JP2012243117A - Storage medium control device, storage device, storage medium control method, and program - Google Patents

Abstract

As a storage medium control device for controlling a plurality of storage media by HSM, it is effective to increase the total capacity of management information for managing data without increasing the size of a data movement unit between storage media By restraining to a low level, efficient data movement control is realized.
A page set including a predetermined number of pages is defined, and a page set (target page set) that may include a page to be stored in a high-speed storage medium 111 is determined based on the access frequency of the page set. Then, out of the pages included in the target page set, a page that is currently stored in the low-speed storage medium 112 and accessed is selected and moved to the high-speed storage medium 111. As management information for this purpose, a page set management table 161 and a page management table 162 are provided.
[Selection] Figure 2

Description

  The present invention relates to a storage medium control device that controls a plurality of storage media having different access performances, and a storage device provided with the plurality of storage media in the storage medium control device. The present invention also relates to a storage medium control method corresponding to the storage medium control apparatus and a program for causing the storage medium control apparatus to realize a function corresponding to the storage medium control method.

  In recent years, the amount of data handled by an information processing system has increased remarkably with the evolution of information communication technology and the expansion of its application fields. Corresponding to such a background, the storage apparatus for storing data is in a situation where the storage capacity is increased and the access performance is increased.

  In order to increase the storage capacity of the storage device, for example, a method of increasing the number of storage media that store data in the storage device is known. This technique is one of the easiest techniques for increasing the storage capacity, and is used in many storage apparatuses. In such a storage apparatus, management is performed so that a plurality of internal storage media can be seen from the outside as a single large-capacity storage area.

  Further, with the increase in capacity of storage devices, it is required to improve immediacy of access, that is, access performance. As one of the easiest methods for this purpose, for example, improving the access performance of the storage medium itself provided in the storage apparatus is widely performed. There is also known a method of improving access performance by adopting striping that simultaneously accesses a plurality of storage media such as RAID 0 and RAID 5.

  However, it is also known that it is difficult to achieve both large capacity and high access performance for the storage medium provided in the storage apparatus due to cost and technical problems. Specifically, HDDs (Hard Disk Drives) provided as storage media in many storage apparatuses have a large storage capacity but a low access performance. The access performance of the HDD is considerably deteriorated particularly when continuous access (random access) to discontinuous data on the storage area is performed.

  Therefore, the use of storage media with higher access performance than HDDs is also being carried out. As one of such storage media, SSD (Solid State Drive) is known. Many SSDs use a semiconductor storage element such as a NAND flash memory and have higher access performance than HDDs. However, a storage medium such as this SSD generally has a smaller storage capacity and is more expensive than an HDD. Not only these HDDs and SSDs, generally, storage media with large capacity and low cost tend to have low access performance, and storage media with high access performance tend to have small capacity and high cost.

  As a countermeasure against the above problem, a technique called HSM (Hierarchical Storage Management) is known. Generally, among data handled by an information processing system, data with high access frequency or data requiring high access performance is a part of the whole. HSM pays attention to this point, and uses a storage medium that has high access performance but is expensive and a storage medium that has low access performance but is inexpensive and easy to increase its capacity, for example, to manage it as one storage area. It is.

  In a storage apparatus to which HSM is applied, data with high access frequency is stored in a storage area of a small capacity storage medium with high access performance, and the remaining data is stored in a storage area of a storage medium with low access performance. Take control. As a result, it is possible to maximize both the advantages of a large capacity and low cost of an inexpensive storage medium and the advantage of high access performance of an expensive storage medium.

  In HSM, data to be stored in a small-capacity storage medium with high access performance is determined based on detecting data with high access frequency in a predetermined data unit. Specifically, in the case of a storage device in which data is accessed in block units, data movement between a high-speed, small-capacity storage medium and a low-speed, large-capacity storage medium is formed by a fixed number of consecutive blocks. This is done in units of pages.

  Patent Document 1 discloses the following configuration as a system including a RAM and a ROM (RAM has a higher access speed). That is, the memory layout conversion information is determined based on information obtained by monitoring the access frequency in units of pages. Then, based on the arrangement conversion information, the memory is rearranged according to the frequency of use of the memory, and the address information of the page table of the memory management unit is rewritten so as to correspond to this (for example, see Patent Document 1). ).

Japanese Patent Laid-Open No. 2005-216053

  Since the above page is defined for all storage areas that the storage device permits access (input / output) to an external host, access frequency must be measured for all pages in the storage area. .

  In the HSM, it is necessary to manage the correspondence between the address for each storage medium and the address space for the entire storage area in which each storage medium is integrated. More specifically, for example, in order to access data of an arbitrary page under the configuration of HSM, an address indicating a storage medium itself storing (storing) each page in the address space, and storage The storage address of the page on the medium is required.

