Federated BFT (fBFT)

Federated Byzantine Fault Tolerance (fBFT) is a consensus mechanism designed to enhance scalability, security, and decentralisation in blockchain, particularly in networks where a federation of trusted nodes participates in validating transactions.

A variation of the traditional Byzantine Fault Tolerance (BFT) model, fBFT uses unique characteristics where a pre-selected group of trusted entities participates in the consensus process, forming a ‘federation’, which could be composed of institutions, organisations, or even individuals. The network relies on these federated nodes to validate transactions, making it more efficient than fully decentralised models, where every node in the network may participate.

The fBFT system uses ‘quorum slices’, which are subsets of nodes that an individual node trusts to validate transactions, allowing consensus without requiring the entire network’s agreement. For the network to reach consensus, quorum slices should overlap. Each node autonomously selects its quorum slice, fostering a decentralised model where nodes operate based on their own criteria.

Typically, a BFT system can tolerate up to one-third of nodes behaving incorrectly or maliciously while still reaching consensus. Conversely, the fBFT model aims to reach consensus even when some of the participating nodes are faulty, compromised, or acting maliciously. This enhances security by assuming that not all participants can be trusted.

The system is designed to withstand the Byzantine problem, where some nodes might attempt to propagate false information or disrupt the network. Federated nodes must agree on the state of the network, and even if some of them act maliciously, the consensus can still be achieved as long as the majority behaves honestly.

Note that, while fBFT improves scalability and transaction speed, it comes at the cost of some decentralisation since the consensus process is handled by a federation of trusted nodes, creating a centralised component.