Hard drive prices - and innovation - decline

[Night Hawk]

Well as you know claims and actual real world wear and tear are two different things entirely. That's why manufacturers offer limited warranties. IT techs will replace an entire rack when a drive goes belly up since they maintain working backups.

 

[Barman58]

gary,

The most important word in the Phrase "Mean Time Between Failures" is "Mean" and it's not the definition of nastiness that applies , but the one of average - the Mean or average time a drive will fail may be 1000000 operations or even years, but there will be a spread of failure times, some may last 2000000 operations but some will last 1 operation. the mean is the most likely time it will last not a guaranteed minimum.

Someone with better math skills, (or more accurately memory, I did used to know the formula for mean ), will explain the actual math but the statistics are valid

 

[jimbo45]

Hi there
All I can say is in over 30 years of using HDD's I've probably only lost TWO and I can't really say if it was due to the HDD's themselves or whether it was due to a bit of "Abuse" on my part.

Laptop HDD's are another matter -- another good reason for fitting SSD's BTW -- since these are fairly small and not as robustly built as their desktop equivalents. A Laptop especially if it's used by several people can take a fair bit of rough handling or even things like coffee spillages etc. If you store data on a laptop HDD then backup becomes even MORE important.

Small external USB devices have the same problem as laptop HDD's (that's what they basically are- a laptop HDD in an enclosure) in that they are often carried around. In fact they can get even more robustly handled than a laptop since these can be stored with your baggage in an aircraft hold or just being carried in your pocket.

Mechanical devices also don't like being continually switched on and off so it's often better to leave computers idling rather than continually powering them on and off.

As SSD's become larger these problems will essentially disappear -- however I'll be keen to see the first CONSUMER 1PB disk. (1000GB = 1TB, 1000TB = 1PB - 1 Petabyte).

(MTBF is very commonly used in Engineering -- of course you can get ridiculous answers if you take a small sample -- say you have 4 Disks -- 1 fails after 1 Hour while 3 go on for 100 hours. In this case the MTBF becomes 301 / 4 hours = approx. 75 hours which quite clearly doesn't give a true performance of this batch of disks. However if you take a LARGE sample then the effect of ONE disk failing after 1 Hour hardly effects the sample and the result is reasonably reliable.

Tough luck if the bad disk slips through QC and you happen to get it.

Calculation is simple -- per disk add up the number of hours before it fails then divide the total number of hours in all the samples by the number of disks in the sample).

Cheers
jimbo

 

[Night Hawk]

Tell me about it! Like I was saying before one laptop seeing 3 drives two go in so many months like a drive every 3 is due to rough handling. But when the kids get their hands on things(or one of the parents hands it over that is) they lack the caution an adult will have like walking with the screen wide open with the laptop carrying the longest ac cord attached to the ac adapter you ever saw and boom down on a table without regard!

Another thing to know about drives in portables as to why they tend to have shorter lives even when "Handled With Care" often due to temps in the confined spaces when they are treated kindly by the owner and never bumped around. Besides the fragility the higher temps tend to see them wear out faster. But you pay for the "convenience" rather then durability.

Now about mean times anyone need a calculator by chance? MTBF Calculator 1.0 isn't specific to hard drives but for general calculations.

Here's one old eweek two page blog you may also want a look at dated 2007. Hard Disk MTBF: Flap or Farce?

That goes into mentioning the data sheets manufacturers have on drives which is brought up right at the top of the article there.

Data sheets for hard drives have always included a specification for reliability expressed in hours: commonly known as MTBF (mean time between failures), or sometimes the mean time to failure. Same difference: One way assumes that a drive will be fixed, and the other, replaced. Nowadays, this number is around a million hours for an "enterprise" hard drive. Some drives are rated at 1.5 million hours. Now, thats a good stretch to time. After all, a year is only 8,760 hours. One million hours comes to a bit more than 114 years. Some may be scratching their heads, since the hard drive itself has only been around for 50 years (IBMs giant 350 Disk Storage Unit for its RAMAC computer). This can be confusing. Instead, the MTBF is a statistical measure based on a calculation extrapolated from less-lengthy readings. It all means that drives are very reliable, with a failure rate well under 1 percent per year. Go Team Storage!

