JP6668785B2 - Information processing system, storage control device, storage control method, and storage control program - Google Patents

Description

  The present invention relates to an information processing system, a storage control device, a storage control method, and a storage control program.

  In recent years, SSDs (Solid State Drives) using NAND flash memories (hereinafter referred to as NAND flash) as storage elements have become widespread as storage media. As for HDDs (Hard Disk Drives) conventionally used as storage media, HDDs employing SMR (Shingled Magnetic Recording) as a technology for increasing the capacity are becoming widespread. It is difficult for storage elements used in these storage media to store data discontinuously in a storage area. Therefore, many of these storage media store data in a storage area in a write-once format in order to facilitate data access from a computer.

  In such a write-once format, data is continuously stored in a free area irrespective of the storage destination address specified by the host computer using the storage medium. Therefore, the storage medium manages the correspondence between the address space referenced by the external host computer and the physical storage area on the storage element. The storage area storing the data invalidated by updating or deletion is reused as a free area that does not correspond to the address space referenced by the external host computer.

  However, it is impossible to reuse an empty area in a storage medium for storing data in a write-once format. For example, in a NAND flash used in an SSD, it is necessary to perform initialization by erasing (Erase) old stored data when updating data. Erase for the NAND flash is performed in units of blocks including a plurality of pages which are the minimum units of read and write. Similarly, in an HDD employing SMR, it is necessary to initialize the entire zone before reusing an area called a zone for continuously storing data. That is, the storage area to be initialized is required to be a continuous free area that does not include valid data.

  For this reason, in a storage medium such as an HDD or the like that employs an SSD or an SMR that stores data in a write-once format, a garbage collection (GC) that collects free space and secures a reusable continuous free space is performed. . In the GC, in order to secure a continuous free area, valid data included in the storage area to be collected is moved to another writable storage area prepared in advance, and the source storage area is changed to a free area. I do. Since the amount of data movement due to GC in the storage medium is proportional to the amount of valid data included in the storage area to be collected, it is necessary to store data from the host computer as the amount of valid data in the collection target area increases. Available capacity is reduced. That is, when the amount of data stored in the storage medium increases, or when the storage and update of data are repeated to fragment free space in the storage medium, the effective data amount included in the storage area to be collected increases. As a result, the frequency of GC for the amount of data stored from the host computer increases. An increase in the GC frequency causes a decrease in access performance from the host computer to the storage medium due to data movement in the storage medium.

  As a method of preventing performance degradation due to an increase in GC frequency in a storage medium, a method of eliminating fragmentation occurring in a storage area in advance can be cited. Specifically, the data stored continuously on the logical address space provided to the host computer by the storage medium is re-stored in a continuous area on the storage element, thereby fragmenting the data and the free area. Can be resolved. For example, in Patent Literature 1 and Non-Patent Literature 1, fragmentation is eliminated by rewriting an area with a low frequency of writing for each area obtained by dividing the logical address space provided by the SSD to the host computer, thereby eliminating internal fragmentation. A method for reducing the frequency of the defragmentation process is described. Eliminating fragmentation of the storage area inside the storage medium to prevent the distribution of free areas and to make continuous free areas likely to occur, thereby reducing the amount of effective data contained in the storage area to be collected. As a result, by reducing the effective data amount in the storage area to be collected, the capacity capable of storing data from the host computer increases, and the frequency of GC is reduced.

  For example, in the technique disclosed in Patent Literature 1, an area in which defragmentation is to be performed is determined by using a write frequency of each area obtained by dividing a logical address space provided to a host computer by a storage medium. Also, in Non-Patent Document 1, when random writing is performed on an SSD, the ratio of the amount of data movement accompanying GC performed inside the SSD to the amount of data written from the outside is calculated using a probability model. The method is shown.

Patent No. 5550741

X.-Y. Hu, E. Eleftheriou, R. Haas, I. Iliadis, and R. Pletka. "Write amplification analysis in flash-based solid state drives." In Proceedings of SYSTOR 2009: The Israeli Experimental Systems Conference, pp. .1-9

  However, in the technology described in the above-mentioned document that performs defragmentation processing on a storage medium that stores data in a write-once format, it is not possible to determine the area to be defragmented based on the effect of the defragmentation processing and the cost estimation. Not implemented. For example, in the technique disclosed in Patent Literature 1, the effect on GC when the fragmentation of each region is eliminated is not considered. In addition, Non-Patent Document 1 assumes that data storage from a host computer is performed uniformly and randomly, changes in Write access patterns from a host computer that vary depending on applications, etc., and data storage conditions inside an SSD. It is difficult to calculate according to.

  An object of the present invention is to provide a technique for solving the above-mentioned problem.

To achieve the above object, a storage control device according to the present invention comprises:
The data of each logical address area obtained by dividing the logical address space used when the host computer accesses the storage medium in units of a fixed size is stored in a plurality of physical address areas in the physical address space used inside the storage medium. Fragmentation management means for estimating whether or not fragmented and stored, and managing the number of the logical address areas estimated to be stored as fragmented,
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds the first threshold value, a defragmentation process for the logical address area estimated to be stored in a fragmented state is performed. Defragmentation processing means to be started;
Equipped with a,
The defragmentation processing means continues the defragmentation process until the number of the logical address areas estimated to be stored in a fragmented state falls below a second threshold,
Wherein the first threshold and the second threshold value, the data move accompanying the interior of garbage collection (GC) of the storage medium by the defragmentation process Ru is set to be reduced.

In order to achieve the above object, a storage control method according to the present invention comprises:
The data of each logical address area obtained by dividing the logical address space used when the host computer accesses the storage medium in units of a fixed size is stored in a plurality of physical address areas in the physical address space used inside the storage medium. Estimating whether or not it is stored in a fragmented manner, a fragmentation management step of managing the number of the logical address areas estimated to be stored in a fragmented manner,
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds a first threshold value, defragmentation is performed on the logical address area estimated to be stored in a fragmented state. A defragmentation processing step for starting processing;
Only including,
In the defragmentation processing step, the defragmentation processing is continued until the number of the logical address areas estimated to be fragmented and stored falls below a second threshold,
The first threshold value and the second threshold value are set such that data movement due to garbage collection (GC) inside the storage medium due to the fragmentation eliminating process is reduced .

In order to achieve the above object, a storage control program according to the present invention comprises:
The data of each logical address area obtained by dividing the logical address space used when the host computer accesses the storage medium in units of a fixed size is stored in a plurality of physical address areas in the physical address space used inside the storage medium. Estimating whether or not it is stored in a fragmented manner, a fragmentation management step of managing the number of the logical address areas estimated to be stored in a fragmented manner,
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds a first threshold value, defragmentation is performed on the logical address area estimated to be stored in a fragmented state. A defragmentation processing step for starting processing;
A storage control program that causes a computer to execute
In the defragmentation processing step, the defragmentation processing is continued until the number of the logical address areas estimated to be fragmented and stored falls below a second threshold,
The first threshold value and the second threshold value are set such that data movement due to garbage collection (GC) inside the storage medium due to the fragmentation eliminating process is reduced .

In order to achieve the above object, a system according to the present invention comprises:
A storage medium,
An access unit that specifies a logical address area of a logical address space and accesses the storage medium;
Storage control means for reading data from the storage medium and writing data to the storage medium based on the logical address area designated by the access means;
An information processing system comprising:
The storage control means,
The data of each logical address area obtained by dividing the logical address space used when the access unit accesses the storage medium by a unit of a fixed size is a plurality of physical addresses in the physical address space used inside the storage medium. Fragmentation management means for estimating whether the area is fragmented and stored, and managing the number of the logical address areas estimated to be fragmented and stored;
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds the first threshold value, a defragmentation process for the logical address area estimated to be stored in a fragmented state is performed. A defragmentation processing unit that starts and continues the defragmentation processing until the number of the logical address areas estimated to be stored in a fragmented state falls below a second threshold value;
Equipped with a,
The defragmentation processing means continues the defragmentation process until the number of the logical address areas estimated to be stored in a fragmented state falls below a second threshold,
Wherein the first threshold and the second threshold value, the data move accompanying the interior of garbage collection (GC) of the storage medium by the defragmentation process Ru is set to be reduced.

  ADVANTAGE OF THE INVENTION According to this invention, the reduction effect of the data movement by GC in a storage medium obtained by defragmentation processing can be improved, and the access performance of a storage medium can be improved and long life can be achieved.

FIG. 1 is a block diagram illustrating a configuration of a storage control device according to a first embodiment of the present invention. FIG. 9 is a block diagram illustrating a configuration of an information processing system including a storage control device according to a second embodiment of the present invention. FIG. 9 is a block diagram illustrating a functional configuration of a storage control device according to a second embodiment of the present invention. It is a figure showing the composition of the update frequency management table concerning a 2nd embodiment of the present invention. It is a figure showing the composition of the fragmentation state management table concerning a 2nd embodiment of the present invention. It is a figure showing composition of a defragmentation object list concerning a 2nd embodiment of the present invention. It is a figure showing composition of a fragmentation presumption table and a fragmentation elimination processing management table concerning a 2nd embodiment of the present invention. It is a figure which shows the structure of the fragmentation estimation condition setting table and fragmentation elimination processing condition setting table which concern on 2nd Embodiment of this invention. FIG. 11 is a block diagram illustrating a hardware configuration of a host computer including a storage control unit according to a second embodiment of the present invention. 9 is a flowchart illustrating a processing procedure of a storage control unit according to a second embodiment of the present invention. It is a flow chart which shows the procedure of Read processing concerning a 2nd embodiment of the present invention. It is a flow chart which shows the procedure of the Write processing concerning a 2nd embodiment of the present invention. It is a flow chart which shows the procedure of fragmentation presumption processing concerning a 2nd embodiment of the present invention. It is a flowchart which shows the procedure of the fragmentation elimination process which concerns on 2nd Embodiment of this invention. FIG. 13 is a block diagram illustrating a functional configuration of a storage control device according to a third embodiment of the present invention. It is a figure showing composition of a fragmentation cancellation object list concerning a 3rd embodiment of the present invention. FIG. 14 is a block diagram illustrating a configuration of an information processing system including a storage control unit according to a fourth embodiment of the present invention. FIG. 14 is a block diagram illustrating a configuration of an information processing system including a storage control unit according to a fifth embodiment of the present invention.

