KR20080071366A - Apparatus and method for backup data considering temperature of nand flash memory in raid system having nand flash memories - Google Patents

Abstract

A device and a method for backing up data in consideration of temperature of NAND flash memories in a RAID(Redundant Array of Independent Disks) system equipped with the NAND flash memories are provided to back up the data stored in a block having high probability converted into a run-time bad block by sequentially storing the data from the block having the most data read/write frequency to a backup area when the temporary of the NAND flash memory is out of a normal operation temperature range. A temperature sensor(102) measures temperature of a NAND flash memory and outputs a measured temperature value. A memory(104) stores includes at least one NAND flash memory and a backup area. A controller(100) checks a data read/write frequency of all blocks forming the NAND flash memory and arranges the blocks in data read/write frequency order when the temperature value is out of a normal operation temperature range. The controller checks the number of blocks included in the backup area and stores the data copied from the arranged blocks as many as the checked number of blocks included in the backup area, and stores the copied data to the backup area.

Description

Data backup device and method considering temperature of NAND flash in RAID system with NAND flash {APPARATUS AND METHOD FOR BACKUP DATA CONSIDERING TEMPERATURE OF NAND FLASH MEMORY IN RAID SYSTEM HAVING NAND FLASH MEMORIES}

1 is an internal block diagram of a RAID system according to an embodiment of the present invention;

FIG. 2 is an internal block diagram of a memory unit including a backup memory device for data backup in the RAID system of FIG. 1; FIG.

3 is a block diagram illustrating an internal block of a memory unit including a virtual backup memory device for data backup in the RAID system of FIG. 1;

4 is a control flowchart illustrating a process for controlling data backup in consideration of a temperature of a NAND flash according to an embodiment of the present invention.

The present invention relates to a device and method for backing up data in a RAID (Redundant Array of Inexpensive Disk) system, in particular, in consideration of the temperature of the NAND flash in a RAID system having at least one NAND flash. An apparatus and method for backing up data are provided.

A general RAID system is a system in which several homogeneous or heterogeneous hard disks are connected and configured as one piece of hardware, and several homogeneous or heterogeneous hard disks are used as a single system memory. Also, in such a RAID system, one hard disk among a plurality of hard disks is used for backup, and the parity of each hard disk is stored in the backup hard disk. Then, if a physical error occurs in a predetermined hard disk, the RAID system may recover data using parity of each hard disk stored in the backup hard disk.

However, since hard disks used in raid systems are expensive, efforts have been made to construct raid systems using NAND flash, which is inexpensive and has high density instead of hard disks.

As described above, NAND flash used in a RAID system is generally treated as a bad block if an error of 2 bits or more occurs in the initial process, and data read / write cannot be performed for the bad block. Produced.

As described above, the NAND flash may have a bad block from an initial process, but in the initial process, a block that was a good block may be converted to a bad block while performing a data write / erase operation of 100,000 or more times. In this way, a block that was a good block in the initial process is later converted into a bad block is called a run-time bad block. On the other hand, since the data read / write operation is restricted for the block that is the bad block from the initial process, no error occurs when reading data. However, during the initial process, the data is stored in the block because it is a good block. If the block is later converted to a runtime bad block, the data stored in the block cannot be read.

When a block used as a good block performs more than 100,000 data write / erase operations and is converted to a run-time bad block, wear-leveling is performed on all blocks constituting each NAND flash. Ensure that data write / erase operations are performed evenly. The data stored in the block which has been written or erased more than 100,000 times is stored in the backup area.

However, since the capacity of the backup area to back up data is limited and cannot be stored for all the blocks that are likely to be converted to runtime bad blocks, the capacity of the backup area in the order of the blocks that are most likely to be converted to runtime bad blocks. Save it.

As described above, in a RAID system having a NAND flash, a backup operation is performed only on a block of the NAND flash which exceeds a preset number of data write / erase times. However, even in a block in which data write / erase operations have not been performed a predetermined number of times as described above, when the NAND flash temperature is out of the normal operating temperature range due to the surrounding environment, it is more likely to be converted to the runtime bad block.

However, at present, a RAID system having NAND flash performs data backup only by considering the number of data write / erases, and there is no method of performing data backup in consideration of NAND flash temperature.

Accordingly, the present invention provides an apparatus and method for backing up data stored in a block that is likely to be converted to a run time bad block in consideration of the NAND flash temperature in a RAID system having a NAND flash.

