Once-great SSD manufacturer OCZ filing for bankruptcy
arstechnica.com
arstechnica.com
They were never great.
They lied to their customers by selling hardware under the same name as previously produced hardware with cheaper components and lesser specs.
They built hardware that was simply off-spec, an example being drives where the connectors were an entire millimeter shifted, such that when installed in certain machines the connectors literally could not make contact with the corresponding metal.
They built drives with extreme speeds while entirely sacrificing longevity and reliability.
At best they had a great marketing department that made it possible for them to peddle their crap to the public for so long.
I'm glad to see them go.
Edit:
For those who must have numbers, return statistics:
http://translate.googleusercontent.com/translate_c?depth=2&h...
Most of the reliability problems I've seen have been self-reported, so I'm not sure what biases are present in what I've seen. ;)
(I agree that many of their practices were bad, which likely lent to bad drives! This isn't trolling, just genuine curiosity if there were published stats)
http://www.behardware.com/articles/862-6/components-returns-...
- Crucial 0.82% (as against 0.8%)
- Intel 1.73% (as against 0.1%)
- Corsair 2.93% (as against 2.9%)
- OCZ 7.03% (as against 4.2%)
Crucial has taken top spot from Intel thanks to a notable increase in Intel’s returns rate. We should say that this time, the Intel sample is only just above the minimum required and that some of the Intel returns are linked to the 8MB bug which has since been resolved. The OCZ rate has got a lot worse, going up to 7%, and only OCZ has models with rates of above 5%:
- 15.58% OCZ Vertex 2 Series SSD 240 GB
- 13.28% OCZ Vertex 2 Series SSD 160 GB
- 11.76% OCZ Vertex 2 Series SSD 80 GB
- 9.52% OCZ Vertex 2 Series SSD 120 GB
- 8.57% OCZ Vertex 3 Series 120 GB
- 7.49% OCZ Vertex 2 Series SSD 60 GB
- 6.61% OCZ Vertex 2 Series 3.5" SSD 120 GB
- 6.37% OCZ Vertex 3 Series 240 GB
- 6.37% OCZ Agility 3 60 GB
- 5.89% OCZ Vertex 2 Series SSD 100 GB
The Vertex 2s have the worst scores but the Vertex 3s have nothing to be proud of either. Note that over the coming period, the Vertex 3s are doing much better thanks to developments in the firmware, with a rate of just 1.01% for the Vertex 3 120 GB as things stand.
"The first question is of course where the stats come from. They’re taken from a large French etailer, whose database we have had direct access to. We were therefore able to extract the stats we wanted directly from source."
(Just a note to keep in mind when seeing all these personal testimonies)
FTFY.
I read this about a year ago and have been unable to find the source. Does anyone know where this is from?
e.g.
http://askville.amazon.com/original-source-quote-plural-anec...
http://freakonomics.com/2010/04/29/quotes-uncovered-whats-th...
ssds are better than spinning media, so what? even the worst one would make you happy. this is not the point. they constantly send one device to reviewers and then put lesser specs on the shelfs, or less immorally, just updated the revision silently to cut costs. i have a couple of them. i regret buying because what i got was something equivalent to something i could have gotten cheaper. also, admins often mocked people on the support forums.
However, the problem is that the retail firmware on the drive shouldn't have had the kinds of problems that would require a firmware fix.
Their original Vertex series had its rough edges but was a real breakthrough - they were the first to really push a "modern" SSD controller in the form of the Indilinx Barefoot. At the time, most other SSD drives on the market used horrible JMicron flash controllers that were more suitable for your camera's memory card; they were not really designed to cope with the traffic a desktop operating system would throw at them.
Back in the Vertex/Vertex2 days they really were the ones to beat for the most part. But then Intel came along with their incredibly stable SSDs and frankly, OCZ's drives started to seem like toys.
Part of OCZ's problem is that, starting with the Vertex 2 era, dozens of other companies used Sandforce controllers just like OCZ did. At that point OCZ obviously tried to compete on price and we can see what it did to their quality. To their credit, they made a really commendable last stab at relevance when they bought Indilinx so they could have their own custom controller and (maybe?) stop competing on price. Those drives got some great reviews although I don't know if they were reliable or what.
You don't say. One example being the OCZ RevoDrives.
Our company investigated replacing our traditional HDDs with SSDs and with RevoDrives to see how that impacted our build-times and development efficiency.
Almost every second user who received a RevoDrive experienced critical data-failures and 100% data-loss.
What they didn't say on the package WITH CAPITAL LETTERS was that it was a RAID0/JBOD of SSDs with zero redundancy. If one drive failed (and they did) you lost all your data with no means of recovery.
We could have asked for replacement units because they were still under warranty (some as new as 2 months), but needless to say, we weren't very eager on more drives from OCZ.
If the rest of their product line-up was like this, I can't say I'm too surprised about them not making it for the long run.
This really shouldn't matter to the consumer - there are lots of ways to couple components in a black box to increase the rate of failure. They don't disclose to you how the controller is built, or how they go about marking bad blocks, so why would they say in bold letters that under the covers they're striping across two disks?
I was hoping someone would bring this up. When shopping around for SSD drives a year or so ago, I stumbled upon stories about this and simply couldn't believe that a company would do something like that.
Basically, you'd read some reviews (with benchmarks) about OCZ drives. Then you would go and purchase the exact specific type that the reviewer had — and the hardware inside would be different (inferior). There was no way to check this without opening the drive up or measuring with benchmarks.