  Hereinafter, embodiments of the present invention will be illustratively described in detail with reference to the drawings. However, the components described in the following embodiments are merely examples, and are not intended to limit the technical scope of the present invention thereto.

  The term “chunk” used in this specification indicates each logical address area obtained by dividing a logical address space in a unit for accessing a storage medium such as a host computer into units of a certain size. For example, when the storage medium is an SSD, each logical address area is divided into several tens of megabytes, and when the storage medium is an HDD employing SMR, each logical address area is divided into several hundreds of megabytes. In this specification, the ratio of the capacity available for storing data written from the host computer to the storage area collected by the GC is defined as “collection efficiency”.

[First Embodiment]
A storage control device 100 according to a first embodiment of the present invention will be described with reference to FIG. The storage control device 100 is a device that controls storage of data in a storage medium.

  As shown in FIG. 1, the storage control device 100 includes a fragmentation management unit 101 and a fragmentation resolution processing unit 102. The fragmentation management unit 101 converts the data of each logical address area obtained by dividing the logical address space used when the host computer accesses the storage medium 110 into units of a certain size into physical addresses used inside the storage medium 110. It is estimated whether or not the space is fragmented and stored in a plurality of physical address areas in the space, and the number of logical address areas estimated to be fragmented and stored is managed. When the number of logical address areas estimated to be stored in a fragmented manner exceeds the first threshold value, the fragmentation resolution processing unit 102 determines the logical address area estimated to be stored in a fragmented state. Of the fragmentation processing for.

  According to the present embodiment, the fragmentation state of each logical address area obtained by dividing the logical address space in units of a fixed size is estimated from the host computer, and the fragmentation eliminating process is performed based on the number of fragmented logical address areas. do. As a result, the effect of reducing data movement by GC in the storage medium obtained by the defragmentation process can be enhanced, and the access performance of the storage medium can be improved and the service life can be extended.

[Second embodiment]
Next, a storage control device according to a second embodiment of the present invention will be described. The storage control device according to the present embodiment stores data of each chunk, which is a logical address area obtained by dividing a logical address space used when a host computer accesses a storage medium in units of a fixed size, in a fragmented manner. The number of chunks estimated to be stored in a fragmented manner is managed. Then, when the number of chunks exceeds the defragmentation processing start condition, the defragmentation processing for the chunk estimated to be fragmented and stored is started, and the number of chunks falls below the defragmentation processing end condition. Until the defragmentation process is continued. In the present embodiment, a case will be described in which the fragmentation elimination processing is continued until the number of chunks falls below the fragmentation elimination processing end condition. However, when the number of chunks exceeds the fragmentation elimination processing start condition, fragmentation is performed. The configuration of starting the fragmentation elimination processing for the chunk estimated to be stored and stored can provide the effect of the present embodiment.

<< Technical outline of this embodiment >>
Hereinafter, a technical outline of the present embodiment will be described.

  The defragmentation process of the storage area in the storage medium is realized by restoring data in the storage element. That is, writing to the storage element in proportion to the amount of data subjected to the fragmentation eliminating processing occurs as a cost. The writing associated with the defragmentation process causes a decrease in performance due to the interruption of the process for the access request from the host computer. For example, when the storage medium is an SSD, there is a lifetime due to the limitation on the number of times of Erase processing of the NAND flash used as the storage element. That is, in the defragmentation process for the SSD, the life is consumed in addition to the performance degradation. In particular, since a NAND flash having a short life is used mainly for a low-cost SSD, a long life is required by reducing the amount of stored data.

  In the present embodiment, the problems of performance degradation of the storage medium and life consumption due to the execution of the defragmentation processing are solved. The defragmentation process has the effect of reducing the amount of data movement accompanying the GC inside the storage medium by increasing the collection efficiency of the free space in the GC, but is proportional to the capacity of the data subjected to the defragmentation process. Re-storage is performed. That is, it is necessary to compare the effect obtained by the defragmentation processing on the storage medium with the cost, and execute the defragmentation processing so that the effect exceeds the cost. Further, in order to enhance the effect of the fragmentation eliminating process, it is necessary to accurately estimate the effect and cost of the data to be subjected to the fragmentation eliminating process.

  In order to enhance the effect obtained by the defragmentation processing on a storage medium such as an HDD or the like that employs an SSD or SMR that stores data in a write-once format, the cost associated with the defragmentation processing is reduced based on the effect of the defragmentation processing on the storage medium. The logical address area to be subjected to the defragmentation processing is determined so that the effect of the subtraction is maximized. Therefore, based on the fragmentation state of the storage data in the storage medium and the state of access from the host computer to the storage medium, the effect of reducing the amount of data movement due to GC caused by the fragmentation elimination process is estimated.

  In the present embodiment, in order to manage storage data writing and fragmentation in the storage medium, a chunk is defined by dividing a logical address space provided by the storage medium to the host computer for each fixed size, Manage the status of fragmentation, update, and deletion of stored data for each chunk.

  First, upon detecting a write request or a delete request for data on a storage medium, the write request specifies a chunk to be stored and updated. The deletion request for the storage medium indicates that data represented by the range of the logical address, such as TRIM (for ATA: Advanced Technology Attachment protocol) or UNMAP (for SCSI: Small Computer System Interface protocol), is unnecessary. This is a command to notify the medium. Then, the information indicating the fragmentation status, the update status, and the deletion status for the specified chunk is updated. Next, by referring to the information indicating the state of fragmentation, update, and deletion of the stored data for each chunk, it is determined whether or not the fragmentation removal processing is necessary for each chunk in the storage medium. Calculate the number of chunks. At the same time, the priority of the fragmentation eliminating process is determined for the chunk determined to require the fragmentation eliminating process. Then, by comparing the number of chunks that need to be defragmented with the number of chunks that is a reference value for determining the start of the defragmentation process determined in advance, the number of chunks where fragmentation has occurred If the value exceeds the reference value, it is determined that the efficiency of free space collection by GC has been reduced as a result of fragmentation of data in the storage medium, and fragmentation elimination processing is started from chunks with higher priorities. After the defragmentation process is started, fragmentation occurs by comparing the number of chunks in a fragmented state with the number of chunks that is a reference value for judging the end of the defragmentation process determined in advance. When the number of performed chunks is smaller than the reference value for termination, the fragmentation elimination processing is terminated.

  Here, the reference value of the number of chunks in the fragmentation state for judging the start and end of the defragmentation processing is determined by the total write amount accompanying the re-storage of the chunks from the data movement reduction effect of the GC processing by the defragmentation processing. The setting is made for each storage medium so that the effect after consideration of the cost of the defragmentation process, which has been subtracted, is maximized. Further, when the present embodiment is used for the purpose of preventing performance degradation due to fragmentation in a storage medium, for example, the number of chunks determined based on an actual measured value of the performance with respect to the number of fragmented chunks of the storage medium is used. It may be set as a reference value.

  The chunk to be subjected to fragmentation processing is determined based on the priority calculated from information indicating the fragmentation, update, and deletion status of each chunk. For example, focusing on the frequency of update and deletion of each chunk, chunks with a high update frequency are likely to cause fragmentation again after performing the fragmentation resolution processing, and thus it can be said that the effect of the fragmentation resolution processing is low. Therefore, the effect of the defragmentation processing can be enhanced by preferentially selecting the chunks that are not updated or deleted frequently as targets of the defragmentation processing. In addition, when updating the data stored in each chunk, by taking into account the number of sectors of the updated data, it is possible to more accurately select a chunk having a high defragmentation processing effect. When the update frequency is constant, fragmentation progresses as the number of updated sectors increases, and it can be said that the effect of the fragmentation elimination process is high.

(Summary of configuration and operation of this embodiment)
In the storage control unit according to the present embodiment, data of chunks, which are logical address areas obtained by dividing a logical address space used when a host computer accesses a storage medium in units of a certain size, are used inside the storage medium. In the physical address space to be stored, it is estimated whether or not a plurality of physical address areas are fragmented and stored, and the number of chunks that are estimated to be fragmented and stored is managed. Then, when the number of chunks estimated to be stored in a fragmented manner exceeds the defragmentation processing start condition (first threshold), the fragmentation for the chunk estimated to be stored in a fragmented state is performed. The defragmentation process is started, and the defragmentation process is continued until the number of chunks estimated to be fragmented and stored falls below the defragmentation process end condition (second threshold).

  Here, the first threshold value and the second threshold value are set so as to reduce data movement due to garbage collection (GC) in the storage medium due to fragmentation eliminating processing. For example, the total number of logical address areas included in the storage medium, the number of logical address areas storing valid data included in the storage medium, and the logical address area corresponding to the amount of data written per unit time to the storage medium And the average update interval of the data of the logical address area included in the storage medium and determined to be stored in a fragmented manner. Further, the first threshold value and the second threshold value are set to the same number, and the defragmentation process is performed so that the number of chunks estimated to be fragmented and stored is the same.

