Regarding point 2, my implementation was to have a duplicate database server where all the search queries are sent. This would ensure that the search wouldn't s

Regarding point 2, my implementation was to have a duplicate database server where all the search queries are sent. This would ensure that the search wouldn't slow down the main database. And most of the duplication would happen quick enough so that the search results were almost up to date.