JPH04315233A - File managing system - Google Patents

Abstract

PURPOSE:To guarantee the reliability of a file and the efficient operation to plural users even when these users simultaneously perform access to the same file. CONSTITUTION:A log part 13 is provided to separately store data actually written by the user among a data substance corresponding to the identifier of these respective data in addition to a table to register pointers showing the storage position of this data substance at a data storage part 11 in the file based on correspondence to the identifier. When the file is opened, the new table is constructed each time based on reference to these tables and the log part 13 so as to register the pointer showing the storage position in the data storage part 11 concerning the not written data and to register the pointer showing the storage position in the log part 13 concerning the actually written data respectively based on the correspondence to the identifiers of the relevant data corresponding to the user opening the file. Regardless of the presence/ absence of write, this newly constructed table can read out the data in the same file to all the users.

Description

[Detailed description of the invention]

0001

[Field of Industrial Application] The present invention relates to a file management system that centrally manages access for reading and writing by multiple users of a computer system to various electronic files shared by these users. The present invention relates to implementing a system that guarantees efficient operation of accessed files for these users even when the same file is accessed at the same time.

0002

[Prior Art] In applications that assume that multiple users access the same file at the same time, it is common for one user to update the contents of a file while another user updates the same file. If you update the contents, this file will no longer produce the desired results. That is, if uncontrolled access is allowed to these shared files, the data content within the files may not be maintained correctly.

[0003] Conventionally, in order to prevent the data in the accessed file from being erased or the consistency of the contents being lost due to simultaneous access by these multiple users, (A) File reference has been carried out. , ie read only, allows simultaneous access by multiple users.

(B) During a user's access to a file, whether for reading or writing, the entire file system is exclusively locked and access by other users is prohibited.

(C) Limiting the number of users who can write to a file to one person, and prohibiting access by other users, even for reading, only while writing is in progress.

[0006] These controls are used to control file access by these multiple users.

[0007] An example of such an application that assumes that multiple users access the same file at the same time is a multi-user database system. ~Page 662, published by Ohmsha on May 30, 1989, introduces several methods based on this database system as methods for controlling file access by the plurality of users.

[0008]

[Problem to be Solved by the Invention] As described above, in an application that assumes that multiple users access the same file at the same time, the above (A) to (
By controlling access to these files in the manner described in C), it is possible to prevent data in those files from being accidentally deleted or
Although it is certainly possible to avoid loss of consistency in the contents, especially as in (B) or (C) above, if the entire file system is accessed by one user, If the system is exclusively locked, the degree of concurrency as a multi-user system will be significantly reduced. That is, once the file is locked in this way, other users will not be able to even reference the file, and if the scale of such a system is large, only the waiting time will increase as a whole, which is not practical.

[0009]Also, in order to avoid such a situation, there is a method of lowering the level of consistency as much as possible to ensure the degree of concurrency, but in this case, it is still necessary to The problem remains as to how to balance the level and degree of concurrency, and the essential problem of maintaining and efficiently operating such file systems with high reliability remains unsolved.

[0010] Furthermore, as a method for making it possible for other users to read these files even when a certain user is writing to the files, and also for maintaining the consistency of the data in these files, (1) When writing, the entire file is first copied, and data changes and the like are written to the copied file.

(2) The entire written file is written back to the file system.

[0012] Although there are methods such as [0012], when the size of the file becomes large, it takes time to perform the above-mentioned copying and space resources to temporarily hold the copied file, that is, Memory space becomes non-negligible, and efficiency still remains a problem.

[0013] The present invention was made in view of these circumstances, and even when the same file is accessed by multiple users at the same time as described above, it is possible to ensure the reliability of the file for all of these users. The purpose is to provide a file management system that guarantees efficient operation.

