TWI718519B - Data storage system and management method thereof - Google Patents

Abstract

The invention discloses a data storage system and management method thereof. The data storage system according to the invention includes N storage devices, an auxiliary memory and a controller where N is an integer equal to or larger than 3. The controller functions in receiving data requested to be stored in the data storage system, performing a predetermined operation for the received data to generate user data and parity data. When the auxiliary memory is not full, the controller writes the user data into a plurality of user data storage areas of the N storage devices, and writes the parity data into the auxiliary memory.

Description

Data storage system and its management method

The present invention relates to a data storage system and its management method, and in particular, to a data storage system with high sequential performance and its management method.

With the increasing amount of data stored by users, data storage systems that conform to the RAID structure (also known as RAID systems) have been widely adopted to store large amounts of data. The RAID system can provide high availability, high performance or high volume data storage space to the host.

The structure of the conventional disk array system includes a disk array controller (RAID controller) and a disk array composed of multiple physical storage devices. The disk array controller is connected to each physical storage device, and the disk array controller defines the disk array as one of RAID 0, RAID 1, RAID 2, RAID 3, RAID 4, RAID 5, or RAID 6. Achieved by multiple logical disk drives. The disk array controller can generate (rebuild) redundant data that is the same as the data to be read.

In practical applications, each physical storage device can be a tape drive, a disk drive, a memory drive, an optical storage drive, or the same A magnetic area in a disk drive corresponding to a single read/write head, or other equivalent physical storage device.

By using different redundancy/data storage schemes for each RAID level, RAID can be implemented at different levels. For example, RAID 1 implements disk mirroring, in which the first storage device stores the stored data, and the second storage device stores an exact copy of the data stored in the first storage device. If any storage device is damaged, because the data in the remaining storage device is still available, no data is lost.

In other RAID systems, each physical storage device is divided into multiple data blocks. From the perspective of fault tolerance, it can be divided into user data blocks and parity data blocks. The user data block stores general user data. The parity data block stores an extra set of parity data for use in inverse calculation of user data when there is a need for fault tolerance. Corresponding user data blocks and parity data blocks existing in different physical storage devices form a storage stripe, where the parity data in the parity data block is used by the user data block The data is the result of the exclusive OR (XOR) operation. If a physical storage device in these RAID systems is damaged, the user data and parity data stored in the remaining undamaged physical storage devices can be used to perform exclusive or (XOR) operations for reconstruction. It should be stated that those who are familiar with this art understand that the calculation of data in a parity data block can not only use exclusive or (XOR) operations, but also include various parity operations or similar operations. Technology, as long as the following relationship exists: the data in any data block in the same storage strip can be calculated from the data in the corresponding data block. The disk array architecture defines the number of redundancy as the number of identical data blocks in a storage stripe.

If the data storage system follows the traditional RAID mode, N storage devices and K redundant messages are used. When this data storage system performs continuous access, only (N-K) storage devices provide sequential performance (read/write).

For example, please refer to Figure 1. The prior art data storage system 1 follows the RAID 5 mode (the number of redundant messages is 1) and uses 4 storage devices (12a~12d). Figure 1 depicts data stored in 4 storage devices ( Within 12a~12d) Storage format. As shown in Figure 1, each storage device (12a, 12b, 12c, 12d) is divided into multiple data blocks, which can be divided into user data blocks and parity data blocks. The user data block stores general user data. The parity data block stores an extra set of parity data for use in inverse calculation of user data when there is a need for fault tolerance. In the case shown in FIG. 1, the storage device 12a includes at least user data blocks D0, D3, D6, and a parity data block P3; the storage device 12b includes at least user data blocks D1, D4, D9 and parity. Data block P2; storage device 12c includes at least user data blocks D2, D7, D10 and parity data block P1; storage device 12d includes at least user data blocks D5, D8, D11 and parity data block P0 Wait. Corresponding user data blocks and co-located data blocks existing in different storage devices form a storage partition, where the data in the co-located data block is subjected to mutually exclusive or (XOR) operation by the data in the user data block The result obtained. In the case shown in Figure 1, the user data blocks D0, D1, D2 and the parity data block P0 form a storage partition S1, and the data in P0 is the mutually exclusive OR operation of the data in D0, D1, and D2 result. Similarly, the user data blocks D3, D4, D5 and the parity data block P1 form another storage partition S2, and the data in P1 is the mutually exclusive OR operation result of the data in D3, D4, and D5. Similarly, the user data blocks D6, D7, D8 and the parity data block P2 form another storage partition S3, and the data in P2 is the exclusive OR (XOR) operation of the data in D6, D7, D8 result. Similarly, the user data blocks D9, D10, D11 and the parity data block P3 form another storage partition S4, and the data in P3 is the mutually exclusive OR operation result of the data in D9, D10, and D11. It should be stated that those who are familiar with this technique understand that the calculation of data in a parity data block can not only use mutual exclusion or arithmetic, but also include a variety of parity operations or similar arithmetic techniques, as long as The following relationship exists: the data in any data block in the same storage partition can be calculated from the data in the corresponding data block. Obviously, the block storing parity data does not provide sequential performance. Therefore, the sequential performance of the data storage system 1 is provided by three storage devices.

