The Apache Iceberg project is taking on the challenge of how to bring database-like agility to petabyte-scale datasets stored in open cloud storage.