Enterprise versus "consumer" grade drives

RVS, RAFF, dual stage actuator.

"miso" wrote:

RVS, RAFF, dual stage actuator.

Since you were including consumer-grade HDDs in your query, RVS is a
non-issue. Are you really expecting to repeatedly smack your HDDs while
they are spinning? Just how many HDDs are you putting inside the same
cabinet? You didn't specify all your criteria; however, you *did* want
a comparison to consumer-grade drives which apparently *was* a choice
for you (i.e., they met your selection criteria) so such technologies,
like RVS, are a non-issue in comparing enterprise to consumer HDDs. If
RVS was a selection criteria then you already know that consumer HDDs
aren't going to be among your choices.

http://enterprise.media.seagate.com/...vs-enterprise/
http://www.hgst.com/tech/techlib.nsf/techdocs/50d6c79f1e3f024b87256c470074569d/$file/wp_rvs_25march.pdf
Written 5 years, or more, ago.

I had mistakeningly assume that you wanted a comparison between
enterprise and consumer HDDs because you actually had some consideration
in purchasing consumer-grade HDDs. If you're going to mandate features
of enterprise-grade HDDs then, well, there's no point in considering
consumer-grade ones. You give no clue on how YOU will actually deploy
the HDDs. For all we know, you're asking for use in your home desktop.

On 10/9/2012 4:13 PM, VanguardLH wrote:
"miso" wrote:

RVS, RAFF, dual stage actuator.

Since you were including consumer-grade HDDs in your query, RVS is a
non-issue. Are you really expecting to repeatedly smack your HDDs while
they are spinning? Just how many HDDs are you putting inside the same
cabinet? You didn't specify all your criteria; however, you *did* want
a comparison to consumer-grade drives which apparently *was* a choice
for you (i.e., they met your selection criteria) so such technologies,
like RVS, are a non-issue in comparing enterprise to consumer HDDs. If
RVS was a selection criteria then you already know that consumer HDDs
aren't going to be among your choices.

http://enterprise.media.seagate.com/...vs-enterprise/
http://www.hgst.com/tech/techlib.nsf/techdocs/50d6c79f1e3f024b87256c470074569d/$file/wp_rvs_25march.pdf
Written 5 years, or more, ago.

I had mistakeningly assume that you wanted a comparison between
enterprise and consumer HDDs because you actually had some consideration
in purchasing consumer-grade HDDs. If you're going to mandate features
of enterprise-grade HDDs then, well, there's no point in considering
consumer-grade ones. You give no clue on how YOU will actually deploy
the HDDs. For all we know, you're asking for use in your home desktop.


I over build my PCs, then run them about 5 years. I don't go cheap, but
I don't get stupid about this either. Thus I am determining what the
marginal increase in money gets me in reliability.

You must have missed this statement:
"So is there any advantage to buying a Seagate Constellation versus a
Baracuda? Or Ultrastar versus Deskstar?"

Since I was referring to both consumer and enterprise drives, to those
that can follow logic, it can be inferred that I was also considering
consumer drives.
QED

"miso" wrote:

On 10/9/2012 4:13 PM, VanguardLH wrote:
"miso" wrote:

RVS, RAFF, dual stage actuator.

Since you were including consumer-grade HDDs in your query, RVS is a
non-issue. Are you really expecting to repeatedly smack your HDDs while
they are spinning? Just how many HDDs are you putting inside the same
cabinet? You didn't specify all your criteria; however, you *did* want
a comparison to consumer-grade drives which apparently *was* a choice
for you (i.e., they met your selection criteria) so such technologies,
like RVS, are a non-issue in comparing enterprise to consumer HDDs. If
RVS was a selection criteria then you already know that consumer HDDs
aren't going to be among your choices.

http://enterprise.media.seagate.com/...vs-enterprise/
http://www.hgst.com/tech/techlib.nsf/techdocs/50d6c79f1e3f024b87256c470074569d/$file/wp_rvs_25march.pdf
Written 5 years, or more, ago.

I had mistakeningly assume that you wanted a comparison between
enterprise and consumer HDDs because you actually had some consideration
in purchasing consumer-grade HDDs. If you're going to mandate features
of enterprise-grade HDDs then, well, there's no point in considering
consumer-grade ones. You give no clue on how YOU will actually deploy
the HDDs. For all we know, you're asking for use in your home desktop.


I over build my PCs, then run them about 5 years. I don't go cheap, but
I don't get stupid about this either. Thus I am determining what the
marginal increase in money gets me in reliability.

