JP2003058519A - Distributed processing system, server device for use in this system, its method, its computer program, and recording medium with program recorded - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a server device which can distributes load so as to optimize the load of each computer. SOLUTION: A load distributing module 22a compares server capability data S1 of a server 20a with client capability data C1 of a client 30a determines distribution of load (processing A to E) to the server 20a and the client 30a. A server module 2a and a client module 31 execute processing determined by the load distribution module 22a. Consequently, load distribution can be performed so as to improve the processing efficiency of the server 20a and the client 30a.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed processing technique in which a plurality of computers connected via a network perform data processing while distributing the processing, and in particular, processing is performed so that loads on the plurality of computers are appropriate. The present invention relates to a distributed processing system for distributing data, a server device used therein, a method thereof, a computer program thereof, and a recording medium recording the program.

[0002]

2. Description of the Related Art In recent years, the complexity and sophistication of software has advanced, and the amount of data that the computer must process and the amount of processing have increased more than the improvement in the processing speed of the computer. In addition, the network environment has been improved and the performance of client computers has improved.

In such a technical background, not only one host computer performs all the processes, but
The distributed processing technology in which a plurality of computers distribute the processing and perform the processing in cooperation with each other has been actively researched.

In the distributed processing technology, the processing speed of the entire system cannot be expected to be improved unless the details of how to distribute the processing among a plurality of computers that perform the processing are designed. That is, which process is to be performed by which computer, which data is transferred to which computer to perform the process, how to call the process between the computers, and the like are appropriately designed, and the appropriate process is performed between the computers. If the processing is not distributed in, some computers will have processing capacity, while others will have data processing that cannot be processed.

Even if any computer has a processing margin, the data transfer amount in the network may become too large, and the computer may be in a data transfer waiting state. On the contrary, although the data transfer in the network is fast, the data transfer amount may be too small and the efficiency of the entire system may decrease.

Therefore, at the time of system design, various performances of a plurality of computers for data processing are sufficiently grasped, data communication speeds between the computers in a network connecting the plurality of computers are also well grasped, and each computer is grasped. In addition, it is necessary to design in consideration of load distribution so that all the networks exert the maximum processing capacity, and to perform a precise design so that the entire system exerts the maximum performance.

However, in general, computers and networks are rarely in the same state as when the system was constructed. For example, a computer may be replaced to improve its performance, a computer may be replaced due to a computer failure, a new computer may be added, or the computer may be allowed to perform other processing. May occur. Therefore, the processing capacity of each part always changes.

In this case, it is possible to avoid a situation in which the load is not optimally distributed, and even if the processing performance of some computers is dramatically improved, the processing speed as a whole is only slightly improved. Can not. As one of the techniques for solving such a problem, Japanese Patent Laid-Open No. 7-282
There is an invention disclosed in Japanese Patent No. 012.

The distributed operation support method in the distributed data processing system disclosed in Japanese Patent Application Laid-Open No. 7-282012 saves information such as data processing capacity in each computer as a database and stores the program for processing in the database. The load distribution status is changed with reference to the information described above. Therefore, when the above-mentioned system change occurs, by referring to the newly constructed database, it is possible to appropriately redistribute the load distribution of each process so as to match the changed system environment. .

In this distributed operation support method, processing capacity data is stored in advance in each computer, and each computer is connected to the network in order to completely construct the database from the beginning. When actually performing the processing, the processing capacity data is first read and registered in the database to avoid an unnecessary increase of the database.

[0011]

However, in the distributed operation support method in the distributed data processing system disclosed in the above-mentioned Japanese Patent Laid-Open No. 7-282012, a database irrelevant to data processing must be constructed and operated. In addition, data of system-dependent data processing capability must be stored in each computer in advance. Such a system can be operated in a relatively static network in a company, but the number of computers that perform distributed processing is very large, or even a small number of computers do not specify distributed processing. In some cases, it is difficult to store the processing capacity data in each computer in advance, which makes it difficult to construct a database.

Further, the system may dynamically change even when the system is in operation due to addition or reduction of computers for performing distributed processing, or increase / decrease of network load. The operation support method has a problem that it cannot cope with such changes.

The present invention has been made to solve the above problems, and a first object of the present invention is to provide a distributed processing system capable of distributing the load so that the load on each computer is optimized, and To provide a server device to be used, a method thereof, a computer program thereof and a recording medium recording the program.

A second object is a server used in a distributed processing system capable of distributing the load so that the load on each computer is optimized in consideration of the method of processing the data transferred via the network. It is to provide a device.

