JP6820060B2 - Storage device, control method and program of storage device - Google Patents

Description

The present invention relates to a storage device, a control method and a program of the storage device. In particular, the present invention relates to a storage device having a data compression function, a control method and a program of the storage device.

In recent years, with the increase in data storage capacity, a storage device having a data compression function has been used (see Patent Document 1). In such a storage device, the controller compresses the write data received from the host terminal into blocks of a certain size, and writes the compressed data to the physical memory which is a recording medium. Further, in response to a read request from the host terminal, the controller reads the compressed data in block units from the physical memory, decompresses it, and then responds to the host terminal as read data.

Special Table 2015-518586

The following analysis was made from the point of view of the present invention. The disclosure of the above prior art documents shall be incorporated into this document by citation.

The storage device disclosed in Patent Document 1 has a problem that the response performance of the storage device to a read request is lowered.

That is, the controller of the storage device may receive a read request for small-sized data such as metadata from the host terminal. In this case, the controller will decompress a size of data that is larger than the size of the requested data. As a result, the load on the controller increases, and the response performance of the storage device to the read request deteriorates.

On the other hand, the controller of the storage device may receive a read request for large-sized data such as sequential data from the host terminal. In this case, the controller needs to read the compressed data divided into a large number of blocks in block units and decompress each data block. As a result, the load on the controller becomes large, and the response performance of the storage device deteriorates.

Therefore, an object of the present invention is to provide a storage device having improved response performance to a read request, a control method and a program of the storage device.

According to the first aspect of the present invention
A management table for managing the input / output patterns of write data received from the host terminal,
An input / output pattern analysis means that analyzes the input / output pattern of the write data and writes the analysis result to the management table.
A block size determining means for determining the size of the block for compressing the write data with reference to the management table, and
A compression means that compresses the write data to the block size determined by the block size determining means and writes the compressed data to the physical memory.
A storage device is provided.

According to the second aspect of the present invention
An input / output pattern analysis step that analyzes the input / output pattern of write data and writes the analysis result to the management table for managing the input / output pattern of the write data received from the host terminal.
A block size determination step for determining the size of the block for compressing the write data with reference to the management table, and
A compression step of compressing the write data to the block size determined in the block size determination step and writing the compressed data to the physical memory.
A method of controlling a storage device including is provided.

According to the third aspect of the present invention
Input / output pattern analysis processing that analyzes the input / output pattern of write data and writes the analysis result to the management table for managing the input / output pattern of write data received from the host terminal.
The block size determination process for determining the size of the block for compressing the write data with reference to the management table,
A compression process in which the write data is compressed to the block size determined in the block size determination process and the compressed data is written to the physical memory.
Is provided with a program that causes the computer to execute.

According to each viewpoint of the present invention, a storage device having improved response performance to a read request, a storage device control method, and a program are provided.

It is a figure for demonstrating the outline of one Embodiment. It is a block diagram which shows one configuration of the storage device of 1st Embodiment. It is a figure which showed an example of the information stored in the management table 120. It is a figure which showed an example of the information stored in the address translation map 130. It is a figure which shows the processing flow by the storage apparatus 10 when the write data is received from the host terminal 30. It is a figure which shows the processing flow about compression and storage of write data by a storage apparatus 10. It is a figure which shows the processing flow by the storage apparatus 10 when the read request is received from the host terminal 30. It is a figure which shows the process flow about the creation and update of the address translation map 130 of the 2nd Embodiment. It is a figure which showed an example of the information stored in the management table 120 of the 3rd Embodiment. It is a figure which shows the processing flow about compression and storage of write data by the storage apparatus 10 of 3rd Embodiment.

Preferred embodiments of the present invention will be described in detail with reference to the drawings. It should be noted that the drawing reference reference numerals added to the following description are added to each element for convenience as an example for assisting understanding, and the present invention is not intended to be limited to the illustrated embodiment. Further, the connecting line between the blocks in each figure includes both bidirectional and unidirectional.