For another example, the data storage system follows the RAID 0 mode (the number of redundant messages is 0) and uses 4 storage devices. Therefore, the sequential performance of the data storage system is provided by 4 storage devices.

For another example, the data storage system follows the RAID 6 mode (the number of redundant messages is 2) and uses 4 storage devices. Therefore, the sequential performance of the data storage system is provided by 2 storage devices.

At present, the prior art still has great room for improvement on how to improve the sequential performance of data storage systems that follow the RAID mode.

The present invention adopts auxiliary memory and changes the position where the parity data is written and the arrangement of the written data to improve the sequential performance of the data storage system.

For the prior art of data storage system using auxiliary memory, please refer to US Patent No. 5,948,110. US Patent No. 5,948,110 discloses a method of using non-volatile memory to provide error correction for disk array systems. Non-volatile memory is used to increase the speed of RAID recovery in case of disk errors. This is achieved by keeping a list of all disk blocks whose parity data may be inconsistent. However, US Patent No. 5,948,110 does not teach the sequential performance of the data storage system, and its purpose is different from the purpose of the present invention.

Therefore, the technical problem to be solved by the present invention is to provide a data storage system with high sequential performance and a management method thereof, especially for a data storage system conforming to the disk array architecture. And in particular, the data storage system according to the present invention uses auxiliary memory and changes the location where the parity data is written and the arrangement of the written data to improve the sequential performance of the data storage system.

A data storage system in a preferred embodiment of the present invention includes N storage devices, auxiliary memory, and a controller, where N is greater than or An integer equal to 3. Each storage device is divided into M storage areas, where M is an integer greater than or equal to 2. Alternatively, the odd-numbered storage area among the M storage areas of each storage device is planned as a user data storage area, and the even-numbered storage area is planned as a parity data storage area, or each storage device Among the M storage areas, the even-numbered storage area is planned as a user data storage area, and the odd-numbered storage area is planned as a parity data storage area. Each storage device has a first read and write speed. The auxiliary memory has a second read and write speed, and in particular, the second read and write speed of the auxiliary memory is higher than the first read and write speed of each storage device. The controller is electrically connected to N storage devices and auxiliary memory. The controller is used to receive multiple pieces of data to be stored in the data storage system, and perform predetermined operations on the multiple pieces of data to be stored to generate multiple pieces of user data and multiple pieces of parity data. When the auxiliary memory is not full, the controller writes multiple pieces of user data into multiple user data storage areas of the N storage devices, and writes multiple pieces of parity data into the auxiliary memory.

Further, when the auxiliary memory is full, the controller writes multiple pieces of user data into multiple user data storage areas of N storage devices, and writes multiple pieces of parity data into as many as N storage devices Within a storage area of parity data. Subsequently, the controller moves the multiple parity data stored in the auxiliary memory to the parity storage area of the N storage devices.

In a specific embodiment, the auxiliary memory may be a non-volatile memory or a volatile memory equipped with an independent power supply.

In a specific embodiment, taking RAID5 as an example, the first storage space range of each user data storage area is 3MB-15GB. The second storage space of each parity data storage area ranges from 1MB to 5GB, and the first storage space and the second storage space maintain the ratio of (N-K):K.

In a specific embodiment, the predetermined operation may be an exclusive OR (XOR) operation.

In a specific embodiment, the controller controls the data access of the N storage devices according to the disk array architecture, and the disk array architecture defines the number of redundant messages as K, where K is a positive integer. The ratio of the storage space occupied by multiple pieces of user data to the storage space occupied by multiple pieces of parity data is equal to (N-K):K.