[0015]

According to one aspect of the present invention, there is provided a server device to which a client device is connected to perform processing while distributing a load, and relates to a processing capability of the client device. The load distribution of the server device and the client device is determined by comparing the data about the processing capacity of the server device and the data about the processing power of the server device with the data about the processing power of the server device. Load balancing means for

The load distribution means compares the data concerning the processing capacity of the server device with the data concerning the processing capacity of the client device to determine the distribution of the load on the server device and the client device, so that the load distribution means for the server device and the client device is determined. It is possible to distribute the load so that the processing efficiency becomes high.

Preferably, the server device further includes first measuring means for measuring the processing capacity of the server device at the time of load distribution, and the load distribution means at the time of load distribution measured by the first measuring device. The processing capacity of the server device and the data related to the processing capacity of the client device acquired by the acquisition unit are compared to determine the load distribution of the server device and the client device.

Since the load balancing means determines the load balancing of the server device and the client device based on the processing capacity of the server device at the time of load balancing measured by the first measuring means, the server device at the time of load balancing It is possible to appropriately distribute the load according to the processing capacity of.

[0019] Preferably, the load distribution means has data on the processing capacity of the server device and data on the processing capacity of the client device acquired by the acquisition means,
Compared to the data on the processing power of the network,
The load distribution of the server device and the client device is determined.

Since the load distribution means determines the distribution of the load on the server device and the client device based on the data regarding the processing capacity of the network, the load distribution means distributes the load so as to improve the processing efficiency of the server device, the client device and the network. Can be performed.

Preferably, the server device further includes second measuring means for measuring the processing capacity of the network at the time of load distribution, and the load distribution means is acquired by the data concerning the processing capacity of the server device and the acquisition means. The load distribution of the server device and the client device is determined by comparing the data related to the processing capacity of the client device with the network processing capability at the time of load distribution measured by the second measuring unit.

Since the load balancing means determines the load balancing of the server device and the client device based on the processing capacity of the network at the time of load balancing measured by the second measuring means, the transfer of the network at the time of load balancing It is possible to appropriately distribute the load according to the speed.

[0023] Preferably, the load balancing means periodically determines the load balancing of the server device and the client device.

Therefore, it is possible to distribute the load so that the load on the server device and the client device is always optimum.

According to another aspect of the present invention, there is provided a distributed processing method in a server device and a client device connected by a network, wherein the distributed processing method acquires data relating to the processing capacity of the client device held by the client device. Comparing the data about the processing capacity of the acquired server device and the acquired data about the processing capacity of the server device, Determining the load distribution of the server device and the client device.

Data relating to the processing capacity of the server device,
Since the load distribution of the server device and the client device is determined by comparing with the data regarding the processing capacity of the client device, it is possible to perform the load distribution so that the processing efficiency of the server device and the client device becomes high. .

According to another aspect of the present invention, there is provided a computer program for causing a computer to execute a distributed processing method in a server device and a client device connected by a network, the distributed processing method comprising: Obtaining data regarding the processing capability of the client apparatus held therein, obtaining data regarding the processing ability of the server apparatus retained by the server apparatus, acquired data regarding the processing ability of the server apparatus, and obtained client apparatus Determining the load distribution of the server device and the client device by comparing the data with the processing capacity of the above.

Data relating to the processing capacity of the server device,
Since the load distribution of the server device and the client device is determined by comparing with the data regarding the processing capacity of the client device, it is possible to perform the load distribution so that the processing efficiency of the server device and the client device becomes high. .

According to still another aspect of the present invention, there is provided a computer-readable recording medium recording a program for causing a computer to execute a distributed processing method in a server device and a client device connected by a network, The distributed processing method includes a step of acquiring data regarding the processing capacity of the client apparatus held by the client apparatus, a step of acquiring data regarding the processing capacity of the server apparatus held by the server apparatus,
Comparing the acquired data regarding the processing capacity of the server device and the acquired data regarding the processing capacity of the client device to determine the load distribution of the server device and the client device.

Data relating to the processing capacity of the server device,
Since the load distribution of the server device and the client device is determined by comparing with the data regarding the processing capacity of the client device, it is possible to perform the load distribution so that the processing efficiency of the server device and the client device becomes high. .

According to still another aspect of the present invention, there is provided a distributed processing system in which a server device and a client device are connected by a network to perform a process while distributing a load, and the client device is a process of the client device. The server device includes a holding unit for holding data regarding the capability, the server device acquires the data regarding the processing capability of the client device retained by the retaining unit, the data regarding the processing capability of the server device, and the acquiring unit. Load balancing means for determining the load balancing of the server device and the client device by comparing with the data regarding the processing capability of the client device obtained by the above.

