Increases in drive size without increases in unrecoverable error rates

I’m trying to decide on drives for a 24 bay chassis (9271-8icc w/ CacheVault & Intel 36 port expander) and I’ve come to the realization that I just can’t have it all.

Among my goals are:

  • lower power (7200 and lower RPM as I’ll use SSD caching to make up some of difference),
  • higher capacity (8TB and higher thinking fewer drives would equate to less power),
  • SAS interface for better protocol support and dual ports, and
  • a 10^16 UER (unrecoverable error rate).

However, I’m finding products that are more like 7200RPM, SATA, 8TB/10TB, and 10^15. This makes me think that these larger drives are not connected to a software or hardware RAID array but more to a storage pool where multiple copies of data are kept on multiple, unassociated drives. I think this because when you calculate the 10^15 UER of 12x 10TB drives in RAID 6 you statistically have an 80% chance of encountering one. (Praying my math is correct.) If drives were 10^16 UER, that number drops to 8% - much more desirable but still significant where data is concerned.

So my initial questions are:

  1. Where are the large SAS drives with 10^16?
  2. At what capacity size does 10^15 become statistically too unreliable?
  3. What is the fabrication cost difference between SATA and SAS? If SAS has many more benefits (wide/dual port, better error detection, etc) and the cost is insignificant, why hasn’t it been adopted for consumer systems?
  4. Why has it taken this long for SSDs to BEGIN to appear with SAS?

Thanks in advance.

Perhaps it would be best to forward your concerns directly to WD Support via E-mail for dedicated assistance:

http://support.wdc.com/support/case.aspx?lang=en