Seagate Constellation versus a Baracuda?

http://www.tomshardware.com/charts/2...5B3017%5D =on

Nothing knocked me out of my chair looking at the benchmark comparison
on those 2 models. Both of those are 7200 RPM drives. There are 15K
RPM Barracuda drives but you never mention WHICH ones you were
comparing. If you want to make an equal comparison between benchmarks
for a 7200 RPM versus another 7200 RPM drive then your choice of Hitachi
models (next) is flawed. You'd have to compare the Seagate
Constellation 7200 RPM against their Cheetah 15K RPM models, like:

Constellation:
http://www.tomshardware.com/charts/2...d%5B4056%5D=on
Cheetah:
http://www.tomshardware.com/charts/e...d%5B2035%5D=on

Alas, they don't use the same set of benchmarks for the enterprise
drives so it's hard to tell what equates to what for benchmark type.

Or [Hitachi] Ultrastar versus Deskstar?"

Couldn't find a page where the benchmarks for both the Ultrastar and
Desktar were listed together (so I could select them and do a
side-by-side compare of benchmarks). You'll have to bounce between the
two charts to compare benchmarks. I picked units with the same cache
size (16MB) which is the largest listed in the chart at that time for
the Ultrastar models; however, the Deskstar has more than twice the
capacity (1TB versus 450GB).

Ultrastar:
http://www.tomshardware.com/charts/e...d%5B2034%5D=on
Deskstar:
http://www.tomshardware.com/charts/2...d%5B2358%5D=on

A big reason the benchmarks are much better for the Ultrastar is you're
comparing a 15K RPM drive to a 7200 RPM drive. It's not anything
"enterprise" for which you're paying a price premium. It's a faster
spinning platter (and noisier drive) that you're paying more for.

If, like for Hitachi, you were to compare Seagate 7200 and 15K RPM,
well, one spins more than twice as much in the same time so there is the
wear factor to consider.

A decade ago I built a PC for a 3-year lifespan because there was enough
ongoing change to warrant a significant enough performance increase (at
least double) to expect a decent replacement at that time. I've been
disappointed with the slow technology changes since then and now
[over]build with a 6-year lifespan in mind - except for mechanical
storage (e.g., HDDs, optical). It's a waste of money. After 3 years,
and regardless that the HDD is in perfect working order, you'll want
more capacity or some tech gotcha will get you itching to get a newer
HDD (like how SSDs are now starting to draw consumers). I might pay
extra for higher performance (if willing to sacrifice on noise and
longevity) but paying for enterprise features when NOT used in an
enterprise scenario doesn't make sense. The problem with using bleeding
edge technology components in your build is that you pay a hell of a lot
more, get little or just a tiny incremental performance gain, and in a
couple years your disappointed with your expense because prices have
radically dropped or something far better came along.

Even with enterprise-grade HDDs having a longer warranty period (maybe 5
years), do you really send in your in-warranty HDD for replacement?
You've only got one that blew, not dozens or hundreds at some RAID-ed
file server datacenter waiting for replacement. Are you really going to
disable your host waiting for an in-warranty replacement? I never found
it cost effective or wasn't willing to stall usability to get a warranty
replacement of an HDD. I just went and got a new one. If you have the
money to spend for an enterprise-grade HDD (just one for your home PC),
are you really going to concern yourself about saving a little to get it
replaced by warranty instead of buying a bigger, faster, or otherwise
better HDD some years from now?

The survivability of the HDD mechanicals, as mentioned in the articles,
is not greater with enterprise-grade HDDs versus consumer-grade HDDs.
They all appear to have the same infant mortality rate along with the
3-4 year peak in failure rate. Unless you can actually use the
enterprise added features, I don't see why you'd waste the money on
them. Rather spend the extra money on a faster HDD. Talk to the forums
on whether they consider the 15K RPM drives as noisier or not. I can't
see how a higher-energy 15K RPM drive is going to outlast, in general, a
7200 RPM drive considering there's twice as much wear for 15K RPM on the
bearing and they run hotter. Also, consider if you want to spend more
money on a 15K RPM drive which will open your programs faster (but
doesn't have much affect thereafter) and copy files faster but is that
the primary focus of your computer? You'll also end up having to
sacrifice drive capacity for higher platter RPM: the faster spinning
platters have more wind resistance so the size is reduced to lower
turbulence while not having to increase motor torque hence why makers
are considering using helium inside the drive. See:

http://arstechnica.com/information-t...er-capacities/

You don't think they'll charge a price premium for helium-filled HDDs?
Will you need it? Not unless you're trying to up the single-unit
capacity of a 15K RPM drive because helium will let them increase
platter size to up the drive's capacity. The rest of us will prefer a
more economical pricing point in the dollar/byte curve and just add
another "old" HDD in the case and for cheaper.

