JPH09231180A - Server division method - Google Patents

Abstract

(57) Abstract: A server division method for dividing a service range and managing location information in a distributed manner when a load on a service location server exceeds a predetermined threshold value. In the client-server system shown in the figure, at least one system (1
The service location server (1013) is provided in (01), and the server (1013) includes service location information that manages location information of processes registered from other processes and service range information that defines the service range of its own server. , Measuring the service load of the local server, and when the service load exceeds a predetermined load level, creates a replica of the local server, divides the service range between the replication source server and the replication server, and Set the service location information for each of the replica servers, transport the replica servers to different processor systems, and arrange them.
For duplication, the location information of the duplication source is set.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a server dividing method in a client server system constructed in a distributed environment.

[0002]

2. Description of the Related Art In the field of computer hardware, parallel machines that use multiple processors and are capable of distributed processing and parallel processing have been attracting attention, and parallel machines equipped with several to several thousand processors are products. Has been converted. IBM's SP2,
An example is nCube2 of N-Cube in the US (Nikkei Computer 1994.9.5, p.93). Nikkei Computer 1994.9.5, p. According to the article described in No. 90, the method of connecting a plurality of processors and performing parallel processing is roughly classified into a tightly coupled multiprocessor, a cluster coupled multiprocessor, and a massively parallel machine. In the tightly coupled multiprocessor system, when the number of processors is increased, memory and bus contention occur, and generally 20 to 30 processors have a performance limit. Therefore, what is attracting attention now is the method of cluster-coupled multiprocessors and massively parallel machines, which has excellent expandability of the number of processors. Also, "What comes after the distributed operating system UNIX"
According to Maekawa, Tokoro and Shimizu, many attempts have been made even in the field of operating systems in order to efficiently manage computer resources of distributed systems including parallel machines as described above. OSF (Open System)
The abbreviation of the organization name "Foundation" minimizes the kernel (makes it into a microkernel) and realizes the UNIX function as a server. As a result, the processes that execute various processes of the operating system operate in concert. One example is the method of configuring as a process-oriented system. In addition, the above-mentioned document “What comes after the distributed operating system UNIX” p. According to No. 13, distributed processing requires distributed transparency such as position transparency, fault transparency, and copy transparency. The name server of Mach is a server that does not make the user aware of the position of a physical task (realization of position transparency) when communicating between tasks (processes) ("Distributed OS Mach book") Inui, Sugawara, p.
65).

Further, in a client server system constructed on a distributed system, in order for a client to issue a service request to a specific server and receive the service, at least the service requested by the client is provided. It is necessary to establish the existence of a server that operates and establish communication means between the client and the server that provides the service requested by the client. As one of the methods to efficiently investigate the existence of the server that provides the service requested by the client and efficiently acquire the information for establishing the communication means with the server,
A server that provides location information of various servers, such as the name server of Mach mentioned above (hereinafter,
There is a method of using a service location server). In this method, the service location server registers the service identifier indicating the service that each server process can provide, and the information necessary for the own server process and other processes to establish communication means, and the client wants to request. By specifying the service identifier indicating the service, the information for establishing the existence of the server and the communication means with the server can be acquired from the service location server. Furthermore, there are load balancing and load balancing as basic means for improving the performance in a distributed system. Basic functions for realizing this include remote execution and process transfer. Regarding remote execution, the distributed operating systems LOCUS and V-System support the function of creating processes for remote nodes. In addition, regarding process transfer, LOCU
It is supported by distributed operating systems such as S, Write, and V-System (the above-mentioned document, "Next to Distributed Operating System UNIX"). Furthermore, in order to use the resources of the distributed system efficiently, global resource management for the entire system is necessary. This management method has a centralized management method in which a specific node manages the entire system and a distributed management method in which the entire system is managed by cooperation of each node. However, both methods are problematic when applied to a large-scale distributed system. Therefore, a method of hierarchically managing system resources has also been proposed (the above-mentioned document "Distributed Operating System UN"
What comes after IX ").

[0004]

DISCLOSURE OF THE INVENTION As described above,
Centralized management methods and distributed management methods are available as global system resource management methods. The centralized management method is relatively easy to manage resources, but since there is only one management node,
There is a problem that load concentration occurs on the management node.
In the distributed management method, there are multiple management nodes, and the problem of load concentration on the management node in the centralized management method can be avoided, but since there are multiple management nodes, communication overhead between the management nodes increases. . In the distributed management method, there is a hierarchical management method that manages resources hierarchically in order to reduce the communication overhead between the management nodes, but when adopting the distributed management method, design such as how to distribute the management nodes. is required.

As described above, there are the above-mentioned management methods for managing global system resources. Which management method should be adopted, and how should the administrator be used when the distributed management method is adopted? It is necessary to design the system such as whether to disperse. In a client-server system in a distributed environment, the location information of each server is one of the global system resources, and it is necessary to design in advance how to manage the location information of each server. But,
Since the management system designed in advance is premised on the target system at the time of design, it is necessary to design for each target system, and also redesign is required when the system configuration is changed. An object of the present invention is to solve the problem that it is necessary to design a global system resource management system in advance, in a client server system in a distributed environment, a service location server that manages the location information of each server is set in advance. It is an object of the present invention to provide a server division method capable of dividing the service range and managing the location information in a distributed manner when the load on the service location server exceeds a predetermined threshold value, instead of distributing the service. . Furthermore, a further object of the present invention is to design a global system resource management method in advance, and since the predesigned management method is premised on the target system at the time of design, it is designed for each target system. In order to solve the problem that it is necessary to redesign even when the system configuration is changed, in a client server system in a distributed environment, service location servers that manage the location information of each server are distributed in advance. Instead, the service range is divided when the load on the service location server exceeds a predetermined threshold,
It is an object of the present invention to provide a server division method capable of managing location information in a distributed manner and specifying a division policy such as how to divide a service range.

[0006]

To achieve the above object, the present invention is constructed on a distributed system in which a plurality of local processor systems having a process transfer means for transferring a process between local processor systems are connected. In the client-server system, a service location server process is provided in at least one of the local processor systems, and the service location server process manages location information of processes registered from other processes, and its own server process. The service range information that defines the service range of the local server process is measured, the service load of the local server process is measured, and when the service load exceeds a predetermined load level, a copy of the local server process is created and The service range information between the replication source server process and the replication server process, setting service location information in each of the replication source server process and the replication server process, and transporting and allocating the replication server process to different local processor systems. I have to. Further, the location information of the replication server process is set in the replication source server process, and the location information of the replication source server process is set in the replication server process. Furthermore, when each of the divided server processes receives a service request from a process outside the service range, the service request is given as being out of the service range and the location information of the server process having a service range different from the own server process. To the process that issued the request, and when each of the server processes after the division receives an inquiry request regarding the location information of the process outside the service range from the process included in the service range, the server process that receives the request is the local server. Another server process having a service range different from the service range of the process is inquired, and the requested location information is notified to the requesting process. Further, the service load is the number of service executions executed by the service location server process within a certain time interval. Further, the service range is a set of local processor systems to be serviced. Further, when the service load exceeds a predetermined load level, the service range is divided according to a predetermined division policy.
Further, in the division policy, a fixed number of local processor systems are selected from the ascending order of the processor number of each local processor system among the local processor systems that are the service range, and the selected local processor system is used as a copy source server process. And the remaining local processor system, which is the service range, is set as the service range of the duplicate server process. Further, the division policy measures the number of service requests for each local processor system included in the service range information, and divides the service range based on the measured number of service requests for each local processor system. ing. Further, in a client-server system constructed on a distributed system in which a plurality of local processor systems having a process transfer means for transferring a process between local processor systems and a server process are connected, the server process comprises:
The service range information that defines the service range of the local server process is provided, the service load of the local server process is measured, and when the service load exceeds a specified load level, a copy of the local server process is created and the replication source server process is created. The service range is divided between the replica server process and the replica server process, and the replica server process is transported and arranged in different local processor systems. Further, the location information of the replication server process is set in the replication source server process, and the location information of the replication source server process is set in the replication server process.

