In my last post on filegroups (it seems like ages ago), I argued that separating tables and their non-clustered indexes between two different filegroups, with each filegroup on a separate I/O path can improve performance. That article was met with skepticism. I think that's a great, healthy thing. So for those that aren't sold on filegroups ...

In my previous post on filegroups, I tried to make the case that using filegroups (please everyone tell me you use other filegroups than PRIMARY...) can improve database performance. I received a number of comments challenging that notion, which personally I find wonderful - disagreement and real arguing ( no Monty Python  ''This isn't an ...