 

[jimbo45]

Hi there
You don't need an MTBF calculator -- nothing wrong in using a little bit of mental arithmetic. However even if you DO use a calculator you should at least have some idea of how the result is derived -- the reason is simple -- you should check that the expected result is at least within an order of magnitude of the result obtained with the calculator. Remember you have said yourself that devices DO fail - no reason to expect that the calculator won't fail at some stage -- or even that you've keyed in the data correctly !!!!.

Perhaps that's another reason why modern education seems to be so BAD almost everywhere these days -- nobody UNDERSTANDS what they are doing but rely totally on technology to do it all for them.

Obviously I don't decry the use of technology -- do we really want to wait 10 days to get a reference book from a Library when we can get the information in seconds via Google etc but I'd hate to rely on travelling in a new High Speed Train if the design Engineers hadn't any idea of how the thing worked, forces involved and in the Proof of Concept (POC) discussions where they could at least mentally check the physics / mechanics of the train they were designing and make sure it worked or was a viable concept.

(Actually it's not often the Youngsters fault -- but usually their teachers who often boast about being totally unable to understand any sort of technology and probably spend more time in various political demonstrations rather than trying to teach their pupils how some of this stuff works from first principles).

Cheers
jimbo

 

[Britton30]

Nigel, yeah I knew that mate. Mean is a figure or point such that there is an equal number lees and more, up or down, etc.

 

[King Arthur]

A "mean" is also known as an "average". For example, the mean of 1 and 5 is 3 and the average of 1 and 5 is 3 (1+5=6/2=3).

So if there were two HDDs, and one failed at 1 hour and another failed at 5 hours, their MTBF is 3 hours.

 

[Britton30]

Arithmetic mean - Wikipedia, the free encyclopedia

Average - Wikipedia, the free encyclopedia

I think median is what I should have said.
http://en.wikipedia.org/wiki/Median

 

[jimbo45]

A "mean" is also known as an "average". For example, the mean of 1 and 5 is 3 and the average of 1 and 5 is 3 (1+5=6/2=3).

So if there were two HDDs, and one failed at 1 hour and another failed at 5 hours, their MTBF is 3 hours.

Hi there
This is why these sorts of stats are meaningless on small samples.

However your calculation explanation could be a bit obscure. Mine I think is easier to follow.

For example you have 10 Disks that are tested say for 10 Hours

1 fails after 1 hr
2 fail after 8 hours
7 pass 10 hours without failure

The mean time before failure of this sample set is ( 1*1 + 2*8 + 7 *10) / 10
87 / 10 which is 8.7 hrs.

On the sample shown here that actually is a representative sample of the bunch.

However just to complicate things further - the fact that 7 have NOT failed within the test period means that the REAL mean time before failure could be MUCH longer.

You haven't always got enough time to test a whole batch of parts until they all fail so you set a target period and base it on that.

All we can guarantee from the statistical analysis shown is that we can reasonably SAY that the disks will work for AT LEAST 8.7 hours before failure -- which is the sort of calculation used for example in Aircraft parts. So the parts would be exchanged after every 8.7 hours regardless of whether they had failed or not.

(I've taught students basic Engineering -- and I'm always a HARD task master -- I question them when they present the result of statistical analysis as to what the results REALLY MEAN in practice -- but at least they leave the classes with a decent understanding !!!!).

@ anybody driving on some UK roads now -- you must have seen those sets of 'Specs speed cameras where they warn you with "Average Speed check" -- this is actually totally incorrect -- what they should say is MEDIAN speed check -- but that would clearly confuse the poor motorist.


Cheers
jimbo

 

[Gary]

WD has come out with a 4TB Hard Drive for $299.00 at NewEgg.

 

[jimbo45]

WD has come out with a 4TB Hard Drive for $299.00 at NewEgg.

Hi there

In Europe these can be had for around 125 - 140 EUR so the price is about the same.

