Slashdot Mirror
Seagate Pushes Hard Drive Platters to 160GB
TheRainDog writes "Although perpendicular recording has yet to make its way into desktop hard drives, Seagate continues to push platter densities the old fashioned way. The company's 160GB platters have the highest areal density in the industry by over 25%, allowing Seagate to create a 160GB Barracuda 7200.9 hard drive that uses a single platter and costs under $90. The single-platter design has lower noise levels and power consumption than multi-platter designs, and a lower probability of a catastrophic head crash. Higher areal densities also allow the drive head access the same amount of data over shorter physical distances, improving performance dramatically in some instances. The Tech Report has an in-depth review of the 160GB Barracuda 7200.9's performance against eight competitors from Hitachi, Maxtor, Seagate, and Western Digital."
"Higher areal densities also allow the drive head access the same amount of data over shorter physical distances, improving performance dramatically in some instances. "
Yes but there is only two heads so it can't be all good.
(although two heads are better than one)
My concern would be that anything that could affect a portion of the disk would destroy more data. I know scratches that aren't noticed on a CD can make a DVD unreadable and, while a drive platter may not have the risk of scratches that optical storage does, the general idea is the same. A physical failure, such as a head alignment issue, that wouldn't be noticed with lower densities may be a factor with the higher densities.
Now, I don't have a solution to the problem, but I just want to point out that getting full performance out of something can raise new risks.
Reading code is like reading the dictionary - you have to read half of it before you can go back and understand it.
It's nice to see these on the SATA drives, but what's keeping things like that from crossing back to SCSI that SATA has taken?
Sure, there are some people who will think cheapness has some good, but I'll take uncompromising quality with speed hands down nearly anytime. 500GB+ SCSI's time is overdue.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
Why, this is the most exciting news I've heard since the last time it happened!
Which was about six months ago!
And six months before that, and six months before that, and six months before that, for more than a decade!
Someone tell me why this is news?
Is Seagate paying for this publicity?
Could anyone knowledgeable care to comment on how reliable this drive can be?
Unless a Seagate engineer that worked on this exact model comes forward and reveals a secret serious flaw, then no, NO ONE, not even other Seagate engineers, can tell you much about this drive's reliability.
You'll hear plenty of anecdotes about reliability, and every company has a hard-core "anti-fan" base who will never buy that company's products again, after losing their porn collection back in 1996.
Even within a drive family, you can't always extrapolate reliability data to other members of that family. One simple example I've seen (to my surprise) a lot here on Slashdot - A lot of people consider Maxtor as good for nothing but paperweights, because some of the earlier members of the DiamondMax line really really sucked. I, however, have half a dozen of the later DiamondMaxes in use today, some as old as five years, without a single failure, ever.
So, buy either the cheapest or the largest (or the inflection in that curve, which IMO Maxtor usually solidly holds, thus my using their drives almost exclusively), and just make sure you have everything backed up. Because eventually, you will have a catastrophic HDD failure. And as much as it sucks to waste a few hours reinstalling your OS of choice, it sucks a LOT more if you don't have all your software, porn, data (but I repeat myself), music, and what-have-you readily available on a backup.
Personally, I wouldn't buy a mere 160GB drive anyway, when you can get nearly twice that for $20 more. But this may have one nice side-effect, in that if Seagate pushes out a 4-platter 640GB drive (hey, no one will ever need more than 640GB, right?), the 400s should finally drop down to the golden $100-$150 range.
Doubtful that this will happen anytime soon, considering that many organizations still use large scale tape drives for backup (example: http://www.exabyte.com/). If tape drives are still around today, who's to say hard disks won't exist 20 years from now? What's more likely is that flash drives may become more viable for mainstream desktop computers but larger-density hard disks could be used in some other market. You'll see the drives fulfilling a different niche, perhaps.
Guess we'll all just see.
--- "To iterate is human, to recurse divine." -- Robert Heller
While in principle I have to agree with you, I think that your post a little off.
Personally, my ultimate setup would be completely mirrored disks. Given that the performance of todays drives outpaces my needs (probably by a few orders of magnitude), and the price of the drives, just using two is good enough for me.
I believe that a mirrored 2 disk array, is much less suseptible to failure than your 5 disk RAID 5. Having only two components, the chances of failure are _MUCH_ lower than 5 disks, and with mirroring you get solid performance with exellent reliability. Plus, are you _SURE_ you could recover from a failed disk in a RAID5 setup? Have you praciced that recovery? RAID 1 is so much easier. A good deal of data is lost while admins try to recover from failed drives, sometimes this loss is caused by the admins actions. Having a simpler solution makes things much safer.
RAID 5 was for a time when small disks were the norm. When you can have a single disk that has 10x the space your needs require, RAID1 makes the most sense to me.
Yeah, I have seen several drives recently for less than $.033/GB at Frys or after rebate. I think the sweet spot is about 250GB right now.
Keep in mind that the most GB/$ will always be in the $80-$150 range regardless of current densities. The premium product always sells for > $150. Also, the manufacturing costs keep the prices from dropping much less than $50. So if a drive only passes on half the heads you get 1/2 capacity for $50 instead of $80 for all the heads and surfaces.
A lot of people consider Maxtor as good for nothing but paperweights, because some of the earlier members of the DiamondMax line really really sucked.
If they hadn't replaced those drives for customers with just as low quality ones that would also fail rather quickly, they probably wouldn't have experienced such a long-lasting backlash from the geek community.
When replacing problem hardware, companies should never send a replacement that they know damned well is likely to be a problem. People will often forgive once, but are not as like to do so twice in a row.
This is not a dream, not a dream...we are transmitting from the year 1-9-9-9.
"make sure you have everything backed up. Because eventually, you will have a catastrophic HDD failure."
When will people learn?
Hard disks fail.
I care little for "my 100 hdds have been running 24/7 for 10 years with out a single failure" anecdotes.
Moving parts fail.
Make your backups.
Don't come crying to me when you lose data.
You just need to ask one question:
Do I care if I lose my data?
No: Fine.
Yes: Backup. Properly. Off your PC. Preferably off-site.
Yes, but my drive will not fail: Take your hands off the keyboard and back away slowly. The men in white coats are on the way.
b3 4phr41d 0f my 4bov3-4v3r4g3 c0mpu73r kn0wI3dg3!
MadDwarf
Don't be afraid to use terms like bandwidth, throughput, seek time, cache access etc.
Single platter solutions result in reduced amount of heads. Less heads = less weight to push across the platter = higher acceleration at same force applied = lower seek times, the head moves faster, can find the place faster.
But the bottleneck point in throughput lies between the surface of the disk and the head, a single head can read just as many bytes per second, the limits are pushed higher but still this is the point that makes read slow once the seek was finished. So all heads read/write at once, a single large file gets spread over all the platters, but at narrow band of cyllinders, so it can be read whole faster, by using all the heads to read parts of it at once, and reassemble the data in the cache. Less heads = less paralell readouts, lower throughput.
I find much more future in big multi-platter drives based on the new tech, than this single-platter thing, that offers little gain and much loss at a very high price.
Anagram("United States of America") == "Dine out, taste a Mac, fries"
Hmm, 1024GB = 1TB.