[0007]

Embodiments of the present invention will be described below. << Embodiment 1 >> Embodiment 1 of the present invention will be described with reference to FIGS. In this embodiment, the process is a unit of execution of a program, the server is a process that provides a certain service in the client server system,
A client is a process that requests a specific service from a certain server in a client server system. Further, server process and server are synonymous, and client and client process are synonymous.

First, the configuration of a client server system to which the present invention is applied will be described with reference to FIGS. FIG.
1 shows an overall configuration of a client server system to which the present invention is applied. In FIG. 1, the client server system is a local processor system (101 to 105).
It is built on a distributed system that interconnects. A plurality of processes can operate in each local processor system, and in the distributed system, servers (1011, 1021, 1031, 104) that provide a certain service are provided.
1, 1051) and clients (1012, 1022, 1032, 1) that issue service requests to the server.
042, 1052) operates. The service location server (1013) exists on a specific local processor system (101), and the server (1011, 102).
1, 1031, 1041, 1051) manages the location information declared, and teaches the client the location information of the server that provides the service requested by the client. Each local processor system (1
01 to 105) are location service server information (1014, 1024, 1034, 1044, 105) for establishing communication means with the service location server.
4) and has a service location server (1013)
Processes other than (1011, 1021, 1031, 10
41, 1051, 1012, 1022, 1032, 10
42, 1052) are processor systems (101-101).
105) Location service server information (101)
4, 1024, 1034, 1044, 1054) can be used to establish and communicate with the service location server (1013). Furthermore, each local processor system (101 to 105) transfers process on its own local processor system to any other local processor system (1015, 1023, 1033, 1043, 105).
3).

FIG. 2 shows a service location server (1
013) is shown. In FIG. 2, the service location server (1013) includes service location information (201) composed of location information declared by each server, load measuring means (207) for measuring the service load status of the server (1013), and Service load information (20) measured by the load measuring means (207)
6), service range information (203) indicating the range of the local processor system serviced by the server (1013), and division criterion information (criterion for determining whether to divide the server (1013). 205)
A server dividing means (202) for dividing the server (1013), and a transfer destination processor system selecting means for selecting a local processor system to which the replication service location server generated from the server (1013) is transferred. (209) and the own server (101
Other than 3), when a service location server that manages the location information of the server process exists, it has other server location information (208) that stores the location information of another service location server.

FIG. 3 shows the service location server (1
Service location information (20
The structure of 1) is shown. In FIG. 3, the service location information (201) is a local processor system in which a service identifier (301) indicating the service content applied to the location server (1013) by an arbitrary server process and a server providing the service are arranged. Service server location processor identifier (302)
And service server communication information (302) for establishing communication means with a server providing a service, and a valid flag (304) indicating whether or not the registered information is valid. The service location server (1013)
At least the service server communication information (303) is responded to the inquiry about the service server location information from the client.

FIG. 4 shows the structure of the service range information (203). In FIG. 4, service range information (203)
Represents a set of processor systems serviced by the service location server (1013), a processor system identifier (401) indicating the processor system to be serviced, and identification information indicating which service location server is in charge of the service ( 402)
(When the service location server is divided, 2 and 3 are mixed as identification information.)
And a duplication source responsible information flag (403) indicating that the service range of the duplication source service location server falls within the service range of the duplication source service location server. In this example, the processor system identifier (401) is a processor number assigned to each local processor system, and the service range information (203) is entered in ascending order of processor number.

FIG. 5 shows the structure of the service load information (206). In FIG. 5, service load information (206)
Is the maximum number of service executions (501) per unit time since the service of the server (1013) was started, and the reference time (50) for determining whether or not the measurement time has passed.
2), a counter (503) of the number of services executed by the service location server after the reference time (502) is set, and a measurement time interval (504). The maximum service execution count (501) is the load measurement means (20
It is set by 7).

FIG. 6 shows the structure of the division standard information (205). In FIG. 6, the division criterion information (205) has a predetermined threshold value (601) of the maximum number of service executions per unit time. By comparing the maximum service execution number (501) and the threshold value (601), it is determined whether or not the server (1013) is to be divided.

FIG. 7 shows a local processor system (1
02) Location service server information (102)
4) shows the configuration. In FIG. 7, the location service server information (1024) establishes a communication means between the service server location processor identifier (701) indicating the local processor system in which the service location server (1013) is arranged and the server (1013). It has communication information (702) for performing. Location service server information (1014, 1034, 10) on another local processor system in FIG.
44, 1054) is also the same.

FIG. 8 shows the service location server (1
Other server location information (208) owned by 013)
Is shown. In FIG. 8, the other server location information (208) holds location information of another service location server that services a service range not included in the service range information (203) of the server (1013). Other server location information (20
8) is for establishing a communication means between a server location processor identifier (801) indicating a local processor system where another service location server exists and a service location server on the local processor system indicated by the processor identifier (801). Communication information (80
2) and the identification information (803) of another service location server. The identification information (803) corresponds to the identification information (402) of the service range information (203).

In the client / server system having the above-described configuration, the service location server (101
3), servers other than the service location server (1013), and clients operate as follows. The servers other than the service location server (1013) store the location information of the own server including the service identifier indicating the service provided by the own server and the communication information for the own server and other clients to communicate with each other. Apply for registration at 1013). The service location server (1013) having received the location information registration application registers the requested location information as service location information.

The client inquires the service location server (1013) of the location information of the server that provides the requested service by designating the service identifier indicating the service requested by the client. Upon receiving the inquiry request from the client, the service location server (1013) searches the service location information by the specified service identifier, and makes an inquiry request for communication information with the server that provides the service indicated by the specified service identifier. Notify the client that issued the. The client, which has obtained the communication information with the server providing the request service of the client, establishes the communication means with the server providing the request service and issues a service request.

Further, the service location server (1
013) receives the request for its own server (1013), the load measuring means (207) measures the service load of its own server (1013), and the service whose service load is defined in the division criterion information (205). When the load becomes larger than the load, the service dividing means (20
The server division processing of the own server (1013) is performed by 2).

Next, the registration processing of the service location information will be described with reference to FIG. FIG. 9 is a flowchart showing the operation of the location information registration processing of the service location server (1013) when an arbitrary process issues a location information registration request. In FIG. 9, step 902 is a process of analyzing the content of the issued request when the location information registration request is issued from an arbitrary process. The location information registration request received by the service location server includes a service identifier indicating the service provided by the request issuing process, a processor system identifier indicating the local processor system in which the requesting process exists, and an arbitrary process including the registration request issuing process. It contains communication information for establishing communication means. By analyzing the received location information registration request, the service location server (101
3) can know which process exists, where, and which service is provided. Steps 903 and 9
Reference numeral 04 is a service range determination process for determining whether or not the process that has issued the registration request is included in the service range handled by the service location server. The service location server executes the above step 902.
It is possible to know the local processor system in which the process that issued the request exists, and further, in the service range information (203), the processor system identifier indicating the local processor system in which the process that issued the request exists is registered, and If the identification information indicating the service range of the service location server is set, it is determined to be the service range.