The load distribution means compares the data concerning the processing capacity of the server device with the data concerning the processing capacity of the client device to determine the distribution of the load on the server device and the client device, so that the load distribution means of the server device and the client device is determined. It is possible to distribute the load so that the processing efficiency becomes high.

[0033]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (Embodiment 1) FIG. 1 is a diagram showing an example of appearances of a server and a client that constitute a distributed processing system according to Embodiment 1 of the present invention. The server and the client include a computer body 1, a display device 2, an FD drive 3 in which an FD (Flexible Disk) 4 is mounted, a keyboard 5, a mouse 6, a CD-
It includes a CD-ROM device 7 in which a ROM (Compact Disc-Read Only Memory) 8 is mounted, and a network communication device 9. Programs such as a server module, a client module, and a load balancing module described below are F
It is supplied by a recording medium such as D4 or CD-ROM 8. As each module is executed by the computer main body 1, distributed processing is performed. Further, each module may be supplied to the computer main body 1 from another computer via a network.

FIG. 2 is a block diagram showing a configuration example of the server and the client in the first embodiment of the present invention. The computer main body 1 shown in FIG. 1 is a CPU (Centra).
l Processing Unit) 10, ROM (Read Only Memor)
y) 11, a RAM (Random Access Memory) 12, and a hard disk 13 are included. The CPU 10 includes a display device 2, an FD drive 3, a keyboard 5, a mouse 6,
CD-ROM device 7, network communication device 9, ROM
11, processing is performed while inputting / outputting data to / from the RAM 12 or the hard disk 13. FD4 or CD
Each module recorded in the ROM 8 is temporarily stored in the hard disk 13 by the CPU 10 via the FD drive 3 or the CD-ROM device 7. CPU10
Distributed processing is performed by appropriately loading the modules from the hard disk 13 into the RAM 12 and executing them.

FIG. 3 is a block diagram showing a functional configuration of the distributed processing system according to the first embodiment of the present invention.
This distributed processing system includes a server 20a, a client 30a, and a LAN (Local Area Net) that connects them.
work) network 40. The distributed processing system shown in FIG. 3 is one in which processing is optimally distributed between two computers connected to a network, but it should be applied to a distributed processing system in which three or more computers are connected via a network. Is also possible.

The server 20a is a computer that controls distributed processing. The client 30a is
It is mainly the computer on the controlled side.
Generally, a computer with high processing capability is assigned to the server 20a and a computer with low processing capability is assigned to the client 30a, but it is not strictly determined.

The server 20a includes a server module 21, which is a program that operates only on the server 20a, a load balancing module 22a, which is a program that compares and evaluates the processing capacity of the server 20a and the processing capacity of the client 30a. It includes processes A to E which are program groups. Further, the client 30a is the client 3
The client module 31 is a program that operates only with 0a, and processes A to E that are a plurality of program groups.

Processes A to E are a group of programs which may be operated by the server 20a or the client 30a, and the process surrounded by the solid line runs on the computer and the dotted line. It means that the process surrounded by does not work on the computer.
Therefore, the server 20a executes the process A, the process B, and the process C, and the client 30a executes the process D and the process E. These processes may consist of many more modules.

A computer that executes many of these processes has a large process load, and a computer that executes few or no processes has a small process load. Therefore, if a computer with high processing capacity executes many processes and a computer with low processing capacity executes few processes, the overall processing efficiency is improved. Server capacity data S1 is stored in the server 20a as data that objectively represents the processing capacity of these computers.
To hold the client capability data C1. Appropriate values evaluated in advance during system design are set for these data.

The load distribution module 22a has a function of optimally distributing the processes A to E according to the processing capability of each computer. The load distribution module 22a compares the server capacity data S1 stored in the server 20a with the client capacity data C1 stored in the client 30a, and distributes the processing so that the load on each computer is optimized.

FIG. 4 is a flow chart for explaining the processing procedure of the distributed processing system according to the first embodiment of the present invention. When the system is activated by a user operation or batch processing, the server 20a first starts the processing of the server module 21 (S11). At this time, in order to distribute the processing load between the server 20a and the client 30a, the server 20a uses the load distribution module 2
The process of 2a is started (S12).

The load balancing module 22a is used by the server 20.
The server capacity data S1 of a is read (S13), the client capacity data C1 of the client 30a is read (S14), and these data are compared (S15). When the server capacity data S1 is larger than the client capacity data C1, that is, when the processing capacity of the server 20a is higher than the processing capacity of the client 30a,
Increase the processing modules of the server 20a (S17),
The processing of the load balancing module 22a is ended (S1).
8). For example, most of the processes A to E are operated by the server 20a, and the remaining very few processes are operated by the client 30a.

