After previous experiences with lucene.net, which weren’t all bad, I’m enjoying experimenting with Solr.

“Solr is an open source enterprise search server based on the Lucene Java search library, with XML/HTTP and JSON APIs, hit highlighting, faceted search, caching, replication, and a web administration interface.”

It was very simple to get an instance running locally with jetty and I had a test python client connected and searching using simplejson within minutes.

Here’s my question. I’m curious how applications are using MultiCore on Solr and what limits they’ve been able to take it to.

Let’s say you’re building an application like SalesForce where you have a lot of clients who technically have no need to search across each other. As well, each client may have multiple indices, let’s say 3 for the sake of having a number. Do you?

1. Use MultiCore to dynamically generate unique indices for each new client resulting in creating and managing 3*(number of clients) indices.
2. Create 3 potentially massive indices which will clearly impact search performance.

Assuming Solr and MultiCore can handle #1 then it seems the clear winner to me. Anyone??

This entry was posted on Tuesday, April 15th, 2008 at 5:46 pm and is filed under Geek. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.