The management method of a preferred embodiment of the present invention is directed to a data storage system. The data storage system includes N storage devices and auxiliary memory, where N is an integer greater than or equal to 3. Each storage device is divided into M storage areas, where M is an integer greater than or equal to 2. Alternatively, the odd-numbered storage area among the M storage areas of each storage device is planned as a user data storage area, and the even-numbered storage area is planned as a parity data storage area, or each storage device Among the M storage areas, the even-numbered storage area is planned as a user data storage area, and the odd-numbered storage area is planned as a homogeneous data storage area. Each storage device has a first read and write speed. The auxiliary memory has a second read and write speed, and in particular, the second read and write speed of the auxiliary memory is higher than the first read and write speed of each storage device. The method according to the present invention first receives a plurality of data to be deposited that are required to be deposited in the data storage system. Then, according to the method of the present invention, multiple pieces of data to be stored are executed by predetermined operations to generate multiple pieces of user data and multiple pieces of parity data. Then, according to the method of the present invention, it is determined whether the auxiliary memory is full. If it is determined that the auxiliary memory is not full, according to the method of the present invention, multiple pieces of user data are written into multiple user data storage areas of N storage devices, and multiple pieces of parity data are written into the auxiliary memory .

Further, if it is determined that the auxiliary memory is full, according to the method of the present invention, multiple pieces of user data are written into multiple pieces of user data storage areas of N storage devices, and multiple pieces of parity data are written into N pieces The multiple parity data storage areas of the storage device are subsequently moved to the parity storage areas of N storage devices.

Compared with the prior art, the data storage system according to the present invention adopts auxiliary memory and changes the location where the parity data is written and the arrangement of the written data, which significantly improves the sequential performance of the data storage system.

The advantages and spirit of the present invention can be further understood from the following detailed description of the invention and the accompanying drawings.

1‧‧‧Data Storage System

12a~12d‧‧‧Storage device

2‧‧‧Data Storage System

20‧‧‧Controller

22a~22n‧‧‧Storage device

220‧‧‧Storage area

222‧‧‧User data storage area

224‧‧‧Parallel data storage area

24‧‧‧Transmission interface

26‧‧‧Auxiliary memory

3‧‧‧Access request application unit

4‧‧‧Management methods

S40~S48‧‧‧Process steps

Figure 1 is a schematic diagram of the storage format of data stored in a prior art data storage system that follows the RAID 5 mode and uses 4 storage devices.

FIG. 2 is a schematic diagram of the structure of a data storage system according to a preferred embodiment of the present invention.

FIG. 3 is a schematic diagram of the storage format of data stored in the data storage system according to the present invention that follows the RAID 5 mode and uses 4 storage devices and auxiliary memory.

4 is a schematic diagram of the storage format in the data storage system according to the present invention when the auxiliary memory of the example shown in FIG. 3 is full.

Fig. 5 is a flowchart of a management method according to a preferred embodiment of the present invention.

Please refer to FIG. 2. The structure of the data storage system 2 according to a preferred embodiment of the present invention is shown in FIG. 2.

As shown in FIG. 2, the data storage system 2 according to the present invention includes N storage devices (22a-22n), a controller 20, and an auxiliary memory 26, where N is an integer greater than or equal to 3. The controller 20 is electrically connected to the N storage devices (22a-22n) and the auxiliary memory 26, respectively.

The controller 20 can generate (rebuild) redundant data that is the same as the data to be read. For example, under the RAID 5 architecture, the controller 20 is mutually exclusive Or (XOR) operation to generate redundant data.

In a specific embodiment, the controller 20 controls the data access of the N storage devices (22a-22n) according to the disk array architecture. The disk array architecture can be RAID 5 or RAID 6, etc.

In a specific embodiment, each storage device (22a~22n) can be a tape drive, a magnetic disk drive, a memory device, an optical storage recording device, or a single read/write head in the same drive One of the sectors, or other equivalent storage devices.

It is also shown in FIG. 2, and FIG. 2 shows the access request application unit 3. The access request application unit 3 is connected to the controller 20 via the transmission interface 24. In practical applications, the access request application unit 3 can be a network computer, a mini computer, a mainframe, a notebook computer, or any electronic device that needs to read data in the data storage system 2 according to the present invention, for example, Mobile phones, personal digital assistants, digital recording equipment, digital music players, etc.

