JPH09319689A - Server selecting system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suitably distribute load between servers and to effectively avoid the fault when the function of any server is stopped by the fault. SOLUTION: In the computer network environment of a client/server system for which a client 200 and servers (1)300, (2)400 and (3)500 are connected to a computer network 600, a name server 100 is provided for dispatching an address list designating the order of servers to be selected to the client 200. In this case, when a client program 24 to be operated on the client 200 is to perform access to any server for providing the same service, the address list designating the order of accessable servers is possessed from the name server 100 by a dynamic link library 23 for address solution for inquiring an address and according to the order of this address list, probe packets are sent out to the respective servers so that the operating conditions of servers can be confirmed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for selecting a server computer (hereinafter also simply referred to as "server") by a program on a client computer (hereinafter simply referred to as "client"), and particularly to a server using a name server. Regarding selection method.

[0002]

2. Description of the Related Art In a computer network environment in which a client computer and a server computer are connected to a network, it is widely practiced to install a plurality of servers that provide the same service for the purpose of load balancing, troubleshooting, etc. .

Under such circumstances, the service requests of the client to the server are often concentrated on a small number of servers, and the responsiveness and the performance to the client using the heavily loaded server may deteriorate.

For this reason, it is necessary for the client and its program to transparently evenly distribute the user load of the client among the servers.

When a server fails and the service from that server stops, it is necessary to transparently select another normal server for the client and its program.

Conventionally, as a server selection method in consideration of such a situation, for example, a simple method of statically evenly distributing clients to each server in advance, or a more practical technique is disclosed in Japanese Patent Laid-Open No. 7-242242. -9323
As described in Japanese Patent No. 8, when the client requests the name server to acquire the Internet address of the server, the name server selects the address of the server having a low load in consideration of the load of the server and instructs the client to do so. A method of notifying and dynamically distributing the load has been proposed.

[0007]

However, although the above-mentioned conventional server selection method can realize the function of load balancing, no consideration is given to the case where the instructed server stops functioning due to a failure. In such a case, there is a problem that the client cannot receive the service from the server.

Since the address of the server is normally stored in the memory of the client, even after the name server recognizes that the server has stopped functioning and updates the address list, this client receives the service from the server thereafter. I can't.

Therefore, the present invention has been made in view of the above-mentioned problems, and it is possible to appropriately perform load distribution among servers and to effectively avoid a failure even when the server stops functioning due to a failure. The purpose is to provide a server selection method that can be performed.

[0010]

In order to achieve the above-mentioned object, the present invention provides a client with an address list designating the order of servers to be selected in a server selection system in a client / server system in a computer network environment. The client has a name server to be passed, and on the client side, an address resolution dynamic link library is used to send out probe packets to each server according to the order of the address list passed from the name server to check the operating status of the server. A server selection method characterized by the above is provided.

Further, the present invention, in the server selection system in the computer network environment by the client / server system, is provided with a name server which includes the metric in the address list of the server to be selected and passes it to the client. An address resolution dynamic link library for inquiring determines the order of servers to be selected based on the metrics included in the address list passed from the name server, and sends a probe packet to each server according to the determined order. A server selection method is provided which is characterized by confirming the operating status of the server.

As described above, according to the server selection method of the present invention, the order of the address list passed from the name server or the address list is included in the address list by the dynamic link library for address resolution for inquiring the address on the client side. By attempting to connect to each server in order according to the order determined based on the metric, even if the server selected by the name server is down, it is possible to connect to the next best server and take countermeasures against the failure.

According to the server selection method of the present invention,
Even if the program on the client is not modified, it can be dealt with only by replacing the address resolution dynamic link library for obtaining the address corresponding to the name, and the transparency to the program on the client can be maintained.

[0014]

Next, embodiments of the present invention will be described with reference to the drawings.

