Will we EVER see 10kRPM ATA drives?!

[ElDonAntonio]

Originally posted by: BD231

Originally posted by: ElDonAntonio
I agree that the HD is the damn slowest thing in my computer and that it bugs the heck out of me. Ever had ~10 programs open? try to switch from one app to the other, and you're certainly going to wait a good dozen of seconds while Windows slaps around the swap file. Same thing when opening/closing a few programs at a time, scanning the drive, running an antivirus, etc.

Those of you saying "I don't care waiting 8ms while Quake3 is loading" obviously don't understand what the hard drive is doing besides storing Quake3.

I dont thinka 10,000rpm drive will compinsate for all of that. I'd much rather wait for solid sate storage because the amount of heat computers are starting to put out is ridiculous, and a 10,000rpm drive will make the heat factor even worse.

A 10k drive won't make everything instantaneous but it will still help nicely. And I think you shouldn't hold your breath for solid state storage amigo
The heat factor isn't that bad, a lot of machines use 10k SCSI drives without any problems.

 

[BD231]

Originally posted by: ElDonAntonio

Originally posted by: BD231

Originally posted by: ElDonAntonio
I agree that the HD is the damn slowest thing in my computer and that it bugs the heck out of me. Ever had ~10 programs open? try to switch from one app to the other, and you're certainly going to wait a good dozen of seconds while Windows slaps around the swap file. Same thing when opening/closing a few programs at a time, scanning the drive, running an antivirus, etc.

Those of you saying "I don't care waiting 8ms while Quake3 is loading" obviously don't understand what the hard drive is doing besides storing Quake3.

I dont thinka 10,000rpm drive will compinsate for all of that. I'd much rather wait for solid sate storage because the amount of heat computers are starting to put out is ridiculous, and a 10,000rpm drive will make the heat factor even worse.

A 10k drive won't make everything instantaneous but it will still help nicely. And I think you shouldn't hold your breath for solid state storage amigo
The heat factor isn't that bad, a lot of machines use 10k SCSI drives without any problems.

<~~~drools- Solid state storage



I know we wont see it for quite a while, but I can dream . I really hate the way you have to sacrifice performance for silence and less heat. My true dream is a computer that runs as cool and in audible as my old K6-2+ machine, but that's as fast as the fastest P4 out there . I had a Atlas 10kIII and it was way loud, it was actually the loudest part in my computer. It even dround out the noise coming from my retail athlon XP fan . It was quite fast, but no where near worth the extra noise.

 

[GoSharks]

you will see 10k drives come into the ide/sata market when scsi moves to the next faster spindle speed (from 15k). 10k scsi is still the major money maker sector for the hard drive manufactures, just as 7200rpm scsi used to be. then 10k came out, and eventually everybody went to that, and ide moved up to 7200. i bet we will see the same thing in a few years... probably 2-4 if you ask me

 

[Leokor]

Originally posted by: dullard
2) The 7200 rpm Western Digital drives with 8 MB buffer run most things faster than almost all 10K SCSI drives. Thus spindle speed is just one part of speed. Performance can be increased by much cheaper means. SCSI isn't all that it is cracked up to be - I switched from SCSI to faster IDE drives (even though the spindle speed dropped) and am completely happy with the results.

A minor correction. That performance comparison holds for many desktop applications, where data locality aspect is high (that is, subsequent seeks are likely to land fairly close to each other). In server applications, where the heads fly all over the drive, SCSI drives with better seek times leave all ATA drives far behind, including those with 8 MB buffer. In addition to the seek times, another factor is that the SCSI interface optimizes the queues, while an ATA drive will always perform the commands in the order they came.

Leo

 

[Xentropy]

Check out StorageReview.com, BD. The 15k.3 drives from Seagate are as quiet and cool-running as 7200rpm drives. Your experience with an earlier generation 10k drive isn't the norm anymore. Taking a *recent* generation 10kRPM drive and slapping an IDE interface on it wouldn't mean higher noise or temps at all, just faster access at a higher cost per gigabyte, a higher cost I'm willing to spend as long as the extra cost of a decent SCSI controller doesn't come with it.

Yes, solid state would be great, but I'm trying to be a realist instead of an idealist. Solid state is just the very highest end there is, far above SCSI both in performance AND price. I'd love to have the money to RAID together a bunch of 8G solid state drives to get 64G or something, but if I had that kind of money, I'd rather buy a Mercedes S-class with 100% cash down--at cheaper than that dream 64GB drive array ;>

That's not to say that technology isn't in the works that could bring solid state hard drives into the mainstream (PFRAM, or polymer memory, for example, is non-volatile, provides approx 1us read and 50us write access, is infinitely rewritable, is extremely dense, extremely cheap per bit, is transistorless, and can be stacked on top of existing silicon to form large "L3 caches"), but since it's still at the very beginning stages of development, it won't hit the mainstream for at least half a decade. Still, the eventuality of PFRAM would be, "terabyte solid state devices with extremely low transistor counts in 'drives' the size of a match book."

