JP2978770B2 - File distribution multiplexing client server system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a client-server system for distributing and multiplexing a virtual file of a server into a file of a client.
The present invention relates to a file distribution multiplexing client-server system that accesses a virtual file of a server and transfers the file access to a client having a real file corresponding to the virtual file.

[0002]

2. Description of the Related Art A client server system (C) connected to a local area network (abbreviated as LAN)
With the spread of the / S system (abbreviated as / S system), a larger capacity and higher reliability of the hard disk of the server are further required.
Therefore, it is common to connect a hard disk with a mirroring function or a disk array device composed of a plurality of hard disks on the server side.

FIG. 8 shows a block diagram of a "file management method for a local area network" described in JP-A-61-286920.
The purpose of the present invention is to improve the frequency of data transfer on the N602, and to distribute data on the local disks 616 to 6n6 of the workstations 601 to 60n to reduce access to the server.
The S system. However, due to recent improvements in LAN performance,
The fact is that the transfer speed of the LAN is higher than that, and even if the file is directly accessed to the local disk of the client instead of the local disk of the server, the LAN does not become a bottleneck.

Further, according to Japanese Patent Laid-Open No. 157541/1992 relating to the prior art of distributed multiplexed files, in a distributed processing system using multiplexed data generated from resources multiplexed in a processor connected to a LAN, the multiplexed data is It discloses that data received first is processed, and data received thereafter is used for consistency check with data received first. In this publication, each client of a transmission source accesses a multiplex resource of a LAN for each client. The client will be aware of the existence of the resource when accessing.

[0005]

In the above-mentioned prior art, the source of file access must access each of the distributed multiplexed files. Also, LAN
The need to reduce traffic due to file access has diminished due to the remarkable improvement in LAN performance. C /
In the S system, in order to increase the capacity and reliability of the disk device of the server, it has been conventionally adopted to cope with the problem by increasing the hardware. The reason is that, in a workstation such as a C / S system or a system mainly composed of a personal computer, the configuration of peripheral devices is poor, so that the capacity and reliability of an expensive hard disk such as a hard disk with a mirroring device and a disk array device are reduced. There is no other way but to give it.

When a plurality of clients access a server disk, the concentration load on the server may increase. The server is usually Small Comp
uter System Interface (hereinafter referred to as
abbreviated as "scsi") to connect a plurality of disk devices, but only one physical path, which is different from distributing files by separating physical paths by a channel device such as a large general-purpose computer. In addition, intensive overload is likely to occur in the file device.

An object of the present invention is to solve the above-mentioned problem in a C / S system, and to install a new disk device in a server by using hard disks of a plurality of clients as virtual disks of a server. Instead, the purpose is to increase the capacity and reliability of the file by distributed multiplexing.

[0008] of the present invention, the file server has a temporary Sofu Airu real file corresponding to the provisional Sofu Airu is stored in the distributed multiplexed into one or more clients connected through a local area network to the server In a distributed multiplexing client-server system, the server includes communication means for communicating with a client, and a client.
Client name and the virtual file stored by the client.
Client management table containing the virtual file name of the
Multiplicity and multiplicity indicating the degree of multiplexing of virtual files
Store the synchronization flag indicating the status of the file and the virtual file
Point to the client management table of the client
Management table that includes the
Storing the multiplicity of the virtual file corresponding to the
File management table containing a pointer to the synchronization management table
And a file access by a virtual file name from a client, and refers to the file management table, the synchronization management table, and the client management table, and stores a virtual file corresponding to the virtual file name and its real file. File access management means for transferring the file access to a client, the client comprising:
A disk device for storing a real file, communication means for communicating with the server and other clients, and sending file access by a virtual file name to the server;
Alternatively, receiving a file access from the server by a virtual file name, converting the virtual file name into a real file name, and accessing the real file on the disk device
File access management means for returning a result of file access to the server ; a file management table including a virtual file name and a pointer to the corresponding real file name; management data of the real file name and the corresponding real file And an actual file name table for storing

In a C / S system, the access speed from a client to a local hard disk of a server is as follows:
In the case of SCSI, it is generally about 5 MB / S. On the other hand, in the case of a 10BASE2 or 10BASE-T type that is compliant with IEEE802.3 (Ethernet type), the transfer speed is 10 MB / S and the transfer speed is twice as large. According to the present invention, a hard disk connected to each client is used as a virtual file of a server by using a high-performance LAN. The virtual disk of the server is composed of the disks of the clients, and the files to be accessed from each client are distributed, whereby the capacity, multiplexing and distribution of the files can be increased.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings.

