JP2000194670A - Client and distributed server system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a flexible client/distributed server system which can simplify a network setting job. SOLUTION: A client process including a GUI control part G1, an object loader L1 and a network address A1 operates on a client machine C1. Meanwhile, a server process including the management objects A and B, an object loader L2 and a network address A2 operates on a server machine S1. Then an acquisition request is created, the client process generates a processing object 01. The object 01 is instanced by the loader L2, etc., and based on the address A1, etc., and also moves successively on the machine C1, etc., to operate as a thread. Then the object 01 moves inversely on the machine S1 to return to the machine C1 after the management objects A and F are acquired.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a client / distributed server system, and more particularly to a technique relating to acquisition of server management information or a processing request to a server.

[0002]

2. Description of the Related Art Conventionally, when constructing a client / distributed server system, either a method in which a client has information on all servers and communicates with each to perform information processing (FIG. 8), One of the systems (FIG. 9) in which the server is a master server, the client communicates only with the master server, and the master server communicates with another server (slave server) to perform processing.

That is, in the former method, in a client / distribution server system system constructed on a network environment as shown in FIG. 7, for example, a client machine C1 is divided into a plurality of server machines S1, S2 distributed on a network. In order to access and obtain information, etc., as shown in FIG. 8, all the client machines directly communicate with all the server machines, such as the client machine C1 accessing the server machines S1 and S2. Things.

In the latter method, as shown in FIG. 9, for example, a client machine C1 accesses a specific server S1 operating as a master server, and the server machine S1 accesses another server S2 operating as a slave server. S3 communicates with the client machine C1.
It provides all the information required by the above process.

[0005]

However, in the conventional system shown in FIG. 8, all the client machines C1, C2,
Since C3 needs to communicate directly with all server machines S1, S2, S3, all routers R1, R2
For 2, it is necessary to make settings for all server machines S1, S2, and S3 to perform communication. Further, all server machines S are connected to all client machines C1, C2 and C3.
It is necessary to give information such as the network address of 1, S2, and S3, and there is a problem that network setting is complicated.

In the conventional system shown in FIG. 9, the server machine S1 operating as the master server needs to mount special processing means for communicating with all the other server machines S2 and S3. There is a problem that the load of S1 increases as compared with that of the server machines S2 and S3, and when a failure occurs in the server machine S1, there is a risk that the entire system may be affected.

Further, the server machine S1 and the server machine S
There is also a problem that a difference occurs in processing between S2 and S3, and the types of software required for the system increase.

Accordingly, it is an object of the present invention to simplify a network setting and to provide a flexible client / client machine without requiring a specific server machine to perform special processing.
It is to provide a distributed server system.

[0009]

According to a first aspect of the present invention, there is provided a client / distribution server system in which a plurality of client machines and server machines are distributed via a network. A processing object with a processing request to a server machine that has been autonomously moved from one distributed server machine to another to perform processing, so that communication between the client machine and the server machine is limited to only between adjacent machines. It is characterized by having done. In the client / distribution server system according to the second aspect of the present invention, in a client / distribution server system in which a plurality of client machines and server machines are distributed via a network, a GUI control for managing a GUI on the client machines is provided. A client process having a unit, an object loader that receives data, receives a data transfer request from another machine and instantiates the processing object by receiving data, and a network address of a server machine that performs communication. On the server machine, a management object having information managed by the client / distributed server system, and an object loader which receives data and receives a data and instantiates the processing object when receiving a transfer request of the processing object from another machine. When operating a server process and a network address of the next server machine acquisition request of the management object by the operation of the GUI control unit occurs,
The client process generates a processing object having the function of acquiring the management object, and the processing object operates as a thread by sequentially moving server machines according to the network address and while being instantiated by the object loader. ,
When the management object is obtained, the server machine moves in the reverse order and returns to the client machine. Further, a preferred embodiment of the present invention
The processing object serializes itself, transfers it to the server machine as byte data, closes the port, deletes itself, and deletes itself.
It is characterized by decoding the received byte data and instantiating the processing object. In short, the present invention moves in a client / distributed server type system by acquiring management information of a plurality of server machines or transferring an object having a function of making a processing request to the server machines to the server machines. By adopting a processing method, it is easy to build a system that adopts a multi-client / multi-server configuration on a network with limited network access, enabling a flexible system to be built. .

