JP2016110175A - Client device, communication system, and data processing method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce a load on a server device even in a server device-alone structure without reinforcing the server device.SOLUTION: When acquiring a request to a server device 105, a representative client decision unit 208 decides whether a representative client role that is a role of handling a request as a representative is assigned to the own device in relation to the acquired request. If the representative client role for the acquired request is assigned, a representative client processing unit 210 decides whether a response to the acquired request is present in a receiving data management unit 211. If the response is absent from the receiving data management unit 211, a server communication unit 212 transmits the acquired request to the server device 105. If the representative client role for the acquired request is not assigned, an other client communication unit 209 transmits the acquired request to any other client device.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a client / server system.

  In a client / server system in which a server device (hereinafter simply referred to as a server) provides various processing and data to a client device (hereinafter also simply referred to as a client), the number of client devices or requests from the client device to the server device When the number increases, the processing load of the server apparatus increases according to the increase rate.

The following methods exist as methods for distributing and leveling the processing load of the server device.
A load balancer is placed between the server device and the client device, the client device sends a request to the load balancer, and which load balancer processes the request according to a certain standard, which server device handles the request Decide.
Then, the load balancer issues an instruction to deal with the request to the server device that performs processing.
Thereby, the processing load of each server apparatus can be distributed and leveled.

  However, in this prior art, it is necessary to secure a plurality of server devices and a load balancer on the server system side, and there is a problem that the technology cannot be used when the server device constructs a single server system.

As a technique for distributing and leveling the processing load of the server device, for example, there is a method described in Patent Document 1.
In the technique of Patent Document 1, in a system configured with a server device and a plurality of client devices and distributing software images to the client devices, the server device classifies the client devices into a plurality of groups, and a representative client device is assigned to each group. The server device distributes the software image to the representative client device.
The representative client device distributes the software image to client devices belonging to the same group based on requests from other client devices belonging to the same group.

International Publication WO2011 / 135629

However, in the technique of Patent Document 1, it is necessary for the server device to perform group classification of a plurality of client devices and determination of a representative client device.
In situations where client devices frequently join or leave the system, for example, in situations where system participation is performed by wireless communication from a high-speed mobile body, the server device frequently performs group management and representative client device determination. There is a problem that the processing load of the server device increases.

  The main object of the present invention is to reduce the load on the server device without increasing the server device and even when the server device is configured as a single unit. And

A client device according to the present invention provides:
A client device included in a communication system including a plurality of client devices,
If a request to the server device generated by any client device is acquired, whether or not a representative client role that is a role for processing the request on behalf of the acquired acquisition request is assigned to the own device A representative client determination unit for determining;
When a representative client role for the acquisition request is assigned to the own device, it is determined whether a response from the server device corresponding to the acquisition request is cached in a cache area accessible by the own device. A representative client processing unit;
A server communication unit that transmits the acquisition request to the server device when a representative client role for the acquisition request is assigned to the own device and a response corresponding to the acquisition request is not cached in the cache area; ,
And an other client communication unit that transmits the acquisition request to another client device when a representative client role for the acquisition request is not assigned to the own device.

In the present invention, since the request processing is consolidated to the client device to which the representative client role is assigned, it is possible to avoid a situation in which the request is transmitted for each client device and the server device individually returns a response. it can.
Further, when the response is cached in the client device to which the representative client role is assigned, the response can be provided to the client device that has generated the request without transmitting the request to the server device.
Therefore, the load on the server device can be reduced without increasing the number of server devices and even when the server device is configured as a single unit.