When did I mention 15k RPM drives or even 10K? Nor did I mention helium.
Why look at 1Tbyte drives?

If anything, I would consider examining a Ultrastar 3T versus a deskstar
3T. The 4T drive prices are not at the sweet spot. The ultrastar has a
2e6 MTBF, and yes that isn't the same as AFR, but it is indicative of
something.

Regarding noise, the deskstar and ultrastar drives are the same. Not so
for WD RE versus red. In fact, the WD RE is one of the noisier drives.

I've had 10krpm cheetah drives in the past. I'm not going that route.
I'm going to SSD the OS and then use magnetic media for file storage.
The C216 mobos have two sata3 and 4 Sata 2. It is not a particularly
generous mobo in that respect. If you mirror the drive, that means the
mobo will support one SSD for OS, and two "drives" for user data.

I may look at getting two smaller SSDs and stripe them. I'm not a fan of
striped raid per se (no redundancy), but the SSD array would be so small
that I could easily ghost it often. I'd have to determine if there is
any speed advantage to striped SSD. It just might end up choking the
system further down the lane.

"miso" wrote:

When did I mention 15k RPM drives or even 10K?

Hitachi Ultrastar. That's a brand and a family line, not a particular
model in that family line. So, for example:

http://www.newegg.com/Product/Produc...82E16822145330

Yep, it qualifies as an Ultrastar. You didn't mention specific models
so all models in a family are up for consideration.

Nor did I mention helium.

Did I say YOU mentioned helium? *I* mentioned it. That's because it
will allow larger platters to spin faster (so teh 15K RPM drives could
have larger platters to up their capacity). It will have a cost
premium. Will YOU pay for it just because it exists or get something
cheaper that does the job for you? This was mentioned because it
follows the same [il]logic from having to pay for enterprise features
that you won't use.

Why look at 1Tbyte drives?

Because, again, you never bothered to mention specifics, like size, so
all sizes are eligible.

Regarding noise, the deskstar and ultrastar drives are the same.

For the same RPM models, that is.

I'm going to SSD the OS and then use magnetic media for file storage.

As long as you realize that SSDs are fast for reads but slower for
writes; however, writes on SSDs are still much faster than for HDDs.
Just be aware that writing is slower than reading on an SSD so consider
how you will use the SSD. For example, consider if your host is used
primarily as a read-based web server or a write-heavy database server.
Apps that do sequential reads benefit less from SSDs than apps that do
random writes. Under heavy write volume, the SSD will slow even further
due to the overhead of the wear levelling mechanism. Over time, the SSD
slows more due to the remapping needed to compensate for oxide stress
that causes cell failure and why SSDs are rated by maximum write cycles.
Flash memory cells wear out; quicker for MLC than for SLC, so SLC is
more durable but far more expensive. When the reserve (remap) space
gets consumed, the device catastrophically fails. For SLC-based
(enterprise-grade, very pricey) SSDs, you can figure a lifespan of about
5 years on a write-heavy setup; see http://tinyurl.com/c9aevxk. Most
SSD makers like to claim their SSD should be usable for a 10 year
lifespan because they're targeting home PC users with their statement
since those users are not stressing their PC all hours of every day and
typically incur a very light load when they do use their PC. The makers
are hoping users don't realize that [the rest of] the 10-year old PC may
very well continue to be very usable after 5 or 10 years.

Once an HDD survives past its infact mortality and 3-year humps in
failure rates, they last a long time but typically users replace them
long before they fail because they want more capacity or a faster HDD.
The device didn't fail but users later change their requirements or
wants. Same for SSDs. While they may have a 5 or 10 year lifespan
depending on how they are used and what type (MLC vs SLC), it's likely
most users will replace the SSD long before it fails because they want
more capacity or a faster SSD. Consumers typically don't wait for [a
hint of] failure before buying something newer, better, bigger, or
improved.

miso wrote:
[...]
So is there any advantage to buying a Seagate Constellation versus a
Baracuda? Or Ultrastar versus Deskstar?
No. Unless you have a hard limit to only use one drive. "Enterprise"
drives are a bit better in reliability, but RAID1 is so massively
better that enterprise drives are laughable. The only place these
pay of (somewhat) is if drive replacement is expensive, e.g. because
somebody has dro drive to the datacenter. There, even a small
imporivement in reliability can justify a larger price increase.

[...]
Given that the OS with be on SSD and the magnetic media is on RAID 0,
would it still make sense to go with enterprise grade drives, presuming
they are more reliable that the consumer grade?

If you use RAID0, your data is basically doomed anyways, unless
you have good backup. If you have good backup, no need to go for
enterprise drives. RAID0 is basically always a bad choice except
for cache and buffer applications that need high throughput, such
as video-capture and editing. But RAID0 should never be used as
actual longer-term storage.

