MD4331-d1G-MECH M-Systems Inc., MD4331-d1G-MECH Datasheet - Page 51

• Manufacturer Part Number: MD4331-d1G-MECH
• Description: Mobile Diskonchip G3 Data Sheet
• Manufacturer: M-Systems Inc.

To overcome this inherent deficiency, TrueFFS uses M-Systems’ patented wear-leveling algorithm.
This wear-leveling algorithm ensures that consecutive writes of a specific sector are not written
physically to the same page in the flash. This spreads flash media usage evenly across all pages,
thereby maximizing flash lifetime. TrueFFS wear-leveling extends the flash lifetime 10 to 15 years
beyond the lifetime of a typical application.
Dynamic Wear-Leveling
TrueFFS uses statistical allocation to perform dynamic wear-leveling on newly written data. This
not only minimizes the number of erase cycles per block, it also minimizes the total number of erase
cycles. Because a block erase is the most time-consuming operation, dynamic wear-leveling has a
major impact on overall performance. This impact cannot be noticed during the first write to flash
(since there is no need to erase blocks beforehand), but it is more and more noticeable as the flash
media becomes full.
Static Wear-Leveling
Areas on the flash media may contain static files, characterized by blocks of data that remain
unchanged for very long periods of time, or even for the whole device lifetime. If wear-leveling
were only applied on newly written pages, static areas would never be cycled. This limited
application of wear-leveling would lower life expectancy significantly in cases where flash memory
contains large static areas. To overcome this problem, TrueFFS forces data transfer in static areas as
well as in dynamic areas, thereby applying wear-leveling to the entire media.
7.1.6 Power Failure Management
TrueFFS uses algorithms based on “erase after write” instead of "erase before write" to ensure data
integrity during normal operation and in the event of a power failure. Used areas are reclaimed for
erasing and writing the flash management information into them only after an operation is
complete. This procedure serves as a check on data integrity.
The “erase after write” algorithm is also used to update and store mapping information on the flash
memory. This keeps the mapping information coherent even during power failures. The only
mapping information held in RAM is a table pointing to the location of the actual mapping
information. This table is reconstructed during power-up or after reset from the information stored
in the flash memory.
To prevent data from being lost or corrupted, TrueFFS uses the following mechanisms:
•
•
7.1.7 Error Detection/Correction
TrueFFS implements a unique MLC-tailored Error Correction Code (ECC) algorithm to ensure data
reliability. Refer to Section 3.7 for further information on the EDC/ECC mechanism.
48
When writing, copying, or erasing the flash device, the data format remains valid at all
intermediate stages. Previous data is never erased until the operation has been completed and
the new data has been verified.
A data sector cannot exist in a partially written state. Either the operation is successfully
completed, in which case the new sector contents are valid, or the operation has not yet been
completed or has failed, in which case the old sector contents remain valid.
Data Sheet, Rev. 1.0
Mobile DiskOnChip G3
91-SR-011-05-8L