provide a feature for clustering data with the same index key around the same location on disk.
The would be to avoid seeks when looking up documents in an one-to-many index query.