If the server capacity data S1 is smaller than the client capacity data C1, that is, if the processing capacity of the client 30a is higher than that of the server 20a, the number of processing modules of the client 30a is increased (S16). , Load balancing module 22a
The process of is finished (S18). For example, most of the processes A to E are operated by the client 30a, and the remaining very few processes are operated by the server 20a.

By appropriately distributing the processes A to E in accordance with the ratio of the high processing capacity of each computer by the process of step S16 or S17, the server 20a restarts the process of the server module 21 and the client 30
The process of the client module 31 is started by a (S19). The server 20a and the client 30a start execution of data processing including processing A to processing E in cooperation with each other.

As described above, in the distributed processing system of this embodiment, the server capacity data S1 held in the server 20a and the client capacity data C1 held in the client 30a are compared and evaluated. Therefore, in a system in which the system configuration does not change from the time of system design, it becomes possible to distribute the processing so that the load on each computer is optimized.

(Second Embodiment) FIG. 5 is a block diagram showing a functional configuration of a distributed processing system according to a second embodiment of the present invention. In the distributed processing system according to the present embodiment, the load distribution module is the server 2 as compared with the distributed processing system according to the first embodiment of the present invention shown in FIG.
The difference is that the load of each computer is distributed in consideration of the network processing capacity data N1 held in 0b.
Therefore, detailed description of overlapping configurations and functions will not be repeated. The reference numerals of the server, the client, and the load balancing module according to the present embodiment are 2 respectively.
This will be described as 0b, 30b, and 22b.

In FIG. 5, processes A to C are the server 2
0b, and processing D and processing E are executed by the client 30b, these modules, the server module 21 and the client module 31 perform processing while communicating with each other via the network 40. There is. In general, communication via the network 40 is more physical than communication between modules via the internal bus of a computer, such as communication between modules within the server 20b and communication between modules within the client 30b. The data transfer rate is very slow. Also, the communication procedure in the network,
For example, since the procedures for generating various protocols of the network, authentication, etc. are complicated, it is general that communication via the network is very slow.

For this reason, in a distributed processing system in which processing is distributed via a network, it is desirable to minimize the amount of data transferred in communication via the network. In order to reduce the amount of data transfer, data compression processing is performed before the data is transferred, the data is decompressed at the transfer destination, and commands are exchanged between modules.
However, since these processes need to be performed separately from the processes A to E in the server 20b and the client 30b, the processing load on each computer increases. If the network is fast enough, these processes are essentially unnecessary.

In data compression, the processing load due to compression generally increases as the compression rate increases. If these processes before the data transfer are highly performed, the amount of data flowing through the network is reduced and the communication can be performed at high speed, but the processing load on each computer is increased.

In the present embodiment, the data transfer rate in the network between the computers performing the processing is appropriately determined, and compared with the processing rate of the computer to determine how much the processing before the data transfer is performed.
That is, when the processing speed of the computer performing the processing is sufficiently high and the data transfer rate in the network is low, the processing such as compression of the data to be transferred is performed by the computer as much as possible to reduce the amount of data flowing through the network and transfer the data. Does not become a bottleneck. On the other hand, if the data transfer rate on the network is high and the processing speed of the computer is low, the conversion processing for data transfer such as data compression should be minimized to reduce the processing load on the computer. Try to transfer.

FIG. 6 is a flow chart for explaining the processing procedure of the distributed processing system according to the second embodiment of the present invention. When the system is activated by a user operation or batch processing, the server 20b first starts the processing of the server module 21 (S20). At this time, in order to distribute the processing load between the server 20b and the client 30b, the server 20b uses the load distribution module 2
The process of 2b is started (S21).

The load balancing module 22b is used by the server 20.
The server capacity data S1 of b is read (S22), the client capacity data C1 of the client 30b is read (S23), and these data are compared (S24). When the server capacity data S1 is larger than the client capacity data C1, that is, when the processing capacity of the server 20b is higher than that of the client 30b,
The number of processing modules of the server 20b is increased (S26).

If the server capacity data S1 is smaller than the client capacity data C1, that is, if the processing capacity of the client 30b is higher than that of the server 20b, the number of processing modules of the client 30b is increased (S25).

After the processes A to E are appropriately distributed in accordance with the ratio of the high processing capacities of the respective computers by the processing of step S25 or S26, the load distribution module 22b reads the network processing capacity data N1. (S27). The load balancing module 22b compares the network processing capacity data N1 with the server capacity data S1 and the client capacity data C1, and the server 20b.
The processing content such as data compression for data transfer is determined according to the network condition between the client and the client 30b (S28).

