Design Solutions and Their Impact on the Execution Costs of Smart Contracts

Abstract

This thesis investigates the impact of design solutions on the execution costs of Ethereum smart contracts, focusing on gas consumption at function level. The work analyzes real-world smart contracts and combines transaction data from Etherscan, function similarity analysis through SmartEmbed and ANTLR, and statistical testing to identify design patterns associated with higher execution costs. Starting from a dataset of Solidity smart contract functions, the study identifies highly similar function pairs with divergent gas costs and manually examines the surrounding contract context to detect recurring design differences. These patterns are then grouped into broader categories and evaluated through statistical analysis to assess their relationship with gas consumption. The results provide practical insights for developers and researchers interested in designing more efficient, sustainable, and cost-aware smart contracts on Ethereum.