$895 for this kind of performance is really encouraging. I know I'm probably being too optimistic but seems like this could be mainstream in 4 or 5 years time.
$895 for this kind of performance is really encouraging. I know I'm probably being too optimistic but seems like this could be mainstream in 4 or 5 years time.
We'll probably see mainstream implementation of tech like this well before 4 or 5 years. $895 feels steep, but if you consider what that card is capable of, and what it'd doing for you if you're a gamer or CG artist, then that price point is very close to being just right. It'll come around sooner than later.
I bet it wont be long until we see some kind of combination unit in a 3.5" form factor. An SSD combined to work in tandem with a hard drive. That could have real potential to speed up your large data transfers while keeping the core of your OS everyday applications on SSD, while dedicating your hard drive for your photo and music collection. You can do that now across a couple of drives, but combining the tech into one unit could make it more palatable to OEM's and mainstream users ultimately making it more affordable.
Long term we all want speedy SSD's that support the massive storage requirements we have, that is obviously some years away, but if you could combine an SSD for speed for large transfers and quick booting the OS, and have the spindle drive there for your storage requirements, it may get us an affordable performance upgrade sooner than we thought.
We were supposed to have those devices. They were called Hybrid Hard Drives. They were supposed to be everywhere on the notebook: A mechanical drive backed by several gigs of NAND. Windows Vista was built specifically for them...
Looking around on Newegg, looks like some 2.5" high speed 64 GB SSD's are hovering around $180 now. Its getting a little better all the time. For guys like us, we can just set up two drives and get the hybrid effect today. $180 might be getting into a range where I can almost justify doing it.
For Joe consumer and the OEM's though, someone will need that hybrid drive. Its probably the next logical step for this tech, and I think that is what Fusion-io is hinting at. We got the fastest SSD on earth, but for today, its not a hard drive replacement, its a hard drive enhancement. (one hell of an enhancement at that)
If they start marketing that way, and getting consumers to accept the fact that they are not going to meet their computers entire storage demand with SSD I think the market may be more receptive to them and start buying them to enhance their performance without expecting it to replace their spindle drive just yet.
So for now, that may be the way to go to market, SSD, its a system performance enhancement, not a replacement storage unit. In four or five years, perhaps we will get to a point where we can all afford blazing fast terabyte SSD's, but if you want to taste the performance now, its getting to a point where you might be able to justify it as an early adopter. Maybe not these $900 Fusion-io offerings that are obviously aimed at professional workstations, but it is getting there.
HHDs are dead. They will never exist. The next logical step is to get SSDs down into a reasonable $/GB range and phase out mechanical disks.
Storage occurs in ~30 year cycles. We've been using the mechanical disk since the 70s. NAND has only been around since 1988 or 1989, which means it has a good 10 years to mature before we even come close to the point hard drives are at.
But there are companies like Fusion-io that want to sell SSD's today, and that means marketing as a performance enhancement and not a storage replacement. SSD is going to be less about storing your data, and more about how it enhances how your data moves from A to B.
Think about this too, if your a company like Western Digital or Seagate, and your schlepping spindle drives in a competitive market for razor thin margin's, and there really isn't much that you can do to advance that tech and demand a reasonable profit, it is what it is, its as good as its going to get. Those companies are probably dying for an opportunity to advance their technology for the next round, and with NAND becoming more affordable, it may provide an opportunity for them to re visit the hybrid drive as a viable option for consumers in the mid term, while SSD becomes a more viable storage replacement, which is at least some years away.
Also, if your the other hardware vendors, the Intel's and AMD's of the world, you have to be hoping that it will get adopted to remove the bottleneck that is already slowing your architecture, and its gotta get bad enough at some point where its going to hamper their development and ultimately their ability to justify new product at a premium.
The only way I see OEM's and consumers eating up NAND over the next year or two is if its marketed in a tidy neat little package with a spindle drive. I would not say hybrid drives are dead, they were just before their time, I will go so far as to predict we see a few in OEM machines alongside the windows 7 release in Oct.
One of the things David stressed was that at the moment, SSDs aren't completely reliable. They have a rapid failure rate of about 1 in 1,000, and each one of those cards have per chip two stacks of arrays of four equalling up to something like a hundred or so chips. I think.
My numbers are probably fudged, I've been partying all week, leave me alone.
You get the idea. So the chance that one of those chips are bad is relatively high right now. Now, the ioXtreme is built with two redundancy chips to prevent data loss at the failure of a chip (and even implements a self-heal technique).
