Seagate Demos World's Fastest Hard Drive, Doubled Performance, HAMR Progress

  • Thread starter Thread starter Guest
  • Start date Start date
Toggle sidebar Toggle sidebar
Status
Not open for further replies.


I do, because I remember having this idea around the same time, but also wondering if it would work combined with an internal RAID controller, thus allowing for something like RAID 0 self-contained within a single drive.
 
The only reason they didn't do this all those years ago is that it was more cost effective to increase the density. Now that they are reaching the limits of platter tech, they can actually afford to put more complex disk controllers in there. It should at least double the cost of the circuitry, if not more. Probably also needs a heap load more cache to keep up the writes.
 
What may be cool with this is make the hard drive an internal raid 1 drive. Halve the capacity/performace (so equal to a standard drive) but have the two actuator arms basically writing the same data on its set of platters. This creates a single drive with Raid 1 reliability.
 


yeah, in a single drive.
How is that anything related to 'reliability'?
That is no different than 2 partitions on a single drive, and calling it a RAID 1.
(And yes, I've seen people here want to do that concept...:pfff: )
 


This isn't a new idea in the slightest. Conner, who was bought out by Seagate after bankruptcy, sold a dual actuator hard drive in the early 90's. Google Conner Chinook. Conner beat your thought by a decade with a commercial product.
 


It doesn't surprise me that this had been considered prior to myself, salgado, and probably others, having the idea.

In fact rather than Seagate's current implementation, with the actuators vertically above each other, I had pictured it more like this (photo of the Chinook from the Wikipedia page)

https://en.wikipedia.org/wiki/Conner_Peripherals#/media/File:Conner_Peripherals_%22Chinook%22_dual-actuator_drive.jpg

with the actuators being separate from each other.

Actually, not realizing the size, I had also wondered if it were possible to have four independent actuators in all four corners.

Still, intriguing that something like this is actually being implemented. The sequential throughput, if accurate, is pretty impressive for a HDD, approaching SATA SSD speeds.

I imagine it's slower in consumer use cases, which I guess is why this is aimed at data centers for now.
 
@Captaincharisma Data Centers still use many mechanical spinning hard drives. Their capacity and cost still make them more effective for storage. NAND Flash SSDs also have a problem of increasingly slower performance and failing faster when they are filled. The only solid state product to show the ability to perform at capacity is the 3D X-point from Micron and Intel. Even so, until prices and density are equivalent, spinning discs still outsell solid state devices. And, SSD or not, even solid state has a density problem when it comes to shrinking chip sizes. The simple fact remains solid state has a long way to go in terms of price and performance and density to totally replace all the spinning drives in data centers around the world.
 
*Performance (at capacity) was the key point I was making; in terms of one of the issues with SSDs in data centers; or for users with smaller SSDs.
 


Depends on your use case.

Also depends on the price per GB of capacity.
 
I don't no when the first commercially available multiple actuator hard drive was available, but it was long before the year 2000. For example, the "History of IBM magnetic disk drives" wikipedia article specifies that the IBM 3380 had:

two hard disk assemblies each with two independent actuators

and that the first of these units shipped in October, 1981.

Good to see multiple actuator technology appearing in small system HDDs, but it appears this first implementation won't improve rotational latency. Ideally the multiple actuators would place the heads at multiple point around the same platter. That way whichever head the data first rotated under would be able to access the data. With two correctly positioned actuators this would allow a halving of the average rotational delay . A big benefit when the rate of database transactions are limited by the rate of persistent random access storage transactions and you can't use SSD because the rate of these transactions are so high that the SSD would quickly wear out.
 
Why do people keep asking why anyone would want a mechanical drive over an SSD? Many mechanical drives are priced in the vicinity of 2.5 cents per GB, while the least expensive SSDs currently cost around 10 times as much, around 25 cents per GB. For anyone who has a lot of data to store, SSDs are simply not very practical for that purpose. You can get a 4TB HDD for as little as $100, while 4TB of SSD storage will set you back around $1000 or more. For most bulk storage uses, the higher performance of an SSD is simply not necessary. SSDs can be great for the OS, applications and maybe some regularly played games, but for things like video, music or backups, hard drives still have a huge cost advantage.
 


I still laugh every time I see your profile pic.

I suppose they are trying to make a point that an SAS flash drive as cache before the slower copying to the spinning drive makes more sense. If speed is the goal, there is already speedy storage. This drive fills a niche of speed and capacity. I can't conceive of a scenario where I would need the bulk storage to have that much throughput. But I am sure someone out there can justify the cost/performance/capacity ratio that this drive offers.

Either way we're not going to rush out and buy an enterprise drive either so most of these comments are pointless.


 
Idk. Unless it is new data that is being written, or if it is able to somehow always split the data between the two heads (Which is what I would think kind of like a RAID 0 almost AND NOT I"M NOT SAYING ITS A RAID 0 just that same kind of thinking) then you could get that double data write. But i mean if ALL your data is on on one platter then you are NOT going to hit those speeds. I just see this drive coming out, they say this, you get it, doesn't even come close because of how the hard drive gets used. Unless it knows to split the data between the two heads you will never see that "More than double bandwidth" it is talking about. If it starts to write data, and you need to pull data from that same platter its going to be just a normal drive.

Now what I think would have been a BETTER idea, yes it would be a LOT harder as it would make the driver bigger (Or i think longer) Is to put in a SECOND full head. Each head acts as they do now. This way each head has access to all the data. Now I could see there being issues as well. They both need to write. Well one head will get to a empty sector before the other etc so the hard drive would then have to do a lot of ahead type thinking possible slowing it down but I would see that as a more valuable product even though it would break HDD Physical size standards.
 
This is already obsolete technology. Super-prone to catastrophic failure and super-slow compared to modern drives. Check out Samsung 960 PRO SSD 2TB:

Sequential Read 3,500 MB/sec
Sequential Write 2,100 MB/sec
RANDOM READ (4KB, QD32) 440,000 IOPS (Thread 4)
RANDOM WRITE (4KB, QD32) 360,000 IOPS (Thread 4)
http://www.samsung.com/semiconductor/minisite/ssd/product/consumer/960pro

Once you try SSD, you do not want mechanical rotational disks, even for free.
 

Yes, we should all just buy $1,250 2TB SSDs to store our cat videos on. Don't forget to pick up a second for backups! Those $60 2TB mechanical hard drives aren't nearly fast enough for playing twenty cat videos simultaneously.
 


These specialty drives aren't for that either.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom