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


View All Comments

  • kmmatney - Wednesday, March 3, 2010 - link

    I've heard that CCleaner's "wipe" function can work as a manual TRIM, and restore performance to OS's that don't support TRIM. I will test this firsthand soon, as I got in on the Intel 80GB G1 deal for $149, which doesn't have TRIM.
  • geddarkstorm - Wednesday, March 3, 2010 - link

    This drive is a piece of junk. Seriously. It's grade A crap. Anand is being way too kind to it. It gets trounced by every other SSD sans the super cheap, nerfed Intel 40GB one, and yet asks us to pay with an arm and a leg? Even if this thing is at half it's price, it'd still be a horrible mistake of a buy.

    And I see people going on about how great it is because it's Western Digital?

    It seems JMicron still can't make a good controller to save its life (it's all that DDR2 memory keeping it afloat!). Though the good news is it's getting better. I guess that's akin to telling a quadriplegic they are improving at wiggling their eyebrows.
  • taltamir - Wednesday, March 3, 2010 - link

    I agree... this drive is utter crap.
    it is the most expensive by far, and it is the lowest performing across the board with the exception of crippled drive like gen 1 jmicron and crippled 40GB intel V.
    they tout "reliability" and "compatibility", but that is just BS. the established drives have already solved all their compatibility and reliability issues. If you buy a current indilinx drive it isn't gonna have any more problems, they were fixed already. This drive however, was never tested en masse by the public.

    1. Most expensive (2x price)
    2. Slowest
    3. Least reliable.

    conclusion: utter crap.
  • The0ne - Monday, March 8, 2010 - link

    I honestly don't know what Anand has been smoking lately but there's quite more of us upset at these types of seemingly bias reviews. Maybe him and his team needs to learn how to properly test and write reports again. Shrug. Disgraceful and shameful imo.
  • v12v12 - Friday, March 5, 2010 - link

    I'm gonna SAY IT!

    I'm actually not glad that they are FINALLY trying to jump on the bandwagon. I think WD makes the best overall consumer drives and I've run them exclusively (on/off) for years... But here's where I differ.

    As many of you know, I've been ranting ad raving on here about metered-technology well guess what? WD finally proves my diatribes b/c they've ran out of ideas and cannot compete with their uber, overpriced/overrated drives, so what do they do? Have a board meeting with execs pulling their hair about "WTF are we going to do with this SSD crap?! They're killing our once herald Raptors and our metered-technology scheme. What are we going to counter with?!"

    As much as I support WD, I'm glad these metered monoliths are facing the heat as they've been pushing junk releases to us for decades. I personally would like to see WD, Sea, Hit, Sam and the like to all fail and hard. All of the years of metered garbage put on for us to review and nitpick about unperceivable performance numbers... while getting charged 10-20-30% for performance advantages that are meaningless; clever marketing schemes and ploys to sell it. etc... The age of the mechanical rotating disk is coming to a close; I hope to see them go extinct!

    The last thing I want to see is another huge conglomerate of metered-chicanery gobble up all these newly inspired innovators, just to yank back the reigns of the wild stallions that blaze a path to future innovation. Just to end up subjugated and attached to the old wooden-wheeled cart, overloaded with slow, underperforming, metered-feed to dump in the toughs of the consumer CATTLE.

    I want these new innovators to stay FREE from large, slow and lumbering technology stalwarts like listed. These corps want to engulf and eventually assimilate these stallions into their cattle-carriage technologically metered assembly line... Just you watch, as soon as these little companies get swallowed up, all the constant and surging innovation will cease.

    You see, a corporation is nothing merely but a giant blob/amoeba; sucking and absorbing anything and everything it comes in contact with, into its slovenly and molasses like body. They'll buy up OCZ and the like and turn them into another pet project to mooch off of and soon another huge boardroom will be filled with greedy executives scheming and plotting on how to meter out even more crap upgrades. "Oh lets see if we can split these SLC to LCs (made up for effect) and come out with a new "green LC" line. We'll justify slow and underperforming drives as "acceptable" b/c they are saving a minuscule amount of energy you could by just turning a light bulb/TV off/unplugging power bricks etc.