[0014]

[Means for Solving the Problems] In order to achieve these objects, the present invention provides: (a) a data storage section in which one or more data as in-file data, that is, data entities are stored;

(b) A first table section in which pointers indicating the storage positions of the one or more pieces of data in the data storage section are registered in correspondence with the identifiers of these pieces of data.

(c) A log section in which data actually written by the user among the data in the file is separately stored in association with its identifier.

(d) A table that is constructed each time the file is opened, corresponding to the user who opened the file, and in which writing is performed with reference to the first table section and the log section. For data that is not stored, a pointer indicating its storage position in the data storage section is
A second table section in which pointers indicating the storage positions of data actually written in the log section are registered in correspondence with identifiers of the corresponding data.

[0018] Even while data is being written through the log section (c), at least the data stored in the data storage section (a) is written through the second table section (d). Allow reading to occur.

[0019]

[Operation] The data storage section (a) and the first table section (b) are normally included in each file as a file system, but when each file is opened for access, In this case, the data storage section (a) in which all data entities are stored remains as is, and only the first table section (b) and log section (c) are stored on a user machine (workstation or terminal device), for example. ) into the system memory, where the second table section (d) is constructed based on the contents of the first table section (b) and log section (c). As mentioned above, the constructed second table section (d) is a pointer indicating the storage position in the data storage section (a) for data that has not been written, and a pointer indicating the storage position of the data that has not been written. The pointer indicating the storage location of the data in the log section (c) is
Each user (more precisely, the machine) is registered in correspondence with the identifier of the corresponding data, and each user (or more precisely, the machine) can register the data for each opened file, even if the data has not been written or has already been written. This constructed second table part (d
), it is possible to read all the data. In other words, this means that even if one user is writing to the file, other users are always guaranteed to be able to read the data in the same file.

[0020]Also, in order to maintain the consistency of the data in the file, it is desirable to limit the number of users who can write to the file at the same time to one person. The writing is
Since only the minimum necessary data that is the unit to be managed in the log section (c) above needs to be handled, the time for performing the above-mentioned copying and the space resources for storing the same file are wasted. Nor.

In this configuration, the data storage section (a), the first table section (b), and the log section (
c) may be provided all at once in one file, or each of these portions may be provided divided into a plurality of files.

[0022] Furthermore, the second table part (d) is
As long as each table is constructed separately for each user, and each user can refer to the corresponding table,
It may be constructed in any part other than the system memory of the user machine mentioned above.

Furthermore, according to the same configuration, as described above, only one user can write to a file at the same time, but the allowable time for writing to a file given to a user is sufficiently short. By doing so, it becomes possible to provide all of these users with the feeling of being able to write at all times.

[0024]

Embodiment FIG. 1 shows an embodiment of a file management system according to the present invention.

As shown in FIG. 1, this embodiment system mainly includes a file 10 that is shared by a plurality of users, and a predetermined exclusive control on simultaneous access to this file 10 by a plurality of users. At the same time, the file management unit 20 that centrally manages reading (reference) and writing (changing or adding) data in the same file 10 and multiple users who share the file 10, Management department 2
The computer is configured by a plurality of first to nth user machines (workstations, etc.) 30 that perform accesses such as read requests and write requests through 0.

[0026] Here, the file 10 identifies the data entity (byte string) contained therein by its identifier (ID number).
As shown in FIG. 1, its configuration includes a data storage unit 11 in which one or more data entities (hereinafter simply referred to as data) are stored;
and a table (hereinafter referred to as a static table in nature) section 12 in which pointers indicating the storage positions of a plurality of data in the data storage section 11 are registered in correspondence with the identifiers of each of these data. , and a log section 13 in which data actually written by the user among the above data is separately stored in correspondence with the identifier thereof.