FIG. 3 is a diagram illustrating a hardware configuration example of a server device and a client device according to the first embodiment. FIG. 3 is a diagram illustrating a configuration example of an internal processing module of the server device according to the first embodiment. FIG. 3 is a diagram illustrating a configuration example of an internal processing module of the client device according to the first embodiment. FIG. 4 is a flowchart showing an operation example of a client device that is a request generation source according to the first embodiment. FIG. 3 is a flowchart showing an operation example of a client device that is a request transmission destination according to the first embodiment. The figure which shows the example of API call which concerns on Embodiment 1, and the example of a request. FIG. 4 is a flowchart showing an operation example of a request generation unit according to the first embodiment. FIG. 6 is a flowchart showing an operation example of a representative client determination unit according to the first embodiment. FIG. 6 is a flowchart showing an operation example of a representative client determination unit according to the first embodiment. FIG. 3 is a flowchart showing an operation example of a representative client processing unit according to the first embodiment. The figure which shows the example of the request | requirement in the received data management part which concerns on Embodiment 1, and a response preservation | save. FIG. 10 is a diagram illustrating an arrangement example of nodes and keys on a virtual space according to the second embodiment. FIG. 10 is a diagram illustrating an arrangement example of nodes and keys on a virtual space according to the second embodiment. FIG. 10 is a diagram illustrating an operation example of a representative client processing unit according to the second embodiment. The figure which shows the example which couple | bonded the information of the backup representative client which concerns on Embodiment 3 with the HTTP request. The figure which shows the example which couple | bonded the information of the backup representative client which concerns on Embodiment 3 with the HTTP request. FIG. 10 is a flowchart showing an operation example of the server apparatus according to the third embodiment. FIG. 10 is a flowchart showing an operation example of a representative client processing unit according to the fourth embodiment. FIG. 10 is a flowchart showing an operation example of a representative client processing unit according to the fourth embodiment. FIG. 10 is a flowchart showing an operation example of a representative client processing unit according to the fourth embodiment. The figure which shows the example of the list | wrist which manages the request which concerns on Embodiment 4, and a request issuer. The figure which shows the relationship between the representative client which concerns on Embodiment 4, and a request issuer client. The figure which shows the relationship between the representative client which concerns on Embodiment 4, and a request issuer client. The figure which shows the relationship between the representative client which concerns on Embodiment 4, a request issuing client, and a representative client clone. The figure which shows the relationship between the representative client which concerns on Embodiment 4, a request issuing client, and a representative client clone. The figure which shows the relationship between the representative client which concerns on Embodiment 4, a request issuing client, and a representative client clone. The figure which shows the relationship between the representative client which concerns on Embodiment 4, a request issuing client, and a representative client clone. The figure which shows the relationship between the representative client which concerns on Embodiment 4, a request issuing client, and a representative client clone. FIG. 10 is a flowchart showing an operation example of a representative client processing unit according to the fourth embodiment. FIG. 10 is a flowchart showing an operation example of a representative client processing unit according to the fourth embodiment. FIG. 10 is a flowchart showing an operation example of a representative client determination unit according to the fifth embodiment. FIG. 10 is a flowchart showing an operation example of a representative client determination unit according to the fifth embodiment. FIG. 9 is a flowchart showing an operation example of a representative client processing unit according to the fifth embodiment. FIG. 9 is a flowchart showing an operation example of a representative client processing unit according to the fifth embodiment. FIG. 18 is a diagram illustrating an example of a table for managing a logical sum request and a response according to the fifth embodiment. FIG. 10 shows an example of a response according to the fifth embodiment.

Embodiment 1 FIG.
This embodiment will be described below.
In the present embodiment, the load reduction effect of the server device can be obtained without increasing the resources of the server device and even when the server device is configured by itself.
For this reason, in the present embodiment, the server apparatus is not allowed to perform load distribution processing, but the client apparatus is caused to perform load distribution processing.

*** Explanation of configuration ***
FIG. 1 shows a configuration example of a client / server system according to the present embodiment.

As shown in FIG. 1, the client / server system according to the present embodiment includes two or more client devices 101, a client device 102, a client device 103, a client device 104, and one or more server devices 105. Are connected via the network 130.
A system including the client device 101, the client device 102, the client device 103, and the client device 104 is referred to as a communication system 1000.
Note that the client device 101, the client device 102, the client device 103, and the client device 104 have a configuration of four clients as an example to show that there are a plurality of client devices.
The number of client devices may be plural, and a configuration other than four may be used.
Hereinafter, the server device 105 is simply referred to as a server, and the client device 101, the client device 102, the client device 103, and the client device 104 are also simply referred to as a client.

Next, a hardware configuration example of the client device 101, the client device 102, the client device 103, and the client device 104 will be described.
Since the client device 101, the client device 102, the client device 103, and the client device 104 have the same hardware configuration example, the hardware configuration example of the client device 101 will be described below as a representative.
The following description applies to the client device 102, the client device 103, and the client device 104 as well.

  The client apparatus 101 is an information processing apparatus, and includes elements such as a CPU (Central Processing Unit) 110, a memory 111, an input interface 112, an output interface 113, a communication interface 114, a persistent storage interface 115, a disk 116, and the like. They are connected by a bus 117.

