Sick of SSD life questions? Lets squash this rumor.

[Magic Carpet]

So why do we see failure rates in the single digit percentage points within a year? These aren't failures where the drive has run out of writes, these are technical failures of parts.

That's a very good point. When an SSD dies... it can be difficult to troubleshoot. And the whole "no movable parts" argument becomes moot. We want something fast, simple, reliable and cheap. Instead, we have fast but complex, unreliable and expensive. Well, of course... not every SSD is like that. Just in general, compared to their mechanical counterparts. And.., in case of data loss, how easy is it... to recover data?

IMO, mechanical disks will still be with us for a very long time.

EDIT: It would be great if mainstream laptops (under $600), would be coming equipped with 2 drive bays, so we could use SSD and HDD at the same time.

 

[Coup27]

I don't remember Anand mentioning wearing out the flash in an SSD. Can you link to the article?

He can't link to it because it doesn't exist. It's another case of not understanding the facts - something unfortunately the SSD industry is plagued with.

Anand has been able to brick SSD's by exploiting firmware bugs. Nothing to do with NAND exhaustion. Infact NAND exhaustion is only an issue for people who want to put them in a server, or people who don't understand the subject.

 

[WiseUp216]

Where do I find my total lifetime writes?

I have SSDLife Pro but I don't see it anywhere! Am I blind?

 

[Blain]

Where do I find my total lifetime writes?

I have SSDLife Pro but I don't see it anywhere! Am I blind?

Google is your friend.

 

[WiseUp216]

Google is your friend.

Says the guy who posted twice in this thread encouraging the posting of duplicate threads.

Seriously?

 

[Coup27]

Where do I find my total lifetime writes?

I have SSDLife Pro but I don't see it anywhere! Am I blind?

I haven't used SSDLife for ages but I am sure in CrystalDiskInfo (which is free) it displays total writes in the top right corner.

 

[WiseUp216]

I haven't used SSDLife for ages but I am sure in CrystalDiskInfo (which is free) it displays total writes in the top right corner.

Thanks, I'll check it out.

 

[jwilliams4200]

He can't link to it because it doesn't exist. It's another case of not understanding the facts - something unfortunately the SSD industry is plagued with.

I suspected that as well, but I wanted to give him a chance to defend his statement.

 

[GotNoRice]

My last X25-M G2 was still at "99% life remaining" when I upgraded to my Crucial M4, and that was after using it for years.

 

[LeftSide]

In many ways I am disappointed with the reliability of SSDs. Solid state in my mind should be approaching the reliability of CPUs or RAM not competing with Hard disk drives. The lack of moving parts should mean they are incredibly reliable devices once they have been quality tested to work at all.

So why do we see failure rates in the single digit percentage points within a year? These aren't failures where the drive has run out of writes, these are technical failures of parts. Until they get to the bottom of why it is so many parts fail long before they should its irrelevant how long the drive will survive being written to.

Lets say an SSD could only survive 1TB of writes and then it would stop working. That is a very short period of time for many people but at least if that was how they behaved you should get a read only drive at the end of it that gave you a warning before it reached the problem point so you could get a replacement. That is a massively better scenario than any HDD because the data is still safe. But it doesn't appear that is what is happening for most people that have SSDs stop working.

The life thing doesn't matter, its the really high failure rates that are the problem, these things should be more reliable than HDDs and yet they aren't.

I agree, that the failure generally has nothing to do with writes, but some people are still worried about it. I think the only reason we have CPUs that operate with such a small failure rates is because of intel's testing mythology. And AMD has to keep just as high of standards to keep up. I've only had one processor go bad, and it was a Cyrix.

 

[Hellhammer]

I haven't used SSDLife for ages but I am sure in CrystalDiskInfo (which is free) it displays total writes in the top right corner.

That depends on the SSD. Not all report host/NAND writes (for example Plextor's SSDs don't).

 

[jwilliams4200]

That depends on the SSD. Not all report host/NAND writes (for example Plextor's SSDs don't).

Actually, the latest M3 Plextor SSDs do appear to track flash writes. With latest firmware (1.04), the M3Ps have smart attribute 177, which from my testing, appears to track flash writes. It is measuring in 64MiB increments.

 

[Zap]

If you haven't seen it already, there's a Xtremesystems forum thread where people go through and try to see how much data can be written to SSDs before they die.

Can anyone decipher the charts in their OP? Maybe I have to wait for my morning coffee to finish kicking in.

