peaq
Search…
ADDBFT Consensus
More detailed information will be added shortly.
ADDBFT stands for AI-Delegated DAG Byzantine Fault Tolerance. Artificial Intelligence (AI) helps to reduce dependance on humans in the consensus process by using randomisation algorithms and optimisation methods. DAG is short for Directed Acyclic Graph, the structure by which nodes communicate, while BFT refers to Byzantine Fault Tolerance by Tendermint, of which we follow the best practices. We developed the ADDBFT consensus algorithm in-house, designed specifically for DAGchains. Here’s how ADDBFT reaches consensus:
AI Delegated The node selection process is done by AI (Artificial Intelligence). There are three types of full nodes: Validator, Generator and Storage nodes. Roles are randomly delegated each round based on node weight by using randomisation algorithms and optimisation methods.
DAG Full nodes communicate using a DAG (Directed Acyclic Graph) structure. Full nodes use messages, hashes and state flags to communicate consensus statuses and to synchronise the ledger.
Byzantine Fault Tolerance The best practices of Byzantine Fault Tolerance by Tendermint are followed in the Transaction Flow and in the Block Flow. Two thirds of chosen Validator Nodes per round must agree for consensus to be reached.
Consensus Consensus is reached without the need for mining, only rewards and fees. ADDBFT leverages Distributed Computing Power, and consensus is reached in 0.5s in private networks and ~1s in the Interoperability Platform.
Last modified 1mo ago
Copy link