Benchmark TPC-C aplicado em replicação máquina de estados

• Ano de defesa: 2023
• Autor(a) principal: Serafim, Kayel Ludtke
• Orientador(a): Dotti, Fernando Luís
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação
• Departamento: Escola Politécnica
• País: Brasil
• Palavras-chave em Português: Replicação Máquina de Estado; Tolerância a Falhas; Benchmark
• Palavras-chave em Inglês: State Machine Replication; Fault Tolerance; Benchmark
• Área do conhecimento CNPq: CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO
• Link de acesso: https://tede2.pucrs.br/tede2/handle/tede/10966

The availability of a system can be impacted by failures or attacks that exploit its vulnerabilities. Increasingly, more people rely on available online systems on the Internet. To minimize downtime risks, State Machine Replication (SMR) is a common approach. SMR is an important strategy for providing highly available services. However, increasing capacity in SMR systems is challenging due to its deterministic model, which requires research efforts to improve its parallelism. Nevertheless, there is a lack of workloads to evaluate different SMR mechanisms according to common and representative criteria for classes of applications of interest. Based on this, important common aspects were identified in the context of online transactions, and the use of TPC-C, the Transaction Processing Performance Council’s benchmark C, this work proposes to evaluate SMRs. Its architecture for the SMR context is discussed and an implementation on a replication platform is provided. TPC-C performance results are presented for the classic SMR model. Additionally, a parallel SMR approach is discussed, and implemented with this workload, and the results obtained reported.