The 1st gen indilinx controller had disgraceful write amplification - typically in the region of 30-70x (compared to 1.2-1.5 for a contemporary Intel controller, and even less for Sandforce).

Crap, I still have five 1st gen Indilinx drives. D: One is 60GB for OS on my HTPC that rarely sees use. The other four are 256GB and typically used for game installations. I guess not much is written to them on a regular basis other than game patches.

This thread is in contrast to Anand's own articles where he has claimed to brick SSDs inside of a week.

Anand has been able to brick SSD's by exploiting firmware bugs. Nothing to do with NAND exhaustion.

There's your answer.

I'd say you are vastly more likely to get firmware issues than NAND issues in the first few years of ownership of an SSD.

EDIT: It would be great if mainstream laptops (under $600), would be coming equipped with 2 drive bays, so we could use SSD and HDD at the same time.

Saw a deal at Fry's for Labor Day weekend sales. I think it was a Lenovo Ultrabook for $500 after rebate that comes with a HDD + SSD cache. I'm thinking it likely that the HDD is a normal 2.5" size (maybe 7mm worst case scenario) and the SSD may be mSATA.

I've only had one processor go bad, and it was a Cyrix.

Lucky you.

 

[Cerb]

In many ways I am disappointed with the reliability of SSDs. Solid state in my mind should be approaching the reliability of CPUs or RAM not competing with Hard disk drives. The lack of moving parts should mean they are incredibly reliable devices once they have been quality tested to work at all.

That is typically true. The writes are limited, but that's not why most fail. Most fail because an SSD is not like a CPU, or RAM. It is, no analogy needed, a complete single-board computer, with its own CPU, RAM, and peripheral interfaces (they just happen to heavy on FTLs and SATAs ). As more bugs get worked out, and the controllers get faster, problems with them will become rarer.

Currently, Intel and Crucial seem to be the closest to HDD-like reliability, but indeed, the technology is not sufficiently mature to replace HDDs across the board.

 

[WiseUp216]

That depends on the SSD. Not all report host/NAND writes (for example Plextor's SSDs don't).

Yup. After looking into it, I see that the Crucial M4 does not have a Total Host Writes attribute.

 

[jwilliams4200]

Currently, Intel and Crucial seem to be the closest to HDD-like reliability, but indeed, the technology is not sufficiently mature to replace HDDs across the board.

I have to disagree with Intel there. The 320 series had (and still have) problems, and the Sandforce models have incompatibility problems and crashing with some systems.

Crucial is good. Samsung is also good.

But the best bet for quality and trouble-free operation is Plextor. The newegg reviews are the best for Plextor, and Plextor has 100% burn-in testing and extensive preproduction testing. Take a look at Plextor's reliability testing for the M5P:

http://www.goplextor.com/asia/index.php/ssd/m5-pro

With the M5Pro, Plextor has focused on developing and supplying one of the most stable drives on the market. To achieve this, extensive pre-production testing took place using the world’s most advanced SSD testing facilities - FLEXSTAR testing chamber. It was only when the design was able to pass strict enterprise-grade Zero Error standard of 400 units surviving 500 hours of the toughest continuous testing without a single error or failure that the drive was approved. To ensure stability and reliability for individual drives all drives undergo rigorous high-temperature burn-in tests and accelerated usage simulation testing.


Preproduction tested for professional use in Plextor’s state-of-the-art testing facilities. Rigorous operational and environmental tests resulting in drive with an impressive MTBF of 2.4 million hours.

48hr Sustained Read and Write Test
Two days continuous 4K file read / write test, zero error pass requirement.
4000 Times Idle Test–Sleep and Hibernation Test
After 24 hrs read & write SSD enters sleep (S3) mode and is woken after 5 minutes, repeated for hibernation (S4) mode, zero error pass requirement.
250 Times Power Cycle–Cold and Warm Boot Tests
Continuous cold boot and warm boot cycles 250 times, zero error pass requirement.

And for Plextor's other models:

http://www.goplextor.com/asia/index.php/ssd/m5s

High Plextor Quality : Stringent testing for every drive

Every Plextor SSD is built to exacting standards and thoroughly tested to ensure that it is manufactured to our impeccable quality demands. These include 20-hour High Temperature burn-in test and the rigors of the industry's premier-level FLEXSTAR testing machine to simulate the real working environment. Plextor SSD devices are of the highest quality, as borne out by the 0.5% average annual failure rate (AFR), the lowest in the industry.

 

[Coup27]

Very impressive. You would have thought Plextor would have wanted AT to include that in their drive reviews as its good publicity.

