Determine the performance differences between taxonomy and ssdv faceting

### Description

Since January of this year (after bringing hierarchical faceting to SSDV - LUCENE-10250), taxonomy and SSDV faceting have feature parity (as far as I'm aware). I'd like to consider moving to a single implementation to simplify things for our users (deprecating one of the existing ones or moving it to the sandbox). As a next step, it would be good to understand specific pros/cons in terms of performance between these two implementations to help guide our thinking. We can start with the existing nightly benchmark tasks, but may want to consider if any additional tasks would be helpful in eliminating blindspots.

Start with Lucene's existing nightly benchmark tasks and compare taxonomy faceting with SSDV faceting, which gained hierarchical faceting in LUCENE-10250. Identify relevant performance differences, document specific pros and cons, and note whether additional benchmark tasks are needed to address blind spots.

Written by the indexing model from the issue text.