Backblaze has released the first SSD edition of its regular drive statistics report, which appears to show that flash drives are as reliable as spinning disks, albeit with surprising failure rates for some models.
The cloud storage and backup provider releases quarterly and annual Drive Stat reports, which focused exclusively on spinning hard drives until last year. Backblaze said it would initially release the SSD edition twice a year, but that could change depending on the value of the drives. The 2021 Drive Stats report was released in February.
In a blog post detailing the latest investigations, Backblaze cloud storage evangelist Andy Klein said that SSDs are all used as boot drives in the company’s storage servers, and that Backblaze doesn’t started using SSDs in this way only from the fourth quarter of 2018. He pointed out that disks do more than just start servers, they also store log files and temporary files produced by servers, and so each SSD will read, write and delete files based on server activity during the day.
At the end of 2021, Backblaze had a total of 2,200 SSDs in its storage servers, but these had been built from scratch as of Q4 2018, and therefore some have been in use longer than others, and some models are also represented by a small number of examples which may skew the results.
As examples, Klein chose the Crucial CT250MX500SSD1 and Seagate ZA2000CM10002, which were found to have high annual failure rates of 43.22% and 28.81% respectively.
However, there were only 20 Crucial drives, all installed in December 2021, and only four Seagate drives, one of which failed in early 2021. In both cases, the annual failure rate calculation is based on very little data. Backblaze generally expects to see a 1-2% annual failure rate, and “anything less is great and anything more is worth watching,” Klein said.
Looking at annual SSD failure rates for 2021 alone, Klein pointed out that there is an extremely wide confidence interval for each drive. He said Backblaze considers a respectable confidence interval to be less than 1%, with 0.6% or less being optimal. Only one SSD model meets the 1% threshold, namely the Seagate ZA250CM10002, although another Seagate model (ZA250CM10003) comes close.
Looking at the cumulative annualized failure rates for all SSDs since the start of 2019, Backblaze finds that the curve on the resulting graph is well below the 2% target, while quarterly failure rates have shown spikes. at or near 3% for some. quarters.
However, the company cautioned that the total number of its SSDs in use represents a relatively small number of drives on which to start basing conclusions.
“That said, the more data the better, and as SSDs age we’ll want to be even more vigilant to see how long they last. We have a lot of data on this for HDDs. , but we’re still learning about SDDs,” Klein said.
Looking at three of the older SSD models that Backblaze has in service, the company found that all three had cumulative annualized failure rates below 1% for 2021, which compares to the cumulative AFR for all SSDs. of 1.07%.
However, while one type of SSD (Seagate ZA250CM10002) appears to follow a classic “bathtub” failure curve – experiencing early failures before settling to an AFR below 1% – the other two drives apparently did not. showed no signs of early failure and have only recently begun to fail, a mode Backblaze said is similar to what it’s seen with working hard drives that no longer fit the curve pattern of the tub.
For comparison, the cumulative annualized failure rate for Backblaze’s large number of hard drives installed was 1.4%, as reported in its Drive Stats 2021 report. Klein was keen to point out that the two groups of drives cannot not be considered to be at the same stage in their life cycle, and the company will continue to build its comparison of SSDs and HDDs over time as it collects more data on both groups.
Player data is made available by Backblaze for anyone to download and analyze on their own. The data contains figures for two additional drive models, Samsung 850 EVO 1TB and HP SSD S700 250GB, which are not included in the report as they were used by Backblaze for testing purposes. ®
Comments are closed.