Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.sys.ibm.pc.hardware.storage > #5525 > unrolled thread

"Consumer-grade SSDs actually last a hell of a long time"

Back to article view | Back to comp.sys.ibm.pc.hardware.storage

Contents

  "Consumer-grade SSDs actually last a hell of a long time" Lynn McGuire <lmc@winsim.com> - 2014-06-17 12:04 -0500
    Re: "Consumer-grade SSDs actually last a hell of a long time" "Rod Speed" <rod.speed.aaa@gmail.com> - 2014-06-18 03:39 +1000
      Re: "Consumer-grade SSDs actually last a hell of a long time" Yousuf Khan <bbbl67@spammenot.yahoo.com> - 2014-06-17 16:09 -0400
        Re: "Consumer-grade SSDs actually last a hell of a long time" Mark F <mark53916@gmail.com> - 2014-06-18 07:47 -0400
          Re: "Consumer-grade SSDs actually last a hell of a long time" Mike Tomlinson <mike@jasper.org.uk> - 2014-06-20 10:52 +0100
            Re: "Consumer-grade SSDs actually last a hell of a long time" Mark F <mark53916@gmail.com> - 2014-06-20 08:24 -0400
              Re: "Consumer-grade SSDs actually last a hell of a long time" RJH <patchmoney@gmx.com> - 2014-06-20 13:31 +0100
                Re: "Consumer-grade SSDs actually last a hell of a long time" GB <NOTsomeone@microsoft.com> - 2014-06-20 14:29 +0100
            Re: "Consumer-grade SSDs actually last a hell of a long time" Rob Morley <nospam@ntlworld.com> - 2014-06-20 20:28 +0100
        Re: "Consumer-grade SSDs actually last a hell of a long time" "Rod Speed" <rod.speed.aaa@gmail.com> - 2014-06-20 06:14 +1000
    Re: "Consumer-grade SSDs actually last a hell of a long time" VanguardLH <V@nguard.LH> - 2014-06-17 18:30 -0500
      Re: "Consumer-grade SSDs actually last a hell of a long time" Arno <me@privacy.net> - 2014-06-18 01:51 +0000
    Re: "Consumer-grade SSDs actually last a hell of a long time" Shadow <Sh@dow.br> - 2014-06-18 13:21 -0300
      Re: "Consumer-grade SSDs actually last a hell of a long time" Yousuf Khan <bbbl67@spammenot.yahoo.com> - 2014-06-19 11:40 -0400
        Re: "Consumer-grade SSDs actually last a hell of a long time" Shadow <Sh@dow.br> - 2014-06-19 19:27 -0300

#5525 — "Consumer-grade SSDs actually last a hell of a long time"

"Consumer-grade SSDs actually last a hell of a long time"
    http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/

I've got four of the Intel 5xx series here in the
office with no failures.  Three of them are over a
year old.

Lynn

[toc] | [next] | [standalone]

#5526

"Lynn McGuire" <lmc@winsim.com> wrote in message 
news:lnpsek$6vo$1@dont-email.me...
> "Consumer-grade SSDs actually last a hell of a long time"
> 
> http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
>
> I've got four of the Intel 5xx series here in the
> office with no failures.  Three of them are over a
> year old.

A year isnt very long. 

[toc] | [prev] | [next] | [standalone]

#5527

On 17/06/2014 1:39 PM, Rod Speed wrote:
>
>
> "Lynn McGuire" <lmc@winsim.com> wrote in message
> news:lnpsek$6vo$1@dont-email.me...
>> "Consumer-grade SSDs actually last a hell of a long time"
>>
>> http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
>>
>>
>> I've got four of the Intel 5xx series here in the
>> office with no failures.  Three of them are over a
>> year old.
>
> A year isnt very long.

It is if they are being torture tested during that entire year.

I've had one of my drives for over 425 days, and it's "only" had 11TB of 
writes to it (and it's only a 240GB drive). So it's nothing compared to 
the 700TB to 1PB of writes that's happened to these drives.

