JP4445783B2 - File management apparatus and computer system using file management apparatus - Google Patents

Description

  The present invention relates to a method for managing information resources stored in a storage apparatus.

  Conventionally, in an environment where multiple storage devices are connected to a network, information resources are migrated between storage devices in consideration of the frequency of access to the information resources stored in the storage devices, the performance and cost of each storage device, etc. As a result, a technique for improving access efficiency has become widespread. Such technology includes, for example, migrating frequently accessed information resources to a storage device with a low operating rate or a high-performance storage device.

  In recent years, super-distributed environments that connect multiple networks via a wide area network such as the Internet and share resources such as files are becoming widespread. Even in such environments, access efficiency by migration and replication of information resources is increasing. Improvement is desired.

“Auto-stor”, [online], Archibi (ARKIVIO), [December 2, 2003 search], Internet <URL: http://arkivio.com/Arkivio_4_Pg_Final_3_2.pdf>

  However, the above-described conventional technology is premised on application on a local area network (LAN), and does not consider the difference in access speed due to the difference in network distance from the client. Therefore, in an ultra-distributed environment, when an information resource with high access frequency is migrated to a storage device with a low operation rate or a high-performance storage device, the distance on the network from clients that frequently accessed the information resource There was a risk that the access efficiency would be reduced.

  The present invention has been made to solve such a problem, and performs migration and replication of information resources based on an access history including information that can specify an access frequency and an access source in a super-distributed environment. An object of the present invention is to provide a technique for improving access efficiency.

  In order to solve at least a part of the problems described above, the present invention has the following configuration as the first configuration. That is, a plurality of storages and information resource management devices that are provided in association with the storage devices and manage information resources stored in the storage devices are connected to the network, and the information resource management devices A storage information management unit that manages storage information that registers in which storage device the information resource is stored in association with identification information allocated based on the identity of the content, and is connected to the network A sending unit for sending the information resource corresponding to the identification information included in the access request sent from another information resource management device to the other information resource management device; and at least the other information resource management An access history storage unit for storing information that can identify a device, and an access history including the identification information; and the access Based on a predetermined condition which is determined based on the gravel, the information resource, and summarized in that and a storage processing unit that performs predetermined processing for the other information resource management device is stored in the storage device to be managed.

  The identification information may be, for example, a number or a file name. When information resources having the same contents are stored in a plurality of storage apparatuses, they may be set to the same file name. When information resources having the same content are managed with different file names for each storage device, the same number may be assigned to the same content file.

  With this configuration, the information resource management device on the accessed side changes from the storage device managed by the own device to the storage device managed by the other information resource management device based on the access history of the own device. Since information resources can be stored, it is possible to autonomously reduce the load on the device itself. Also, other information resource management devices that make access requests store information resources that they want to access in the storage devices managed by their own devices, which can suppress network load and improve access efficiency. it can.

  In the information resource management apparatus of the present invention, the predetermined processing may be migration or replication of the information resource. For example, information resources accessed from a plurality of devices may be targeted for replication, and information resources accessed from a single device may be targeted for migration. In this way, it is possible to arrange the information resources in other storage devices while storing the necessary information resources in the storage device managed by the own device, and the convenience is improved.

  In the information resource management device of the present invention, the predetermined condition may be that the frequency of access exceeds a predetermined value. The predetermined value may be, for example, 10 accesses for 3 days. By doing so, the load can be distributed by storing frequently accessed information resources in another storage device, and convenience is improved.

  In the information resource management device of the present invention, the transmission unit further sends a reply to a user who has requested access to the information resource via another information resource management device, and the access history management unit It is also possible to manage an access history including user information that can specify the user ID, and the storage processing unit may perform a predetermined process based on a change in the information resource management device used by the user.

  In this way, information resources can be stored in an appropriate storage device in accordance with the movement of the physical position of the user, and convenience is improved. In addition, for example, when a user's movement is detected and the user is also accessing other information resources, if the information resources are also stored together, the access efficiency is improved. Can be preferred.

  The present invention has the following configuration as the second configuration. That is, a plurality of storage devices that store information resources and a plurality of information resource management devices that manage the storage locations of the information resources are connected to the same network, and all access histories to the information resources are integrated and managed. In the access history management device, the information resource management device holds an access history to the information resource stored in a storage device designated in advance. The access history management device is allocated based on the information that can identify the source information resource management device that has transmitted an access request to the information resource from the plurality of information resource management devices, and the identity of the information resource. In addition, a history acquisition unit that acquires an access history including identification information that can identify the information resource at a predetermined timing, and an instruction to change the storage device in which the information resource is stored in the information resource management device based on the access history And an instruction sending unit for sending the message.

  For example, the instruction sending unit may send an information resource acquisition instruction to the information resource management apparatus on the information resource acquisition side, or send an instruction to the information resource management apparatus on the information resource transmission side. May be sent out. The change instruction may be, for example, a migration instruction or a replication instruction.

  With such a configuration, the access history management device collectively manages the access history to all information resources on the network, and the information resource is allocated to each information resource management device based on the access history. Since an instruction to change the storage device to be stored can be sent, the load on the information resource management device can be reduced, and convenience is improved.