[0010]

Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 schematically shows a client / distribution server system of the present invention. In the present invention,
The processing object O1 generated by the client machine C1 moves to the server machine S1 to perform the processing requested by the client machine C1, and then moves to the server machine S2 to perform the processing. By performing the process by moving between the machines, the communication between the client machine and the server machine is performed only between the adjacent machines, and the setting of the router can be simplified.

When the server machine S4 is added to the system, it is necessary to add information of the added server machine S4 to all routers and all client machines in the conventional method of FIG. Occurs, and
In the conventional method shown in FIG. 9, it is necessary to set the server machine S1 operating as the master server and the router R2, but in this method, as shown in FIG.
3 only need to be set, and all client machines and all routers do not need to know information about the server machine S4.

[0012] In this manner, by processing objects moving autonomously distributed between servers and performing information collection processing, communication between the client and the server and between the server and the server can be simplified in the client / distributed server system. Can be As a result, the network settings of the system can be simplified, and changes in the case of system expansion can be reduced.

An example in which a processing request is sent from the client machine C1 to the servers S1, S2, S3 in the network configuration of FIG. 1 will be specifically described with reference to FIG.
In FIG. 2, a client process operates on the client machine C1, and this process has the following processing and information. A GUI control unit G1 having a function of managing a GUI (Graphic User Interface); receiving a transfer request of a processing object from another machine;
Object loader L1 having a function of receiving data and instantiating (classifying) a processing object ・ Network address A1 of a server machine with which client machine C1 communicates

On the server machine S1, a server process operates, and this process has the following processing and information. A managed object A and a managed object B having information managed by the client / distributed server system; an object loader L2 having a function of receiving a transfer request of a processing object from another machine, receiving data, and instantiating the processing object.・ Network address A2 of the next server machine

Similarly, the server processes on the server machine S2 are managed objects C, managed objects D,
The server process on the server machine S3 has an object loader L3 and a network address A3, and has a management object E, a management object F, an object loader L4, and a network address A4.

