... which is going to be quite different to what you see with RAID 5.
That's not quite true. Regardless of the RAID level in use, RAID 0, RAID 1, RAID 5, or RAID 10, underlying drive performance will impact system performance. For example, in RAID 5, total performance will be limited by the slowest drive in the array. One advantage of RAID, however, is that it is commonly implemented with a hardware controller which offers substantial performance improvements through buffering, asynchronous reading/writing, etc.
Also, the earlier point about lower numbered cylinders (closer to the center of the drive) having faster performance is also true. Most UNIXes, if given half a chance, will locate the swap partition, which is used to augment physical memory with disk storage, as the first partition on the drive. I don't own a Mac, but as the underlying OS is a UNIX I'm willing to bet that this applies there, too.
Finally, all file systems use some sort of indexing to both locate files on the drives and to handle free space. As file systems fill up--with a large number of "small files" versus a few huge files--the indexing gets more involved. It also gets harder to allocate contiguous free space, which is worsened when files are added and deleted. NTFS, used by Windows, is particularly prone to this latter issue, which is why Windows has a defragmentation routine. Other file systems--UFS, ZFS, Rieser, etc--suffer from this issue less.
Installing a solid state drive for your scratch disk and possibly your catalog--remembering to back up your catalog frequently--is the best way to improve performance. After all, most of your disk activity deals with reading previews and filtering on metadata rather than actually reading RAW files and writing output files.
Your mileage may very.
Tom Frerichs