  In the access history management apparatus of the present invention, the predetermined timing may be set at predetermined intervals. “Every predetermined interval” may be collected every day, for example, every 10 minutes, every 6 hours, or at night, for example, at 12:00, when the network usage rate is relatively low. It is also possible to set the number of information resource management devices to be collected at a time and collect them sequentially at predetermined intervals. For example, if it is collected at dense time intervals, it is preferable that the state of access to information resources can be judged early, and if it is collected at sparse time intervals, the load on the network is reduced. This can be reduced and is preferable.

  For example, the predetermined timing may be an arbitrary timing in the information resource management apparatus. The “arbitrary timing” may be a predetermined time interval such as every 10 minutes, or may be every time an access occurs. This is preferable because the access history management apparatus can acquire the access history without collecting the access history.

  In the access history management device of the present invention, the history acquisition unit further acquires user information that can identify the user who has requested access to the information resource as part of the access history, and the instruction sending unit further includes: It is also possible to send out a change instruction for collectively changing storage devices that should store a plurality of information resources accessed by the same user. By doing so, it is possible to change the storage apparatus in which the information resource is stored according to the movement of the user, which is preferable. For example, when a user moves, information resources that were frequently used before moving can be expected to be used even after moving. Therefore, together with the information resource that triggered the detection of user movement, the information resources If the storage device in which the data is to be stored is changed, the convenience is improved.

  In the access history management apparatus of the present invention, the instruction sending unit may send information that can identify the information resource management apparatus that manages the storage apparatus before the information resource is changed in the change instruction. Since it is possible to determine the distance on the network, the congestion state of the network, the access status to the information resource, and the like based on the access history, for example, a change instruction in consideration of these may be sent. In this way, it is possible to flexibly determine from which storage device the information resource is changed, and convenience is improved. For example, the storage device before the change may be a storage device in which the distance on the network is arranged closest to the storage device after the change. This is preferable because the load on the network can be reduced.

  In addition to the information resource management apparatus and access history management apparatus described above, the present invention can also be configured as a computer system, an information resource management method, and the like composed of these. Further, the present invention can be realized in various modes such as a computer program for controlling the above-described information resource management apparatus, a recording medium that records the program, and a data signal that includes the program and is embodied in a carrier wave. In each embodiment, the various additional elements shown above can be applied.

  When the present invention is configured as a computer program or a recording medium on which the program is recorded, it may be configured as an entire program for controlling the information resource management device and the access history management device, and fulfills the functions of the present invention. It is good also as what comprises only a part. The recording medium includes a flexible disk, a CD-ROM, a DVD-ROM, a punched card, a printed matter on which a code such as a bar code is printed, an internal storage device of a computer (memory such as ROM or RAM), an external storage device, and the like. Various computer-readable recording media can be used.

Hereinafter, embodiments of the present invention will be described in the following items.
A. First embodiment:
A1. System overview:
A2. Function block:
A3. Access processing:
A4. Access history:
A5. Migration and replication processing:
A5 (1). Migration / replication judgment process:
A5 (2). User access file detection process:
A5 (3). File location information update processing:
B. Second embodiment:
B1. System overview:
B2. Function block:
B3. Access history collection processing:
B4. Migration and replication processing:
C. Variations:

A. First embodiment:
A1. System overview:
FIG. 1 is an explanatory diagram illustrating an overview of a system in the first embodiment. The computer system 1000 includes three local area networks LAN1, LAN2, and LAN3 via the Internet INT. A control node CN1 and a client CLN1 are connected to the local area network LAN1, and a storage node SN1 is connected to the control node CN1. As shown in the figure, the storage node SN1 stores files “file — 1a” and “file — 1b”, and “file — 1c” in the directory “dir — 1c”. The client CLN1 is used by the user foo. Similarly, in the local area networks LAN2 and LAN3, control nodes CN2 and CN3 and clients CLN2 and CLN3 are respectively connected, and the storage nodes SN2 and SN3 are respectively connected to the control nodes CN2 and CN3. In the storage node SN2, a file “file_1b” and “file_2a” are stored in the directory “dir_2a”. The file “file_1b” is the same file as the file “file_1b” stored in the storage node SN1. The storage node SN3 stores the file “file — 3a”. Each control node manages access to a file stored in a storage node connected thereto.

  When the user foo desires access to the file stored in the storage node SN1, the user foo makes an access request to the control node CN1 and receives the desired file via the control node CN1. When the user bar wants to access a file stored in the storage node SN1, the user bar is connected to the local area network LAN2 of the same network as the network to which the client CLN2 used by the user bar is connected. An access request is made to the control node CN2. The control node CN2 determines that the requested file is stored in the storage node SN1, and makes an access request to the control node CN1 that manages the storage node SN1. As described above, in this embodiment, when a file stored in the storage node SN1 is accessed, an access request is made to the control node CN1 that manages the storage node SN1. Further, as shown in the figure, the control node CN1 manages the history of accesses made to the files stored in the storage node SN1 via its own device. Similarly, the control node CN2 and the control node CN3 manage access histories to the files stored in the storage node SN2 and the storage node SN3, respectively.