On the other hand, each of the user machines 30, as shown in FIG. An access management unit 31 manages arbitration and granting of permissions for access requests such as data changes and data additions in collaboration with the file management unit 20, and each machine an arithmetic processing unit 32 that centrally executes various controls and processes based on events and user inputs that occur each time;
A system memory 33 as an internal memory in which various processing data etc. by the arithmetic processing unit 32 are temporarily stored;
Information input through the input device 35 consisting of a keyboard, mouse, etc. is decoded and transmitted to the arithmetic processing section 32, and display information given from the arithmetic processing section 32 is decoded and its contents are displayed on a CRT display. and an input/output control unit 34 that visually displays information through a display device 36 consisting of a display device 36, etc., and in particular, when the user issues an instruction (input) to open the file 10, , the contents of the table section 12 and log section 13 in the file 10 are transferred to the system memory 33 via the access management and file management section 20.
Based on the read contents, a pointer indicating the storage position in the data storage unit 11 for data that has not been written is set, and a pointer indicating the storage position in the data storage unit 11 for data that has actually been written is set in the log unit. A table 40 (hereinafter referred to as a dynamic table in nature) is constructed in the system memory 33, in which pointers indicating storage positions within the system memory 33 are registered in correspondence with corresponding data identifiers. It works like that.

That is, although the static table section 12 exists in the file 10 as well, in which a pointer indicating the data storage position in the data storage section 11 and the identifier of each data are registered in association with each other. , each user machine 30 that has actually opened this file 10 uses the dynamic table 40 built in its system memory 33 based on the contents of the static table section 12 and the contents of the log section 13 to The desired data will be referenced each time.

FIG. 2 schematically shows a mechanism for referring to data in the file 10 in the system of this embodiment through the data storage section 11, static table section 12, log section 13, and dynamic table 40.

That is, the data storage unit 1 shown in FIG.
1, its addresses "01", "02", "03"
”, “04”, etc. The data “Data A0”, “Data B0”, “Data C1”, “Data D0”, etc. are not changed at that time (exactly, as will be explained later). (prior to update processing based on file compression), and the storage addresses of each of these data are stored by the static table unit 12 also shown in the figure.
Its identifier (ID) “A”, “B”, “C”, “D”,
It is pointed through...

On the other hand, in the log section 13 shown in FIG.
Data “Data A1” and “Data D” are registered corresponding to the identifiers (ID) “A”, “D”, and “S”, respectively.
1" and "Data S0" are data that has been changed or added by any user after data update processing based on file compression, which will be described later. Here, for example, "Data A1" is data that has been added to the data storage unit. "Data D1" is the modified data of the "data A0" stored in the data storage section 11, and "data D1" is the modified data of the "data D0" stored in the data storage section 11. and "data S0" respectively indicate data newly added to the data storage unit 11 as data of the file. Note that the data registered in the log section 13 for the above change, that is, "data A
", "Data D], etc., are taken into the log unit 13 in units that maintain the minimum consistency of these data contents when writing them. related issues will be resolved.

The above dynamic table 40 constructed in the system memory 33 of each user machine 30 has an identifier stored in the log section 13 based on the contents of the static table section 12 and the contents of this log section 13. For registered data, the identifier is used as a pair with a pointer indicating the registration position of the same data in the log section 13, and for data whose identifier is not registered in the log section 13, it is stored in the static table section 12. Similarly, as a pair of the identifier and a pointer indicating the storage position of the same data in the data storage unit 11,
Each is constructed as a data reference table in which these identifiers and pointers are registered.

In other words, in this example, the log section 1
Assuming that this dynamic table 40 has been constructed with the above-mentioned changed or added data registered in 3, the pointing mode is shown in FIG. A points to the latest "L1" registered in the log section 13 by pointing to "L1" which is the registered position in the log section 13.
Data A1" is referred to.

◆Data B identified by the identifier “B” is stored in the data storage unit 11 by pointing to the storage address “02” in the data storage unit 11.
"Data B0" stored in "Data B0" is referenced.

◆Data C identified by the identifier “C” is stored in the data storage unit 11 by pointing to its storage address “03” within the data storage unit 11.
"Data C1" stored in "Data C1" is referred to.