As shown in FIG. 1, the storage device 10 is connected to the physical memory group 20 and has a storage unit 100 that stores data necessary for information processing and a processing unit 200 that performs information processing. The physical memory group 20 has a plurality of physical memories 21 for storing compressed data, and each physical memory 21 is connected to the storage device 10 via a bus or the like. The storage unit 100 includes a management table 120 for managing input / output patterns of write data received from the host terminal.

The processing unit 200 is also connected to a host terminal (not shown) and has an I / O (input / output) pattern analysis unit 210, a block size determination unit 220, and a data compression / decompression unit 230. The I / O pattern analysis unit 210 is also referred to as an input / output pattern analysis means, and the block size determination unit 220 is also referred to as a block size determination means. Further, the data compression / decompression unit 230 executes a part of the function as the compression means.

The I / O pattern analysis unit 210 analyzes the input / output pattern of the write data received from the host terminal and writes the analysis result in the management table 120. The block size determination unit 220 determines the size of the block for compressing the write data with reference to the management table 120. The data compression / decompression unit 230 compresses the write data to a size determined by the block size determination unit 220. Then, the storage device 10 writes the compressed data to the physical memory 21.

To give a specific example, the management table 120 stores the access pattern (sequential access count, random access count, average I / O size) from the host terminal for the write data received from the host terminal. The I / O pattern analysis unit 210 analyzes the access pattern from the host terminal for the write data received from the host terminal, and writes the analysis result in the management table 120.

The block size determination unit 220 refers to the management table 120, and determines, for example, the block size as "average I / O size * average sequential access count" for write data whose sequential access count is larger than the random access count. To do. Further, the block size determination unit 220 determines the block size to be "average I / O size" for write data whose sequential access count is less than the random access count.

The data compression / decompression unit 230 compresses the write data to the block size determined by the block size determination unit 220. Then, the storage device 10 writes the compressed data to the physical memory 21. Further, when the storage device 10 receives a read request from the host terminal, the storage device 10 reads the compressed data from the physical memory 21, decompresses the compressed data by the data compression / decompression unit 230, and then responds to the host terminal as read data. To do.

As described above, the storage device 10 tends to receive a read request with a small size, in other words, the write data having a large number of random accesses is compressed into a small size block and written to the physical memory 21. On the other hand, the storage device 10 tends to receive a read request with a large size, in other words, the write data having a large number of sequential accesses is compressed into a large size block and written to the physical memory 21.

Therefore, when the storage device 10 responds to a read request for data having a small size, the storage device 10 has a high tendency to decompress data having a size as required. Further, when the storage device 10 responds to a read request for large-sized data, the number of blocks of compressed data read from the physical memory 21 can be reduced. Therefore, the load on the data compression / decompression unit 230 is reduced, and as a result, the response performance of the storage device 10 to the read request is improved.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

[First Embodiment]
As shown in FIG. 2, the storage device 10 is connected to the physical memory group 20 and the host terminal 30, and has a storage unit 100, a processing unit 200, a host control unit 300, and a memory control unit 400. The physical memory group 20 has a plurality of physical memories 21 such as an HDD (hard disk drive) and an SDD (solid state drive) for storing compressed data, and each physical memory 21 is connected to a storage device 10 via a bus or the like. Will be done. The physical memory 21 can also be mounted inside the storage device 10. The host terminal 30 is a computer or the like connected via a network, and transmits write data, a read request, and the like to the storage device 10.

The storage unit 100 is a ROM (read only memory), a RAM (random access memory), or the like, and specifically includes a cache memory 110 that stores pre-compressed data 111 and post-compressed data 112. The storage unit 100 also stores a management table 120 for managing the input / output patterns of write data received from the host terminal, and an address translation map 130 for associating the addresses of the logical volume and the physical volume. The management table 120 and the address translation map 130 can also be stored in the cache memory 110.

As shown in FIG. 3, the management table 120 stores the number of sequential accesses, the number of random accesses, and the average I / O size for each area of each logical volume 311. The number of sequential accesses and the number of random accessess are also referred to as access patterns. The average I / O size is also referred to as an average input / output size.