So anyway, I digress... I'm not looking for super-high-end to come down to earth anytime soon, but I see a recently-formed vacuum in the midrange that's just begging to be filled by 10kRPM SATA drives. Since there's no TECHNICAL reason why even 15kRPM can't move to ATA, asking for 10kRPM to do so isn't particularly crazy of me.

 

[Leokor]

Originally posted by: Xentropy
Thanks, that was my main question. Is there any *technological*reason why not. And the answer is no, as I suspected. I can think of plenty of commercial reasons why not, but can refute those as well, which I'll do below.

Yup, there is no technological reason why not. As a matter of fact, there had been precedents. Quantum Fireball LM (though not Fireball AS) was the ATA reincarnation of an older SCSI drive.

Leo

 

[McCarthy]

I still stand against heat and noise and am for reliability.... and, well, sure economy

You guys forgot to include the economy part when you quoted me. Those nice new cool and quiet Cheetahs are running what, a bit over $300 for 36 gigs? Our platter densities would go down with a speed increase which would mean more platters to get the same amount of space as we have with 7200's, more heat and noise and expense for the same capacity. I'm plum happy with my 200gigs of nice cool quiet cheap and snappy 7200rpm drives here, with one being a WD SE 8meg cache which is still the change I'd favor seeing better exploited before we "rev it up some more!" For the video and audio editing I've been doing my bottleneck of my drives still doesn't hold a candle to the hours of CPU time required.

The part I never understand about this topic when it comes up is the people saying they'd be glad to pay more for a 10K IDE drive, but if you say "go to scsi, it's where you want to be right now" they say it's too expensive. What, they think the HD companies are going to develope 10K drives for charity's sake, the same companies these people accuse of ripoff prices for scsi? But then relating to what I said above, it seems this topic has more to do with people wishing for unicorns than anything. When will we see 10K drives? When the cost premium of manufacturing them is low enough that they'll be marketable. Not before.

In the end, I stand by what I've said on the topic previously. First and foremost I want my drives to be reliable. If the drive requires active cooling, it'd better damn well say so right on the drive in red letters, not show up in a flurry of RMAs later like seems to be the case with the IBM 75GXPs.

Leo V - just think how quiet a 8086 system was....though that'd kinda ruin your current experiments and fun. Gonna show us your new videocard fanless? That is one benefit of heat and noise, makes for good articles to read about combating it.

 

[JHeiderman]

Does anyone remember a couple months ago HardOCP had a link on their front page to a case at Newegg that in it's description listed a feature of the harddrive bay being manufactured to the 10,000rpm IDE harddrive spec because it acted as a heatsink. What ever happened with that?

- J

 

[Athlon4all]

Well, there certainly needs to be some breakthroughs in terms of drive techjnolgoy that would allow these to be 10K RPM Consumer level ATA drives to be quiet and cool and reach the same capacities of current high-end 7200RPM IDE drives. This is a major factor, but another issue is that, these drive makers definately make a vbery good amount of money on their 10 and 15K SCSI drives and the minute they add 10K ATA Drives, many people will start to take them over 10K SCSI's cause, while I will certainly admit that SCSI does have its advantages still over ATA, ATA and SCSI are very even in single drive situations (when the drives are the same)

 

[Xentropy]

Originally posted by: McCarthy

I still stand against heat and noise and am for reliability.... and, well, sure economy

You guys forgot to include the economy part when you quoted me. Those nice new cool and quiet Cheetahs are running what, a bit over $300 for 36 gigs? Our platter densities would go down with a speed increase which would mean more platters to get the same amount of space as we have with 7200's, more heat and noise and expense for the same capacity. I'm plum happy with my 200gigs of nice cool quiet cheap and snappy 7200rpm drives here, with one being a WD SE 8meg cache which is still the change I'd favor seeing better exploited before we "rev it up some more!" For the video and audio editing I've been doing my bottleneck of my drives still doesn't hold a candle to the hours of CPU time required.

When an increased cache size can decrease access times to 5ms and lower, I'll be quite happy to use the miracle 7200RPM drive that'll do it. As you said yourself, for video/audio editing, even a "mere" 7200RPM drive will handle things from a throughput standpoint. My issue is seek speeds, and I know of no way other than faster spindles to improve seek. If you look at "real world" benchmarks, seek's what makes the difference, not throughput (past a certain point; obviously too-low throughput will hurt you, but you'd have to drop to 5400rpm to get there).

And once again I'll point out www.storagereview.com for proof that even 15000RPM can be attained with sound and heat in the same range as 7200RPM drives. The 73GB 15K.3 is 0.1dB louder than the 80GB WD drive w/ 8M cache you tout above (and 2dB *quieter* than the 120GB version of the WD drive). Runs hotter, but still well under the range where active cooling would be necessary. Therefore, I consider those sound/heat "issues" completely moot. Technology has overcome them already.

