Breaking News
News & Analysis

SanDisk Finds Profit Behind the Tech Curve

4/18/2014 12:30 PM EDT
3 comments
NO RATINGS
Page 1 / 2 Next >
More Related Links
View Comments: Newest First | Oldest First | Threaded View
HangLai
User Rank
Rookie
San Disk share price is way over valued!
HangLai   4/18/2014 7:05:52 PM
NO RATINGS
SanDisk/Toshiba stock is flying high because analysts' recommendation, but the revenue, thus the profit level will be stagnate due to price erosion.  Sandisk/Toshiba is not a low cost company, with Samsung, MU/Intel and Hynix coming to the market in a big way, the NAND price will be coming down in a hurry.  SanDisk does not have any secret source to the SSD for memory nor NAND manufacturing process, why should SNDK share price deserve 20 times earning while the others are in low teen's?

Wilco1
User Rank
CEO
Re: Managing SSD sector rewrites
Wilco1   4/18/2014 4:16:09 PM
NO RATINGS
Flash can only write 16KB blocks and erase 128KB blocks, ie. much larger than disk sectors. SSDs maintain an internal mapping of the disk blocks, combine multiple writes into a single block write and perform garbage collection to reclaim dead blocks. Typically 5-10% of the total flash is reserved to minimize the wear on the flash blocks.

To give you an idea how well this works in practice, my main drive is an 128GB SSD, and after 1TB written has a wear level count of just 25. It has a guaranteed wear level of 3000, and so can still write 120 times as much.

Hmm, if I don't start writing a LOT more data soon then I will wear out long before my SSD does!

DrQuine
User Rank
CEO
Managing SSD sector rewrites
DrQuine   4/18/2014 1:34:59 PM
NO RATINGS
I note the 'SATA SSDs designed to stand up to 10 full-drive writes per day, are guaranteed for five years'. My recollection is that this technology is susceptable to failure if written too many times. If there are specific sectors on the SSD that are experiencing frequently rewritten data, will they fail or is there a mechanism to protect the disk data integrity? This might be accomplished by using logical mapping which move the activity elsewhere on the disk or by limiting the number of write cycles to a given sector. Alternatively, the system might simply monitor sectors for failures and then remove them from the logical record pool once they're identified as bad (while remapping the data elsewhere).

Radio
NEXT UPCOMING BROADCAST
EE Times Senior Technical Editor Martin Rowe will interview EMC engineer Kenneth Wyatt.
Top Comments of the Week
Like Us on Facebook

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Times on Twitter
EE Times Twitter Feed
Flash Poll