JPH1141252A - Client server system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a client server system capable of distributing load without impairing flexibility, reliability and expandability. SOLUTION: Plural authentication servers 22a-22z are provided with busy server list tables 28a-28z composed of the number of servers under operation and a busy server list arranging the addresses of servers to be preferentially connected among the servers under operation according to priority, each time access is performed from gateways 24a-24z, the tables are transmitted and the priority is exchanged. Then, the access from the gateways 24a-24z to the respective authentication servers 22a-22z is performed according to the priority in these tables so that the load is equalized at the respective authentication servers 22a-22z.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a client-server system in which a plurality of server devices and a plurality of client devices are connected via a communication path.

[0002]

2. Description of the Related Art A typical example of a client / server system is a LAN system of a company or the like, or the Internet spread worldwide. When the scale of the system increases in these networks, there is a strong demand for increasing the processing capacity by installing a plurality of servers and ensuring reliability against device failure.

In a client-server system, a request such as a data search request is issued from a client device to a server device, and the server device performs processing such as data search according to the request, and returns a result to the client device. Perform the operation.

When requests from a large number of client devices concentrate on one server device, a bottleneck occurs in the processor processing capability of the server device or an interface portion with a communication path, and the processing is delayed or a timeout occurs. There was a problem that normal processing became impossible.

Conventionally, as a system for solving the above problem, as shown in FIG. 1, a plurality of client devices 1a,
1b,..., 1z, a plurality of, here two, server devices 3a, 3b having the same function are installed (here, server devices other than the original server device are called mirror servers). There has been a system in which which server device is used is left to the user of the client device.

In this system, for example, when the user of the client device 1a issues a request to one of the server devices 3a (4), the server device 3a is congested and a timeout occurs (5). The user 1a issues a request to the other server device 3b again (6).

[0007]

However, in the case of the above system, since the user of the client device selects the server device arbitrarily, it is difficult to equalize the load on each server device. However, there is a problem that it cannot be used in a client device that operates automatically without human operation.

As another system for equalizing the load, as shown in FIG. 2, among a plurality of client devices 12a, 12b,... 12z accommodated in a network 11, client devices 12a to 12m are server devices. 1
There is a system in which a parent server device is uniquely determined for each client device so that the client devices 12n to 12z access the server device 13b so that the client device 12n to 12z accesses the server device 13b so that the load is not concentrated.

In this system, the load of each server device can be equalized, but the flexibility of the combination of the server device and the client device is poor, and if one server device fails, the service to all the subordinate client devices can be provided. Has a reliability issue, such as an outage,
Further, when the number of client devices increases sequentially, there is a disadvantage that the configuration is not always optimal.

[0010] It is an object of the present invention to provide the disadvantages of the conventional client / server system, such as the problem of poor system expandability, the problem that the loads of a plurality of server devices are not equalized, and the reliability of the server device against failure. It is an object of the present invention to solve the problem of lack of security and to realize a flexible and highly reliable client-server system.

[0011]

In order to solve the above-mentioned problems, according to the present invention, a plurality of server apparatuses having the same function, a plurality of client apparatuses, and all of the server apparatuses and the client apparatuses are connected to each other to connect each other. In a client-server system comprising a network enabling communication and each server device and client device having a unique address for identifying each server device, the server device operates based on response data to a request from the client device. Among the plurality of server devices having the same function, information indicating a server device to be preferentially connected is added and transmitted, and when a connection request to the server device is generated, each client device And determining a server device to be accessed.

According to the above configuration, each time the client device accesses the server device, the client device can receive information on the server device to be preferentially connected, and each client device receives this information when the next connection request occurs. Based on this, it is possible to equalize the load on each server device.

Each server holds information as to whether or not all the servers contained in the network are in operation, and if a change occurs in the state of the server, the servers exchange with each other. By updating the information by performing communication, all the server devices share the same information, and a list in which the total number of operating server devices and the addresses of the server devices to be connected with priority are arranged in accordance with the priority order. This table is stored as a table, and this table is added to the response data to the request from the client device and transmitted. Each time the transmission is performed, the address in the table is assigned the highest priority to the lowest, and The objects are characterized in that they are exchanged so that the rank rises one by one.