More EXPENSE per capacity exists, yes, but I'm willing to pay more per gigabyte for these drives, PROVIDED I don't also have to pay for another redundant controller all because they decided to solder on a SCSI instead of IDE chip and make the drive not work with the controller I've got. (More on this point below.)

The part I never understand about this topic when it comes up is the people saying they'd be glad to pay more for a 10K IDE drive, but if you say "go to scsi, it's where you want to be right now" they say it's too expensive. What, they think the HD companies are going to develope 10K drives for charity's sake, the same companies these people accuse of ripoff prices for scsi?

No, I think the HD companies will sell the DRIVES for the same price, or only slightly less. But even if they charged 10% *more* for the ATA versions, it'd be a cost SAVINGS for the end user, because they could use the controller built right into their motherboard instead of having to, after just coating the suits at Seagate in 24K gold, then turn around and do the same for the execs at Adaptec. It's the controller costs I'm really concerned with, and (now it's my turn to play idealist, though not quite as far out there as solid state) would like to see SCSI completely vanish, replaced altogether by ATA/SATA. EVERY new drive should be SATA by this time next year, period. There is *no* reason why that couldn't happen. There'd be SATA drives ranging from 7200RPM to 15kRPM, at every range of price/performance points, every amount of noise/heat from large to small, the entire gamut, ALL RUNNING ON THE CONTROLLERS EVERYONE ALREADY HAS. But this doesn't happen, because somehow "SCSI" is supposed to be "better", so the drives are made better for it instead of bringing the same drive tech to ATA and widening the range of the market.

In the end, I stand by what I've said on the topic previously. First and foremost I want my drives to be reliable. If the drive requires active cooling, it'd better damn well say so right on the drive in red letters, not show up in a flurry of RMAs later like seems to be the case with the IBM 75GXPs.

Good point there. ATA drives are less reliable than SCSI on the whole as well. Either every motherboard should have SCSI built-in and ATA should just go away, or SCSI should go away and every drive use ATA/SATA. There is *no* reason why this couldn't happen other than the greed of the manufacturers driving the status quo.

Those of you who say 10kRPM shouldn't come to ATA act as if 10kRPM would replace every 7200RPM solution out there, and that's simply not true. It'd just be another OPTION. Those of you who couldn't stand 45.1dB 15kRPM drive instead of a 45.0dB 7200RPM drive could stick with your lower performing 7200RPM and save a few bucks, too. I'm not saying that everyone should own 10kRPM ATA drives, I'm saying they should EXIST. Not sure why such an argument is being responded to so defensively, as if the new product would be shoved down your throats. My argument is for more choices in the marketplace, not less.

 

[mastertech01]

I just cant see how the average Joe needs 200-300 GB of drive space on their main array. I find it far more satisfying to have the speed and dependability of SCSI. Storage has many mediums and that is where I belive ecomony should be considered. But on your main performing data subsystem, why settle for less if the rest of your system is the fastest money can buy. And Im no spend thrift, I get a lot of my gear used off the forums. IMHO, Id rather have a used 36LP Cheetah 15K than a couple brand new ATA drives. And my Raid controller I got half price by diligently watching Ebay for months. And it carries over from setup to setup. Initial costs are substantially higher than ATA, but then Ive never had a quality SCSI drive die on me...which I honestly cannot say for the best ATA drives. But of course this is all my personal experience and preferrence and I wouldnt say your ideas of the best setup are any less valid.

 

[CQuinn]

Originally posted by: Xentropy

Originally posted by: Leokor
Technically speaking, any drive can be equipped with either the SCSI or the ATA interface and still have the same internal mechanics. Therefore, 10K ATA drives are certainly possible (and 15K ATA drives, too, by the way). Just strip their respective SCSI twins of the extra SCSI chip.

Thanks, that was my main question. Is there any *technological*reason why not. And the answer is no, as I suspected. I can think of plenty of commercial reasons why not, but can refute those as well, which I'll do below.

Actually that is not quite correct. If you are thinking that something like the WD1200JB and the Cheetah 15K, or even the
Atlas 10k III have the "same" internal mechanics, then you are mistaken. Ironically SCSI designs tend towards speed and
reliability, while the IDE market has moved forward more on capacity and low cost. Both can be considered to follow the
same ideas, technically speaking. But thats like saying my Jeep Cherokee is the same as a Lan Evo because they are
both four-wheel drive vehicles.

The Quantum Fireball LM was the same mechanical design for its SCSI and IDE offerings, this was also back in the day
when 7200 RPM drives were considered the leading edge even for SCSI. IIRC, the Fireball LM was also a short lived
drive, because it cost them more to manufacture than other IDE drives being offered at the time.

Second point, an increased cache size will not bring access time down, period. All increasing the cache does is give you
a better chance to maximize cache hits during consistent data access, and an increase in the burst transfer for all of the
half-second that it would take to flush the cache. To bring the average access time down, you must A) improve the
drive mechanism for the read-write arm, to improve seek time, or B) increase the rotation of the platters, improving
(lowering) latency - both of which increase the cost of manufacture at a time when companies in the IDE market
are running as loss leaders with the drives they currently offer.

