Disks for Data Centers

Google released a whitepaper on their wishlist for HDDs.  It's somewhat amusing that they appear to think that ECC is still used.

"An obvious question is why are we talking about spinning disks at all, rather than SSDs, which have higher IOPS and are the 'future' of storage. The root reason is that the cost per GB remains too high, and more importantly that the growth rates in capacity/$ between disks and SSDs are relatively close (at least for SSDs that have sufficient numbers of program-erase cycles to use in data centers), so that cost will not change enough in the coming decade."

mhddfs & GlusterFS

I haven't posted any links for a while.  They've been piling up.

There's a FUSE plug-in for Linux called mhddfs that aggregates storage devices into a single virtual filesystem.

I'm thinking I'd like to use a SATA port multiplier to pull together a bunch of drives.  (Such as those made by DAT Optic.)  Unfortunately, that system wouldn't have any redundancy, although one could still copy to individual drives manually to establish multiple copies.

I also found reference to a distributed filesystem called GlusterFS that appears worthy of further investigation.

Addendum: SATA-IO has an article about port multipliers and the difference between implementations, i.e. command-based switching (CBS) vs. FIS-based switching (FBS or just FIS).