Once-great SSD manufacturer OCZ filing for bankruptcy(arstechnica.com)
arstechnica.com
Once-great SSD manufacturer OCZ filing for bankruptcy
http://arstechnica.com/gadgets/2013/11/once-great-ssd-manufacturer-ocz-filing-for-bankruptcy/
91 comments
Probably the best thing you can say about OCZ is that they played a huge role in establishing the consumer SSD market.
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.
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.
> They lied to their customers by selling hardware under the same name as previously produced hardware with cheaper components and lesser specs
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.
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.
Sounds like what Packard Bell computers did. They were inexpensive. And if you bought two of the same model and opened them up you'd see different components. Different modems, ram, etc.
The problem here though is that they were shipping off their spicier models to review, while the shrink wrapped version you would buy was off-spec.
Apple does the same with its display manufacturers.
Their drives were better than Crucial. I had a Crucial SSD that due to a firmware bug became unusably slow (far slower than a mechanical drive) after a couple of months. They released a firmware update that supposedly fixed it (which also required tweaking BIOS settings and booting in to a weird minimal version of Windows 95 - in 2013...), but it only made it slightly better, still a totally useless drive. I'd take an OCZ drive any day - been using one in my main dev desktop for a while now and it's kept up just fine.
Surprised it's not Crucial going under after that experience.
Surprised it's not Crucial going under after that experience.
I bought 2x4GB new Crucial memory sticks to upgrade a Mac this year, and they failed memtest. They are no longer a premium brand IMO. Replaced them with used Kingstons from eBay. No problems.
> They built drives with extreme speeds while entirely sacrificing longevity and reliability.
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.
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.
> What they didn't say on the package WITH CAPITAL LETTERS was that it was a RAID0/JBOD of SSDs with zero redundancy
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?
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?
> At best they had a great marketing department that made it possible for them to peddle their crap to the public for so long.
They also chose the most susceptible target market - teen gamers and overclockers. "Being cool" is their mantra and OCZ delivered.
They also chose the most susceptible target market - teen gamers and overclockers. "Being cool" is their mantra and OCZ delivered.
"They built drives with extreme speeds while entirely sacrificing longevity and reliability." This is so true, I had their SSD before, and it broken ==> I lost most of my data in that drive...
I'm not sure it's so bad. I bought my OCZ Vertex 3 Max IOPS 2.5 years ago, it's been great and I'm still happy with it. They also have had many good reviews.
No, it is so bad. OCZ was a terrible company and nobody is going to miss them.
My Vertex 4s are doing well after 18 months and my crucial m4s died within 6 (power system failures). I'll miss ocz.
Crucial is only marginally better than OCZ in terms of reliability. You got lucky. If you care about drives surviving, buy Intel or Samsung:
http://www.hardware.fr/articles/911-7/ssd.html
http://www.hardware.fr/articles/911-7/ssd.html
Either explain why or save your breath :|
I've had to have them replace my OCZ Vertex 2 three times.
Both of my OCZ Vertex 2 died on me. I was able to bring one back from the dead with firmware upgrade and repartitioning and reformatting. Second one serves as interesting although bit thick coaster. Couldn't return them to manufacturer because they were not bought from one of their listed retailers. I couldn't return them to the shop either because they worked few years before any trouble happened.
One of my Vertex2's died and I need to send it in on warranty (if it isn't too late). The other one is doing fine for now.
I was lucky I got RMA and got a replacement. I sold the replacement on ebay once I got it and ever since I couldn't trust another SSD myself. Though I am still very happy with the SSD that comes with my Macbook Pro 15"
Same here. Swapped out the stock drive in my MPB with a Vertex 3 in April 2011, and its been rock solid ever since with everyday-use.
Everyone on HN is glad you're happy, but your single data point does fly in the face of statistically significant reliability problems with OCZ drives (ie, the only real way to evaluate storage manufacturers).
The plural of anecdote is not data.
(Just a note to keep in mind when seeing all these personal testimonies)
(Just a note to keep in mind when seeing all these personal testimonies)
> "The plural of anecdote is not data" -unknown source.
FTFY.
I read this about a year ago and have been unable to find the source. Does anyone know where this is from?
FTFY.
I read this about a year ago and have been unable to find the source. Does anyone know where this is from?
In an era of high-quality search, you could plug this into the search engine of your choice and get a better answer than you're likely to find on the HN comments.
e.g.
http://askville.amazon.com/original-source-quote-plural-anec...
http://freakonomics.com/2010/04/29/quotes-uncovered-whats-th...
e.g.
http://askville.amazon.com/original-source-quote-plural-anec...
http://freakonomics.com/2010/04/29/quotes-uncovered-whats-th...
Do you have actual statistics? I'd truly like to see them.
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)
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)
Enjoy:
http://translate.googleusercontent.com/translate_c?depth=2&h...
http://www.hardware.fr/articles/911-7/ssd.html
http://translate.googleusercontent.com/translate_c?depth=2&h...
http://www.hardware.fr/articles/911-7/ssd.html
And from 2012:
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.