 

[Pariah]

"You guys forgot to include the economy part when you quoted me."

I don't think anyone forgot, you're just right, nough said.

"The part I never understand about this topic when it comes up is the people saying they'd be glad to pay more for a 10K IDE drive, but if you say "go to scsi, it's where you want to be right now" they say it's too expensive. What, they think the HD companies are going to develope 10K drives for charity's sake, the same companies these people accuse of ripoff prices for scsi? But then relating to what I said above, it seems this topic has more to do with people wishing for unicorns than anything. When will we see 10K drives? When the cost premium of manufacturing them is low enough that they'll be marketable. Not before."

I agree completely.

"In the end, I stand by what I've said on the topic previously. First and foremost I want my drives to be reliable."

I don't think there is any debate that SCSI drives are more reliable, and they are all 10k and 15k RPM drives. Spindle speed, noise, and heat are not indications of reliability.

"My issue is seek speeds, and I know of no way other than faster spindles to improve seek."

Faster head actuators and reducing platter size will both decrease seek time. Current IDE drives still don't have the lower access times that the now obsolete field of 7200RPM SCSI drives had.

"The Quantum Fireball LM was the same mechanical design for its SCSI and IDE offerings, this was also back in the day
when 7200 RPM drives were considered the leading edge even for SCSI. IIRC, the Fireball LM was also a short lived
drive, because it cost them more to manufacture than other IDE drives being offered at the time."

All the Fireball+ line drives were like that (KA+,KX+LM+) except the AS+ which was designed from the ground up as an IDE drive. To date, the ATA drive with the fastest tested access time is still Quantum's first 7200RPM ATA drive the KA+ released back in early 1999. Sidenote, Seagate was already on their third generation of 10K drives and the Atlas 10KII was out by the time the KA+ was released so 7200 was no longer leading edge for SCSI when the first gen 7200 ATA drives were out, let lone the third.

 

[McCarthy]

For me, and I'm sure I'm not alone here, $220 for a 36G 10kRPM drive would DEFINITELY be a price I'd be willing to pay, IF it would work with my existing controllers instead of having to add $300 for a RAID SCSI on top of $880 for four of the drives to RAID together, not to mention try to find a controller for a 33Mhz 32-bit PCI slot instead of 64-bit and/or 66Mhz, a chore in itself.

I need RAID because I don't want 100 partitions, I want one large drive. I got burned out on partitions when using the 511MB partitions you pretty much had to use with FAT to avoid large clusters. Now that I'm using NTFS, I want one large C drive, period, and the only way to do that with smaller per-drive SCSI sizes is to RAID stripe them, else obviously each drive will be a partition. The seek time increase with RAID is minimal (actually tests the exact same with my particular drives before and after striping), a tenth of a second or less most of the time.

Ok.

Sounds like what would be ideal for you is a plain controller and a piece of software to make all your drives look like one big volume. There is such stuff around, dunno but maybe the software RAID in W2K (or was it XP? Just read something in passing without much interest) would do it. Packard Bell used to do a joined volume that way back when, they had a few machines with a 600meg and a 400 meg joined together to look like a single 1gig drive, using up some old hardware they had in stock I guess. Might try looking for something like that, though it only takes $150 off your estimates. You're after something pretty specific with access time being your goal here so you can't expect your unique wishes to be shared by the buying public as a whole. Personally access time doesn't mean much to me, it's the overall package of access time, sustained transfer rates, noise, heat, reliability and cost that makes a drive attractive or not and that's why you're seeing me and others as defensive about the topic.

As for other sidetopics, the day someone comes out with a 10K ATA then everyone will be expected to. They'll switch over and design for 10K, unless some company really jumps the gun and brings it out too soon. Point is, who's designing 4200rpm desktop drives now? Who's designing 5400's? There are a few 5400's around, but far and few between. I just don't see a landscape where 10K, 7200 and 5400 all share their own pieces of the market. Is iNTEL still making 486s? Nah. Not even making PIII's anymore, though a lot of people would love to pick one up cheap. But when you split things up you lose the economies of scale with unique product lines so it rarely happens. If we go 10K, then we'll be 10K, with some older 7200's being made with old designs. If they're realiable, quiet, cool and cheap then there's no problem. That's where we have different visions though.

Heck, I'd like a 40ms 100rpm 3000gig drive and be done with CD binders for my VCDs of stuff I grab off TV. We all want different things

--Mc

 

[Goosemaster]

Originally posted by: BD231

Originally posted by: ElDonAntonio
I agree that the HD is the damn slowest thing in my computer and that it bugs the heck out of me. Ever had ~10 programs open? try to switch from one app to the other, and you're certainly going to wait a good dozen of seconds while Windows slaps around the swap file. Same thing when opening/closing a few programs at a time, scanning the drive, running an antivirus, etc.

Those of you saying "I don't care waiting 8ms while Quake3 is loading" obviously don't understand what the hard drive is doing besides storing Quake3.

