The useful life of solid-states drives (and how to improve their endurance) is a topic that pops up occasionally. Now, you could speculate on SSD life spans until the cows are out of write cycles, or you could embark on a real-world test, something one outlet has done and at 500TB in, things are getting interesting.
Image: Yutaka Tsutano / Flickr, licensed under Creative Commons 2.0
Started back in August of last year, Tech Report’s SSD endurance test recently hit the 500TB mark. In a nutshell, the test subjects five drives — a Corsair Neutron GTX 240GB, Intel 335 Series 240GB, Kingston HyperX 3K 240GB, Samsung 840 Series 250GB and a Samsung 840 Pro 256GB — to a constant stream or writing tasks using Anvil’s Storage Utilities. The objective is to see how long it takes for the drives to fail.
Of those drives, only the Samsung 840 Series has seen significant wear, which occurred 100TB in, though if you check out the reallocated sector graph, this wear has continued in a linear fashion.
Before you get your SATA cables in a twist, the drive’s spare flash, put aside specifically to compensate for the 840’s less hardy triple-level cell memory, has compensated completely for the wear and as such, has not lost any capacity.
On top of this, as the Tech Report’s Geoff Gasior explains, all the drives have been subjected to the equivalent of 140GB of writes a day, a level of use the average user would struggle to hit, so despite the 840 Series’ apparent fragility, in the grand scheme of things, it’s not really an issue.
The SSD Endurance Experiment: 500TB update [The Tech Report]
Comments
10 responses to “How Do Various SSDs Fare After 500 Terabytes Of Activity?”
I read some recent article (sorry no link) that SSDs don’t handle power outages very well compared to hard disks.
Something to consider.
i don’t know about that, my area has frequent brown outs in the summertime and my ssd based pc has not had ant trouble.
Our electric hot water system was on a dodgy circuit that we didn’t get around to fixing for like a year. We would lose power at exactly 8:21am roughly 2 days a week. Never caused an issue with any of the plethora of SSD based PCs we have.
I would think ssds would fair better because the disk isn’t being shut off while at full rpm and the arm out on the platter.
Someone correct me if I’m wrong (Pffft – this is the internet, that would never happen)… but if a platters drive looses power while its spinning, doesn’t it use the remaining energy in the drive to park the head?
If there is no power to the motor how does it park the head? The platter will keep spinning from momentum.
The heads aren’t ‘powered’ as such. It electromagnetic force, so in the event of a power outage the heads are parked automatically.
http://www.techarp.com/showarticle.aspx?artno=84&pgno=2
(Pffft – this is the internet, that would never happen)
Ha ha ha ha I just spit my beer out.
agreed, i thought one of the main selling points of SSD (after speed) was its resilience to these sort of events due to the lack of moving parts
Agree that SSD should be far more resilient to power fluctuations and other physical trauma than spinning HDD’s. I have moved all my computers to SSD, especially the portable ones.
Spinning HDD are my favourite and most underrated piece of modern engineering. I read somewhere a long time ago that when it comes to spinning HDD’s the tolerances are monumentally tight and mind blowing. When spinning at working speed, the distance and speed of the r/w head above the platter is the equivalent of flying a 747 jumbo at cruising speed of 900kph… 5 feet off the ground… and not hitting anything… ever. Until some human drops the laptop while the drive is spinning, causing a head crash of roughly the same outcome as would befit the 747 described above. Hence why your HDD dies. Except in a vast majority of cases…. it doesn’t… completely. Again the engineering comes into play and there is enough 10011011 1010 10101 voodoo redundancy built into these things they can usually RECOVER from all but a full speed 90 degree nose dive. AMAZING.