  In addition, for the chunk estimated to be stored in a fragmented state, the priority order for performing the defragmentation process is further set, and the defragmentation process is performed according to the priority order. Here, the priorities are set in ascending order of update frequency for updating chunk data.

  In addition, it is estimated that chunk data is fragmented and stored based on the occurrence distribution of the write process or the delete process for the subdivision area obtained by subdividing each chunk. That is, the number of times that the first Write process or the deletion process to the re-divided area has occurred is accumulated as the number of fragmentation occurrences. If the number of fragmentation occurrences exceeds the fragmentation estimation condition (third threshold), the chunk data is fragmented. It is presumed that it is stored in the form of

  In the defragmentation process, data of a chunk that has been fragmented and stored is read out and re-stored in the same chunk. In that case, other writing is restricted during the fragmentation eliminating process. The storage medium to which this embodiment is applied is a storage medium that stores data in a storage element in a write-once format, and includes an SSD and an HDD that employs an SMR.

《Information processing system》
FIG. 2 is a block diagram illustrating a configuration of an information processing system 200 including a storage control unit 220 as a storage control device according to the present embodiment.

  Referring to FIG. 2, the information processing system 200 of the present embodiment includes one or more host computers 210, a storage medium 240, and a network 250 connecting the host computer 210 and the storage medium 240. Note that the storage medium 240 may be distributed to a plurality of storage media and connected to a network.

  The host computer 210 includes an OS (Operating System) 211, an application 212 that is software operating on the OS 211, and the storage control unit 220 of the present embodiment. The storage control unit 220 controls access processing including read and write to the storage medium 240 in processing of the OS 211 and the application 212.

  The storage medium 240 is a write-once storage medium such as an SSD using a NAND flash as a storage element or an HDD using an SMR. Note that the storage medium 240 may include a medium main body 241 and a cache memory 242. The form of the network 250 is not particularly limited, and may be a form in which the host computer 210 and the storage medium 240 are directly connected. Further, the host computer 210 may be constituted by a plurality of devices or a system connected by the network 250.

<< Functional configuration of storage controller >>
FIG. 3 is a block diagram illustrating a functional configuration of the storage control unit 220 as the storage control device according to the present embodiment.

  Referring to FIG. 3, the storage control unit 220 includes an access execution unit 310, a fragmentation management unit 320, and a fragmentation resolution processing unit 330.

《Access execution section》
The access execution unit 310 receives an access request from the OS 211 or the application 212 running on the host computer 210, and performs necessary access processing on the storage medium 240. In the case of a Write request, a Write process is executed on the storage medium 240 via the Write queue 311. In addition, the access execution unit 310 reads data from the area on the storage medium 240 specified by the fragmentation resolution processing unit 330 and performs a re-storage process. Reading of data from the area of the storage medium 240 can be performed by making a request to a cache memory for the storage medium 240 (not shown), instead of making a request directly to the storage medium 240.

《Fragmentation management》
The fragmentation management unit 320 monitors update by a write request or a deletion request to the storage medium 240 by the access execution unit 310, estimates fragmentation of each chunk, and generates a list of chunks estimated to be fragmented. And manage.

  3, the fragmentation management unit 320 includes an update monitoring unit 321, an update frequency management table 322, a fragmentation state management table 323, a fragmentation removal target estimation unit 324, and a fragmentation removal target list 325. Prepare.

(Update monitoring section)
The update monitoring unit 321 monitors software operating on the host computer 210, a Write request and a deletion request from the OS to the storage medium 240, and determines the access destination logical address and the access destination chunk in the storage medium 240 based on the access size. To identify. Then, the update frequency information stored in the entry of the update frequency management table 322 corresponding to the specified chunk is updated. In addition, for the entry corresponding to the chunk in the fragmentation state management table 323, information indicating whether or not fragmentation has occurred in the chunk and the fragmentation state are updated by the Write request received by the access execution unit 310.

(Update frequency management table)
The update frequency management table 322 has an entry for each chunk in the storage medium, and performs an update frequency management function of managing information on the update frequency represented by the frequency of updating or deleting data in the chunk. . The update frequency management table 322 has entries for storing information on the update frequency of each chunk in the logical address space of the storage medium 240 at regular time intervals, and the update frequency information of each entry is reset at regular time intervals.

  FIG. 4A is a diagram illustrating a configuration of the update frequency management table 322 according to the present embodiment. Each entry of the update frequency management table 322 stores an update frequency 412 indicating the frequency of update or deletion within a certain time in association with the chunk ID 411.

  The format of the data stored in each entry of the update frequency management table 322 is not limited to the format shown in FIG. 4A. For example, the data may have a last update time instead of the update frequency 412, and may have a slightly different update frequency for each chunk. May be determined. In this case, it is not necessary to reset the update frequency every fixed time. Further, a format may be used which has information other than the update frequency 412 and the last update time, and manages the update history of each chunk in more detail than the update frequency.

(Fragmentation state management table)
The fragmentation state management table 323 has an entry for each chunk in the storage medium, and after the continuous storage of the entire chunk or the execution of the defragmentation processing, the fragmentation associated with the update or deletion of the data in the chunk is performed. It performs a fragmentation state management function of managing information indicating whether or not an error has occurred. The fragmentation state management table 323 has entries for storing information on the number of occurrences of fragmentation for the data of each chunk in the logical address space of the storage medium 240 and information on the distribution of addresses where fragmentation has occurred. In addition, each entry includes a flag indicating a change in the fragmentation state, which indicates whether the fragmentation removal target estimation unit 324 determines the fragmentation removal target for the corresponding chunk, and indicates whether the priority of the fragmentation removal target needs to be prioritized. Have.

  FIG. 4B is a diagram showing a configuration of the fragmentation state management table 323 according to the present embodiment. Each entry of the fragmentation state management table 323 is associated with the chunk ID 421, and indicates the number of fragmentation occurrences 422 indicating the number of fragmentation occurrences in the chunk, and whether or not fragmentation has occurred for each area obtained by further subdividing the chunk. And a state change flag 424 indicating a change in the fragmentation state.

  The update monitoring unit 321 specifies an area obtained by subdividing a chunk corresponding to an update / deletion destination address for a write or deletion target chunk, determines that the chunk is an area in which fragmentation has occurred, and determines in the fragmentation distribution information 423 Set if the corresponding bit is not set. Also, the number of fragmentation occurrences 422 is updated by the number of bits newly set in the fragmentation distribution information 423. In addition, the update monitoring unit 321 determines whether the fragmentation occurrence number 422 and the fragmentation distribution information 423 have been changed for the entry corresponding to the update / deletion destination chunk, and the status change flag 424 has not been set. , The status change flag 424 is set.

  Note that the format of the data stored in each entry of the fragmentation state management table 323 is not limited to the format shown in FIG. 4B. For example, the fragmentation distribution information 423 indicates the fragmentation for each divided area obtained by subdividing the chunk. In addition to the format in which the presence / absence is represented by a bitmap, a format in which information that can identify the fragmented region may be listed for the subdivided divided regions may be used. Also, for example, the number of fragmentation occurrences 422 is omitted and the number of bits set in the fragmentation distribution information 423 is not used to indicate the degree of fragmentation in a chunk by the number of fragmentation occurrences 422 and the fragmentation distribution information 423. A format may be used instead of the number of fragmentations 422. Alternatively, a format may be used which simply indicates only the presence / absence of fragmentation based on the presence / absence of Write access to the chunk. Further, although the fragmentation state management table 323 is shown independently of the update frequency management table 322 in FIG. 4B, it may be configured as a single integrated table.

(Fragmentation removal target list)
The defragmentation target list 325 stores the chunks in the logical address space of the storage medium 240 that have been determined as defragmentation targets by the defragmentation target estimation unit 324 in order of the priority of the defragmentation processing. Further, the fragmentation resolution target list 325 manages the number of chunks to be fragmentation resolution stored by the fragmentation resolution target estimation unit 324.

  FIG. 4C is a diagram illustrating a configuration of the fragmentation cancellation target list 325 according to the present embodiment. The defragmentation target list 325 includes a defragmentation target number 431 indicating the number of chunks to be defragmented, a chunk ID 432 uniquely indicating each chunk of the storage medium 240 for each chunk to be subjected to defragmentation processing. , And a priority 433 indicating the priority of the defragmentation processing of each chunk.

  Note that the format of the data of the fragmentation removal target list 325 is not limited to the format shown in FIG. 4C. For example, by creating a list in which the chunk IDs 432 are arranged according to the priority of the fragmentation removal processing, the priority 433 is changed. Can be omitted.

(Fragmentation resolution target estimation unit)
The fragmentation elimination target estimating unit 324 refers to the update or deletion history and the fragmentation state information of each chunk managed by the update frequency management table 322 and the fragmentation state management table 323, and refers to the logical address of the storage medium 240. It performs a defragmentation target estimation function of estimating whether or not the defragmentation process is necessary for each chunk in the space. In addition, the priority order of the fragmentation eliminating process is determined for the chunk estimated to require the fragmentation eliminating process, and the number of chunks to be subjected to the fragmentation eliminating process is calculated.

  First, the fragmentation elimination target estimating unit 324 refers to the fragmentation occurrence number 422 stored in the fragmentation state management table 323 for each chunk, and performs a fragmentation elimination process on chunks exceeding a predetermined reference value. Is determined to be necessary. Next, when it is determined that defragmentation processing is necessary for the chunk, the number of occurrences of fragmentation 422 and the update frequency 412 and the last update time stored in the corresponding entry of the update frequency management table 322 are referred to for the chunk. Thus, the priority of the fragmentation eliminating process is determined. Then, the defragmentation target estimation unit 324 stores the number representing the chunk and the determined priority in the defragmentation target list 325 as the chunk ID 432 and the priority 433, respectively, and sets the value of the number of defragmentation targets 431. Is added.

