Sunday, February 28th 2016

High-end SLC SSDs No More Reliable than MLC SSDs: Google Study

A FAST '16 paper titled "Flash Reliability in Production: The Expected and the Unexpected," by Professor Bianca Schroeder of the University of Toronto, and Raghav Lagisetty and Arif Merchant of Google, studied the reliability data from millions of SSD drive-days over a period of 6 years, to come up with some very interesting conclusions on SSD reliability. One of the study's biggest findings is that high-end (read: enterprise) SSDs with single-level cell (SLC) NAND flash memory are no more reliable than cheaper multi-level cell (MLC) drives. Besides millions of drive-hours, the group also studied 10 different models of enterprise and consumer SSDs, from three different memory types - MLC NAND, SLC NAND, and eMLC NAND.

The study also shows that RBER (raw bit error rate) is a more dependable measure of reliability than UBER (uncorrectable bit error rate) mentioned in drive specs or datasheets. RBER increases slower than expected from wearout, and isn't correlated with UBER. However, the measured/real-world UBER is higher for SSDs than HDDs. This means that while SSDs are less likely to "fail" than HDDs, they're more likely to lose portions of their data. Keep your SSDs regularly imaged. Age, rather than usage, affects reliability of SSDs. A disturbing 30-80% of SSDs in the study developed at least one bad-block, and 2-7% of the SSDs developed at least one bad chip, within the first 4 years of deployment.
Source: ZDNet
Add your own comment

30 Comments on High-end SLC SSDs No More Reliable than MLC SSDs: Google Study

#26
yogurt_21
for consumers that makes sense. The trouble with enterprise level is that single drive failure isn't really concerning. As long as the average drive lasts 5 years it doesn't really matter if even 10% die within the first 6 months. You throw those out or warranty them and pop in a replacement drive into the array. No downtime or data loss. Plus as fast as raid 10 rebuilds on a ssd, you will barely have any time with the array in a degraded state. Advanced SAN's are also protected against that by simply ignoring the degraded array until it comes back to full speed.

The main gist of what I'm seeing in this study and others is that our propensity for reliability may not be in line with how cloud services are moving. As long as it has "good enough" reliability the real focus should be on speed. After 3-5 years that SAN will be retired and something faster will come online. They really won't care how much longer the drives last after retirement. That's the worry of the person buying the old used tech, not amazon or google.
Posted on Reply
#27
RejZoR
That's the thing. Corporations buy them en mass. They are expendable for them. Not for consumers who run single drives and demand durability and reliability.
Posted on Reply
#28
Parn
Guess I've made the right choice to stick to HDDs for my home file server. SSDs will be strictly for OS/Program on my desktops/laptops.
Posted on Reply
#29
xorbe
Reliable as in data retention, or reliable as in also reliable performance?
Posted on Reply
#30
xtremesv
30-80%? Don't you think that's a rather large margin to come up with a conclusion. Well anyhow, if that study is valid then the part of SSD age being more critical than use worries me.
Posted on Reply
Add your own comment
Jul 18th, 2024 03:28 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts