Improved Consensus Mechanisms Against Censorship Attacks

Abstract

The censorship attack is ubiquitous in prevailing blockchains, such as Bitcoin and Ethereum, yet has not been resolved well so far. According to known vulnerabilities of different consensus algorithms, censorship attacks can always achieve great benefits at low costs by various means. In this paper, we propose improvements to existing POS and POW algorithms to resolve censorship attacks in two methods. The first method improves the Tendermint consensus mechanism by introducing three new types of messages and an auxiliary role of network nodes. The method can automatically defend against censorship attacks and organize an honest chain. The second method is based on the CasperFFG consensus mechanism and can accurately identify the attackers by evaluating the suspicious scores of all validators. However, an honest validator has to manually choose the honest chain. Theoretical analyses show their effectiveness.