A couple thoughts here....
First, under 7.2.2 I was able to make a 39 disk aggregate using 500 GB SATA disks - (3) RAID groups of 13 disks each. Yes, that's beyond the 16 TB space limit but it did work....go figure.
Second, under 7.3 the rules have changed in that the parity disks aren't counted -- you can now have 39 500 GB data disks in an aggregate. Given (3) RAID groups of 13 data disks each (with 2 parity disks each), the largest aggregate you could make with 500 GB SATA disks would actually be 45 disks total.
http://now.netapp.com/NOW/knowledge/docs/ontap/rel731_vs/html/ontap/rnote/rel_notes/reference/r_oc_rn_feat73_aggr-size-max-drives.html
This link actually confirms the setup I mentioned under 7.2.2 (although I couldn't find any reason why it worked at the time).
If you're looking at 3 shelves though, I'd probably do (2) 13 disk RAID groups and one 14 disk RAID group -- 40 disks total and keep 2 spares. If 6 shelves, you could do (2) 41 disk aggregates -- would mean each aggregate would have (1) 13 disk RAID group, (2) 14 disk RAID groups, and would leave 2 hot spares overall. If doing 6 shelves via MPHA, 4 GB FC backend bandwidth would be sufficient I think.
Side-note: doing big SATA aggregates does help alleviate the IOP limitations.
