Archived from groups: comp.sys.ibm.pc.hardware.storage (
More info?)
Arno Wagner <me@privacy.net> wrote:
> Previously Peter <peterfoxghost@yahoo.ca> wrote:
>
>> "Arno Wagner" <me@privacy.net> wrote in message
>> news:3olloqF6ifp1U1@individual.net...
>>> Previously Peter <peterfoxghost@yahoo.ca> wrote:
>>>>> Solid state is massively more reliable
>>>
>>>> Not always:
>>>>
http://www.bitmicro.com/products_edisksan_2U_RAID_fc.php
>>>> ...
>>>> Pure Solid State/Non-Volatile
>>>> MTBF - 400,000 Hours
>>>> +5 to + 35 degrees C
>>>> 10,000 feet Altitude
>>>
>>> Well, yes.
>> Bitmicro E-Disks are not cheap, yet not massively reliable. Popular
>> hard drives have better reliabilty.
> Actually MTBF is only part of the story.
He didnt say anything about MTBF.
> Bit-error rate is another. There may also be SSDs that are
> optimised for speed only, I agree. SSDs used to be only an
> option for special reliability needs, I guess there is a whole
> set of products targetted at different usage scenarios now.
Corse there are.
> Also reliability of SSDs is usually used under conditions like
> shock and vibration (nearly indestructable) or for example in
> non-standard gas mixes, low/no air pressure, humindity or the like.
> SSDs will do pretty well under thse conditions while HDDs die very fast.
Irrelevant to the normal use of both types of storage.
> One other thing: For HDDs component life
> (the time the MTBF is valid) is usually 5 years.
Wrong, that is an entirely different concept, the design life.
> For SSDs it can be 30 years or around.
Not with mass market retail SSDs.
> Very important for equipment that cannot be easily replaced.
Tiny part of the market now.
> Also note thet the eqipment you reference is not an SSD, but a whole
> box with PSU and the like. The SSD itself may actually have a
> better MTBF then the whole box. It is especially difficult to
> get PSUs with > 500.000h MTBF. It is also expensive to
> measure MTBFs in this range so since you need to operate
> many units for a long time. 500.000h is just about 2% failure
> rate ber year.
MTBF wasnt even being discussed.
>>> Also a cheap USB dongle is technically "solid state"
>>> but does not match classical SSDs in reliability.
>> That comes with no surprise. Do you have any reliabilty info on "USB
>> dongles"?
> No. Just that I and people I know had failures on not too much used
> ones. I solved this by buing branded dongles where the brand name
> may have something to loose. I also copy everything important (like
> slides for a talk I am about to give) to two different dongles.