  The number of pages in the storage area of the storage device increases in proportion to the storage capacity that the storage device permits to be provided to an external host. The number of pages in the storage area is inversely proportional to the increase in capacity (size) per page.

  In order to suppress the increase in capacity per page and realize HSM in a small capacity page unit in the storage apparatus, it is necessary to measure access frequency and manage address information for a huge number of pages. That is, when the page size is reduced, there arises a problem that the capacity of addresses and access frequency information necessary for page management increases.

As a specific example, assuming that the management information capacity per page is 64 bytes and the total storage capacity of the storage is 100 TB, if the page size is 1 GB, the total management information capacity is
100TB / 1GB × 64byte = 6400KB
It is about 6.4 MB. On the other hand, when the size of one page is 1 MB, the total capacity of the management information is
100TB / 1MB × 64byte = 6400MB
And the size increases to about 6.4 GB.

  In order to avoid the increase in the size of the management information accompanying the increase in the capacity of the storage medium as described above, for example, the size per page may be increased to reduce the number of pages in the storage area. This method can be realized only by changing the page size, and does not require complicated control. However, since the size per page increases as the number of pages is reduced, there arises a problem that fine data control cannot be performed. For example, not all blocks in a frequently accessed page have a high access frequency, and a block having a high access frequency may actually be a part of the page. Even in such a case, since data movement between storage media is in units of pages, the entire page with only a part of frequently accessed blocks is moved to a small-capacity storage medium at high speed. It will be. That is, there is a problem that the storage area of a high-speed and small-capacity storage medium cannot be effectively used.

  Accordingly, an object of the present invention is to provide a storage medium control device, a storage device, a storage medium control method, and a program that can solve the above-described problems.

  The present invention has been made in order to solve the above-described problem. Data is transferred between a first storage medium having a predetermined access speed and a second storage medium having an access speed lower than that of the first storage medium. Access monitoring means for detecting presence / absence of access by a data unit when moving, and measuring access frequency by a data unit set in which a predetermined number of the data units are continuously formed, and access measured by the access monitoring means A target page set determining means for determining a target page set including a page to be stored in the first storage medium based on a measurement result of the frequency, the target page set included in the target page set, and the second A transfer for detecting a page stored in a storage medium and accessed as movement target data to be moved to the first storage medium. And object data detecting means, a storage medium control apparatus, characterized in that it comprises a data moving means for moving the moving object data from said second storage medium to said first storage medium.

  In addition, the present invention provides a storage unit including a first storage medium having a predetermined access speed and a second storage medium having an access speed lower than that of the first storage medium, the first storage medium, Access monitoring means for detecting presence / absence of access by a data unit when moving data to / from a second storage medium and measuring access frequency by a data unit set in which a predetermined number of the data units are continuously formed A target page set determining unit that determines a target page set including a page to be stored in the first storage medium based on a measurement result of the access frequency measured by the access monitoring unit; and A page that is included, stored in the second storage medium, and accessed is to be moved to the first storage medium And moving object data detecting means for detecting a moving object data, a storage device, characterized in that it comprises a data moving means for moving the moving object data from said second storage medium to said first storage medium.

  According to the present invention, as a storage medium control device that controls a plurality of storage media corresponding to HSM, management information for managing data without increasing the size of a data movement unit between storage media is provided. The effect that the increase in the total capacity can be effectively suppressed is obtained.

It is a figure which shows the structural example of the information processing system as embodiment of this invention. It is a figure which shows the structural example of the storage 100 as embodiment of this invention. It is a figure which shows the structural example of the page set management table 161 and the page management table 162 as embodiment of this invention. It is a figure which shows the structural example of the arrangement order data 163 as embodiment of this invention. It is a figure which shows typically the data movement from the high speed storage medium 111 to the low speed storage medium 112 as embodiment of this invention. It is a figure which shows the example of a process timing for the data movement from the high-speed storage medium 111 as an embodiment of this invention to the low-speed storage medium 112. FIG. It is a figure which shows the example of a process sequence which the storage 100 as embodiment of this invention performs until it measures access frequency and selects the page set of high speed storage medium arrangement | positioning object. It is a figure which shows the example of a process sequence for the page movement to the high-speed storage medium which the storage 100 as embodiment of this invention performs. It is a figure which shows the example of the page management table 400 assumed in the storage corresponding to HSM with the size of each item. It is a figure which shows the example of the page set management table 161 and the page management table 162 as embodiment of this invention with the size of each item.

<Configuration of information processing system>
FIG. 1 shows a configuration example of an information processing system in an embodiment of the present invention. The information processing system shown in this figure is configured by connecting a storage 100 and a plurality of hosts 200 via a network 300 so that they can communicate with each other.

