Who makes the best hard drives?

Who makes the best hard drives?

Summary: I'm just hoping to avoid the worstVendors and large users won't tell us who makes the best hard drives. So I decided to figure it for myself.

TOPICS: Hardware

I'm just hoping to avoid the worst Vendors and large users won't tell us who makes the best hard drives. So I decided to figure it for myself. Surprising findings: Over all quality is up. Western Digital is doing much better. Seagate/Maxtor has work to do.

Disk drives are marvels of engineering and precision volume manufacturing for which the industry gets far too little credit. But when a drive dies without warning - as 2 of mine did last week - I get grumpy.

Beyond the 2 failed drives I also upgraded the StorageMojo mainframe from a notebook to a tower with 4 SATA bays. I'm thinking 10k Raptor for system disk.

Scientific method? The method is simplicity itself. Total up the number of hits that "[vendor name] sucks" gets on Google. For companies who aren't exclusively disk vendors search on "[vendor name] drives suck."

Please note this is only a proxy for real numbers. One brand may attract the clueless and another the easily enraged. Some vendors sell more to OEMs who don't complain online. One thing I did figure out is that we can look at all-time numbers and numbers for the just the last year.

A final point about the numbers: rather than a direct measure of drive quality, this method is more an index of customer (dis)satisfaction. Someone's drives may be great, but if the warranty process turns people into keyboard-pounding Reavers, that's a bug.

Why "suck"? I needed a standard term of derision. There is no strong pejorative in the 3000 most popular American words. So I went with my default term.

For the drive companies I added both the "drives" and the "vendor name" totals. For diversified companies like Hitachi I only used the "drives" total.

Market share weighted suck factors Then I used the Q1, 2006 market vendor market shares - by total shipments - as calculated by iSuppli and compiled by DigiTimes online (no link due to DigiTimes subscription requirements.)

Dividing the Raw Suck by market Share gives a Weighted Suck factor. The lower the number the better.

So here's the score: lower is better all-time-drive-complaints3.png

Maxtor bites it!

This last 12 months After I did all that, I realized that the method found all complaints over all time. What about lately? I'm not buying a drive 5 years ago, I'm buying one now.

So I ran the searches limiting the results to the last year. The good news is that the vendors have improved customer satisfaction. i-year-online-drive-complaints.png

Western Digital snags the "most improved" award. They equaled Toshiba after being second worst to Maxtor over all.

The still-strong anti-Maxtor number may be people repeating old news. Maxtor is now the Seagate consumer brand selling Seagate-designed drives exclusively, so there shouldn't be a big difference between the two. Maybe Maxtor was the drive company people loved to hate.

The Storage Bits take What is surprising is that once you get past Maxtor, all the other drive companies are doing a pretty good job. And Maxtor is no longer making Maxtor drives: they're now Seagate and thank goodness.

Given all the other variables in drive operation, there isn't a single vendor I wouldn't buy from based on these numbers. Instead, I'll be looking at warranty - Seagate shines here - price and, for external drives, features.

Comments welcome, as always. Other ideas for a good proxy? BTW, I learned of the search idea from Vaughn Aubuchon and added a few wrinkles.

Update: After my morning coffee I reconsidered my late-last-night conclusion. The data didn't support my comments as well as I like, so I changed 'em (the comments, not the data!) so they did.

Update II: Savvy reader Combrink pointed out that I'd left out IBM's Deskstar (aka "Deathstar") and Travelstar drives. Good point! I went back and got the numbers for them and added them to the first table. There was only 1 IBM, Travelstar or "Deskstar drives suck" in the last year so I didn't alter the second table. Sounds like people are happy with their Hitachi drives.

Topic: Hardware

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • I cannot think of a time where a blogger ....

    ... at zdnet has spread more bunk about an industry then you have with your vendetta against disk drive manufacturers. For the most part there are very few issues with disk drive reliability. Even your own dubious statistics bear this out. Try dividing the the number of disk drives sold by the number of complaints you found and see for your self. I would expect better from someone claiming to be a storage expert.
    • Some Flaws

      I have read several articles about Hard Drive test that claim to say one drive is better then another. Where this test is that I have problems is with Hitachi drives and the manner he came up with the rated suck factor.

      I dont know about you but I have had more issues with Hitachi drives know as the Desk Star series. Most IT Professionals I know refer to them as a Death Star drive. Though they may have become better but in the last 5 years they where sued for quality problems. I can not understand how this slipped by and got such a low suck factor for 5 years ago.

      I agree with how you say they should have tested these drives. I have seen Western Digitals fail expecially last year with 320 gig sata drives first hit. They should have gotten stats from the company about RMAs and the Number of Drives sold. This would have provided a percentage of drive faliure and wouldnt be biased based off a web search.