What you MUST have to get any sort of decent performance out of these large spinners is A LARGE CACHE otherwise the system will forever be calculating the area / sectors of the disk to acces, wait for the drive to position to the correct physical address and then retrieve the data.

The SATA speed of course becomes important ONCE THE DATA HAS BEEN READ as it's then handled by the computer bus rather than the disk's hardware. Having a large cache means that disk accesses are much less as the data is already in storage, and with a decent "pre-fetching" algorithm the disk controller can read the next bunch of data that it thinks the user will want while the computer is doing something else.

So my advice to anybody thinking of investing in large capacity spinners is ALWAYS to pick the one with the largest Cache as well as having a decent spin rate - typically 7200 rpm or faster.

Cheers
jimbo

 

[Lady Fitzgerald]

...What about cases where it's not actually an HDD failure -- but say Nature -- Floods / Earthquakes / Volcanoes / Tsunamis / even simple traffic accidents where a truck crashes through your front wall and destroys your computer -- to say nothing of Burglars or even accidental data erasure.

Cheers
jimbo

Still your own fault for not having a reliable off-site back up scheme.

 

[Night Hawk]

WD has come out with a 4TB Hard Drive for $299.00 at NewEgg.

Hi there

In Europe these can be had for around 125 - 140 EUR so the price is about the same.

What you MUST have to get any sort of decent performance out of these large spinners is A LARGE CACHE otherwise the system will forever be calculating the area / sectors of the disk to acces, wait for the drive to position to the correct physical address and then retrieve the data.

The SATA speed of course becomes important ONCE THE DATA HAS BEEN READ as it's then handled by the computer bus rather than the disk's hardware. Having a large cache means that disk accesses are much less as the data is already in storage, and with a decent "pre-fetching" algorithm the disk controller can read the next bunch of data that it thinks the user will want while the computer is doing something else.

So my advice to anybody thinking of investing in large capacity spinners is ALWAYS to pick the one with the largest Cache as well as having a decent spin rate - typically 7200 rpm or faster.

Cheers
jimbo

Nothing truer said! When looking at the WD Black editions as well as the cache I didn't want those compared to the Blue and GP models seeing a 32mb cache when I could grab a 1tb drive with 64mb instead! That can be similar to some extent as how you would explain the differences between the 32bit and 64bit OS in that why is more ram taken with the 64bit like 2.5gb minimum rec over the 2gb rec for the 32bit?

The 64bit kernel preloads more data into the active ram for the most frequently used programs meaning less swapping out to the paging file created on the drive by the 32bit which relies more heavily on that. Likewise the larger the cache the more data is stored in that form of active memory providing a smoother operation as data is accessed faster. It makes for a more efficient startup of apps just the faster bus speed can utilize the data already preloaded without delays.

Obviously there is a vast difference between the cache and 64bit one being OS and the other hardware access times. But the similarities are evident once you "do the math" in that sense. With mean times on the other hand looking at the manufacturer's own white paper data sheets generally provides a more accurate assessment since these are reached under controlled circumstances the average user wouldn't see like having a batch of 1,000 drives to work with out of let's say a model line of 100,000 or more drives on a large production scale.

You can't sample each but gain the average expectancy from extensive testing with a handful out of a product line. Once those data sheets are confirmed a company will then know what to expect when something goes into mass production. The 3-5yr. limited warranties you see for various makes and models only reflect what the manufacturer already knows. And with anything mass produced you know there will a few "ut ohs" along the way.

 

[jimbo45]

...What about cases where it's not actually an HDD failure -- but say Nature -- Floods / Earthquakes / Volcanoes / Tsunamis / even simple traffic accidents where a truck crashes through your front wall and destroys your computer -- to say nothing of Burglars or even accidental data erasure.

Cheers
jimbo

Still your own fault for not having a reliable off-site back up scheme.

Hi there
Absolutely --there's no such thing as "too much backup".
I was answering primarily answering one of those people who were saying that the HDD's should NEVER fail so why should I take backup.