Referring to the block diagram of one embodiment of the present invention, a file distributed multiplexing C / S system includes a server 1, a LAN 2, and clients 3 to n.
The server 1 includes a communication unit 11 for communicating with the clients 3 to n, a file access management unit 12 for transferring file access of the clients 3 to n to a client having a file, a virtual file name designated by the file access, and a number of virtual files. A file management table 13 including a pointer indicating a table indicating the severity, a multiplicity of the virtual file and a synchronization management table 14 including a pointer indicating a management table of a client holding the virtual file, a name of the client having the virtual file and the virtual A client management table 15 including the name of the file.

The clients 3 to n include a communication unit 31 for communicating with the server 1 or another client, and a file access management unit 32 for accessing a file owned by another client as a virtual file of the server 1.
A file management table 33 including a pointer indicating a real file name from a virtual file name, a real file name table 34 including a real file name and its management data,
Is provided.

In FIG. 1, file access to the server 1 by the clients 3 to n is performed by using a virtual file name, assuming that a virtual disk device accommodating a virtual file exists in the server 1. Server 1
The file access is transferred to the clients 3 to n, and the clients 3 to n convert the virtual files into the real files and access the files of the disk devices 35 to n5.

In this embodiment, referring to FIG. 2, which illustrates the distributed duplication of files, duplicated virtual files 161 to 164 exist in the virtual disk 160 of the server 1, and the virtual files 161 to 164 are provided.
The entity 4 is located in the disk devices 35, 45, 55, and 65 connected to the clients 3 to 6. Client 3 and the client 4 has the server 1 of duplicated virtual file 161 and 162 of the one, respectively, it is updated synchronization both files capital I by the synchronization management table 14. On the virtual disk 160 of the server 1, 1
Although it appears that the files are stored in the disk, the files are actually distributed to access the clients 3 and 4, respectively.

FIG. 3A illustrates a table group used by the file access means 12 of the server 1 and its association, and FIG. 3B shows a table group used by the file access means of the clients 3 to n and the table group. Is illustrated. FIG. 3A illustrates the file management table 130, the synchronization management tables 141 and 142, and the client management tables 151 and 152, which are examples of the file management table 13, the synchronization management table 14, and the client management table 15, which are managed by the server 1. 153 are shown. Since the synchronization management table 141 points to the respective client management tables 151 and 152, the multiplicity is 2. Further, synchronization management is performed by a synchronization flag indicating the synchronization state of the duplicated file. When access to the virtual file indicated by one client management table 152 fails, the synchronization flag is set to N.
G.

FIG. 3B illustrates a file management table 33 of the client 3 and an actual file name table 34. The client 3 obtains the real file name and the management data by the pointer from the virtual file name of the file management table 33 to the real file name table 34 in response to the file access from the server 1 by the virtual file name, and Access to.

The data flow of this embodiment will be described with reference to FIGS. Referring to FIG. 4 showing the flow of data at the time of data input, when data is updated from the client 4 to the virtual file 165, first, the updated data is transferred from the client 4 to the server 1 (step 101). The server 1 searches the file management table 13 and searches the synchronization management table 14 corresponding to the access target file by using the pointer. Server 1
Searches for a plurality of client management tables 15 to which the synchronization management table 14 points according to the multiplicity of the file to be accessed. The server 1 transfers the update data to the client 3 having the access target file based on the client management table 15 (Step 102). The client 3 accesses the real file 351 of the disk device 35 based on the virtual file name of the virtual file 165 received from the server 1 and writes the update data “X”. 1 (step 103). Based on the processing result, the server 1 returns a result indicating that the writing processing has been normally performed to the data input source client 4 (step 104).

Referring to FIG. 5 showing the flow of data at the time of restoration processing of a duplicated file, the server 1 searches the synchronization flag of the synchronization management table 14 for a file that is NG, and the client 1 that has written normally. From 3, NG
A copy process is started through the server 1 to the client 4 having the file (step 111, step 112). The file to be copied is locked so as not to be updated during copying. Upon completion of the copying, the server 1 returns the synchronization management table 14 to a normal state upon normal termination from the client 4 (step 113), and notifies the client 3 of the release of the file lock (step 114). If the synchronization flag in the synchronization management table 14 is NG, the file access unit 12 of the server 1 automatically checks the synchronization flag in the synchronization management table 14 by monitoring the timer, and when synchronization is lost, the server 1 Print a message on the console.

