Scaling / optimizing search on netlog

This document discusses using Sphinx to optimize search and content filtering for the social network Netlog. It describes Netlog's search needs, types of searches, and challenges around scaling search across distributed content. It then evaluates solutions using MySQL, external services, and Sphinx, ultimately selecting Sphinx. The document outlines how Sphinx works, how Netlog has implemented it using main and delta indexes, and future plans including indexing shards and tag clouds. It concludes with tips for using Sphinx effectively.