When the access request application unit 3 is an independent electronic device, it can be accessed through storage area network (SAN), local area network (LAN), serial advanced technology (serial ATA, SATA) interface, fiber channel (FC) ), a transmission interface such as a small computer standard interface (SCSI), or an input/output (I/O) interface such as PCI Express is connected to the data storage system 2 according to the present invention. In addition, when the access request application unit 3 is a special integrated circuit component, or other equivalent device capable of sending I/O reading requests, it can send data reading requests based on commands (or requests) from other devices To the controller 20, and then read the data in the storage devices (22a-22n) through the controller 20.

The controller 20 and the storage devices (22a-22n) in the data storage system 2 according to the present invention can be placed in a single box at the same time, or can be separated and placed in different boxes. In practice, the controller 20 can use Fibre Channel (FC), small computer system interface (SCSI), serial connection SCSI (SAS), serial ATA (SATA), parallel ATA (PATA, or IDE) The transmission interface is connected with the storage devices (22a-22n). If the storage devices (22a~22n) are disk drives, each storage device (22a~22n) can be FC, SAS, SCSI, SATA, PATA and other disk drives with various interfaces. The controller 20 can be a disk array controller or a controller with a data storage system capable of generating redundant data.

Also shown in FIG. 2, each storage device (22a-22n) is divided into M storage areas 220, where M is an integer greater than or equal to 2. Alternatively, the odd-numbered storage area 220 of the M storage areas 220 of each storage device (22a-22n) is planned as the user data storage area 222, and the even-numbered storage area 220 is planned as the parity data The storage area 224 is shown in FIG. 2. Or, among the M storage areas 220 of each storage device (22a-22n), the even-numbered storage area 220 is planned as a user data storage area 222, and the odd-numbered storage area 220 is planned as a parity data storage area 224 It is also possible.

Each storage device (22a-22n) has a first read and write speed. The auxiliary memory 26 has a second read and write speed, and in particular, the second read and write speed of the auxiliary memory 26 is higher than the first read and write speed of each storage device (22a-22n).

When the user terminal wants to write data into the data storage system 2 according to the present invention, the controller 20 receives through the transmission interface 24 a plurality of data to be stored transmitted by the application unit 3 of the access request. The controller 20 executes predetermined calculations on multiple pieces of data to be stored to generate multiple pieces of user data and multiple pieces of parity data. Different from the prior art, when the auxiliary memory 26 is not full, the controller 20 writes multiple pieces of user data into multiple user data storage areas 222 of N storage devices (22a-22n), and more The parity data is written into the auxiliary memory 26.

Please refer to Figure 3. The data storage system 2 according to the present invention follows the RAID 5 mode (the number of redundant messages is 1) and uses 4 storage devices (22a-22d), Figure 3 depicts the storage format of data stored in 4 storage devices (22a-22d) and auxiliary memory 26. The example shown in FIG. 3 is that of the M storage areas 220 of each storage device (22a-22n), the odd-numbered storage area 220 is planned as a user data storage area 222, and the even-numbered storage area 220 is planned as a parity element. Data storage area 224. The storage device 22a stores the user data D0, D4, and D8 from the start address of the logical address; the storage device 22b stores the user data D1, D5, D9 from the start address of the logical address; the storage device 22c stores the user data D1, D5, D9 from the start address of the logical address. The starting address of the logical address stores the user data D2, D6, and D10; the storage device 22d stores the user data D3, D7, and D11 starting from the starting address of the logical address. The auxiliary memory 26 stores the parity data P0, P1, P2, P3 from the starting address of the logical address. The parity data storage area 224 of the four storage devices (22a-22d) has not yet stored parity data.

Each user data storage area 222 is divided into multiple user data blocks. Each parity data storage area 224 is divided into a plurality of parity data blocks. In a specific embodiment, the first storage space of each user data storage area 222 ranges from 3 MB to 15 GB. The second storage space of each parity data storage area 224 ranges from 1MB to 5GB.

In a specific embodiment, the predetermined operation may be an exclusive OR (XOR) operation. The data in the parity data block is the result of the exclusive OR (XOR) operation performed on the data in the user data block. It should be stated that those who are familiar with this technique understand that the calculation of data in the parity data block can not only use exclusive or (XOR) operations, but also include various parity operations. Or similar computing technology, as long as the following relationship exists: the data in any data block in the same storage partition can be calculated from the data in the corresponding data block.

In a specific embodiment, the auxiliary memory 26 may be a non-volatile memory or a volatile memory equipped with an independent power source (for example, a battery). The storage space of the auxiliary memory 26 can be smaller than that of each storage device (22a~22n) storage space, but not limited to this.