  The storage 100 includes a storage unit that is a combination of a high-speed storage medium and a low-speed storage medium corresponding to HSM (Hierarchical Storage Management).

  The host 200 is a terminal device connected to the network 300, for example, and can access the storage 100 via the network 300 and acquire data stored in the storage 100. For this purpose, the host 200 transmits an access request to the storage 100 via the network 300. In response to this access request, the storage 100 reads the requested data from the storage unit and transmits it to the requesting host 200. That is, the storage 100 functions as a data distribution server.

<Storage configuration>
FIG. 2 shows a configuration example of the storage 100. The storage 100 includes a storage unit 110, an access monitoring unit 120, a target page set determination unit 130, a movement target page detection unit 140, a data movement unit 150, a working memory 160, a communication unit 170, and an access control unit 180.

  The storage unit 110 stores data to be transmitted in response to access from the host 200. The storage unit 110 is compatible with HSM and includes a high-speed storage medium 111 and a low-speed storage medium 112. The high-speed storage medium 111 is a storage medium that can be accessed at a certain speed or higher as access performance. Further, the low speed storage medium 112 is a storage medium whose access speed is lower than that of the high speed storage medium 111.

  The storage devices employed in the high-speed storage medium 111 and the low-speed storage medium 112 are not particularly limited. As an example, it is conceivable to employ an SSD (Solid State Drive) for the high-speed storage medium 111. An SSD using a NAND flash memory is widely known. In addition, it is conceivable to employ an HDD (Hard Disk Drive) as the low-speed storage medium 112.

  As can be seen from the above example, the high-speed storage medium 111 generally has a higher capacity unit price than the low-speed storage medium 112, and thus the capacity mounted as the high-speed storage medium 111 is smaller than the low-speed storage medium 112. Become. In addition, data is moved between the high-speed storage medium 111 and the low-speed storage medium 112 based on the access frequency. This data movement is performed in units of pages. It is assumed that data access to the storage unit 110 is performed in units of blocks. A page is formed as a set of a fixed number of blocks having consecutive addresses.

  Further, in the present embodiment, the entire storage capacity of the storage unit 110 composed of the high-speed storage medium 111 and the low-speed storage medium 112 is divided into a page set composed of a set of a certain number of pages with consecutive addresses. The data can be managed also by.

  The access monitoring unit 120 monitors a processing status in response to an access request in the access control unit 180, and detects an access in units of pages as one access status. The access monitoring unit 120 updates the access flag of the page registered in the page management table 162 in the work memory 160 based on the detection result. The access flag indicates whether or not the corresponding page is accessed.

  Further, the access monitoring unit 120 measures the access frequency not in units of pages but in units of page sets as the access status. The access monitoring unit 120 writes the measured access frequency information to the page set management table 161 in the work memory 160. Note that structural examples of the page set management table 161 and the page management table 162 will be described later. Further, the access monitoring unit 120 notifies the movement target page detection unit 140 of a page (access page) corresponding to the accessed data in response to the occurrence of access.

  The target page set determination unit 130 reads and acquires the access frequency for each page set stored in the page set management table 161 at an arbitrary time interval. Then, based on the acquired access frequency, a page set (target page set) including a page in which the high-speed storage medium 111 is stored is determined. The target page set determination unit 130 executes a process of arranging the page sets in descending order of access frequency in the process of determining the target page set. Then, sort order data 163 in which page sets are sorted in descending order of access frequency is generated from the sort result and stored in the work memory 160. Further, the target page set determination unit 130 also updates the contents of the page set management table 161 according to the determination result of the target page set. The procedure for determining the storage destination storage medium by the target page set determination unit 130 will be described later.

  The movement target page detection unit 140 detects a movement target page. The movement target page is a page to be moved from the low-speed storage medium 112 to the high-speed storage medium 111 for the storage destination. The movement target page detection unit 140 detects, as a movement target page, a page that satisfies the condition that it is included in the target page set and stored in the low-speed storage medium 112 as a page for which an access request has been made. The movement target page detection unit 140 instructs the data movement unit 150 to move the detected movement target page.

  For example, it is possible to perform data movement control in which data is moved to the high-speed storage medium 111 in units of page sets. However, even if the access frequency of the page set is high, the access frequency of each page included in the page set is not uniform, and therefore a page with low access frequency may be included in the page set with high access frequency. For this reason, when data is moved in units of page sets, a considerable number of pages with low access frequency are also stored in the high-speed storage medium 111, and the small-capacity high-speed storage medium 111 is used. It leads to lower efficiency. Therefore, in the present embodiment, a page estimated to have a higher access frequency is selected as a movement target page from a page set having a high access frequency.

  The data moving unit 150 executes control for moving the movement target page from the low speed storage medium 112 to the high speed storage medium 111 in accordance with an instruction from the movement target page detection unit 140. In addition, when the moving target page is moved, if the free space in the high-speed storage medium 111 is insufficient, the low-speed storage medium 112 in order from the page with the lowest access frequency among the pages stored in the high-speed storage medium 111. The control for moving to is also executed. Then, the contents of the page management table 162 are updated so that the page movement result is reflected.

  The work memory 160 is a part that stores management information necessary for moving data stored in the storage unit 110 between the high-speed storage medium 111 and the low-speed storage medium 112. Specifically, the work memory 160 stores a page set management table 161, a page management table 162, and alignment order data 163 as the management information.

  The communication unit 170 is a part for performing communication via the network 300. The communication unit 170 can respond to access from the host 200.

  The access control unit 180 is a part that performs access control on the storage unit 110 in response to data access performed by the host 200 via the network 300. For example, when the access control unit 180 receives an access request transmitted from the host 200 from the communication unit 170, the access control unit 180 reads the requested data from the storage unit 110. Then, the read data is transmitted from the communication unit 170 to the requesting host 200 as an access response. For example, the access monitoring unit 120 can detect the page or page set corresponding to the identified data by identifying the data specified by the access request received by the access control unit 180.

<Management information structure>
Next, an example of the structure of management information stored in the work memory 160 will be described. First, the page set management table 161 will be described with reference to FIG. The page set management table 161 shown in this figure has a structure in which items of storage area address, high speed storage medium storage destination address, low speed storage medium storage destination address, inclusion page, and access frequency are associated with each page set.

  The storage area address indicates the address of the corresponding page set in all storage areas formed by integrating the storage areas of the high speed storage medium 111 and the low speed storage medium 112 of the storage unit 110. The high-speed storage medium storage destination address indicates the address of a page stored in the high-speed storage medium 111 in the corresponding page set. When there is no page stored in the high-speed storage medium 111 in the corresponding page set, a specific address indicating invalidity is designated (indicated as “----” in the figure). The effective value stored in the high-speed storage medium storage destination address indicates that at least one page in the corresponding page set is stored in the high-speed storage medium 111. Instead of the high-speed storage medium storage destination address, a flag indicating whether or not a page in the corresponding page set is stored in the high-speed storage medium 111 may be considered.

  The low-speed storage medium storage destination address indicates the address of a page stored in the low-speed storage medium 112 in the corresponding page set. In the included page, page IDs for all pages included in the corresponding page set are indicated. The access frequency indicates the value of the access frequency for the corresponding page set detected by the access monitoring unit 120.

  Next, a structural example of the page management table 162 will be described with reference to FIG. The page management table 162 shown in this figure has a structure in which a page ID, a page set address, a storage destination storage medium, a storage destination address, and an access flag are associated with each page.

  The page ID is an identifier for identifying the corresponding page. The page set address indicates an address in the storage area of the storage unit 110 where the page set including the corresponding page is stored. Depending on the page set address, the page set including the corresponding page can be uniquely identified. The storage destination storage medium stores a value indicating which of the high-speed storage medium 111 and the low-speed storage medium 112 is the storage medium in which the corresponding page is stored. Here, the high-speed storage medium 111 and the low-speed storage medium 112 are identified by binary values “1” and “0”.

  The storage destination address indicates the address of the high speed storage medium 111 or the low speed storage medium 112 in which the corresponding page is stored. The access flag is a flag indicating whether or not the corresponding page has been accessed when the page set including the corresponding page becomes the target page set. Here, “1” indicates that there is an access, and “0” indicates that there is no access. This access flag is updated according to the page access detection result by the access monitoring unit 120.

  FIG. 4 shows an example of the structure of the sort order data 163. For example, the sort order data has a structure in which page set addresses for each page set are arranged according to the order of high access frequency for each sort order in ascending order from 1.

<Overview of page movement control>
With reference to FIG. 5, an overview of page movement control executed in the storage 100 of the present embodiment will be described. FIG. 5A schematically shows the entire storage area in the storage unit 110. In the storage area of the storage unit 110 in FIG. 5A, the storage areas of the high-speed storage medium 111 and the low-speed storage medium 112 are divided in units of page sets PS.

  Assume that the target page set determination unit 130 determines two page sets PS1 and PS2 in the storage area of the low-speed storage medium 112 as target page sets from the page sets in FIG. In this case, the page sets PS1 and PS2 include pages with high access frequency.

