Aphelion — Artificial Intelligence Governed Blockchain Consensus
Aphelion is AI-governed blockchain consensus algorithm which, the same as any other consensus algorithms, is used to consistently reach an agreement about the state of the blockchain between different participants on a blockchain network. However, Aphelion, opposite to other consensus algorithms, is not a trade-off system — it represents a solution for issues with speed, cost, security, decentralization, efficiency, scalability and the lack of interoperability without introducing any drawbacks.
Aphelion is a consensus engine used in Libonomy blockchain, but it presents characteristics to be used in developing all blockchains in future.
• Autonomous — Artificial intelligence engine makes decisions to improve the blockchain network;
• Interoperable — Ability to communicate with former blockchains (like Bitcoin and Ethereum) in a decentralized manner;
• Dynamic — Updates determined by AI engine are executed seamlessly, without interruption on the network and without forking issues;
• Hybrid — Can be utilized both in private and public solutions;
• Scalable — Sophisticated in distributing resources in the network; can sustain a fast network with a wide variety of nodes;
• Honest — Better at distributing rewards to a wide variety of network participants than other consensus algorithms;
• Efficient — Fast network (6000+ TPS) and that is also energy efficient;
• Secure — Combination of AI engine and a multi-pool system ensures absolute network security.
In Aphelion all the nodes that join the network are divided into special pools:
• Power Pool is where all the transactions and activities take place;
• Exploit Finding Pool runs a security and performance check on the nodes;
• Audit Pool runs an audit of the nodes in the network and broadcasts the rules over the whole network;
• Maintenance Pool is where extra nodes are waiting to be distributed to other pools.
Part of the nodes is involved in the main work, powering the whole network. After Aphelion examines the node’s possibilities, it gets assigned tasks that it will be capable of completing in a short, set amount of time. So, a faster node completes more transactions in a second than a slower node, but no bottleneck is created, as there is no need to wait on the slower node.
The second pool is the security pool. Nodes that have been checked and verified by Aphelion and are deemed to be trusted are constantly looking for various exploits and vulnerabilities. With Aphelion’s help they are surveying the big picture and ready to examine and solve all suspicious activities.
To increase security even more, the third pool is introduced that watches over the watchmen in the second pool. The audit pool conducts tests and checks to the other nodes without those nodes knowing if they are solving a real issue or being subjected to a test. If a node becomes suspicious through its actions, it is the audit pool nodes that will flag them and, in dire cases, remove them from the network.
As the system is fast and self-reliant, Aphelion decides where each node will work and what tasks will it do. If a node leaves the network or is considered untrustworthy, a stand-by node from the fourth pool, the maintenance pool, takes its place, thus ensuring a constant flow of uninterrupted work.
Aphelion works in the following steps:
1. Nodes join the blockchain network.
2. Nodes go through an extensive evaluation process which will decide their place in different pools of the network. This data will help to predict the performance of the node in the network.
3. After the power index is gathered from nodes, Aphelion algorithm assigns them to one of the four pools.
4. Transaction is received and announced to power pool nodes for verification.
5. Aphelion generates the rules for transaction verification among the nodes which allow the protocol to make an autonomous decision in adding the transaction to the respective block or it makes the decision based on a prediction from the previous dataset using classification agents.
6. After the transaction has passed the verification round from the nodes and the autonomous path directed by the protocol itself, the verified data is then forwarded to the audit pool in order to ensure that the verification carried out is valid.
7. Audit pool scores the whole process carried out by the nodes and classification agent using a scoring mechanism for future improvements.
8. The nodes from the respective pool instantly transmit information to the whole swarm with multi-layered gossiping.
9. The nodes will receive rewards for voting and validating the blocks.
10. Aphelion consensus learns through each dataset, and the more it learns the better it is able to predict and detect the performances, and hold the decision-making power.