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.
Wow thanks for the data! In 10 years of interest in acquiring data in hardware return stats, this is the first time I come across such a source.
"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."
"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."
Oh good! I'd posted specifically to ensure just you were happy, but it warms my heart to know that the entirety of HN is officially glad that I'm happy too!
I'll provide a dissenting data point. After going through 5 RMA's with Vertex 2 drives, I finally convinced them to just upgrade me to a Vertex 3, due to reports of better reliability. The Vertex 3 failed within a week. Considered the drive a lost cause, swapped in a Crucial M4 -- been rock solid ever since.
I had a similar experience with the Vertex 2, and switched to Intel 320 and subsequently Samsung 840 Pro. Intel and Samsung drives have done a great job, while I had 3 Vertex 2 failures.
please, unless you compared their offering with other ssds or used their support forums, this comment is as empty as it can be.
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.
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.
Also using actual buyers/users for testing their drives and firmware. Not cool at all...
Those numbers are actually incredibly skewed and not very accurate - I can confidently say that a large amount of the returns in the Vertex 2/3 and Agility 2/3 lines were due to firmware problems which were fixed very early on. Customers simply didn't understand that upgrading firmware was not optional, and the support team was not large/capable enough to teach this to every customer. Many of the drives which were returned were simply re-flashed to the new firmware and sent back, working perfectly.
I have 4 OCZ SSDs of various models. When I bought them I'd read the forums and there was almost always a notice about upgrading the firmware to fix issues. I always kept up on firmware updates and I've never had a problem with a drive.
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.
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.
In the "hardware junkies" mailing list at work (Microsoft) the OCZ drives were a periodic source of derision.
"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.
"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.
I am pleased that Microsoft has a list called "hardware junkies" :). I'd definitely be on it if I were there.
Nice to know they agree with the rest of us that Samsung is the way to go nowadays, too.
Nice to know they agree with the rest of us that Samsung is the way to go nowadays, too.
On Newegg, OCZ SSDs would consistently have a larger number of 1-star reviews than other drives. Whereas Intel and Samsung drives would consistently have a smaller number of 1-star reviews.
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.
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.
The newest model is TLC, but it is likely that they figured out a way to make it last long enough for average consumers.
We had about 40 OCZ drives (Vertex 2/3/3 Max IOPS) and 6 of them failed. Our 8 Vertex 3 Max IOPS were in RAID 5 for a huge calculations which didn't require reliability but wrote a lot of data and therfore we tried to save time with this experimental RAID. The RAID was fine for about a couple of month and seeing almost 3 GB/s throughput was mindblowing. But suddenly we we saw drives randomly failing. But the drives did not completly fail, we were able to rebuild the RAID with the same hard drive. We did that a couple of times until we thought it was too much hassle and used the drives somewhere else. Now we are buying Crucial M4s and they are totally fine. In the first place a hard drive must be reliable.
This was a firmware bug originating from the Sandforce firmware which the releases were based on - I think this one was completely fixed at least 8 months ago.
We had to replace 50+ Agility 128s as around half of them went bad within weeks of install.
A year later and none of the replacement Intels have gone bad...
A year later and none of the replacement Intels have gone bad...
I wasn't aware of it before reading into this whole story, but OCZ seemed to have quite a problem with QA. My vertex 2 barely lasted a year before it started going funny, which I put down to bad luck. But now it seems I wasn't the only one.
They are notorious for pushing out low quality products. You really should be a more informed consumer before you buy stuff, it's not like it takes a long time to research this kind of stuff.
At the consumer level, it is quite difficult to research a lot of this kind of thing. There is so much contradictory information out there on almost every product, and trying to ferret out the right information can be a gruelling task.
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".
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".
Well, Newegg at least makes it possible to distinguish an OCZ from an Intel/Samsung. When an OCZ drive has 3-stars and 30% of them are 1-star reviews, then you know something is wrong, compared to the Intel drive that has 4.5-stars with only 5% being 1-star.
It's in the finer gradations that it gets murky. Is a 4.5-star drive really better than a 4-star drive?
It's in the finer gradations that it gets murky. Is a 4.5-star drive really better than a 4-star drive?
I have OCZ drives in a couple of my machines, all of which are at least a year old. I'd heard bad things about OCZ, but I didn't realize their reputation was that shot. I'm wondering if I should replace my drives, or if the bathtub curve is in full effect here and I've dodged a bullet...
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.
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.
> I'm wondering if I should replace my drives, or if the bathtub curve is in full effect here\
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.
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.
Depends on the model of the SSD.
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.
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.
Some of this writing had been on the walls:
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...
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...
However, a long string of failures across several of its product lines (most notably the high-performance Vertex family) took a lot of the shine off of OCZ's name
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
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
In the early SSD days, I only heard two adjectives to describe OCZ drives - "fastest" and "unstable". It was inevitable the latter would catch up to them.
When I bought my SSD, I narrowed in on Intel very quickly. I don't trust "gamer" / benchmark-focused brands at all for this kind of product.