I dont thinka 10,000rpm drive will compinsate for all of that. I'd much rather wait for solid sate storage because the amount of heat computers are starting to put out is ridiculous, and a 10,000rpm drive will make the heat factor even worse.

Some of you just don't get it.

They CAN but WON"t


I have a MAxtor 10kII ultra 160 SCSI drive and it is relatively cool. I would even go as far as to say that it is on par with my 60GXP..writes aren't to bad either....
I have active cooling,but just as a precaution and because they c ame with my case...and those for those latenights when I leave it crunching...


Anyways.....IT is perfectly easy for them to make an IDE 10k DRive....but is inpratical...I now many people who think I am holding a toy when I show them a HD.


As for the speed. SCSI is DEFINITELY WORTH THE MONEY. Just becasue you don't fit the bill don't blame a great set of products. You might as well attack the caliber of a Sonicwall or the viability of SONET rings...

SCSI is THE FASTEST AT BOTHE TRANSFER RATE AND ACCESS TIMES BAR NONE.

"But..oh no...my Western Digital is awesome"

well Frank, not everybody live on your farm.


Just you put a WD in a HOME enviroment(its designation) and put a SCSI drive there and do the same at work...


Perhaps without the number one SCSI drive, the transfer rate maybe higher, but that is in a home eviroment.

In workstation operations or better, in server operations, it will get brutaly slaughered.


Access times is where scsi excells. I think the fastest drive in the world has about a 3.6ms access time, and about 5-6ms with latency taken into account
(mine has 4.9ms and 7-8ms with latency )


The WDs, legendary IDE drives of their own, the times are easily around 12-13 ms with latency...


I put an 8GB Maxtor against a 9GB ultra2wide scsi drive.....using windows and specific progs was definitely faster on the scsi..especially in stressful situations


Put that in a situation where 100 clients are accessing the system drive to retrieve files of slave drives or a webpage.....access times for the 100 indivual clients seem more important now, don't they.....


Don't dish something that is not in your league.


Just as you never inssulted the football captain in his face.


SCSI is for professionals and people who can appreciate it*.











*Rich bastids or people who save up for a really long time



<---proud Ultra160 owner and lover

 

[Goosemaster]

WOW!


I shut everbody up!


<---deserves a cookie for his valiant efforts

 

[ElDonAntonio]

nobody's asking for a free ride here guys, but I agree when someone said there currently is a huge gap in the price-performance of HDDs. Your choices actually are pretty much IDE ~1.5$/GB, or SCSI (probably Ultra160 now) which is WAY more expensive, all costs considered.

When you think about processors or graphics cards, you see manufacturers taking care of every segment of the market. Not so with HDDs. I'd like to see too some intermediate products, with performance and price between IDE and SCSI.

 

[Xentropy]

Originally posted by: ElDonAntonio
nobody's asking for a free ride here guys, but I agree when someone said there currently is a huge gap in the price-performance of HDDs. Your choices actually are pretty much IDE ~1.5$/GB, or SCSI (probably Ultra160 now) which is WAY more expensive, all costs considered.

When you think about processors or graphics cards, you see manufacturers taking care of every segment of the market. Not so with HDDs. I'd like to see too some intermediate products, with performance and price between IDE and SCSI.

ElDon definitely seems to be the only person on the same wavelength as me, here. I think introducing 10kRPM to the IDE world, using SATA as an excuse, would be the perfect offering to fill the midrange niche. 7200RPM would still be alive and kicking for another few years, and by the time they got phased out completely (like 5400RPM's pretty much are now), the 10kRPM's sound and heat properties would be even MORE optimized to the point that they'd be the same as or better than 7200RPM is today, quelling the fears of those like McCarthy.

It's how tech works. Older stuff gets optimized, miniaturized, and cheaper, and new stuff comes in at the top which at first has big drawbacks (including high cost), but those slowly get taken care of as it moves down through the mainstream toward the low end again.

The problem is that, right now, in a mid-2002 snapshot of the hard drive market, THERE IS NO MAINSTREAM PRODUCT. 7200RPM is rediculously old in tech terms, on a match with "mere" 1Ghz systems and below, I wouldn't pair a 5400RPM drive with anything Pentium III or better at all, and there's nothing above 7200RPM until you hit 10k SCSI, which represents a 4x or more increase in cost when factoring in lower densities and specialized controllers. A 4x cost increase just to have a drive that doesn't hold back 2Ghz/2000+ or better systems is rediculous, and at least one step, if not several, between those points is needed.

EVERY other aspect of the PC market has a gentle slope going all the way from sub-$500 systems to $3000+ systems and then all the way into $6000+ multiprocessor "dream" machines and servers. CPU's range all the way from Celerons through Thunderbirds and P4's through soon 333 bus Bartons and 3Ghz P4's, and then finally into multiprocessor solutions. Graphics goes from 7500's and GF2MX's through 9000/GF4MX to 8500/Ti4200 and Ti4600's/9700's/soon-NV30's, and then at the very high end, boards like the FireGL and Quadro lines, and even higher.

