I just want to share a recent discovery of mine: A Lucene Directory based on AWS S3 SDK https://github.com/albogdano/lucene-s3directory and it's coming to Lucene: https://github.com/apache/lucene/pull/13949
Of course, by itself, the performance is atrocious (see the footer on the page) but if our general purpose "blockcache" https://github.com/apache/solr/tree/main/solr/modules/hdfs/src/java/org/apache/solr/hdfs/store/blockcache (hiding out as a dependency-free package inside our hdfs module) were layered on top; it'd be something interesting. I have a year+ old proposal somewhere on how Solr might use a remote Directory (be it HDFS or S3 or whatever; doesn't matter) for indexing but use a standard MMapDirectory PULL replica for query. ~ David Smiley Apache Lucene/Solr Search Developer http://www.linkedin.com/in/davidwsmiley