    As much as I like WD, Hitch, Sea etc... these places are evil, overtly greedy and frankly a poor model of technological innovators (WD being the top of out them). They'll discover a new breakthrough in the lab, cover it up, then the tech and marketing board meet up and try to chop up said innovation into 3-4-5-6; or 10 diff product lines, with tailore-specific marketing to sell to the cattle. This type of business model HAS TO GO! *You guys need to stop supporting this model* if you want them to actually start innovating Vs pretending.

    Why are there enterprise drives ranging from 300-600GB, 15Krpm out? What about 10Krpm drives with the same capacity? What do we, the desktop cattle receive over the same years? Some (comparatively) slow 10Krpm, tiny 36,74.... then 150, then 300GB Raptors (Oh and the PRICE!)? WHY did the public not receive the next major spindle speed upgrade like enterprise users did? 5400 ---- 7200... STOP! Now come out with a "special" 10krpm, small capacity, uber expensive boot/OS drive, give it a catchy name (I wonder how many boardroom hours were billing conceiving this new ploy) and see if it sells... It sure did! What about the rest of the public? No 10Krpm upgrade for us, unless you wanna drop $100-200 more for 500% less capacity on a Raptor… HOGWASH! They over extended their metering plan and SSDs BLEW IT UP in their loser, lame faces. Jokes on you WD!

    Now all of a sudden before they could even contemplate releasing a standardized 10Krpm for the desktop consumer... SSDs jump the gun and completely rule out this metered strategy as MOOT! HAHA those poor raptors are near worthless now and will soon be dumped on ebay for pennies on the dollar b/c the GIG IS UP! The Raptor hold-out, con-game iS FINALLY BROKEN! Farwell rape-tor, no more price premium for YOU!

    This to me paints a clear portrait of the chicanery and metered non-sense that’s been covered up and going on for decades; SSD releases have deciphered this flaw in the marketing con-game… GOOD RIDDANCE! Soon, one day… there’ll BE no more lag, and access times (meaningful), and spinning garbage to suddenly fail on your w/o SMART warning etc… GOOD BY mechanical junk…. And to think, we all could have had this technology and watched it long surpass the 2TB mark of mech drives today, IF this metered ploy had not been allowed to be in place for so long. Enthusiasts have been hooking up CF adaptors and such for years before SSDs. They were the pioneers of said solid-state… What a wonderful break though era; again, I hope to high hell OCZ and the like, out compete the mech-marketers, completely take over the their market and keep WD, Sea, Hit, Sam and the rest of the con-artists OUT… Stick to spinning disks and actuator science LAMES!

    PS- And of course I HAVE to call out Anand (no offense, just business) staff for the constant wrist slapping “conclusions” that we’re all so used to… Excuses, apologies and semantic games, used to run around the facts of this obviously evident shill-game? YOU guys work in the business, just like I do; I KNOW you’ve long known this crap… but to make a buck ya gotta be quiet about it or if ya want to keep getting free Golden-samples to review… better hush-hush on the shitty products we’re left to purchase from a sub-par, and inconclusive review. How many times has a review site built up a nice machine, only for use to mimic the same build and get nothing near the results you attain? GOLDEN-SAMPLES that’s why; the best that the factory can produce, given you to review and speculate on; thus we go out from all the built up hype, only to get avg to slightly above avg performance… Who’s to blame, nobody just a few writers/bloggers shrugged their shoulders as if they don’t know the game… Get outta here, many of us know about golden-samples and the like… If a product doesn’t finish a test or meet the requirements like it’s competition in the line up… DID NOT FINISH, DNF should be awarded and NO excuses offered, other than it didn’t work like the rest… Then again…*sigh* you run a business, and what’s good for business, isn’t so much for the consumer; Grow some spine guys… Seesh.

    Straight-talk-express comin’ through! CHOOOO CHOOO!

  • Iketh - Saturday, March 6, 2010 - link

    you lost me on the metered-technology... please inform
  • BelardA - Wednesday, March 3, 2010 - link

    Its rather shocking and yet understandable that the HDD companies of Western Digital, Seagate, Hitachi and Toshiba (anyone else?) ignored the coming of SSDs. Instead have little tiny companies coming out with products and learning the tech.

