This is needed for synchronisation with a file system. Then we need to address
each bucket by key (e.g. filename). But we also need to address by the order of
ingestion for processing with progress indicator.