Referring to FIG. 1, the present invention, in an embodiment thereof, is a client / server in which a client computer 200 and server computers (1) 300, (2) 400 and (3) 500 are connected to a computer network 600. Server computer (1) 300, (2) 400 and (3) 5 to be selected in the computer network environment according to the method
A name server computer 10 that passes an address list specifying the order of 00 to the client computer 200
0 and the client program 24 operating on the client computer 200 tries to access any of the server computers (1) 300, (2) 400 and (3) 500 which provide the same service. , An address list that specifies the order of accessible server computers from the name server computer 100 is acquired by the address resolution dynamic link library 23, and a probe packet is sent to each server computer according to the order of the address list. By checking the operating status of the server computer, only the address of the operating server computer can be obtained by the client program 24.
I will return it to.

[0016]

Next, in order to more specifically describe the above-described embodiment of the present invention, an embodiment of the present invention will be described in detail with reference to the drawings.

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

Referring to FIG. 1, in one embodiment of the present invention, a computer network 600 includes a name server computer 100, a client computer 200, and three server computers (1) 300.
A server group consisting of (2) 400 and (3) 500,
Is connected. Although only three server computers and one client computer are shown in FIG. 1 for convenience of illustration, it is not limited to this, and it is needless to say that a plurality of server computers and client computers can be connected.

The name server computer 100 includes an input device 11 and an output device 12 for transmitting / receiving data via a computer network 600, a decision logic 13 for determining a priority order of a plurality of server computers according to a certain standard, and a server name. To a corresponding address list, a writing means 15 for updating the data in the storage device 17, a reading means 16 for reading the data from the storage device 17, and a list of pairs of server names and addresses. And a storage device 17 that holds the.

The client computer 200 includes an input device 21 and an output device 22 for transmitting / receiving data via the computer network 600, an address resolution dynamic link library 23, and a client program 24.

Next, the operation of the embodiment of the present invention will be described. FIG. 2 is a flowchart showing the flow of processing by the address resolution dynamic link library 23 shown in FIG.

Referring to FIGS. 1 and 2, the client program 24 operating on the client computer 200 provides a group of servers (server computers (1) 300, (2) 400 and (3) 500) which provide a uniform service. When trying to access the server, in order to obtain the correct address of the server computer, the dynamic link library for address resolution 2
Pass the server name to 3 and request processing to return the correct address.

At this time, the client computer 20
Dynamic link library 23 for address resolution of 0
Receives the server name from the client program 24 (step S1), and inquires about the address from the name server computer 100 via the output device 22 (step S2). The communication protocol at this time is known to those skilled in the art.

The input device 11 of the name server computer 100 receives the server name transmitted from the client computer 200 and passes it to the conversion mechanism 14. A list of pairs of server names and addresses is held in the storage device 17 of the name server computer 100. When the conversion is started, the conversion mechanism 14 creates an address list of data that matches the received server name. Reading means 16
Is read from the storage device 17, and the read address list is returned to the client computer 200 via the output device 12.

The address resolution dynamic link library 23 of the client computer 200 receives the address list via the input device 21 (step S3). According to the conventional technique, this address list is passed to the client program 24 as it is, but in this embodiment, it is given before that in order to confirm whether or not the server computer is actually performing the service. The probe packet is transmitted to the address indicated at the beginning of the address list (step S4). The probe packet contains the ECH of ICMP, which is in the communication protocol.
It is possible to use an O packet or a protocol defined in advance between the server and the client.

When a response is returned from the server computer of the destination of the probe packet for a certain period of time (step S5) and a response is returned (step S6), it is considered that the service of the server computer is operating,
The value of the leading address is returned to the client program 24 as the address corresponding to the requested server name (step S10).

On the other hand, if there is no response from the server computer (step S6), it is considered that the service of the server computer is stopped, and the leading address is removed from the address list (step S6).
7) If the next address remains (step S
8) Then, returning to step S4, the operation status of the service of the server computer corresponding to the address is continuously confirmed.

