SSDs already have internal RAIDs, kind of. An SSD is made out of a few memory chips and a controller chip. The memory chips are used in a fashion similar to RAID, coordinated by the controller. That's the reason for SSDs' high performance. Larger SSDs tend to be faster, too, for this reason. They need multiple memory chips anyways, might as well access them in RAID.
Assembly is bound to be more expensive, even on SSDs I think. I can't see how this would be otherwise. So, in my opinion, an integrated disk would almost certainly be more expensive in terms of $/Mb when compared to a normal one. I don't want to draw a straw man argument, but can't stop thinking about the reasons it wasn't done already...
Good RAID solutions support hotplugging. SATA supports hotplugging.
Originally Posted by happyclown
Just get 3 drives and make a RAID-5 array. You get 2/3 the total capacity (the other 1/3 is the xor of the first 2 parts, to provide redundancy), with theoretically higher read speed than a single drive, and quite a bit lower write speed (since 2 drives need to be updated everytime you write). RAID-5 guarantees that your data will be intact if at most 1 of the 3 drives die.
If a drive existed whereby drive failure resulted in no(or minimum) downtime, with no loss of data, I would be willing to pay double(even tripple) the price of a normal drive.
RAID provides all those things :).
4. Better would be if the drive had no shared electronics - if one media dies, the drive instantly switches to the second media, with no disruption. But as I said previously, I'll leave the fine technical details as to how it may be achieved to someone much smarter than myself.