Peers and discovery
Production nodes will need a way to discover compatible peers and verify that they are speaking the intended network and protocol version. DNS seeds, boot nodes, peer exchange, static peers, or other discovery mechanisms remain implementation choices.
Conceptual message classes
Message names and binary encodings are not yet final.
Propagation
Nodes should validate inexpensive structural rules before relaying data and should not amplify obviously malformed or invalid messages. The exact gossip strategy must balance fast propagation with bandwidth, privacy, and denial-of-service resistance.
Synchronization
A node joining after being offline needs a secure method to determine the accepted finalized state and obtain any data required for its operating mode. Snapshot and proof-based synchronization may be evaluated, but verification rules must be explicit.
Network separation
Mainnet, public testnet, and local development networks should use distinct network identifiers and address or transaction protections so that data created for one environment cannot be mistaken for another.
Resilience and abuse controls
Peer limits, request budgets, message-size limits, duplicate suppression, ban or reputation policies, and resource accounting are part of the production threat model. These controls should defend nodes without turning a few infrastructure operators into gatekeepers.