KR-20260063662-A - SSD Dedicated RAID System for Efficient Resource Management
Abstract
The present invention relates to a RAID system dedicated to SSDs for resource management, and more specifically, to a RAID system dedicated to SSDs for resource management that improves resource optimization and data recovery functions in a RAID system including SSD devices of various capacities. According to the present invention, the problem of wasted storage space occurring when SSD devices of different capacities are combined into a RAID5 system can be resolved by including: a block information collection unit that collects block information from each SSD device and provides SSD device resource information, such as page size, block size, number of blocks, and number of chips, to a RAID system; a grouping unit that configures a logical storage device by allocating block resources to be included in each RAID group among n SSD devices based on the block information collected through the block information collection unit; and a block allocation unit that allocates blocks using a circular allocation method, which distributes data including parity to each SSD device in a RAID5 manner when the block resources are identical, and distributes resources sequentially and evenly when the resources are different, in order to distribute resources evenly according to the resource status of each SSD device.
Inventors
- 이현섭
Assignees
- 백석대학교산학협력단
Dates
- Publication Date
- 20260507
- Application Date
- 20241030
Claims (6)
- A block information collection unit (100) that collects block information from each SSD device and provides SSD device resource information, such as page size, block size, number of blocks, and number of chips, to a RAID system, A grouping unit (200) that configures a logical storage device by allocating block resources to be included in each RAID group among n SSD devices based on block information collected through the block information collection unit above, and A RAID system dedicated to SSDs for resource management, characterized by including a block allocation unit (300) that, in order to distribute resources evenly according to the resource status of each SSD device, distributes data including parity to each SSD device using a RAID5 method when the block resources are the same, and allocates blocks using a circular allocation method that distributes resources evenly in sequence when the resources are different.
- In paragraph 1, The above-described grouping unit utilizes block resources within each SSD device to form m RAID groups from n SSD devices, wherein each RAID group is first formed with n RAID groups based on the RAID size that can be allocated by SSD devices with insufficient resources, and subsequently, mn groups are allocated from each SSD in a circular manner from SSD devices with remaining resources to additionally form logical RAID groups, characterized by an SSD-dedicated RAID system for resource management.
- In paragraph 1, The above block allocation unit is characterized by configuring a RAID group by sequentially allocating resources from each SSD device according to resource status for RAID group configuration, and allocating additional resources from another SSD device with sufficient resources when resources are insufficient, in a RAID system dedicated to SSD resource management.
- In paragraph 1, The above block information collection unit is, A RAID system dedicated to SSDs for resource management, characterized by receiving chip and flash memory information as shown in the figure via the Get Block Information command, and setting up Internal RAID on each SSD using the Set RAID Chip command when Internal RAID is required.
- In paragraph 1, The above grouping unit is, A RAID system dedicated to SSDs for resource management, characterized by including a command system capable of setting or controlling the resources of an SSD device in a RAID system so as to form additional RAID groups for internal RAID configuration based on the resources within the SSD.
- In a method using an SSD-dedicated RAID system for resource management, (a) A step in which the RAID system collects block information from each SSD device and transmits SSD resource information, such as page size, block size, number of blocks, and number of chips, to the RAID system; (b) a step of creating a logical storage device by selecting an SSD resource to be allocated to each RAID group based on the block information received by the RAID system; (c) A step of optimizing resources by allocating blocks using a circular allocation method that distributes data including parity to each SSD device in a RAID5 manner when the RAID system has the same block resources according to the resource status of each SSD device, and distributes resources sequentially and evenly when the resources are different; and (d) A method using an SSD-dedicated RAID system for resource management, characterized by including the step of configuring a RAID group by minimizing the allocation to SSD devices with insufficient resources based on block resources included in the RAID group and allocating additional resources to SSD devices with available resources.