Further, when the auxiliary memory 26 is full, the controller 20 writes multiple pieces of user data into the multiple user data storage areas 222 of the N storage devices (22a-22n), and writes multiple pieces of parity data Write into a plurality of parity data storage areas 224 of N storage devices (22a-22n). Subsequently, the controller 20 moves the multiple parity data stored in the auxiliary memory 26 to the parity storage area 224 of the N storage devices (22a-22n). Please refer to FIG. 4, the structure of the example shown in FIG. 4 is the same as the structure of the example shown in FIG. In the example shown in FIG. 4, the auxiliary memory 26 is full. The multiple parity data P0, P1, P2, and P3 stored in the auxiliary memory 26 are moved to the parity data storage area 224 of 4 storage devices (22a-22d).

As described above, the controller 20 controls the data access of the N storage devices (22a-22n) according to the disk array architecture, and the disk array architecture defines the number of redundant messages as K, where K is a positive integer. The ratio of the storage space occupied by multiple pieces of user data to the storage space occupied by multiple pieces of parity data is equal to (N-K):K.

Please refer to FIG. 4. FIG. 4 is a flowchart of a management method 4 according to a preferred embodiment of the present invention. The management method 4 according to the present invention is directed to, for example, the management method of the data storage system 2 in FIG. 2. The structure of the data storage system 2 according to the present invention has been described in detail above, and will not be repeated here. However, it should be emphasized that in the data storage system 2 of the present invention, the even-numbered storage area 220 among the M storage areas 220 of each storage device (22a-22n) is planned as a user data storage area 222, and the odd-numbered storage area 220 The storage area 220 may also be planned as a parity data storage area 224.

As shown in FIG. 4, the management method 4 according to the present invention first executes step S40, which is to receive a plurality of data to be deposited in the data storage system 2 according to the present invention.

Then, according to the management method 4 of the present invention, step S42 is executed, It performs predetermined operations (for example, exclusive or (XOR) operations) on multiple data to be stored to generate multiple user data and multiple parity data.

Next, according to the management method 4 of the present invention, step S44 is executed to determine whether the auxiliary memory 26 is full.

If the result of step S44 is that it is determined that the auxiliary memory 26 is not full, according to the management method 4 of the present invention, step S46 is executed to write multiple user data into multiple users of N storage devices (22a~22n) In the data storage area 222, multiple par-bit data is written into the auxiliary memory 26.

Further, if the result of step S44 is that it is determined that the auxiliary memory 26 is full, according to the management method 4 of the present invention, step S48 is executed to write multiple user data into multiple use of N storage devices (22a~22n) In the data storage area 222, and write multiple parity data into the multiple parity data storage areas 224 of N storage devices (22a-22n), and then store the multiple parity data in the auxiliary memory 26 The metadata is moved to the same location storage area 224 of N storage devices (22a-22n).

Please refer to Table 1 and Table 2. The reading speeds measured by a test method for a number of comparative examples and an example of the present invention are shown in Table 1, and a number of comparative examples and an example of the present invention are measured by the test method. The obtained write speed series are shown in Table 2. The units of read speed and write speed are both MB/sec. A number of comparative examples include a single disk, 4 disks conforming to the traditional RAID 0 structure, and 4 disks conforming to the traditional RAID 5 structure. An example of the present invention is that 4 disks conform to the RAID 5 structure. The tool used in the test method is IOmeter, and the parameters: IO size is 1MB, outstanding number is 16, 100% sequential write or 100% sequential read.

Table 1

The data in Table 1 and Table 2 confirm that the read speed and write speed of 4 disks according to the present invention and follow the RAID 5 architecture are significantly higher than the read speed and write speed of 4 disks that conform to the traditional RAID 5 architecture. , Even close to 4 disks meet the read speed and write speed of the traditional RAID 0 architecture. This proves that, compared with the prior art, it is obvious that the data storage system according to the present invention has a high sequential performance.

Through the above detailed description of the preferred embodiments of the present invention, it can be clearly understood that the data storage system according to the present invention adopts auxiliary memory and changes the location where the parity data is written and the arrangement of the written data, which significantly improves data storage Sequential performance of the system.

Based on the above detailed description of the preferred embodiments, it is hoped that the characteristics and spirit of the present invention can be described more clearly, rather than limiting the aspect of the present invention by the preferred embodiments disclosed above. On the contrary, the purpose is to cover various changes and equivalent arrangements within the scope of the patent for which the present invention is intended. Therefore, the aspect of the patent scope applied for by the present invention should be interpreted in the broadest way based on the above description, so that it covers all possible The change of energy and the arrangement with equality.