  FIG. 5B and FIG. 5C show the page sets PS1 and PS2 divided in units of page PG, respectively. Then, in response to the notification from the access monitoring unit 120, the movement target page detection unit 140 performs the movement target page detection process. As a result, the pages PG11, PG12, and PG13 in the page set PS1 shown in FIG. Is detected as a movement target page. Also, assume that pages PG21, PG22, and PG23 in the page set PS2 shown in FIG. 5C are detected as pages to be moved.

  The data moving unit 150 executes control to move the pages PG11, PG12, PG13, pages PG21, PG22, and PG23 detected as the movement target pages from the low speed storage medium 112 to the high speed storage medium. At this time, if the high-speed storage medium has a free space, the data moving unit 150 moves the page to be moved with respect to the free space. When the free space is insufficient, the pages stored in the high-speed storage medium 111 are moved to the low-speed storage medium 112 in ascending order of access frequency to secure the free space, and the page to be moved is placed in this free space. Move to store.

  Thus, in the present embodiment, the access frequency is measured in units of page sets, not in units of pages. And based on this measurement result, after determining the page set (target page set) including the page to be stored in the high-speed storage medium 111, the movement target page is determined according to the access status to the actual page, Data is moved in units of pages. As a result, in the present embodiment, as will be described later, it is possible to efficiently perform data movement control while suppressing an increase in the size of the management information (page set management table 161 and page management table 162). .

  FIG. 6 shows an example of processing timing of page movement control according to the present embodiment. As the processing timing, as shown in the figure, a certain movement control execution period continues. For example, in response to the start of the n-th movement control execution period, first, the access monitoring unit 120 starts monitoring the access status. Then, the monitoring of the access status is finished at a certain time before the end of the n-th movement control execution period. Subsequent to the end of the access status monitoring, the target page set determination unit 130 executes a process for determining the target page set. When determining the target page set, information on the access frequency of the page set obtained as an access status monitoring result in the same nth movement control execution period is used.

  The page movement control to the high-speed storage medium 111 based on the access status monitoring result for the page in the n-th movement control execution period and the determination result of the target page set is performed in the next (n + 1) movement control execution period. Executed in Further, in the (n + 1) th movement control execution period, each process of next access status monitoring and target page set determination is executed. Thus, the page movement control in the present embodiment is dynamically executed according to the current access status.

<Processing procedure>
A processing procedure example for page movement control executed in the storage 100 will be described with reference to the flowcharts of FIGS. 7 and 8. First, an example of a processing procedure until the target page set is determined will be described with reference to FIG. The processing shown in FIG. 7 corresponds to FIG. 6 and corresponds to access status monitoring and subsequent target page set determination processing in one movement control execution period.

  The target page set determination unit 130 starts a new movement control execution period (step S101), and then resets the access frequency value in the page set management table 161 (step S102).

  Next, the access monitoring unit 120 executes a process of monitoring the data access status over a certain time with the start of the movement control execution period (step S103). As one of the results of the access status monitoring process, the access frequency for each page set is measured. Therefore, the access monitoring unit 120 updates the access frequency value in the page set management table 161 so that the access frequency measurement result is reflected (step S104).

  Next, the target page set determination unit 130 refers to the access frequency of the page set management table 161 updated in step S104, and executes processing for arranging page sets in descending order of access frequency (step S105). . The target page set determination unit 130 generates a result of arranging the page sets as the arrangement order data 163 and stores it in the work memory 160.

  Next, the target page set determining unit 130 executes processing for determining the target page set (step S106). For example, the process for determining the target page set is as follows. The target page set determination unit 130 refers to the sort order data 163, and sequentially determines the page set as the target page set according to the order of high access frequency. For example, this determination process stores the target page set at high speed. Assuming that the data has been stored in the medium 111, the process is executed until there is no remaining capacity to store more page sets. Alternatively, a method of selecting a higher-level page set corresponding to a predetermined number from the sort order data 163 and determining it as a target page set can be considered.

  Next, the target page set determination unit 130 updates the page management table 162 so that the determination result of the target page set in step S106 is reflected (step S107). That is, for example, the target page set determination unit 130 tentatively determines an address when the target page set determined this time is stored in the high-speed storage medium 111, and the value of this address is determined in the page management table 162. Store in the high-speed storage medium storage destination address associated with the set. Further, an invalid value is stored in the high-speed storage medium storage destination address corresponding to the page set that has not been determined as the target page set.

  Next, with reference to the flowchart of FIG. 8, a description will be given of a processing procedure example after the target page set is determined until the movement target page is detected and the page movement control is completed. The processing shown in this figure corresponds to page movement control in one movement control execution period in correspondence with FIG.

  First, the access monitoring unit 120 notifies the movement target page detection unit 140 of the page (access page) requested to be accessed from the host 200 during the access status monitoring execution period in step S103 of FIG. S201).

