Originally Posted by ZenerDiode
So I've seen this quoted a lot, but I haven't seen any proof of this. Could you please elaborate...this issue has caused me a lot of concern.
Rough calculations of SSD lifespan with NAND having an estimated 3,000 write cycles (estimated for 25nm) at 100GB writes per day. Formula's not the most accurate but it makes for a good enough guess.
Total Host Writes: 768,000 GB or 750 TB
768,000 GB / 100 GB/day / 365 days/yr = ~21 years
Total Host Writes: 384,000 GB or 375 TB
384,000 GB / 100 GB/day / 365 days/yr = ~10.5 years
Total Host Writes: 192,000 GB or 188 TB
192,000 GB / 100 GB/day / 365 days/yr = ~5.3 years
40GB (@25nm 3,000 PE)
Total Host Writes: 120,000 GB or 117 TB
120,000 GB / 100 GB/day / 365 days/yr = ~3.3 years
40GB (@34nm 5,000 PE)
Total Host Writes: 200,000 GB or 195 TB
200,000 GB / 100 GB/day / 365 days/yr = ~5.5 years
From crazy people determined to kill their SSDs:http://www.xtremesystems.org/forums/...e-25nm-Vs-34nm
The Kingston SSDNow 40GB (rebadged Intel X25-V 40GB 34nm) appears to be crazy reliable. Even after 466TB written, it still passes the data retention test when SSDs with higher capacity using different controllers have already failed.
The Intel 320 40GB also performed beyond expectations.
You'd likely see some random controller failure before you hit the limit of NAND endurance.