In accordance with an aspect of the present invention, there is provided a device for backing up data in consideration of a temperature of the NAND flash in a RAID system having at least one NAND flash and a backup area for data backup, wherein the temperature of the NAND flash is measured. A temperature sensor for outputting a temperature value, the memory unit including the at least one NAND flash and the backup area, and the temperature value input from the temperature sensor is outside the normal operating temperature range of the NAND flash. By checking the number of data write / erase operations for all blocks constituting the flash, the data write / erase operations are sorted in order of the largest number of data write operations, and the number of blocks in the backup area is checked after checking the number of blocks in the backup area. Block order with the most data write / erase operations among the sorted blocks To copy the data of the block is characterized in that a control unit for storing in the backup area.

The present invention also provides a method for backing up data in consideration of a temperature of the NAND flash in a RAID system having at least one NAND flash and a backup area for data backup, the method comprising: measuring the temperature of the NAND flash, Checking whether the measured temperature is within the normal operating temperature range of the NAND flash, and writing / clearing data for all blocks constituting each of the NAND flashes when the NAND flash temperature is outside the normal operating temperature range. Checking the number of operations and sorting the data in the order of the largest number of operations, checking the number of blocks of the backup area, and writing / writing data among the sorted blocks by the number of blocks of the checked backup area. The backup by copying the data of the block in order of the block with the most erase operation And storing the data in the area.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description and the annexed drawings, detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted.

The present invention proposes a method for backing up data stored in blocks that are likely to be converted to a runtime bad block in consideration of the NAND flash temperature in a RAID system having a NAND flash.

Next, referring to FIG. 1, a block configuration of a RAID system capable of backing up data stored in blocks that are likely to be converted into a runtime bad block according to the NAND flash temperature will be described. First, in order to control the data backup according to the temperature of the NAND flash according to an embodiment of the present invention, the temperature of the NAND flash must be measured, but it is difficult to directly measure the temperature of the NAND flash. Therefore, in the embodiment of the present invention, since the temperatures of the NAND flash and the printed circuit board (PCB) on which the NAND flash is mounted are almost similar, the temperature of the NAND flash is determined by measuring the temperature of the PCB.

The temperature sensor 102 may be composed of a thermistor, an IC temperature sensor, and measures the temperature of a specific device. In particular, in the present invention, the temperature sensor 102 measures the temperature of the PCB on which the NAND flash constituting the memory unit 104 is mounted, and outputs the measured temperature value to the controller 100.

The controller 100 performs an overall control operation of the raid system. In particular, the control unit 100 in the present invention checks whether the temperature value of the PCB input from the temperature sensor 102 is within the normal operating temperature range of the NAND flash. In this case, information on a temperature range in which the NAND flash can be normally operated is stored in a portion of the NAND flash during the initial process of the NAND flash constituting the memory unit 104. In addition, the normal operating temperature range of a typical NAND flash is -30 degrees to 85 degrees. Therefore, the controller 100 checks whether the temperature value is within the normal operable temperature range by using the temperature range previously stored in the NAND flash. At this time, if the temperature value of the PCB is outside the normal operating temperature range of the NAND flash, the controller 100 performs wear leveling on all blocks constituting each NAND flash to perform data write / erase operations evenly. . Thereafter, the controller 100 checks the number of data write / erase operations for all blocks constituting each NAND flash, and arranges the data write / erase operations in the order of the blocks having the most data write / erase operations. The controller 100 checks the number of blocks of the backup area for data backup, copies the data in the order of the blocks with the most data write / erase operations among the sorted blocks as many as the number of the checked backup area, and stores them in the backup area. do. In this case, copying a block means copying data stored in the block. In addition, when storing the copied block in the backup area, the controller 100 sequentially stores the first block constituting the backup area in order of the highest number of data write / erase operations.

Under the control of the controller 100, the buffer 106 temporarily stores the order information of the blocks arranged in the order of the blocks with the largest number of data write / erase operations confirmed by the controller 100. In addition, the buffer 106 functions to output the stored block order information in response to a request of the controller 100.

On the other hand, the memory unit 104 has at least one NAND flash, and these NAND flashes are mounted on the PCB to constitute the memory unit 104. The memory unit 104 according to the embodiment of the present invention includes a backup area for backing up data. In this case, the backup area may use a separate backup memory device or use a single virtual backup memory device configured by collecting reserved valid blocks of the NAND flash. The backup memory device separately provided herein may be one NAND flash or NOR flash. In addition, the memory unit 104 may collect the reserved valid blocks of the NAND flash and configure it as a virtual backup memory device to use as a backup area. Also, a homogeneous NAND flash refers to a NAND flash having the same number of reserved valid blocks, and a heterogeneous NAND flash refers to a NAND flash having different numbers of reserved valid blocks. The valid block refers to a good block, and the reserved valid block refers to a good block that is not arbitrarily used among the minimum valid blocks. In addition, the minimum valid block is a value suggested by each semiconductor manufacturer's specification (Spec: Specification), and refers to a value that the manufacturer guarantees that the number of blocks is a good block.