  Note that the criterion for determining the necessity of defragmentation processing for each chunk in the defragmentation target estimation unit 324 is not necessarily limited to comparison with a reference value. For example, referring to the update frequency 412 and the last update time of each chunk before determining the priority, chunks whose update frequency is higher than a certain level may cause fragmentation again even if fragmentation is eliminated. Since the effect of defragmentation is high and low, it may be excluded from chunks requiring fragmentation.

  Further, the reference value to be compared with the fragmentation occurrence number 422 of each chunk can be limited to, for example, a chunk that can be a data movement target by the GC algorithm inside the storage medium 240. Specifically, when the GC algorithm used by the storage medium 240 selects the area to be collected so as to minimize data movement, the ratio of the surplus capacity of the storage element to the size of the logical address space of the storage medium 240 is P. Then, an area in which the ratio of the empty area is equal to or more than P is always selected as a collection target. Therefore, by referring to the number of fragmentation occurrences 422, the ratio of the amount of fragmented data in the subdivided area of each chunk is calculated, and when the calculation result is less than P, it is excluded from the fragmentation removal target. By setting the reference value to, chunks that cannot be targeted for data movement by GC can be excluded from the fragmentation resolution targets.

《Defragmentation processing section》
In FIG. 3, the defragmentation processing unit 330 includes a defragmentation target list 325, a defragmentation condition setting unit 331, a defragmentation processing control unit 332, and a re-storage control unit 333. Note that the fragmentation removal target list 325 is shared with the fragmentation management unit 320, and receives the fragmentation removal target chunk information.

(Fragmentation resolution condition setting section)
The defragmentation condition setting unit 331 determines the number of chunks in the fragmented state included in the storage medium 240 that are necessary for the defragmentation processing control unit 332 to determine the start and end of the defragmentation process for the storage medium 240. It fulfills the function of setting fragmentation elimination conditions for managing and setting reference values. The defragmentation condition setting unit 331 calculates the number of chunks of the reference value by subtracting the total write amount accompanying the re-storage of chunks in the defragmentation process from the reduction amount of data movement due to GC in the storage medium 240. The calculation is performed so that the estimated result of the effect after the cost regarding the resolution processing is considered is maximized. Alternatively, a calculation result of the number of chunks that satisfies the performance condition of the storage medium specified by the user is set.

  The calculation of the number of chunks in which fragmentation has occurred, which is a criterion for determining whether to start or stop execution of fragmentation resolution processing, is based on the efficiency of collecting free space when GC is performed on each chunk on the storage medium. decide. The collection efficiency of the free space in the GC indicates the ratio between the amount of write data from the host computer that can be newly stored by the GC and the amount of data movement inside the storage medium. The data to be moved along with the GC in the storage medium is data stored in a fragmented state by updating and deleting. On the other hand, chunk data that has not been updated or deleted after the defragmentation processing is performed is continuously stored without fragmentation, and thus is not included in data to be moved by GC. . That is, the total amount of data that can be moved in the GC is represented by the number of chunks in the storage medium in a state where fragmentation has occurred.

  In addition, the effect of reducing the data movement amount of the GC by the fragmentation eliminating process can be calculated from the efficiency of collecting the free space by the GC. First, the occurrence frequency of GC with respect to the amount of write data from the host computer is calculated using the collection efficiency of the free area. By using the calculated occurrence frequency and information on the period from the new storage of data to the update or deletion, it is possible to calculate the amount of data movement accompanying GC. More specifically, using the information on the update frequency 412 stored in the update frequency management table 322 for each chunk, an average period from when the data of each chunk is stored to when it is updated or deleted is obtained. By dividing the average period by the number of chunks in the fragmented state and the frequency of occurrence of GC processing, that is, the average occurrence interval of GC processing, the data of each chunk is associated with GC in the form of an expected value of the number of times of movement in the GC processing. The amount of data movement is required.

  Furthermore, by calculating the difference in the amount of data movement due to GC before and after the defragmentation processing, the amount of reduction in the amount of data movement in the defragmentation processing can be calculated. In addition, the cost required for the defragmentation processing is storage of the chunk size data in the storage medium 240. By taking into account the cost, the value obtained by subtracting the total amount of writing due to the re-storage of chunks in the defragmentation processing from the amount of reduction in data movement due to the GC inside the storage medium due to the defragmentation processing is calculated. The effect after consideration is required. The number of chunks in a fragmented state, which is a criterion for starting and stopping execution, may be set so that this effect is maximized.

(Fragmentation resolution processing control unit)
In a state where the defragmentation processing has been stopped, the fragmentation resolution processing control unit 332 acquires information on the number of chunks serving as a start condition of the fragmentation resolution processing from the fragmentation resolution condition setting unit 331, and performs the fragmentation resolution processing. If the required number of chunks exceeds the number of chunks in the start condition, the fragmentation elimination process is started in order from the chunk with the highest priority. On the other hand, when the defragmentation process has been started, information on the number of chunks which is the end condition of the defragmentation process is obtained from the defragmentation condition setting unit 331, and the number of chunks requiring the defragmentation process is determined as the stop condition. If the number of chunks is smaller than the number of chunks, the function of controlling fragmentation resolution processing to end the fragmentation resolution processing is performed.

  In a state where the defragmentation processing is stopped, the defragmentation processing control unit 332 refers to the number of defragmentation targets 431 of the defragmentation target list 325 and sets the defragmentation set by the defragmentation condition setting unit 331. Compare the number of chunks that are the processing start conditions. As a result of the comparison, if the calculated number of chunks exceeds the number of chunks that are the conditions for starting the fragmentation resolution processing, the fragmentation resolution processing is set to a start state, and the priority is set higher with reference to the chunk ID 432 and the priority 433. The information of the chunks is sequentially notified to the re-storage control unit 333 from the chunks to be subjected to the defragmentation processing. On the other hand, when the defragmentation process is started, the number of defragmentation targets 431 in the defragmentation target list 325 is referred to, and the defragmentation process stop condition set by the defragmentation condition setting unit 331 is set. Compare the number of chunks. As a result of the comparison, when the calculated number of chunks is smaller than the number of chunks serving as the condition for stopping the fragmentation eliminating process, the fragmentation eliminating process is stopped and the fragmentation eliminating process is terminated.

(Restorage control unit)
The re-storage control unit 333 re-stores a read request for reading the data of the chunk targeted for fragmentation resolution processing from the storage medium, and the chunk data read by the read request in the same chunk of the storage medium. And performs a re-storage control function of executing fragmentation elimination processing.

  The re-storage control unit 333 instructs the access execution unit 310 to make a Read request for reading the corresponding data from the storage medium 240 for the chunk to be defragmented, which is notified from the defragmentation processing control unit 332. Instruct. Also, by instructing the access execution unit 310 to issue a Write request for restoring the data read by the Read request to the storage medium 240, the chunk defragmentation processing is performed. Reading of data from the storage medium 240 can be substituted by making a request to the cache memory 242 for the storage medium 240 other than making a direct read request to the storage medium 240.

(Fragmentation estimation table and fragmentation resolution processing management table)
FIG. 4D is a diagram showing a configuration of the fragmentation estimation table 440 and the fragmentation resolution processing management table 450 according to the present embodiment. The fragmentation estimation condition (third threshold value), the fragmentation resolution process start condition (first threshold value), and the fragmentation resolution process end condition (second threshold value) in the fragmentation estimation table 440 and the fragmentation resolution process management table 450 are as follows. , May be a fixed value initially set, or may be an updated value calculated as needed according to an environmental change.

  The fragmentation estimation table 440 is used by the defragmentation target estimating unit 324 to estimate the chunk to be defragmented based on the fragmentation estimation condition (third threshold). In addition, the fragmentation estimation table 440 is used to set the priority in which the fragmentation removal target estimating unit 324 performs the fragmentation removal processing on the chunk estimated to be fragmented.

  The fragmentation estimation table 440 stores parameters for fragmentation estimation, such as the chunk capacity 442 and the number of fragmentation occurrences 443, in association with the chunk ID 441. Then, a fragmentation presence / absence flag 445 estimated based on the fragmentation estimation condition (third threshold) 444 is stored. In the present embodiment, the fragmentation estimation condition (third threshold) 444 is a reference value of the number of fragmentation occurrences, and the chunk fragment is determined by the magnitude of the fragmentation occurrence number 443 and the fragmentation estimation condition (third threshold) 444. Is estimated.

  In the present embodiment, the fragmentation estimation table 440 stores the fragmentation resolution based on the comparison of the update frequency 446 of the chunk estimated to be fragmented with the update frequency of the chunk estimated to be other fragmentation. The processing priority 447 is set.

  The defragmentation process management table 450 includes a defragmentation process start condition (first threshold) and a defragmentation process end condition (second threshold) from the defragmentation condition setting unit 331. Is used to control the start or end of the defragmentation process based on the

  The fragmentation resolution processing management table 450 corresponds to the comparison between the number of fragmentation chunks 451 from the number of fragmentation resolution targets 431 in the fragmentation resolution target list 325 and the fragmentation resolution processing start condition (first threshold value) 452. , The fragmentation elimination processing start flag 453 is set. Further, the fragmentation resolution processing management table 450 sets a fragmentation resolution processing end flag 455 corresponding to the comparison between the number of fragmentation chunks 451 and the fragmentation resolution processing end condition (second threshold value) 454.

