A method for surrogate key generation performed by a physical computing system includes creating a lookup record for a production key of an input record, a key of the lookup record including the production key and a value of the lookup record including both a record identifier for the input record and a unique identifier of the production key within the input record. The method further includes sending the lookup record to a first node of a distributed computing system, the first node determined by hashing the production key with a first hash function, and with the first node, determining a surrogate key for the production key.