The CPU 110 implements an operation as the client apparatus 101 by performing an interrupt and data receiving process from programs and data stored in the memory 111 and various elements.
More specifically, the CPU 110 includes an application logic unit 206, a request generation unit 207, a representative client determination unit 208, another client communication unit 209, a representative client processing unit 210, a received data management unit 211, and a server communication unit 212, which will be described later. By executing a program that realizes the function, processing described later of the client device 101 is performed.
The input interface 112 is an interface to which an input device such as a mouse or a keyboard is connected. The user of the client apparatus 101 operates the client apparatus 101 by operating the input device connected to the input interface 112.
The output interface 113 is an interface to which an output device such as a display or a speaker is connected. The user of the client apparatus 101 perceives the situation of the client apparatus 101 based on a stimulus such as light or sound output to the output device. Is possible.
The communication interface 114 is an interface for the client device 101 to exchange data with the server device 105, the other client device 102, the client device 103, and the client device 104, and is connected to the network 130.
The permanent storage interface 115 is an interface to which a non-volatile storage device such as a disk is connected, and data to be stored and temporary data are stored and read after the power is turned off.

Next, a hardware configuration example of the server device 105 will be described.
The server device 105 is an information processing device equivalent to the client devices 101 to 104.
The server device 105 includes elements such as a CPU 120, a memory 121, an input interface 122, an output interface 123, a communication interface 124, a permanent storage interface 125, and a disk 126, which are connected via a bus 127.

The CPU 120 performs an interrupt and data receiving process from programs and data stored in the memory 121 and various elements, and realizes an operation as the server device 105.
More specifically, the CPU 120 executes a program that realizes the functions of the communication unit 213 and the request processing unit 214, which will be described later, so that the process described later of the server device 105 is performed.
The input interface 122 is an interface to which an input device such as a mouse or a keyboard is connected, and the server administrator operates the server device 105 by operating the input device connected to the input interface 122.
The output interface 123 is an interface to which an output device such as a display or a speaker is connected, and the server administrator can perceive the situation of the server device 105 based on a stimulus such as light or sound output to the output device. It becomes.
The communication interface 124 is an interface for exchanging data with the client device 101, the client device 102, the client device 103, and the client device 104, and is connected to the network 130.
The permanent storage interface 125 is an interface to which a non-volatile storage device such as a disk is connected, and saving and reading of data and temporary data to be stored even after the power is turned off are performed through this interface.

The client device 101, the client device 102, the client device 103, and the client device 104 perform cooperation processing with the server device 105 via the network 130 in order to provide a certain service to the user of the client device.
When the client device 101, the client device 102, the client device 103, and the client device 104 request the server device 105 to perform some processing, the client device 101, the client device 102, the client device 103, and the client device 104 issue a request describing the request content as data to the server device 105.
For example, in a Web system, communication between a client and a server is performed using HTTP (Hyper Text Transfer Protocol), and the request is expressed as an HTTP request method, a URI (Unified Resource Identifier), a body, and the like.
Note that communication between the server and the client is essentially the same even if another protocol such as WebSocket API (Application Programming Interface) is used in addition to HTTP, and there is no difference.

The server device 105 that has received the request transmitted by the client device 101, the client device 102, the client device 103, and the client device 104 analyzes the content of the request and determines what processing should be performed.
When receiving a request using HTTP, generally, the server apparatus 105 uniquely identifies a resource to be processed by a URI, and identifies a specific operation for the resource by a method.
When other parameters required for processing are specified from the client device, the parameters are given as a URI query part or body and transmitted from the client device. Therefore, the server device 105 uses the information (parameters) in a timely manner. Process.
The processing provided by the server device 105 to the client device is diverse, such as transmission of file contents, file rewriting, control of other processes such as a database, and control of various devices connected to the server device 105.

  The internal processing module of the server apparatus 105 according to the present embodiment will be described with reference to FIG.

The server device 105 includes a communication unit 213 for communicating with the client device 101, the client device 102, the client device 103, and the client device 104 existing in the system.
The communication unit 213 can establish a communication connection in parallel with a plurality of client devices.
When receiving a request from the client device, the communication unit 213 transfers the request to the request processing unit 214.

The request processing unit 214 analyzes the request and identifies what kind of processing is requested.
Thereafter, the request processing unit 214 executes the specified request processing and obtains the processing result.
Upon obtaining the processing result, the request processing unit 214 converts the processing result into a format that can be interpreted by the client device, and generates a response.
After generating the response, the request processing unit 214 uses the communication unit 213 to send a response back to the client device that sent the request.

Next, the internal processing module of the client device according to the present embodiment will be described with reference to FIG.
Each client device has an equivalent internal processing module configuration, and all client devices are basically equivalent.
In FIG. 3, only the client device 101 and the client device 102 describe the internal processing module, but the client device 103 and the client device 104 do not have the description because the description is omitted. 103, the client apparatus 104 has an internal processing module equivalent to the client apparatus 101 and the client apparatus 102.
Hereinafter, as a representative, the internal processing module of the client apparatus 101 will be described.
The following description applies to the client device 102, the client device 103, and the client device 104 as well.