  The user foo who has been using the client CLN1 moves to the client CLN3 connected to the local area network LAN3, and as indicated by the solid line arrow in the figure, via the control node CN3 to the control node CN1 and to the storage node SN1 An access request is made to the stored file “file_1a”. The control node CN1 transfers the file “file — 1a” to the client CLN3 via the control node CN3 as indicated by a one-dot chain line. Each time the control node CN3 accesses the file “file_1a”, the access history is accumulated. When the control node CN1 refers to the access history and detects that the access from the control node CN3 occurs frequently, the control node CN1 determines whether or not the file “file_1a” is accessed from another device, and is indicated by a thick arrow. As described above, the file “file_1a” is replicated or migrated to the storage node SN3 managed by the control node CN3. It is assumed that replication is performed when there is access from another device, and migration is performed when there is no access.

A2. Function block:
FIG. 2 is an explanatory diagram illustrating a functional block of the control node CN1 in the present embodiment. The control node CN1 is configured as a microcomputer including a CPU 100, a network interface 101, a storage interface 102, and a ROM 110. The ROM 110 includes an access history management unit 103, an access request processing unit 104, a file movement processing unit 105, and a location information management unit 106 as functional blocks that execute each function. Further, the hard disk 120 in the control node CN1 stores an access history 107 for storing a history of access requests to the control node and file location information 108 indicating a storage node in which the file is stored. The control node CN1 is controlled by the CPU 100.

  The network interface 101 has a function of communicating with the local area network LAN1 and the Internet INT. The storage interface 102 has a function of communicating with the storage node SN1. The access history management unit 103 stores, in the access history 107, the history of access to the storage node SN1 from other devices. Details of the access history 107 will be described later.

  When the access request processing unit 104 receives an access request from another device, the access request processing unit 104 refers to the file location information 108 and confirms the storage node in which the requested file is stored. If the requested file exists in the storage node SN1 under its management, the file is acquired from the storage node SN1 and sent to the request source apparatus. When the requested file exists in another storage node, an access request having itself as a request source is sent to the control node that manages the storage node.

  When the file movement processing unit 105 refers to the access history 107 and detects that a specific file is frequently accessed from a specific control node, the file movement processing unit 105 moves the file to a storage node managed by the control node. To do. There are two types of migration, migration and replication. If there is an access to the file from other than the control node, replication is performed, and if there is no access, migration is performed.

  The position information management unit 106 updates the file position information 108 when the file movement processing unit 105 moves the file. Further, the location information management unit 106 updates the file location information 108 held by the own control node in accordance with the update contents of the file location information 108 notified from the location information management unit 106 of another control node. The contents of the file position information 108 are also shown in the figure. The file location information 108 includes a file ID, a management CN, a storage SN, and a file path. The file ID is a number assigned based on the identity of the file, and the storage SN indicates the storage node where the file is stored. The management CN indicates a control node that manages each storage node, and the file path indicates the position of a file under the storage storage node. For example, the files having the file ID “1002” have the same contents and are stored in the storage nodes SN1 and SN2. In the present embodiment, as shown in the file path, the file name of the file ID “1002” is the same as “file — 1b”, but may be different. The file path “/ dir_2a / file_1c” of the file ID “1005” indicates that the file “file_1c” is stored in the directory “dir_2a”.

A3. Access processing:
FIG. 3 is a flowchart for explaining access processing in the present embodiment. In this figure, the case where the client CLN3 accesses the file ID “1001” stored in the storage node SN1 is illustrated.

  The client CLN3 sends the access request 150 to the control node CN3 of the local area network LAN3 to which the client CLN3 is connected (step Sa100). The access request 150 includes the file ID “1001” of the requested file, the user “foo” who made the access request, and the access source “CLN3”. The user “foo” is a user who is logged in to the client CLN3.

  Upon receiving the access request 150, the control node CN3 refers to the file location information 108, and acquires information on the storage node in which the file with the file ID “1001” is stored and the control node that manages the storage node. (Step Sa101). Since the file ID “1001” is stored in the storage node SN1 and is managed by the control node CN1, the control node CN3 sends an access request 160 to the control node CN1 (step Sa102). The access request 160 includes the file ID “1001” of the requested file, the user “foo” who made the access request, and the access source “CN3”. Since the access request 160 is sent by the control node CN3, the access source is the control node CN3.

  Upon receiving the access request 160, the control node CN1 refers to the file location information 108, and acquires information on the storage node in which the file with the file ID “1001” is stored and the control node that manages the storage node. (Step Sa103). Since the file ID “1001” is stored in the storage node SN1 and is managed by the control node CN1, that is, the device itself, an acquisition request for the file ID “1001” is sent to the storage node SN1 (step Sa104).

  The storage node SN1 receives the file acquisition request, and transfers the file with the file ID “1001” to the control node CN1 (step Sa105).

  The control node CN1 receives the file from the storage node SN1, transmits the file to the access source control node CN3 (step Sa106), and stores the access history (step Sa107).

  The control node CN3 receives the file from the control node CN1, transmits the file to the client CLN3 that is the access request source (step Sa108), and stores the access history (step Sa109).

