JPH1145203A - File management device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the file management device which efficiently gather access frequency information to one place as to a virtual file management system which evade concentration of access on a server. SOLUTION: The file management device arranged on a server/client type computer network is equipped with a virtual file management table 103, and virtual file identifiers given uniquely over servers, the server identifiers of the servers stored with respective files, and the file identifiers of the files are grouped and stored corresponding to the respective files. Further, an access request process part 106 is provided; when a specific terminal device makes an access request including a virtual file identifier, a server 101 and a file to be accessed are specified by referring to the virtual file management table 10 and an access request is sent to the specified terminal device.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file management apparatus for centrally managing a plurality of files existing in a plurality of servers, and more particularly, to a file management apparatus for centrally managing access information to each of a plurality of files. It is about.

[0002]

2. Description of the Related Art In recent years, in order to handle a large amount of data such as video data, a storage capacity required for a system tends to increase more and more. For this reason, in the server / client type computer network system, a plurality of servers are arranged on the network in order to secure more storage capacity and distribute access requests from terminal devices to improve the processing capacity of the entire system. In addition, there is a device for distributing and storing data (files) on a plurality of servers.

In such a server / client type computer network system in which a plurality of servers exist, each computer terminal device (hereinafter simply referred to as a terminal device) individually mounts a server using a mount table (server). And the correspondence table of the files stored there)
And file management. File management with this mount table requires rewriting the mount table when a new file is registered in the server, or when a file that has not been managed by itself has been needed. There is a disadvantage that management becomes complicated.

Therefore, the applicant of the present application has filed Japanese Patent Application No. 7-301.
At 798, a file management device for centrally managing files stored in a plurality of servers is arranged on a network,
A system that simplifies file management in the above case is proposed.

FIG. 6 is a configuration diagram of a virtual file management system in Japanese Patent Application No. 7-301798. The network 110 includes a plurality of computer terminals 109a, 109
b and a plurality of servers 101a and 101b are connected, and the server 101a has a configuration in which a file management device 102 described below is mounted.

[0006] The file management device 102 has a virtual file identifier centrally assigned to the plurality of servers corresponding to each of a plurality of files stored in the plurality of servers. The management table 103 stores a server identifier storing a file and a file identifier of each file in the server as a set.

[0007] Further, an access request processing unit 106 is provided to process an access request in the procedure described below. Here, it is assumed that the contents of the management table 103 of the file management device 102 are, as shown in FIG. 2, the server identifier 101b and the real file identifier R100 corresponding to the file of the virtual file identifier V001, and the terminal device 109a Consider a case where a read access request has been issued to a file corresponding to the virtual file V001.

First, the receiving unit 107 of the file management device 102 receives a read access request from the terminal device 109a. This access request is input to the access request processing unit 106, and the access request processing unit 10
6, the management table 103 is searched, and the server identifier 101b of the file corresponding to the virtual file identifier V001 is retrieved.
And the real file identifier R100.

The access request processing unit 106 further creates a read access request for the file identifier V001 as if the access request was issued from the terminal device 109a, and sends it out from the transmission unit 108 to the server 101b. Upon receiving the read access request for the file identifier R100, the server 101b
5 is the file identifier R1 from the file system 104b.
00 is read and sent to the terminal device 109a.

[0010] As described above, in Japanese Patent Application No. 7-301798, a specific file server 101a equipped with the file management apparatus 102 receives a file access request from a terminal device, and a virtual file identifier managed by the file management apparatus 102. By issuing a file access request to the server where the actual data of the file exists, as if there was a request from the terminal device, the terminal device can communicate to multiple servers regardless of the actual arrangement state of the file. It allows access to distributed files.

[0011]

However, in the above-described configuration, a file access request from a plurality of terminal devices is generated even though a plurality of servers are used.
When concentrated on a specific file that is in high demand, access is concentrated on a specific server and the response speed to the terminal device is reduced.

In order to avoid concentration of access to the server, it is necessary to predict the frequency of file access and store the files in a distributed manner when storing the files in the server. In order to realize this, it is necessary to collect accurate file access frequency information. However, at present, access information to the above system, specific server or specific file is not centrally managed. Had issues.

SUMMARY OF THE INVENTION In view of the above-mentioned problems, an object of the present invention is to provide a virtual file management system that provides a file management apparatus that collects access frequency information in a centralized and efficient manner.

