Racetrack Memory -- 100,000 Times Faster Than Today's -- May Arrive in 5-7 Years | Hacker News Reader

Racetrack Memory -- 100,000 Times Faster Than Today's -- May Arrive in 5-7 Years

itworld.com

"100,000 times faster than today's" ... hard drives. Not RAM. Or Flash.

Plus, I'm very skeptical of claims that contain weasel-words like "may" and "could" and "possibly".

It's like beta software. Until it ships, the best anyone can say is "It may." When it ships, then they can say, "It does."

Did it strike anyone else as intellectually dishonest (or really sloppy?) that that entire article seemed to come from a parallel reality where SSDs don't exist?

I get that this would be an improvement over current SSDs. But to not even mention them while at the same time saying thigns like:

"In a first step, one could envisage attacking the Flash market and replacing the USB sticks with something much faster..."

"One could imagine the technology being useful for both small mobile devices (such as smartphones) and large systems (such as servers)."

Yes! If only there were some sort of non-volatile, low power, shock resistant, storage technology which were orders of magnitude faster than rotating hard disks! We could call it some sort of Solid State Drive, and it would totally change the storage market!

http://en.wikipedia.org/wiki/Racetrack_memory is worth a read.

The article did not make this clear, but this is non-volatile memory (according to Wikipedia). Anyone know this compares to memristors? Or is that apples and oranges?

It's certainly in the same general category of nanotechnology-enabled non-volatile solid-state memory technologies. Off the cuff, I would guess (emphasizing guess) that memristors will have better access times while racetrack looks like it might be denser. But the big question is manufacturability. There's already a pilot project to produce memristor memory. Racetrack, even if it's superior in the abstract (and I'm not sure it is), might just be too late to the party. It's not even completely clear that memristors will edge out flash, which after all has massive momentum at this point, but I think the odds are good -- memristors are (or can be) much faster.

You have to wonder about someone working on "memory of the future" who makes analogies like this:

" "It’s like reading an entire VHS cassette in less than a second."

I don't know about the rest of you but I think we're close to done using magnets to store data...

This used the word "spin" prominently, which leads me to believe that it is using spintronics: http://en.wikipedia.org/wiki/Spintronics , which may involve the word "magnetic" but is not actually based on magnets in the way you meant it.

Wow, that was an awesome read; thanks for passing it along!

I was mostly just poking fun at the reference to VHS but you are correct in your assessment of the "license" I was taking in my reference to magnetism.

This reminds me a little of reel-to-reel tape storage. It seems to me that the throughput could be enormous but the seek time would vary according to how physically long the race track was, so if you double the capacity you double the seek time. Or am I getting the wrong end of the stick?

It seems like they were right about Singularity!

Reply on news.ycombinator.com