A4. Access history:
FIG. 4 is an explanatory diagram illustrating the access history 107 in this embodiment. The access history 107 is composed of six items: ID, request date, request time, file ID, user ID, and access source. “ID” is a unique number assigned to one access history, and “request date” and “request time” indicate the date and time when the access request is received. The file ID, user ID, and access source reflect the contents of the access requests 150 and 160 described with reference to FIG. For example, the ID “3” has an access request at “17:32:20 on July 11, 2003”, and the access request is sent to the file ID “1001” by the user “foo” using the client CLN1. This indicates that the access request is for the following. For IDs “1” to “7”, the user “foo” has accessed the file IDs “1001” and “1002” from the client CLN1, and the control node CN3 for IDs “8” to “10”. To access the file ID “1001”. That is, it can be determined that the user ID “foo” has moved from the client CLN1 to the client CLN3.

A5. Migration and replication processing:
FIG. 5 is a flowchart illustrating the file movement process in the present embodiment. This process is performed by exchanging information between the control node CN1 and the control node CN3.

  The control node CN1 refers to the access history, and performs migration or replication processing of a file when an arbitrary file is frequently accessed via the control node CN3 (step S11). Such processing will be described later. Since the file storage node is changed by file migration or replication processing, the control node CN1 updates the file location information 108 (step S12).

  When the control node CN3 receives the file by migration or replication processing of the control node CN1 (step S20), the control node CN3 stores the received file in the managed storage node SN3 (step S21) and updates the file position information 108 (step S21). Step S22).

  Finally, either the migration destination / replication destination control node or the migration source / replication source control node, once the migration / replication processing is completed, the other control node that is not involved in migration / replication (this embodiment Then, the update contents of the file position information are notified to the control node CN2), and the other control nodes update the respective file position information 108 accordingly.

A5 (1). Migration / replication judgment process:
FIG. 6 is a flowchart for explaining file migration / replication processing in this embodiment. This process is a process executed by the file movement processing unit 105 of the control node CN1 with reference to the access history 107, and is the process shown in step S10 of FIG.

  The control node CN1 refers to the access history (step S30), and determines whether there is a predetermined access or more to the same file from any user via another control node (step S31). In the present embodiment, the predetermined access or more is assumed to be access three or more times in three days. If there is no access exceeding the predetermined level (step S31: NO), the process is terminated.

  If there is an access greater than or equal to the predetermined value (step S31: YES), the control node CN1 determines whether there is an access from another device to the file being accessed (step S32). If there is an access from another device (step S32: YES), the file itself needs to remain in the storage node SN1, so that the file is set as a replication target (step S33). If there is no access from another device (step S32: NO), the file is set as a migration target (step S34).

  In this embodiment, as shown in FIG. 4, the user ID “foo” is three times in two days from July 13, 2003 to July 14, 2003 as IDs “8” to “10”. Since access has occurred and access from another device has not occurred, the file ID “1001” is set as the migration target.

  Next, if there is a file accessed by this user in addition to such a file, the control node CN1 performs a process for setting these files as a replication or migration target (step S35). Such processing will be described later.

  The control node CN1 executes replication / migration of the file that is to be replicated or migrated (step S36).

A5 (2). User access file detection process:
FIG. 7 is a flowchart for explaining user access file detection processing in this embodiment. This is a process corresponding to step S35 in FIG.

  The control node CN1 refers to the access history (step S40), and determines whether or not there is a change in the user access source (step S41). Specifically, in the access history 107, when the user ID is the same and the access source is changed, for example, as shown in FIG. 4, when the access source is changed from the client CLN1 to the control node CN3. , It is determined that the user has moved from the client CLN1 to the client CLN3 connected to the local area network LAN3.

  If there is no change in the access source (step S41: NO), the process ends. When there is a change in the access source (step S41: YES), the control node CN1 determines whether there is another file accessed before the change based on the access history 107 (step S42). If the accessed file does not exist (step S42: NO), the process ends.

  If there is a file accessed before the change (step S42: YES), it is determined whether there is an access from another device to the file (step S43). When there is an access from another device (step S43: YES), the file itself needs to be left as it is in the storage node SN1, so that the file is set as a replication target (step S44). If there is no access from another device (step S43: NO), the file is set as a migration target (step S45).

  In this embodiment, the user ID “foo” has accessed the file ID “1002” before the change, and the file ID “1002” is also not accessed from another device. Target migration.

A5 (3). File location information update processing:
FIG. 8 is an explanatory diagram schematically showing the update processing of the file position information 108 in the present embodiment. This process corresponds to steps S12 and S22 of FIG.

  FIG. 8A shows the update status of the file position information 108 when the file is moved by the migration process. This is a process executed by the location information management unit 106 of the control node CN1.

  The file position information 108 indicates before migration processing, and the file position information 108a indicates after migration processing. When the migration is completed, the location information management unit 106 changes the management CNs of the file IDs “1001” and “1002” from “CN1” to “CN3” and the storage SN “SN1” as indicated by the broken line in the file location information 108a. To “SN3”.

FIG. 8B shows the update status of the file position information 108 when the file is moved by the replication process. In this embodiment, the file IDs “1001” and “1002” are migration processing, but in this figure, both files are described as being moved by replication processing.