  Further, the access monitoring unit 120 updates the access flag in the page management table 162 based on the access result for the page detected in the access status monitoring execution period in step S103 of FIG. 7 (step S202). That is, “1” indicating that there is an access is stored in the access flag corresponding to the page detected as the access page. On the other hand, “0” indicating that there was no access is stored in the access flag corresponding to the page for which there was no access request.

  Receiving the notification of the access page in the previous step S201, the movement target page detection unit 140 first collates the notified access page with the movement target condition as the detection process of the movement target page (step S203). Here, the movement target condition is a condition for an access page that corresponds to the movement target page. Specifically, the condition is that the page is included in the target page set determined in step S106 of FIG. 7 and is currently stored in the low-speed storage medium 112. In other words, the movement target page is a page that is stored in the low-speed storage medium 112 among the pages included in the target page set and satisfies the condition that access has been performed.

  In order to identify a page that satisfies the above condition, the movement target page detection unit 140 refers to the inclusion page and the high-speed storage medium storage destination address of the page set management table 161. As a result, the access page included in the target page set is identified from the notified access pages. The access page specified in this way becomes a candidate for a movement target page.

  Next, the movement target page detection unit 140 refers to the storage destination storage medium corresponding to the page ID of the access page in the page management table 162, and selects the movement target page from the candidates for the movement target page to the current low-speed storage medium 112. Filter the stored access pages. Access pages that remain as a result of this narrowing down are detected as pages to be moved.

  Next, the movement target page detection unit 140 determines whether or not at least one movement target page has been detected by the collation in step S203 (step S204). If it is determined that the movement target page is not detected (step S204: NO), the process is terminated without executing the page movement control. On the other hand, if it is determined that a movement target page has been detected (step S204: YES), the movement target page detection unit 140 moves the movement target page to the high-speed storage medium 111 with respect to the data movement unit 150. An instruction is given (step S205).

  In response to this movement instruction, the data moving unit 150 checks the free capacity of the high-speed storage medium 111 (step S206), and determines whether or not all the movement target pages can be stored in the recognized free capacity. (Step S207). Here, when the free capacity is insufficient and all the movement target pages cannot be stored (step S207: NO), the data moving unit 150 selects one page stored in the low-speed storage medium 112 and performs high-speed processing. Processing for moving to the storage medium 111 is executed (step S208).

  In step S208, the data moving unit 150 refers to the sorting order data 163 and recognizes the page set having the lowest access frequency. Next, the pages included in the page set with the lowest access frequency are recognized with reference to the inclusion page of the page set management table 161. Next, the data migration unit 150 refers to the storage destination storage medium of the page management table 162 and stores it in the high-speed storage medium 111 from the pages recognized as being included in the page set with the lowest access frequency. Select one of the pages that has been selected. Since the selected page is included in the page set having the lowest access frequency, it is stored in the high-speed storage medium 111, but the access frequency becomes low when compared with the current access status. Can be estimated. Therefore, the data moving unit 150 moves the selected page to the low-speed storage medium 112 and stores it. Then, the process returns to step S206. As a result, until the free capacity of the high-speed storage medium 111 is secured, the pages with the lowest access frequency are sequentially moved to the low-speed storage medium 112 by the above procedure.

  When it is determined that the free capacity of the high-speed storage medium 111 capable of storing the movement target page has been secured (step S207: YES), the data moving unit 150 performs the free capacity secured in the high-speed storage medium 111. The movement target page is moved (step S209). Next, the data moving unit 150 updates the page set management table 161 and the page management table 162 so that the page movement results in steps S208 and S209 are reflected (step S210). In the page set management table 161, the high-speed storage medium storage destination address corresponding to the page set including the moved page is updated. In the page management table 162, the storage destination storage medium for the moved page is updated.

<Management information size>
A comparison is made between the management information in the present embodiment and the management information assumed to be used in a normal HSM system. As an example, here, the page size is 1 MB, and the size of the page set is 256 MB. Further, the total storage capacity of the storage unit 110 is 1 PB (= 1048576 GB).

  First, FIG. 10 shows one page management table 400 as management information assumed to be used in a normal HSM system. In a normal HSM system, management is performed only in units of pages. For this reason, the management information for managing the page set as in the present embodiment is not prepared.

  The page management table 400 shown in this figure has a structure in which storage destination storage media, storage destination addresses, and access frequency items are associated with each page. The items in the page management table 400 are arranged according to the page address order.

  In addition, the storage destination storage medium is 1 bit because it is sufficient to obtain a binary value corresponding to the high-speed storage medium 111 and the low-speed storage medium 112. The storage destination address is set to 4 bytes according to the capacity of the storage medium. The access frequency is also 4 bytes.

