Question I am over my SSD TBW - should I be concerned?

[alexb75]

Hi,

I have a 240GB Intel 730 SSD that I use for my Windows 10 Boot drive. It's been my SSD since 2014, and now sitting at the current read/write:

Total Host Reads: 24,632GB
Total Host Writes: 101,407GB

Now, based on specs, the TBW on the drive is 91TB, which I am currently over! Yet, the Intel SSD utility says my SSD life is at 75%, so which is it?

Should I be worried and maybe change the SSD for something else? I did get the Intel as it was MLC, and generally known to be more reliable than standard SSDs on the market.

Thanks

 

[Gam3r01]

The TBW is just an estimate for warranty reasons. Most drives will likely go much further than their ratings in proper conditions.
So long as you back up your data regularly (which should always be done anyway) you can keep using the SSD until it does die.
That said, 9 years is a very respectable run from any drive, so you may consider upgrading just for increased capacity/speed given the current value of drives on the market.

 

[USAFRet]

Any storage device is subject to die, from the moment you take it out of the box.

10 minutes, 10 years....
Be prepared for that.

 

[alexb75]

The TBW is just an estimate for warranty reasons. Most drives will likely go much further than their ratings in proper conditions.
So long as you back up your data regularly (which should always be done anyway) you can keep using the SSD until it does die.
That said, 9 years is a very respectable run from any drive, so you may consider upgrading just for increased capacity/speed given the current value of drives on the market.

Thanks, I have backup on Synology... but generally, is this drive at this amount of data write, close to die? Trying to figure out what people have experienced. I've had hard drives fail and I know how it goes, but not SSD.

Would be 8 years plus (bought end of 2014, installed in 2015), I have TWO HDDs in the same desktop that are older even! One WD Red, one Hitachi drive... they're 10-12yrs old easily.

 

[alexb75]

Any storage device is subject to die, from the moment you take it out of the box.

10 minutes, 10 years....
Be prepared for that.

Thanks for your incredible insight!

 

[Lafong]

generally, is this drive at this amount of data write, close to die? Trying to figure out what people have experienced. I've had hard drives fail and I know how it goes, but not SSD.

Would be 8 years plus (bought end of 2014, installed in 2015), I have TWO HDDs in the same desktop that are older even! One WD Red, one Hitachi drive... they're 10-12yrs old easily.

The so-called "bathtub curve" is said to apply to electronic components generally, but I've never seen a proper study of the extent to which it applies to SSDs.

I bought an Intel SSD 320 series, 80 GB in early 2011. I don't know the TBW on it...most likely under 30 somewhere?

It failed completely at about 3.8 years of age for reasons totally unrelated to data writes. I suspect I had written no more than 15 TB to it at the time.

Some SSDs are said to be designed to become "read only" at some point. Not sure how true that is.

Too many variables and too much randomness to evaluate. I don't know how you could have much faith in any conclusions about your specific drive as long as it appears to be working OK.

You can certainly replace a 250 or even 500 gb SSD for well under $50. Maybe you'd rather spend that much to avoid anxiety? I have no idea.

As said above, the TBW is largely a warranty-related value. If you exceed TBW in the warranty period, Intel might deny your warranty claim. You are well out of warranty.

As far as I know, the details of the algorithm used to generate that 75% figure in the Intel toolbox is a mystery. I assume the corresponding WD or Seagate tools might generate some other number.

 

[alexb75]

The so-called "bathtub curve" is said to apply to electronic components generally, but I've never seen a proper study of the extent to which it applies to SSDs.

I bought an Intel SSD 320 series, 80 GB in early 2011. I don't know the TBW on it...most likely under 30 somewhere?

It failed completely at about 3.8 years of age for reasons totally unrelated to data writes. I suspect I had written no more than 15 TB to it at the time.

Some SSDs are said to be designed to become "read only" at some point. Not sure how true that is.

Too many variables and too much randomness to evaluate. I don't know how you could have much faith in any conclusions about your specific drive as long as it appears to be working OK.

You can certainly replace a 250 or even 500 gb SSD for well under $50. Maybe you'd rather spend that much to avoid anxiety? I have no idea.

As said above, the TBW is largely a warranty-related value. If you exceed TBW in the warranty period, Intel might deny your warranty claim. You are well out of warranty.

As far as I know, the details of the algorithm used to generate that 75% figure in the Intel toolbox is a mystery. I assume the corresponding WD or Seagate tools might generate some other number.

Thanks. Well, that’s why I bought the Intel 730. Wasn’t the fastest at the time, but was deemed to have good reliability and being more industrial than commercial (per Anandtech), and it’s MLC which also should help.

I guess it’s cheap enough just to change it up and I might as well do.

 

[Lafong]

Thanks. Well, that’s why I bought the Intel 730. Wasn’t the fastest at the time, but was deemed to have good reliability and being more industrial than commercial (per Anandtech), and it’s MLC which also should help.

I guess it’s cheap enough just to change it up and I might as well do.

Those were my thoughts at the time I bought Intel as well.

There's more to "reliability" than the NAND itself or MLC or TLC or SLC.

There's circuit boards and controllers and several distinct components.

SSDs are still "assembled", which introduces points of failure.

I was working inside my PC and had removed the Intel 320. I was holding it in my right hand. I looked down as I was about to re-attach it in the mount and noticed a shiny object on the carpet.

The copper connector on the SSD that mates with the cable.

Hello New Egg, send me a Crucial MX 100.

Random chance. Did not deter me from later buying an Intel M.2.

 

[BFG-9000]

Intel was the only major SSD manufacturer that programmed their firmware to brick their drives when predicted SSD life dropped to 0%. And they did it in a quite stupid way too--I could understand if they made it go read-only permanently, but no--it went read-only until there was a power-cycle, then they didn't let you ever read the SSD again either so you would just lose the data.

Technically, a TBW rating is only the number of writes after which data is supposed to still be retained for at least a year when unpowered, and stored at elevated temperatures (the data retention for a new drive with one-write only is estimated to be over 10 years under such conditions, so it degrades to 1 year after a TBW number of writes). If you leave the drive powered-on, then the firmware should be smart enough to refresh your data often enough for this to never be a problem even well after many TBW are written. And if you think about it there are many uses for a SSD with only short-term memory like a scratch disk, so Intel's bricking-is-better strategy was really inexplicable.

 

[fzabkar]

I expect that Intel is calculating the remaining life of the SSD based on the number of P/E cycles. There should be a SMART attribute that records the wear levelling counts (average, min, max). It could be that the average value is now sitting at 75% of the rated number of cycles.

https://www.intel.com/content/dam/w...roduct-specifications/ssd-730-series-spec.pdf

Media Wearout Indicator
This attribute reports the number of cycles the NAND media has undergone. The normalized value declines linearly from 100 to 1 as the average erase cycle count increases from 0 to the maximum rated cycles. Once the normalized value reaches 1, the number will not decrease, although it is likely that significant additional wear can be put on the device.