AleoBFT: A new hybrid architecture for consensus.
Today I am writing again about the Aleo project. Just yesterday, the first phase of the third test network of this project began. The Aleo team decided to change the type of consensus in the network. Let's look into this and what were the reasons for this.
So, Aleo is the world's first full-featured platform for developing private applications (with zero knowledge), which is characterized by speed, integrity and scalability. Aleo achieves this by using decentralized systems and zero-knowledge cryptography to secure user data online.
Aleo plans to switch to AleoBFT's hybrid consensus architecture during Testnet 3. Previously, Aleo used the concept of Proof-of-Succinct Work as a mechanism to incentivize the market to lower the cost and increase the efficiency of creating zero-knowledge proofs. But as practice has shown, block production has become very centralized, as a small number of network participants began to use specialized equipment for zero-knowledge cryptography. This exposed the network to the risk of centralization and censorship for users.
Before looking at how AleoBFT works, let's remember all the participants in the network.
Validators produce blocks and are rewarded for doing so.
Provers solve Coinbase puzzles (PoSW) in order to get rewarded for doing so.
AleoBFT combines the instant finality of new blocks from validators with the processing power of provers to inject a supply of monetary base into the network. The design of AleoBFT incentivizes validators to maintain viability by creating blocks, and also incentivizes provers to scale proving opportunities for the Aleo ecosystem.
What are the benefits of AleoBFT?
First, AleoBFT guarantees instant finality for each block.
Transactions are confirmed once validators produce and achieve consensus for each block. This approach resolves issues of shallow forking for validators and improves node stability. In addition, this creates a smooth application experience for developers on Aleo. And for users, this improves UI/UX by enabling fast confirmation in wallets, browsers, explorers and applications. Lastly, this guarantee makes interoperability with other ecosystems a much simpler proposition.
Second, AleoBFT ensures the network remains sufficiently decentralized.
By using AleoBFT, we are able to decouple the role of block production from coinbase generation. Validators are now in charge of producing blocks, and provers are now able to compute proofs at their desired scale. AleoBFT is designed to incentivize provers to become validators, by staking at least 1 million Aleo credits on the network. This guarantees provers can preserve their censorship-resistance by being a representative for all provers in consensus.
Third, AleoBFT incentives provers to scale proving capacity for the Aleo ecosystem over time.
In AleoBFT, provers solve and produce coinbase proofs (PoSW) on the network in exchange for a share of the coinbase reward in each block. With this new model, provers are now computing on the core subcomponents of zero-knowledge proofs – namely multi-scalar multiplications (MSMs) and fast-Fourier transforms (FFTs) – while earning their percentage share in proofs from each block reward.
For provers, AleoBFT allows operating any hardware at any scale to produce zero-knowledge proofs, without competing for a “winner-take-all” effect in each block. Thus, unlike in Testnet 2, this process means that even provers running on commodity hardware will be able to participate and earn coinbase rewards. These coinbase rewards are shared between provers and validators to incentivize provers to become validators as they accumulate tokens over time. By incentivizing a broader reward distribution for participants, we believe these new mechanics will scale the Aleo network to achieve far greater proving capacity for Aleo developers and users.
As we can see, this is quite an interesting consensus that should greatly improve the Aleo network. Therefore, we are closely following Testnet-3 Aleo, based on the results of which it will be possible to draw conclusions, backed up by practice.