To compare the hard drive market to the graphics market, it would be as if no cards between a $150 Ti4200 and a $900 Quadro video card existed. Maybe calling the missing step "midrange" isn't accurate, but upper-mid might perhaps get me a little sympathy? I'm just looking for a step between abysmal 7200RPM ATA drives and 24K gold-plated 10kRPM SCSI drives. I'm still firmly a believer that if SCSI were just *gone* ATA/SATA drives all the way up the range would exist, and with much less of a price jump between 7200 and 10k RPM. The price/performance curve doesn't slope at that transition, it turns into a sheer cliff, and there are no *technical* (only greedy) reasons for it to be that way.

 

[ElDonAntonio]

amen Xentropy!

 

[Pariah]

"and there's nothing above 7200RPM until you hit 10k SCSI, which represents a 4x or more increase in cost when factoring in lower densities and specialized controllers.

True, but you also have to realize that for under $250 you can own the fastest hard drive on the planet, while the fastest P4 available will set you back $575. You can't have that many options when top of the line is $250. Look how many people are drooling over dropping $350+ on the Radeon 9700. People seem to want miracles out of hard drives. If you want top of the line you have to pay for it, but with the case of hard drives, top of the line is quite a bit cheaper than top of the line in a number of other components for your computer but people still seem shellshocked when the see the price of SCSI drives. To me the hard drive is the most important part of a computer. If I fry my video card or CPU, it's a annoyance but no big, I just get a replacement. If my hard drive dies then I lose the only irreplaceable part or my computer, the data that was on the drive. Yes, everyone should make backups, but how many people regularly do? For that reason I have no problem spending the extra money on SCSI.

"CPU's range all the way from Celerons through Thunderbirds and P4's through soon 333 bus Bartons and 3Ghz P4's, and then finally into multiprocessor solutions."

That's not a valid comparison because you are comparing a mechanical product to an electronic one. Take the Athlon for example, the only difference between the 1600+ and the 2100+ is a few cut bridges. AMD makes one product and either speed bins them or purposely drops their performance by cutting the bridges and limiting the multipliers. You can't do this with hard drives. There is no way to speed bin hard drives. If you want different hard drive spindle speeds and data densities, you have to redesign the electronic boards each time, modify the read heads, and modify the platters themselves and modify the manufacturing process. You can't just make one drive and then alter it to run at a different speed.

"I'm just looking for a step between abysmal 7200RPM ATA drives and 24K gold-plated 10kRPM SCSI drives."

10K drives are the midrange now. 15K are the highend. For $127 you can get an 18GB Atlas 10KIII which is currently the fastest 10K drive around. $127 is dirt cheap, you can't even get a 1.7GHz P4 from Newegg for that. SCSI simply isn't that expensive anymore. For a one drive setup you don't need a dual channel Ultra 320 controller. Hypermicro sells an LSI U2W controller for $69, add that the Atlas for a total cost of under $200.

 

[Goosemaster]

Originally posted by: Pariah
"and there's nothing above 7200RPM until you hit 10k SCSI, which represents a 4x or more increase in cost when factoring in lower densities and specialized controllers.

True, but you also have to realize that for under $250 you can own the fastest hard drive on the planet, while the fastest P4 available will set you back $575. You can't have that many options when top of the line is $250. Look how many people are drooling over dropping $350+ on the Radeon 9700. People seem to want miracles out of hard drives. If you want top of the line you have to pay for it, but with the case of hard drives, top of the line is quite a bit cheaper than top of the line in a number of other components for your computer but people still seem shellshocked when the see the price of SCSI drives. To me the hard drive is the most important part of a computer. If I fry my video card or CPU, it's a annoyance but no big, I just get a replacement. If my hard drive dies then I lose the only irreplaceable part or my computer, the data that was on the drive. Yes, everyone should make backups, but how many people regularly do? For that reason I have no problem spending the extra money on SCSI.

"CPU's range all the way from Celerons through Thunderbirds and P4's through soon 333 bus Bartons and 3Ghz P4's, and then finally into multiprocessor solutions."

That's not a valid comparison because you are comparing a mechanical product to an electronic one. Take the Athlon for example, the only difference between the 1600+ and the 2100+ is a few cut bridges. AMD makes one product and either speed bins them or purposely drops their performance by cutting the bridges and limiting the multipliers. You can't do this with hard drives. There is no way to speed bin hard drives. If you want different hard drive spindle speeds and data densities, you have to redesign the electronic boards each time, modify the read heads, and modify the platters themselves and modify the manufacturing process. You can't just make one drive and then alter it to run at a different speed.

"I'm just looking for a step between abysmal 7200RPM ATA drives and 24K gold-plated 10kRPM SCSI drives."

10K drives are the midrange now. 15K are the highend. For $127 you can get an 18GB Atlas 10KIII which is currently the fastest 10K drive around. $127 is dirt cheap, you can't even get a 1.7GHz P4 from Newegg for that. SCSI simply isn't that expensive anymore. For a one drive setup you don't need a dual channel Ultra 320 controller. Hypermicro sells an LSI U2W controller for $69, add that the Atlas for a total cost of under $200.