 

[Hellhammer]

Very impressive. You would have thought Plextor would have wanted AT to include that in their drive reviews as its good publicity.

I did mention it briefly in the M5S review:

Plextor says that all their SSDs go through the same validation process, regardless of the series. I don't know the exact specifics of their testing methods, but according to their website all SSDs are tested for 20 hours in a high temperature burn-in test. Plextor is also claiming that their average annual failure rate is 0.5%.

http://www.anandtech.com/show/6090/plextor-m5s-256gb-review/2

Personally I don't really like quoting manufacturers' testing methods. They are created by marketing people and their only aim is to sell the product. Essentially it's the same as looking at the speeds manufacturers report and saying which SSD is the best based on them. I'm not saying Plextor is lying but without knowing how others test their SSDs, I find it pointless to praise one manufacturer's testing methods.

I have to agree with jwilliams4200 that Plextor is definitely among the most reliable SSD manufacturers, though. Whatever their testing methods are, they seem to be pretty darn effective.

 

[Cerb]

I have to disagree with Intel there. The 320 series had (and still have) problems, and the Sandforce models have incompatibility problems and crashing with some systems.

Crucial is good. Samsung is also good.

Samsung 830 with GPT.

If you can bring up the Intel 320, then Crucial doesn't belong on the list, either, because of their C300 drives breaking (not just slowing down), due to needing idle time that most PCs wouldn't give them.

AFAIK, M3 series on certain MPBs do seem to be the only major Plextor problem, though, which is about as good as most HDDs (HDD compatibility and reliability is not perfect, and they've had their bad firmware bugs, too).

 

[jwilliams4200]

Personally I don't really like quoting manufacturers' testing methods. They are created by marketing people and their only aim is to sell the product. Essentially it's the same as looking at the speeds manufacturers report and saying which SSD is the best based on them. I'm not saying Plextor is lying but without knowing how others test their SSDs, I find it pointless to praise one manufacturer's testing methods.

Actually, preproduction (aka, qualification) and burn-in test specifications are almost always written by engineers. I've never heard of them being written by marketing people. Marketing people may suggest a standard that the product should meet, but in all cases I am familiar with the engineers write the specifications.

Perhaps you meant that the description on the web page was written by marketing people. While that is likely the case, the marketing people almost surely just asked the engineers for a description of the tests and then excerpted it or used it verbatim. I recognize the language of an engineering test specification (I've written some myself), and it looks legitimate.

And it is far from pointless to praise a manufacturer's PUBLISHED testing methods. At a minimum, it is good that the manufacturer makes an effort to inform the customer -- that shows at least a minimal commitment to quality and pride enough to let customers know about it. That alone is a good thing. :thumbsup:

 

[jwilliams4200]

Samsung 830 with GPT.

If you can bring up the Intel 320, then Crucial doesn't belong on the list, either, because of their C300 drives breaking (not just slowing down), due to needing idle time that most PCs wouldn't give them.

Was there ever a definitive test showing that the Samsung 830 was actually INTERNALLY (i.e., not their supplied software) corrupting data previously written to legitimate LBAs? I saw the thread about it here, but I did not see any clear demonstration that it was happening internally.

As for the C300, I'm not familiar with the problems you mentioned, but note that the Intel 320 was introduced about the same time the Crucial m4 was (March 2011), and the C300 is older.

Anyway, I don't recommend the m4 or the C300. I think choosing an SSD these days is simple. For the best value (good price, good performance, good quality) choose the Samsung 830. For the best SSD (best quality, best performance) choose the Plextor M3P or Plextor M5P.

 

[Sherman Bay]

If they only had a "meter" showing how near failure the drive is, it would make me feel safer, and give a warning. And is a failure a soft or hard one? I have a 125GB SSD, but am afraid to use it for anything more than short-term backup or temporary work files.

 

[Coup27]

If they only had a "meter" showing how near failure the drive is, it would make me feel safer, and give a warning. And is a failure a soft or hard one? I have a 125GB SSD, but am afraid to use it for anything more than short-term backup or temporary work files.

Either you haven't done any research into SSDs or you need to start again.

 

[Sherman Bay]

Either you haven't done any research into SSDs or you need to start again.

I am doing the research as we speak. One element of my research is checking forums for experiences and ideas. But it appears that your reply hasn't provided any useful information. Do you have any?

 

[Blain]

But it appears that your reply hasn't provided any useful information.

Ofttimes things are not always as they appear on the surface.