Hard Drives Not Needed

[Fox_granit]

Funny, this isn't coming up on any major tech. sites that I've seen, but this article states we may have RAM completely eliminated from computer architecture's soon. I'm wondering how soon this will come into the market, and what difference this will make overall in computing.

We may have direct access RAM only instead of a HD and RAM. This is something that the electronics industry has been trying to develop for years. This will completely revolutionize all that we do with computers. Think about what this will do for storage potential, about durability, SAN storage capabilities……. everything.


http://msnbc.msn.com/id/13798680/

 

[mpjesse]

Doh! I just made a news post about this.

Anyways, I wouldn't count on seeing MRAM in very high capacities (certainly not in hard drive capacities).

 

[Heyyou27]

Yeah, right.

Freescale, which was spun off of Motorola Inc. in July 2004, said Monday it has been producing the 4-megabit MRAM chips at an Arizona factory for two months to build inventory.

What can we do with 4mb?

 

[Fox_granit]

Not initially, but I see it completely eliminating hard drives. The capacity is all going to depend on the manufacturer and their process.

It only makes sense this would happen. But in the next five years, I don't seeing it make a huge impact on desktops or laptops, but on phones, PDAs, and other portable devices are going to really be changing. The speed of flash drives and MP3 things are going to be extreme.


SAN's could potentially see the greatest impact if they can start building it in masses. If it could change everything concerning storage if they can get the capacities high enough.

 

[Fox_granit]

How big was RAM when computer's first started? in the K's right? I'm looking at the future man. They've taken RAM and pushed it further. besides, 4Mbs on a chip is nothing right now, but how big were most RAM chips on the old 64MBs when SDRAM came out?


When the manufacturing changes, it will only grow larger and faster.

 

[FITCamaro]

It has potential but it'll be probably 2-3 years before we see capacities at the level of flash memory now.

 

[mpjesse]

Yeah but the thing is MRAM doesn't increase capacity. The largest capacity NAND flash chips are like 2-4GB. I can see MRAM replacing SDRAM (DDR)... but not hard drives.

MRAM is about speed more than it is capacity. With hard drives approaching the 1TB mark, NAND, NOR, and SDRAM are still way behind in capacity. And far more expensive to make.

 

[Fox_granit]

the question then remains, why create it? To me, the idea of having ram that can store information without corruption, which can access faster than current flash memory has only 2 uses, to replace flash memory in moble machines and for supplimenting, and then eventual replacement of larger storage devices. Look at solid state drives now; look at the size of an ipod shuffle, how many of those could sit inside a blade server case? if they were ever to get into the density of what current flash memory has evolved, then designing circuit boards to house these chips would be easy. Less physical space would be needed, and less heat would be generated, becuase of no moving parts. They never described the physcial size of these chips, so if used in larger quantities, multiple gigs can easily be placed on an atx size circuit board.

This will only be expanded beyond the limits of current flash memory. And even if it is under the current restraints of flash memory, what provents its use in mass on a circuit board for larger storage? The only thing that would be its restriction would be the memory controller.

To me its only a matter of time. That time, however, is upto debate. Think of how much more complicated it is to build a hard drive than it is to manufacture a chip. Last look at Seagate and Samsung who have already incoporated flash memory into their harddrives for production in early 2007. They have started the transition.

 

[mpjesse]

For sure solid state memory drives will eventually replace hard drives. No one can argue against that. What I'm saying is that it won't come in the form of MRAM.

The only thing stopping "RAM drives" right now is cost. And I think that with the current form of flash memory (even MRAM) it will remain so. Hard drives are exponentially cheaper to make and have a solid record of increased capacity and reliability.

I think the only thing that can displace the hard drive's reign is some form of new nanotechnology based memory that doesn't require the complicated process of creation like flash based and MRAM chips do. Huge advancements have been made in nanotech since MRAM was first demostrated. Though MRAM is a "nanotech" based device, it still doesn't have the "exponential value" that we'd need to replace hard drives once and for all.

But hey, you could be absolutely right. MRAM could replace hard drives. No one really knows for sure. ;-)

 

[mpjesse]

I believe this has strong potential to replace hard drives.

Someone (tom's or anandtech) did a write up on it. Very interesting stuff. I think Seagate is working on it now. I'll try to find the article.

 

[Fox_granit]

I completely agree with most points, but as far as it not being MRAM, well I dunno, out of all out there it has the most potential. If its not going to be it, I think it would be a desendant of it. Too much at stake for them not to develop it further.

 

[Fox_granit]

I would like to learn more about that paticular nano switch. It could prove itself useful, but we'll see overall what happens.

 

[mpjesse]

I spent probably 30 minutes looking for the article that was written about it. I never found it... irks me because it was pretty in-depth, but easy to understand for sods like us.

 

[Fox_granit]

Well looking at it overall, I'd like just to see what else is being developed. I think the solid state hard drive is just around the corner, esp. if they can get either of these 2 technologies ramped up.

I like the idea behind Gigabyte's I-Drive where you just add ram. Personally I believe this could be applied to the MRAM, esp. when it gets into larger quantities. If this could then be applied for a SAN, what a difference that would make for all enterprise servers, major webpages, etc.