In this way, steps S4 to S8 are performed.
If there is no response from the server computer corresponding to all the addresses on the address list (step S8), an error indicating that the address is unknown is returned to the client program 24, or the error processing is executed by the client. It entrusts it to the computer 24 side and returns any address in the address list (step S
9).

It should be noted that the name server computer 100 is provided with decision logic 13 for determining the priority order of a plurality of server computers according to a certain standard.
The decision logic 13 uses the metric string
Determine the preferred order of server computers by querying g) to each server computer. Here, as the optimum order, the order of the server computers on the basis that the load at the present time is smaller is adopted. The decision logic 13 launches a number of child processes, each process sending a metric string to a particular server computer and receiving a response from the server computer to which the metric string was sent. Decision logic 13
Determines the order in which the server computer should be preferred based on the received response, creates a list of pairs of server names and addresses, and writes the data in the storage device 17 by the writing means 15 while maintaining the preferred order. Update.

As described above, according to this embodiment, it is possible to realize the dynamic load distribution and the avoidance of the failure which are completely transparent to the client program 24.

Although one embodiment of the present invention has been described above, the present invention is not limited to such an embodiment and includes various modified embodiments according to the principle of the present invention.

For example, instead of returning the ordered address list from the name server computer 100 to the client computer 200 as in the above-described embodiment, the metric is included in the address list and returned to the client computer 200 to resolve the address of the client computer 200. The dynamic link library for use 23 may determine the priority order by itself based on the metric. It should be noted that in the case of such a change, the decision logic 13 of the name server computer 100 does not need to determine the priority order based on the metric, and only adds the metric to the address list and writes it in the storage device 17. Have advantages.

Further, as in the above-described embodiment, the address resolution dynamic link library 23 of the client computer 200 sends a temporary address to the client program 24 without sending a probe packet (a start address of the list or the client can communicate). Addresses not normally used) may be returned. That is,
Recognizing the necessity of checking the operating status of the server computer by modifying the existing communication library and detecting the temporary address when the client program 24 actually opens communication with this temporary address. Then, the operation status of the server computer is confirmed by the address resolution dynamic link library 23 according to the same procedure as in the above-described embodiment. It should be noted that such a change has an advantage that a communication opening request packet for the service of the server computer can be used as the probe packet.

[0034]

As described above, according to the present invention,
The dynamic link library for address resolution of the client sends the probe packet to check the operating status of the server computer, thus realizing dynamic load balancing between servers and avoiding failures without modifying the program on the client. can do.

[Brief description of drawings]

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

FIG. 2 is a flowchart showing a flow of processing by a dynamic link library for address resolution according to an embodiment of the present invention.

[Explanation of symbols]

 11 input device 12 output device 13 decision logic 14 conversion mechanism 15 writing means 16 reading means 17 storage device 21 input device 22 output device 23 address resolution dynamic link library 24 client program 100 name server computer 200 client computer 300 server computer 1 400 server Computer 2500 server computer 3 600 computer network

Claims (3)

[Claims] 1. In a server selection method in a computer network environment by a client / server method, a name server for passing an address list designating the order of servers to be selected to the client is provided, and an address for inquiring the address on the client side With the dynamic link library for resolution,
A server selection method characterized in that a probe packet is sent to each server according to the order of the address list passed from the name server to check the operating status of the server. 2. A server / server selection method in a computer network environment based on a client / server method, comprising a name server which includes a metric in an address list of a server to be selected and passes it to the client, and an address for inquiring the address on the client side. With the dynamic link library for resolution,
The order of the servers to be selected is determined based on the metric included in the address list passed from the name server, and the probe packet is sent to each server according to the determined order to confirm the operating status of the servers. Characteristic server selection method. 3. When the client actually opens communication with the server, the address resolution dynamic link library is called via a communication library, and a communication opening request packet to the server is used as the probe packet. 3. The server selection method according to claim 1, wherein the operating status of the server is confirmed by sending the server.