# On transaction parallelizability in Ethereum

## TL;DR

This paper investigates the potential for parallel execution of Ethereum transactions, showing that significant performance gains are possible, especially at higher transaction rates, by relaxing sequential constraints when safe.

## Contribution

It provides an empirical analysis of transaction parallelizability in Ethereum, demonstrating potential performance improvements through parallel execution strategies.

## Key findings

- Parallelization can significantly improve Ethereum transaction throughput.
- Potential for parallelizability increases with higher transaction rates.
- Simulations using real data support the feasibility of safe parallel execution.

## Abstract

Ethereum clients execute transactions in a sequential order prescribed by the consensus protocol. This is a safe and conservative approach to blockchain transaction processing which forgoes running transactions in parallel even when doing so would be beneficial and safe, e.g., when there is no intersection in the sets of accounts that the transactions read or modify. In this work we study the degree of transaction parallelizability and present results from three different simulations using real Ethereum transaction data. Our simulations demonstrate that notable gains are achievable with parallelization, and suggest that the potential for parallelizability improves as transaction rates increase.