data-productsUsage Records

Usage Records

The v0.2 Accountability Ledger (Layer 4) records methodological decisions by named people. A traded data product also needs a record of use: who activated it, for what, under which policy, when. v0.3 adds a usage_records entry type to Layer 4.

accountability_ledger:
  usage_records:
    - record_id: "..."
      timestamp: "..."
      dataset_id: "https://..."
      dataset_hash: "sha256:..."
      policy_hash: "sha256:..."
      counterparty: "principal id or DID"
      purpose: "advertising_targeting"
      context: { geo, media, execution_id, volume }
      fee: { amount, currency }
      signatures: [ { kid, alg, sig, role: rail | licensee | licensor } ]

Receipt equivalence. A rail’s signed clearing receipt that carries the dataset and policy hashes is a usage record by construction. The ledger may live at the rail rather than inside the document — in which case Layer 4 carries a pointer (usage_ledger: { url, query }).

From a rail receipt

from orp_validator.usage import usage_record_from_receipt, verify_usage_record
 
record = usage_record_from_receipt(
    receipt,                       # the rail's signed clearing receipt
    dataset_id="https://ds/1",
    dataset_hash="sha256:...",
    policy_hash="sha256:...",
    purpose="advertising_targeting",
)
 
# verify the record's signatures with the same offline machinery as attestation
verify_usage_record(record, resolve_key=my_resolver)["valid"]  # True

This is the concrete integration point between ORP and a clearing rail: the receipt the rail already signs, reinterpreted as a provable record of use that either counterparty can check.

Next steps

LineageConformance