|
| Consensus | Main idea | Solution effect | Existing problems |
|
| dBFT | Participants vote on nodes according to the tokens they hold and elect bookkeepers, and all bookkeepers run the BFT algorithm to reach a consensus to generate a new block | The newly generated block can be confirmed immediately, a new block is generated every 15–20 s, and throughput can reach 1000 TPS | The number of consensus nodes is limited, and the degree of centralisation is too high |
|
| Tendermint | On the basis of counting votes by a node, a weight is assigned to each vote, and the deposit is forfeited if a node behaves maliciously | It is robust against double-spending attacks and can resist up to one-third of the saboteurs in the network | There is no corresponding real-world trust model. It has high time complexity |
|
| HoneyBadger | The consensus is based on an atomic broadcast that can achieve asymptotic validity and can process tens of thousands of TPS on hundreds of nodes in the WAN | Extended by hundreds of nodes, achieving a throughput of tens of thousands of TPS | |
|
| VBFT | All legitimate voters have the right to vote and be voted, and the legitimacy of voter identity is verified by a VRF | Solves the bifurcation problem, realises the rapid consensus of the network, resists malicious attacks, and has a high degree of decentralisation and security | With the increase in consensus nodes, the scalability decreases |
|
| Algorand | VRF randomly elects block nodes and verification nodes and determines a verifier set and leader in the form of an encrypted lottery (the one with the smallest credential value is elected) | Resists malicious attacks, has low resource consumption and a high degree of decentralisation, and can ensure security and activity during synchronisation | It has higher calculation complexity and communication overhead than those of PBFT |
|
| OmniLedger | PoS selects verification nodes; VRF assigns verification nodes to the shard chain; and the shard chain reaches an agreement through BFT | Can resist Sybil attacks; the throughput increases linearly with the number of shard chains | |
|
| Dfinity | BLS generates random numbers, and a group of people generate signatures. A single person cannot prevent the issuance of the signature, and no individual can predict the result of the signature | Effectively prevents double-spending attack, selfish mining attack, long-range attack, and noninterest attack | There are adaptive attacks; it uses broadcast communication and has high time complexity |
|