Then, the processing of the load distribution module 22b is terminated (S29), the server 20b restarts the processing of the server module 21, and the client 30b starts the processing of the client module 31 (S30). The server 20b and the client 30b cooperate with each other to start execution of data processing including processing A to processing E.

As described above, in the distributed processing system of this embodiment, the server capacity data S1
Since the transfer data processing method is set by comparing the client capacity data C1 and the network processing capacity data N1, distributed processing can be performed in consideration of the transfer speed of the network and the processing speed of each computer. The load on each computer and the network 40
It has become possible to determine the distribution of processing and the method of processing transferred data so that the data transfer amount in the above is optimal.

(Third Embodiment) FIG. 7 is a block diagram showing a functional configuration of a distributed processing system according to a third embodiment of the present invention. The distributed processing system according to the present embodiment is different from the distributed processing system according to the second embodiment of the present invention shown in FIG. 5 in that a server capacity evaluation module SD1 for measuring the processing capacity of the server is added to the server. The difference is that a client capability evaluation module CD1 for measuring the processing capability of the client is added to the client. Therefore, detailed description of overlapping configurations and functions will not be repeated. The reference numerals of the server, the load balancing module and the client in the present embodiment will be described as 20c, 22c and 30c, respectively.

The server capacity evaluation module SD1 measures the processing capacity of the server 20c and redetermines the value of the server capacity data S1 set at the time of system design. For example, the server 20c is caused to perform a predetermined process, and the processing time is measured to redetermine the value of the server capacity data S1 at the start of the distributed process.

Client ability evaluation module CD1
Measures the processing capacity of the client 30c and redetermines the value of the client capacity data C1 set at the time of system design. For example, the value of the client capability data C1 at the time of distributed processing is redetermined by causing the client 30c to perform a predetermined process and measuring the processing time.

In this way, the server capacity evaluation module SD1 and the client capacity evaluation module CD1
By executing before distributed processing, if the hardware configuration is changed from the system design time,
When the server 20c or the client 30c is used for a purpose different from the distributed processing and the processing capacity allocated to the distributed processing is lower than the initial design,
It is possible to grasp the actual processing capacity of each computer during distributed processing.

FIG. 8 is a flow chart for explaining the processing procedure of the distributed processing system according to the third embodiment of the present invention. When the system is activated by a user operation or batch processing, the server 20c first starts the processing of the server module 21 (S31). At this time, in order to distribute the processing load between the server 20c and the client 30c, the server 20c uses the load distribution module 2
The process of 2c is started (S32).

The load balancing module 22c executes the server capacity evaluation module SD1 to measure the current server capacity (S33), and then the re-determined server 2
The server capability data S1 of 0c is read (S34). Then, the load balancing module 22c executes the client capacity evaluation module CD1 to measure the current client capacity (S35), and then reads the re-determined client capacity data C1 of the client 30c (S36).

Next, the load distribution module 22c compares the server capacity data S1 with the client capacity data C1 (S37). When the server capacity data S1 is larger than the client capacity data C1, that is, the server 2
If the processing capacity of 0c is higher than that of the client 30c, the number of processing modules of the server 20c is increased (S39).

If the server capacity data S1 is smaller than the client capacity data C1, that is, if the processing capacity of the client 30c is higher than that of the server 20c, the number of processing modules of the client 30c is increased (S38).

After the processes A to E are appropriately distributed in accordance with the ratio of the high processing capacities of the computers by the processing of step S38 or S39, the load distribution module 22c reads the network processing capacity data N1. (S40). The load balancing module 22c compares the network processing capacity data N1 with the server capacity data S1 and the client capacity data C1, and the server 20c.
The processing content such as data compression for data transfer is determined according to the network status between the client and the client 30c (S41).

Then, the processing of the load distribution module 22c is terminated (S42), the server 20c restarts the processing of the server module 21, and the client 30c starts the processing of the client module 31 (S43). The server 20c and the client 30c cooperate with each other to start execution of data processing including processing A to processing E.

As described above, in the distributed processing system according to the present embodiment, the current server capacity data S1 and client capacity data C1 are measured at the start of distributed processing. When the hardware configuration is changed, the server 2
0c or client 30c is used for a purpose different from distributed processing, and the processing capacity allocated to distributed processing is lower than the initial design, etc., the actual processing capacity of each computer at the time of distributed processing is grasped. Can load each computer and network 40
It has become possible to determine the distribution of processing and the method of processing transferred data so that the data transfer amount in the above is optimal.

