JPH07281938A - File system - Google Patents

Abstract

PURPOSE:To simplify the complicated management of files by giving the same names to the files which are used for the same purpose. CONSTITUTION:When a client computer 100 has an access to a file prog.obj stored in a storage 1, an identifier stored in an identifier storage part 130 is pointed together with the file name prog.obj. A remote file controller 321 placed at the server side retrieves a directory 10 with the file name and the identifier used as keys and has an access to an OS-a file 71 which is proper to the computer 100.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file management method in a system in which a plurality of computers of different types are connected by a communication network.

[0002]

2. Description of the Related Art When storing computer data in a storage device such as a magnetic disk, a file system is created in the storage device, and the data is divided into files for storage and is used to identify each file. Manage by assigning a file name to. As such a file system, the locals described on pages 5 to 6 of "Design of UNIX Kernel (Kyoritsu Publishing Co., Ltd.)" by Maurice J. Bach, Sakamoto, Yoshikatsu Tada, Jun Murai In the file system, files are identified by directory names and file names.

A network file system can be used in a system in which a plurality of computers are connected by a communication network. In the network file system described by Hal Stern, translated by Akira Kurahoshi, translated by Hideki Sunahara, pages 178 to 184 of “NFS & NIS (Ascii Publishing),” the local file system of the server computer is used by multiple client computers. Share as a remote file system from. The file on the server computer is identified by the file name given by the server computer, and is also identified by the same method when accessed from the client computer.

[0004]

When a storage device is connected to only a single computer and is not shared by a plurality of computers, a conventional local file system technique provides a management method of giving a file name to each purpose of use of a file. It was possible.

However, when a plurality of client computers share the file system of the server computer, some files cannot be shared depending on the purpose of use. For example, C
A program file including instruction data for a PU (processor) cannot be shared by computers with different CPUs.

In the conventional network file system technology, even when a file owned by the server computer is accessed from the client computer, the file name assigned by the server computer identifies the file, so that the correspondence between the file and the file name is limited to one-to-one. Therefore, it is impossible to give the same file name to files such as program files that have different contents even if they have the same purpose of use.

For this reason, the user must manage by giving a different file name for each computer type or a different directory name for each computer type.
The management work was complicated.

Alternatively, in order to simplify the management work, files having the same file name but different directory names have been created on the server computer. In this method, even if the file is sharable, there are as many copies as there are client computers, and the storage device is wasted.

[0009]

In order to achieve the above object, the file control method of the present invention comprises: (1) communication between a plurality of client computers of different types and a server computer having a storage device for storing a file system. In a system connected by a network, an identifier corresponding to each type is assigned to each type of client computer, each client computer is provided with an identifier storage unit to store the identifier, and the file stored in the file system is identified. The file name and the identifier are stored in the file system as a key, and when the client computer accesses the file, the file to be accessed is selected from the identifier stored in the file name and the identifier storage unit and the identifier stored in the file system. .

(2) In the file system described in (1) above, when the client computer type is composed of a plurality of type elements, the identifier is divided for each type element, and the file contents and The dependency relationship of the type elements is stored in the file system, and the file to be accessed is selected from the file name, the identifier of the client computer, and the dependency relationship stored in the file system.

(3) In the file system described in (1) or (2) above, when the client computer accesses the file, the identifier is transmitted to the server computer together with the file name, and the server computer Select the file to access.

[0012]

According to the present invention, in the shared file system, the file is identified not only by the file name but also by the computer type such as the CPU, OS (operating system) and character type of the client computer to be accessed. As a result, even if the computer type is different, the files used for the same purpose can be given the same directory name and file name, without wasting the storage device.
The complexity of management work can be reduced.

[0013]

FIG. 1 is a block diagram showing an embodiment of a file system to which the present invention is applied, a system comprising client computers 100 and 200, a server computer 300, and a communication network 90 connecting these computers. It is shown. The server computer 300 has the storage device 1, and the respective applications 110 and 210 of the computers 100 and 200 access the storage device 1 via the communication network 90.

Each of the computers 100 and 200 is a CPU
It has 101, 201 and memories 102, 202. Applications 110 and 210 and OS 12 are stored in the memory.
0, 220 are placed, application programs 111, 211 and file buffers 112, 212 are placed in the application, and client side remote file controls 121, 221 are placed in the OS. Further, the client side remote file control is performed by the identifier storage units 130 and 230.
And communication controls 140 and 240.

In this embodiment, the identifier consists of CPU, OS, and character type classification.