The address translation map 130 is composed of a logical volume map (FIG. 4A) and a physical volume map (FIG. 4B). The logical volume map stores the address of the logical volume and the address of the physical volume in association with each other, and the physical volume map stores the address of the physical volume and the address of the physical memory 21 constituting the physical volume in association with each other. To do. Note that FIG. 4C is a diagram conceptually showing the information stored in the logical volume map and the physical volume map.

The processing unit 200 includes a CPU (Central Processing Unit) and a chip, is connected to a storage unit 100 and a host control unit 300, and has an I / O pattern analysis unit 210, a block size determination unit 220, and data compression / It has an extension portion 230.

The I / O pattern analysis unit 210 analyzes the input / output pattern of the write data received from the host terminal 30, and writes the analysis result in the management table 120. The block size determination unit 220 determines the size of the compressed data block with reference to the management table 120. The data compression / decompression unit 230 compresses the write data to a size determined by the block size determination unit 220.

The host control unit 300 has a plurality of logical volumes 311 managed by the logical volume management unit 310. The logical volume 311 is recognized as an I / O target by the host terminal 30. The memory control unit 400 has a function of controlling the physical memory group 20, and has a plurality of physical volumes 411 managed by the physical volume management unit 410. The memory control unit 400 executes a part of the function as the compression means.

Hereinafter, a series of processing flows by the storage device 10 and processing by each functional unit will be described with reference to the drawings.

FIG. 5 is a diagram showing a processing flow by the storage device 10 when write data is received from the host terminal 30. The host control unit 300 receives a write command and write data from the host terminal 30 (step S11). For example, the host control unit 300 receives a write command from the host terminal 30 to write 4KB (kilobytes) of write data to the area 3 of the logical volume 311 (# 0). When the host control unit 300 receives the write data, the I / O pattern analysis unit 210 analyzes the I / O pattern of the received received light data (step S12). Specifically, the I / O pattern analysis unit 210 analyzes the number of sequential accesses, the number of random accesses, and the average I / O size as the I / O pattern. For example, when the I / O pattern analysis unit 210 receives the write data from the host terminal 30, it is a sequential access when the address has continuity as compared with the write data previously received from the same source. Is determined. On the other hand, the I / O pattern analysis unit 210 determines that the access is random when the addresses are not continuous.

Then, the I / O pattern analysis unit 210 stores the I / O pattern in the management table 120 for each area of each logical volume 311 (step S13). Specifically, the I / O pattern analysis unit 210 reflects the above determination result in the management table 120. Then, the I / O pattern analysis unit 210 stores the write data in the cache memory 110 as uncompressed data 111 in association with the logical volume address (step S14). After that, the I / O pattern analysis unit 210 responds to the host terminal 30 with the normal end of the writing process (step S15).

FIG. 6 is a diagram showing a processing flow related to compression and storage of write data by the storage device 10. When the memory control unit 400 detects that the uncompressed data is stored in the cache memory 110 (step S21, YES), the memory control unit 400 instructs the block size determination unit 220 to determine the block size. Upon receiving an instruction from the memory control unit 400, the block size determination unit 220 acquires the logical volume address associated with the uncompressed data 111 (step S22). Then, the block size determination unit 220 determines the size of the compressed data block (steps S24 to 26) with reference to the management table 120 (step S23).

For example, in the management table 120 shown in FIG. 2, the block size determination unit 220 determines that the number of sequential accesses is less than the number of random accesses for the area 3 of the logical volume 311 (# 0) (step S24, YES). .. Here, the block size determination unit 220 determines the block size to 4KB, which is the average I / O size of the region 3 of the logical volume 311 (# 0) (step S25).

