DBFT and super block: security, finality and scalability
Leaderless security
DBFT is a leaderless consensus algorithm, meaning that the nodes of the network are able to reach consensus without relying on a single central authority or designated leader to make decisions. Many traditional consensus algorithms require a leader or central coordinator to propose new blocks that the other nodes validate. This leader not only introduces a single point of failure, but often a performance bottleneck as well.
With DBFT, on the other hand, there is no central authority or designated leader. Instead, all nodes in the network participate equally in the consensus process. Each node has the ability to propose new transactions, validate the proposed transactions of other nodes, and collectively agree on the state of the ledger. The algorithm aims to ensure that all nodes reach a consistent agreement on the order and validity of transactions, even in the presence of faulty or malicious (Byzantine) nodes.
DBFT can tolerate faults up to t < n/3 , where t is the fault tolerance threshold and n is the number of nodes participating in consensus on the network. That is, even if strictly less than a third of the nodes on the network are malicious (Byzantine) then the network can guarantee it will be able to reach consensus.
The advantage of leaderless consensus algorithms like DBFT is their resilience and decentralisation. They are less susceptible to single points of failure. Additionally, since every node can participate in the consensus process, the network is more scalable and maintains higher levels of security.
Furthermore, DBFT has been formally verified meaning that a rigorous mathematical proof has been performed to demonstrate the algorithm’s correctness and adherence to its intended specifications.
Instant finality
DBFT achieves quasi-instant finality. This means that transactions can be deterministically finalised in a matter of seconds, unlike the probabilistic finality of most blockchains based on Nakamoto consensus.
On Redbelly, you can have confidence that your transaction will be finalised in seconds not minutes which is crucial for real time, high value applications which require certainty, predictability, and reliability. This also means that Redbelly blockchain cannot fork, which is the main cause of double spending attacks.
Redbelly also embeds predictability in the form of its gas model which allows users to know ahead of time how much it will cost to perform any given operation on the network. A fixed price of US$0.01 for the gas required to perform a native transfer. This gives us a unit gas price of US$0.000000476190476190. For more information, see Network fees.
Scalability through collaboration
Redbelly has also implemented a super block optimisation of the EVM which takes advantage of DBFT's characteristics to optimise the way in which blocks are committed to the blockchain.
Thanks to the super block optimisation, proposals from all governor nodes can be combined and committed at the same time, increasing the amount of throughput possible significantly.
In this way, Redbelly is collaborative rather than competitive. Nodes on the network work together to solve consensus and add blocks to the blockchain rather than compete for the right to append a block, as is the case with most blockchains.
Further reading
- Crain, Gramoli, Larrea, Raynal. DBFT, Proceedings of the 17th IEEE International Symposium on Network Computing and Applications 2018.
- Crain, Natoli, Gramoli. Red Belly: A Secure, Fair and Scalable Open Blockchain. Proceedings of the 42nd IEEE Symposium on Security and Privacy S&P 2021.
- Natoli, Gramoli. The Balance Attack or Why Forkable Blockchains Are Ill-Suited for Consortium. Proceedings of the 47th IEEE/IFIP International Conference on Dependable Systems and Networks. p.579-590, 2017.