Well said.

Data Integrity and speed, not sheer size were my main reason behind buying my drive

 

[Xentropy]

Originally posted by: Pariah
"and there's nothing above 7200RPM until you hit 10k SCSI, which represents a 4x or more increase in cost when factoring in lower densities and specialized controllers.

True, but you also have to realize that for under $250 you can own the fastest hard drive on the planet, while the fastest P4 available will set you back $575. You can't have that many options when top of the line is $250.

If 15K.3 drives were $250 and didn't require a $300+ U320 controller, this thread wouldn't exist. Where are you finding your prices if you don't see SCSI drives over $250?! (Flagship 15K.3's are nearly a grand MSRP, and I saw them as cheap as $530, but $250?!)

And granted that Atlas sounds like a good deal, but at 18GB I'd need four of them, so how much would a RAID0-capable U160 SCSI controller that'd accept 4 drives set me back? We come right back to the 4x increase. Four 18GB drives at $127 ea. is $508, about four times the price of a single 72GB 7200RPM ATA drive, BEFORE adding the cost of the controller.

But, again, I'd actually be willing to pay $508 for four 18GB 10kRPM IDE drives that I could plug right into my IDE RAID controller and be good to go. Closer to $1000 for those drives as SCSI *plus* a RAID controller is a much harder pill to swallow. I should also bring up that unlike more expensive 10k SCSI offerings, the Atlas 10K III *does* have noise and heat issues and would require active cooling, another price consideration.

Edit: Doing a quick search, the best price I can find for a RAID-capable U160 SCSI controller that'll fit in a 32-bit 33Mhz PCI slot is $428 shipped for an Adaptec 2100S.
http://www.pcprogress.com/product.asp?m1=pw&pid=AD2100SKIT

That brings total cost of 72GB of 10kRPM SCSI space to $936. Hardly the $250 cost of entry you claim and I'd love to see :> Again, I'd be *ecstatic* for a $508 cost of entry instead of that $936. And then by the time I build a new system in a couple of years, PCI-X'll become standard and I'll need a U320 controller for a good SCSI at that time, so it's not even a one-time cost.

If they'd just bring 10kRPM to SATA, bam, $508 and I'd be set until the next time around when SATA3G or 6G'll be out but THOSE will be built into my motherboard for no significant additional cost.

 

[GoSharks]

but how many people like you are in the world? not many. putting two hard drives into one OEM computer will be a big step. it will be required as the capacities of IDE drives increase. people will want to see/say that they have the huge amounts of space that are not provided by 10k drives (at a reasonable price) at this point....

*ergh.. that doesnt really make sense.. its too late*

 

[Goosemaster]

Xentropy,


Like is was stated, SCSI is a joy to deal with, especially with access times in mind. I bought a $45 dollar 9 GB Eagate ultra2wide scsi for my s. Dad's machine and a 60GB IDE drive for storage.



THe key here is that although the older scsi drive is well, older and cheaper, it still holds its own in access times. THat is why it is great as an OS drive.

For example, on my"workstation" I am using scsi. A Quantum(now Maxtor)10kII 10,000rpm ultra160 SCSI drive.

here is the essence of all this: it is 18GB


Still, with its 10krpm it serves as an azazing OS drive..I even have most of my progs on there.


Since the duties of the OS are ongoing, it provides me with quick OS/prog access while still simultaneously and falwlessly running the barrage of apps I throw at it.

And since scsi is so expenive, I take into account the stregth of IDE..raw sustained speed.

Since they are considerably cheap to buy, and since they come in such hige varieties, I use it for storage, where at most two files are being accessed simultaneously, and I can keep quickly perform simple tasks.

AS for RAID, don't even bother getting into that disscussion. IF you have ever seen, even actually viewed performance on a well equiped IDE machine and SCI machine, you would know that RAID is overkill for anything fall short of a HEAVY-duty or REDUNDANT workstation or server.


Regarless, the Adaptec 2100s it a HORRIBLE scsi RAID card in that its performance sometimes is on par with the performance of one drive. Perhaps for mirroring(RAID 1) it will do but for striping (RAID 0), RAID 0+1 or even RAID 5, it is utterly useless and a waste of money.

MYLEX and some even better oens are where it's at.





FOR reference:

18GB seagate ---18GB 15k.3 (15krpm) Ultra160---Faster , colder-running and quiter--been told it rivals many IDE drives COST :$ 205

TEKRAM ULTRA 160 SCSI ADAPTER CARD DC390U3W--amazingly good car and company. --LSI chipset sometimes outperform Adaptec cards-- COST: $ 160
yet cost much much less


Knowing you have the fastest HD on earth: COST: Priceless



MATERCARD: Eat your heart out

 

[Goosemaster]

and anyways...consumer products, due to prohibitive business measures, usually never outgun professional components...