	Yousuf Khan

[toc] | [prev] | [next] | [standalone]

#5530

On Tue, 17 Jun 2014 16:09:21 -0400, Yousuf Khan
<bbbl67@spammenot.yahoo.com> wrote:

> On 17/06/2014 1:39 PM, Rod Speed wrote:
> >
> >
> > "Lynn McGuire" <lmc@winsim.com> wrote in message
> > news:lnpsek$6vo$1@dont-email.me...
> >> "Consumer-grade SSDs actually last a hell of a long time"
> >>
> >> http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
> >>
> >>
> >> I've got four of the Intel 5xx series here in the
> >> office with no failures.  Three of them are over a
> >> year old.
> >
> > A year isnt very long.
Another factor to consider it how long the data is retained;
consider both the case of un referenced data when the power
is on and all data when the power is off.

(As the memory gets worn out the data retention time decreases.
 So un referenced data can be lost.  Most, if not all, current
 devices will automatically check and refresh things if the
 are powered up, but powered down retention time might
 have originally been 15 years or more, but less than 1 year on
 a used device.)
> 
> It is if they are being torture tested during that entire year.
> 
> I've had one of my drives for over 425 days, and it's "only" had 11TB of 
> writes to it (and it's only a 240GB drive). So it's nothing compared to 
> the 700TB to 1PB of writes that's happened to these drives.
> 
> 	Yousuf Khan

[toc] | [prev] | [next] | [standalone]

#5535

In article <3ou2q9p97haauak10o0dj7amk5njr6q2h5@4ax.com>, Mark F
<mark53916@gmail.com> writes

>Another factor to consider it how long the data is retained;
>consider both the case of un referenced data when the power
>is on and all data when the power is off.

This is actually quite an important question.

If data are written to an SSD and that SSD is then unplugged, what is
the guaranteed retention time without corruption, of that data?

Hard disks are well known to suffer from "bit-rot", but do we have any
data on whether it's a problem on SSDs?

-- 
 (\_/)
(='.'=)
(")_(")

[toc] | [prev] | [next] | [standalone]

#5536

On Fri, 20 Jun 2014 10:52:40 +0100, Mike Tomlinson
<mike@jasper.org.uk> wrote:

> In article <3ou2q9p97haauak10o0dj7amk5njr6q2h5@4ax.com>, Mark F
> <mark53916@gmail.com> writes
> 
> >Another factor to consider it how long the data is retained;
> >consider both the case of un referenced data when the power
> >is on and all data when the power is off.
> 
> This is actually quite an important question.
> 
> If data are written to an SSD and that SSD is then unplugged, what is
> the guaranteed retention time without corruption, of that data?
> 
> Hard disks are well known to suffer from "bit-rot", but do we have any
> data on whether it's a problem on SSDs?
I found various theoretical items using Google that support the idea.

I couldn't find any current articles from flash or SSD manufacturers,
but in the past I have seen 10 year initial powered down retention
time and on the same page end of life determined by 1 year un powered
retention time.

I did find some references from SSD and flash memory:

SanDisk has or had a product with 100 year powered down retention
time:
 http://www.sandisk.com/go/preserve/
the links to details are broken.

SanDisk also has an article with byline NetApp:

http://www.sandisk.com/assets/docs/NetApp%20Write-Intensive%20SSD%20White%20Paper.pdf
that indicates:
 Expected lifetime is a calculated number based on both theoretical
and demonstrated flash endurance limits.
 "the device still meets its specified error rate after being powered
  off for three months"
and mentions refreshing data in the background when power is on.

This is from a user of SSDs (Dell) in a report from 2011:

http://www.dell.com/downloads/global/products/pvaul/en/Solid-State-Drive-FAQ-us.pdf
that says (on "6 | Page")
 " 6. I have unplugged my SSD drive and put it into storage. How long
      can I expect the drive to retain my data without needing to plug
      the drive back in?
   NAND Technology Data Retention @ rated P/E cycle 
   SLC             6 Months 
   eMLC            3 months 
   MLC             3 Months 


After doing these searches today, I agree with
Yusaf Kahn (6/19/2014 11:40 AM) that SSDs probably have a shorter
powered down data retention time than flash memory keys {and cameras
memory cards.)

