Machine Learning View on Blockchain Parameter Adjustment

Abstract

A fundamental problem in distributed computing is achieving agreement among many parties for a single data value in the presence of faulty processes–to get consensus. The consensus mechanism is an underlying part of blockchain design and commits new blocks and changes protocol itself. In addition to classic correctness requirements, blockchains need specific ones: high performance regarding transactions per second, fast transaction confirmation, etc. Blockchains control the requirements with parameters. But how to meet qualitative and optimize quantitative requirements? Typically we have the main blockchain network without access to try different parameters and the test network to do whatever we want. In the paper, we provide a machine learning view on the blockchain parameter adjustment. We list the blockchain parameters for Solana blockchain and apply feature importance to select the most significant parameters during the forthcoming optimization.