Lucene vs Sphinx
Lucene and Sphinx are both free, open-source search engines.
Sphinx is maintained by Sphinx Technologies Inc and is distributed under GPL license.
Lucene is maintained by Apache Foundation and is distributed under Apache license.
[edit] Architecture
Sphinx is a stand-alone server, implemented in C++ and comprised of batch indexing tool (indexer), a searching daemon (searchd), and several other maintenance utilities.
Lucene project includes several subprojects, most notably Lucene Java, which is a library, and Solr, which is a stand-alone server. Both Lucene Java and Solr are implemented in Java. Solr is built around Lucene as its core.
[edit] Features
Lucene supports live updates to the index. Sphinx versions 0.9.x only support rebuilding the whole index. (However, one can only rebuild tiny "delta" index that would contain just the recent updates.) Sphinx versions 1.x support two index backends, "disk" one that can only be rebuilt from scratch, and "rt" one that supports live updates.
Sphinx supports direct imports from MySQL, PostgreSQL, or any ODBC compliant database for faster indexing speeds.
[edit] Performance
Lucene and Sphinx are both very scalable. Lucene backs the search features of Wikipedia and Digg. Sphinx backs search at Craigslist, and search and browsing at Slashdot.
