SSD Reliability In The Real World

Megalith

Megalith

24-bit/48kHz
Staff member
Joined
Aug 20, 2006
Messages
13,000
There is a new paper out that offers insight into how SSDs fare. One major conclusion is that age has a much higher effect on an SSD than usage.

Two standout conclusions from the study. First, that MLC drives are as reliable as the more costly SLC "enterprise" drives. This mirrors hard drive experience, where consumer SATA drives have been found to be as reliable as expensive SAS and Fibre Channel drives. The paper's second major conclusion, that age, not use, correlates with increasing error rates, means that over-provisioning for fear of flash wearout is not needed. None of the drives in the study came anywhere near their write limits, even the 3,000 writes specified for the MLC drives.
 
i still have Data with heavy i/o write load on a raid 0 mechanical drive
but all software ( mostly reads) on SSD...

Raid 5 mechanical drive for longtime storage
 
My Intel 335 SSD (MLC) from 2012 has 2.54 TB written and has yet to move off the 100 mark for Media Wearout Indicator (which counts down from 100% to 0%). I've made no effort to protect it from writes, and it has the pagefile on it.
 
I've had my Crucial M4's (2 in raid-0) for over three years now and haven't had a problem with them at all. Best upgrade I ever made. I've had a couple mechanical drives fail during this time.
 
have a few tablets that are all flashed based and they have lasted despite heavy usage... none of them are even close to the reigning champ though a 8+ year old WDC Caviar Black that is still in use
 
evilsofa said:
My Intel 335 SSD (MLC) from 2012 has 2.54 TB written and has yet to move off the 100 mark for Media Wearout Indicator (which counts down from 100% to 0%). I've made no effort to protect it from writes, and it has the pagefile on it.
Click to expand...

Similar experience here. Picked up an Intel 320 series 80GB drive back in 4/11 and used it as my boot drive for a little over two years until I installed a Samsung 840 Pro 128GB as the boot drive and moved the Intel to scratch duty. Since I wanted to keep the load on the new Samsung low I moved the page file, temp folders, and browser caches to the Intel. As of now the Intel has nearly 25,000 power-on hours and 4.86TB read / 4.06TB written and the media wear out indicator is at 99, so it looks like it is still in pretty good shape. The Samsung has far lower values across the board so it too should be in pretty good shape.

The question now is at what "age" do error rates begin to increase? That Intel drive is approaching 5 years old - actually probably already there since it was most likely "born" at least a couple months before I bought it.
 
I have written 33.74 TB to my 1TB Samsung 850 Pro and have 0 UBER count. The wear leveling count is at 34, but I have read that these drives can go way beyond what SMART reports.
 
I still back up to a spinning disk. If there's one thing I like more than a reliable storage medium it's two reliable storage mediums.
 
The article does emphasize that it's even more important to back up SSDs because their UBER rates are higher than those for HDDs. But we already knew that running without a backup on any medium is stupid.
 
Killerxp100 said:
How long are we talking here? 5, 10, 20 years?
Click to expand...

Looks like 6 years.

They did say this though:
  • 30-80 percent of SSDs develop at least one bad block and 2-7 percent develop at least one bad chip in the first four years of deployment.
"30-80%" .. They couldn't narrow that down a bit? I mean that's a pretty big range.
 
I've purchased around 100 over the past 5 years. So far 1 has failed. It was a Sandisk.
 
the only ssd that's failed on me had a sandforce controller, all the rest have been fine. still do a weekly backup of my boot drive though.
 
I've got a OCZ Agility 3 from 2011 and still runs perfectly. My 2nd oldest drive is from 2012, a corsair Force GT. No issues with that one either.
 
Its been known that NAND ages, its also expected that the smaller the process it is on the quicker it ages.

That's one of the reasons Samsung went back to such a higher/older process with with their 850 line.
 
My Crucial m4 512MB SSD from May 2012 is still going strong. Hard Disk Sentinel reports 54 power cycles, 33173 power on hours, 95% wear leveling and 5% of the rated lifetime used. There are 0 errors from any of the other SMART values like "Reallocated Sectors Count", "Erase Fail Count", "Reallocation Event Count", etc.. I see no total bytes read or written, unfortunately.
 
evilsofa said:
My Intel 335 SSD (MLC) from 2012 has 2.54 TB written and has yet to move off the 100 mark for Media Wearout Indicator (which counts down from 100% to 0%). I've made no effort to protect it from writes, and it has the pagefile on it.
Click to expand...

But what they are saying is that media wearout indicator and SSD failure don't correlate strongly, anyway, right?
 
You must log in or register to reply here.
Back
Top