On the other hand, in the management table 120 shown in FIG. 2, the block size determination unit 220 determines that the number of sequential accesses is larger than the number of random accesses for the area 2 of the logical volume 311 (# 0) (step S24, NO). ). Here, the block size determination unit 220 determines the block size to be the “average I / O size * average number of sequential accesses” of the area 2 of the logical volume 311 (# 0) (step S26). Here, the average number of sequential accesses corresponds to the average value of the number of command issuances that occur in one sequential access. For example, it is assumed that a file of 100 MB or 200 MB is stored in the area 2 of the logical volume 311 (# 0), and 1 MB of data is read each time a command is issued. At this time, reading a 100 MB file requires issuing 100 commands, and reading a 200 MB file requires issuing 200 commands. The average value of the number of issuances of these commands corresponds to the average number of sequential accesses, and the block size determination unit 220 determines the block size of the area 2 of the logical volume 311 (# 0) to 512KB * 150≈75MB.

Then, the block size determination unit 220 returns the size of the block determined as the logical volume address to the memory control unit 400. When the memory control unit 400 receives the block size from the block size determination unit 220, the memory control unit 400 instructs the data compression / decompression unit 230 to compress the uncompressed data 111 to the logical volume address and the size of the received block. When the data compression / decompression unit 230 receives an instruction from the memory control unit 400, the data compression / decompression unit 230 reads the uncompressed data 111 from the cache memory 110, compresses the data 111 to a specified block size, and stores the compressed data 112 in the cache memory 110. (Step S27). When the compressed data 112 is stored, the memory control unit 400 refers to the address conversion map 130 to the compressed data 112 to the address of the physical memory 21 associated with the logical volume address received from the block size determining unit 220. Is stored (step S28).

FIG. 7 is a diagram showing a processing flow by the storage device 10 when a read request is received from the host terminal 30. The host control unit 300 receives a read request from the host terminal 30 (step S31). For example, the host control unit 300 receives a read request from the host terminal 30 to read the data stored in the area 3 of the logical volume 311 (# 0). When the host control unit 300 receives the read request, the I / O pattern analysis unit 210 analyzes the I / O pattern of the requested data (step S32) and updates the management table 120 (step S33). Specifically, the I / O pattern analysis unit 210 analyzes whether the read request is sequential access or random access, and updates the access pattern of the management table 120. Further, the I / O pattern analysis unit 210 analyzes the size of the read-requested data and updates the average I / O size of the management table 120.

The memory control unit 400 reads the compressed data from the address of the physical memory 21 corresponding to the logical volume address specified in the read request with reference to the address conversion map 130, and stores the compressed data in the cache memory 110 as the compressed data 112. (Step S34). Then, the memory control unit 400 instructs the data compression / decompression unit 230 to decompress the compressed data. When the data compression / decompression unit 230 receives an instruction from the memory control unit 400, the data compression / decompression unit 230 reads the compressed data 112 from the cache memory 110 and decompresses it (step S35), and stores the data in the cache memory 110 as pre-compression data 111. Then, the host control unit 300 transmits the uncompressed data 111 stored in the cache memory 110 to the host terminal 30 (step S36).

As described above, the storage device 10 of the first embodiment has a high tendency to decompress data of a size as needed when responding to a read request for data having a small size. Further, when the storage device 10 responds to a read request for large-sized data, the number of blocks of compressed data read from the physical memory 21 can be reduced. Therefore, the load on the data compression / decompression unit 230 is reduced, and as a result, the response performance of the storage device 10 to the read request is improved.

[Second Embodiment]
In the first embodiment, the address translation map 130 is created and updated as appropriate. Therefore, as the second embodiment, the creation and update processing of the address translation map 130 will be described in detail. As shown in FIG. 8, the memory control unit 400 creates a new entry for the write data when the entry regarding the logical volume address of the uncompressed data 111 does not exist in the address translation map 130 (step S41, NO). (Step S42). Further, when the entry exists in the address translation map 130 (step S41, YES), the memory control unit 400 determines whether or not the address of the corresponding physical memory 21 has a capacity for storing the compressed data of the write data. (Step S43). If it is determined that there is no capacity (step S43, NO), the memory control unit 400 allocates a new address of the physical memory 21 and updates the address translation map 130 (step S44). When it is determined that there is a capacity (step S43, YES), the memory control unit 400 updates the address of the existing physical memory 21 to an address having a capacity that can store the compressed data of the write data (step S45).

In this way, the address translation map 130 is appropriately created and updated.