Steps 901, 908, 909, 90
Reference numerals 5 and 906 denote the processing when the process that has issued the registration request in steps 903 and 904 is included in the service range of the service location server.
Step 907 is a process when it is determined that the process that issued the registration request is out of the service range of the service location server. Step 901 is
This is service load measurement processing for measuring the service load of the service location server (1013). In this embodiment, it is assumed that the number of services executed at regular time intervals is measured. Details of this process will be described later with reference to FIG. Steps 908 and 909 are the same service range determination processing as steps 903 and 904. This process is step 9
Since the server may be divided at 01, this processing is performed assuming this division. Steps 908, 90
If it is determined in 9 that the service is out of the service range, the registration request is ignored. In step 905, the location information requested to be registered is converted into service location information (201).
Register with. Service location server (1013)
Is service location information, a service identifier (301) indicating the service content, a service server location processor identifier (302) indicating the local processor system in which the process that issued the registration request exists, and a communication means with the process that requested the registration. The communication information (303) for establishing the communication is registered, and the valid flag (304) is set. Step 906 is a registration completion notification process for notifying the process that has requested the registration of the completion of the registration process. The service location server notifies the process that issued the registration request that the registration process has been completed.

If it is determined in steps 903 and 904 that the process that issued the registration request is outside the service range of the service location server,
The registration failure notification process of step 907 is performed. Step 9
07 indicates that the process that issued the registration request is outside the service range of the service location server, and that the identification information (401) of the processor system identifier (401) corresponding to the local processor system in which the process that issued the registration request exists. The registration request issuance process is notified of the location information of the other service location server having the identification information (803) of the other server location information (208) that matches 402).

Next, referring to FIG. 10, the inquiry processing of the service location information will be described. FIG. 10 is a flowchart showing the operation of the location information inquiry processing of the service location server (1013) when an arbitrary process issues a location information inquiry request. In FIG. 10, step 10001 includes
When a location information inquiry request is issued from an arbitrary process, it is a process of analyzing the content of the issued request. The location information inquiry request received by the service location server includes a service identifier indicating a service to be received by the inquiry request issuing process, a processor system identifier indicating a local processor system in which the request process exists, and a process issuing the inquiry request. And a flag indicating whether the service location server has a service range different from that of the service location server. By analyzing the received location information inquiry request, the service location server (1013) can know where the request issuing process is, and which service is inquiring about the location information of the server process providing the service. Further, it is possible to know whether or not the process that issued the inquiry request is a service location server having a service range different from that of the own service location server.

In step 10016, it is determined whether the process that issued the inquiry request is a service location server other than the service location server. If the process that issued the inquiry request is a service location server other than the service location server, the process proceeds to step 10007,
Otherwise, the process proceeds to step 10002.
Steps 10002 and 10003 are service range determination processing for determining whether or not the process that issued the inquiry request is included in the service range handled by the service location server. The service location server can know the local processor system in which the process that issued the request exists in step 10001, and further, in the service range information (203),
If the processor system identifier indicating the local processor system in which the process that issued the request exists is registered and the identification information indicating the service range of the service location server is set, it is the service range. To determine. Step 1 above
If it is determined in 0002 or 10003 that the service is out of the service range, the out-of-range report notification process in step 10015 is performed. In step 10015, the process that issued the inquiry request is outside the service range of the service location server, and further, the identification information of the processor system identifier (401) corresponding to the local processor system in which the process that issued the inquiry request exists. The inquiry request issuing process is notified of the location information of the other service location server having the identification information (803) of the other server location information (208) that matches (402).

Step 10004 to Step 10014
Is processing when the inquiry request is issued in steps 10002 and 10003 and it is determined that the process is included in the service range of the service location server. Step 10004 is a service load measuring process for measuring the service load of the service location server (1013). In this embodiment, it is assumed that the number of services executed at regular time intervals is measured. Step 100
05 and 10006 are steps 10002 and 10003.
This is the same service range determination processing as. Since this process may divide the server in step 10004, this process is performed assuming this division. Step 10
If it is determined in 005 and 10006 that the service is out of the service range, the inquiry request is ignored. This is to prevent confusion that occurs when the division is performed. When it is determined in the above steps 10005 and 10006 that the service is within the service range, the processing of steps 10007 to 10014 is executed.

In step 10007, the service location information (201) is searched for location information having a service identifier that matches the service identifier designated by the process that issued the inquiry request. Step 10
Reference numeral 008 denotes a process of determining whether or not the location information retrieved in the above step 10007 exists in the service location information (201). Step 10008 above
If it is determined that the requested location information does not exist, the process proceeds to step 10012, and if the requested location information is found, step 1012 is performed.
The process proceeds to 009. Step 10009, 10010
Is a process of determining whether or not the location information found in step 10007 is the location information managed by the self-service location server. Based on the service server location processor identifier of the location information found in the above step 10007 and the service range information (203) of the service location server, it is determined whether or not it is within the service range of the service location server. When it is determined in the above steps 10009 and 10010 that the location information retrieved in the above step 10007 is out of the service range, step 10012
Processing proceeds to If it is determined that the location information retrieved in step 10007 is included in the service range, the process proceeds to step 10011. In step 10011, the retrieved location information is
This is inquiry information notification processing for notifying the process that issued the inquiry request. The service location service server notifies the process that issued the inquiry request of the location information.

Step 10012 is a process for inquiring location information to another service location server when an inquiry request for location information other than the location information managed by the service location server is received. The service location server uses the location information registered in the other server location information to inquire of the other service location server the location information of the server that executes the requested service. Step 10
013 is obtained by the processing result of step 10012.
This is a process of determining the result of the inquiry to the other service location server. When the location information requested by the other service location server is found by the above step 10012, by the above step 10011,
The service location server is notified of the location information to the process that issued the inquiry request. If the location information requested by the other service location server is not found, step 10014
Processing proceeds to Step 10014 is an unregistered notification process when the location information requested by the service location server is not found. The service location server notifies the process that issued the inquiry request that the location information is not registered.

Next, step 901 in FIG. 9 and FIG.
The service load measuring process which is step 10004 in 0 will be described with reference to FIG. In this embodiment, it is assumed that the number of services executed at regular time intervals is measured. FIG.
In step 11001, in step 11001, the service location server acquires the time when it receives a request. Steps 11002 and 11003 are measurement time interval determination processing for determining whether or not the current load measurement time interval has passed. Steps 11002, 11
In 003, the time when the reference time (502) (time to start measurement) is advanced by the measurement time interval (504),
The times acquired in step 11001 are compared. If the time obtained by advancing the reference time (502) by the measurement time interval (504) is smaller than the time obtained at step 11001, the process proceeds to step 11004. In other cases, it is determined that the load is being measured, and the processing proceeds to processing 11008.

Steps 11004 to 11007
Is processing when it is determined in steps 11002 and 11003 that the current load measurement time interval has passed. Steps 11004 and 11005 are update determination processing for determining whether to update the maximum service execution count (501) per unit time included in the service load information (206). Step 11004, 11005
Compares the value of the maximum service execution number (501) with the value of the execution service counter (503) to determine whether the value of the maximum service execution number (501) needs to be updated. Step 11006, 11007
Is processing when it is determined in steps 11004 and 11005 that the value of the maximum service execution number (501) needs to be updated. Step 11006 is
This is a process of updating the maximum execution service number (501), and the value of the execution service number counter (503) is copied to the maximum execution service number (501). Step 11007 is a measurement environment initialization process, the content of the reference time (502) is changed to the time acquired in step 11001, and the execution service number counter (503) is cleared. If it is determined in steps 11004 and 11005 that the value of the maximum service execution count (501) does not need to be updated, the process proceeds to step 11008.

In step 11008, the execution service number counter (503) is updated, and the value of the execution service number counter (503) is incremented by one. Steps 11009 and 11010 are processing for determining whether to divide the service location server. The determination is made by comparing the value indicated by the maximum service execution number (501) included in the service load information (206) and the threshold value (601) included in the division criterion information (205). The value of maximum service execution count (501) is the threshold value (601)
If it is larger than the value of, it is determined that the division is necessary, and otherwise it is determined that the division is not necessary. Step 11 above
If it is determined by 009 and 11010 that division is necessary, server division processing (step 11011) is executed.