Next, the operation of this embodiment will be described in detail with reference to FIGS. 1. First, it is assumed that an operation of the GUI control unit G1 of the client machine C1 in FIG. 2 causes a request to acquire the managed objects A and F held by a plurality of server machines. 2. Then, the client process of the client machine C1 generates a processing object O1 having a function of acquiring a management object (processing P1 in FIG. 2), and adds the management object A and the management object F to the processing object O1.
And the own network address A1 as the address of the activation source machine. The processing object O1 holds the given condition (A, F) and the network address A1 of the client machine C1 in its own attribute. 3. The processing object O1 operates as a thread in the process of the client machine C1 and moves to the server machine S1 (processing P2). The processing at this time will be described with reference to the flowcharts of FIGS. . 4. The processing object O1 determines whether the machine on which the processing object O1 is operating is the activation source machine based on the network address (processing F1 in FIG. 4). 5. Since the processing object O1 is operating on the activation source machine (client machine C1), it is determined whether the management object acquisition processing has been completed (processing F).
2). 6. Since the acquisition processing of the management objects A and F has not been completed yet, the transfer processing to the next server is started. 7. Currently running machine (client machine C)
From the process of 1), the address of the destination server machine
The server machine S1 is acquired, stored in its own attribute (process F10), the port of the server machine S1 is opened (process F12), and the process shifts to the process of the transfer destination machine. 8. As shown in FIG. 5, in the server machine S1, the object loader L2 opens a port (processing F1 in FIG. 5).
6) Monitor. Since there is a connection request from the processing object O1 on the client machine C1, connection is made (processing F17). 9. Processing The object O1 serializes (serializes) itself and transfers it to the server machine S1 as byte data (processing F13). Object loader L2
Receives the transferred byte data (process F1).
8). 10. After the transfer, the processing object O1 closes the port (processing F14), deletes itself (processing F15), and ends the processing. 11. After the connection is disconnected (process F19), the object loader L2 of the server machine S1 decodes the received byte data and instantiates the process object O1 (process F20). 12. Generated on server machine S1 (process P in FIG. 2)
3) The processing object O2 starts processing as a thread as shown in FIG. 13. Again, the server machine S1 is not the activation source machine (process F1), and the management objects A and F have not been obtained (process F5), so the server machine S1 starts the acquisition process. 14. The management object is searched for in the server machine S1, and the management object A that meets the condition is acquired (process F6). 15. The network address A2 of the next server machine is obtained (Process F7), and the network address of the server machine S2 is obtained (Process F9), so the destination is determined by the server machine S2 (Process F10). . 16. Move to the server machine S2 in the same manner as when moving from the client machine C1 to the server machine S1 (process P5 in FIG. 2). 17. Server machines S2 and S3 in the same manner as 12 to 16
Above, repeat the acquisition process and movement of the management object,
The management object F is acquired by the server machine S3 and stored in its own attribute (processing P6, P7, P8, P
9). 18. When the processing object O4 acquires the management object F on the server machine S3 (processing F6), the processing condition is satisfied (processing F7), and the processing object O4 moves to the activation source machine. 19. In order to reverse the network address to which the processing object has moved, the network address of the source machine is set as the transfer destination (processing F11), and the processing object is moved to the server machine S2 (processing P10 in FIG. 3). 20. Execute the same processing as in 7 to 12, and move to the client machine C1 (processing P11, P12, P1 in FIG. 3).
3, P14). However, since the management object acquisition processing has been completed, the management object acquisition processing is not performed in the server machine, and only the movement is performed (processing F5). 21. The process object O8 instantiated in the client machine C1 (process P15 in FIG. 5) determines from the network address that the machine executing the process is the same as the activation source machine (process F1), Also,
Since the process of acquiring the managed objects A and F has been completed, the acquired managed objects A and F are delivered to the client process (process P16). 22. Thereafter, the processing object O8 deletes itself and ends the processing. 23. In this way, the processing object O1 generated on the client machine C1 moves between the server machines, executes the request of the client machine C1, acquires the management objects A and F, and sends information to the client machine C1. It provides.

Next, an application example of the present invention will be described.
Of the software written in Java, an object-oriented programming language developed by American Microsystems, software that is embedded in a file, downloaded from a web server, and executed on a web browser is Java
It is called an applet.

In this Java applet, as shown in FIG. 6, when the client machine C1 downloads the Java applet from the server machine S1 (P1
7), Java applet A1 can communicate only with downloaded server machine S1 (P18, P19, P2
0). Therefore, when the server machines holding the management objects are distributed, a distributed server system using an applet as a client process cannot be constructed. Alternatively, as shown in FIG. 9, the process on the server machine which has downloaded the applet needs to implement a special processing means for accessing another server.

However, in the present invention, if the server machines S1 and S2 and S2 and S3 can communicate, the processing object moves between the servers (processing P21, P2).
22), it is possible to access the management objects of all servers. In the case where the operation of the client application is limited, such as a case where a Java applet is used as a client, the present invention is used to construct a distributed server environment without special implementation on a server machine. It becomes possible.

In the embodiment described above, the processing executed by each processing object in each server is acquisition of a management object, but the processing performed by the processing object is
The present invention is not limited to this, and may be, for example, DB access or the like, and there is no limitation on the processing content.

Further, in the above-described embodiment, the client process and the server process themselves have the object loader unit, but may be independent processes having a function of generating an object. In this case, the object loader process performs necessary processing by performing inter-process communication with another process running on the same server machine.

Further, in the above-described embodiment, the server is sequentially circulated by moving the code. However, after copying to another server, the server may continue processing without deleting the server itself. As a result, the processing object executes processing on all servers.
When performing time-consuming processing, processing time can be reduced by performing processing in parallel on all servers.
Alternatively, processing can be continued on all servers, and work such as failure monitoring of server processes can be performed.

Further, in the above embodiment, three servers are described. However, the number of servers may be any number n.
It may be.

