting the benefits of parallelization [00:32:58].
Monad’s Solution: A Custom Database and Parallel Execution
To address the state access bottleneck, the Monad team embarked on a nearly year-long journey to build a new custom database [00:32:03]. This database is designed with two key properties:
- Natively Optimized Merkle Tree Storage: It stores Merkle tree data in a custom data structure specifically optimized for that kind of data [00:32:14].
- Asynchronous Reads and Writes: The database supports asynchronous operations, which is crucial for parallel execution [00:32:29].
With this custom database, Monad enables parallel access to different regions of the Merkle tree, allowing multiple virtual machines to progress much more efficiently and achieve significantly higher performance [00:33:13]. The approach is rooted in benchmarking and identifying bottlenecks to engineer specific solutions for performance [00:33:37].
Impact and Future Outlook
Monad’s architecture, particularly its parallel EVM, is poised to deliver a throughput of 10,000 transactions per second (TPS) [00:11:27]. This high performance can enable applications that are not feasible elsewhere, such as:
- Fully on-chain order book trading [00:25:58].
- Highly interactive social applications and games [00:26:01].
- Decentralized finance (DeFi) becoming the standard for personal finance [00:11:42].
High throughput is critical for crypto to compete with traditional financial systems, especially in areas like payments and personal finance [00:11:11]. Reducing transaction costs and slippage (currently common at 1-5% in DeFi) to single-digit basis points requires a performant environment where market makers can quote tightly and compete the spread down, leading to better execution for users [00:12:06].
The goal is to shift user and developer expectations, normalizing extremely low transaction costs (e.g., $0.00001 per trade like on Solana) [00:49:29]. This change in “normalcy” will not only attract more users but also allow developers to build more robust protocols without fine-tuning for gas costs, potentially mitigating security risks [00:50:28]. This renormalization will pave the way for a new wave of applications and users to fully leverage decentralized technology [00:51:01].