(Fourth Embodiment) FIG. 9 is a block diagram showing a functional configuration of a distributed processing system according to a fourth embodiment of the present invention. In the distributed processing system according to the present embodiment, network capacity evaluation modules SN1 and CN1 that measure the processing capacity of the network 40 are compared with the distributed processing system according to the third embodiment of the present invention shown in FIG. The difference is that it was added to. Therefore, detailed description of overlapping configurations and functions will not be repeated. The reference numerals of the server, the load balancing module and the client in the present embodiment will be described as 20d, 22d and 30d, respectively.

The network capacity evaluation modules SN1 and CN1 measure the processing capacity of the network 40 between the server 20d and the client 30d, and redetermine the value of the network processing capacity data N1 set at the time of system design. For example, transfer predetermined data between the server 20d and the client 30d,
By measuring the transfer time, the value of the network processing capacity data N1 at the time of distributed processing is redetermined.

FIG. 10 is a flow chart for explaining the processing procedure of the distributed processing system according to the fourth embodiment of the present invention. When the system is activated by a user operation or batch processing, the server 20d first starts the processing of the server module 21 (S50). At this time,
In order to distribute the processing load between the server 20d and the client 30d, the server 20d starts the processing of the load distribution module 22d (S51).

The load distribution module 22d executes the server capacity evaluation module SD1 to measure the current server capacity (S52), and then the re-determined server 2
The server capability data S1 of 0d is read (S53). Then, the load distribution module 22d executes the client capacity evaluation module CD1 to measure the current client capacity (S54), and then reads the re-determined client capacity data C1 of the client 30d (S55).

Next, the load distribution module 22d compares the server capacity data S1 with the client capacity data C1 (S56). When the server capacity data S1 is larger than the client capacity data C1, that is, the server 2
If the processing capacity of 0d is higher than that of the client 30d, the number of processing modules of the server 20d is increased (S58).

If the server capacity data S1 is smaller than the client capacity data C1, that is, if the processing capacity of the client 30d is higher than that of the server 20d, the number of processing modules of the client 30d is increased (S57).

After the processes A to E are appropriately distributed according to the ratio of the high processing capacities of the respective computers by the processing of step S57 or S58, the load distribution module 22d determines that the network processing capacity evaluation module SN1.
And CN1 is executed to measure the current network processing capacity (S59), and then the re-determined network processing capacity data N1 of the network 40 is read (S60).

The load distribution module 22d compares the network processing capacity data N1 with the server capacity data S1 and the client capacity data C1, and the server 20d.
The processing content such as data compression for data transfer is determined according to the network condition between the client and the client 30d (S61).

Then, the processing of the load distribution module 22d is terminated (S62), the server 20d restarts the processing of the server module 21, and the client 30d starts the processing of the client module 31 (S63). The server 20d and the client 30d cooperate with each other to start execution of data processing including processing A to processing E.

As described above, in the distributed processing system according to the present embodiment, the current network processing capacity data N1 is measured at the start of distributed processing.
The actual processing capacity of 0 can be grasped, and it becomes possible to decide the processing distribution and the transfer data processing method so that the load of each computer and the data transfer amount in the network 40 are optimized.

(Fifth Embodiment) FIG. 11 is a block diagram showing a functional configuration of a distributed processing system according to a fifth embodiment of the present invention. The distributed processing system according to the present embodiment is different from the distributed processing system according to the fourth embodiment of the present invention shown in FIG. 9 in that the load distribution module 22d is replaced with a dynamic load distribution module 22e. Therefore, detailed description of overlapping configurations and functions will not be repeated. In addition, the server in the present embodiment,
Reference numerals of the load balancing module and the client will be described as 20e, 22e, and 30e, respectively.

In the distributed processing system according to the fourth embodiment of the present invention, the server capacity evaluation module SD1
Since the client capacity evaluation module CD1 and the network processing capacity evaluation modules SN1 and CN1 measure the processing capacity of each computer and the processing capacity of the network 40 at the start of distributed processing, We were able to optimize load distribution. However, the longer the time required for data processing, the higher the possibility that the conditions of each computer and network that make up the distributed processing system will change.

For example, the server 20d and the client 3
If the processing capacity of the server 20d or the client 30d with respect to the target distributed processing is relatively reduced by causing 0d to perform other processing, or if the performance is reduced due to a failure of a part of the system, C, which is an element of the processing capacity of the computer
The processing speed may be improved by changing the PU, memory, hard disk, etc.