Description
SSD Dedicated RAID System for Efficient Resource Management The present invention relates to a RAID system dedicated to SSDs for resource management, and more specifically, to a RAID system dedicated to SSDs for resource management that improves resource optimization and data recovery functions in a RAID system including SSD devices of various capacities. IoT (Internet of Things) technology has become diverse and advanced, evolving across various fields ranging from small embedded systems to those processing large volumes of data. Recently, numerous IoT devices have been generating vast amounts of data, providing a new dimension of technological environment for corporate data collection and analysis. In enterprise environments, highly reliable fault recovery systems are being implemented to reliably store and manage this massive volume of collected data. In enterprise environments, RAID (Redundant Array of Independent Disks) systems are applied to reliably recover data from data loss and failures. In particular, RAID 5 ensures space efficiency and stability by distributing parity across multiple storage devices. However, when storage devices have different capacities, the RAID is configured based on the smallest capacity device, resulting in wasted storage space. In enterprise storage systems, it is essential to reliably store and manage large volumes of data. In such systems, a structure capable of rapid recovery is required in the event of data loss or failure. To this end, RAID (Redundant Array of Independent Disks) systems are widely used, and they are designed to simultaneously provide data reliability and storage space efficiency. Figure 1 shows the structure of RIAD 5. RAID is a method of storing data by configuring multiple storage devices into a single logical storage device. There are various types of RAID depending on how the logical storage devices are configured. In the structure of Figure 1, four SSDs are grouped into a single logical storage device. In Group A, parity data is assigned to SSD No. 4. In Group B, parity data is assigned to SSD No. 3. Thus, RAID 5 stores parity data by allocating one of the storage devices in the logical groups. Figure 2 illustrates the parity data generation and fault recovery process of RAID. RAID parity data is generated through XOR operations. As shown in the figure, the parity of data 1, 2, and 3 (111111, 111000, and 101010) becomes 101101 through XOR operations. Then, when data 3 is lost, data 3 is recovered by XORing data 1 and 2 with the parity data. In the example in the figure, 101010 was recovered by XORing 111111, 111000, and 101101. Figure 3 illustrates the problems that occur when storage devices of different capacities are combined into a RAID. Figure 3 illustrates the resource waste problem that occurs when storage devices of different capacities—100GB, 150GB, 200GB, and 200GB—are combined into a RAID. RAID requires grouping identical data spaces into a single logical group. Therefore, as shown in Figure 3, the RAID group is configured with the smallest capacity storage device. Although a total of 650GB of resources are combined into a RAID, the RAID is configured in units of 100GB, which is the smallest capacity within the group. Consequently, a problem arises where 400GB of logical storage is created, leaving 250GB of resources unused. Among the various levels of such RAID systems, RAID 5 is widely adopted, particularly for its ability to guarantee space efficiency and stability. RAID 5 protects data by distributing and storing parity information across each storage device. This parity information is necessary to reconstruct the original data; in the event of an error in a specific storage device or data loss, this parity data can be utilized to perform recovery. Consequently, data remains intact and can be restored, thereby enhancing system reliability. However, since RAID systems are configured by combining multiple storage devices into a single array, problems can arise when the capacities of the devices differ. Although RAID consists of storage devices of the same capacity, some systems may necessitate the use of a mix of devices with different capacities. For example, capacities may differ when updating an existing RAID5 configuration or adding new storage devices. In such cases, the RAID system configures the array based on the storage device with the smallest capacity. For example, if three storage devices of 1TB, 2TB, and 3TB are configured as RAID 5, the RAID is set based on the 1TB device. Consequently, even though the 2TB and 3TB devices have more than 1TB of capacity remaining, the remaining capacity is wasted without being used. This reduces storage space efficiency and leads to the problem of inefficient resource usage. Therefore, there is a need to resolve the issue of wasted storage space by optimizing RAID5 configurations that include SSDs of various capacities. Figure 1 shows the structure of RIAD 5. Figure 2 illustrates the pa