I like this idea. The Flash Transaction Layer (FTL) which is programmed into the firmware of current SSDs (for compatibility) is rubbish. You basically wind up with a device that has all the pitfalls of both pancake-stack storage and flash.

I'm hoping that we can re-flash our old firmwares when this is stable. There's also some Open Sauce SSD filesystems in the works :)
It might be right that there is "No seek time" or very little. The slow down was caused on the File System and Operating system level, not the hardware level. There are many checking and correcting procedures during a write. For different FS, writing speed and results are different. Try it for yourself.

So, changing/improving a file system for flash based memory may help as described by Sandisk. It makes sense. Whether the new file system works as described is another matter, but believe me, file system does affect performance.
Maybe SanDisk should have been bought, as write speed, u[p till now, has been 1/2 read speed. So SanDisk1, All Others 0.

Also, Like Idea of Pulling Up Its'(sandisk)Technology By its' Own BootStraps.
TS Drashek
The FFS stuff will be done at a firmware level and can no doubt be done pretty cheaply. The major cost of a decent sized flash drive will still be the flash memory itself.

The point is that existing filesystems (NTFS, FAT, ext3 for you Linux boys) were written for hard disks. Sandisk is just trying to balance that. Flash's one weakpoint is lots of small random writes.

Just one more nail in the hard disk coffin. I was argueing this with Seagate investors a year ago. Since then their shares fell 60%+ and I made a mint!
Intel already has this technology, or something very close to it. and OCZ's core 3 series will likely be very very fast as well.
How long will it take their competetors to reverse engi.. I mean "EXAMINE" it and make their own version?
I like this idea. The Flash Transaction Layer (FTL) which is programmed into the firmware of current SSDs (for compatibility) is rubbish. You basically wind up with a device that has all the pitfalls of both pancake-stack storage and flash.

I'm hoping that we can re-flash our old firmwares when this is stable. There's also some Open Sauce SSD filesystems in the works :)
It might be right that there is "No seek time" or very little. The slow down was caused on the File System and Operating system level, not the hardware level. There are many checking and correcting procedures during a write. For different FS, writing speed and results are different. Try it for yourself.

So, changing/improving a file system for flash based memory may help as described by Sandisk. It makes sense. Whether the new file system works as described is another matter, but believe me, file system does affect performance.
It doesn't make any difference WHERE on the disk you put a block. It's not a hard drive--there's no "seek time".
Err, its a file system. It should be free.
Maybe SanDisk should have been bought, as write speed, u[p till now, has been 1/2 read speed. So SanDisk1, All Others 0.

Also, Like Idea of Pulling Up Its'(sandisk)Technology By its' Own BootStraps.
TS Drashek
The FFS stuff will be done at a firmware level and can no doubt be done pretty cheaply. The major cost of a decent sized flash drive will still be the flash memory itself.

The point is that existing filesystems (NTFS, FAT, ext3 for you Linux boys) were written for hard disks. Sandisk is just trying to balance that. Flash's one weakpoint is lots of small random writes.

Just one more nail in the hard disk coffin. I was argueing this with Seagate investors a year ago. Since then their shares fell 60%+ and I made a mint!
I can only imagine how much more this will cost...