Rumors first appeared about two weeks ago that Western Digital planned on releasing 1.5TB and 2TB hard drives. As of today, the rumors are official. WD is introducing its third-generation GreenPower drive series and WD manages to reach the two terabyte milestone first. The latest Caviar Green (WD20EADS) hits the 2TB mark with four 500GB platters, each rated with an areal density of 400Gb/in².
Of course, the first question that comes to mind is spindle speeds. Western Digital does not comment on exact rotational speeds with the GreenPower drives, only to say that it is close to 5,400RPM. While that is important for some, we see it a bit differently as this latest Green's high areal density, combined with 32MB of cache, and new electronics will provide very good performance.
This particular series of Green Drives features an update to WD's Intelligent Drive Technology. Those include: StableTrac, which secures the motor shaft at both ends to reduce system-induced vibration and stabilize platters for accurate tracking during read and write operations; IntelliPower, which fine-tunes the balance of spin speed, transfer rate and caching algorithms; IntelliSeek, which calculates optimum seek speeds to lower power consumption, noise, and vibration; and NoTouch ramp load technology, which is designed to ensure the recording head never touches the disk media.
Western Digital says the 2TB Caviar Greens will ship late this week, so expect to see them at e-tailers soon. WD's suggested list price for the drive is $299, which is certainly more than two of their top performing Caviar Black 1TB drives . However, you end up with a single drive featuring improved acoustics and power consumption along with performance that should satisfy most users. The 1.5TB drive should ship later this quarter. We decided to take a break from the firmware carousel and will provide an in-depth review of the Caviar Green 2TB drive once our retail unit arrives.
Facebook
Twitter
RSS
29 Comments
View All Comments
GenRabbit - Friday, January 30, 2009 - link
So how long before motherboards support EFI so we can boot from drives using GPT rather than MBR? I see a problem coming as we pass the 2TB barrier..gleblanc - Thursday, January 29, 2009 - link
OK, might be stretching a bit here, but I was trying to figure out what was required to make these drives work, and I've miscalculated something. If we take 400Gb/in^2 to mean 400,000,000,000, and run the calculation 1/sqrt(400,000,000,000 Gb/in^2) to get the linear density in one dimension, I come up with approximately 1.5 micro-inches between bits, or right about 40nm. What am I missing here?I ran some more calculations quickly, and it seems that 10 square inches (500GB/400Gb/in^2) is about right per platter. I just can't quite grok a positional accuracy on that scale.
Yawgm0th - Wednesday, January 28, 2009 - link
Is anyone else a little irritated that advances in spindle rate and general performance of HDDs have been virtually halted for years, while capacity has increased at a healthy rate?Don't get me wrong, I love capacity. I am using about 2TB of data across multiple RAID arrays and over a dozen hard drives right now. I have nothing against big drives. But capacity is cheap. Even capacity with redundancy is affordable. Performance is not.
To get substantially improved performance in all areas, right now one must use multiple drives in RAID 0, 0+1, or 10 or use pricey enterprise-class drives. A 10,000 RPM drive is pretty nice, but they're still extremely expensive and frankly the performance improvement isn't good enough. RAID 0 is a crappy option because it increases the chance of data loss, and RAID 0+1/10 brings costs much higher.
It's appalling that speeds and capacities in other areas (RAM, CPU, GPU) are doubling every 1 to 2 years, but over the last five years or so HDDs have maybe increased 50% overall, with virtually no increase in seek time. An 80GB 7200RPM was the standard back then, and the same amount of money now probably gets you a 7200RPM 500GB that performs 30-40% better overall but nearly identically on seek times.
Most will agree the overall computer experience these days is slowed by the hard drive more than anything. We're basically using bigger versions of the same drives we had just past the turn of the millennium, but RAM, CPU, and GPU are anywhere from 500% to 2000% faster. That's ridiculous. Instead of buying multiple GPUs or 8GBs of RAM, I build a high-end systems with 4 or 8 hard drives in RAID 0+1. I, for one, think that's ridiculous, but the overall experience is better than giving someone 2TB of storage or triple SLI.
Is there a good technological reason they can't make 10,000 RPM or 15,000 RPM drives mainstream? These drives have always been cost-prohibitive for the consumer or forced an unacceptable sacrifice in storage capacity. Where are the affordable 500GB 10,000 RPM drives? That's what we need, not 2TB 5,400 RPM nonsense.
/rant
Jeff7181 - Wednesday, January 28, 2009 - link
There is a market for large, slow drives. (I'm in the market for one, I just wish they were cheaper) Disk to disk backup for instance... archiving... bulk media storage.Sure, speed is good, but I also like big slow and preferably cheap drives for things that require space, not speed.
erple2 - Wednesday, January 28, 2009 - link
I think that you're looking at the limitations of the current technology of a spinning metal disk. You can't read the data accurately enough at higher spindle speeds.Ultimately what's important is the rate at which the read head can read data off the platter.
Note that the 15k RPM drives are really only 2.5" sized, but put inside a 3.5".
I think that the only way to boost speeds is going to be with heavy usage of flash memory.
We're not really able to manufacture the drives to within tighter tolerances (physically constructing the platters and read mechanism, that is) than we could when the 15k RPM drives came out. That, I think, is what's really limiting the drive market.
Notice that for everything else that's doubling in speed, NONE of them have any moving parts in them? I think that's the problem. Get rid of the moving parts, and you'll see a speed increase with drives.
Ultimately, it all boils down to the mighty manufacturing tolerances of a device spinning very fast is harder to achieve than the latest die process tech Intel/AMD is using. Improving the areal density is about the only way to boost the actual read speed of current tech, I think.
Baov - Tuesday, January 27, 2009 - link
I'd rather have a 1TB drive with just two platters. Are they gonna make any of those?Jansen - Tuesday, January 27, 2009 - link
Yes, the WD10EADS will transition to a two platter design as a cost saving measure.Stonedofmoo - Wednesday, January 28, 2009 - link
Any source for that information, or as to when it will happen?Presumably the model number will change to reflect this change or it will be damn confusing?
VaultDweller - Wednesday, January 28, 2009 - link
Probably will just be damn confusing.WD didn't change model numbers last time they transitioned drives to models with higher density platters.
mczak - Wednesday, January 28, 2009 - link
You can bet it will be confusing. Just like there exist 3 platter and 4 platter 1GB EACS now, there will almost certainly be 2 and 3 platter 1GB EADS (and possibly 2 platter EACS too making that 3 different drives with quite different performance/power/noise with the exact same model number except the batch number is different - remember the C vs D there is only for 16 vs 32 MB cache). As much as I like WD disks, I really think they shouldn't do this. Worse, there's no obvious transition date to new models since WD appears to produce two different versions at the same time.