摘要 |
A fault tolerance system for one or two failed disks in a disk array includes a CPU, a disk array, and a bus. The disk array includes disks, each of which is logically divided into multiple blocks, wherein the blocks include data blocks, P parity blocks and Q parity blocks. The CPU, which is connected to the disk array through the bus, includes: an exclusive-or (XOR) unit for performing XOR operations on blocks of the disk array when generating P/Q parities or reconstructing failed data; a modulus operation unit for performing modulus operations; a shift operation unit for performing shift operations on the blocks of the disk array; and an address conversion unit for converting a logic address into a physical address. Related methods are also provided.
|