Next, the server dividing process will be described with reference to FIG. FIG. 12 is a flowchart showing the operation of the server division process. In FIG. 12, step 12001 includes
This is the duplication source service range setting process for setting the division state of the service range that the service location server is in charge of. In step 12001, the duplication source charge information flag (403) included in the service range information (203) of the service location server is set. When the server is divided, the copy source charge information flag (403) includes the local processor system indicated by the processor system identifier in which the copy source charge information flag (403) is set in the service range of the copy source server. Indicates that. The division policy of the service range is predetermined, and in this example, the division policy that divides the number of systems of the processor system registered in the service range information and the processor system in charge of its own service location server into two. It will be. That is, if the number of local processor systems in charge of the own service location server is k,
When the number of systems k is an even number, the number of systems in charge is divided by (k / 2) by the replication source server and the replication server, and when the number of systems k is an odd number, ((k-1) / 2
+1) and ((k-1) / 2) are assumed to be divided by the replication source server and the replication server. In step 12001,
The duplication source responsible information flag (403) of the local processor system in charge of the duplication source server is set by the number of divided processor systems.

Step 12002 is a server replica generation process for generating a replica of the current service location server on the same local processor system. Step 1
In 2002, the operation environment including the internal data of the service location server which is the copy source at the time when the copy is generated is copied. Step 12008 is a process of determining whether the replication server and the replication source server are replicated servers or replication source servers. Step 12004 is a transfer destination processor system selection process for selecting a local processor system in which a duplicate server is arranged. The copy source server selects one of the local processor systems that the copy server is in charge of by the transfer destination processor system selection means (209) of the copy source server. Step 12005 is a process transfer process in which the copy source server transfers the copy server to the local processor system selected in step 12004 above. Step 12006 is another server location information setting process in which the replication source server and the replication server set their location information in the other server location information. The duplication source server and the duplication server can communicate by setting the location information of each other. Step 12007 is a service range information setting process in which the replication server and the replication source server respectively modify the service information managed by themselves. Of the service range information of the self server, the duplication server identifies the identification information (402) of the information that is the local processor system that the self server is in charge of and the duplication source responsible information flag (403) is set. It is updated so as to match the content of the identification information (803) of the other server location information set in step 12006. The duplication source server uses the identification information (402) of the information in the service range information of the self server that is the local processor system that the self server is in charge of and the duplication source responsible information flag (403) is not set. It is updated so as to match the content of the identification information (803) of the other server location information set in step 12006. Furthermore, the duplication server and the duplication source server clear the duplication source charge information flag (403) after the setting of the identification information (402) of the service range information (203) managed by the own server is completed.

In the examples described with reference to FIGS. 1 to 12, the present invention is applied to a service location server that manages location information between processes. As shown in the above example, by applying the present invention, the service location server can divide its service range according to its load situation, and increase the number of servers in a self-proliferating manner. Furthermore, since the number of servers increases according to the load, it is not necessary to fixedly determine the number of servers. Further, the server can be duplicated according to the load, and global system resources can be distributedly managed by the duplication source server and the duplicate server.

Further, in the above example, after the service range of the service location server is divided, the service location server that receives the registration / inquiry request regarding the location information performs the following processing. When the location information registration request process exists in the local processor system within the service range, the service location server registers the requested location information in the location information of its own server.
If the location information registration request process exists in the local processor system that is out of the service range, the process that issued the registration request is out of the service range and the location information of the service location server having a service range different from that of the own service location server is displayed. , Notify the process that issued the registration request. As a result, the process notified of being out of service can issue a registration request to another service location server. If the location information inquiry process exists on the local processor system within the service range, and the requested location information is the location information of the process on the local processor system within the service range, the service location server has the location information owned by itself. Of the requested location information. If the location information inquiry process exists in the local processor system within the service range, and the requested location information is the location information of the process in the local processor system outside the service range, the service location server is different from its own server. The other service location server having the service range is inquired about the requested location information, and the location information inquiry process is notified of the location information provided by the other service location server. When the location information inquiry process exists in the local processor system that is out of the service range, the process that issued the inquiry request is out of the service range, and the location information of the service location server having a service range different from that of the own service location server is displayed. Notify the process that issued the inquiry request. As a result, the process notified of being out of the service range can issue an inquiry request to another service location server.

<< Embodiment 2 >> Next, in the above embodiment,
An embodiment in which the service load is measured for each local processor system and the division range is determined based on the measurement result for each processor will be described with reference to FIG. FIG. 13 is a diagram showing the configuration of service load information for measuring the service load of each local processor. In FIG. 13, the service load information (206) is the maximum number of service executions (501) per unit time since the service of the server (1013) was started, and a criterion for determining whether or not the measurement time has passed. The time (502), the counter (503) of the number of services executed by the service location server after the reference time (502) is set, the measurement time interval (504), and the service load of each processor system requesting the service. It has information (13001). The service load information (13001) for each processor system is
A processor identifier (13002) indicating a local processor system, a counter for executing service number per unit time (13003) for each processor system, and a maximum per unit time since the service of the server (1013) for each processor system is started. It has a service execution count (13004).

Next, the processing for measuring the service load of the server (1013) using the service load information (206) having the configuration shown in FIG. 13 will be described. Basically, the same processing as the service load measurement processing described with reference to FIG. 11 is performed. In step 11006 in FIG. 11, the execution service number counter (1300) of each local processor system of the processor service load information (13001)
The value of 3) is compared with the value of the maximum service execution number (13004) per unit time of each local processor system, and the value of the execution service number counter (13003) is larger than the value of the maximum service execution number (13004). Sometimes, a process of copying the value of the execution service number counter (13003) to the value of the maximum service execution number (13004) is added. Further, step 11007 in FIG.
In addition, processing for clearing the value of the execution service number counter (13003) for each local processor system is added. Further, in step 11008 in FIG. 11, a process of increasing the value of the execution service number counter (13003) of the local processor system in which the process issuing the request exists by 1 is added. As described above, the service load of each local processor system can be measured as the service load information. Further, by using the service load information (13001) in the duplication source service range setting process of step 12001 in the server division process shown in FIG. The number of services can be used. In other words, by request from each local processor system,
The service range can be set.

In the embodiment described above, the service load level of the server is measured, and when the load level exceeds a predetermined level, the processor of each local processor system among the local processor systems included in the service range of the server is measured. Select a fixed number of local processor systems from the smallest number, set the selected local processor system as the service range of the replication source server process, and set the remaining local processor systems that are the service range as the service range of the replication server process. The first embodiment and the second embodiment in which the service load level is measured for each local processor within the service range have been described.

<Third Embodiment> Next, as a third embodiment, an embodiment in which the server is divided according to the contents of the file in which the division policy is described will be described. Example 3 will be described with reference to FIGS. 14 to 19 and 24. FIG. 14 is a diagram showing a part of the configuration of a client server system to which the present invention is applied. FIG.
Reference numeral 4 is a part of the client-server system shown in FIG. 1, and shows the configuration of the local processor system (101). The configuration of the client-server system to which the present invention is applied in the third embodiment conforms to the system configuration shown in FIG. 1, and the components other than those shown in FIG. 14 are the same.

In FIG. 14, the processor system (1
01) has an external storage device (14001). Further, the external storage device (14001) has a division policy description file (14002) describing the division policy of the service location server (14003). The service location server (14003) is similar to the service location server (1013 (FIG. 1)) in the servers (1011, 1021, 1031, 1041, 105).
The location information declared by 1) is managed, and the location information of the server that provides the service requested by the client is taught to the client. Further, the service location server (14003) reads the distribution policy description file using the external storage device interface (14004).

FIG. 15 is a diagram showing the structure of the service location server (14003). In FIG. 15, the service location server (14003) has service location information (201), service range information (203), division criterion information (205), service load information (206), and load measuring means (207). Other server location information (208), server dividing means (202), dividing policy information (15001), dividing level information (15002), and transfer destination processor system selecting means (209). Division level information (150
02) indicates the division level of the server (14003).
The initial value of the division level information (15002) is 0, and the division level information (15
002) is incremented by 1. Division policy information (1500
1) is server division policy information that reflects the contents of the division policy description file (14002).

