Get to know the features of the Kingston Industrial cards

Bad Block Management

Bad blocks contain one or more bits that have lost reliability. The bad blocks appear during the manufacturing process (Early Bad Blocks) or throughout the lifetime of the card (Later Bad Blocks). Both types of bad blocks are inevitable, making Bad Block Management a necessity to manage errors in NAND flash devices. Bad Block Management will identify and flag bad blocks, it will then use the free extra capacity to replace the invalid blocks. It will stop data from writing to the bad blocks which strengthens the reliability of the product. If the bad block has data, it will move the data to a valid block to prevent data loss.

ECC Engine

NAND flash memory must maintain data integrity as data moves from the host PC to the NAND storage via the flash controller. The data transfers from the host to the card are often referred to as “data in flight” or “data in transit” before they are actually written to the NAND flash storage. Flash controllers incorporate Error Correction technology (called ECC, which stands for Error Correction Code) to detect and correct most errors that can affect data along this trajectory. Flash memory chips incorporate additional error correction information along with every block of data that is written. This information allows the flash controller to simultaneously correct errors when reading a data block. NAND flash memory, like hard disk drives, will encounter bit errors during normal operation that it will correct on the fly with its ECC data. If a NAND device has excessive errors in a data block, then that block will be marked as a Bad Block, retired, then replaced with one of the spare blocks rotated into service. During this process, the data will be corrected if needed using ECC. The use of Spare Blocks extends the useful life and endurance of SSDs.

Power Failure Protection

Power loss is unavoidable and can cause havoc in a work environment if the proper hardware is not being used. Power Failure Protection is necessary to prevent data loss. A supported host device can send a command to the card that will halt any of its operations if it detects any drop in power. This allows the card time to save any data currently being written at the time of power loss.

Auto-Refresh Read Distribution Protection

The auto-refresh function reads the data on the flash memory, including those where data are rarely read out, and performs automatic error correction as required to prevent data losses caused by read disturb errors, data holding errors and other errors. The auto-refresh function is performed in the background so that it causes little delay in the response to commands even during the correction process.

Dynamic Data Refresh

Dynamic Data Refresh is employed to make sure that during read only operations, blocks with a high number of errors can be removed and refreshed for the next use. During each read command, the controller will perform a three stage check on the target block:

• First stage is to check for a “need to refresh” mark
• Second stage is to check for the number of error bits currently present
• Third stage is to check retry for the number of retry counts currently present

Garbage Collection

Garbage Collection is key for NAND flash to remain durable and maintain its speed. NAND flash-based devices cannot overwrite data that is already there. They must go through a Program/Erase cycle to write to an already used block of data. A NAND flash controller would first copy all valid data (that is still in use) and write it to empty pages of a different block. It will then erase all the cells in the current block (both valid and invalid data), to then start writing new data to the newly erased block. This process is called Garbage Collection

Wear Leveling

Kingston flash storage devices incorporate controllers utilizing advanced wear leveling technology which distributes the number of P/E cycles (program/erase) across the flash memory evenly around all blocks. When a block is needed to store data, the empty block with the lowest erase count is used. Wear-leveling thus extends the useful life of a flash memory card.