[Third Embodiment]
The size of the compressed data block may be determined based on the type of write data. In this embodiment, the management table 120 stores the type of write data in association with the input / output pattern. Specifically, as shown in FIG. 9, the management table 120 stores the number of sequential accesses, the number of random accesses, the average I / O size, and the data type for each area of each logical volume 311. The data type is information indicating the type of write data, such as "metadata" and "streaming data".

The I / O pattern analysis unit 210 analyzes the type of write data received from the host terminal 30. Then, the I / O pattern analysis unit 210 writes the analysis result as a data type in the management table 120.

When the write data is metadata, the block size determination unit 220 determines the block size to the average input / output size. If the write data is streaming data, the block size is determined by the average input / output size * average number of sequential accesses.

FIG. 10 is a diagram showing a processing flow relating to compression and storage of write data by the storage device 10 of the third embodiment. The processing related to the compression and storage of the write data by the storage device 10 of the third embodiment is the same as the processing flow shown in FIG. 6 except that the size of the compressed data block is determined based on the type of the write data. The differences will be described below.

For example, in the management table 120 shown in FIG. 9, the block size determination unit 220 determines that the write data type is metadata for the area 3 of the logical volume 311 (# 0) (steps S54, YES). Here, the block size determination unit 220 determines the size of the block to be 4 KB, which is the average I / O size of the region 3 of the logical volume 311 (# 0) (step S55).

On the other hand, in the management table 120 shown in FIG. 9, the block size determination unit 220 determines that the write data type is streaming data for the area 2 of the logical volume 311 (# 0) (steps S54, NO). .. Here, the block size determination unit 220 determines the size of the block to be the “average I / O size * average number of sequential accesses” of the area 2 of the logical volume 311 (# 0) (step S56).

As described above, when the storage device 10 responds to the read request for metadata, the storage device 10 has a high tendency to decompress data having a size as required. Further, when the storage device 10 responds to a read request for streaming data, the number of blocks of compressed data read from the physical memory 21 can be reduced. Therefore, the load on the data compression / decompression unit 230 is reduced, and as a result, the response performance of the storage device 10 to the read request is improved.

The storage device 10 is realized by, for example, a computer equipped with a processing module. The processing module is realized, for example, by the CPU executing a program stored in the memory. In addition, the program can be downloaded via a network or updated using a storage medium in which the program is stored. Further, the processing module may be realized by a semiconductor chip. That is, the function performed by the processing module may be realized by some hardware and / or software.

Some or all of the above embodiments may also be described, but not limited to:

(Appendix 1)
A management table for managing the input / output patterns of write data received from the host terminal,
An input / output pattern analysis means that analyzes the input / output pattern of the write data and writes the analysis result to the management table.
A block size determining means for determining the size of the block for compressing the write data with reference to the management table, and
A compression means that compresses the write data to the block size determined by the block size determining means and writes the compressed data to the physical memory.
Storage device with.

(Appendix 2)
The storage device according to Appendix 1, wherein the management table stores an input / output pattern including the number of sequential accesses, the number of random accesses, and the average input / output size.

(Appendix 3)
The storage device according to Appendix 2, wherein for write data in which the number of sequential accesses is larger than the number of random accesses, the block size determining means determines the block size as the average input / output size * the average number of sequential accesses.

(Appendix 4)
The storage device according to Appendix 2, wherein the block size determining means determines the block size to the average input / output size for write data in which the number of sequential accesses is less than the number of random accesses.

(Appendix 5)
The management table stores the type of write data in association with the input / output pattern.
The storage device according to Appendix 2, wherein the input / output pattern analysis means analyzes the type of write data and writes the analysis result in the management table.

(Appendix 6)
The storage device according to Appendix 5, wherein when the write data is metadata, the block size determining means determines the block size as the average input / output size.

(Appendix 7)
The storage device according to Appendix 5, wherein when the write data is streaming data, the block size determining means determines the block size by the average input / output size * the average number of sequential accesses.