(Fragmentation estimation condition setting table and fragmentation elimination processing condition setting table)
FIG. 4E is a diagram showing a configuration of the fragmentation estimation condition setting table 460 and the fragmentation elimination processing condition setting table 470 according to the present embodiment.

  The fragmentation estimation condition setting table 460 is used by the fragmentation elimination target estimation unit 324 to set a fragmentation estimation condition (third threshold). The fragmentation estimation condition setting table 460 stores a fragmentation estimation condition (third threshold) 463 generated according to the fragmentation estimation condition setting algorithm 462 based on the fragmentation estimation condition setting parameter 461. Although the fragmentation estimation condition setting parameter 461 indicates the chunk capacity, the number of chunk divisions, the division area capacity, and the like, the present invention is not limited thereto.

  The defragmentation processing condition setting table 470 is used by the defragmentation condition setting unit 331 to set a defragmentation processing start condition (first threshold) and a defragmentation processing end condition (second threshold). . The defragmentation processing condition setting table 470 includes a defragmentation processing start condition (first threshold) 473 and a defragmentation processing generated according to the defragmentation condition setting algorithm 472 based on the defragmentation condition setting parameter 471. An end condition (second threshold value) 474 is stored. As the fragmentation elimination condition setting parameter 471, the total number of chunks (C) of the storage medium, the number of effective data storage chunks (A), the number of fragmentation state chunks (B), the amount of write data per unit time (W), the chunks Are shown, but the present invention is not limited to this.

《Hardware configuration of host computer》
FIG. 5 is a block diagram illustrating a hardware configuration of a host computer including the storage control unit 220 according to the present embodiment. In FIG. 5, the storage control unit 220 is illustrated as software processing of the host computer 210, but the storage control unit 220 may be implemented by a one-chip computer independent of the host computer 210.

  In FIG. 5, a CPU (Central Processing Unit) 510 is a processor for arithmetic control, and implements the functional components of the storage control unit 220 in FIG. 3 by executing a program. A ROM (Read Only Memory) 520 stores fixed data such as initial data and programs and programs. The communication control unit 530 communicates with the storage medium 240 and other devices via the network 250. Note that the number of CPUs 510 is not limited to one, and may include a plurality of CPUs or a GPU (Graphics Processing Unit) for image processing. Further, it is desirable that the communication control unit 530 has a CPU independent of the CPU 510 and writes or reads transmission / reception data in an area of a RAM (Random Access Memory) 540. It is desirable to provide a DMAC (Direct Memory Access Controller) for transferring data between the RAM 540 and the storage 550 (not shown). Further, it is desirable that the input / output interface 560 has a CPU independent of the CPU 510, and writes or reads input / output data in / from the area of the RAM 540. Therefore, CPU 510 recognizes that the data has been received or transferred to RAM 540 and processes the data. Also, the CPU 510 prepares the processing result in the RAM 540, and leaves the subsequent transmission or transfer to the communication control unit 530, the DMAC, or the input / output interface 560.

  RAM 540 is a random access memory used by CPU 510 as a work area for temporary storage. In the RAM 540, an area for storing data necessary for realizing the present embodiment is secured. The logical address 541 is an address at which the storage medium 240 is accessed from the OS 211 or the application 212. The fragmentation estimation condition 463 is a threshold value for estimating that the defragmentation target estimation unit 324 has fragmented each chunk. The fragmentation estimation table 440 shown in FIG. 4D is a table used by the defragmentation target estimation unit 324 to estimate that each chunk has been fragmented. The defragmentation processing condition 473/474 is a threshold value for the defragmentation processing control unit 332 to start or end the defragmentation processing. The defragmentation process management table 450 is a table used by the defragmentation process control unit 332 shown in FIG. 4D for starting or ending the fragmentation process. The read information 545 is information read from the storage medium 240 based on a request from the OS 211 or the application 212. The read information 546 is information to be written to the storage medium 240 based on a request from the OS 211 or the application 212. The input / output data 547 is data input / output via the input / output interface 560. The transmission / reception data 548 is data transmitted / received via the communication control unit 530. The application use area 549 is an area used by an application in processing other than storage control.

  The storage 550 stores a database, various parameters, or the following data or programs necessary for realizing the present embodiment. The update frequency management table 322 is a table for managing the update frequency of each chunk shown in FIG. 4A. The fragmentation state management table 323 is a table for managing the fragmentation state of each chunk shown in FIG. 4B. The defragmentation target list 325 is a list shown in FIG. 4C that stores the number of chunks to be defragmented and the priority of the defragmentation processing. The fragmentation estimation condition setting algorithm 462 is an algorithm for generating the fragmentation estimation condition shown in FIG. 4E. The fragmentation elimination condition setting algorithm 472 is an algorithm for generating the start condition or the end condition of the fragmentation elimination process shown in FIG. 4E.

  The storage 550 stores the following programs. The OS 211 is a basic program that controls the entire host computer 210. The application 212 is a program currently being executed by the host computer 210. The storage control program 552 is a program that receives access from the OS 211 or the application 212 to the storage medium 240 and implements access control according to the present embodiment. The read control module 553 is a module that controls reading from the storage medium 240 in the storage control program 552. The write control module 554 is a module that controls writing to a storage medium in the storage control program 552. The fragmentation management module 555 is a module for monitoring the fragmentation state of each chunk and managing the number of chunks estimated as the fragmentation state. The defragmentation processing module 556 is a module that controls the defragmentation processing based on the number of chunks estimated to be in a fragmented state, and performs rearrangement in each chunk to eliminate the fragmented state.

  The input / output interface 560 interfaces input / output data with input / output devices. The display unit 561 and the operation unit 562 are connected to the input / output interface 560. Further, the storage medium 240 may be connected to the input / output interface 560. Furthermore, a speaker as an audio output unit, a microphone as an audio input unit, or a GPS position determination unit may be connected.

  Note that programs and data related to general-purpose functions and other realizable functions of the host computer 210 are not illustrated in the RAM 540 and the storage 550 of FIG.

<< Processing procedure of storage control unit >>
FIG. 6 is a flowchart illustrating a processing procedure of the storage control unit 220 according to the present embodiment. This flowchart is executed by the CPU 510 of FIG. 5 using the RAM 540, and implements the functional components of FIG.

  In step S601, the storage control unit 220 of the host computer 210 performs a threshold setting process for setting a threshold value for estimating fragmentation of each chunk and a threshold value for starting or ending fragmentation resolution processing based on the number of chunks. Execute. The threshold may be a fixed threshold calculated in advance or a threshold dynamically calculated based on the current state of the host computer 210 or the storage medium 240.

  In step S611, the storage control unit 220 determines whether the request is a Read request from the OS 211 or the application 212. If it is a Read request, the storage control unit 220 executes a read process from the storage medium 240 in step S613. If it is not a Read request, the storage control unit 220 determines in step S621 whether it is a Write request from the OS 211 or the application 212. If it is a write request, the storage control unit 220 executes a write process to the storage medium 240 in step S623.

  If neither a read request nor a write request, the storage control unit 220 determines in step S631 whether or not it is the timing of fragmentation management. If it is the fragmentation management timing, the storage control unit 220 executes fragmentation management processing in step S633. If it is neither a Read request nor a Write request, nor is the fragmentation management timing, the storage control unit 220 determines in step S641 whether it is a fragmentation elimination timing. If the timing is for fragmentation elimination, the storage control unit 220 executes fragmentation elimination processing in step S643.

  In FIG. 6, the processing is executed in sequence, but each processing may be executed independently in parallel. Further, the fragmentation management timing and the fragmentation elimination timing may be arbitrary timings or timings related to other processes. For example, the fragmentation management timing may be when the Write process in step S623 is completed (performed with reference to the Write queue 311). Further, the fragmentation elimination timing may be a timing subsequent to the fragmentation management processing in step S633 (performed with reference to the write queue 311).

(Threshold setting processing)
In the threshold setting process in step S601, a fragmentation elimination process start condition (first threshold), a fragmentation elimination process end condition (second threshold), and a fragmentation estimation condition (third threshold) are set. The defragmentation processing start condition (first threshold) and the defragmentation processing end condition (second threshold) may be the same threshold.

  The defragmentation processing start condition (first threshold) is a threshold (upper reference value of the number of chunks estimated to be fragmented) for starting the defragmentation processing in the defragmentation processing (S643). The fragmentation elimination process end condition (second threshold) is a threshold value (lower reference value for the number of chunks estimated to be fragmented) for ending the fragmentation elimination process. Further, the fragmentation estimation condition (third threshold) is a threshold for estimating whether or not each chunk is a target for defragmentation in the fragmentation management process (S633) (a reference for the number of fragmentation occurrences of each chunk). Value).

  In the threshold value setting process in step S601, a fixed threshold value calculated in advance may be set, or the threshold value may be dynamically calculated and set based on the acquired condition parameters according to a calculation algorithm described later. Is also good.

(Read processing)
FIG. 7 is a flowchart illustrating a procedure of the read process (S613) according to the present embodiment. FIG. 7 shows a procedure in which, for example, a read request is received from the OS and the requested data is returned in step S613 of FIG.

  Upon receiving a Read request from software or an OS running on the host computer 210, the access execution unit 310 of the storage control unit 220 reads data from the storage medium 240 at the requested address, and returns the result to the access source. The process is returned to the software or the OS, and the process ends (step S701).

  By performing the processing in step S701 described above, the operation of receiving a Read request from the software or OS operating on the host computer 210 and returning the requested data in this embodiment is completed.