[0014]

SUMMARY OF THE INVENTION The present invention relates to a file management system arranged on a server / client type computer network comprising a plurality of servers and a plurality of terminal devices sharing a file service provided by these servers. The device is assumed.

The file management apparatus includes a virtual file management table 103, and a virtual file identifier that is assigned to each of a plurality of files and is centrally assigned to the plurality of servers; Are stored as a pair with the server identifier of the server in which is stored and the file identifier of each of the above files.

Further, an access request processing unit 106 is provided. When an access request including the virtual file identifier is received from a specific terminal device, the server to be accessed is referred to by referring to the virtual file management table 103. 1
01 and a file, and sends an access request to the specified terminal device.

Further, the access recording unit 111 monitors the access request processing unit 106, and records the access information to each of the files for each virtual file identifier corresponding to each of the files.

With this configuration, access information to files existing in a plurality of servers can be centrally managed. As the access information to each file, the start time of access to each file and the end time of access to each file can be used. Also,
The number of access requests to each file can also be used.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a file management apparatus according to a first embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing a network system using a file management device according to a first embodiment of the present invention. As in the conventional example shown in FIG. 6, the terminal device 109a,
109b and a plurality of servers 101a and 101b are arranged, and a file management device 102 is arranged for the network 110. The file management device 102 is, as in the prior art shown in FIG.
01a or 101b, or an independent file management device as shown in FIG.

As shown in FIG. 2, the file management apparatus 102 is provided with a file management table 103, which stores a virtual file identifier corresponding to a specific file and a server identifier of a server storing the specific file. ,
A set of the specific file and a file identifier is stored.

Further, the file management device 102 is provided with an access request processing unit 106, which controls access between the terminal devices 109a and 109b and each of the servers 101a and 101b as in the prior art. Further, in the present invention, an access recording unit 111 and an access recording area 112 are provided, and access information such as the time of each access is managed as described below. Hereinafter, a more detailed configuration of the present embodiment will be described together with its operation.

The receiving unit 107 receives a file access request from the terminal device 109a or 109b via the network 110 and outputs the file access request to the access request processing unit 106. The access request processing unit 106 includes a receiving unit 107
The access request from the server is analyzed, and first, the virtual file identifier included in the request and the contents of the access request (access start or end) are output to the access recording unit 111, and the management table 103 is searched to find the virtual file identifier. Extract the corresponding server identifier and real file identifier.

Based on the server identifier and the actual file identifier thus extracted, the transmitting unit 108 is instructed to transfer the access request to the server corresponding to the server identifier. With this, the transmitting unit 108
The access request is transferred to the server 101a or 101b based on the instruction from the access request processing unit 106. The access request at this time is as if the terminal device 109
From a (109b), it is as if there was a direct access request to the server.

On the other hand, the access recording unit 111 examines the content of the access request received from the access request processing unit 106, and if the content is an access start request, stores the current time as the access start time for the received virtual file identifier. It is recorded in the access recording area 112 of the means. If the content of the access request is an access end request, the current time is used as the access end time for the received virtual file identifier.
Record at 12.

The request processing unit 105 provided in the server 101a or 101b processes the access request output from the transmission unit 108, processes the corresponding file using the file system 104a or 104b, and executes the terminal device 109a or 101b. Responds to 109b.

Next, a specific application example of this embodiment will be described. FIG. 2A shows the state of the management table 103.
, There are virtual file identifiers V001, V002, and V003, and the file corresponding to the virtual file identifier V001 is stored in the server 101a by the real file identifier R1.
00, and the file corresponding to the virtual file identifier V002 is stored in the server 101b as the real file identifier R2.
00, and the virtual file identifier V003 is managed by the server 101a as a real file identifier R101. It is also assumed that there is no record in the access recording area 112 related to the virtual file identifier V001 as shown in FIG.

In the above case, the virtual file identifier V00
Request to start accessing the file corresponding to terminal 1
The operation in the case where the number is from 09a will be described. First, the receiving unit 107 receives a request to start access to a file from the terminal device 109a, and outputs the access start request to the access request processing unit 106. Next, access request processing unit 1
06 analyzes the access start request received from the receiving unit 107 and outputs a request to start access to the virtual file identifier V001 to the access recording unit 111, and at the same time, based on the contents of the management table 103, the virtual file identifier V001. Of the real file identifier R100 stored in the server 101a, and instructs the transmitting unit 108 to transfer an access start request for the real file identifier R100 to the server 101a. Upon receiving this instruction, the transmitting unit 108
The access request to the real file identifier R100 is transmitted to the server 101a.