      All drive manf. have good drives and bad drives that come off the line. That is okay because the average life of a hard drive still is longer then most people keep their machine nowadays. Remember it is the only constantly spinning and working device in the computer that is mechanical. It is going to fail at some point.
      • Good point about IBM

        So I updated the post and the table to add IBM, Travelstar and Deskstar. IBM wasn't
        as good as Hitachi, and yet looking at the last year Hitachi has done a good job.

        I'd love it if companies released more statistics. They don't and they don't want to
        and nobody can make them do it. So I wouldn't hold my breath. Why the array
        vendors or independent PC companies don't is beyond me.

        Also agree that good and bad come off the line. That's why for the low-volume
        buyer, drive life is a matter of luck.

        R Harris
    • Sounds a little defensive, ShadeTree

      I like disk drive vendors. I think they are the unsung heroes of the storage industry,
      and I've said so numerous times over the last several years.

      I think it is a shame that array vendors have been padding their margins on the
      backs of disk vendors for over a decade. I've also cautioned against irrational
      exuberance over flash drives, which some of the less-informed have been touting
      as disk replacements.

      However, independent research has shown that there is a big difference between
      what the industry sells and what customers actually see. Specifically, the CMU and
      Google research found significant differences between the failure rates spec'd by
      vendors and seen by users of some 200,000 drives in the field.

      Would <i>you</i> care to explain why? Seems to me there should be greater
      industry transparency around drive lifetimes, operating conditions, failure rates and

      If asking for that makes me a bad guy, then so be it. I think the thousands of
      people who angrily commented on the web - and the many tens of thousands who
      thought the same and didn't - would like to know more.

      R Harris
    • Load of nonsense

      Hard drives have become less and less reliable as new models are introduced. There are some reasonable causes for this such as high density, more platters, faster spin rates, etc., but the fact is that they are not as reliable as they used to be and consumers don't seem to know this.

      I've got 10 year old SCSI drives with a 1,000,000 hour MTBF. Yeah, a million. I am actually researching a bit for this response, I see that some Samsung Spinpoint drives have a MTBF of 330,000 hours. Still sounds not bad but don't believe it.

      I've found the best way to predict the reliability of the drive is by the warranty. If a drive has a 1-year warranty, consider it temporary storage and unreliable. A 5-year warranty points to the manufacturer's greater confidence that it isn't a POS.

      I don't think the author's statistics necessarily bear anything out because I think relatively few people bother to post that this or that hard drive "sucks." The results, therefore, are likely to be skewed by whoever was most pissed off. The way a drive fails might get owners of certain drives more angry than others, while the drive sucks no more or less. I've had Samsung drives that simply gave lots of write errors. The damned thing appeared to be working fine, but in reality you couldn't get the data back without errors. That REALLY pissed me off, even more than the Maxtors that went clicking off to hard drive hell because at least that was obvious.

      I agree with you that the problem should be approached with a more rigorous methodology, but I disagree with your apparent belief belief that the drives are reliable.
      • Several problems with what you say.

        There are less platters in the new drives then the old ones. In many cases only one. There is no coorelation between the number of platters and drive failures. The higher density has been achieved by flying the heads closer to the platter and by turning the magnetic particles on end. I am not aware of any study showing a coorelation betwen these two factors and reduced reliability. There is also no coorelation between spin rates and drive failure.

        You start of your post talking about a SCSI drive. It is unfair to compare and Enterprise class drive to a Consumer class drive which is what you are doing by introdsucing the SCSI drive. Warranty has nothing to do with drive life and has more to do with marketing then anything else. It is not an accurate predicter of drive reliability. Finally your conclusion that drive reliability has went down is simply wrong! Drive reliability has steadily improved.
        • Again, don't be ridiculous...

          Stop trying to argue with people's own experience of drive failure rates. We know how many drives we have had fail. Those of us in the industry that have had to buy the drives, install them and then replace them when they go bad know which drives suck. The author clearly stated that drives were getting better but that some drives suck worse than others. He states his conclusions clearly and I for one believe he has it right. I know from my own experience in replacing hard drives. Oh, and I also disagree that warranty has nothing to do with drive life. The difference between a 1 year and 6 year warranty is quite OBVIOUSLY an indication of predicted drive life. Don't be stupid.
      • Warrenties

        I will not sayone way or the other, I am not a bean counter with a HD manufactor, but I can tell you that in other industries, the higher price you pay for a product with a longer warrenty is oft times the additional costs to replace the failures, the units come off the sam eline, the ones labled X go in one box with one warrenty, he ones labled Y go into the othe rbox with the other.
        The authors statistics are accurate, just not properly performed. The number of people who write is a statistic of the total number of people that have views of thatever you are looking at. that is why calling your congressman gets more action than sending an email. If you are upset enough to call you may be a reprentative of another 5000 constituants, a letter may be 3000, and an email 1000 (ficticious examples).
        The polling people actually have better idea of what the proper ratios are, its what they recieve money for.
        Last, as much as we might say we want all the data for MTBF, infant mortality , expected life span, unless its regulated or is falling under a gov sepc for sale, you will probably not ever see it. Why would an independent vendor open their books to competition if they can help it? The number sold is information available for stockholders, the defects are buried in operating expenses. bottom line is whats truely public in a publicly held company.
      • Who makes the best hard drives?

