| Abstract [eng] |
Decentralised smart contracts based on blockchain technology and secured by cryptographic means have been developed to address the problems of traditional - centralised - systems. As the Ethereum platform is the first and main platform for smart contracts, an analysis of smart contracts on this platform has been carried out. The analysis, including technical specifications, helped to clarify execution fees problems arising from unoptimized smart contracts code. A few high performance computing techniques, whose practical application in the context of smart contracts has not yet been extensively explored and which could help to address the optimization deficiencies in complex smart contracts, were analysed. Introduced high-performance computing techniques were discussed in the context of a sample optimisation problem - the Knapsack problem. Different Knapsack problem algorithms and their time and space complexities were compared, with further experiments being carried out with the most optimal one. Alternative solutions of cross-chain implementation using “Chainlink CCIP“ and off chain implementation using “Chainlink Functions“ were proposed. A simulated smart contract, complying with a discussed optimisation problem, was introduced for the experiment. Proposed technical solutions were implemented and deployed on the Sepolia network. Different datasets were used to compare execution fees in relation to the algorithm time complexity and the volume of data between the solutions. The results showed, that both cross-chain and off-chain solutions reduce execution fees for example smart contract. Regarding other covered aspects and limitations, integration with “Chainlink Functions“ to implement off-chain solutions was selected as a better alternative to implementing complex smart contracts solely on-chain. Using the proposed off-chain solution may not only reduce the execution fees on the Ethereum platform but may also increase its efficiency by allowing more complex smart contracts to be implemented. |