  The file position information 108 indicates before replication processing, and the file position information 108b indicates after replication processing. When replication of the location information management unit 106 ends, as indicated by a broken line in the file location information 108b, records with file IDs “1001” and “1002” are added, the management CN is “CN3”, and the storage SN is “SN3”. Set to.

  Note that after the process shown in FIG. 5, either the control node CN1 or the control node CN3 has another control node that is not involved in the migration / replication process (control node CN2 in this embodiment) The update contents of the file position information 108 are notified, and the other control nodes update the respective file position information 108 accordingly. At this time, the update of the file position information 108 executed in the other control nodes is the same as the process shown in FIG.

  According to the control node CN1 of the first embodiment described above, the file migration or replication destination can be specified based on the access frequency and access source information specified from the access history. In other words, the accessed control node can move the file in consideration of the load state, so that the load can be reduced autonomously and the file transfer on the network is suppressed. Can improve convenience.

B. Second embodiment:
In the first embodiment, the control node manages the access history to itself and performs file migration and replication processing based on the access history. In the second embodiment, an access history manager that centrally manages all access histories of control nodes on the network is arranged, and the access history manager instructs the control node to acquire a file based on the access history. It was decided.

B1. System overview:
FIG. 9 is an explanatory diagram illustrating a system overview in the second embodiment. The computer system 2000 includes three local area networks LAN1, LAN2, and LAN3 and an access history manager 200 via the Internet INT. The configurations of the control nodes CN1 to CN3, the clients CLN1 to CLN3 connected to the local area networks LAN1 to LAN3, and the storage nodes SN1 to SN3 connected to the respective control nodes are the same as those in the first embodiment. Is omitted.

  Each control node manages the access history to the files stored in the storage node managed by each control node, and the access history manager 200 collects and manages the access history as shown by the solid line in the figure. To do. The access history collection process will be described later. When the access history manager 200 refers to the access history and detects that the access to the file “file — 1b” from the control node CN3 is frequently occurring, the access history manager 200 sends the file “ send instruction for “file_1b”. The control node CN3 receives such an instruction, refers to the file position information 108, and, as indicated by the one-dot chain line arrow, the file “file — 1b” exists and the file from the storage node SN2 having the shortest network distance from the own device. “File — 1b” is acquired. Such file acquisition processing will be described later.

B2. Function block:
FIG. 10A is an explanatory diagram illustrating functional blocks of the access history manager 200 in the present embodiment. The access history manager 200 is configured as a microcomputer including a CPU 201, a network interface 202, and a ROM 210. The ROM 210 includes an access history collection unit 203 and a file movement instruction unit 204 as functional blocks for executing each function. The hard disk 220 in the access history manager 200 stores an access history 205 collected from each control node and CN management information 206 for managing the control node. The access history manager 200 is controlled by the CPU 201.

  The access history collection unit 203 collects the access history of each control node at a predetermined interval and stores it in the access history 205. The file movement instruction unit 204 sends out a file movement instruction based on the access history 205 and the CN management information 206.

  The contents of the CN management information 206 are also shown in the figure. The CN management information 206 includes a CN name and an address. “CN name” indicates the name of the control node, and “address” indicates the IP address of the control node. For example, the IP address of the control node CN1 is “192.168.10.12”.

  FIG. 10B is an example of an access history 300 that the access history manager 200 collects from the control node CN1. The access history 300 includes a request date, a request time, a file ID, a file mediation CN, and an access source. “Request date” and “request time” indicate the date and time when the access request is received. The access source indicates the client or control node of the access request source, and the file mediation CN indicates the control node that manages the storage node that stores the file for which the access request has been made. For example, in the access history 301 surrounded by a broken line, the client CLN1 sends an access request to the file ID “4001” to the control node CN1, and the control node CN1 stores the file with the file ID “4001”. It shows that an access request is sent to the control node CN2 that manages the storage node SN2 and a file is acquired. Since the access histories 302 and 303 are access requests to files stored in the storage node SN1 managed by the control node CN1, the own device becomes the file mediation CN.

B3. Access history collection processing:
FIG. 11 is an explanatory diagram illustrating an access history collection process in the present embodiment. This process is performed while exchanging information between the access history manager 200 and each control node. In the present embodiment, for convenience of explanation, processing between the access history manager 200 and the control node CN1 is shown.

  The access history manager 200 refers to the time (step S50), and determines whether it is time to acquire the history (step S51). In this embodiment, the history is acquired every 6 hours. If it is not the time to acquire the history (step S51: NO), the process ends.

  If it is time to acquire a history (step S51: YES), the access history manager 200 sends a history acquisition instruction for the previous 6 hours to the control node CN1 (step S52). For example, if the history acquisition time is every day, 0:00, 6 o'clock, 12 o'clock, 18 o'clock and the date of the day is July 1, 2003, the history acquisition instruction sent at 6 o'clock includes “2003 / 07/01 00: 00: 00-2003 / 07/01 05:59:59 "includes an instruction to acquire an access history.

  When receiving the instruction, the control node CN1 refers to the access history (step S60), and extracts the access history that occurred at the designated time (step S61). Next, the extracted access history is transmitted to the access history manager 200 (step S62).