Arno
--
Arno Wagner, Dr. sc. techn., Dipl. Inform., CISSP -- Email:
GnuPG: ID: 1E25338F FP: 0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
----
Cuddly UI's are the manifestation of wishful thinking. -- Dylan Evans

"Arno" wrote:

miso wrote:

Given that the OS with be on SSD and the magnetic media is on RAID 0,
would it still make sense to go with enterprise grade drives, presuming
they are more reliable that the consumer grade?

If you use RAID0, your data is basically doomed anyways, unless
you have good backup. If you have good backup, no need to go for
enterprise drives. RAID0 is basically always a bad choice except
for cache and buffer applications that need high throughput, such
as video-capture and editing. But RAID0 should never be used as
actual longer-term storage.

Another problem with stripping (RAID-0) is that MTBF goes down as you
add more drives. As the drive count goes up, MTBF goes down. The
minimum RAID-0 config uses 2 HDDs which means your MTBF is half of the
whichever HDD has the lowest MTBF.

http://en.wikipedia.org/wiki/Standar...0_failure_rate

If you can afford 2 HDDs for RAID-0, and especially after choosing NOT
to waste money on "enterprise" HDDs for a personal-use host, you can
probably afford 3 HDDs even if they are smaller but result in an equal
or increase in total capacity compared to the 2 HDDs. Then you could
put them in a RAID-5 configuration.

It makes no sense to be spending money on enterprise-grade HDDs for
features you won't use but then go cheap on the RAID config. Forget the
"enterprise" lure and go with a better RAID config.

On 13/10/12 08:27, VanguardLH wrote:
"Arno" wrote:

miso wrote:

Given that the OS with be on SSD and the magnetic media is on RAID 0,
would it still make sense to go with enterprise grade drives, presuming
they are more reliable that the consumer grade?

If you use RAID0, your data is basically doomed anyways, unless
you have good backup. If you have good backup, no need to go for
enterprise drives. RAID0 is basically always a bad choice except
for cache and buffer applications that need high throughput, such
as video-capture and editing. But RAID0 should never be used as
actual longer-term storage.

It's okay to use RAID0 on top of other raids - such as a RAID0 stripe of
RAID1 pairs. But RAID0 on its own is seldom a good idea.


Another problem with stripping (RAID-0) is that MTBF goes down as you
add more drives. As the drive count goes up, MTBF goes down. The
minimum RAID-0 config uses 2 HDDs which means your MTBF is half of the
whichever HDD has the lowest MTBF.

http://en.wikipedia.org/wiki/Standar...0_failure_rate

If you can afford 2 HDDs for RAID-0, and especially after choosing NOT
to waste money on "enterprise" HDDs for a personal-use host, you can
probably afford 3 HDDs even if they are smaller but result in an equal
or increase in total capacity compared to the 2 HDDs. Then you could
put them in a RAID-5 configuration.


If you are using Linux, you can use its RAID10 modes to get RAID1 safety
and better than RAID0 performance from hard disks.

It makes no sense to be spending money on enterprise-grade HDDs for
features you won't use but then go cheap on the RAID config. Forget the
"enterprise" lure and go with a better RAID config.


Absolutely.

On 10/12/2012 9:58 PM, Arno wrote:
miso wrote:
[...]
So is there any advantage to buying a Seagate Constellation versus a
Baracuda? Or Ultrastar versus Deskstar?
No. Unless you have a hard limit to only use one drive. "Enterprise"
drives are a bit better in reliability, but RAID1 is so massively
better that enterprise drives are laughable. The only place these
pay of (somewhat) is if drive replacement is expensive, e.g. because
somebody has dro drive to the datacenter. There, even a small
imporivement in reliability can justify a larger price increase.

[...]
Given that the OS with be on SSD and the magnetic media is on RAID 0,
would it still make sense to go with enterprise grade drives, presuming
they are more reliable that the consumer grade?

If you use RAID0, your data is basically doomed anyways, unless
you have good backup. If you have good backup, no need to go for
enterprise drives. RAID0 is basically always a bad choice except
for cache and buffer applications that need high throughput, such
as video-capture and editing. But RAID0 should never be used as
actual longer-term storage.

Arno


I meant mirror, i.e. raid 1, for the hard drives.

I'll consider RAID5. I've done that before. Not with the greatest
results though. When the mobo died, I was able to put the RAID 10 array
in another PC and all the data was still there. The RAID5 array couldn't
be recovered. I had a pretty good backup.

Raid 5 at least buys you something in the way of more storage with just
as good security, well provided the mobo doesn't get Chinese cap disease.