Definately, Intel fabs their own NAND devices in the USA! The 320/520 also has Intel's own controller in it. I have a 320 & a 330 striped in my laptop (SATA II). The have been working flawlessly for over a year, nearly two for the 320.
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.]
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.]
The 520 is SandForce[1], not Intel's own controller. It looks like Intel has given up on using their own controllers for their consumer SSDs, given that the 330 is also SandForce. Their own next generation controller is in their DC series SSDs marketed towards enterprise & datacenter usage.
[1]: http://www.anandtech.com/show/5508/intel-ssd-520-review-cher...
[1]: http://www.anandtech.com/show/5508/intel-ssd-520-review-cher...
Thanks for the clarification, my memory was off. Link to Intel/Micron site: http://imflash.com/
The one thing that will kill a hardware company much faster than a software company is managing supply chain and inventory costs. I don't know the full backstory of OCZ other than seeing how the SSD segment of the storage industry has gone and the prices have obviously gone a lot lower and are more competitive. Depending on how OCZ managed their production, it would be very easy to be left holding the bag on millions of dollars of inventory. Unless you have billions in the bank, that's enough to sink a company every time.
They had a high rate of warranty replacements for several years...so much so that it cost the founder and CEO his job.
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.
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.
Well that would totally do it. High failure rate on a low margin product is nearly as bad as having too much inventory because the effect is very much the same. Example:
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.
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.
In addition to OCZ's reliability issues, it must have been tough to compete on price with companies like Samsung and Intel that fab their own flash memory.
I wonder why Toshiba wants OCZ given Toshiba is already making SSD themselves and been doing fine. As someone has already mentioned down below, OCZ to me was a RAM company before they sold first SSD. Is Toshiba trying to compete in the SSD market too? I thought Samsung and Intel were pretty much the winners in this market. If OCZ exist because of quality, what can Toshiba can get out of the acquisition? The factory? The machines? The top engineers?
I assume just the branding, really. They did very well with the gamer market.
I'd guess factory, brand, and patents.
I remember them being something other than an SSD manufacturer. I've got a few sticks of DDR memory from them. In fact I was surprised to see them producing flash memory products lately (for a very long definition of lately).
Not surprised, their installation instructions should have had "update firmware" as a step...
The first three steps were 1)hold breath 2)cross fingers 3)pray.
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.
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.
I have avoided them since I started buying SSDs because 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 for my SSD needs.
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.
> All other things being equal (quality of NAND, workload, etc) the NAND on Sandforce drives will last longer because of their write compression.
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.
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.
Intel's consumer SSDs use Sandforce controllers.
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.
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.
Almost unrelated but every SSDs with the old SF-1200 (such as Corsair F60) are not compatible with the current Intel Haswell platform. The drive randomly disappears from the system and this on ALL motherboards.
Google "sandforce 1200 haswell"
Google "sandforce 1200 haswell"
I had a OCZ petrol drive that failed recently, it was just over a year old, but have had a vertex drive in another macbook pro that's still alive after 2 years. I just try and back up regularly. I'm surprised at the bankruptcy but they did have huge problems, their forums were inundated with issues.
This is upsetting as we are moving towards a tech world filled with only a few big brands. Will we only be able to buy SSDs from Intel or Samsung in the future? Time will tell, but I am upset that the small guy can't compete in this space anymore.
I would definitely expect the market to vertically integrate so that most SSDs are made by NAND vendors. That would mostly leave us with Samsung, Intel, Micron, and Toshiba.
But by the time the SSD market becomes totally boring PCM may arrive...
But by the time the SSD market becomes totally boring PCM may arrive...
I'm glad I stuck with Intel ssds even though they eventually switched to SandForce I have drives with both controllers and they are both awesome. I almost bought an OCZ but they were garbage especially now Samsung is a player.
When I got my first SSD over a year ago, I bought a 250gb OCZ Agility 3 drive. Not knowing about the very high drive failure rate. It died last week, completely without warning.
Bought a Samsung SSD to replace it.
Bought a Samsung SSD to replace it.
I don't understand why in the world would you spend more money on a new drive - if you only bought that drive a year ago you could have it replaced under warranty. Even if it only works for another year, it's still infinitely better(in my mind) than spending money on a new drive.
I own a vertex 3, never had an issue with it going on 2 years. It is used in an osx laptop that is used aggressively for devops work. Guess I got one of the good ones...
A few years ago, I went through at least two different sets of faulty OCZ RAM before exchanging for the vastly superior Corsair memory.
Good riddance.
Good riddance.
good. I picked up a 30GB ssd on eBay that was NIB for like 30 bucks and it failed while installing arch linux. I contacted the seller and he happily refunded me. That was my first experience with OCZ as I have always used Intel or Kingston SSD.
I have an OCZ disk.... in the bottom of a drawer, failed.
Same, all it's good for is being a paperweight now.
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.
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.
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...
http://www.hardware.fr/articles/911-7/ssd.html