This blog post describes concurrent query execution in Lucene (at the IndexSearcher
level), and claims:
Concurrent query execution is a surprisingly little known sleeper feature in Lucene, since it is not yet exposed in Elasticsearch nor Solr, two popular distributed search applications that build on Lucene. Their concurrency model is instead concurrent search across index shards (usually on different servers) for a single query, but using single-threaded search within each shard.
Is that still true (the blog post is from 2019)? Is there anything written about this in detail - how does a shard process a search, and handle querying across multiple segments?