Garbage collection algorithm for NAND flash memory based on separation of hot and cold logical region

Aiming at the problem that the existing NAND flash garbage collection algorithms have insufficient considerations on collection efficiency and wear leveling and large memory overhead, a NAND flash garbage collection algorithm based on logical interval cold and hot separation is proposed. The proposed algorithm considers both collection efficiency and wear leveling, and adopts a victim block selection strategy which can dynamically adjust the ratio between collection efficiency and wear leveling. At the same time, a novel method of calculating the heat of the logical region is presented, and the valid data in victim block are divided into hot data, warm data and cold data according to the heat of the logical region. Thus, the valid data in the victim block can be effectively separated. The simulation results show that compared with GR, CB, CAT, FaGC and LRGC algorithms, the proposed algorithm not only achieves better results in wear leveling, total erasure times and total copy times, but also reduces the memory consumption greatly.