FIG. 16 shows a division policy description file (1400
It is a figure which shows the structure of 2). In FIG. 16, the division policy description file (14002) is composed of division policy description records (16001 to 16005).

FIG. 17 shows division policy description record 1 (16
It is a figure which shows the structure of (001). In FIG. 17, the division policy description record 1 (16001) includes a division level designation field (17001) that designates a division level and a server allocation processor designation field (17 that specifies a local processor system in which a duplicated server after division is allocated.
002), of the records having the same division level, a priority designation field (17003) for designating a priority for determining which record is to be preferentially selected, and a service range of the duplicated server after division. It has a service range designation field (17004).

FIG. 18 shows a division policy description file (140
It is a figure which shows the example of description of 02). In FIG. 18, the division policy description file (14002) is composed of four division policy description records, and each field (17001, 17002, 17003, 1700) of the division policy description record.
4) is separated by a colon (:). The first record indicates a local processor system group having a division level of 0, the number of the local processor system in which the duplicate server is arranged, the priority of 1, and the service range of the processor system numbers 1 to 16. The second record indicates that the local processor system group has the division level 1, the number of the local processor system in which the duplicate server is arranged is 9, the priority is 1, and the service range is the processor system numbers 9 to 16. The third record is
This shows that the local processor system has a division level of 1, the number of the local processor system in which the duplicate server is arranged is 5, the priority is 2, and the service range is the processor system numbers 5 to 8. The fourth record indicates that the division level is 2, the number of the local processor system in which the duplicate server is arranged is 13, the priority is 1, and the service range is a processor system number 13 to 16 group. Further, division level 0 indicates the initial state of the server, division level 1 indicates that the server divides from the division level 0 state, and division level 2 indicates that the server divides from division level 1. further,
The value of the priority specification field means that the smaller value has priority.

Therefore, according to the division policy description file of FIG. 18, a service location server, which is a group of local processor systems having service numbers of processor numbers 1 to 16, is arranged in the local processor system of processor number 1 as an initial state. At this time, the service location server is referred to as server 1. When the service load of the server 1 exceeds the predetermined load level, the server 1 creates a duplicate server (the duplicate server (replica service location server) at this time is referred to as the server 2) and divides the service range from the server 2. . The service range of the server 1 after the division is the local processor system group of processor numbers 1 to 8, and the service range of the server 2 is the local processor system group of processor numbers 9 to 16.
Further, the arrangement of the server 2 after the division of the server 1 is on the local processor system with the processor number 9.

Further, when the service load of the server 1 exceeds a predetermined load level, the server 1 further generates a duplicate server (at this time, the duplicate server is the server 3),
The service range is divided from the server 3. The service range of the server 1 after this division is the local processor system group of processor numbers 1 to 4, and the service range of the server 3 is the local processor system group of processor numbers 5 to 8. The server 3 is arranged on the local processor system with the processor number 5.

When the service load of the server 2 exceeds a predetermined load level, the server 2 creates a duplicate server (the duplicate server at this time is referred to as the server 4) and divides the service range from the server 4. The service range of the server 2 after this division is the local processor system group of processor numbers 9 to 12, and the service range of the server 4 is the local processor system group of processor numbers 13 to 16. The server 4 is arranged on the local processor system with the processor number 13. When the duplicate server is created, the division policy information of the duplicate source server is also copied to the duplicate server.

FIG. 19 is a diagram showing the structure of the division policy information (15001). In FIG. 19, the division policy information (15001) includes division level (19001), server allocation processor identifier (19002), priority information (19003), and service range information (19004).
And a use flag (19005). The division level (19001) has a value designated in the division level designation field (17001) of the division policy description record,
The server allocation processor identifier (19002) has the value specified in the server allocation processor specification field (17002) of the division policy description record, and the priority information (1
9003) has the value specified in the priority specification field (17003) of the division policy description record. The service range information (19004) indicates the service range specified in the service range specification field (17004) of the division policy description record. In FIG. 19, the total service range is 1
6 shows an example of a client-server system composed of 6 local processor systems, in which each local processor system is associated with 1-bit data,
The service range information (19004) has a data length of 16 bits. Each bit of the service range information (19004) indicates whether or not the corresponding local processor system is included in the service range. When it is 1, it indicates that it is within the service range, and when it is 0, it is outside the service range. Indicates that. The use flag (19005) is a flag indicating whether the division policy information is used or unused.
When the use flag (19005) is 1, it indicates that the corresponding division policy information has been used, and the use flag (190
When 05) is 0, it indicates that the corresponding division policy information is unused. The division policy information (15001) is set by the service location server (14003) reading the division policy description file (14002) after the service location server (14003) is activated.

As described above, in the client-server system having the configuration described with reference to FIGS. 14 to 19, the service location server (14003), the server other than the service location server (14003), and the client operate as follows. Service location server (1
Operations of the server and the client other than 4003) are the same as those in the first embodiment. Further, in the processing of the service location server (14003), the division level information (15
002), division policy information (15001) setting processing, and processing other than server division processing are the same as the processing of the service location server (1013) described in the first embodiment.

The setting process of the division policy information (15001) and the division level information (15001) is executed by the service location server initialization process and the server division process. Initialization processing of the service location server (14003) (executed before the service location server (14003) starts processing location information registration request from server other than service location server (14003) and location information inquiry request from client Process), the service location server (14003) determines the division policy description file (1400
2) is read, the contents of the file (14002) are interpreted, and division policy information (15001) is set. Of the division policy information (15001), the division level (1900
1), server placement processor identifier (190002),
Priority information (19003), service range information (190
04) is set by the content of the division policy description file (14002), and the use flag (19005) is initialized to 0. Further, of the division policy information (15001), the use flag of the division policy information whose division level is 0 is set to 1. Further, the division level information (15002) is set to 0.

Next, the server dividing process in the third embodiment will be described with reference to FIG. FIG. 24 is a diagram illustrating a flowchart of server division processing according to the third embodiment. FIG.
4 is step 2400 in the flowchart of FIG.
1, Step 24002, Step 24003, Step 24004, Step 24005, Step 2400
6 is added, and Step 2 is replaced with Step 22005.
It is the flowchart figure which changed into 4007. In FIG. 24, step 24001 is the division policy information (150
01) is a process of determining whether or not the division policy information of the server division process exists. If the division policy information exists, the process proceeds to step 24006. If the division policy information does not exist, the process proceeds to step 12001.

In the third embodiment, in step 24001, if there is division policy information that satisfies all of the following four conditions, it is determined that division policy information exists. The first condition is 1 for the content of the division level information (15002).
Is to have a division level (19001) of a large value. The second condition is that service range information (1900) is set in the service range of the service location server.
That is, it includes all the service ranges indicated by 4).
The third condition is that the use flag (19005) is 0. The fourth condition is that the duplicate server of the service location server (14003) does not exist on the local processor system indicated by the server placement processor identifier (19002). Step 24006 is a division policy information selection process of selecting the division policy information having the smallest priority information (19003) among the division policy information (15001) satisfying the above four conditions in step 24001. Furthermore, in step 24006, the use flag (19005) of the selected division policy information is used.
Is set to 1.

In step 24002, the division status of the service range in charge of the service location server is
The division policy information (15
(001) is a duplication source service range setting process of the division policy information base set based on (001). Step 240
02, the service range information (19004) of the division policy information selected in step 24006 among the local processor systems included in the service range information (203) of the service location server and included in the service range of the service location server. ), The copy source charge information flag (40
Set 3).