Further, even if the processing capability of the computer does not change, another computer not related to the distributed processing is connected to the network 40, and the data transfer rate via the network 40 is increased by increasing or decreasing the processing of the computer. It may also increase or decrease. In such a situation, the effect of initially distributing the load is reduced, the processing efficiency in each computer is reduced, and the processing that maximizes the efficiency of the entire system cannot be performed.

The dynamic load distribution module 22e in the present embodiment has steps S72 to S8 of FIG. 12 described later.
The processing of No. 1 is performed periodically, the status of the distributed processing system is grasped in a form close to real time, and the load is distributed according to it.

FIG. 12 is a flow chart for explaining the processing procedure of the distributed processing system in the fifth embodiment of the present invention. When the system is activated by a user operation or batch processing, the server 20e first starts the processing of the server module 21 (S70). At this time,
In order to distribute the processing load between the server 20e and the client 30e, the server 20e starts the processing of the dynamic load distribution module 22e (S71).

The dynamic load balancing module 22e executes the server capacity evaluation module SD1 to measure the current server capacity (S72), and then reads the re-determined server capacity data S1 of the server 20e (S72). S73).
Then, the dynamic load distribution module 22e executes the client capacity evaluation module CD1 to measure the current client capacity (S74), and then the re-determined client capacity data C1 of the client 30e.
Is read (S75).

Next, the dynamic load balancing module 22e
The server capability data S1 and the client capability data C1 are compared (S76). When the server capacity data S1 is larger than the client capacity data C1, that is, when the processing capacity of the server 20e is higher than the processing capacity of the client 30d, the number of processing modules of the server 20d is increased (S78).

If the server capacity data S1 is smaller than the client capacity data C1, that is, if the processing capacity of the client 30e is higher than that of the server 20e, the number of processing modules of the client 30e is increased (S77).

After the processing A to the processing E are appropriately distributed according to the ratio of the processing capacities of the respective computers by the processing of step S77 or S78, the dynamic load balancing module 22e causes the network processing capacity evaluation module S to perform processing.
After executing N1 and CN1 to measure the current network throughput (S79), the re-determined network throughput data N1 of the network 40 is read (S80).

The dynamic load distribution module 22e compares the network processing capacity data N1 with the server capacity data S1 and the client capacity data C1 and compares the data with the server 2
0e and the client 30e according to the network condition, the processing contents such as data compression for data transfer are determined (S81).

The server 20e is caused to restart the processing of the server module 21, and the client 30e is caused to start the processing of the client module 31 (S82). Server 20
The e and the client 30e start execution of data processing including processing A to processing E in cooperation with each other. After this, the dynamic load distribution module 22e, the step S72
The process from S81 to S81 is periodically repeated.

As described above, in the distributed processing system according to the present embodiment, the dynamic load distribution module 22e regularly measures and grasps the actual processing capacity of each computer and the network 40. It has become possible to determine the processing distribution and transfer data processing method so that the load on each computer and the data transfer amount in the network 40 are always optimized.

The embodiments disclosed this time must be considered as illustrative in all points and not restrictive. The scope of the present invention is shown not by the above description but by the claims, and is intended to include meanings equivalent to the claims and all modifications within the scope.

[0092]

According to one aspect of the present invention, the load balancing means compares the data relating to the processing capability of the server device with the data relating to the processing capability of the client device to balance the load of the server device and the client device. Therefore, the load can be distributed so that the processing efficiency of the server device and the client device is improved.

Further, since the load balancing means determines the load balancing of the server device and the client device based on the processing capacity of the server device at the time of load balancing measured by the first measuring means, at the time of load balancing It became possible to appropriately distribute the load according to the processing capacity of the server device.

Further, since the load distribution means determines the distribution of the load of the server device and the client device based on the data regarding the processing capacity of the network, the load is increased so that the processing efficiency of the server device, the client device and the network is improved. It has become possible to disperse.

Further, since the load balancing means determines the load balancing of the server device and the client device based on the processing capacity of the network at the time of load balancing measured by the second measuring means, the network at the time of load balancing It became possible to appropriately distribute the load according to the transfer speed of.

Further, since the load distribution means periodically determines the distribution of the load on the server device and the client device, it is possible to distribute the load so that the load on the server device and the client device is always optimum. became.

[Brief description of drawings]

FIG. 1 is a diagram showing an example of external appearances of a server and a client included in a distributed processing system according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration example of a server and a client according to the first embodiment of the present invention.

FIG. 3 is a block diagram showing a functional configuration of the distributed processing system according to the first embodiment of the present invention.

FIG. 4 is a flowchart for explaining a processing procedure of the distributed processing system according to the first embodiment of the present invention.

