| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.14 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
State machine replication (SMR) is a classical technique to implement consistent and faulttolerant
replicated services. This type of system is usually built on top of consensus protocols that have high
throughput but have problems scaling to settings with a large number of participants or widearea sce narios due to the required number of messages exchanged to reach a consensus.
We propose ProBFT (Probabilistic Byzantine Fault Tolerance), a consensus protocol specifically de signed to tackle the scalability problem of BFT protocols. ProBFT is a consensus protocol with optimal
latency (three communication steps, as in PBFT) but with a reduced number of messages exchanged
in each phase (O(n
√
n) instead of PBFT’s O(n
2
)). ProBFT is a probabilistic protocol built on top of
wellknown primitives, such as probabilistic Byzantine quorums and verifiable random functions, which
provides high probabilities of safety and liveness when the overwhelming majority of replicas is correct.
We also propose a state machine replication protocol called PROBER (PRObabilistic ByzantinE
Replication) that builds on top of two consensus protocols, ProBFT and PBFT. PROBER makes use
of ProBFT to provide fast and probabilistic replies to the clients and uses PBFT to eventually determinis tically commit the history of operations guaranteeing that the system will not roll back the requests after
such commit. This periodic deterministic commit allows the clients to enjoy the low latency provided by
ProBFT while still having the guarantees provided by a deterministic protocol.
We provide a detailed description of both protocols and analyse the probabilities for safety and live ness depending on the current number of Byzantine replicas.
Descrição
Tese de mestrado, Segurança Informática, Universidade de Lisboa; Faculdade de Ciências, 2022
Palavras-chave
Consenso Tolerância a faltas bizantinas Quóruns probabilísticos Replicação máquina de estados Teses de mestrado - 2022