Step 24003 is step 2400.
The determination process is the same as 1. If the corresponding division policy information exists, the process proceeds to step 24004, and if the corresponding division policy information does not exist, the process proceeds to step 12004. Step 24004 is a process for setting the local processor system in which the duplicate server is arranged. In step 24004, the server allocation processor identifier (190 of the division policy information selected in step 24006).
The local processor system corresponding to the processor number registered in 02) is selected as the processor system in which the duplicate server is arranged. Step 24007
Is a process transfer process for transferring the replication server to the local processor system selected in step 12004 or step 24004. Step 2400
A division level setting process 5 sets division level information (15002) of the replication server. Step 2400
In 5, the content of the division level information (15002) of the replication server is incremented by 1. Nothing is done to the division level information (15002) of the copy source server.

As described above, in the third embodiment described with reference to FIGS. 14 to 19 and 24, when dividing the service location server, the division policy is determined according to the contents of the file describing the division policy, and the service location server is divided. I explained an example. Further, in the third embodiment, when the division policy is not described in the division policy description file, the division policy described in the first embodiment is adopted. As shown in the third embodiment, by applying the present invention, it is possible to specify the division policy when dividing the server.
Further, by designating the division policy, it is possible to designate the local processor in which the duplicate server after the server division is arranged. Further, by specifying the division policy, the service range of the copy source server and the copy server can be specified.

<< Fourth Embodiment >> Next, as a fourth embodiment, an example in which the server is not divided when the division policy is not described in the division policy description file will be described with reference to FIG.
FIG. 25 is a flowchart showing the server dividing process in the fourth embodiment, and is a modification of the flowchart of the server dividing process in the third embodiment shown in FIG. In FIG. 25, in step 24001, if there is no corresponding division policy information, nothing is done in the server division processing. The other steps described in FIG. 25 are the same as the processing contents described above. The fourth embodiment is the same as the third embodiment except for the server dividing process described in FIG. As shown in the fourth embodiment, by applying the present invention, it is possible to specify the division policy when dividing the server. Further, by designating the division policy, it is possible to designate the local processor in which the duplicate server after the server division is arranged. Furthermore, by specifying the division policy,
The service range of the replication source server and replication server can be specified. Furthermore, the server division process can be stopped.

<< Fifth Embodiment >> Next, as a fifth embodiment, when the service location server is divided, division criterion information (20
An example of setting 5) will be described with reference to FIGS.
FIG. 26 is a diagram showing the structure of a division policy description record in the fifth embodiment. In FIG. 26, the division policy description record in the fifth embodiment is the same as the division policy description record described in FIG.
Has been added. The division criterion designation field (26001) is the maximum per unit time as the division criterion information (205) of the service location server arranged on the local processor system corresponding to the processor number designated in the server arrangement processor designation field (17002). Service execution threshold (60
This is a field for specifying 1).

FIG. 27 is a diagram showing the structure of the division policy information held by the service location server in the fifth embodiment. In FIG. 27, the structure of the division policy information in the fifth embodiment is the division policy information (1500) described in FIG.
This is a configuration in which division criterion information (27001) is added to 1). The division standard information (27001) has a value designated in the division standard designation field (26001).

Next, the division standard setting process for setting the division standard information (205) will be described. Division criteria information (205)
Is the division criterion information (2
05) is set. The server dividing process for setting the dividing criterion information (205) is performed in step 2400 described in FIG.
In FIG. 2, a process of setting the division standard information (27001) included in the division policy information as the division standard information (205) of the replication server may be added. In the fifth embodiment, the configuration of the division policy description record, the configuration of the division policy information, the configuration other than the setting process of the division criterion information (206), the processing, etc. are the same as those in the third embodiment. As described in the fifth embodiment, by applying the present invention, it is possible to specify the division policy when dividing the server. Further, by designating the division policy, it is possible to designate the local processor in which the duplicate server after the server division is arranged. further,
By specifying the division policy, the service range of the copy source server and the copy server can be specified. further,
The division standard information (205) of the duplicate server can be designated for each duplicate server.

<Sixth Embodiment> Next, as a sixth embodiment, an example in which the location information is multiplexed by the service location server which is the replication source and the replication service location server will be described. In the sixth embodiment, the multiplicity of location information is 2. Further, in the sixth embodiment, the backup server means the replication source server when the service range is divided between the replication source server and the replication server.

The sixth embodiment will be described with reference to FIGS. 20 to 23 and 28. In the sixth embodiment, the configuration of the client server system to which the present invention is applied conforms to the client server system described in the first embodiment. FIG. 20 shows location service server information (101
It is the figure which showed the structure of 4). In FIG. 20, the location service server information (1014) has a location server information table (20002) for registering the location information of the service location server and current service location server information (20001). The location server information table (20002) is a server location processor identifier (2000) indicating the local processor system in which the service location server exists.
4), communication information (20005) for establishing communication means with the service location server, and entry number (2) used as an identifier of the location server information.
0003) and a valid flag (20006) indicating whether or not the location information is valid. Further, the location server information table (20002) is capable of registering at least location information having a multiplicity of location information or higher. In the sixth embodiment, since the multiplicity of location information is 2, at least two pieces of location information can be registered in the location server information table (20002). When the multiplicity of location information is 3, at least three pieces of location information can be registered in the location server information table (20002).

Current service location server information (20
001) is the location server information table (200
This is an identifier for identifying the location information of the service location server currently in charge of the local processor system of the location information included in 02). In the sixth embodiment, the entry number of the location information registered in the location server information table (20002) is used as the current service location server information (20
001). That is, it is assumed that the location information of the entry number (20003) that matches the value of the current service location server information (20001) is the location information of the service location server currently in charge of its own local processor system. Note that FIG.
The configuration of the location service server information (1014) has been described above. Location service server information (1024, 10) present in other local processor systems.
34, 1044, 1054) also has the same configuration as the location service server information (1014).

FIG. 21 is a diagram showing the structure of the service location server (1013) in the sixth embodiment.
In FIG. 21, a service location server (101
3) includes multiplicity information (21001) indicating the multiplicity of location information and service location information (210
03), server dividing means (202), service range information (203), division reference information (205), service load information (206), load measuring means (207),
Other server location information (21002) and transfer destination processor system selecting means (209) are included.

FIG. 22 is a diagram showing the structure of the other server location information (21002) in the sixth embodiment.
In FIG. 22, other server location information (2100
2) is identification information (803), server location processor identifier (801), and communication means (802) with the server.
And backup server information (22001) indicating whether the server in which the location information is registered is the replication source of the self server, the replication server, or another server. In the sixth embodiment, 1 is set in the backup server information (22001) when the registered location information is the location information of the replication source server of the own service location server, and the registered location information is the own service location server. If it is the location information of the replication server having the replication source server of 2, the registered location information is not the location information of the replication source server of the local server, and is not the location information of the replication server having the local server as the replication source server. 0 is set in the case. For example, in FIG.
The identification information is 2, the server location processor identifier is 50,
The location information having the communication information with the server of 3 is the location information of the duplicate server having the self-service location server as the duplication source server because the location information of 2 is set as the backup server information.

FIG. 23 is a diagram showing the structure of the service location information (21003) in the sixth embodiment.
In FIG. 23, the service location information (2100
3) indicates that the service identifier (301), the service server location processor identifier (302), the communication information (303) with the service server, the valid flag (304), and the location information are supervised. It has a multiple data flag (23001) shown. The location information in which the valid flag (304) and the multiplex data flag (23001) are set indicates that the location information is managed by the service location server as a backup server. In FIG. 23, the service identifier (301) is 1800, 0010, 021.
The location information of 0,0000 indicates that the service location server manages the location information as a backup server.

