Replication & Peer Discovery (Future)

Replication and peer discovery is the planned subsystem that lets the nodes hosting a space find each other and synchronize authorization state as a first-class peer-to-peer network — modeled on how Bitcoin and Ethereum nodes discover peers and gossip — rather than relying on manually configured hosts.

Why

A space's consistency model already defines how events are ordered (epoch-based, hash-linked DAG, LWW CRDT); replication defines how those epochs propagate across hosts, and discovery defines how a host finds the other hosts to propagate with. Making both first-class is what turns "trusted nodes you point at each other" into a self-organizing host network, so a user's data stays available and consistent without a central coordinator.

Current status

Actively under development. The replication/ module exists in tinycloud-core but is not declared in lib.rs — it is not compiled or mounted — and peer discovery is not implemented as of 2026-06. The whitepaper sketches discovery via host-delegation multi-addresses, DID-document service endpoints, and a manifest registry, but no gossip/broadcast or epoch-sync is wired up. Treat the discovery mechanism as design intent.

See also

The live concept tracking the unmounted module is replication; it builds on consistency, the epoch DAG, and conflict resolution. The "Recon" light-client direction is a related but distinct line of work. Sits in Layer 1; see the roadmap.

Sources

• tinycloud-node: replication/mod.rs, replication/recon.rs (present, not mounted in lib.rs)
• whitepaper: README.md §5 (peer-to-peer replication / host discovery)