The identifier storage units 130 and 230 are C of the computer.
The PU, the OS, and the identifier indicating the character type classification are stored. Identifier CPU fields 131 and 231 and identifier O, respectively
It is divided into S fields 132 and 232 and identifier character type fields 133 and 233. Identifier storage unit 130
Is the CPU of the client computer 100 is CPU-p, O
S is OS-a, character type is SJIS, identifier storage unit 230
Is the CPU of the client computer 200 is CPU-q, O
It is indicated that S is OS-b and the character type is SJIS.

The storage device 1 has a management table 2 and O
The S-a file 71, the OS-b file 72, and the common file 70 are stored. The management table 2 includes a directory 10 and a file table 30. The directory 10 has a file name field 11, an identification method field 12, and a link field 13. The identification method field 12 further includes an identification method CPU field 14, an identification method OS field 15, and an identification method character type field 1.
It is divided into six. The file name field 11 stores the file name. The identification method field 12 stores the dependency relationship between the identifier and the file. In the present embodiment, the dependency relationship is shown in a table, and whether the file depends on the CPU, OS, or character type classification is determined by the CPU field and OS, respectively.
Store in the field and character type field. The link field 13 stores a link to the file table.

The file table 30 has an identifier value field 31, a pointer field 32 and a list field 33, and the identifier value field 31 further includes an identifier value.
It is divided into a CPU field 34, an identifier value OS field 35, and an identifier value character type field 36. The pointer field 32 stores a pointer to the file.

In FIG. 1, the file name prog. src depends only on the character type identifier of the identifier by the directory entry 17, and by the file table entry 37 indicated by the link 80, the character type identifier SJIS
The file corresponding to is the common file 70 pointed to by the pointer 85.

The file name prog. obj depends on only the CPU identifier and the OS identifier among the identifiers by the directory entry 18, and indicates a file to which the file table entry group starting from the file table entry 38 indicated by the link 81 corresponds. In the figure, the file table entries 38, 3 are represented by the list 82.
9 corresponds. The file corresponding to the CPU-p and OS-a by the file table entry 38 is the OS-a file 71 pointed to by the pointer 86, and the file table entry 39 indicates the CPU-q and OS-a.
This indicates that the file corresponding to b is the OS-b file 72 pointed to by the pointer 87.

FIG. 2 is a flow chart showing the initialization operation of the identifier storage section. First, the value determined by the execution result of the CPU instruction, the value set in the read-only memory, the switch, the nonvolatile memory, or the like, or O
A determination value such as a value preset in S or a value stored in a file for defining the type is obtained (1000). Based on the judgment value, CPU, OS,
The type of character type is determined (1001). The discrimination result is set as an identifier in the identifier CPU field, the identifier OS field, and the identifier character type field (100
2).

FIG. 3 is a flow chart showing the operation of registering a file name in the directory. When the file name to be registered is received (1100), a new entry is secured (11
01), set each field. A file name is set in the file name field (1102). In the identification method field, it is set whether the file contents depend on the CPU identifier, the OS identifier, and the character type identifier (110).
3). Set the link field to empty (110
4).

FIG. 4 is a flow chart showing a file registration operation in the file table. The file name and identifier to be registered are received (1200) and a new entry is secured (1201). The directory is searched for a directory entry that matches the file name to obtain the identification method and link (1202).

Determine whether the link is empty (120
3) If it is empty, the secured file table entry is set so that the link field points (1204). The directory entry 17 is an example of this case.

If the link is not empty, the secured file table entry is connected to the end of the list starting from the file table entry pointed to by the link (1205).
The file table entry 39 is an example of this case.

The contents of the identifier corresponding to the field that is independent of the identification method field of the directory entry are masked (1206). In the figure, the field in the identifier value field that is "-" is masked. The masked identifier is set in the identifier value field of the secured file table entry (1207).

A new file is created (1208), and the pointer field of the reserved file table entry is set to point to the new file (1209).

The list field of the reserved file table entry is set to empty (1210).

FIG. 5, FIG. 6, FIG. 7, FIG. 8 and FIG. 9 show the file prog. It is a flowchart which shows the file access operation with respect to obj. Hereinafter, they will be described in order.

FIG. 5 is a flowchart showing the operation of the application 110 in file access.
The application program 111 uses an access method,
File name to access prog. obj, memory location of file buffer, remote file control on client side 12
1 (1300), and the access result is obtained from the client side remote file control 121 (1301).