2‧‧‧Data Storage System

20‧‧‧Controller

22a~22n‧‧‧Storage device

220‧‧‧Storage area

222‧‧‧User data storage area

224‧‧‧Parallel data storage area

24‧‧‧Transmission interface

26‧‧‧Auxiliary memory

3‧‧‧Access request application unit

Claims (10)

A data storage system comprising: N storage devices, N is an integer greater than or equal to 3, each storage device is divided into M storage areas, M is an integer greater than or equal to 2, alternatively, Among the M storage areas of each storage device, the odd-numbered storage area is planned as a user data storage area, and the even-numbered storage area is planned as a homogeneous data storage area, or the M storages of each storage device The even-numbered storage area in the area is planned as a user data storage area, and the odd-numbered storage area is planned as a single-bit data storage area. Each storage device has a first read and write speed; an auxiliary memory has a second data storage area. Two reading and writing speeds, the second reading and writing speed is higher than the first reading and writing speed; and a controller, which is electrically connected to the N storage devices and the auxiliary memory, respectively, for receiving the data requested to be stored Store multiple data to be stored in the system, and perform a predetermined operation on the multiple data to be stored to generate multiple user data and multiple parity data. When the auxiliary memory is not full, the The controller writes the plurality of user data into the plurality of user data storage areas of the N storage devices, and writes the plurality of parity data into the auxiliary memory. The data storage system according to claim 1, wherein when the auxiliary memory is full, the controller writes the plurality of user data into the plurality of user data storage areas of the N storage devices, And write the plurality of parity data into the storage areas of the plurality of parity data of the N storage devices Then, the controller moves the parity data stored in the auxiliary memory to the parity storage area of the N storage devices. The data storage system according to claim 2, wherein the auxiliary memory system is a non-volatile memory or a volatile memory equipped with an independent power supply. In the data storage system described in claim 3, the first storage space of each user data storage area ranges from 3MB to 15GB, and the second storage space of each parity data storage area ranges from 1MB to 5GB. The data storage system according to claim 3, wherein the predetermined operation is an exclusive or (XOR) operation, the controller controls the data access of the N storage devices according to a disk array architecture, and the disk array The architecture defines a number of redundant messages as K, K is a positive integer, and a ratio of the storage space occupied by the multiple pieces of user data to the storage space occupied by the multiple pieces of parity data is equal to (NK):K. A management method for a data storage system, the data storage system includes N storage devices and an auxiliary memory, N is an integer greater than or equal to 3, each storage device is divided into M storage areas, M is one An integer greater than or equal to 2, alternative between the two, the odd-numbered storage area in the M storage areas of each storage device is planned as a user data storage area, and the even-numbered storage area is planned as a single-bit data storage area Area, or the even-numbered storage area in the M storage areas of each storage device is planned as a user data storage area, and the odd-numbered storage area is planned as a single-bit data storage area, and each storage device has a first A read and write speed, the auxiliary memory has a second read and write speed, the second read and write speed is higher than the first read and write speed, the management method includes the following The following steps: (a) Receive multiple data to be deposited in the data storage system; (b) Perform a predetermined operation on the multiple data to be deposited to generate multiple user data and multiple parity Metadata; (c) Determine whether the auxiliary memory is full; and (d) If the result of step (c) is negative, write the multiple user data into the multiple of the N storage devices In a user data storage area, and write the multiple parity data into the auxiliary memory. The management method according to claim 6, further comprising the following steps: (e) If the judgment result of step (c) is affirmative, then write the multiple user data into the multiple use of the N storage devices In the data storage area, and write the multiple parity data into the multiple parity data storage areas of the N storage devices, and then move the multiple parity data stored in the auxiliary memory to In the co-bit storage area of the N storage devices. The management method according to claim 7, wherein the auxiliary memory system is a non-volatile memory or a volatile memory with an independent power supply. According to the management method described in claim 8, wherein the first storage space of each user data storage area ranges from 3MB to 15GB, and the second storage space of each parity data storage area ranges from 1MB to 5GB. The management method according to claim 8, wherein the predetermined operation is an exclusive or (XOR) operation, and the data access of the N storage devices is controlled according to a disk array architecture, and the disk array architecture defines a The number of redundant messages is K, and K is a positive integer. The storage space occupied by the multiple user data is equal to The storage space occupied by the multiple parity data is equal to (N-K):K.

Images