Vulnerability of block reorganization on Ethereum 2.0

Abstract

Block reorganization (reorg) may occur when a blockchain fork is deliberately instigated and remains poorly solved on Ethereum. We utilize an agent-based modeling approach to simulate the block generation and decision-making processes of reorg attackers and ordinary validators. We analyze the impact of six validator network structures (i.e., random, regular, small-world, scale-free, hierarchical, and community networks) and four fork selection rules (i.e., LMD GHOST, HLMD GHOST, Goldfish, and RLMD GHOST) on the success rate of reorg attacks. The results show that the community network is more vulnerable to reorg attacks, while the hierarchical network is more resilient to such attacks. In addition, the Goldfish fork-choice rule can significantly reduce the success rate of reorg attacks.