Sign in to follow this  
Kevin OBrien

Kingston Says Lower NAND Prices Will Quicken Adoption of SSDs Discussi

Recommended Posts

The high cost of flash has been a major obstacle for the mainstream implementation of SSDs in the market place. According to Kingston Technology, however, the use of SSDs are expected to expand significantly with falling NAND flash prices. The average price for 1GB of NAND flash memory is predicted to drop to $1 USD and will help trigger the replacement of HDDs with SSDs by late 2012, claims Nathan Su, flash memory sales director of Kingston APAC.

Read Full Article

Share this post


Link to post
Share on other sites

Everyone already learned that lesson when HP dropped their tablet prices to $99.

I still think that the real issue here will be with the lifespan of 19nm TLC flash. I think it will be a major problem.

Edited by Nihility

Share this post


Link to post
Share on other sites

You've mentioned that in a previous post. Can you give a bit more information?

As far as I can tell, even the best write amplification is at around 0.7 (for Sandforce SSDs which do on-the-fly compression). Any other SSD will have a >1 write amplification, so it would actually be worse than the stated lifespan of TLC Flash.

Besides compression (which can't be done on media files or other compressed files) and flash over provisioning (keeping a chunk of flash on the side side 7-15%, what else could they do in software?

Share this post


Link to post
Share on other sites

Very valid question, Nihility. Brian's link essentially says "Indilinx nDurance", again. That's something the OCZ Octane supposedly uses, but where (as for as I know) we have absolutely no idea what it is.

One option that I can think of to increase write endurance via software (which just came to my mind after reading your question..) is: if a write to an existing file is to be performed, read the data first, compare to what's to be written and only actually (re-)write the modified blocks. That would trade write performance for endurance. This trade off may be worth it with the read speeds being as high as they currently are, and with TLC on smaller processes getting less and less "write enduring". One could even go to smaller blocks than 4k to increase the granularity and gain larger benefits by this technique. It would also make the 4k write performance more important, again (assuming 4k blocks). Which manufacturers may like, as it would make SSD speed important again.

I should probably launch a patent investigation immediately, maybe someone beat me to it. But of well.. Brian, feel free to forward this suggestion to your contacts (if you think it's worth it). Just don't forget to tell it was from me ;)

MrS

Share this post


Link to post
Share on other sites

We're actually trying to get on the schedule with the Indilinx guys now to get a more in-depth understanding of their processor.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this