In this paper we extend the Multidimensional Byzantine Agreement (MBA) Protocol, a leaderless Byzantine agreement for lists of arbitrary values, into a protocol suitable for wide gossiping networks: Cob. This generalization allows the consensus process to be run by an incomplete network of nodes provided with (non-synchronized) same-speed clocks. Not all nodes are active in every step, so the network size does not hamper the efficiency, as long as the gossiping broadcast delivers the messages to every node in reasonable time. These network assumptions model more closely real-life communication channels, so Cob may be applicable to a variety of practical problems, such as blockchain platforms implementing sharding. Cob has the same Bernoulli-like distribution that upper-bounds the number of steps as the MBA protocol. We prove its correctness and security assuming a supermajority of honest nodes in the network, and compare its performance with Algorand.

Cob: a leaderless protocol for parallel Byzantine agreement in incomplete networks

Andrea Flamini;Riccardo Longo;
2024-01-01

Abstract

In this paper we extend the Multidimensional Byzantine Agreement (MBA) Protocol, a leaderless Byzantine agreement for lists of arbitrary values, into a protocol suitable for wide gossiping networks: Cob. This generalization allows the consensus process to be run by an incomplete network of nodes provided with (non-synchronized) same-speed clocks. Not all nodes are active in every step, so the network size does not hamper the efficiency, as long as the gossiping broadcast delivers the messages to every node in reasonable time. These network assumptions model more closely real-life communication channels, so Cob may be applicable to a variety of practical problems, such as blockchain platforms implementing sharding. Cob has the same Bernoulli-like distribution that upper-bounds the number of steps as the MBA protocol. We prove its correctness and security assuming a supermajority of honest nodes in the network, and compare its performance with Algorand.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11582/355308
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
social impact