Snark: P2P propagation Network Node
After storage and execution are addressed, we are left with networking. Forks increase as block size increases. Forks increase as block time decreases.
We extend Boot Nodes to provide further functionality as first point of contact Network Nodes. Boot Nodes are the cornerstone of the ecosystem. They are assumed at least to some extent, as always online. They are the entry point for new nodes into the system. P2P propagation using gossip, is a slow process that spreads across a network.
With Consensus First Protocol, we no longer needs blocks. So instead our focus is on transactions. Transactions are sent to Network Nodes as a point of first contact, and then P2P propagation occurs as required or as a fallback should Network Nodes be forced offline. Transactions are first pulled from Network Nodes, before neighbor requests occur.
Network Nodes are configured on an infrastructure layer. Using data center infrastructure to connect Network Nodes and provide minimum distance routing. Network Nodes are designed to propagate amongst each other based on the shortest configured route. As a secondary mechanism they begin their gossip protocol.
By reducing time to transfer processing can occur faster and potential forks can be decreased.
With the design of Consensus First Protocol, we also see the Node Info struct as well as the Node Info transaction list. Transactions are encoded with short references to avoid full transfer. Network Nodes are also a first point of contact for the full Node Info struct.
The above implementation allows for standard propagation with a first point of contact mechanism to achieve faster transactional volume.