I tought RAID10 was mirrored RAID5

 

Nope. The whole POINT of raid5 is that you don't have to mirror it. The redundancy is built into the layout of the data on the disks. This layout is one reason raid5 is SO horribly slow writing data. Remember, there's really no such thing AS raid 10 or raid 01. It's really raid 0 then together WITH raid 1 OR raid 1 then together WITH raid 0. And actually raid 0 isn't technically r.a.i.d (Redundant Array of Inexpensive Disks). The key word is how to get some redundancy. Raid 0, striping, is a performance enhancing thing. As someone already pointed out, there's raid 1 through 5. But because disks have gotten so fast and cheap, the only thing people use anymore is raid 1 (which maximizes performance at the expense of capacity) or raid 5 (which maximizes capacity at the expense of performance).

Also, you can have "raid levels" built into your disk controller, commonly called "hardware raid" or you can implement "raid levels" on the OS level, commonly called "software raid". I'm using Sun Solaris with Veritas, which implements raid from the OS. This is also a trade off, disk arrays that have "hardware raid" are generally more expensive than "dumb" disk arrays. But, raid on the OS level takes CPU cycles that could otherwise be used for application programs. So, just a tradeoff thing.

Kinda like maximizing bimbo attraction with a viper, vs eating perhaps. LOL!!!!!!!!

 

Off all the solutions RAID 1 or 2 are the slowest. RAID 1 and/or Mirroring is the slowest because the same data must be written TWICE. RAID 0 is by far the fastest solution, but it does not offer any level of redundancy and possiblility of a failure scales with the number of the drives. RAID 3 is multiple drives with a single parity drive. RAID 2 and 4 are less cost efficient than other designs and really are not worth considering.

RAID 5 is considered the best of both worlds. You achieve speed with a large stripe set, and you have inherant redundancy because you can lose any drive in the array and still have access to all of the data. Why doesn't everyone do it? Its the most expensive of the viable solutions.