◆The data D identified by the identifier “D” is registered at the “L” position in the log section 13.
2", the latest "data D1" registered in the log section 13 is referred to.

◆The data S identified by the identifier “S” is registered at the “L” position in the log section 13.
3", "data S0" which is additional data registered in the log section 13 is referred to.

[0038] This is the case.

FIG. 3 mainly shows the file management section 20 and each user machine 30 of the file management system of this embodiment.
The file management mechanism of the system executed through the access management unit 31 of the system is shown as a flowchart, and the operation of the system of this embodiment as a multi-user system will be described in further detail below with reference to FIG. do.

In other words, suppose that an open request for the file 10 is issued through the access management unit 31 of one of the user machines 30 (FIG. 3).
In step S1), the file management unit 20 receives this request and permits this open request on the condition that the file 10 is not being compressed (step S2 in FIG. 3). The access management unit 31 that has been granted this open request constructs the dynamic table 40 in its system memory 33 through the arithmetic processing unit 32 of its own machine (step S3 in FIG. 3), and opens the file 10. do. When constructing this dynamic table 40, the file 10
As described above, the static table section 12 and log section 13 at that point in time are referred to.

[0041] When the opening of the file 10 is completed in this way, the access management unit 31 handles "data reference requests", "
In response to a ``request for migration to refer to the latest data,'' ``request to change or add data,'' and ``request for file compression,'' the file management unit 20 performs the following processes, respectively.
This will be carried out in collaboration with

The operations for each of these requests will be listed below.

◇Data reference request If the request by the user is a request to reference the opened file 10 (step S20 in FIG. 3), the access management unit 31 determines the location of each data through the dynamic table 40 constructed above. (Step S21 in Figure 3),
Read the data in the file 10 based on the recognition (
FIG. 3 step S22). The data read in this way is temporarily held in the system memory 33 and displayed on the display device 36 through the arithmetic processing section 32 and the input/output control section 34.

◇ Migration request to refer to the latest data If the request by the user is a transfer request to refer to the latest data for the opened file 10 (step S30 in FIG. 3), the access management unit 31, an inquiry is made to the file management unit 20 to confirm that the file 10 is not being modified or added to by another user (user machine) (step S3 in FIG. 3).
1), and the contents of the static table section 12 and log section 13 in the file 10 on the condition that the migration is not in progress for other users to refer to the latest data for the same file 10 (step S32 in FIG. 3). is read again, and the contents of the dynamic table 40 are updated based on the newly read contents (step S33 in FIG. 3). That is, as a result, the latest data in the file 10 at that point in time is referenced.

◇ Data change/addition request If the request by the user is a data change or data addition request for the opened file 10 (step S40 in FIG. 3), the access management unit 31 requests the file management unit 20 to confirm that this file 10 is not being modified or added to by another user (step S41 in FIG. 3), and that other users can refer to the latest data regarding the file 10. The user is permitted to make these requests on the condition that the user is not in the process of transitioning (step S42 in FIG. 3). In other words, if this request is, for example, to change the data A shown in FIG. "Data A1" is registered in the log unit 13 along with its identifier "A", and if this request is also for the addition of data S as shown in FIG. ”
is additionally registered in the log section 13 along with its identifier "S" (step S43 in FIG. 3). Then, based on the contents registered in the log section 13, the contents of its own dynamic table 40 are updated (step S44 in FIG. 3).

◇File Compression Request If the request by the user is a request for compression processing of the opened file 10 (step S50 in FIG. 3), the access management section 31 makes an inquiry to the file management section 20, and 10 is not opened by another user (user machine) (step S51 in FIG. 3), and the same file 10 is not compressed by another user (step S52 in FIG. 3). In cooperation with the unit 20, (1) remove unnecessary data from the data storage unit 11, and sequentially fill in each piece of data accordingly (step S53 in FIG. 3);

(2) The contents of the static table section 12 are updated based on the organized data storage contents (storage positions) in the data storage section 11 (step S54).