FIG. 28 is a flow chart showing the service location information registration processing in the sixth embodiment. In FIG. 28, step 28001 is a process of determining whether or not the process that issued the registration request is a duplicate server of the own service location server. In step 902, the process that issued the registration request can be identified.
In step 28001, if the process that issued the registration request is a duplicate server of its own server, step 28004
If the process that has issued the registration request is not the duplicate server of its own server, the process proceeds to step 903. Step 28004 is a location information registration process of registering the location information requested for registration in the service location information (21003). The service location server itself has, as the service location information, a service identifier (301) indicating the service content and a service server location processor identifier (302) indicating the local processor system in which the process that issued the registration request exists.
And the communication information (303) for establishing the communication means with the process requested to register, and the valid flag (304)
Is set. Furthermore, in step 28004, if the process that issued the registration request is a duplicate server of the own service location server, the multiple data flag (23
001) is set.

Step 28002 is a process for judging whether or not it is necessary to register the location information requested for registration in the copy source server of the own service location server. In step 28002, if the process that has issued the registration request is not the duplicate server of the own service location server, it is determined that the location information requested to be registered needs to be registered in the duplicate source server of the own service location server. Step 28003 is
This is a process of requesting the backup server to register the location information requested to be registered in the own service location server. The self-service location server issues a location information registration request to the duplication source server of the self-service location server, and the duplication source server that receives the location information registration request from the duplication server requests the service location information that the self-service location server has. The registered location information is registered. Step 28005 is a registration completion notification process for notifying the process that requested the registration of the completion of the registration process. If the process that issued the registration request is a duplicate server of the self-service location server, it notifies the process that issued the registration request that the registration process has been completed. If the process that issued the registration request is a process other than the duplicate server of the local service location server, whether registration to the local service location server is complete and whether the request to register the local service location server to the backup server has succeeded To notify you.

Next, the inquiry processing of the service location information in the sixth embodiment will be described. The service location information inquiry processing in the sixth embodiment is similar to the service location information inquiry processing described in the first embodiment. In the sixth embodiment, steps 10002 and 1
In 0003, it is determined to be within the service range when either of the following two conditions is satisfied. The first condition is that the service range information (203) is registered with the processor system identifier indicating the local processor system in which the process that issued the request exists, and is within the service range of the service location server. That is, the identification information is set. The second condition is that the service range information (203) is registered with the processor system identifier indicating the local processor system in which the process that issued the request exists, and the identification information (corresponding to the registered processor identifier ( The backup server information (2) of the other server location information (21002) having the identification information (803) matching with the (402).
2001) is 2.

Next, server division processing in the sixth embodiment will be described. The server division process according to the sixth embodiment is similar to the server division process described in the first embodiment. In the sixth embodiment, the other server location information setting process of step 12006 is performed as follows. For the source server,
The duplication server identification information (803), the server location processor identifier (801) in which the duplication server exists, and the communication information (802) with the duplication server, other server location information (2100 2) of the own service location server.
The backup server information (22001) of the registered duplicate server location information is set to 2 indicating that it is the duplicate server location information of its own service location server. In the case of a duplicate server, the identification information (803) of the duplication source server, the server location processor identifier (802) in which the duplication source server exists, and the communication information (802) with the duplication source server are stored in the other service location server. Server location information (21002)
Registered to the backup server information (22001) of the location information of the registered duplication source server, which is 1 indicating that it is the location information of the duplication source server of its own service location server, and the already registered location information. , The value of the backup server information (22001) of the location information is set to 0.

Next, the setting processing of the location service server information (1014) in the sixth embodiment will be described. Location service server information (1014, 1024,
The contents of (1034, 1044, 1054) are changed when the service location server notifies that it is out of the service range. The process notified by the service location server that it is out of the service range is an information table (20002) of location service server information on its own local processor system.
, The location information included in the out-of-service notification from the service location server is registered in the entry that does not match the contents of the current service location server information (20001), and the valid flag (2000
6) is set, and the entry number of the newly registered location information is set in the current service location server information (20001).

In the sixth embodiment, the configuration and the processing other than those described above are the same as those in the first embodiment. As described above, in the sixth embodiment, the example in which the location information is duplicated by the service location server that is the replication source and the replication service location server has been described. Also, between the replication service location server, the replication source service location server, and the replication source server of the replication source service location server,
In other words, it is possible to triple the location information between the duplicate service location server and the duplicate origin service location server for two generations in the same manner as in the sixth embodiment. Furthermore, it is possible to realize high multiplexing between the duplicated service location server and the duplicated original service location server. As described in the sixth embodiment, by applying the present invention, the location information managed by the service location server can be multiplexed between the service location server that is the duplication source and the duplicated service location server. By multiplexing the location information between the source service location server and the replicated service location server, the client is the source service location server even if the replicated service location server fails. You can get the service of. Further, the client can issue a location information inquiry request for a server that provides a specific service to both the replication service location server and the replication source service location server, and receive the inquiry information notification, whichever comes first.

Further, like the division policy description file described in the third embodiment, a file for describing the multiplicity of location information is provided, and the file is read during the initial setting processing of the service location server, and the multiplicity information ( Two
By setting 1001), the multiplicity of location information can be designated. Furthermore, in the embodiments of the present invention described in the first embodiment, the second embodiment, the third embodiment, and the fourth embodiment, the service location server is replicated according to the service load level of the service location server, and the replication server and the replication source server. In this example, the service range is divided between two servers and the service range is divided. However, by dividing the service range, the location information, which is a global resource that spans between local processor systems, is shared by multiple servers. It will be managed hierarchically. By using the example in which the location information managed by the service location server described in the sixth embodiment is multiplexed between the service location server that is the replication source and the duplicated service location server and the embodiment described in the fourth embodiment, a plurality of The server can perform distributed management of location information having location information of the entire system.

In each of the above-described embodiments, between processes in a client server system constructed on a distributed system in which a plurality of local processor systems having process transfer means for transferring processes between local processor systems are connected. Although the example of the server division method for dividing the service location server managing the location information has been described, it is constructed on a distributed system in which a plurality of local processor systems having a process transfer means for transferring a process between the local processor systems are connected. In the client-server system described above, the invention is also applicable to a server that manages global resources extending between local processor systems.

Further, in the above embodiment, when the service location server process is divided and the duplicate server process is generated on different local processor systems, the duplicate server process is created on the local processor system provided with the duplication source server process. Then, the service range is divided between the replication source and the replication server process, service location information is set for each of the replication source and replication server processes, and then the replication server process is transferred to a different local processor system. However, the replication source server process is duplicated on the memory of a different local processor system to create the replication server process, and then the service range is divided between the replication source and the replication server process. , The information of the divided service range is different from the above. That by transferring to the local processor system, it may be set the service location information to each of the replicated server processes as the source server process.

[0073]

According to the present invention, in a client-server system constructed on a distributed system, when the load on the server process exceeds a predetermined threshold value, the server process is not distributed in advance. , The server can be dynamically divided, and the flexibility of the system can be increased. In addition, it is possible to avoid performance degradation due to an increase in the number of processors. Furthermore, the division policy when dividing the server can be specified. Furthermore, the location information managed by the service location server can be multiplexed between the service location server that is the duplication source and the duplicated service location server.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a client server system to which the present invention is applied.

FIG. 2 is a diagram showing a configuration of a service location server to which the present invention is applied.

FIG. 3 is a diagram showing a structure of service location information.

FIG. 4 is a diagram showing a structure of service range information.

FIG. 5 is a diagram showing a structure of service load information.

FIG. 6 is a diagram showing a structure of division criterion information.

FIG. 7 is a diagram showing a configuration of location service server information.

FIG. 8 is a diagram showing a configuration of other server location information.

FIG. 9 is a diagram showing a flowchart of location information registration processing.

FIG. 10 is a diagram showing a flowchart of location information inquiry processing.

FIG. 11 is a diagram showing a flowchart of load measurement processing.

FIG. 12 is a diagram showing a flowchart of server division processing.

FIG. 13 is a diagram showing a configuration of other service load information.

FIG. 14 is a diagram showing a part of the configuration of another client-server system to which the present invention is applied.

FIG. 15 is a diagram showing the configuration of another service location server to which the present invention has been applied.

