Exploração de algoritmos de consenso no Quorum

• Autor(a) principal: Lopes, João Pedro Ascensão
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/23439
• topic: Blockchain; Quorum; Consensus Algorithms; IBFT; QBFT; Clique; Raft

As blockchain technology matures, more industries are becoming interested in evaluating if the technology can answer their needs for decentralized systems that guarantee data immutability and traceability. Quorum is a blockchain platform that accommodates enterprise use-cases by extending Ethereum to support private transactions and a higher transaction throughput. To achieve this, Quorum replaced Ethereum’s proof-of-stake consensus mechanism with proof-of-authority ones, supporting four different algorithms: Raft, Clique, IBFT 1.0, and QBFT. This work explores Quorum’s consensus algorithms and how they affect performance and fault-tolerance, in order to assess the best use cases for each and what should drive their choice. A GoQuorum network was set up, and benchmarks were run against this system under different scenarios while only changing the consensus algorithm for each scenario. Results showed that Raft is the most performant consensus algorithm in Quorum in both private and public transactions. Additionally, QBFT achieved the same performance as IBFT, and Clique was the worst performer across the board, particularly due to having high resource-usage. Regarding fault-tolerance, it was found that bringing validator nodes down at random, when the network has high-availability, had no impact on networks under any of the consensus algorithms.