Realize raid in solid state memory

The present disclosure relates to systems and techniques include fault tolerance data stored in solid state memory. In some implementations, the method comprising: receiving data to be stored; data into the logical data block; blocks allocated to the logical block grouping, grouping the logical block comprises two memory devices from a plurality of solid-state physics of the at least one or more of physical data storage block; the physical data storage block in said block storage; determining persistent data corresponding to the code; and a memory storing code data corresponding to the logical block packet. Logical block and the logical block through a physical data storage block corresponding to the identifying corrupted packets, and the read data codes are stored in the identified packet, and the data stored in the code differs from the damaged block comparing the read data, the block data of the stored recoverable damage.