  When the access history manager 200 acquires the access history from the control node CN1 (step S53), the access history manager 200 stores it in the access history 205 together with the information of the acquisition source control node (step S54). The access history collected in this way is shown in FIG.

  FIG. 12 is an explanatory diagram illustrating the access history 205 in this embodiment. The access history 205 includes seven items: ID, request date, request time, file ID, file mediation CN, access source, and history acquisition CN. “ID” is a unique number assigned to one access history, and “request date” and “request time” indicate the date and time when the access request is received. The contents of the access history 300 shown in FIG. 10B are reflected in the file ID, file mediation CN, and access source. The history acquisition CN indicates the control node that acquired the access history. For example, the access history of the ID “3” is a history acquired from the control node CN1, and the client CLN1 accesses the file ID “1002” at “17:32:20 on July 11, 2003”. This indicates that the file has been transferred via the control node CN1.

  The access history indicated by the IDs “5” to “7” is the access history collected from the control node CN3, and the client CLN3 frequently requests access to the file ID “1002” via the control node CN2. It shows that it is going.

B4. Migration and replication processing:
FIG. 13 is a flowchart for explaining migration / replication processing in this embodiment. As described with reference to FIG. 12, the access history manager 200 detects that the client CLN3 frequently requests access to the file ID “1002” via the control node CN2, and sends the control request to the control node CN3. , Send out an instruction to acquire the file. In this embodiment, it is assumed that the network distance is closer to the control node CN2 and the control node CN1 in this order when viewed from the control node CN3.

  The access history manager 200 refers to the access history 205 (step S70), and the client CLN3 frequently requests access to the file ID “1002” via the control node CN2 as described in FIG. When this is detected, in order to move the file to the control node CN3, the file ID “1002” is set as a migration or replication process target (step S71). Next, an instruction to acquire the file ID “1002” is sent to the control node CN3 (step S72).

  When receiving the acquisition instruction, the control node CN3 refers to the file position information 108 and identifies the storage node in which the file ID “1002” is stored (step S80). In the present embodiment, the file having the file ID “1002” is stored in the storage nodes SN1 and SN2. Next, the control node CN1 identifies the control node CN having the file ID “1002” and the shortest network distance from the control node CN3 to the storage node storing the file (step S81). In this embodiment, the control node CN2. The control node CN3 sends an acquisition request for the file with the file ID “1002” to the control node CN2 (step S82).

  Upon receiving the file acquisition request from the control node CN3 (step S90), the control node CN2 determines whether there is an access from another device to the file and performs migration or replication of the file ID “1002”. A file is transmitted by the processing (step S91).

  When receiving the file from the control node CN2, the control node CN3 stores it in the storage node SN3 (step S83), and sends a file reception completion notice (step S84). Next, the file position information is updated (step S85).

The control node CN2 receives the file reception completion notification from the control node CN3, and updates the file position information (step S92).

  Finally, either the migration destination / replication destination control node, or the migration source / replication source control node, once the migration / replication process is complete, to other control nodes not involved in migration / replication. Thus, the update contents of the file position information 108 are notified, and the other control nodes update the respective file position information 108 accordingly.

  In this embodiment, the user ID is not collected as the access history, but if it is managed as the access history including the user ID, the movement of the user can be detected and the movement of the file can be instructed. Convenience is improved. For example, when the user accesses the files with the file IDs “1001” and “1002” via the CN1 from the client CLN1 before moving, and repeatedly accesses the file ID “1002” after moving to the client CLN3. In addition, the access history manager 200 determines from the access history that the user has also accessed “1001” before moving, and obtains both the file IDs “1001” and “1002” from the control node CN3. It is also possible to send an instruction to do so. At this time, the file ID “1002” may be acquired from the control node CN2, and the file ID “1001” may be acquired from the control node CN1, or both may be acquired from the control node CN1 that can be acquired at once. It is good.

  In this embodiment, the access history manager 200 collects access histories from each control node at a predetermined interval. However, the present invention is not limited to this. The control node may transmit the access history to the access history manager 200 at an arbitrary timing. The arbitrary timing may be various timings such as the time when the access history is generated and every hour.

  In this embodiment, the access history manager 200 designates a file move instruction and sends it to the control node on the file acquisition side. It is also possible to give a movement instruction. In addition, the acquisition instruction may be included in the acquisition instruction from which control node the file acquisition side control node is acquired.

C. Variations:
Although various embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and it goes without saying that various configurations can be adopted without departing from the spirit of the present invention.

It is explanatory drawing which illustrates the system outline | summary in 1st Example. It is explanatory drawing of the functional block of the control node in 1st Example. It is a flowchart explaining the access process in 1st Example. It is explanatory drawing which illustrates the access history in 1st Example. It is a flowchart explaining the file movement process in 1st Example. 6 is a flowchart illustrating file migration / replication processing in the first embodiment; It is a flowchart explaining the user access file detection process in 1st Example. It is explanatory drawing which showed typically the update process of the file position information in 1st Example. It is explanatory drawing which illustrates the system outline | summary in 2nd Example. It is explanatory drawing which illustrates the functional block of the access history manager in 2nd Example. It is explanatory drawing which illustrates the access history collection process in 2nd Example. It is explanatory drawing which illustrates the access history in 2nd Example. It is a flowchart explaining the migration / replication processing in the second embodiment.