(Write processing)
FIG. 8 is a flowchart illustrating the procedure of the write process (S623) according to the present embodiment. FIG. 8 shows a procedure for performing processing of a write request or a deletion request issued from the software or the OS running on the host computer 210 in step S623 in FIG. 6, for example.

  Upon receiving a write request or a deletion request from the software or OS running on the host computer 210, the access execution unit 310 of the storage control unit 220 notifies the update monitoring unit 321 of the information of the write or deletion request (step S801). . The update monitoring unit 321 specifies a chunk in the logical address space to be accessed from the logical address and the size of the access destination of the storage medium 240 with respect to the access request received from the access execution unit 310 (step S803).

  The update monitoring unit 321 adds one to the value of the update frequency 412 stored in the entry of the update frequency management table 322 corresponding to the chunk specified in step S803 (step S805). The update monitoring unit 321 specifies the access destination in units of the area into which the chunk is subdivided from the logical address and the size of the access destination of the storage medium 240 regarding the access request received from the access execution unit 310. Then, in the fragmentation distribution information 423 stored in the entry of the fragmentation state management table 323 corresponding to the chunk specified in step S803, it is confirmed whether a bit corresponding to the access destination area is set (step S807). .

  As a result of the confirmation in step S807, if all the bits of the fragmentation distribution information 423 corresponding to the access destination area stored in the fragmentation state management table 323 corresponding to the chunk specified in step S803 have been set ( The processing in step S813 is performed (No determination in step S809). As a result of the confirmation in S807, the bits that are not set in the bits of the fragmentation distribution information 423 corresponding to the access destination area and stored in the entry of the fragmentation state management table 323 corresponding to the chunk specified in step S803. If it exists (Yes in step S809), the update monitoring unit 321 sets a bit corresponding to the access destination area of the fragmentation distribution information 423. Further, the number of bits newly set in the fragmentation distribution information 423 is added to the value of the fragmentation occurrence number 422 of the fragmentation state management table 323 stored in the corresponding entry. Further, the update monitoring unit 321 sets the status change flag 424 of the corresponding entry (Step S811).

  The access execution unit 310 issues the access request received in step S801 to the storage medium 240, and upon receiving a completion response from the storage medium 240, returns a completion response to the software operating on the host computer 210 that is the access request source and the OS. To end the process (step S813).

  The processing from steps S801 to S813 described above is performed, and the operation of processing the software running on the host computer 210 and the write request or deletion request issued from the OS is completed. The processing order of steps S801 to S813 does not necessarily need to be the order shown in FIG. For example, the processes of steps S801 to S811 and step S813 can be executed at the same time or executed in a different order. Also, the processing in step S805 and steps S807 to S811 can be executed at the same time or executed in a different order.

(Fragmentation estimation processing)
FIG. 9 is a flowchart illustrating the procedure of the fragmentation estimation process (S633) according to the present embodiment. In the fragmentation estimating process, the chunks that need to be defragmented are estimated, and the priorities of the defragmentation processes are set.

  The fragmentation elimination target estimating unit 324 of the storage control unit 220 pays attention to the first chunk in the logical address space of the storage medium 240 at an arbitrary timing, and starts determining whether fragmentation elimination processing is necessary (step S901). ). The fragmentation elimination target estimation unit 324 refers to the state change flag 424 of the entry of the fragmentation state management table 323 corresponding to the currently focused chunk (step S903).

  If the state change flag 424 corresponding to the focused chunk has not been set (No in step S905), the processing from step S919 is executed. When the state change flag 424 corresponding to the chunk of interest is set (Yes in step S905), the fragmentation elimination target estimating unit 324 sets the entry in the fragmentation status management table 323 corresponding to the chunk of interest at present. With reference to the fragmentation occurrence number 422, a reference value of a predetermined fixed number of fragments is compared (step S907). Note that the reference value of the number of fragments is set in advance by the method described above regarding the fragmentation elimination target estimation unit 324.

  If the fragmentation occurrence number 422 is smaller than the reference value as a result of the comparison in step S907 (No determination in step S909), the processing from step S919 is executed. If the number of occurrences of fragmentation 422 is equal to or greater than the reference value as a result of the comparison in step S907 (Yes in step S909), the fragmentation elimination target estimating unit 324 stores the entry in the update frequency management table 322 corresponding to the chunk of interest. The stored update frequency 412 is referred to (step S911).

  The fragmentation elimination target estimating unit 324 refers to the update frequency 412 stored in the update frequency management table 322 corresponding to each chunk represented by the chunk ID 432 stored in the fragmentation elimination target list 325, and Then, the chunks currently focused on are sorted in the order of the least updated frequency (step S913). Then, the defragmentation target estimation unit 324 adds the number corresponding to the currently focused chunk as a new chunk ID 432 in the defragmentation target list 325. Further, based on the sorting result in step S913, the priority order 433 corresponding to each chunk number registered in the fragmentation elimination target list 325 is updated (step S915).

  The fragmentation elimination target estimating unit 324 resets the state change flag 424 of the entry of the fragmentation state management table 323 corresponding to the currently focused chunk (step S917). The fragmentation elimination target estimating unit 324 checks whether the chunk currently focused on is the last chunk in the logical address space of the storage medium 240 (step S919).

  If the currently focused chunk is not the last chunk in the logical address space of the storage medium 240 (No in step S921), the fragmentation elimination target estimating unit 324 sets the chunk of interest in the logical address space of the storage medium 240. The chunk is changed to the chunk corresponding to the address immediately after, and the process is executed again from step S903 (step S923). If the currently focused chunk is the last chunk in the logical address space of the storage medium 240 (Yes in step S921), the chunk that needs to be defragmented is determined, and the priority of the defragmentation process is determined. The process ends.

  The above-described steps S901 to S923 are performed to determine the chunk that needs to be defragmented, and the operation of determining the priority of the defragmentation process is completed. The processing order from steps S901 to S923 does not necessarily have to be the procedure shown in FIG. For example, the state change flag 424 of the entry of the fragmentation state management table 323 is checked in steps S901 and S923, and if the chunk whose fragmentation state has changed can be directly specified as the chunk of interest, the determination of No in steps S903 and S905 is made. Processing can be omitted.

(Defragmentation processing)
FIG. 10 is a flowchart illustrating the procedure of the fragmentation eliminating process (S643) according to the present embodiment. In the defragmentation process, the defragmentation process for the chunk to be defragmented is started, executed, and stopped.

  The defragmentation process control unit 332 of the storage control unit 220 refers to the number of defragmentation targets 431 from the defragmentation target list 325 and obtains the number of chunks to be defragmented in the storage medium 240 at an arbitrary timing. (Step S1001). The defragmentation processing control unit 332 calculates the number of chunks to be defragmented and acquired in step S1001 and the number of chunks set as the defragmentation processing start conditions set by the defragmentation condition setting unit 331 (first threshold). Are compared (step S1003).

  When the number of chunks to be defragmented is smaller than the number of chunks as the condition for starting the defragmentation process (No in step S1005), the defragmentation process is terminated. If the fragmentation elimination processing is executed independently and in parallel, the processing from step S1001 is executed again. If the number of chunks to be defragmented is equal to or greater than the number of chunks in the start condition of the defragmentation processing (Yes in step S1005), the defragmentation processing control unit 332 sets the priority 433 from the defragmentation target list 325 to the highest. The high chunk ID 432 is extracted and notified to the re-storage control unit 333 (step S1007).

  The re-storage control unit 333 executes the Read request for reading the data of the chunk notified in step S1007 and the Write request for re-storage the data read by the Read request as the fragmentation resolution processing as the access execution unit 310. (Step S1009).

  When the re-storing process for the chunk notified in step S1007 is completed, the fragmentation resolution processing control unit 332 deletes the information of the corresponding chunk ID 432 and priority 433 from the fragmentation resolution target list 325, and the fragmentation resolution target number 431. Is subtracted by one (step S1011). The defragmentation processing control unit 332 includes the number of fragmentation occurrences 422 stored in the entry corresponding to the chunk for which the defragmentation processing has been completed in the fragmentation state management table 323, each bit of the fragmentation distribution information 423, and the state change flag. 424 are all reset to 0 "zero" (step S1013).

  The defragmentation processing control unit 332 calculates the value of the number of defragmentation targets 431 subtracted in step S1011 and the number of chunks (second threshold values) set by the defragmentation condition setting unit 331 as the conditions for stopping the defragmentation processing. Are compared (step S1015). When the fragmentation resolution target number 431 exceeds the number of chunks in the condition for stopping the fragmentation resolution processing (No in step S1017), the fragmentation resolution processing from step S1007 is executed again.

  When the fragmentation resolution target number 431 is equal to or smaller than the number of chunks in the fragmentation resolution process stop condition (Yes in step S1017), the fragmentation resolution process ends. If the fragmentation elimination processing is executed independently and in parallel, the processing from step S1001 is executed again.

  The processing from the above steps S1001 to S1017 is performed, and the operation of executing and stopping the fragmentation eliminating process for the chunk to be subjected to the fragmentation eliminating process is executed. The processing order from steps S1001 to S1017 does not necessarily need to be the order shown in FIG. For example, each processing of steps S1007 to S1009, step S1011 and step S1013 can be executed at the same time or executed in a different order. In addition, in the determination of step S1005, a Yes determination is made when the number of chunks to be defragmented exceeds the number of chunks as a condition for starting the defragmentation process. Is also good. Similarly, if the number of chunks to be defragmented is smaller than the number of chunks for stopping the defragmentation processing in the determination of step S1017, the determination is Yes, and if the number of chunks is equal to or larger than the number of chunks for the stop condition for fragmentation resolution processing, the determination is No. Is also good.