This was the moment I decided I would never buy anything from OCZ, as I can't do business with a company that lies to me. I am glad they are going away.
Surprised it's not Crucial going under after that experience.
They also chose the most susceptible target market - teen gamers and overclockers. "Being cool" is their mantra and OCZ delivered.
"My OCZ drive failed..." / "I'm on my fourth RMA, what should I do?"
"Real junkies buy Intel." (Or Samsung). And they don't buy TLC flash, either.
Nice to know they agree with the rest of us that Samsung is the way to go nowadays, too.
This was one situation where you actually got what you paid for. Yeah, you do pay more for Intel and Samsung -- but that pays for the QA engineers. If you bought OCZ, then you got to be your own QA engineer.
A year later and none of the replacement Intels have gone bad...
Go to a few review sites, have the reviews say "OCZ looks fast and good", go away with a good impression. You have to look elsewhere - on forums and the like, all of which are highly variant in quality - to get information about long-term performance. It's compounded when you have a new player or new product lines. Here in this very thread, it has several people saying "OCZ is great".
It's in the finer gradations that it gets murky. Is a 4.5-star drive really better than a 4-star drive?
With regards to other brands, I spent some time at the startup where I used to work putting together manufacturing PCs meant for programming serial numbers into devices, assembled from Intel SSDs, cheap Foxconn nettop computers, and the cheapest sticks of RAM we could find. I must have put together around 15 or so of those machines, and although a lot of them failed due to factory conditions/rough handling/power cuts, I don't think any of the Intel drives ever broke down.
You should back them up, but you should do that regardless of the drive manufacturer. Replacing working drives won't likely gain you anything except a lighter wallet. As you mentioned, you're "in the bathtub" now, and a newer drive is probably more likely to fail. Maintain good backups and replace the drives if/when they fail.
Personally, I have had 3 different Revodrives, all of which have failed. The last one I sent in, and asked to get a more reliable disk back, and I got a Vector. So far so good. Keep away from Revodrives.
If you're a fan of conference calls, then you probably already know that OCZ isn't in the same rosy position it has been in years past. Fortunately, its enterprise-oriented offerings are really helping the company's bottom line. But the situation is darker on the desktop. It's still in the position of needing to source NAND from the fabs manufacturing it, which means it's paying more for the flash it uses and perhaps unable to ship as many units as it'd like.
But again, if you listen to earnings calls, you might already know all of this.
http://www.tomshardware.com/reviews/vector-150-ssd-review,36...
Although only anecdotal, we ditched OCZ at my last company because of their high failure rate; never risked going back. No such issue with Intel or others
After my issues with Western Digital, Asus, OCZ (power supplies) and EVGA (mobo/gfx) I will never trust "gaming" hardware again. Overpriced junk. [OTOH all my Intel, MSI, and AMD hardware is still working after 6+ years in service.]
[1]: http://www.anandtech.com/show/5508/intel-ssd-520-review-cher...
It's pretty hard to make money when you have to provide two products for the price of one. It went on for so long that it was scary.
Say you make $10 profit on a $100 product. For every replacement product, you need to sell 9 more to make your money back on the 1 failure. So, a 10% failure rate means you are basically selling the other 90% just to break even and try to stay in business.
I don't know about OCZ's margins or their failure rate, but with prices squeezing downward, I'd imagine they found themselves in a situation where they maybe had a high enough margin initially that the failure rate wasn't high enough to sink them or the failures didn't occur early enough to burn through the overall margin right away. That's not a sustainable business and the only option is to get people to float you a loan that is maybe big enough to buy time to fix the quality problems if possible. In the event that it's unfixable, no amount of money can save an upside down business model.
Always not a good thing when dealing with such a vital system component.
I hated flashing the things, since there was always a real possibility of turning them into lightweight plastic bricks, and often the firmware was needed to fix show-stopping bugs. so flashing wasn't optional.
Whether these associations are informed by and backed by data or not, they are among the points that steered me into the arms of Intel and Samsung for my SSD needs.
I associate their brand with Sandforce controllers, and I
associate Sandforce controllers with drive failure.
Whether these associations are informed by and backed
by data or not, they are among the points that steered me
into the arms of Intel and Samsung
All other things being equal (quality of NAND, workload, etc) the NAND on Sandforce drives will last longer because of their write compression. Of course, most drives will be replaced or retired before that point anyway.
I think Sandforce had a bit of a bad rap for a while because they were the go-to controller for brands like OCZ who competed on price. Their actual silicon is great. In the hands of a manufacturer like Intel (or even OWC - not to be confused with OCZ) who puts quality and stability first, Sandforce controllers really shine.
The write compression is the reason why I associate Sandforce with "I'm being punished for using full-disk encryption". I could never convince myself to buy any drive with a controller where one of the main advertised "advantages" was completely useless for my usecase.
However, Intel seems to have a more thorough QA team than the other companies. They made a lot of changes to the SF-2281 firmware before they accepted it for their drives. And then after it shipped, they discovered an AES-256 bug that every other manufacturer missed for over a year.
Google "sandforce 1200 haswell"
But by the time the SSD market becomes totally boring PCM may arrive...
Bought a Samsung SSD to replace it.
Good riddance.
Their firmware is ridiculously buggy. I was using one in my primary laptop, and got hit by the bug where, if you sleep the drive too many times, the whole drive fails. Total unrecoverable data loss.
Their forums are full of stories like that.
Next drive, went Intel, haven't looked back since.