: In tests for our implementation (25 concurrent connections generating : search/sort requests), we've seen performance in terms of requests/second : drop by a factor of 10, compared to similar tests executing only search
In case it's not clear from Yonik's response: reuse the same IndexReader/IndexSearcher as much as possible. The biggest factor in sorting tends to be initializing the FieldCache -- which is keyed off of the IndexReader. If you aren't reusing the same IndexReader for all of your concurrent connections your requests/second will be seriously sub-optimal. -Hoss --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
