An airport ops DC synchronizes model artifacts bidirectionally with the cloud and must detect corruption in transit. Which data-synchronization control should be evaluated?

An airport ops DC that synchronizes model artifacts bidirectionally with the cloud and must detect corruption in transit should evaluate checksum or integrity verification on that bidirectional model-artifact synchronization. Weights and graphs must match what was published; silent bit-rot or truncated transfers can poison inference on either side of the hybrid link. File-size and timestamp reconciliation on bidirectional model-artifact synchronization only compares metadata; it cannot detect bit-rot, truncation, or silent corruption inside the transferred weights, so it is not an integrity control for transit corruption. Version identifier comparison on bidirectional model-artifact synchronization only confirms that both sides report the same version label, but it does not verify that the artifact bits transferred intact, so it will not catch corruption in transit. Replication acknowledgment tracking on bidirectional model-artifact synchronization only confirms that the peer acknowledged receipt, not that the payload arrived intact, so it cannot detect corrupted model artifacts in transit. Exam caveat: Domain 2.5 evaluates data synchronization quality for hybrid AI, including integrity of artifacts moved between sites—not merely that a copy job exited zero. Operational check: enable checksum or equivalent verify on push and pull of model artifacts, inject a deliberately truncated file in a test lane, and confirm the sync path rejects it before airport inference loads it. Airport model promotion pipelines should treat failed integrity checks as release blockers, the same way they treat failed functional tests. Bidirectional sync without verify can silently reintroduce a bad artifact from cloud back to on-prem after a rollback, so both directions need the same checksum policy and retained evidence of matches.