Have you guys and gals seen this: http://blog.macsales.com/25650-mac-pro-2013-three-thunderbay-4-drives-breaks-3900mbs
?
?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
2.6GB/s (2600MB/s) RAID array transfer rate via Thunderbolt 2
- Thread starter FireWire2
- Start date
- Sort by reaction score
Link probably should have been
http://blog.macsales.com/25850-additional-testing-with-three-thunderbay-4-drives-and-a-mac-pro
or the "previous" one pointed to by that article. [ Not sure if there later tweaked and revised something before/after those two. ]
I though it has some sort of redundand, but it's a RAID0
It's quite a risk when use it
I don't see any other solution without raid0? I believe that if you really want the performance talked about in this thread you should really see this speedy array as nothing but a temporary drive that you use for editing. Storage and backup should still be part of your workflow so that a failure in the raid0-setup should only give you some downtime, not lost content.
They are obviously shooting for maximum drag racing numbers so RAID 0 is pretty much how you'd do that. There are Thunderbay enclosures that ship with SoftRAID bundled ( it is the RAID 5 versions ). You probably can stripe the SoftRAID across/intra these enclosures to get slower but still much higher than a single TB v2 link numbers out of it.
They are hitting 3600MB/s and if only need 2600 MB/s layering RAID 5 on top probably isn't going to be a 1000MB/s hit on overhead. Coupled to double digit TB capacity demands part of the problem's solution is going to draw in a larger number of drives.
As far as safety goes if pulling static, archival RAW from pool the data should be fully backed up anyway. Plain everyday general read/write usage, no. But 3600 MB/s isn't general read/write usage either.