How fast can SQL Server write to disk, and why am I not reaching those speeds?
A question that regularly comes up in ETL scenarios is whether SQL Server can read and write faster if you divide your tables and indexes over multiple filegroups and your filegroups over multiple files. My standard thinking on this was always as follows: Depending on whether you can write those files to different disks or not this will be the case. In a single disk scenario it should not matter at all. I started out…