opc.three (11/24/2012)
It's quite possible that leaving the FILLFACTOR at 100 would make the most sense. With the SSD in play, the page splits and ensuing fragmentation may be less of a concern than absorbing a 25% space-hit to the buffer pool for each index where the FILLFACTOR is lowered.
At first, and even though it goes against my inner data-troll, I thought that would be fine especially considering the blazing speed of SSDs. It might still be fine if you have a reorg going on a regular basis. My concern is that when you have a lot of pages of GUIDs, inserting a lot more GUIDs will cause page splits possibly on almost every page because of the extremely random distribution. In theory, if you have 10,000 pages of GUIDs where the CI is on the guid and you insert just 10,000 new GUIDs, it could cause a page split on each and every page. Now you suddenly have the space equivalent of a 50% FILL FACTOR on some relatively very expensive hardware.
Considerig the speed that a reorg or rebuild might happen on the SSD, that might still be the way to go, though.
--Jeff Moden
Change is inevitable... Change for the better is not.