An internal configuration of the memory unit 104 having a backup memory device separately provided as described above may be illustrated in FIG. 2. In addition, the internal configuration of the memory unit 104 having one virtual backup memory device configured by collecting reserved valid blocks of the NAND flash may be illustrated as shown in FIG. 3.

On the other hand, if the backup area has the number of blocks equal to the maximum number of blocks calculated as in Equation 1 below, data stored in all blocks that are likely to be converted into a runtime bad block can be stored. However, in <Equation 1>, since the minimum number of effective blocks of one NAND flash depends on the type of NAND flash, in the above Equation 1, when the NAND flash having the same number of minimum effective blocks is used. Assume that

Maximum number of blocks in the backup area

= Number of NAND flashes * (the total number of blocks that one NAND flash has-minimum number of effective blocks that one NAND flash has)

The most ideal number of blocks of the backup area is to have the maximum number of blocks calculated as in Equation 1 above. However, since the size of the memory is limited, the backup area using a part of the memory is limited. Therefore, according to an exemplary embodiment of the present invention, a method of selecting and backing up the backup area having a limited number of blocks by the size of the backup area in the order of the blocks having the greatest probability of generating a runtime bad block will be proposed below.

The process of storing the data stored in the block in the order of the blocks which are likely to be converted to the runtime bad block in the backup area of the memory unit 104 configured as shown in FIGS. 2 and 3 will be described with reference to FIG. This will be described in detail in the description.

Next, a process of controlling the data backup by considering the temperature of the NAND flash in the RAID system configured as shown in FIG. 1 will be described with reference to FIG. 4.

First, in step 400 the control unit 100 receives the temperature value of the PCB measured by the temperature sensor 102 to check the temperature value of the PCB. Thereafter, the controller 100 proceeds to step 402 to check whether the temperature value of the PCB is outside the normal operating temperature range of the NAND flash. That is, the controller 100 checks whether the temperature of the NAND flash is outside the temperature range in which the NAND flash can operate normally. Here, the normal operating temperature range of the NAND flash is stored in the NAND flash constituting the memory unit 104 during an initial process, and the temperature range value is -30 degrees to 85 degrees.

If the test result in step 402, the PCB temperature value is out of the normal operating temperature range, the control unit 100 proceeds to step 403, otherwise proceeds to step 413 to perform a normal backup operation and then ends the data backup operation. In this case, the general backup operation means that if a block in which the number of data write / erase operations exceeds a preset number of blocks constituting each NAND flash occurs, wear leveling is performed for all blocks constituting each NAND flash. Copy the data saved in the file and save it in the backup area.

On the other hand, in step 402, the control unit 100 performs wear leveling on all blocks constituting each NAND flash in step 403 where the temperature value of the PCB is outside the normal operating temperature range of the NAND flash. Ensure write / erase operations are performed evenly. Thereafter, the controller 100 proceeds to step 404 to check the number of data write / erase operations of all blocks constituting each NAND flash. In operation 406, the controller 100 sorts all blocks constituting each NAND flash in the order of the blocks having the most data write / erase operations, and stores the order information of the sorted blocks in the buffer 106.

In operation 408, the controller 100 checks the number of blocks of the backup area included in the memory unit 104. If the memory unit 104 includes a separate backup memory device as a backup area, the controller 100 checks the size of the backup memory device in step 408. For example, assume that the memory unit 104 includes four NAND flash NAND # 1, NAND # 2, NAND # 3, NAND # 4, and one backup memory device as a backup area as shown in FIG. In this case, in FIG. 2, if the number of blocks of the backup memory device has four minimum valid blocks, the backup memory device may back up data stored in the four blocks.

In another embodiment, it is assumed that the memory unit 104 collects the reserved valid blocks of each NAND flash as shown in FIG. 3 and configures it as a virtual backup memory device to use as a backup area. Check the total number of valid blocks reserved for each NAND flash. For example, M NAND flashes NAND # 1, NAND # 2, NAND # 3, ... which have N blocks as shown in FIG. , NAND # M, and assume that the number of reserved valid blocks included in each NAND flash is three. The number of blocks of the virtual backup memory device is then 3 * M.