        Not everyone complains or blogs about such and such HD sucks! It's rather childish to begin with. I have only had one Maxtor drive fail on me. Maxtor replaced the drive free of charge with a new, not refurbished drive that was actually larger than the one I bought. (Including shipping I paid nothing) Out of the nine computers running Maxtor hey only one failed. On the other had 22 computer have WD and 12 of them failed. Not a good percentage. But to WD's credit all of the drives were replaced and none have failed since. Drives fail get over it!
    • You call 5 failures out of 10 drives "bunk"?

      Over the past 5 years I have purchased about 10 hard drives for my personal use and 5 of them have died before the warranty expired. The only other place I know of where a 50% failure rate is considered acceptable is weather forcasting. Incidentally, all the drives were Maxtors (I finally admitted defeat and switched to Seagate late last year) and no, I haven't complained online so my results aren't figured into the tables in the article.
      • Yeah Maxtor drives truly do suck

        We had 10 maxtors out of 10 fail. And seagate is now seagate/maxtor. I believe they have their act together now as we haven't had a single new seagate/maxtor drive fail. However, we haven't purchased as many seagates as we did WD. My own experience of WD is that THEY are the most reliable, though. We have yet to have a single Western Digital drive ever fail. I myself have purchased over 100 of them in the last 5 years and not a single drive has had a problem or failed. That may be luck but I don't think so.
        • i guess i'm lucky i have yet to have a maxtor drive fail

          i guess i'm lucky i have yet to have a maxtor drive fail in this box i'm writing this on has 3 maxtor sata 2 drives in it.

          i've all ways had good luck with them.
          SO.CAL Guy
        • Correction: Maxtor IDE Drives Suck!!

          Hold on a minute mate thats way to sweeping a statement to make.

          I boot off 4xRAID-0 U320 15K rpm SCSI Maxtor Atlas II SCSI in my Opteron Server. I've blown many a SCSI drive up with accidental rewiring of PSUs (OK I got 5V and 12V mixed up :-) and there was no hassle with the Maxtor 5 year warranty (no doubt they could put a new board on the drive anyways bless them... :-) )... As I recall Seagate let me stack up my RMAs but Maxtor was one return at a time (presume that policy still holds today even with them both being under one umbrella)... But anyway thats Enterprise drives right!! All killed off by my own incompetence...

          When it comes to IDE drives (PATA/SATA) I have a couple of PATA drives from a Lacie IEEE1394 external drive that run hot as hell and died horribly sometime ago... Only drives I didn't kill myself so far (touch an earth line, etc.)!!

          So in summary Maxtor Enterprise drives are pretty good!!

          Bob Wya
          Bob Wya
    • another 2 failed Segates

      I know another 2 Segate HDD's that failed just after warrenty period. so I think author of this artcle has a point
    • Don't be ridiculous

      This entry is completely valid. And your opinion of drive reliability is contrary to the numbers. I myself experienced a 100% failure rate on a batch of 10 maxtor drives that our department ordered 2 years ago. All 10 drives failed within 8 months. So get your own facts straight before slamming someone else.
      • Nothing ridiculous about it.

        As an engineer for one of the top three PC OEM's I can assure you my experience with disk drives involves a lot larger sample size then yours. I stand by my comments.
      • ridiculous

        Like anything else perhaps you just got a bad batch! Hey it happens!
  • Hitachi (IBM) TravelStar

    I have had good results with Hitachi drives.
    FYI, I found an interesting study produced by Google on 'Disk Failures'.

    Download and read the pdf [url=http://labs.google.com/papers/disk_failures.pdf]here[/url]

    Apparently SMART isn't necessarily so smart in sensing errors which lead to failure.

    Thanks Robin for a good article!
    D T Schmitz
    • I agree _dietrich

      I use Travelstars in my notebooks, and Deathstars in my desktops.

      IBM / Hitachi drives are superb, quiet, fast and not too expensive. I'd recommend them to anyone.
      • Finally...

        ...someone agrees with me!
        Thanks Scrat!
        D T Schmitz