Routers

Routers are the network's traffic managers. They sit between labs and nodes, decide which node handles which request, seal traffic, meter it, and enforce policy.

Responsibilities

Node selection. For each request a router picks a node that is online, in the right region, has spare allocation, meets the lab's reputation threshold, and has not recently been used for a related task. Selection is randomised within those constraints so that traffic spreads across the network.

Sealing. The router encrypts the request for the destination and arranges for the response to be encrypted for the lab. Nodes only ever handle envelopes.

Metering. The router records bytes in and bytes out for every delivery and issues a signed routing receipt that a validator can compare against the node's delivery receipt.

Policy. Destination checks, rate limits, the acceptable use policy and per-lab spend caps are all enforced at the router. A request that fails policy is rejected before it reaches any contributor.

Latency. Routers are deployed in multiple regions and requests are handled by the router closest to the selected node, keeping the added latency low.

Trust model

A router can see which lab sent a request, which node relayed it, the destination, and the byte counts. It cannot read payloads. It cannot bill a lab or pay a node on its own; only a validator can write to the ledger, and only when the router's receipt matches the node's.

This means a malicious router could at worst refuse service or misreport a byte count, and misreported counts are caught at verification because the node independently signs its own receipt.

Operation

Routers are operated by Root Network today. The protocol reserves a role for independent router operators, selected by stake and paid a share of network fees, once volume justifies it. Independent routers will be subject to slashing for reporting invalid traffic, censoring participants or violating policy.