Lucene feedback from Atlassian JIRA and Confluence
Mike Cannon-Brookes from Atlassian has posted (some moons ago) two interesting presentations related to Lucene and Atlassian's feedbacks from Confluence and JIRA.
My favorite is the first one: Lucene: Generic Data Indexing. It's a nice introduction to the benefits of Full Text search engines, as well as the gotchas you will face.
I found their use of FilterQuery as a cross-cutting concern implementation for security fairly interesting.
The first presentation also quickly address some of the indexing strategies (synchronous / asynchronous) depending on the product requirements. Mike goes a bit deeper in the second one by describing some clustering solutions and the one they choose.
JIRA has gone very far in its use of Lucene, I am not sure I would have gone that far, but that's definitively a very interesting extreme use case, and very successful :)
Comments