The point is, high capacity NAND requires a lot of chips, which results in a riskier failure rate. Though not typically catastrophic, it is still a nuinsence. For massive storage amounts, the HDD is still a far better choice. Until NAND matures and becomes more reliable, we probably won't see too many massive storage options.
We'll probably see mainstream implementation of tech like this well before 4 or 5 years. $895 feels steep, but if you consider what that card is capable of, and what it'd doing for you if you're a gamer or CG artist, then that price point is very close to being just right. It'll come around sooner than later.
I won't lie, @ $895 my interest is perked. There is a CAM program that I use and I'm wondering if a faster drive could make it generate complex tool paths faster. It's a CPU intensive task for sure, but I don't know if a faster hard drive makes a difference. If it does and it could significantly reduced times, then this could be something I would give a serious look at.
UPSLynx
:KAPPA:Redwood City, CA Icrontian
Comments
We'll probably see mainstream implementation of tech like this well before 4 or 5 years. $895 feels steep, but if you consider what that card is capable of, and what it'd doing for you if you're a gamer or CG artist, then that price point is very close to being just right. It'll come around sooner than later.
Somewhere between 9-20, depending on the drive.
Long term we all want speedy SSD's that support the massive storage requirements we have, that is obviously some years away, but if you could combine an SSD for speed for large transfers and quick booting the OS, and have the spindle drive there for your storage requirements, it may get us an affordable performance upgrade sooner than we thought.
Yeah, that never happened.
For Joe consumer and the OEM's though, someone will need that hybrid drive. Its probably the next logical step for this tech, and I think that is what Fusion-io is hinting at. We got the fastest SSD on earth, but for today, its not a hard drive replacement, its a hard drive enhancement. (one hell of an enhancement at that)
If they start marketing that way, and getting consumers to accept the fact that they are not going to meet their computers entire storage demand with SSD I think the market may be more receptive to them and start buying them to enhance their performance without expecting it to replace their spindle drive just yet.
So for now, that may be the way to go to market, SSD, its a system performance enhancement, not a replacement storage unit. In four or five years, perhaps we will get to a point where we can all afford blazing fast terabyte SSD's, but if you want to taste the performance now, its getting to a point where you might be able to justify it as an early adopter. Maybe not these $900 Fusion-io offerings that are obviously aimed at professional workstations, but it is getting there.
Storage occurs in ~30 year cycles. We've been using the mechanical disk since the 70s. NAND has only been around since 1988 or 1989, which means it has a good 10 years to mature before we even come close to the point hard drives are at.
Think about this too, if your a company like Western Digital or Seagate, and your schlepping spindle drives in a competitive market for razor thin margin's, and there really isn't much that you can do to advance that tech and demand a reasonable profit, it is what it is, its as good as its going to get. Those companies are probably dying for an opportunity to advance their technology for the next round, and with NAND becoming more affordable, it may provide an opportunity for them to re visit the hybrid drive as a viable option for consumers in the mid term, while SSD becomes a more viable storage replacement, which is at least some years away.
Also, if your the other hardware vendors, the Intel's and AMD's of the world, you have to be hoping that it will get adopted to remove the bottleneck that is already slowing your architecture, and its gotta get bad enough at some point where its going to hamper their development and ultimately their ability to justify new product at a premium.
The only way I see OEM's and consumers eating up NAND over the next year or two is if its marketed in a tidy neat little package with a spindle drive. I would not say hybrid drives are dead, they were just before their time, I will go so far as to predict we see a few in OEM machines alongside the windows 7 release in Oct.
My numbers are probably fudged, I've been partying all week, leave me alone.
You get the idea. So the chance that one of those chips are bad is relatively high right now. Now, the ioXtreme is built with two redundancy chips to prevent data loss at the failure of a chip (and even implements a self-heal technique).
The point is, high capacity NAND requires a lot of chips, which results in a riskier failure rate. Though not typically catastrophic, it is still a nuinsence. For massive storage amounts, the HDD is still a far better choice. Until NAND matures and becomes more reliable, we probably won't see too many massive storage options.
...I'm exhausted.
I won't lie, @ $895 my interest is perked. There is a CAM program that I use and I'm wondering if a faster drive could make it generate complex tool paths faster. It's a CPU intensive task for sure, but I don't know if a faster hard drive makes a difference. If it does and it could significantly reduced times, then this could be something I would give a serious look at.