FIG. 16 is a diagram showing the structure of a division policy description file.

FIG. 17 is a diagram showing the structure of a division policy description record.

FIG. 18 is a diagram showing a description example of a division policy description file.

FIG. 19 is a diagram showing a structure of division policy information.

FIG. 20 is a diagram showing a configuration of location service server information.

FIG. 21 is a diagram showing the configuration of another service location server.

FIG. 22 is a diagram showing the configuration of another other server location information.

FIG. 23 is a diagram showing the structure of other service location information.

FIG. 24 is a diagram showing a flowchart of another server division process.

FIG. 25 is a diagram showing a flowchart of another server division process.

FIG. 26 is a diagram showing the structure of another division policy description record.

FIG. 27 is a diagram showing the configuration of other division policy information.

FIG. 28 is a diagram showing a flowchart of another location information registration process.

[Explanation of symbols]

1013 service location server 201 service location information 202 server dividing means 203 server range information 205 division standard information 206 service load information 207 load measuring means 208 location information of other service location servers 209 transfer destination processor system selecting means 101, 102, 103, 104, 105 Local processor system 1011, 1021, 1031, 1041, 1051
Servers 1012, 1022, 1032, 1042, 1052
Clients 1015, 1023, 1033, 1043, 1053
Process transfer means 1014, 1024, 1034, 1044, 1054
Location service server information 1016, 1025, 1035, 1045, 1055
Communication interface 14001 External storage device 14002 Division policy description file 14003 Service location server 14004 External storage interface 15001 Division policy information 15002 Division level information 16001-16005 Division policy description record 17001 Division level designation field 17002 Server allocation processor designation field 17003 Priority designation Field 17004 Service range designation field 26001 Division criterion designation field 20002 Location server information table 21001 Multiplicity information

Claims (20)

[Claims] 1. A client-server system constructed on a distributed system in which a plurality of local processor systems having a process transfer means for transferring a process between local processor systems are connected, and at least one of the local processor systems is serviced. A location server process is provided, and the service location server process includes service location information that manages location information of processes registered from other processes, and service range information that defines the service range of the own server process. When the service load exceeds a predetermined load level, a copy of the local server process is created and the service range between the copy source server process and the copy server process is measured. Dividing the, set the service location information to each of the replicated server processes and said replication source server process, and transfer the duplicated server processes on different local processor system, the server divides wherein placing. 2. A client-server system constructed on a distributed system in which a plurality of local processor systems having means for copying a process to another local processor system among the local processor systems are connected to each other. A service location server process is provided in at least one system, and the service location server process defines service location information that manages location information of processes registered from other processes and a service range that defines the service range of its own server process. It has information and measures the service load of the local server process, and when the service load exceeds a predetermined load level, copies the local server process to a different local processor. Generated on the system, divides the service range between the replication source server process (self server process) and the replication server process, transfers the information of the divided service range to the different local processor system, and replicates it. A method for dividing a server, wherein service location information is set for each of the server process and the duplicate server process. 3. The server division method according to claim 1, wherein the replication source server process is set with location information of the replication server process, and the replication server process is configured with location information of the replication source server process. A server partitioning method characterized by the above. 4. The server dividing method according to claim 3, wherein when each of the divided server processes receives a service request from a process outside the service range, it is out of the service range and is different from the own server process. The location information of the server process having the service range is notified to the process that issued the service request, and each of the divided server processes receives an inquiry request regarding the location information of the process outside the service range from the process included in the service range. When receiving a request, the server process receiving the request inquires of another server process having a service range different from the service range of the own server process and notifies the requested process of the requested location information. Server partition method. 5. The server partitioning method according to claim 1, wherein the service load is the number of service executions executed by the service location server process within a fixed time interval. . 6. The server partitioning method according to claim 1, wherein the service range is a set of local processor systems to be serviced. 7. The server division method according to claim 1, wherein when the service load exceeds a predetermined load level, the service range is divided according to a predetermined division policy. Method. 8. The server partitioning method according to claim 7, wherein the partitioning policy selects a fixed number of local processor systems from the smallest processor number of each local processor system among the local processor systems that are the service range. And set the selected local processor system as the service range of the replication source server process,
A server partitioning method, characterized in that the partitioning policy is such that the remaining local processor system, which is the service range, is set as the service range of the duplicate server process. 9. The server partitioning method according to claim 7, wherein the partitioning policy measures the number of service requests for each local processor system included in the service range information, and measures the service for each local processor system. A server division method characterized by a division policy of dividing a service range based on the number of requests. 10. A client-server system constructed on a distributed system in which a plurality of local processor systems having a server process and a process transfer means for transferring a process between local processor systems are connected to each other, wherein the server process is its own server. Equipped with service range information that defines the service range of a process, the service load of the local server process is measured, and when the service load exceeds a specified load level, a copy of the local server process is created, and the copy source server process and the copy A method of dividing a server, characterized in that the service range is divided between a server process and a duplicate server process, which is transferred to a different local processor system and arranged. 11. A client-server system built on a distributed system in which a plurality of local processor systems having a means for a local processor system to duplicate a process to another local processor system among local processor systems having a server process are connected. The server process includes service range information defining the service range of the own server process, measures the service load of the own server process, and when the service load exceeds a predetermined load level, replicates the own server process differently. It is generated on the local processor system, the service range is divided between the replication source server process (local server process) and the replication server process, and the information on the divided service range is transferred to different local processor systems. , The server dividing wherein placing. 12. The server division method according to claim 10 or 11, wherein location information of the replication server process is set in the replication source server process, and location information of the replication source server process is set in the replication server process. A server partitioning method characterized by the above. 13. The server partitioning method according to claim 1, wherein one local processor system in which the service location server is arranged is provided with an external storage device, and the external storage device has a predetermined service load. Provide a division policy description file that describes the policy for dividing the service range when the load level is exceeded, and divide the service range according to the contents of the division policy description file when the service load exceeds the predetermined load level. A server partitioning method characterized by. 14. The server division method according to claim 13, wherein a division level specification field for the service location server to determine to which service location server the division policy description file has a division policy, Consists of a set of partitioning policy description records having a server allocation processor specification field that specifies the local processor system where the subsequent replication server is allocated, and a service range specification field that specifies the service range in which the replication server after the partition is in charge A method for dividing a server, which is characterized by being performed. 15. The method of dividing a server according to claim 14, wherein the division policy description record includes a duplication server after division.
A server partitioning method comprising a partitioning criterion designating field for designating a service load level for determining whether or not to partition the own server. 16. The server partitioning method according to claim 14, wherein when the service load exceeds a predetermined load level in the partitioning policy description record and the service range is partitioned between the replication server and the replication source server, A server partitioning method comprising a priority specification field for designating a priority for determining which partitioning policy is adopted when there are a plurality of partitioning policies that can be adopted. 17. The server partitioning method according to claim 13, wherein when a corresponding partitioning policy does not exist in the partitioning policy description file, the processor number of each local processor system among the local processor systems in the service range is set. A division policy in which a fixed number of local processor systems are selected from the ascending order, the selected local processor system is set as the service range of the replication source server process, and the remaining local processor systems that are the service range are set as the service range of the replication server process. A method of dividing a server, characterized in that 18. The server dividing method according to claim 13, wherein, when a corresponding dividing policy does not exist in the dividing policy description file, a duplicate server is not generated and the service range is not divided. How to divide. 19. The server division method according to claim 1, wherein service location information registered in the replication server process is registered in the replication server process and the replication source service location server, and the replication service is registered. A method of dividing a server, wherein service location information is multiplexed and managed between a location server and the duplication source service location server. 20. The server division method according to claim 1, wherein the service location information registered in the duplicate server process is registered in the duplicate server process and a duplicate origin service location server over a plurality of generations, A method of dividing a server, wherein the service location information is multiplexed and managed among a plurality of service location servers that have registered the service location information.