Each client device holds a table received from the server device, and when a connection request to the server device is generated next, in accordance with the priority shown in the table, the client devices sequentially start with the server device with the highest priority. It is characterized in that access is repeated until connection to a server device with a low priority is successful or all server devices are exhausted.

The most different point from the conventional technique is that the servers and the server and the client cooperate to distribute the load.

According to the above configuration, the server device transmits the addresses of all available server devices to the client device together with information indicating the priority in which the connection should be made preferentially. In addition, the client device accesses available servers in order of priority based on the information, and the information is updated for each access, so that the load on each server device is equalized.

Further, when the number of server devices increases due to an increase in the scale of the system, the increased information of the server device is reflected in the information of the operating server, so that the system may be stopped or the client device may receive the information. It is possible to increase the number of server devices without changing the access procedure.

When a server device breaks down or is to be temporarily stopped for maintenance, the information of the failed or stopped server device is reflected in the information of the operating server. Does not have to wait for a long time before accessing a failed or stopped server for a timeout. Even in the event that a faulty server device is accessed by passing, the client device is given the address of the next server device in the case where access failed due to a response from the previous server device, A retry process can be executed, and a highly reliable system can be constructed even for a failure.

[0019]

FIG. 3 shows an example of an embodiment of a server / client system according to the present invention, in which the present invention is applied to an Internet telephone / FAX system. In FIG. 3, reference numeral 21 denotes a network, here, the Internet; 22a, 22b,..., 22z, a plurality of authentication servers (AS) accommodated in the Internet 21;
a, 23b,... 23z are telephone networks (PSTN), 24
24z are gateways (GW) having a signal conversion function between the Internet 21 and the telephone networks 23a to 23z, 25a, 25b,... 25z are a plurality of telephones accommodated in the telephone network, 26a, 26b, 26z
Are a plurality of faxes accommodated in the telephone network.

Each of the authentication servers 22a to 22z respectively
Subscriber files 27a, 27b,... 27z containing at least Internet telephone / fax subscriber information;
Active server list tables 28a, 28b,..., 28z
And

FIG. 4 shows the configuration of the active server list table. As shown in FIG. 4A, the number of active servers 281 representing the total number of active server devices and the number of active server devices are shown. A list of operating servers 282 is arranged in accordance with the order of priority of the server devices (IP addresses and the like) to be connected preferentially (however, in the drawing, the code is used instead). In the server list 282, the upper part has a higher priority, and the lower part has a lower priority. In FIG. 4, four authentication servers 22a, 22
2b, 22c, and 22d are operating, and the other servers are not operating.

In this embodiment, the gateways 24a to 24z correspond to client devices. Also, the gateway 24a
-24z, telephones 25a-25z and FAX 26a-2
6z can be connected via telephone networks 23a to 23z.

FIG. 5 shows an outline of a processing flow in the server device, and FIG. 6 shows an outline of a processing flow in the client device.

Next, the operation will be described.

It is assumed that the GW 24a holds the AS 22b as a server device to be accessed in the initial state. When a call is made from the telephone set 25a and connected to the GW 24a, the GW 24a performs AS2 to authenticate the subscriber.
Inquire 2b. In the AS 22b, collation with the subscriber file 27b is performed, and the result is returned to the GW 24a. At this time, the AS 22b sets the return signal (response data) to the active server list table 28 as shown in FIG.
b is given.

When the response to one inquiry (request) is completed, the AS 22b changes the priority order of the active server list 282 in the active server list table 28b. Specifically, as shown in FIG. 4B, the AS 22a with the highest priority (top) is dropped to the lowest (bottom), and the ASs 22b, 22c, and 22d with second or lower priority are ranked.
Are changed so that the ranking of each is raised one by one. This AS2
When the next inquiry is made to 2b, the list table shown in FIG. 4B is transmitted, and then the same priority order is replaced as described above.

On the other hand, the GW receiving the response from the AS 22b
24a starts the operation based on the authentication result at that time (executes the process) and holds the received active server list table (or receives the active server list table held by itself at that time). Rewrite to the running server list table.)

Thereafter, the GW 24a is connected to any of the ASs.
When it is necessary to make an inquiry to, access is performed according to the priority shown in this table. Specifically, first, the AS having the highest priority, that is, 22a is accessed, and if the AS fails, the AS having the second priority, that is, 2A is accessed.
2b, and if it fails, the third place AS, that is, 22c,..., Etc., the order is moved down by one, and the access is repeated. If the access succeeds, the selection of the AS is completed, and the process proceeds to execution. If access to all servers in the list table has failed, the process is considered to be impossible to continue and abnormal processing is started.

