In the beginning was Lucene...
It’s still Lucene underneath. Just like SOLR.
The bread and butter of elasticsearch is in the aggressive sharding, shard allocation, replication and promotion. It essentially “clusterises” lucene.
I can’t wait for something similar to what elastic did for lucene to happen to redis or Postgres.
Did you look at ZomboDB? It allows you to use Elasticsearch as a native Postgres index type.