In the above case, the information size required to manage one page is 65 bits = 1 bit + 4 bytes × 2
It becomes. The number of pages in the 1PB storage area is
1PB / 1MB = 1073741824
It becomes. Therefore, the size of the page management table 400 is
1073741824 × 65 bits = 8724152320 bytes = 8320 MB
It becomes.

  On the other hand, FIG. 9A and FIG. 9B respectively show the page set management table 161 and the page management table 162 of the present embodiment. First, the page set management table 161 shown in this figure has a structure in which items corresponding to each page set are arranged in the order of the address of the page set, and the storage area address and the included page in FIG. 3A are omitted. ing. In this case, since the page set can be uniquely specified by the column position in the page set management table 161, the storage area address item can be made unnecessary. Further, since the number of pages forming the page set is fixed, the item of the included page can be made unnecessary. As a result, the page set management table 161 in FIG. 9A can include only the three items of the high-speed storage medium storage destination address, the low-speed storage medium storage destination address, and the access frequency.

  In addition, the high-speed storage medium storage destination address and the low-speed storage medium storage destination address are each 3 bytes. These storage destination addresses are 1 byte smaller than the storage destination address of 4 bytes in the page management table 400 of FIG. This is due to the following reasons. One page set is 256 MB, which is 256 times the size of 1 MB, which is the size of one page. Therefore, the number of page sets stored in the storage area is (1/256) for pages stored in the same storage medium. As a result, as compared with the case where addresses are managed in units of pages, the amount of information necessary for address management of page sets can be (1/256). Therefore, the high-speed storage medium storage destination address and the low-speed storage medium storage destination address indicating the page set address can be 1 byte less than 4 bytes. Further, the access frequency is 4 bytes as in the page management table 400 of FIG. That is, the amount of information corresponding to one page set is 10 bytes (= 3 bytes + 3 bytes + 4 bytes).

  Further, the page management table 162 shown in FIG. 9B has a structure in which items of storage destination storage medium, storage destination address, and access flag are associated with one page, and the page in FIG. ID and page set address are omitted. The page management table 162 arranges the above items according to the page address order. Thereby, a page ID for identifying a page can be made unnecessary. Further, by arranging according to the page address order, the page set including each page can be uniquely identified, so that the page set address can be made unnecessary.

  In addition, the storage destination storage medium is 1 bit, and the storage destination address is 1 byte. The access flag is set to 1 bit because it is sufficient to take a binary value in accordance with the presence or absence of access.

  The storage destination address is 1 byte for the following reason. In other words, the area for storing a page in each storage area of the high-speed storage medium 111 and the low-speed storage medium 112 is restricted so as to be a predetermined range corresponding to an address for storing a page set including the page. For this reason, only the position of the page in the page set needs to be managed as the page address. Specifically, there are 256 positions where one page of 1 MB is stored for a 256 MB page set. Therefore, the storage address of the page management table 162 may be 1 byte. As a result, the information amount corresponding to one page in the page management table 162 is 10 bits (= 1 byte + 1 bit + 1 bit).

The total size of the management information in this embodiment is the sum of the sizes of the page set management table 161 and the page management table 162 shown in FIGS. 9A and 9B. The number of page sets in the storage area of the storage unit 110 is
1PB / 256MB = 4194304. Further, the number of pages is 1073741824 as in the case of FIG. Therefore, the sum of the page set management table 161 and the page management table 162 is
10 bytes x 4194304 + 10 bits x 1073741824 = 1320MB
It becomes. On the other hand, the capacity of the page management table 400 in FIG. 10 is 8320 MB, and it can be seen that the capacity of management information is greatly reduced in this embodiment. As described above, in this embodiment, an increase in the capacity of management information is suppressed, thereby enabling efficient data movement control.

The sort order data 163 has a structure in which page set addresses are simply arranged in accordance with the access frequency order of the page set. Therefore, its capacity is very small and hardly affects the increase in management information. Specifically, the page set address may be 3 bytes, similar to the high-speed storage medium storage destination address and the low-speed storage medium storage destination address. Therefore, the sort order data 163 is:
3 bytes × 4194304 = 12582912 bytes = 12 MB
It becomes.

  In the above embodiment, the movement target page detection unit 140 detects the movement target page based on the notification of the access page from the access monitoring unit 120. In addition to this, for example, the movement target page detection unit 140 may detect the movement target page by recognizing the access page with reference to the access flag of the page management table 162 updated by the access monitoring unit 120. Good. Further, the movement target page detection unit 140 holds information indicating the detection result in the work memory 160. In addition, for example, instead of the movement instruction from the movement target page detection unit 140, the data movement unit 150 refers to the information indicating the detection result, and moves the movement target page according to the timing at which this information is updated. You may do it. In addition, the process in which the target page set determination unit 130 arranges the page sets in the order of high access frequency does not necessarily have to be performed every time the access status monitoring is completed, for example, at the timing when the access frequency is updated. It may be. Further, the page movement control by the data moving unit 150 does not necessarily need to be executed at the start timing of the movement control execution period. For example, in a situation where the load of the storage 100 is high, until the load drops below a certain level. It may be executed after waiting.