The GW 24a is connected to one of the AS2
Since a new active server list table is received when the access to 2a to 22d is successful, the new list table is used for the next access.

In this manner, accesses from the GWs 24a to 24z to the ASs 22a to 22z are distributed.

FIG. 7 shows how the list table changes when the AS 22c fails. When detecting that the active AS 22c has failed by communication between the ASs, the respective ASs maintain their own running server list table as shown in FIG. 7A, as shown in FIG. 7B. Rewrite as follows. Specifically, the address corresponding to the AS 22c is deleted from the priority order 282 in the active server list table, and the number of active servers 281 is reduced by one.

In this way, the inefficient operation in which each GW accesses the failed AS many times and fails can be reduced. Also, when the AS is artificially stopped by replacing software or the like, the operation is the same as that in FIG.

FIG. 8 shows how the list table changes when the number of operating ASs increases. According to the communication between each AS and the instruction from the maintenance function, 2
Upon detecting that 2e has been added, each AS rewrites the running server list table as shown in FIG. 8A, which is held by itself, as shown in FIG. 8B. Specifically, the priority order 282 in the active server list table
The address corresponding to the AS 22e is added to the lowest order of, and the number of active servers 281 is increased by one.

By doing so, the access from each GW is equally applied to the added AS.

[0035]

As described above, according to the present invention,
The loads on a plurality of server devices having the same function can be equalized.

When a server is added in accordance with an increase in the scale of the system, information on the added server is reflected in the information on the operating server. It is possible to add server devices without changing the procedure.

When a certain server device breaks down or is temporarily stopped for maintenance, the information on the active server reflects the information on the failed or stopped server device. Alternatively, it is not necessary to access the stopped server and wait for a long time until a timeout occurs. Even in the event that a faulty server device is accessed by passing, the client device is given the address of the next server device in the case where access failed due to a response from the previous server device, A retry process can be executed, and a highly reliable system can be constructed even for a failure.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing an example of a conventional server / client system.

FIG. 2 is a configuration diagram showing another example of a conventional server / client system.

FIG. 3 is a configuration diagram showing an example of an embodiment of a server / client system of the present invention.

FIG. 4 is a configuration diagram showing an example of a running server list table;

FIG. 5 is a flowchart showing an outline of processing in a server device;

FIG. 6 is a flowchart showing an outline of processing in a client device.

FIG. 7 is an explanatory diagram of a change in a running server list table when a running server device breaks down.

FIG. 8 is an explanatory diagram of a change in a running server list table when the number of running server devices increases;

[Explanation of symbols]

21: Internet, 22a to 22z: Authentication server (AS), 23a to 23z: Telephone network (PSTN), 24
a to 24z ... gateway (GW), 25a to 25z ...
Telephones, 26a to 26z FAX, 27a to 27z subscriber files, 28a to 28z table of active servers.

Claims (3)

[Claims] 1. A plurality of server devices having the same function,
A plurality of client devices and a network connecting all of the server devices and the client devices to enable mutual communication, wherein each server device and the client device have a unique address for identifying each; In the server system, each server device transmits, to response data to a request from a client device, information indicating a server device to be preferentially connected among a plurality of active server devices having the same function. Each client device determines a server device to be accessed based on the information when a connection request to the server device is generated. 2. Each server device holds information on whether or not all the server devices accommodated in the network are operating, and when a change occurs in the state of the server device, the server devices communicate with each other. The server updates the information by communicating with the server, and all the server devices share the same information. A list in which the total number of operating server devices and the addresses of the server devices to be connected with priority are arranged in order of priority. This table is stored as a table, and this table is added to the response data to the request from the client device and transmitted. Each time the transmission is performed, the address in the table is assigned the highest priority to the lowest, and 2. The client-server system according to claim 1, wherein the items are exchanged so that the order increases one by one. 3. Each client device retains a table received from the server device, and when a connection request to the server device is generated next, according to the priority shown in the table, the client devices sequentially start with the server device with the highest priority. 3. The client-server system according to claim 2, wherein the access is repeated until the connection to the low-priority server device succeeds or all the server devices are exhausted.