Next, the operation flow of this embodiment will be described with reference to FIGS. Referring to FIG. 6 showing the flow of the operation of the data input process, the client 3 transfers the update data to the server 1 (Step 201). The file access management means 12 determines from the file management table 13 whether the update data from the client 3 is a request for a virtual file or a request for a local file (step 202). If the request is for a local file (YES in step 202), the local file of the disk device connected to the server 1 is updated as it is (step 203), and the processing result is returned to the client 3 (step 209). If the update data is a request for a virtual file ( step 2
02), the synchronization management table 14 and the client management table 15 pointed to from the file management table 13 are referred to (step 204). The client name and the virtual file name are obtained from the client management table 15 by referring to the multiplicity flag in the synchronization management table 14 (step 205). A write request is sent to the client with the obtained client name (step 2
06). Upon receiving the write request, the client 4 converts the received virtual file name from the file management table 43 into a real file name, and executes a write process to a specified storage location (step 210). If the writing process of the client 4 is performed normally, the result of the writing process is returned to the server 1. If the writing process is normal (YES in step 207), the client 4 is sent to another client corresponding to the multiplicity number referred to in the synchronization management table 14. Then, a write request is made (step 208). If the write request to all the file multiplexing clients is successful (YE in step 208)
S), the normal result of the write processing is returned to the client 3 that has issued the data update request (step 209). When the write processing result becomes abnormal in the client 4 (NO in step 207), the server 1 to which the abnormal client belongs
Is set to NG in the synchronization management table 14 of FIG. When a read request is made, the same steps are performed. However, in the case of triple or more multiplexing, the server 1 returns a processing result to the requesting client if the first read request is successful. If the read request fails in the first client, the server 1 sets the synchronization flag to NG and issues a read request to another multiplexed client. The synchronization flag is updated is the time of system processing start and read / write execution, then the synchronization flag to display the abnormal message if Runara it the NG to the console of the server 1. The operator can select either to perform the recovery process manually based on the message or to perform the recovery automatically by monitoring the timer. In the recovery processing shown in FIG.
To find the client management table 15 of the client that has not been synchronized (step 221). A read request is issued to a normal client, and a write request is issued to a client whose synchronization flag "NG" is not synchronized (step 222). This is repeated until synchronization for the multiplicity is obtained (step 223).

[0020]

According to the above description, the first effect of the present invention is that a new disk device needs to be added to a server by providing a virtual disk in a server using a disk device of a client. It is gone.
As a result, excessive capital investment in the server can be suppressed.

A second effect of the present invention is that the file multiplexing of the server is performed by the virtual file of the virtual disk.
It is to realize. Thus, file multiplexing can be performed without adding a disk device to the server.

A further advantage of the present invention is that files of the server can be distributed. Since the server uses and manages the disk devices of a plurality of clients on the LAN, the read / write requests from the clients can be distributed to the read / write of the disk devices of the plurality of clients using the LAN.

[Brief description of the drawings]

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

FIG. 2 is a diagram illustrating duplication and decentralization of a file.

FIG. 3 is a diagram showing the configuration of a server / client table group. FIG. 3 (a) is a diagram for explaining the configuration and links of the server file management table, synchronization management table, and client management table. FIG. 4 is a diagram for explaining the configuration and links of the file management table and the actual file name table of the client.

FIG. 4 is a diagram illustrating the flow of data in data input.

FIG. 5 is a diagram illustrating the flow of data when synchronizing files.

FIG. 6 is a diagram showing a flow of an operation of a file access management unit of a server / client in data input.

FIG. 7 is a diagram showing an operation flow of a file access management unit of a server / client in a synchronous process.

FIG. 8 is a block diagram illustrating a conventional technique.

[Explanation of symbols]

 1 server 2 local area network 3 to n client 11 communication means 12 file access management means 13 file management table 14 synchronization management table 15 client management table 31, 41 communication means 32, 42 file access management means 33, 43 file management table 34, 44 Real file name table 35, 45 Disk unit 751, 451 file

Continuation of the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G06F 12/00

Claims (1)

(57) [Claims] 1. A server has a temporary Sofu Airu, the virtual
The actual file corresponding to file In the file distribution multiplexing client-server system to be stored in distributed multiplexed into one or more clients connected through a local area network to the server, communication means for communicating with the client And the client name and the temporary
Client management data including the virtual file name of the virtual file
And the multiplicity and multiplicity indicating the degree of multiplication of virtual files.
Store the synchronization flag indicating the status of the file and the virtual file
Point to the client management table of the client
Synchronization management table including the virtual file name and the corresponding virtual file.
Pointer to the synchronization management table that stores the multiplicity
A file management table including the file management table, a file access by a virtual file name from a client, and referring to the file management table, the synchronization management table, and the client management table, and a virtual file corresponding to the virtual file name and its actual File access management means for transferring the file access to a client for storing the file, wherein the client is a disk device for storing an actual file; and a communication means for communicating with the server and other clients. Sending a file access by the virtual file name to the server, or receiving a file access by the virtual file name from the server, converting the virtual file name into a real file name, and accessing the real file on the disk device. And Seth, and file access management means for returning <br/> results file access to the server, the virtual file name and the file management table containing a pointer to the actual file name corresponding thereto, the actual file name and the corresponding And a real file name table storing management data of the real file.