Also, in 2010 I felt that a new SSD used for backup would probably
retain memory while powered for more than 5 years, but I think that
even SSDs that have only been written to a small amount and then
kept normally powered down should be powered up for a day or so
every year.

[toc] | [prev] | [next] | [standalone]

#5537

On 20/06/2014 13:24, Mark F wrote:
> On Fri, 20 Jun 2014 10:52:40 +0100, Mike Tomlinson
> <mike@jasper.org.uk> wrote:
>
>> In article <3ou2q9p97haauak10o0dj7amk5njr6q2h5@4ax.com>, Mark F
>> <mark53916@gmail.com> writes
>>
>>> Another factor to consider it how long the data is retained;
>>> consider both the case of un referenced data when the power
>>> is on and all data when the power is off.
>>
>> This is actually quite an important question.
>>
>> If data are written to an SSD and that SSD is then unplugged, what is
>> the guaranteed retention time without corruption, of that data?
>>
>> Hard disks are well known to suffer from "bit-rot", but do we have any
>> data on whether it's a problem on SSDs?
> I found various theoretical items using Google that support the idea.
>
> I couldn't find any current articles from flash or SSD manufacturers,
> but in the past I have seen 10 year initial powered down retention
> time and on the same page end of life determined by 1 year un powered
> retention time.
>
> I did find some references from SSD and flash memory:
>
> SanDisk has or had a product with 100 year powered down retention
> time:
>   http://www.sandisk.com/go/preserve/
> the links to details are broken.
>
> SanDisk also has an article with byline NetApp:
>
> http://www.sandisk.com/assets/docs/NetApp%20Write-Intensive%20SSD%20White%20Paper.pdf
> that indicates:
>   Expected lifetime is a calculated number based on both theoretical
> and demonstrated flash endurance limits.
>   "the device still meets its specified error rate after being powered
>    off for three months"
> and mentions refreshing data in the background when power is on.
>
> This is from a user of SSDs (Dell) in a report from 2011:
>
> http://www.dell.com/downloads/global/products/pvaul/en/Solid-State-Drive-FAQ-us.pdf
> that says (on "6 | Page")
>   " 6. I have unplugged my SSD drive and put it into storage. How long
>        can I expect the drive to retain my data without needing to plug
>        the drive back in?
>     NAND Technology Data Retention @ rated P/E cycle
>     SLC             6 Months
>     eMLC            3 months
>     MLC             3 Months
>
>
> After doing these searches today, I agree with
> Yusaf Kahn (6/19/2014 11:40 AM) that SSDs probably have a shorter
> powered down data retention time than flash memory keys {and cameras
> memory cards.)
>
> Also, in 2010 I felt that a new SSD used for backup would probably
> retain memory while powered for more than 5 years, but I think that
> even SSDs that have only been written to a small amount and then
> kept normally powered down should be powered up for a day or so
> every year.
>

Not as long as books, then ;-)

It's kind of ironic that as we move through the most culturally and 
technologically rich period of time (on earth), a good deal of our media 
will, in the not too distant future, be lost. Perhaps.

-- 
Cheers, Rob

[toc] | [prev] | [next] | [standalone]

#5538

On 20/06/2014 13:31, RJH wrote:

> Not as long as books, then ;-)
>
> It's kind of ironic that as we move through the most culturally and
> technologically rich period of time (on earth), a good deal of our media
> will, in the not too distant future, be lost. Perhaps.
>

It's safer than the library at Alexandria, though.

[toc] | [prev] | [next] | [standalone]

#5539

On Fri, 20 Jun 2014 10:52:40 +0100
Mike Tomlinson <mike@jasper.org.uk> wrote:

> This is actually quite an important question.
> 
> If data are written to an SSD and that SSD is then unplugged, what is
> the guaranteed retention time without corruption, of that data?
> 
> Hard disks are well known to suffer from "bit-rot", but do we have any
> data on whether it's a problem on SSDs?
> 
Just in case it is you can generate PAR files for anything important, so
minor bit rot is less likely to stop you recovering your files.

