A hosted search service for Drupal

Jay just announced Acquia's plans to get into the hosted search business. By doing so, we hope to provide Drupal sites a (i) faster, (ii) more scalable search solution that (iii) can work across multiple sites, that (iv) can handle broader content, (v) that offers faceted navigation and more.

Content management systems with basic keyword search are going nowhere but out of style. Content management systems that do search really well can improve a site's navigation in new and significant ways. We want Drupal sites to be in the latter category.

As we've learned with big sites like drupal.org, it is extremely difficult to do search well from within Drupal, hence the need for a powerful hosted search service. By offering a hosted search service, Acquia continues to tear down barriers for adoption so Drupal can fulfill its full potential.

We're still working on this new service, and we don't know yet when it will be ready, but let us know if you're interested in being an early private beta tester.

Comments

Webseiter (not verified):

Sounds interesting, but this raises the question whether or not the Drupal search capabilities need to be reworked to gain more flexibility (both for frontend users and administrators) and performance.

Dries:

The Drupal 7 search module already has some improvements over the one in Drupal 6, and I hope we can continue to make improvements to the built-in search module. A hosted search service does not change that desire.

The search industry is reported to be bigger than the CMS industry, so Drupal's search module is only scratching the surface of what is possible. One has to realize that search is a really hard problem with many different use cases -- it is going to be hard for Drupal to compete with commercial-grade search engines, to keep up with the advances in search technology and to continue to be able to run on shared hosting environments.

That said, the Drupal search module has its purpose and target audience, and because of that, we need to keep investing and improving Drupal's built-in search.

greggles (not verified):

And 6 has vast improvements over 7. So, I agree that improvements to Drupal core search are likely to continue. However, we also have to recognize the benefits of these external search systems.

This is an interesting additional service - well done, Dries and Acquia.

Robert Douglass (not verified):

Google can never know as much about your site as you do. Google would have a very hard time coming up with faceted search that lets you filter by content type, for example. Google also cannot respect node access privileges, meaning your private content is unsearchable by Google. These are all problems that the ApacheSolr module solves.

Bojhan Somers (not verified):

So how are you competing with other offerings who are quality wise better? I think it's a very intresting move, but hard to tackle - unless there are some rigour efforts to improve the search.

Fabio Varesano (not verified):

Well.. they, as a Drupal based company, probably want to deeply integrate with Drupal.

Knowing it's core structures such as how contents are stored, displayed etc.. it probably would work pretty good.

ken (not verified):

I'd be interested in a beta test. Get in touch with me

Ken

PWolanin (not verified):

Using Zend Lucene you have the advantage of immediate and intimate access to the index. However, it's using an older Lucene version/API, it's limited (most likely) to interacting with a site on a single server, it requires additional resources on the local server in terms of PHP execution, memory, etc. Since Solr uses the newest java Lucene version, it's capable of supporting more advanced features including geo searches as well as combined indexes from sites running on multiple servers as well as very large indexes.

Both Solr and search Lucene are obviously built on Lucene and at some level can support similar features. There are some trade-offs with either option, but Acquia is offering a hosted option for Solr is that it removes a lot of the potential friction points for using Solr.