  In the above-described embodiment, each storage unit 110 includes one high-speed storage medium 111 and one low-speed storage medium 112. However, at least one of the storage units 110 may include a plurality. It is also conceivable that the high-speed storage medium 111 and the low-speed storage medium 112 are provided in different storage devices.

  The above-described storage 100 has a computer system inside. The processing of the storage 100 described so far is stored in a computer-readable recording medium in the form of a program, and is executed by the computer reading and executing this program. Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the program may be distributed to a computer via a communication line, and the computer that has received the distribution may execute the program.

DESCRIPTION OF SYMBOLS 110 Storage part 111 High-speed storage medium 112 Low-speed storage medium 120 Access monitoring part 130 Target page set determination part 140 Movement target page detection part 150 Data movement part 160 Work memory 161 Page set management table 162 Page management table 163 Arrangement order data 170 Communication part 180 Access control unit 200 Host 300 Network 400 Page management table

Claims (5)

Detecting presence / absence of access based on a data unit when moving data between a first storage medium having a predetermined access speed and a second storage medium having a lower access speed than the first storage medium, and Access monitoring means for measuring access frequency by a data unit set in which a predetermined number of data units are continuously formed; and
A target page set determining unit that determines a target page set including a page to be stored in the first storage medium based on a measurement result of the access frequency measured by the access monitoring unit;
A page included in the target page set, stored in the second storage medium, and accessed is detected as movement target data to be moved to the first storage medium. Moving object data detection means;
Data moving means for moving the data to be moved from the second storage medium to the first storage medium;
A storage medium control apparatus comprising: A storage destination address of the first storage medium, a storage destination address of the second storage medium, data movement unit set management information in which access frequency information is associated with each data unit set, and a storage medium of the storage destination Further comprising a working memory for storing data unit management information in which storage destination storage medium information, a storage destination address, and access presence / absence information indicating the presence / absence of access are associated with each data unit;
The access monitoring means updates the access presence / absence information according to the detection result of presence / absence of access of the data unit, and updates the access frequency information according to the access frequency measurement result of the data unit set,
The target page set determining means determines the target page set based on the access frequency information,
The movement target data detection unit detects a page stored in the second storage medium based on the storage destination storage medium information, and detects a page that has been accessed based on the access presence / absence information. ,
The storage medium control device according to claim 1. Storage means comprising a first storage medium having a predetermined access speed and a second storage medium having an access speed lower than that of the first storage medium;
By detecting the presence / absence of access by a data unit when data is moved between the first storage medium and the second storage medium, and by a data unit set in which a predetermined number of the data units are continuously formed Access monitoring means for measuring access frequency;
A target page set determining unit that determines a target page set including a page to be stored in the first storage medium based on a measurement result of the access frequency measured by the access monitoring unit;
A page included in the target page set, stored in the second storage medium, and accessed is detected as movement target data to be moved to the first storage medium. Moving object data detection means;
Data moving means for moving the data to be moved from the second storage medium to the first storage medium;
A storage device comprising: Detecting presence / absence of access based on a data unit when moving data between a first storage medium having a predetermined access speed and a second storage medium having a lower access speed than the first storage medium, and An access monitoring procedure for measuring access frequency by a data unit set in which a predetermined number of data units are continuously formed;
A target page set determination procedure for determining a target page set including a page to be stored in the first storage medium based on a measurement result of the access frequency measured by the access monitoring procedure;
A page included in the target page set, stored in the second storage medium, and accessed is detected as movement target data to be moved to the first storage medium. A procedure for detecting data to be moved;
A data movement procedure for moving the data to be moved from the second storage medium to the first storage medium;
A storage medium control method comprising: In the storage medium control device,
Detecting presence / absence of access based on a data unit when moving data between a first storage medium having a predetermined access speed and a second storage medium having a lower access speed than the first storage medium, and An access monitoring procedure for measuring access frequency by a data unit set in which a predetermined number of data units are continuously formed;
A target page set determination procedure for determining a target page set including a page to be stored in the first storage medium based on a measurement result of the access frequency measured by the access monitoring procedure;
A page included in the target page set, stored in the second storage medium, and accessed is detected as movement target data to be moved to the first storage medium. A procedure for detecting data to be moved;
A data movement procedure for moving the data to be moved from the second storage medium to the first storage medium;
A program for running

Images