《Setting method of threshold value》
Hereinafter, a specific example of the threshold setting method will be described. However, the threshold value of the present embodiment is not limited to these setting methods.

(Estimation conditions for fragmentation of each chunk)
In step S907 of FIG. 9, the fragmentation estimation condition (third threshold value) 444 to be compared with the fragmentation occurrence number 443 is determined by, for example, the following procedure.

  First, the ratio α of the used capacity of the storage medium to the total physical capacity of the storage elements inside the storage medium is calculated. When the data stored in the storage medium is stored in a state of being uniformly distributed on the storage elements, the ratio of valid data included in each area of the storage medium is represented by α. That is, when the GC operates efficiently in the storage medium, the ratio of valid data included in the collected storage area is always represented by α or less.

  Therefore, with respect to the divided areas included in the chunk for which the fragmentation is determined, the chunk data in which the ratio of the divided areas that have become fragmented by the update is less than (1-α) has a low degree of fragmentation, and is subject to collection by GC. Does not. Assuming that the number of divided regions included in a chunk is N, the fragmentation proceeds if the fragmentation estimation condition 444, which is the number of fragments for judging fragmentation of the chunk, is {(1−α) × N}. Only chunks can be determined to be in a fragmented state.

(Fragmentation resolution processing start condition / fragmentation resolution processing stop condition)
In steps S1005 and S1017 of FIG. 10, the number of chunks of the start condition and the stop condition of the defragmentation process to be compared with the number of defragmentation targets 431 is determined by the defragmentation condition setting unit 331 in the following procedure, for example. . In the following procedure, an example will be described in which the number of chunks in the start condition and the stop condition of the fragmentation elimination process is the same.

  The total number of chunks obtained by converting the total capacity of the physical storage elements included in the storage medium 240 into the number of chunks is defined as C. First, the fragmentation elimination condition setting unit 331 calculates the number A of chunks for storing valid data from the usage status of the logical address space of the storage medium 240. If the number of chunks in the chunk included in the storage medium 240 that has been fragmented due to the update is B, valid data is stored and the update is not performed, that is, collected in the GC inside the storage medium 240. The number of non-target chunks is determined by (AB). Therefore, the total capacity of storage elements to be collected by GC in the storage medium 240 is expressed as {C- (AB)} = CA + B when converted into the number of chunks. That is, the data included in the B chunks in the fragmented state is stored in the storage element corresponding to the total capacity of C−A + B chunks in a write-once format. Assuming that updating is performed equally for each chunk, if the collection efficiency in the GC processing of the storage medium 240 is defined as E, E = (CA + BB) / (CA + B) = (C− A) / (CA + B).

  Assuming that a value obtained by converting the amount of write data per unit time to the storage medium 240 of the host computer 210 into the number of chunks is W, collection processing for W / E chunks per unit time occurs in the GC inside the storage medium 240. I do. In the case where the storage area is equally collected by the GC for the physical elements of the storage medium 240, the time interval in which the data included in each fragmented chunk included in the storage area 2 is subjected to the movement processing accompanying the GC Let R be R = (CA + B) / (W / E) = (CA) / W.

  Subsequently, the fragmentation elimination condition setting unit 331 updates each fragmented chunk included in the storage medium 240 from the average update frequency of data of each chunk included in the storage medium 240, and then updates the chunk in the fragmented state. The average update interval T up to is calculated. The expected value of the number of times the data included in the chunk in the fragmented state is subjected to the movement processing accompanying the GC in the storage medium 240 until the data is updated next is T / R = T · W / (C -A). Also, when the total amount of data that is moved along with the GC before being updated for each chunk in the fragmented state is represented by the number of chunks, one data movement process by the GC results in 1-E chunks. Data moves. Therefore, when the total data movement amount M by GC until the data of each chunk is updated is represented by the number of chunks, M = (1−E) · (T / R) = B · T · W / ｛(C− (A + B). (CA)}.

  On the other hand, when the entire chunk is re-stored in the storage medium 240 in order to eliminate the fragmentation state for the chunk in the fragmentation state, a Write equal to the chunk size occurs in the storage medium 240. In other words, when M> 1 is satisfied with respect to the total data movement amount M by GC for the fragmented chunks before being updated, the amount of data stored in the storage medium 240 can be reduced by performing fragmentation elimination processing. . Therefore, the number of chunks C representing the total capacity of the storage elements in the storage medium 240 is known, and the number of chunks A for storing valid data in the storage medium 240 and the amount of write data per unit time from the host computer 210 are chunked. When the value W converted into the number and the average update interval T from the time when each chunk in the fragmentation state is updated to the time when the chunk is next updated are determined in advance, the fragmentation elimination condition setting unit 331 updates the fragmentation by updating. With respect to the number B of the chunks in the state, the number of chunks for the start and stop conditions of the fragmentation elimination process used in steps S1005 and S1017 of FIG. 10 can be determined based on the minimum integer value that satisfies M> 1.

  The fragmentation elimination condition setting unit 331 not only calculates the number of chunks of the start condition and the stop condition using A, W, and T using the measurement results in advance, but also requests the update monitoring unit 321 to write to the storage medium 240. It is possible to optimize the start condition and the stop condition of the defragmentation processing in accordance with the state of the write request from the host computer 210 to the storage medium 240 by dynamically calculating based on the monitoring result of the above. For example, the number A of chunks storing valid data and the W representing the amount of write data per unit time from the host computer 210 can be calculated by monitoring the write / erase request for the storage medium 240 by the update monitoring unit 321. It is. In addition, the average update interval T of each chunk can be dynamically calculated by acquiring the update frequency 412 of each chunk in the storage medium 240 from the update frequency management table 322.

  The calculated number of chunks of the fragmentation elimination condition can be used as a common threshold for the start condition and the end condition, and can be controlled to be the threshold chunk number B. The threshold value) and the end condition (second threshold value) can be set to different numbers of chunks as follows.

(1) For example, a value (BX) obtained by subtracting a predetermined number X (0 ≦ X ≦ B) from the number of chunks B for start determination is set as the number of chunks for end determination. For example, if X = 10, fragmentation elimination processing will be performed on 10 chunks from the start to the end.
(2) The number (B × p), which is a value obtained by multiplying the predetermined number p (0 ≦ p ≦ 1) of the number B of chunks for start determination, is set as the number of chunks for end determination.

  The setting of the end condition is not limited to the above two types of methods, but can be set by another method. For example, in the above example, the calculated number of chunks of the fragmentation elimination condition is directly used as the threshold value of the start condition. However, the start condition is set to (B + y) and the end condition is set to (B−y) so as to be controlled by the calculated number of chunks. z), y + z = 10, and the like.

  In the present embodiment, the defragmentation processing start condition and / or the defragmentation processing stop condition and the fragmentation estimation condition are calculated according to the condition parameters. However, a history of combinations of past condition parameters and conditions (also referred to as thresholds and reference values) including the number of chunks and chunk capacities according to the OS and applications of the host computer and the effects of the condition setting are stored in association with each other. In advance, appropriate condition settings may be set with reference to new condition parameters.

  According to the present embodiment, the fragmentation state of each logical address area obtained by dividing the logical address space in units of a fixed size is estimated from the host computer, and the fragmentation eliminating process is performed based on the number of fragmented logical address areas. Control the start and end of the This makes it possible to estimate the storage area to be defragmented so as to minimize the frequency of defragmentation processing on the write-once storage area, which is necessary for maintaining a constant access performance in the storage medium. Therefore, in consideration of the cost of data re-storage accompanying the fragmentation eliminating process, the effect of reducing data movement by GC in the write-once storage medium obtained by the fragmentation eliminating process can be enhanced, and the access of the storage medium can be improved. It is possible to improve the performance and extend the service life.

[Third embodiment]
Next, a storage control device according to a third embodiment of the present invention will be described. The storage control device according to the present embodiment is different from the second embodiment in that the chunks to be defragmented have no priority. Other configurations and operations are the same as those of the second embodiment, and thus the same configurations and operations are denoted by the same reference numerals and detailed description thereof will be omitted.

<< Functional configuration of storage controller >>
FIG. 11 is a block diagram illustrating a functional configuration of the storage control device 1100 according to the present embodiment. In FIG. 11, the same components as those in FIG. 3 are denoted by the same reference numerals, and duplicate description will be omitted.

  The fragmentation management unit 1120 does not have the update frequency management table 322. In addition, the fragmentation elimination target estimating unit 1124 generates a fragmentation elimination target list 1125 in which the priorities based on the update frequency are not set for the chunks to be defragmented.

  In addition, the fragmentation resolution processing control unit 1132 of the fragmentation resolution processing unit 1130 selects an arbitrary chunk from the fragmentation resolution target list 1125 with no priority set according to the condition of the fragmentation resolution condition setting unit 1131. , And instructs the re-storage control unit 333 to perform the rearrangement.

  Note that the defragmentation condition setting unit 1131 may set only the start condition of the defragmentation process, and the defragmentation process control unit 1132 may determine only the start of the defragmentation process.

(Fragmentation removal target list)
FIG. 12 is a diagram showing the configuration of the fragmentation removal target list 1125 according to the present embodiment. In FIG. 12, the same components as those in FIG. 4C are denoted by the same reference numerals, and redundant description will be omitted.

  The defragmentation target list 1125 lists the chunk IDs 432 of the chunks estimated to be defragmentation targets, and does not list the priority of the defragmentation processing for each chunk.