FIG. 5 is a block diagram showing a functional configuration of a distributed processing system according to a second embodiment of the present invention.

FIG. 6 is a flowchart for explaining a processing procedure of the distributed processing system according to the second embodiment of the present invention.

FIG. 7 is a block diagram showing a functional configuration of a distributed processing system according to a third embodiment of the present invention.

FIG. 8 is a flowchart for explaining a processing procedure of the distributed processing system according to the third embodiment of the present invention.

FIG. 9 is a block diagram showing a functional configuration of a distributed processing system according to a fourth embodiment of the present invention.

FIG. 10 is a flowchart for explaining a processing procedure of the distributed processing system according to the fourth embodiment of the present invention.

FIG. 11 is a block diagram showing a functional configuration of a distributed processing system according to a fifth embodiment of the present invention.

FIG. 12 is a flowchart for explaining a processing procedure of the distributed processing system according to the fifth embodiment of the present invention.

[Explanation of symbols]

1 computer main body, 2 display device, 3 F
D drive, 4 FD, 5 keyboard, 6 mouse,
7 CD-ROM device, 8 CD-ROM, 9 network communication device, 10 CPU, 11 ROM, 12 R
AM, 13 hard disks, 20a-20e servers,
21 server module, 22a-22d load balancing module, 22e dynamic load balancing module, 30a-
30e client, 31 client module, CD1 client capability evaluation module, SD1
Server capacity evaluation module, CN1, SN1 Network processing capacity evaluation module.

Claims (9)

[Claims] 1. A server device to which a client device is connected and which performs processing while distributing a load, the acquisition device for acquiring data regarding the processing capability of the client device, which the client device holds, and the server. A load balancing unit for determining load balancing of the server device and the client device by comparing data relating to the processing capability of the device with data relating to the processing capability of the client device acquired by the acquiring unit. , Server equipment. 2. The server device further includes first measuring means for measuring a processing capacity of the server device during load balancing, and the load balancing means measures the load measured by the first measuring means. 2. The load distribution of the server device and the client device is determined by comparing the processing capacity of the server device at the time of distribution with the data about the processing capacity of the client device acquired by the acquisition unit. Server device. 3. The server according to claim 3, wherein the load distribution unit compares the data regarding the processing capability of the server device and the data regarding the processing capability of the client device acquired by the acquisition unit with the data regarding the processing capability of the network. 3. The server device according to claim 1, wherein load distribution of the device and the client device is determined. 4. The server device further includes a second measuring unit for measuring the processing capacity of the network at the time of load distribution, and the load distribution unit includes the data concerning the processing capacity of the server device and the acquisition unit. The data concerning the processing capacity of the client device acquired by the above is compared with the processing capacity of the network at the time of load distribution measured by the second measuring means to determine the distribution of the load on the server device and the client device. Claim 1
The server device according to any one of 1 to 3. 5. The server device according to claim 1, wherein the load distribution unit periodically determines load distribution of the server device and the client device. 6. A distributed processing method in a server device and a client device connected by a network, the distributed processing method comprising: acquiring data regarding processing capacity of the client device held by the client device; Acquiring the data on the processing capacity of the server apparatus held by the apparatus, comparing the acquired data on the processing capacity of the server apparatus, and the acquired data on the processing capacity of the client apparatus to obtain the server apparatus And a step of determining load distribution of the client device. 7. A computer program for causing a computer to execute a distributed processing method in a server device and a client device connected by a network, wherein the distributed processing method has a processing capability of the client device held by the client device. The data regarding the processing capacity of the server apparatus held by the server apparatus, the data regarding the acquired processing ability of the server apparatus, and the acquired processing ability of the client apparatus Comparing the data with data to determine a load distribution of the server device and the client device. 8. A computer-readable recording medium recording a program for causing a computer to execute a distributed processing method in a server device and a client device connected by a network, wherein the distributed processing method is the client device. Acquiring the data regarding the processing capacity of the client device held by the server, acquiring the data regarding the processing capacity of the server device held by the server device, the acquired data regarding the processing capacity of the server device, and the acquisition Computer-readable recording medium, the step of determining load distribution of the server device and the client device by comparing the stored data regarding the processing capability of the client device. 9. A distributed processing system in which a server device and a client device are connected by a network and performs processing while distributing the load, wherein the client device holds data regarding the processing capacity of the client device. And an acquisition unit for acquiring data regarding the processing capacity of the client device held by the holding unit, data regarding the processing capacity of the server device, and the acquisition unit. A distributed processing system comprising: load balancing means for determining load balancing of the server device and the client device by comparing data regarding processing capacity of the client device.