On the other hand, since the information received from the access request processing unit 106 is a request to start access to the file corresponding to the virtual file identifier V001, the access recording unit 111 records the information relating to the virtual file identifier V001 in the access recording area 112. The access start time is recorded from the state shown in FIG. 3A as shown in FIG. 3B.

Similarly, the status of the management table 103 is changed to the virtual file identifiers V001, V0 as shown in FIG.
02 and V003 are managed, and when the access record content corresponding to the virtual file identifier V001 is the content shown in FIG. 3B, the file corresponding to the virtual file identifier V001 The operation when the access end request to be made is received from the terminal device 109a will be described.

First, the receiving unit 107 receives a request for terminating access to a specific file from the terminal device 109a, and outputs the request for terminating access to the access request processing unit 106. Next, the access request processing unit 106 analyzes the access end request received from the receiving unit 107, outputs the access end request of the virtual file identifier V001 to the access recording unit 111, and
3 instructs the server 101a that the virtual file identifier V001 is the real file identifier R100, and instructs the transmitting unit 108 to transfer an access end request for the real file identifier R100 to the server 101a.

As a result, the transmitting unit 108 transmits the server 1
01a is transmitted to the real file identifier R100. On the other hand, the access recording unit 111 confirms that it is a request to end access to the virtual file identifier V001 received from the access request processing unit 106, and stores the access record of the virtual file identifier V001 in FIG.
From the state (b), the access end time is recorded as shown in FIG.

Further, when the access to a specific file is large (small), the specific file is moved to a server having better (bad) read performance. Therefore, the management table 103 stores the recorded contents related to the virtual file identifier V001 in FIG.
When the content is as shown in (c), the virtual file identifier V
Assume that the file corresponding to 001 has been moved from the server 101a to the server 101b.
03 is changed from FIG. 2A to FIG. 2B.

In such a state, the virtual file identifier V
If the terminal device 109b issues an access start request and an end request to the file corresponding to 001, the server to be accessed is the server 101b, but the contents recorded in the access record area 112 as the access record are only virtual files. Since the content is based on the identifier V001, the contents are recorded with the start time and the end time as shown in FIG.

In this way, even if a file managed by the virtual file management apparatus is moved between servers, access records are managed centrally by using virtual file identifiers, so that access records need not be changed. become.

Also, since the access record can be recorded for each virtual file identifier, the period and time during which access from the terminal is concentrated can be checked. Enables you to transfer the file to a high-performance server.

Further, here, the access start time and the end time are recorded as the access record, but the number of access requests may be recorded. In other words, when the access is started but the actual reading is stopped for a long period of time, or when a terminal device with high processing capacity reads a large amount of data, the load on the server is reduced only by the access start and end times. It is inaccurate as information for considering dispersion. Therefore, by recording the number of access requests, it is possible to accurately understand how much data has been read from the server between the start and end of access, and to perform more appropriate load distribution based on server performance become. Although the number of access requests was accessed once,
Those that have been stopped halfway are excluded.

In the above description, an access start request or an access end request is specified. However, the case where the NFS protocol used for a typical file sharing service using TCP / IP is used as a network communication protocol is described. In some cases, an access start request or an access end request may not be specified, and will be described below in such a case.

A virtual file management system according to a second embodiment of the present invention will be described with reference to the drawings.
FIG. 4 is a block diagram showing a file management device according to the second embodiment of the present invention. In addition to the access recording unit 111 and the access recording area 112 shown in FIG. 1, a last access time recording area 401 and an end time monitoring unit 402 are provided.

The operation of the virtual file management system configured as described above will be described. Receiver 107
Receives an access request for a file from the terminal device 109a or 109b via the network 110, and outputs it to the access request processing unit 106.

The access request processing unit 106 includes the receiving unit 10
7 is analyzed, the arrival of the access request to the virtual file identifier corresponding to the specific file is output to the access recording unit 111, the management table 103 is searched, and the server corresponding to the virtual file identifier is searched. Extract the identifier and the real file identifier. Next, access request processing section 106 instructs transmission section 108 to transfer the access request to the server corresponding to the server identifier, and in response to this instruction, transmission section 1
08 transfers an access request to the server 101a or 101b according to an instruction from the access request processing unit 106.

