WD 2TB drive benchmarked

RAID 0 (striping) shouldn't really be called a RAID. But it is, and is given the name of 0 because it is less than RAID 1 (mirroring) which does have redundancy. I would say that a single drive having internal striping makes sense and should be possible on a SATA hard drive as it would increase read/write performance.

Internal RAID 1 doesn't make sense though, because it definitely is intended for redundancy where the idea is to be able to quickly restore the data in the event of hardware failure by simply swapping out the drive that has failed and allowing the RAID to be rebuilt. Wouldn't be good if you had to replace platters and/or other internal components of a single drive though...

Personally, I can't wait to buy a couple of these 2TB drives to go into my home theater PC and back up caddies. The 1.5 TB drives seemed like an OK idea, but really didn't give enough of a storage size increase over the 1TB drives, especially for the price. I am also ending up with too much data to store and not enough SATA ports or drive bays to contain it all...

hmm so create raid that doesn't offer redundancy mainly striping. To raid anything we need the host controllers bios that was on the mb telling the system what kind of setup you wanted i.e 128kb, 64kb, all integrated into the hd, obviously there is no way the mb is going to be able to override internal settings all this sound barely feasible and no doubt expensive

Interesting about r/w simultaneity.

But RAID is very much about redundancy - that is what makes it important in a data centers and with servers. Hot swap disks and RAID 5,6, 50, 60 and so on, it lets me replace failed disks without data loss or downtime.

I think redundancy is important at home too: what are you going to do if your multi-disk stripe set loses a disk? Your entire stripe set is dead. All that data, gone.

Forget about redudancy. Single RAID0 spindle can give you faster rates, today data is not read/write simulataneous across all platters. If it would, we would see double performance when switching from single to double platter HDDs, and four platter strip should completely saturate SATA interface.

Anyone remember the CD-ROM that came out 1998-99? The kenwood 72x comes to mind. It used a splitter to split the laser and then employed several controller DAC's to increase the speed.

I purchased this product, and I have to say it is a really fast cd-rom. It's just companies are cheap bastards, and they don't always give the people innovative tech. They would rather keep everything the same, but the market forces them to change (slowly). So they cripple their own products and slowly make them faster or better, inch by inch, percent by percent, every year. So people think its new and improved, when really it is the same IP they aquired 10 years ago.

I just wanted to let you all know that it is completely possible to RAID (or store data in specific way as to take advantage of multiple data read/writes) then to use the current controllers way of doing it.

The issue here is one of expense. The HD companies are like fat over-bloated steering committees. No innovation, just a slight "tweak" every few years so they can keep milking the same cash cow.

Currently it would require adding controllers for each 'line' of data. So 4 platters, then probably 4 controllers. So it would increase cost. That is not to say that if a HDD company WAS interesting in make HD's faster then why not ask the controller chip company to design a controller chip capable of multi-streams, and firmware that takes advantage of all the separate read/write heads to increase data in/out speed???
I think the answer is - because it would cost money to develop, and we have been riding on ancient technology as 'forced' status quot for 100's of years now.

Single spindle RAID cannot give you redundancy: you can't replace platters you can only replace spindles.

Single spindle RAID cannot give you faster read/write: the platters are all locked at same angle wrt the head. AFAICT data is read/written from/to all platters simultaneously anyway.

So RAID in a single disk does not work.

It is possible and sometimes useful to make "RAID of RAIDs". You can make software RAID on hardware RAIDs, and harware RAID on software RAIDs. But thats not the point. Just don't drink and comment.

In each HDD drive you have separate heads for each side of platter. You have also separate "control electronics" for each head. You dont have to have separate motors for platters or/and heads to make "internal RAID" work - it's all about spliting data across interfaces. It is theoreticaly possible to make "internal RAID" (0 or 1) using just platters instead of disks.

On has to follow dedicated path, set by software variable transistors & comparisson unit. Theres just one volt from main controller to HDD at time. You'd RAID HDD & RAID Again at Controller, or multiple speeds would erupt over all sata inputs. there can be one god controller of peaquat & its on Mainboard, adjustable to abilities of HDD, & thats it, you'd have to have two sets of HDD in ONE Case, & So What, you'd be out Mb, yet maybe way to add SATA to NON Sata System? it'd figure there probably are such units, if its' possible at all, its just no use to todays mainboard, break it be my beat. STeWie Drashek

The whole idea behind RAID 0 is spreading data across two physically different devices to improve performance and get double the storage. To implement this in one drive you would need separate read/write heads for each platter and separate control electronics for each read/write head.

The point of RAID 1 is data mirroring in case a hard drive fails. To do this in one drive you would need separate read/write heads for each platter, each platter would have to have its own motor and each read/write head would also need to have its own control electronics for redundancy.

So yeah, learn the point of RAID first: http://en.wikipedia.org/wiki/RAID

And obviously you have no slight idea on how RAID works.

I though in RAID0/1 configuration disk heads are in the same position at given time and the difference is in how data feeds to them (strip or mirror). I see no reason it couldn't be done with platters instead of separate disks. In case of RAID0 data would be stripped not to separate disks by controller logic, but by disk logic to separate platters.

TVE or terrabyte Vol engine helps for mass of data. 120 Mb/S is bit high, Use as Survielence Camera is Top Rated, can be Gamer, as its reasonably fast, ~80 Mb/S. Now why two internal platters don't R.A.I.D. IS simply because all hooked together. looking for file in haze of Static & Lightning, oNE cATALOG 32 mB CACHE & aLL pICK uPS DUMP INTO SAME CONTROLLER, EACH hdd IS one uNIT, not sEPERABLE. Also, basis for RAID is two platters THAT arn't in exact same point at any moment, giving one advantage in finding file first. when fixed theres is never any variation between items put down, might as well be one huge platter is net effect, as its all ridgidly stable, NOT Like RAID array of floating address points in forced time sequence, due to variance of spin locations at each moment.tHEREFORE no aDVANTAGE till 2 HDD+ & eLECTROINICS ARE AT FASTEST CONSTANTLY in Each. nO rOOM tO iMPROVE FROM. ON JUST oNE hdd & not pOSSIBLE to RAID Alone. Drashek

You are right, you could not replace faulty platter, but in case of single block drive logic could mark bad sector on both platters and logical sector with data can be moved to another part of platter. But I think most benefit would be from RAID0 - we could have 2 times faster disk at cost of half size. Most people do it with normal two-disk setups, so why they didn't make RAID from 2 platters it in one drive to save energy, costs and maintain same noise level.

Drive-internal RAID is an interesting idea; RAID1 would certainly offer the security of data redundancy. The problem is that if one platter craps out, it could not be swapped out, requiring the user to replace the whole drive anyway to restore redundancy. So, data would be preserved, but there's no real cost benefit to the user.

I think it would be quite easy for HDD makers (and much more anticipated by customers) to create "internal RAID" from 2 or more platters inside one drive. In case of this drive we have 4 platters, so theoriticaly we could have 1TB drive with twice read/write rates (RAID0) or 500GB drive with double speed and mirrored data. It is just a matter of different internal logic of HDD. I think most consumers don't care anymore if they have 1 or 2TB, but they care of speed, and it seems to be quite easy to sacrifice size for speed within one drive. Copyright - allright.