    I guess they can do the wait-and see and let those smaller companies suffer the birthing pains of a new tech... then buy them out or reverse engineer the tech.

    So, WD and Seagate both buy out a tech, spend many months learning how it works and then come out with product that is pretty much based on what is out on the market now... how many millions were spent?

    I'm not impressed with WD's performance or its price... this blue drive is constantly on the bottom end of things. Yes there are problematic drives - mostly in the past. And as shown with a little bit of research on the consumer end, they know what to avoid... like certain OCZ and Kingston models, Corsair. That same bit of research will show that the current blue-SSD drives are not worth buying.

    In the end with my limited SSD experience, I have gone with intel with their X25-G2 drives. I'm not an intel fan (their business practices), nor an AMD fan-boy as I own computers with both types of CPUs. But I bought their drives because of the tools, support and overall-performance of their drives. Yes, they are NOT the fastest seq. R/W drives on the market... but they are among the fastest Random R/W and have a proven track record (not perfect). When you pickup an Intel or re-badged G2 drive, you know exactly what you're getting. I've bought intel and A-Data's G2 (intel) drives, same firmware... and A-Data was $70 cheaper than intel's sticker.

    With OCZ, theres confusion of what drive is what. 4~6 different versions of a 64 or 128GB drives. etailer A has some, but maybe not the others... some drives support trim, some don't. I'm glad OCZ is at least coming out with a lot of products, but there is too much gray area... new models (or discontinued) every 3~4 months it seems.

    The day will come when we, the consumer can go to a store (or website) and have a dozen good drives to choose from that are fast, reliable and low-cost. For desktop users, I think the hybrid setup works best for costs (SSD boot & progs, HDD for data).. but todays $100 32~40GB drives tend to be slow crippled versions of faster drives. About 5 years ago a 60GB 2.5" HDD was $200. Today, a 640GB 2.5" HDD goes for $100 ($50 for 250GB).

    So yeah, in less than 2 years... $100 should get you a 128~160GB SSD with SATA3/6GBs interface and able to pump at least 450MB/s into your computer and have an easy 5year 7-day usage life span.

    My experience with intel-X25s. Amazing... Windows7, fully loaded with apps and such... boots in about 10seconds (after POST). A fast HDD would boot Win7 in about 35~40 seconds. The SSD & computer are not SATA 3... :(

  • jwilliams4200 - Wednesday, March 3, 2010 - link


    You should add a couple spaces to the end of the names of the drives that you expect to have low performance (like the HDDs). Then in your graphs when the benchmark text collides with the drive name text, it will be spaces colliding with the numbers, and the numbers will be readable.
  • shangshang - Wednesday, March 3, 2010 - link

    I'm glad to see WD entering this market. People need to realize that eventually, SSD will become commoditized (but it won't ever going to be as cheap as mechanical drives!), and that the NAND chips (or whatever flash chips in them) will also be commoditized.

    Commodity is not important.

    What is/will be important is the interface and the optimization of the SSD drives. WD has the best team and the best experience when it comes to interface. Makers like OCZ, Micron, and even the mighty Intel,... DO NOT have extensive experience when it comes to interface and optimization.

    I predict that eventually, WD will become the dominant SSD maker for desktops, followed by Seagate. Intel will be there simply of its massive size. But smaller makes like OCZ, Supertalent... will go the way of the Dodo.
  • PandaBear - Wednesday, March 3, 2010 - link

    I disagree.

    The SSD team at Marvell came from the old Maxtor/Quantum hard drive team.

    The SanDisk firmware team came from the Maxtor/Quantum/Seagate hard drive team.

    The SandForce firmware VP came from SanDisk.

    The Micron team came from Lexar.

    There are plenty of veterans that can do SSD correctly, basically people who've done hard drive and some flash memory before. The biggest problem is the closeness to a FAB that build and design the NAND, so far WD doesn't have a fab, so they can't get their hands on the latest generation's nand chip problem.

    Plus they are behind by a few years.

Log in

Don't have an account? Sign up now