The baseline:

• Locations (nodes) are indexed as sub documents of the master content
• Some location meta data is indexed with the main content object where the multivalued nature of it (for example priority) is indexed as single valued fields to enable sorting in some edge cases
• The locations returned in the response apply the location specific filters as well, so only the relevant ones are returned
• As in Solr locations can be searched for independently, it is also possible to construct for example tree menus in a very efficient way (performance wise)