If anything, access times may continue to increase as PLATTERS get denser and rpms are left alone.


Seriosuly, with enough ram..usally 512MB, most of your active progs are in RAM so your efforts are futile...



What is so wrong about the expense anyways....you are basically comparing you "home use" to this drives capabilities and complaining why an ENTERPRISE piece of hardware is not cheap.


if anything PAY ATTENTION TO THIS:


SCSI IS DESIGNED TO HAVE A HORRIBLE LIVE FULL OF STRESS AND MISERY


5 YEAR WARRANTIES ON EVERY SCSI DRIVE THERE IS....

THEY ARE DESIGNED FOR SERVER AND WORKSTATIONS WHERE THEIR COMPLETE POWER IS NECESSARY ALL THE TIME...


NOT YOUR HOME GAMING MACHINE...




Personally...sometimes I fail to see the HD light to turn on becasue everything is in the RAM...I just like saving money and getting the best...I ocassionally push it to the limit when I need it for work....



Seriosuly, you are just like someone I found on a forum complaining about the expensive CISCO customer support..


Seriosuly, if you are buying a OC-3 capable router, you most likely know what you are doing...

He kept on insisting that they should have the right to free assitance because as a "consumer" the company has to MEET his EVERY NEED






Here is what I wrote...read it when you get a chance.....


John,

My gosh;you sound like confuscious through your language...please don't bother groping my consciousness for you have no business in there...


Anyways, here it the breakdown. Cisco caters to a SPECIFIC audience, and nothing more. Linksys does the same. They have allocated their finances to corressponding requirements and such. Just because someone has ventured into their territory does not mean that a company has to modulate their business plan to best fit YOUR needs. They do not cater to the end user, or at least not on purpose. Check their dsl products. Most of them have business-like features such as vpn and expansion slots. These are meant for IT/IST employees, who have studied the field, to install, not some guy named BOb that skimmed through "some book" he bought at Borders. Although the may resemble consumer products, they are not, and as such, there business model still remains appropriate. This is not in violation of consumer rights, because that is, once again, NOT their intended audience. Why do people complain that there is so much crap on tv? "Oh no, the entertainers are at fault," they yelp to the nearest listner(Yeah i said yelp). But in fact it is the parents fault because they purposely or inadvertantly exposed their children to something meant for another audience.

Let me give you a relevant example.

I just bought Suse Linux 7.3 Pro. It is not the "average-consumer-product." therefore I still bought it, realizing that I would have to learn how to use it on my own, or pay to learn on the phone. Is Suse an "evil" company for doing so? Definitely not. In fact, they are one of the greatest developers for the linux platform along with Mandrake and Red hat, Of which, ALL give you either limited(around 90 days of) telephone support or none at all. But as as responsible companies, they have taken the time to write extensive articles on absolutly EVERYTHING that someone might be in search of. As a non consumer-tier'd product(at least not yet anyway), I am responsible for knowing what to do or finding out how to do it correctly on my own, which is still very easy compared to some other things in life(Birds and the bee's anyone?)

It is all about audience. Really, would you expect your IT guy to be on tech support all day. Seriously, they tier to an audience which is SUPPOSED to know what to do. Just because you delve into that area does not mean you run things. That is the way busineess works...that is the way business thrives. If you don't agree, then you are not paying attention. This is not my opinion..this is the way capatilism works.

Another exmaple for my friend(?) John and others...

I am doing a design project here at college and I was in need of some heavy duty thermoplastics that could withstand 300-450 degrees C. Once I was on the phone with a company named Ticona(Ticona.com), I was instantly aware that was in another league(I dealin comps and electronics..this was a mechanical Eng. group...anyways, it was difficult for me to get support. Now, they did do an excellent job of adapting to my lack of knowledge on the subject, but clearly there services were tiered towards a different audience. Finally, now that I have been in the group for 4 months, I know what I am talking about, and when I called them back, the conversation went very well, and I got what I needed, and actually realized how follish I must have sounded the first time I called.(I had been told degrees c when Fairenheit was necessay...thermoplastics will never get that high by a long shot...I never bohtered to check the numbers...and I was interested in thermosets...as in I did not know that they are not flexible..they are used to make telephone molds by the way...)

Basically this is my entire point. Stop complaining about a company's lack of support if you aren't knowledgable enough to be in the product-audience's
knowledge-requirement bracket.

Basically, if you go out on a ledge, don't expect someone else, especially a business to go rescue you if you weren't smart enough not go out.

Realize what is this country's truths and stop chastizing it for not being idealist. Life is life, if you are not happy with it, work for a better outcome...as in read a book so you can use the product or start your own company, but just don't fill the world with frilvalous complaints.


By the way John, stop calling people immature. This forum is meant to help individuals for free. Every service I or someone else provides here is free, even if in some other areas it may cost you.

to All : By the way, I don't know if you railzed it but you are looking for answers in this forums..and that is why you will get them for free.....go to cisco's documentation pages, or Linux etc, and you can do the same.