Explanation of symbols

1000 ... computer system 100 ... CPU
DESCRIPTION OF SYMBOLS 101 ... Network interface 102 ... Storage interface 103 ... Access history management part 104 ... Access request processing part 105 ... File movement processing part 106 ... Location information management part 107 ... Access history 108, 108a, 108b ... File position information 110 ... ROM
120 ... Hard disk 150, 160 ... Access request 2000 ... Computer system 200 ... Access history manager 201 ... CPU
202 ... Network interface 203 ... Access history collection unit 204 ... File movement instruction unit 205 ... Access history 206 ... CN management information 210 ... ROM
220: Hard disk 205, 300, 301, 302: Access history

Claims (11)

Provided in association with the first storage device, a first file management apparatus for managing a plurality of files stored in the first storage device,
A plurality of file management devices provided for each of a plurality of storage devices different from the first storage device, wherein a plurality of file management devices for managing files stored for each of the plurality of storage devices, Connected to the network in addition to the first file management device,
A storage information management unit that manages storage information that registers in which storage device each file is stored in association with identification information allocated based on the identity of the contents of each file ;
Among the plurality of files stored in the first storage device, the identification information included in the access request sent from the second file management device among the plurality of file management devices connected to the network A sending unit that returns a specific file corresponding to the second file management device;
An access history storage unit for storing a history of access from the plurality of file management devices to the first file management device, including at least information that can identify the plurality of file management devices, and the identification information;
When it is detected that the frequency of the access from the second file management apparatus to the specific file determined based on the access history exceeds a predetermined value , the specific file is changed to the second file. A storage processing unit configured to set the specific file as a migration processing target or a replication processing target and execute a storage process based on the setting in order to store the second storage device managed by the management device. and, with a,
The storage processing unit
Based on the access history, when there is an access request to the specific file from a third file management device different from the second file management device, the specific file is set as the replication processing target,
A file management apparatus that sets the specific file as a migration process target when there is no access request to the specific file from the third file management apparatus based on the access history . The file management apparatus according to claim 1,
The feed out section is further to the user performing the access request to the second file management apparatus said specific file via performs reply of the specific file,
The access history management unit manages an access history including user information that can identify the user,
The storage processing unit
Whether the user had accessed another file other than the specific file stored in the first storage device via the first file management device before executing the storage process of the specific file If there is an accessed access file among the other files, the access file is stored in the second storage device managed by the second file management device. , Set the access file as either the migration process target or the replication process target,
The setting of the processing target for the access file of the storage processing unit is the same as the setting of the processing target of the specific file.
Based on the access history, when there is an access request to the access file from a third file management device different from the second file management device, the access file is set as the replication processing target,
Based on the access history, when there is no access request to the access file from the third file management device, the access file is set as a migration process target,
Furthermore, the storage processing unit executes storage processing for the specific file and the access file after setting a processing target for the specific file and the access file . The network, a plurality of storage devices for storing files, the plurality of storage and multiple file management apparatus for managing the files between devices, the access history management apparatus that manages the entire access history access history for all files Is a computer system connected to
Each file management device
Of the plurality of storage devices, holding an access history for a file stored in a storage device designated in advance,
The access history management device
Identify each file allocated based on the sameness of the contents of each file and information that can specify at least the source file management device that sent the access request to each file from the plurality of file management devices possible containing the identification information, the access history, a collection unit for collecting at a predetermined timing,
An all access history storage unit for storing the access history collected by the collection unit as the all access history;
Based on the entire access history, extract a specific file whose frequency of access to a file stored in another storage device managed by another file management device from the first file management device exceeds a predetermined value; An instruction sending unit that sends a change instruction including an instruction to change the storage location of the specific file from the other storage device that stores the specific file to the first storage device to the first file management device ; Have
When the first file management apparatus receives the change instruction, the first file management apparatus requests the second file management apparatus that manages the second storage apparatus, which is one of the other storage apparatuses, to acquire the specific file. Send
The second file management device has a second storage processing unit that receives the acquisition request and transmits the specific file to the first file management device by migration processing or replication processing,
The second storage processing unit
When there is an access request to the specific file from a file management device different from the first file management device based on the access history held by the second file management device, the specific file is stored in the first file by replication processing. 1 to the file management device,
Based on the access history held by the second file management apparatus, if there is no access request to the specific file from a file management apparatus different from the first file management apparatus, the specific file is transferred to the first file by a migration process. 1 to the file management device
Computer system. A computer system according to claim 3, wherein
The computer system in which the predetermined timing at which the collection unit of the access management apparatus collects is every predetermined interval set in advance. A computer system according to claim 3 , wherein
The computer system in which the predetermined timing at which the collection unit of the access management apparatus collects is an arbitrary timing in the file management apparatus. A computer system according to any one of claims 3 to 5 ,
The collection unit of the access history management device further collects user information that can identify a user who has requested access to the specific file as a part of the entire access history,
The instruction sending unit extracts the specific file and is accessed by a file management device different from the first file management device used by the user before using the first file management device. Of the files other than the specific file, an access file that is not stored in the first storage device is extracted,
The change instruction sent by the instruction sending unit to the first file management apparatus is an access file change that changes the storage location of the access file from another storage apparatus that stores the access file to the first storage apparatus. Including instructions,