[0024]

The first effect of the present invention is that the setting of the client machine and the server machine and the equipment constituting the network can be simplified. This point will be specifically described with reference to examples. In FIG. 1, the server machine with which the client machine C1 directly communicates is only S1, and the server machines S2 other than S1
Since there is no need to know the existence of S3, it is not necessary to set the network addresses of the server machines S2 and S3 in the client machine C1.
There is no need to set the network address of the client machine C1 in S3. Further, as shown in FIG. 4, a server machine S4 is added to the system,
4 also holds the management information, only the server machine S3 needs to set the network,
Network settings can be simplified as a system. Furthermore, if the client machine C1 can communicate only with the server machine S1, the client machine C1 can access information of all servers.
Even if it is not possible to directly communicate with the server machine S3, the processing can be performed without any problem.

The second effect is that the programs on all servers can use the same program.
This means that a distributed server environment can be constructed without imposing a special load on a specific server. It is the processing object created on that server that obtains information from all server processes, and a specific server process, which is positioned as a master server, has special information to communicate with other server machines. This is because there is no need to perform processing and settings.

[Brief description of the drawings]

FIG. 1 is a schematic diagram of a client / distribution server system of the present invention.

FIG. 2 is a diagram showing a state in which a processing object generated on a client machine moves on a server machine to acquire a management object in the client / distributed server system of the present invention.

FIG. 3 is a diagram showing a state in which a processing object that has acquired a target management object moves a server machine toward a client machine of an activation source in the client / distributed server system of the present invention.

FIG. 4 is an operation flowchart of a processing object starting machine according to the present invention.

FIG. 5 is an operation flowchart of a processing object transfer destination machine according to the present invention.

FIG. 6 is a diagram illustrating an application example of the present invention.

FIG. 7 is a configuration diagram of a general client / distribution server system.

FIG. 8 is a diagram showing an example of access from a client machine to a server machine in a conventional client / distributed server system.

FIG. 9 is a diagram showing another example of access from a client machine to a server machine in a conventional client / distributed server system.

[Explanation of symbols]

 A, B, C,. D, E, F Management objects A1, A2, A3, A4 Network address C1, C2, C3 Client machine G1 GUI control unit L1, L2, L3, L4 Object loader N1, N2. N3 network O1, O2, O3, O4 processing object R1, R2 router S1, S2, S3, S4 server machine

Claims (7)

[Claims] A client / server in which a plurality of client machines and server machines are distributed via a network.
In a distributed server system, communication between a client machine and a server machine is performed by a processing object accompanying a processing request to the server machine generated by the client machine autonomously moving the distributed server machines one after another to perform processing. Is a client / distribution server system characterized in that it is only between adjacent machines. 2. A client / server in which a plurality of client machines and server machines are distributed via a network.
In the distributed server system, a GUI for managing a GUI on the client machine
An I control unit, an object loader that receives data when receiving a transfer request for a processing object from another machine and instantiates the processing object, and a client process having a network address of a server machine that performs communication, and A management object having information managed by the client / distributed server system on the server machine, an object loader receiving data to receive a transfer request of a processing object from another machine and instantiating the processing object, When a server process having a network address of the next server machine is operated and a request to acquire the management object is generated by operating the GUI control unit, the client process executes a processing object having a function to acquire the management object. The processing object operates as a thread by sequentially moving server machines according to the network address and while being instantiated by the object loader, and operates as a thread. A client / distributed server machine system, which moves and returns to the client machine. 3. The processing object serializes itself, transfers the converted byte data to a server machine, closes a port, deletes itself, and the transfer destination object loader decodes the received byte data. 3. The client / distribution server system according to claim 1, wherein the processing object is instantiated by performing the processing. 4. The client according to claim 1, wherein the processing executed by the processing object on each server machine is a database access instead of acquiring the management object.・
Distributed server system. 5. The method according to claim 1, wherein the object loader is separated from the client process or the server process to be an independent process having a function of generating an object, and communicates with another process operating on the machine. 5. The client / distribution server system according to claim 2, wherein 6. The server machine according to claim 1, wherein, instead of traversing the server machine by moving the code of the processing object, the server machine does not delete itself after copying to another server machine, but performs a predetermined process in the server machine. A client / distribution server system according to any one of claims 1, 2, 4 and 5. 7. The client / distribution server system according to claim 1, wherein a Java applet is used as the client process.