On the other hand, the access recording unit 111 confirms the recorded contents of the last access time recording area 401 for recording the last received time of the virtual file identifier received from the access request processing unit 106. At this time, if the last access time is not recorded, the current time is recorded in the access recording area 112 as the access record of the corresponding virtual file identifier as the access start time for the file corresponding to the received virtual file identifier. .
At the same time, the access recording unit 111 records the current time as the last access time of the virtual file identifier received in the last access time recording area 401.

The end time monitoring unit 402 checks the recorded contents of the last access time recording area 401, and when the last access time has passed a certain time (for example, 10 seconds) from the current time, the end of access to the virtual file identifier is completed. The last access time as the time is stored in the access recording area 1
Record at 12.

The request processing unit 105 provided in the server 101a or 101b processes the access request output from the transmission unit 108 as described above, and processes the access request.
The corresponding file is processed by using a or 104b, and a response is made to the terminal device 109a or 109b.

Next, a specific use example of the present embodiment will be described. FIG. 2A shows the state of the management table 103.
, There are virtual file identifiers V001, V002, and V003, and the file corresponding to the virtual file identifier V001 is stored in the server 101a by the real file identifier R1.
00, and the file corresponding to the virtual file identifier V002 is stored in the server 101b as the real file identifier R2.
00, and the file corresponding to the virtual file identifier V003 is stored in the server 101a by the real file identifier R1.
01.

The recorded contents corresponding to the virtual file identifier V001 in the access recording area 112 are the contents shown in FIG. 3D, and the recorded contents corresponding to the virtual file identifier V001 in the last access time recording area 401 are shown in FIG.
The operation when a terminal device 109a requests access to the file corresponding to the virtual file identifier V001 when nothing is recorded as shown in FIG.

First, the receiving unit 107 receives an access request for a file corresponding to the file identifier V100 from the terminal device 109a, and outputs the access request to the access request processing unit 106. Next, access request processing unit 1
06 analyzes the access request received from the receiving unit 107, outputs to the access recording unit 111 that the access request to the virtual file identifier V001 has arrived, and corresponds to the virtual file identifier V001 from the contents of the management table 103. It confirms that the file is stored in the server 101a as the real file identifier R100, and instructs the transmitting unit 108 to transfer an access request for the real file identifier R100 to the server 101a.

Upon receiving this instruction, the transmitting unit 108 transmits a request to start access to the real file identifier R100 to the server 101a. On the other hand, the access recording unit 111 extracts the virtual file identifier V received from the access request processing unit 106 from the recorded contents of the last access time recording area 401.
Confirm the recorded contents regarding 001. At this point, nothing is recorded in the last access time recording area 401, so the current time is recorded in the access recording area 112 as the access start time of the virtual file identifier V001, and the recorded contents of the access recording area 112 are shown in FIG. 3 (e)
Like

At the same time, the contents of the last access time recording area 401 of the virtual file identifier V001 are updated with the same contents as those written in the access recording area 112, as shown in FIG. Subsequently, when a new access request to the virtual file identifier V001 arrives, first, it is notified to the access recording unit 111 that a new access request has arrived, as described above.

At this time, the access recording unit 111 confirms the recorded contents (FIG. 5B) relating to the file identifier V001 in the last access time recording area 401. At this time, since the last access time of the file identifier V001 is recorded in the last access time recording area 401, the recorded content of the last access time recording area 401 is set as the current time as the last access time in FIG. The content shown in b) is updated to the content shown in FIG.

Thereafter, after a lapse of a predetermined time (for example, 10 seconds), the end time monitoring unit 402 records the virtual file identifier V001 for which the predetermined time has passed since the last access time recording area 401 was written. The content is detected, and the last access time is recorded in the access recording area 112 of the virtual file identifier V001 as the access end time of the virtual file identifier V001 as shown in FIG.

In this way, even when a request is made from a terminal that does not explicitly issue an access end request when accessing a file shared on the network, access records can be managed.