When the first file management device receives the change instruction, the first file management device sends a third storage device that is one of the other storage devices that store the access file to the third file management device. Send out an access file acquisition request,
The third file management device that has received the access file acquisition request has a third storage processing unit that transmits the access file to the first file management device by migration processing or replication processing,
The third storage processing unit
When there is an access request to the access file from a file management device different from the first file management device based on the access history held by the third file management device, the access file is stored in the first file by replication processing. 1 to the file management device,
Based on the access history held by the third file management device, if there is no access request to the access file from a file management device different from the first file management device, the access file is transferred to the first file by migration processing. A computer system for transmitting to one file management apparatus. The computer system according to claim 3,
The computer system in which the first file management device and the second file management device are arranged at the closest distance on the network. The first file management apparatus that is provided in association with the first storage apparatus, a file management method for managing a plurality of files stored in the first storage device,
A plurality of file management devices provided for each of a plurality of storage devices different from the first storage device, wherein a plurality of file management devices for managing files stored for each of the plurality of storage devices, Connected to the network in addition to the first file management device,
(A) in association with identification information allocated on the basis of the identity of the contents of each file, the steps of the each file, managing storage information registered or stored in any storage device,
(B) a specific file corresponding to the identification information included in the access request sent from the second file management device among the plurality of file management devices connected to the network is transferred to the second file management Sending to the device;
(C) storing a history of access from the plurality of file management devices to the first file management device, including at least information that can identify the plurality of file management devices, and the identification information;
(D) When it is detected that the frequency of the access from the second file management apparatus to the specific file determined based on the access history exceeds a predetermined value , the specific file is stored in the second file . A storage step of executing either a migration process or a replication process for the specific file to be stored in the second storage device managed by the file management device ;
The storing step includes
(D1) executing a replication process as the storage process when there is an access request to the specific file from a third file management apparatus different from the second file management apparatus based on the access history;
(D2) A file management method including a step of executing a migration process as the storage procedure when there is no access request to the specific file from the third file management device based on the access history . The network, a plurality of storage devices for storing files, a plurality of file management apparatus for managing the files between the plurality of storage, and access history management apparatus for managing access history for all files as full access history Is a file management method in a computer system connected to
(A) Each file management device holds an access history for a file stored in a storage device designated in advance among the plurality of storage devices;
(B) The access history management apparatus is based on at least information that can identify a transmission source file management apparatus that has transmitted a file access request from the plurality of file management apparatuses, and the identity of the contents of each file. Collecting the access history at a predetermined timing, including identification information capable of identifying each of the allocated files;
(C) The access history management apparatus determines that the frequency of access from the first file management apparatus to a file stored in another storage apparatus managed by another file management apparatus is a predetermined value based on the entire access history. Including an instruction to extract a specific file having a frequency exceeding the frequency and to change the storage location of the specific file from the other storage device storing the specific file to the first storage device to the first file management device Sending a change instruction;
(D) When the first file management apparatus receives the change instruction, the specific file is transmitted to a second file management apparatus that manages a second storage apparatus that is one of the other storage apparatuses. Sending an acquisition request for
(E) the second file management device receiving the acquisition request and transmitting the specific file to the first file management device by migration processing or replication processing,
The step (e)
(E1) Based on the access history held by the second file management device, when there is an access request to the specific file from a file management device different from the first file management device, the specific file is obtained by replication processing. Sending to the first file management device;
(E2) Based on the access history held by the second file management device, if there is no access request to the specific file from a file management device different from the first file management device, the specific file is obtained by migration processing. Transmitting to the first file management device,
File management method. A first computer provided in association with the first storage device, a computer program for managing a plurality of files stored in the first storage device,
A plurality of computers provided corresponding to a plurality of storage devices different from the first storage device, wherein the plurality of computers managing files stored in the plurality of storage devices are the first storage device. Connected to the network in addition to the computer,
(A) a function of managing storage information that registers in which storage device each file is stored in association with identification information allocated based on the identity of the contents of each file ;
(B) Among the plurality of files stored in the first storage device, the identification information included in the access request sent from the second computer among the plurality of storage devices connected to the network A function of sending a specific file corresponding to the second computer ;
(C) a function of storing at least information capable of specifying the plurality of computers and a history of the access from the plurality of computers to the first computer , including the identification information;
(D) When it is detected that the frequency of the access from the second computer to the specific file determined based on the access history exceeds a predetermined value , the specific file is stored in the second computer. A function of setting the specific file as a migration processing target or a replication processing target and executing a storage process based on the setting in order to store the file in the second storage device managed by ,
The function (d) further includes:
(D1) Based on the access history, when there is an access request to the specific file from a third computer different from the second computer, the function of setting the specific file as a replication processing target;
(D2) based on the access history, when there is no access request to the specific file from the third computer, having a function of setting the specific file as a migration processing target,
A computer program for causing the first computer to execute the functions (a) to (d) . A recording medium in which the computer program according to claim 10 is recorded so as to be readable by a computer.

Images