After that, the controller 100 proceeds from step 408 to step 410 by copying the data of the corresponding block in the order of the blocks with the most data write / erase operations among the sorted blocks as many as the number of blocks in the backup area. Stores the backup areas sequentially from the first block that makes up the backup area. For example, in the memory unit 104 including a backup memory device as a backup area as shown in FIG. 2, the number of data write / erase operations is the highest among the blocks of NAND # 1, NAND # 2, NAND # 3, and NAND # 4. It is assumed that the order of sorting blocks from small blocks to small blocks is block # 1-1> block # 2-3> block # 3-1> block # 4-5. At this time, the controller 100 stores the data stored in block # 1-1 in the first block of the backup memory device, the data stored in block # 2-3 in the second block, and the block # 3-1 in the third block. The stored data is stored in the fourth block, and the data stored in block # 4-5 is stored.

As another embodiment, as shown in FIG. 3, regions A, B, C, and D of the memory unit 104 that collect reserved valid blocks of the plurality of NAND flash units constituting the memory unit 104 and use them as one virtual backup memory. Suppose you store data in order. At this time, the control unit 100 sequentially stores the data stored in the M * (N-3) blocks arranged in the order of the smallest blocks in the block with the most data write / erase operations from the block with the most data write / erase operations. Store in A, B, C, D area.

After that, the controller 100 proceeds to step 412 and stores the backup address management table in which the block address of the NAND flash data is backed up and the block address of the backup area where the data of the NAND flash is backed up are stored in the memory unit 104. do. After the data backup process ends, the controller 100 continues the process from step 400 to step 412 again.

Meanwhile, in the above description of the present invention, the embodiments have been described, but various modifications may be made within the scope of the present invention. In particular, in the exemplary embodiment of the present invention, the above-described data backup method may be applied to a raid system composed of a plurality of heterogeneous NAND flashes.

As described above, in the RAID system having the NAND flash, when the NAND flash temperature is outside the normal operating temperature range, the backup area is sequentially sequentially starting from the block with the largest number of data write / erase operations by the number of blocks in the backup area. Store in Therefore, the present invention has an advantage of backing up data stored in a block that is likely to be converted to a runtime bad block in consideration of the NAND flash temperature.

Claims (10)

An apparatus for backing up data in consideration of a temperature of the NAND flash in a RAID system having at least one NAND flash and a backup area for data backup, the apparatus comprising: A temperature sensor for measuring the temperature of the NAND flash and outputting the measured temperature value; A memory unit including the at least one NAND flash and the backup area; If the temperature value input from the temperature sensor is outside the normal operating temperature range of the NAND flash, the number of data write / erase operations is checked by checking the number of data write / erase operations for all blocks constituting the NAND flash. After sorting in block order, checking the number of blocks in the backup area, copying the data of the corresponding block in the order of the blocks with the most data write / erase operations among the sorted blocks by the number of blocks in the checked backup area. And a control unit for storing in the area. The method of claim 1, wherein the control unit, The storing of the data of the copied block in the backup area further includes sequentially storing the first block of the backup area in the order of the highest number of data write / erase operations. . The data backup apparatus of claim 1, wherein the at least one NAND flash unit provided in the memory unit is mounted on a printed circuit board (PCB). The method of claim 3, wherein the temperature sensor, And determining a temperature of the NAND flash by measuring a temperature value of the PCB on which the NAND flash is mounted. The method of claim 1, wherein the control unit, And performing a wear leveling operation on all blocks of each NAND flash when the temperature of the NAND flash is out of the normal operable temperature range to perform data write / erase operations evenly. The method of claim 1, wherein the backup area, A data backup device, characterized in that a separate backup memory device composed of one of NAND flash or NOR flash. The method of claim 1, wherein the backup area, And a virtual backup memory device configured by collecting reserved valid blocks of each NAND flash. A method for backing up data in consideration of a temperature of the NAND flash in a RAID system having at least one NAND flash and a backup area for data backup, the method comprising: Measuring a temperature of the NAND flash and checking whether the measured temperature is within a normal operating temperature range of the NAND flash; When the temperature of the NAND flash is outside the normal operating temperature range, the number of data write / erase operations is checked for all blocks constituting each NAND flash, and the data write / erase operations are arranged in order of the largest number of blocks. Process, Checking the number of blocks in the backup area, copying the data of the corresponding block in order of the blocks with the most data write / erase operations among the sorted blocks as many as the number of blocks in the backup area, and storing the data in the backup area; Data backup method comprising the. 10. The method of claim 8, wherein the storing of the copied block data in the backup area is sequentially performed from the first block of the backup area in the order of the highest number of data write / erase operations. . The method of claim 8, And performing a wear leveling operation on all blocks of each NAND flash when the temperature of the NAND flash is outside the normal operating temperature range so that a data write / erase operation is evenly performed. .

Images