(3) Move the latest data registered in the log section 13 to the data storage section 11 (step S5 in FIG. 3).
5).

(4) Furthermore, the contents of the static table section 12 are updated based on the contents of the data storage section 11 to which this data has been transferred (step S56 in FIG. 3).

(5) Finally, the contents of the log section 13 are cleared and emptied (step S57 in FIG. 3).

A series of processes such as [0051] are executed. In this empty log section 13, data newly changed or added by a subsequent data change/addition request will be registered together with its identifier.

[0052] Regarding this file compression process,
The process may be automatically executed through the file management unit 20 on the condition that no user has accessed the file.

As described above, according to the file management system according to this embodiment, even if writing and reading are performed simultaneously on the same file, file 10, only data that has been changed or added is processed. File 1 is managed separately through the log section 13.
It is no longer necessary for all data in 0 to be moved or copied when accessed such as writing. That is, it is only necessary to process data that is a unit of management in the log section 13. Note that the data that is managed as a unit in this log section 13 is a unit that maintains the consistency of these data contents at the minimum, so that problems related to consistency of data contents can be resolved satisfactorily. What happens is as described above. Moreover, according to the operations of the file management unit 20 and the access management unit 31 described above, only one user can write to the file 10, so although some limitations remain as a multi-user system, the same file 10 reliability is maintained well. However, if the time allowed for such write processing is set sufficiently short, it is possible to give all users the feeling of being able to write at all times.

When referring to the file 10, the data managed through the log unit 13 is referred to as that data, and the other data is referred to as the data stored in the data storage unit 11 in the dynamic table 40. Even if a user is currently writing to this file 10, all other users can freely read relatively new data (while writing). Otherwise, you can refer to the latest data through the migration process described above.

Further, on the condition that the file 10 is not opened by any user, the file 10 is compressed, and its data storage section 11 and static table section 1 are compressed.
Since the contents of 2 can be organized, the same file 10
It will also be possible to maintain it in a suitable state at all times without any waste.

By the way, in the above embodiment, the file management section 20 and the access management section 31 each have (
A) When a certain user makes a request to read data regarding the shared file 10, based on the contents of the dynamic table 40 that is constructed corresponding to the user, the data is registered corresponding to the identifier of the data requested to be read. Means for reading the data at the storage location indicated by the pointer and displaying it to the same user.

(B) Shared file 10 from a certain user
When there is a request to write data to the same file, the request is permitted on the condition that no other user has written to the same file, and the written data is stored in the log section 13 along with the identifier of the data. means of storage.

(C) When a user who has opened the shared file 10 requests migration to the latest data contents of the file, it is confirmed that no other user has written to the file, and that no other user has written to the file. A means for permitting a migration request on the condition that there is no migration request from a user, and updating the pointer registration contents of the dynamic table 40 based on the contents of the log section 13 at that time.

(D) The data storage unit 11 is automatically updated on the condition that the shared file 10 is not accessed by any user, or in response to a command from the only user who has the file open. The data contents and the pointer contents of the static table section 12 are updated according to the data stored in the log section 13, and the log section 1
A means to clear the contents of 3.

[0060] However, if only the basic management functions related to reading and writing of files are to be realized, the above (A) means for controlling reading and the above (B) means for controlling writing are required. It is sufficient to have the means to do so.

Furthermore, in the above embodiment, the case where each user machine directly handles the file 10 in cooperation with the file management section 20 has been described, but this is not limited to this example. A system based on a server-client system in which the clients (workstations, etc.) C1 to Cn collectively refer to and modify data in the file 10 through a file server S as shown in FIG. 4, or a system as shown in FIG. Each terminal T1 to Tn
The host computer H collectively processes data reference/change requests in the file 10 by the host computer H.
The file management system according to the present invention can also be applied to a terminal-based system.