  In the present embodiment, the priority order is not set based on the update frequency 412 in FIG. 4A. Therefore, in the fragmentation management processing flowchart shown in FIG. 9, steps S911 to S915 are skipped. In the flowchart of the fragmentation eliminating process shown in FIG. 10, in step S1007, an arbitrary fragmentation eliminating target chunk is notified to the re-storage control unit 333 without considering the priority.

  According to this embodiment, the configuration of the storage control device can be simplified, and the fragmented state of chunks can be eliminated by rearrangement with a simple configuration.

[Fourth embodiment]
Next, an information processing system including a storage control device according to a fourth embodiment of the present invention as a storage control unit in a host computer will be described. The information processing system according to the present embodiment is different from the second embodiment and the third embodiment in that the information processing system includes a buffer for substituting a Write process for an area where fragmentation is estimated to be remarkable. Other configurations and operations are the same as those of the second embodiment and the third embodiment. Therefore, the same configurations and operations are denoted by the same reference numerals, and detailed description thereof will be omitted.

《Information processing system》
FIG. 13 is a diagram illustrating a configuration of an information processing system 1300 including the storage control unit 1320 according to the present embodiment. Note that, in FIG. 13, the same components as those in FIG. 2 are denoted by the same reference numerals, and redundant description will be omitted.

  The information processing system 1300 of FIG. The buffer 1360 stores data corresponding to a Write request for a chunk with a high fragmentation frequency among Write data to the storage medium 240. As the buffer 1360, in addition to using a storage medium having a higher performance than the storage medium 240 such as a DRAM (Dynamic Random Access Memory), it is also possible to use an HDD or the like that employs an SSD or an SMR similarly to the storage medium 240. The buffer 1360 may have a form in which a partial area of the storage medium 240 is allocated.

  The host computer 1310 has a storage control unit 1320 that controls a write process to the storage medium 240 or the buffer 1360 in addition to the process of the second embodiment.

  According to the present embodiment, by storing Write data for an area where fragmentation occurs frequently in a storage area prepared as a buffer and avoiding storing in a chunk in which fragmentation is estimated to occur frequently in advance. Further, the access performance of the storage medium can be improved and the service life can be extended.

[Fifth Embodiment]
Next, an information processing system including a storage control device according to a fifth embodiment of the present invention as a storage control unit in a host computer will be described. The information processing system according to the present embodiment is different from the second to fourth embodiments in that the storage control unit is provided on the storage medium side or is provided outside the host computer via a network. . Other configurations and operations are the same as those of the second to fourth embodiments. Therefore, the same configurations and operations are denoted by the same reference numerals, and detailed description thereof will be omitted.

《Configuration of information processing system》
FIG. 14 is a diagram illustrating a configuration of an information processing system 1400 including the storage control unit 1443 or 1420 according to the present embodiment. In FIG. 14, the same components as those in FIG. 2 are denoted by the same reference numerals, and redundant description will be omitted.

  In FIG. 14, the host computer 1410 does not have the storage control unit of the present embodiment. As an example, the storage control unit 1443 of the present embodiment is mounted on a storage medium 1440. As another example, the storage control unit 1420 of the present embodiment is connected as an independent device via the network 250. Such a storage control unit 1443 or 1420 may be realized by software or a one-chip processor.

  According to this embodiment, the processing is not limited to the processing by the host computer as in the second to fourth embodiments, but the fragmentation state of the chunk in writing to the storage medium is estimated, and the fragmentation state is eliminated by relocation. By doing so, it is possible to improve the access performance of the storage medium and extend the life.

[Other embodiments]
As described above, the present invention has been described with reference to the exemplary embodiments. However, the present invention is not limited to the exemplary embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention. In addition, a system or an apparatus in which different features included in each embodiment are combined in any way is also included in the scope of the present invention.

  Further, the present invention may be applied to a system including a plurality of devices, or may be applied to a single device. Further, the present invention is also applicable to a case where an information processing program for realizing the functions of the embodiments is directly or remotely supplied to a system or an apparatus. Therefore, a program installed in the computer to realize the functions of the present invention by the computer, a medium storing the program, and a WWW (World Wide Web) server for downloading the program are also included in the scope of the present invention. . In particular, at least a non-transitory computer readable medium that stores a program that causes a computer to execute the processing steps included in the above-described embodiments is included in the scope of the present invention.

Claims (13)

The data of each logical address area obtained by dividing the logical address space used when the host computer accesses the storage medium in units of a fixed size is stored in a plurality of physical address areas in the physical address space used inside the storage medium. Fragmentation management means for estimating whether or not fragmented and stored, and managing the number of the logical address areas estimated to be stored as fragmented,
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds the first threshold value, a defragmentation process for the logical address area estimated to be stored in a fragmented state is performed. Defragmentation processing means to be started;
Equipped with a,
The defragmentation processing means continues the defragmentation process until the number of the logical address areas estimated to be stored in a fragmented state falls below a second threshold,
Wherein the first threshold and the second threshold value, set Ru storage controller so that data movement is reduced due to the interior of the garbage collection of the storage medium by the defragmentation process (GC). The first threshold and the second threshold are at least a total number of the logical address areas included in the storage medium, a number of the logical address areas in which valid data included in the storage medium are stored, and The number of the logical address areas corresponding to the write data amount per unit time to the storage medium, and the average update interval of the data of the logical address areas included in the storage medium and determined to be stored in a fragmented manner; The storage control device according to claim 1 , wherein the storage control device is calculated in consideration of: The first threshold and the second threshold are set to the same number,
3. The defragmentation processing unit according to claim 1, wherein the defragmentation processing unit performs the defragmentation processing such that the number of the logical address areas estimated to be stored in a fragmented state is the same number. Storage controller. The fragmentation management means further sets a priority for performing the defragmentation process for the logical address area estimated to be stored fragmented,
The defragmentation processing means, the priority in accordance performing the defragmentation process, the storage control device according to any one of claims 1 to 3. 5. The storage control device according to claim 4 , wherein the priorities are set in ascending order of update frequency for updating data in the logical address area. The fragmentation management unit estimates that the data of the logical address area is fragmented and stored based on the distribution of occurrence of Write processing or deletion processing on the subdivision area obtained by subdividing each of the logical address areas. to storage control device according to any one of claims 1 to 5. The fragmentation management means accumulates, as the number of fragmentation occurrences, the number of times the first write processing or deletion processing has occurred in the subdivision area, and, when the number of fragmentation occurrences exceeds a third threshold, the logical address area. The storage control device according to claim 6 , wherein it is estimated that the data is stored in a fragmented manner. The defragmentation processing means performs the defragmentation process by reading data of the logical address area stored in a fragmented manner and restoring the data in the same logical address area. The storage control device according to any one of claims 1 to 7 . 9. The storage control device according to claim 8 , wherein the fragmentation eliminating processing unit restricts other writing during the fragmentation eliminating process. The storage medium is a storage medium for data storage for the memory element in the write-once format, SSD (Solid State Drive) and and a HDD adopting SMR (Shingled Magnetic Recording), one of the claims 1 to 9 1 A storage control device according to the item. The data of each logical address area obtained by dividing the logical address space used when the host computer accesses the storage medium in units of a fixed size is stored in a plurality of physical address areas in the physical address space used inside the storage medium. Estimating whether or not it is stored in a fragmented manner, a fragmentation management step of managing the number of the logical address areas estimated to be stored in a fragmented manner,
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds a first threshold value, defragmentation is performed on the logical address area estimated to be stored in a fragmented state. A defragmentation processing step for starting processing;
Only including,
In the defragmentation processing step, the defragmentation processing is continued until the number of the logical address areas estimated to be fragmented and stored falls below a second threshold,
The storage control method, wherein the first threshold value and the second threshold value are set such that data movement accompanying garbage collection (GC) in the storage medium due to the fragmentation eliminating process is reduced . The data of each logical address area obtained by dividing the logical address space used when the host computer accesses the storage medium in units of a fixed size is stored in a plurality of physical address areas in the physical address space used inside the storage medium. Estimating whether or not it is stored in a fragmented manner, a fragmentation management step of managing the number of the logical address areas estimated to be stored in a fragmented manner,
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds a first threshold value, defragmentation is performed on the logical address area estimated to be stored in a fragmented state. A defragmentation processing step for starting processing;
A storage control program that causes a computer to execute
In the defragmentation processing step, the defragmentation processing is continued until the number of the logical address areas estimated to be fragmented and stored falls below a second threshold,
A storage control program wherein the first threshold value and the second threshold value are set such that data movement accompanying garbage collection (GC) inside the storage medium due to the fragmentation eliminating process is reduced . A storage medium,
An access unit that specifies a logical address area of a logical address space and accesses the storage medium;
Storage control means for reading data from the storage medium and writing data to the storage medium based on the logical address area designated by the access means;
An information processing system comprising:
The storage control means,
The data of each logical address area obtained by dividing the logical address space used when the access unit accesses the storage medium by a unit of a fixed size is a plurality of physical addresses in the physical address space used inside the storage medium. Fragmentation management means for estimating whether the area is fragmented and stored, and managing the number of the logical address areas estimated to be fragmented and stored;
When the number of the logical address areas estimated to be stored in a fragmented manner exceeds the first threshold value, a defragmentation process for the logical address area estimated to be stored in a fragmented state is performed. A defragmentation processing unit that starts and continues the defragmentation processing until the number of the logical address areas estimated to be stored in a fragmented state falls below a second threshold value;
Equipped with a,
The defragmentation processing means continues the defragmentation process until the number of the logical address areas estimated to be stored in a fragmented state falls below a second threshold,
Wherein the first threshold and the second threshold value, the information processing system that will be set as internal data movement accompanying the garbage collection (GC) is reduced in the storage medium by the defragmentation process.

Images