[toc] | [prev] | [next] | [standalone]

#5533

"Yousuf Khan" <bbbl67@spammenot.yahoo.com> wrote in message 
news:53a0a06c$1@news.bnb-lp.com...
> On 17/06/2014 1:39 PM, Rod Speed wrote:
>>
>>
>> "Lynn McGuire" <lmc@winsim.com> wrote in message
>> news:lnpsek$6vo$1@dont-email.me...
>>> "Consumer-grade SSDs actually last a hell of a long time"
>>>
>>> http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
>>>
>>>
>>> I've got four of the Intel 5xx series here in the
>>> office with no failures.  Three of them are over a
>>> year old.
>>
>> A year isnt very long.
>
> It is if they are being torture tested during that entire year.

His wasn't.

> I've had one of my drives for over 425 days, and it's "only" had 11TB of 
> writes to it (and it's only a 240GB drive). So it's nothing compared to 
> the 700TB to 1PB of writes that's happened to these drives.
 

[toc] | [prev] | [next] | [standalone]

#5528

Lynn McGuire wrote:

> "Consumer-grade SSDs actually last a hell of a long time"
>     http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
> 
> I've got four of the Intel 5xx series here in the
> office with no failures.  Three of them are over a
> year old.

So 10 months under torture testing of writes is how long in years under
a normal load of writes?

[toc] | [prev] | [next] | [standalone]

#5529

VanguardLH <V@nguard.lh> wrote:
> Lynn McGuire wrote:

>> "Consumer-grade SSDs actually last a hell of a long time"
>>     http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
>> 
>> I've got four of the Intel 5xx series here in the
>> office with no failures.  Three of them are over a
>> year old.

> So 10 months under torture testing of writes is how long in years under
> a normal load of writes?

Depends, but for a system drive, e.g. something like 10-100 years.
For a data-drive possibly longer.

Arno

[toc] | [prev] | [next] | [standalone]

#5531

On Tue, 17 Jun 2014 12:04:07 -0500, Lynn McGuire <lmc@winsim.com>
wrote:

>"Consumer-grade SSDs actually last a hell of a long time"
>    http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/

	Fantastic. My 64Mb USB is only 10 years old, and it still
works fine. Excited it might still be working in another 10 or more
years.
	;)
	[]'s
-- 
Don't be evil - Google 2004
We have a new policy  - Google 2012

[toc] | [prev] | [next] | [standalone]

#5532

On 18/06/2014 12:21 PM, Shadow wrote:
> On Tue, 17 Jun 2014 12:04:07 -0500, Lynn McGuire <lmc@winsim.com>
> wrote:
>
>> "Consumer-grade SSDs actually last a hell of a long time"
>>     http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
>
> 	Fantastic. My 64Mb USB is only 10 years old, and it still
> works fine. Excited it might still be working in another 10 or more
> years.


A USB flash drive is not the same thing as an SSD, even though they both 
use flash storage. The SSD would have much more sophisticated load 
balancing features, and much more redundancy.

	Yousuf Khan

[toc] | [prev] | [next] | [standalone]

#5534

On Thu, 19 Jun 2014 11:40:07 -0400, Yousuf Khan
<bbbl67@spammenot.yahoo.com> wrote:

>On 18/06/2014 12:21 PM, Shadow wrote:
>> On Tue, 17 Jun 2014 12:04:07 -0500, Lynn McGuire <lmc@winsim.com>
>> wrote:
>>
>>> "Consumer-grade SSDs actually last a hell of a long time"
>>>     http://arstechnica.com/gadgets/2014/06/consumer-grade-ssds-actually-last-a-hell-of-a-long-time/
>>
>> 	Fantastic. My 64Mb USB is only 10 years old, and it still
>> works fine. Excited it might still be working in another 10 or more
>> years.
>
>
>A USB flash drive is not the same thing as an SSD, even though they both 
>use flash storage. The SSD would have much more sophisticated load 
>balancing features, and much more redundancy.
>

	I was just pointing out that sizes get redundant much faster
than durability. Ten years is more than enough.
	[]'s
-- 
Don't be evil - Google 2004
We have a new policy  - Google 2012

[toc] | [prev] | [standalone]

Back to top | Article view | comp.sys.ibm.pc.hardware.storage

csiph-web