That is, when accessing a specific file stored in the server using the NFS protocol used for a typical file sharing service using TCP / IP as a communication protocol of the network, the terminal device:
First, the file identifier of the corresponding file is obtained from the server, and a read instruction (RE
AD) and the offset position from the beginning of the file (OFF
SET) and the read size (SIZE) are specified, and the access to the specific file is performed. Therefore, the server is not explicitly notified of the start of access to the file or the end of access. Even when file access to a server is performed using such a protocol, access information can be easily collected by using the present invention.

Although the number of servers managed by the file management apparatus 102 has been described as two, the number of servers may be three or more, or only one may be considered. Also, the case where only one file management device 102 exists in the network has been described, but a case in which a plurality of file management devices exist may be considered. Also,
The file management device 102 includes servers 101a and 10
Although described as being different from 1b, it may be included in any server.

Further, it is conceivable that information on the name of the user who accessed the terminal and the name of the terminal as the access information are also recorded at the same time. Further, each of the above embodiments may be configured by hardware, but similar functions may be configured by software.

Further, when each function is configured by software, a program for realizing all or a part of the software may be provided by a recording medium such as a floppy disk, a CD-ROM, or the like.

[0057]

As is apparent from the above description, the present invention records file access information for each virtual file identifier in a file management device that centrally manages files stored in a plurality of servers. By doing so, access information to files that exist on multiple servers can be centrally managed, and even if the server that stores the actual data of the file is changed without changing the virtual file identifier, it has been moved. It is possible to continuously manage the access information to the file.

Further, by recording the access start time and the access end time as the access information, in addition to the above-mentioned effects, it is possible to collect information for knowing the date / time during which access is concentrated. By using this information, it is possible to move and copy files in a time and a period to enable efficient use of a plurality of servers.

Also, by recording the number of access requests to access files as access information, it is possible to record the actual data transmission amount from the server to the terminal device. Utilizing this information enables more effective load distribution of the server. In addition, it can be used as effective data for constructing an optimal configuration such as server equipment.

When there is no access request to a specific file as an access start time before that,
By recording the time at which the above access request first arrives as the access start time, even in a system where the start of access from the terminal device is not explicitly performed, the processing in the terminal device is not changed at all. The access start time can be recorded.

Further, when the access request for the file does not arrive for a predetermined time as the end time of the access, the time when the access request arrives last is set as the end time. Even in a system in which an access end notification from a device is not explicitly made, the access end time can be recorded without changing the terminal device.

[Brief description of the drawings]

FIG. 1 is a block diagram of a virtual file management system according to a first embodiment of the present invention.

FIG. 2 is an internal configuration diagram of a management table according to the first and second embodiments.

FIG. 3 is an internal configuration diagram of an access record according to the first and second embodiments.

FIG. 4 is a block diagram of a virtual file management system according to a second embodiment of the present invention.

FIG. 5 is an internal configuration diagram of a last access time record according to the second embodiment.

FIG. 6 is a configuration diagram of a conventional file management system.

[Explanation of symbols]

 102 file management device 103 management table 106 access request processing unit 107 reception unit 108 transmission unit 111 access recording unit 402 end time monitoring unit

Claims (8)

[Claims] 1. A file management device arranged on a server / client type computer network including a plurality of servers and a plurality of terminal devices sharing a file service provided by these servers, comprising: Corresponding to, a virtual file identifier that is unified across the plurality of servers, a server identifier of a server where each of the files is stored,
A virtual file management table storing a pair of the file identifier of each of the files as a pair, and a server and a file to be accessed are specified based on an access request including the virtual file identifier from a specific terminal device, An access request processing unit that issues an access request to the specified terminal device; and an access that monitors the access request processing unit and records access information to each of the files for each virtual file identifier corresponding to each of the files. A file management device comprising a recording unit. 2. The file management device according to claim 1, wherein the access information to each file is a start time of access to each file and an end time of access to each file. 3. The file management device according to claim 1, wherein the access information to each file is the number of access requests to each file. 4. The file management apparatus according to claim 2, wherein the access start time for each file is a time when an access request for each file arrives for the first time. 5. The file management apparatus according to claim 2, wherein the access start time to each file is a time at which a request for notifying the start of access to each file arrives. 6. The method according to claim 2, wherein the end time of access to each file is a time at which the last access request arrives when an access request to each file does not arrive for a predetermined time or more. File management device. 7. The file management device according to claim 2, wherein the access end time for each file is a time when a request for notifying the end of access to each file arrives. 8. The file management apparatus according to claim 3, wherein the number of access requests to each file is the number of data read requests to each file.