Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide


  1. 1. Apache solr module vs Searchlight Comparison of two modules to provide faceted browsing capability via a solr back end.
  2. 2. Term Clarification I will refer to the apache Solr module, in this document this refers to the Drupal module of the same name not the apache Solr software itself. If I am referring to Apache Solr itself I will not use the word module
  3. 3. Overview <ul><li>Both modules act as an interface between Drupal and a apache Solr search system.
  4. 4. This takes some load off Drupal which only includes a SQL based search by default, which can't scale.
  5. 5. Helps Scalability in terms of content indexed and number searches performed.
  6. 6. Assumption is that you will use Solr and stick with it for some time. </li></ul>
  7. 7. Metrics <ul><li>6.x-1.2 is stable version
  8. 8. Installed on 5,000 sites
  9. 9. 5 Active developers
  10. 10. Average bug lifetime of 3 days 8 hours in last month
  11. 11. Over 20 Add on modules
  12. 12. Created Dec 2007 </li></ul><ul><li>No Stable version
  13. 13. No install metrics
  14. 14. 4 Active developers
  15. 15. Average bug lifetime of 18 hours 28 minutes in last month
  16. 16. No known add on modules
  17. 17. Created Mar 2010 </li></ul>
  18. 18. Safe option For a standard site which was looking for a faster search the Apache Solr module is the safe option. It is used on a large and it has a release which is marked as stable.
  19. 19. Digging deeper The Apache Solr module was built as a replacement for core search. So its' root is in Drupal terminology “ Among the extra features is the ability to have faceted search on facets ranging from content author to taxonomy to arbitrary CCK fields ” This may be a big limitation for a company which is big on content aggregation
  20. 20. Drupal as a Distributed CMS <ul><li>Traditionally a CMS has sat on one site where content is entered or imported and stored in a single database.
  21. 21. Drupal has the ability to be a distributed CMS, where content is displayed from a number of sources and not stored locally.
  22. 22. The Apache Solr module, works on the assumption of local data
  23. 23. Searchlight, works from views so can index data from any source. </li></ul>
  24. 24. Views in Views out <ul><li>Drupal views is a module which allows data to be searched and exposed in an extremely flexible way.
  25. 25. Views 3, allows plug ins to be written from any accessible data source, including SPARQL and Solr.
  26. 26. Any content aggregation system which uses Drupal should leverage views as much as possible.
  27. 27. Hence for content aggregation searchlight is a better match </li></ul>
  28. 28. Flexible Architecture <ul><li>It is important to keep architecture options open
  29. 29. At the moment I would recommend
  30. 30. searchlight – Solr - searchlight
  31. 31. In the future this may be
  32. 32. custom perl – Solr – solr views
  33. 33. Keeping the architecture loosely coupled will allow for further expansion in the future.
  34. 34. Drupal, Solr, Views and Searchlight will allow for a lot of flexibility </li></ul>Search Indexing Aggregation Result Display
  35. 35. Provisos <ul><li>The above is a recommendation based on research but would need a spike solution to prove the concept before a final decision
  36. 36. Searchlight is still under development and will take extra effort to keep in sync
  37. 37. Some effort may be needed to fix searchlight bugs, or implement features we need
  38. 38. May not be as stable as we would like initially
  39. 39. Load test your solution constantly </li></ul>