Why Test SSD Write Endurance?

The SSD Endurance Test site is a result of my personal childish curiosity, creativity and the challenge of finding consumer level SSDs suitable for server use. The data collected in these tests might become a useful tool for predicting SSD failures and choosing replacement intervals.

An SSD is probably the single best upgrade a computer can get. I hope these endurance tests will show that we don't need to worry too much about our aging SSDs. The average user will most likely never run out of P/E cycles. Only power users that do a lot of I/O intensive tasks need to take the SSD's NAND endurance into account.

Writing to NAND-based memory causes microscopic wear on the oxide layer within the circuit. The number of times a NAND based memory can be written to are called Program/Erase cycles or P/E cycles.

Manufacturers are constantly trying to shrink the NAND circuit die to cut production costs. The first consumer level SSDs used 34nm dies. Some recent SSDs use 19nm dies. From an endurance perspective that would be a drop from 5000 p/e cycles down to 1500 p/e cycles.
True or false? Let's find out!

- Do you think most SSDs will fail gracefully (read only) or just stop working?
- How much further than the specified TBW ratings can we push the drives?
- Will the S.M.A.R.T data prove reliable?
- Should we perform a data retention test at some point?

I'd appreciate your thoughts and feedback on this experiment.

Christian Löfstedt
Google+


Comments and Feedback

Is this really normal?

No. The tests preformed here must not be confused with normal use. They are designed to stress SSDs and simulate a really busy environment.
Normal workstation use would be more like 10-20GB written daily.

Why Test Smaller SSDs?

They use fewer NAND circuits and will wear out faster. Once we know the write endurance of these units, it'll be easy to estimate the equivalent numbers for the larger units.

Testing SSD Endurance in RAID Arrays »