Scrambling Data to Improve Reliability?

I secure erase all SSDs before I run any test, unless I'm testing used performance. On a modern SSD, performing an ATA Secure Erase takes anywhere from 30 - 50 seconds in most cases. The first time I went to secure erase the SiliconEdge Blue it took almost 17 minutes. I asked Western Digital what was going on, and got an interesting explanation.

The smaller NAND flash gets, the more prone the cells are to developing bit errors. That's why SandForce made such a big deal out of its controllers making 34nm possible thanks to extra error correction and redundancy.


SandForce's controllers attempt to fix bit errors in order to enable the use of cheaper MLC NAND

Western Digital found that writing certain combinations of data were more prone to generating these bit errors on smaller dimension flash (< 50nm). To prevent these sorts of errors from cropping up, the controller reorganizes (or scrambles) the data it's fed before writing it to the flash. When you go back to read the data off the flash, it's unscrambled before sending it back to the host controller so you get the data you requested.

As a side effect, the controller can't simply perform a flash erase to zero out the contents of the drive. Instead the SiliconEdge Blue has to physically write (a F or 0) to every NAND flash cell on the drive itself, which takes some time. On the bright side, WD does this to improve data integrity (or more likely to use cheaper MLC NAND without sacrificing reliability). The only downside is that a secure erase, which you shouldn't need to do that often, will take a long time to complete.

The Controller Sequential Read/Write Speed
Comments Locked

64 Comments

View All Comments

  • Frallan - Friday, March 5, 2010 - link

    Performance isn't average its abyssimal. It can only compete with the other brands low-performance offers and is priced hicher then the High performance offers.

    I don't understand why WD turned up at the party at all - this deflates the good will that they still have.

    /F
  • capeconsultant - Wednesday, March 3, 2010 - link

    Bigfoot? Nope. Smallfoot :)
  • vol7ron - Wednesday, March 3, 2010 - link

    Anand,

    You're really pumping out the articles lately, you must be swamped!

    I just wanted to comment on your closing remarks: "It just strikes me as odd for hard drive manufacturers with decades of experience in firmware development and data access patterns, to not come out of the gates swinging."

    I'm sure you realize that the two technologies are almost completely different. Sure they have common components and similar storage logistics, but WD made it's way on the quality of the physical aspect and head speeds of hard disks.

    The comparison is almost the same as how floppy disks have been phased out by thumb drives, or film cameras being replaced with digital cameras. We didn't see those companies that excelled in the in the first generation of their technology come out swinging, did we? Verbatim and Kodak are still struggling.

    What's more surprising is not seeing a company with a big wallet to come out aiming for the fences. That's what I'd like to see, just like Intel.

    Cheers for another good article,
    vol7ron
  • 7Enigma - Thursday, March 4, 2010 - link

    Unfortunately you must have missed the FIRST sentence of the article. They purchased a large company that had a great deal of experience in the SSD arena.

    That kind of makes your comment moot.
  • The0ne - Monday, March 8, 2010 - link

    IF everything stayed the same than yes the point would be moot but sadly when acquisition happens changes are always in effect and mostly for the bad reasons. You can acquire the best SSD house in the world and bring it into your team with little experience in it and have them run it. Give it a few days and that "best" becomes WTF, guarantee.
  • vol7ron - Friday, March 5, 2010 - link

    Haha, you do go me there. :)

    Except, SiliconSystems is really a mock company and I use that term loosely. Even if you say they focused on CF, although it is more similar to SSD NAND, as a storage product it is still significantly different. I would say, though, that the combination of WDs current storage management, combined with some experience with the NAND controllers should have put out something better than this WD SSD.

    Still, WD acquired Silicon Systems in March of last year and (even then) I thought SS was an inexperienced company. They put out Enterprise/OEM products, that didn't really deal with speed, but more with reliability; more specifically, being able to recover from unforeseen power downs, data errors, etc. While that is very important, SSDs are supposed to provide speed and that's not something SS is known for.

    So, yes, I still stand by my statement. Still different technologies and a hard transition.

    vol7ron
  • Soltis - Wednesday, March 3, 2010 - link

    lol... these drives are on the lower end of the performance spectrum and the higher end of the price spectrum..

    But who knows? maybe with this new "reliability" WD drives will now survive the trip to your house! ~zing
  • CTT - Wednesday, March 3, 2010 - link

    A comprehensive/insightful article, as always. I would have a couple of remarks about the graphs though: the low values are printed over the drive names and the the WD drive is listed as JM618.

    I see the TRIM support is given the due attention, but there are quiet a few users that don't benefit from it (e.g. using Windows XP, TrueCrypt). Would it be too much trouble to ask for some tests with drive full/some percent free and TRIM disabled?

    Did the WD experienced any significant (read abysmal) drop in write performance after some use (such as HD Tune Pro benchmarks at Legit Reviews and StorageReview)?
  • Anand Lal Shimpi - Wednesday, March 3, 2010 - link

    I've been trying to figure out the best way to test performance in non-TRIM aware OSes. I'm playing around with some things and will eventually present my findings :)

    My drive didn't show any significant drop in write performance after use. That appears to be an issue with the HD Tune benchmark itself.

    Take care,
    Anand
  • Iketh - Saturday, March 6, 2010 - link

    Yes it's an issue with HD Tune. It sometimes reports my mechincal drives as reading 1/3 of what they normally do.

Log in

Don't have an account? Sign up now