Mike, one of the co-founders of Atlassian has posted some interesting slides on Lucene, a Java OSS reverse indexing project. The slides give some insight into how Jira and Confluence make use of Lucene’s capabilities.
About two years ago, I was driving to the Snowy Mountains with Atlassian’s other founder Scott, who was waxing lyrically about the virtues of Lucene and how it could help solve Campaign Masters then current performance problem.
Two years on, Atlassian is five years old and Lucene.Net has long since been the solution to one of our biggest teething issues. Again we are looking to Lucene to do for us what our database can not, on a clustered scale. Mike’s slides cover their scaling issues and what their options and problems were for clustering. Fortunately, our situation seems much less complex and we will be testing the speed of Lucene using indexes on network storage.
Lucene.Net, AFAIK, is up to 1.9 and doesn’t currently support the Update/Delete features of it’s 2.1 Java sister.