The above cases were for the situation assuming the HDD's NEVER failed there's still scenarios where you would need to recover your data from backups !!!!.

Cheers
jimbo

 

[Britton30]

Even without a HDD or SSD failure backups are needed, anything can happen, and will if there's no backup. Windows can decide to quit working for no obvious reason at times. Mine has been fine since I started making backup images though.

 

[Lady Fitzgerald]

Even without a HDD or SSD failure backups are needed, anything can happen, and will if there's no backup. Windows can decide to quit working for no obvious reason at times. Mine has been fine since I started making backup images though.

It's like rain and umbrellas. If you have an umbrella (bumbershoot for those on the east side of the pond) with ypu, it won't rain. Forget the darned thing and it is guaranteed to rain.

 

[Night Hawk]

...What about cases where it's not actually an HDD failure -- but say Nature -- Floods / Earthquakes / Volcanoes / Tsunamis / even simple traffic accidents where a truck crashes through your front wall and destroys your computer -- to say nothing of Burglars or even accidental data erasure.

Cheers
jimbo

Still your own fault for not having a reliable off-site back up scheme.

Hi there
Absolutely --there's no such thing as "too much backup".
I was answering primarily answering one of those people who were saying that the HDD's should NEVER fail so why should I take backup.

The above cases were for the situation assuming the HDD's NEVER failed there's still scenarios where you would need to recover your data from backups !!!!.

Cheers
jimbo

I didn't think you could have too much backup while you may use up a lot of drive space or removable media at times. Hard drives never failing? HAR! That's a joke! You can have a brand new as well as old clunker drive decide to give you a major headache!

Brand new drive comes in with bad sectors or something else goes and without any backup(s)... oops! "having a bad hair day" is someone learning a valuable lesson by an unfortunate mishap with hardware as well as not having any foresight about the countless problems that might come up.

Even without a HDD or SSD failure backups are needed, anything can happen, and will if there's no backup. Windows can decide to quit working for no obvious reason at times. Mine has been fine since I started making backup images though.

It's like rain and umbrellas. If you have an umbrella (bumbershoot for those on the east side of the pond) with ypu, it won't rain. Forget the darned thing and it is guaranteed to rain.

That's one of those: "gee if I had only...." type of days when you seem to be all prepared with every precaution and ? nothing happens! But as soon as you turn around and don't tend to some type of preparation all H breaks loose and you get caught between a rock and a hard place!

 

[Lady Fitzgerald]

...That's one of those: "gee if I had only...." type of days when you seem to be all prepared with every precaution and ? nothing happens! But as soon as you turn around and don't tend to some type of preparation all H breaks loose and you get caught between a rock and a hard place!

Ok, I'm getting old. The first time I read that, I saw Preparation H.

 

[Britton30]

...That's one of those: "gee if I had only...." type of days when you seem to be all prepared with every precaution and ? nothing happens! But as soon as you turn around and don't tend to some type of preparation all H breaks loose and you get caught between a rock and a hard place!

Ok, I'm getting old. The first time I read that, I saw Preparation H.

Me too. It's appropriate, computers, or building them, can be a PITA.

 

[Night Hawk]

You can get "slam dunked" fast when putting everything into a system and lose it all from not using the old noggin ahead of time. Whether drive or some hardware fails you can be stuck without some form of backups accessible if needed from another machine. That would apply to removable media and usb flash drives where surprisingly people carry photos and other file types like work around with them.

As for having one external drive to keep everything on as for laptop users without any other means of backing things up? that can easily turn into a "recipe for dsaster" if the drive fails or you have an oops and bump into it and awwwwayyy she goessss.. cccrasshhhh... and you out of a drive! Even with extreme care having the backup of the backup for the... may just save the day.

Besides second drives like externals another you can never fully count on for backups is removable media like data disk you burn. Once you get into the next optical drive or into a new system premade or other the disks can be unreadable at times! I've tossed plenty fo cd-rs and several dvd-rs used for backups of folders for various file types often enough to never completely rely on optical either. Spare drive to back up another secondary drive to back up...