(Appendix 8)
An input / output pattern analysis step that analyzes the input / output pattern of write data and writes the analysis result to the management table for managing the input / output pattern of the write data received from the host terminal.
A block size determination step for determining the size of the block for compressing the write data with reference to the management table, and
A compression step of compressing the write data to the block size determined in the block size determination step and writing the compressed data to the physical memory.
How to control storage devices, including.

(Appendix 9)
Input / output pattern analysis processing that analyzes the input / output pattern of write data and writes the analysis result to the management table for managing the input / output pattern of write data received from the host terminal.
The block size determination process for determining the size of the block for compressing the write data with reference to the management table,
A compression process in which the write data is compressed to the block size determined in the block size determination process and the compressed data is written to the physical memory.
A program that causes a computer to run.

The disclosure of the above patent documents shall be incorporated into this document by citation. Within the framework of the entire disclosure (including the scope of claims) of the present invention, it is possible to change or adjust the embodiments or examples based on the basic technical idea thereof. Further, various combinations or selections of various disclosure elements (including each element of each claim, each element of each embodiment or embodiment, each element of each drawing, etc.) within the scope of the claims of the present invention. Is possible. That is, it goes without saying that the present invention includes all disclosure including claims, and various modifications and modifications that can be made by those skilled in the art in accordance with the technical idea.

10 Storage device 20 Physical memory group 21 Physical memory 30 Host terminal 100 Storage unit 110 Cache memory 111 Pre-compressed data 112 Post-compressed data 120 Management table 130 Address conversion map 200 Processing unit 210 I / O pattern analysis unit 220 Block size determination unit 230 Data compression / decompression unit 300 Host control unit 310 Logical volume management unit 311 Logical volume 400 Memory control unit 410 Physical volume management unit 411 Physical volume

Claims (6)

A cache memory for storing the write data received from the host terminal and the compressed data after the write data is compressed, and a storage unit for storing a management table for managing the input / output pattern of the write data .
An input / output pattern analysis unit that analyzes the input / output pattern of the write data and writes the analysis result to the management table.
A block size determination unit that determines the size of the block related to the compressed data with reference to the management table,
A compression unit that reads the write data from the cache memory, compresses the data, and stores the created compressed data in the cache memory .
When it is detected that the write data is stored in the cache memory, the block size determination unit is instructed to determine the block size for the compressed data, and the compression unit is instructed to determine the block size. The size of the block determined by is specified to create the compressed data of the specified block size, and when the compressed data is subsequently stored in the cache memory, the compressed data is stored in the physical memory. Memory control unit and
Storage device with. The storage device according to claim 1, wherein the management table stores an input / output pattern including the number of sequential accesses, the number of random accesses, and the average input / output size. The storage device according to claim 2, wherein for write data in which the number of sequential accesses is larger than the number of random accesses, the block size determining unit determines the block size as the average input / output size * the average number of sequential accesses. The storage device according to claim 2, wherein for write data in which the number of sequential accesses is less than the number of random accesses, the block size determining unit determines the block size to be the average input / output size. An input / output pattern analysis step that analyzes the input / output pattern of write data and writes the analysis result to the management table for managing the input / output pattern of the write data received from the host terminal.
When it is detected that the write data is stored in the cache memory, the block size determination step of determining the block size of the compressed data after compressing the write data with reference to the management table, and
The write data is read from the cache memory, and the write data is compressed to the block size determined in the block size determination step to create compressed data, and the compressed data is stored in the cache memory. With compression steps
A memory control step for storing the compressed data stored in the cache memory in the physical memory, and
How to control storage devices, including. Input / output pattern analysis processing that analyzes the input / output pattern of write data and writes the analysis result to the management table for managing the input / output pattern of write data received from the host terminal.
When it is detected that the write data is stored in the cache memory, the block size determination process of determining the block size of the compressed data after compressing the write data with reference to the management table, and
The write data is read from the cache memory and the write data is compressed to the block size determined in the block size determination process to create compressed data, and the compressed data is stored in the cache memory. Compression processing and
Memory control processing that stores the compressed data stored in the cache memory in the physical memory, and
A program that causes the computer that controls the storage device to execute.

Images