The Speed Trap: Does Parallel Execution Actually Solve Ethereum’s Scaling Problem? | Markdown Memoirs

Introduction: Beyond the Sequential Bottleneck

If you’ve been following the evolution of the Ethereum Virtual Machine (EVM), you’ve likely encountered the buzz surrounding Monad and Sei. These Layer 1 blockchains are positioning themselves as the "autobahn" of the crypto world by leveraging parallel execution.

But is parallel execution inherently better? While Ethereum, the world’s second-most popular blockchain, processes transactions in a strict sequential order, many newer competitors claim that processing multiple transactions simultaneously is the only way to achieve true global scale.

However, we should be cautious of marketing hype. While the architecture is theoretically superior, many of these "fast" chains currently lack the actual traffic to prove they can handle the massive volumes they claim to support.

Defining the Mechanism: Sequential vs. Parallel

At its core, a blockchain is a distributed, immutable ledger—essentially a massive, unchangeable database.

Sequential Execution: Think of older chains like Bitcoin and Ethereum. They process data one piece at a time. It is reliable and predictable, but it creates a bottleneck during high traffic.
Parallel Execution: Used by Solana, Sei, Sui, and Monad, this handles transactions concurrently. It’s the difference between one chef cooking an entire five-course meal dish-by-dish versus five chefs working on different parts of the meal at once.

The Architecture of Parallelization

Moving from sequential to parallel isn’t just about adding "more workers"; it requires a fundamental shift in technical architecture to manage complexity.

1. Multi-Core Distribution

Blockchains utilizing parallel execution typically separate nodes (the computers maintaining the network) into groups or "instances" that process transactions across multiple CPU cores.

Sei utilizes roughly 500 workers (dynamically adjusted based on demand).
Monad optimizes the EVM specifically to handle these simultaneous threads.

2. The Conflict Problem (Double Spending)

In a parallel system, if User A has 10 tokens and tries to send 10 to User B and 5 to User C at the exact same time, a naive parallel system might accidentally allow both. To prevent this, blockchains use Optimistic Parallel Execution.

Assumed Validity: The system assumes transactions don't conflict and processes them in a temporary state.
Conflict Review: The system scans for overlapping data or "double spends."
Resubmission: If a conflict is found, the specific transaction is aborted and resubmitted in sequential order.

3. Developer Access Lists

Modern EVM upgrades now support access lists. This allows developers to specify exactly which smart contract storage areas a transaction will touch. By knowing the "path" of a transaction ahead of time, nodes can sort and execute them in parallel much more efficiently.

"accessList": [
     {
        "address": "0x4869a4c7657cef5e5496c9ce56dde4cd593e4923",
        "storageKeys": [
           "0x360894a13ba1a3210667c828492db98dca3e2076cc3735a920a3ca505d382bbc",
           "0x0000000000000000000000000000000000000000000000000000000000000002"
         ]
       }
]

Execution Data: Reality vs. Marketing

While the theoretical throughput (TPS) of parallel chains is impressive, the "real-world" data tells a more nuanced story.

Metric	Ethereum (Sequential)	Sei / Monad (Parallel Claims)
Claimed TPS	15 – 30	10,000 – 15,000
Finality Time	~15 Minutes	400ms – 800ms
Cost (ERC-20)	~$0.03	~$0.000098

The "Ghost Chain" Challenge

The main issue with TPS claims is adoption. While Monad claims 10,000 TPS, its mainnet launch saw peaks closer to 100 TPS. Sei, despite reaching 4 million daily transactions recently, still averages well below its theoretical limit. In contrast, Ethereum L2s like Base have already hit averages of 150 TPS with 20 million daily transactions.

Parallel execution improves speed and costs significantly, but we have yet to see if these architectures can maintain stability when they reach the same density of users as the Ethereum mainnet.

Conclusion: More Than "Barely" Better?

Parallel execution definitely enhances the EVM ecosystem. While sequential execution remains the gold standard for predictability and "battle-tested" security, it is simply too slow and costly for mainstream global adoption.

Sei and Monad are leading the charge in proving that the EVM can be modernized. Even if their full TPS potential remains untapped, the drastic reduction in finality time and gas fees proves that parallel execution is the necessary path forward for the next generation of decentralized applications.

Sources

Liu et al. (2020). Parallel and Asynchronous Smart Contract Execution. https://arxiv.org/pdf/2306.05007
Marsh et al. (2025). Sei Giga v0.2. Sei Labs. https://arxiv.org/pdf/2505.14914
Das et al. (2025). Accelerating Blockchain Scalability: New Models for Parallel Transaction Execution in the EVM. https://arxiv.org/html/2504.01370v1