FIG. 6 is a flow chart showing the operation of the client side remote file control 121 in file access. It receives the access method, the name of the file to be accessed, and the memory location of the file buffer (140
0), the access method is transmitted to the server side remote file control 321 of the server computer 300 via the communication control 140 (1401).

The access method is checked (1402), and in the case of reading, the file name and identifier are passed through the communication control 140 and the remote file control 3 on the server side of the server computer 300
21 (1410), receives the data and the access result (1411), and stores the data in the file buffer (1412).

When the access method is write, the write data, which is the file name, the identifier, and the contents of the file buffer, is sent to the server computer 300 via the communication control 140.
To the remote file control 321 on the server side (142
0), the access result is received (1421).

Finally, the access result is returned to the application program (1430).

7 and 8 are flowcharts showing the operation of the server side remote file control 321 in file access.

Receive an access method (1500). The access method is checked (1501), and if read, the file name and identifier are received (1502). If you write
Receives file name, identifier, and write data (15
03).

The directory entry that matches the file name is searched, and the identification method and the link are obtained by referring to the identification method field and the link field (1504). File prog. For obj, search directory entry 18
And get the link 81.

By the same method as (1206) in FIG. 4, the contents of the identifier corresponding to the field that is independent of the identification method field of the directory entry are masked (1505).

In the list starting from the file table entry pointed to by the link, a file table entry whose identifier value matches the identifier generated by masking in (1505) is searched and a pointer is obtained by referring to the pointer field (1506). From the client computer 100 to the file prog. When accessing obj, the list starting from the file table entry 38 pointed to by the link 81 is searched to obtain the file table entry 38, and the pointer 86
To get

The access method is passed to the local file control 350 (1510). Check the access method (1511),
For reading, the pointer is passed to the local file control 350 (1520), the data and the access result are received (1521), and the data and the access result are controlled by the communication control 340.
It returns to the client side remote file control 121 on the client computer 100 via (1522).

If the access method is write, the write data and pointer are passed to the local file control 350 ((1
530), receives the access result (1531), and returns the access result to the client-side remote file control 121 on the client computer 100 via the communication control 340 (1532).

FIG. 9 is a flow chart showing the operation of the local file control 350 in file access.
Receive an access method (1600). The access method is checked (1601), and if read, a pointer is received (1610), the content of the file pointed to by the pointer is read (1611), and the read data and the access result are returned (1612). If the access method is write, the pointer and the write data are received (1620), and the write data is written to the file pointed to by the pointer (162).
1), the access result is returned (1622). File prog. From client computer 100. In the case of accessing obj, the target pointer is the pointer 86.

[0043]

According to the present invention, files shared by computers having different types such as CPU, OS, and character type can be handled with the same directory name and file name.
The complexity of file management can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a flow chart showing initialization of an identifier storage unit.

FIG. 3 is a flowchart showing directory initialization.

FIG. 4 is a flowchart showing file table initialization.
To.

FIG. 5 is a flow chart showing an example of application processing.

FIG. 6 is a flow chart showing a client side remote file control process.

FIG. 7 is a flowchart showing a remote file control process on the server side.

FIG. 8 is a flowchart showing a server-side remote file control process.

FIG. 9 is a flow chart showing a local file control process.

[Explanation of symbols]

1 ... Storage device, 2 ... Management table, 70, 71, 72 ...
File, 90 ... Communication network, 100, 200,
300 ... Computer, 121, 221 ... Client side remote file control, 130, 230 ... Identifier storage unit, 14
0,240,340 ... Communication control, 321 ... Server side remote file control.

Claims (3)

[Claims] 1. In a system in which a plurality of client computers of different types and a server computer having a storage device for storing a file system are connected by a communication network, an identifier corresponding to the type is assigned to each type of the client computer. An identifier storage unit is provided for each of the client computers to store the identifier, and a file name and the identifier are stored in the file system as a key for identifying a file stored in the file system. When accessing the file, the file system selects a file to be accessed from a file name, the identifier stored in the identifier storage unit, and the identifier stored in the file system. 2. In claim 1, when the type of the client computer is composed of a plurality of type elements, the identifier is divided for each type element, and the contents of the file and the type element dependency relationship are determined. A file system which is stored in the file system and selects a file to be accessed from a file name, the identifier of the client computer, and the dependency relationship stored in the file system. 3. The method according to claim 1, wherein when the client computer accesses the file,
A file system for transmitting the identifier together with a file name to the server computer and selecting a file to be accessed by the server computer.