Re: SSD TBW

On Thu, 20 Aug 2026 16:55:09 +0100, mm0fmf wrote:

> On 20/08/2026 05:42, Lawrence D’Oliveiro wrote:
>>
>> So why not divide the TBW by the actual capacity of the drive? Then
>> you end up with a ratio of how much can be written in total, to the
>> drive capacity -- call it, say, the “cumulative write ratio”.
>>
>> That number, it seems to me, correlates better to the overall quality
>> of the unit than TBW does.
>
> When I compare my NVME drive endurance for a 2TB device, it is quoted as
> being 1200TBW and the 1TB version is quoted as 600TBW.

There you go -- both drives have a cumulative write ratio of 600.

> I was a little concerned whether 1200TBW would be enough considering
> how much "chatter" there was on the drive. After 1 year of use, the
> drive has seen 4TBW meaning it will last me 299 more years. I think
> that will be enough.

Trouble is, we don’t know what the shape of the failure curve is like.
To take an extreme case, if the shape is skewed enough so that 99% of
drives fail with a write ratio within, say, 10 (i.e. your drive fails
within 5 years), while 1% last as long as 60000 rewrites, the
arithmetic mean would still stay the same, but you would be rather
more likely to suffer data loss ...