Incidentally, in each of these cases, there is no need to construct the above-mentioned dynamic table on each user machine such as a client or a terminal, and as shown in FIG. 4 or FIG. Or host computer H
The files for each user may be collectively constructed in the file management section 20 disposed within the system.

In these cases, the data storage section 11, static table section 12, and
And the log section 13 need not be arranged all together in one file, but may be divided into a plurality of separate files and arranged.

[0064]

As explained above, according to the present invention, the pointer indicating the storage position of the data entity in the data storage section in the file is registered in the table in correspondence with the identifier of each data. In addition, a log section is provided in which the data actually written by the user among the above data entities is stored separately based on the correspondence with the identifier, and when a file is opened, the data is stored separately corresponding to the user who opened the file. For data that has not been written, a pointer indicating its storage position within the data storage section is used, and for data that has actually been written, a pointer indicating its storage position within the log section is used for the corresponding data. A new table that is registered based on the correspondence with the identifier of Even if you do so, other users are always guaranteed to be able to read data in the same file through this newly constructed table.

[0065] Furthermore, when writing to a file, only the minimum necessary data is imported into the log section, and additions and modifications to the same data are performed only through this log section, so there is no wastage of time. Efficient and reliable write processing is achieved without wasting space resources.

Therefore, even if the same file is accessed by multiple users at the same time, the degree of concurrency as a multi-user system will not decrease, and the reliability of the file will not be impaired. This will ensure smooth system operation.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of a file management system according to the present invention.

FIG. 2 is a schematic diagram schematically showing a reference mechanism for data in a file in the system of the embodiment.

FIG. 3 is a flowchart showing a file management mechanism by the system of the embodiment.

FIG. 4 is a block diagram schematically showing a configuration when the file management system according to the present invention is applied to a server-client type computer system.

FIG. 5 is a block diagram schematically showing a configuration when the file management system according to the present invention is applied to a host/terminal type computer system.

[Explanation of symbols]

Claims (2)

[Claims] Claim 1: A file management system that exclusively controls simultaneous access to a file shared by a plurality of users and centrally manages reading and writing of data within the file, wherein the data within the file includes: A data storage section in which one or more of the data is stored, and a pointer indicating the storage position of the one or more data in the data storage section are registered in correspondence with the identifiers of each of these data. 1, a log section in which the data actually written by the user among the data in the file is stored separately based on the correspondence with the identifier, and a log section in which the data actually written by the user among the data in the file is separately stored, and A table that is constructed correspondingly each time, and indicates the storage position of data that has not been written in the data storage section based on reference to the first table section and the log section. a second table section in which a pointer indicating a storage position in the log section for data actually written is registered in correspondence with an identifier of the corresponding data; When a user makes a request to read data regarding the file, the data is registered in accordance with the identifier of the data requested to be read based on the contents of the second table section that is constructed corresponding to the user. A first control means reads data at a storage location indicated by a pointer and displays it to the same user; and when a user requests data writing for the file, another user writes data to the same file. a second control means that permits the request on the condition that the request is not made and stores the written data in the log section together with an identifier of the data. . 2. In a file management system that exclusively controls simultaneous access to a file shared by a plurality of users and centrally manages reading and writing of data within the file, the data within the file includes: A data storage section in which one or more of the data is stored, and a pointer indicating the storage position of the one or more data in the data storage section are registered in correspondence with the identifiers of each of these data. 1, a log section in which the data actually written by the user among the data in the file is stored separately based on the correspondence with the identifier, and a log section in which the data actually written by the user among the data in the file is separately stored, and A table that is constructed correspondingly each time, and indicates the storage position of data that has not been written in the data storage section based on reference to the first table section and the log section. a second table section in which a pointer indicating a storage position of actually written data in the log section is registered in correspondence with an identifier of the corresponding data; Even while data is being written through the log section, at least the data stored in the data storage section is read out through the second table. File management system.