Using Iceberg allows us to pick the optimal "big data" compute environment for the specific requirements we have. There's no need to limit yourself to a single solution.