The application logic unit 206 describes specific logic of services provided by the client device 101 to the user of the client device 101.
For example, if the service to be provided is a game application that operates in conjunction with the server device 105, software in which data necessary for the game application, game logic, and the like are programmed corresponds to the application logic unit 206.
In addition to this, if the service provides plant monitoring control via the server device 105, the application such as software programmed with the display procedure for displaying the control procedure necessary for monitoring control and data acquired from the plant, etc. This corresponds to the logic unit 206.
The application logic unit 206 may be any application that operates in conjunction with the server device 105.
When the application logic unit 206 needs to be linked to the server device 105, the application logic unit 206 performs a process of transmitting specific processing contents to be linked to the server device 105 to the request generation unit 207.

The request generation unit 207 provides the application logic unit 206 with an API corresponding to the processing content provided by the server device 105 to the client device 101.
The application logic unit 206 calls an API corresponding to the processing content to be linked.
Further, the request generation unit 207 provides parameters to the API as necessary.
Further, the request generation unit 207 performs processing without being aware of the protocol and request data format when the application logic unit 206 communicates with the server device 105, and without being aware of the response data format from the server device 105. Has the role of abstraction layer so that it can be performed.
The request generation unit 207 also serves as an abstraction layer for making the behavior related to the present embodiment independent of the application logic unit 206.
The request generation unit 207 that has generated a request that can be interpreted by the server device 105 based on the processing content corresponding to the called API and the provided parameters, outputs the generated request to the representative client determination unit 208.

When the representative client determination unit 208 acquires a request to the server device 105 generated by one of the client devices, the representative client role, which is a role of processing the request on behalf of the acquired request, is the own device ( It is determined whether it is assigned to the client apparatus 101).
A client device to which a representative client role is assigned is referred to as a representative client device or simply a representative client.
The “processing” in the “processing the request” includes processing for determining whether a response to the request is cached, processing for transmitting a request to the server device 105 when the response is not cached, A process of receiving a response from the server device 105, a process of returning the received response to the request issuer, and a process of caching the received response are included.
As described above, the representative client determination unit 208 may acquire a request generated by the request generation unit 207 of its own device, or acquire a request generated by the request generation unit 207 of another client device. There is also.
A request generated by the request generation unit 207 of the own device and acquired by the representative client determination unit 208, a request generated by the request generation unit 207 of another client device and acquired by the representative client determination unit 208 Both requests will be referred to as acquisition requests below.
The representative client determination unit 208 determines whether or not a representative client role for an acquisition request is assigned to the own device (client device 101) using a distributed hash table shared among a plurality of client devices. .
The representative client determination unit 208 outputs the acquisition request to the representative client processing unit 210 when the representative client role is assigned to the own device (client device 101) for the acquisition request.
On the other hand, when the representative client role is not assigned to the own device (client device 101) for the acquisition request, the representative client determination unit 208 outputs the acquisition request to the other client communication unit 209.

As described above, when the representative client role for the acquisition request is not assigned to the own device (client device 101), the acquisition request is output from the representative client determination unit 208, and the other client communication unit 209 outputs the acquisition request. Send to other client devices.
In addition, the acquisition request is a request generated by another client device, a representative client role for the acquisition request is assigned to the own device (client device 101), and a response corresponding to the acquisition request is cached in the cache area. The other client communication unit 209 transmits the response cached in the received data management unit 211 to the client device that has generated the acquisition request.
The other client communication unit 209 provides an API (Application Programming Interface) for communicating with other client devices using the communication interface 114.
For example, when the system according to the present embodiment is constructed as a Web system and the client apparatus 101 is mounted on a Web browser, an API such as WebRTC is provided.
When a data transmission request is notified via the API, the other client communication unit 209 transmits data to the client apparatus as a transmission destination according to the request.
Further, when some data is transmitted from another client device, the client device 101 is notified in the form of an event or the like that the data has been transmitted from another client device, and the event is monitored. The received module will process the received data.

The representative client processing unit 210 performs an operation for acting as a representative client when a representative client role is assigned to the own apparatus (client apparatus 101).
Specifically, in the representative client processing unit 210, the acquisition request is a request generated by the request generation unit 207, the representative client role (client device 101) for the acquisition request is assigned to the own device, and the received data When the response from the server device 105 corresponding to the acquisition request is cached in the management unit 211, the response cached in the received data management unit 211 is passed through the representative client determination unit 208. To the request generation unit 207.
Further, the acquisition request is a request generated by the request generation unit 207 of another client device, and the representative client role (client device 101) for the acquisition request is assigned to the own device, and is acquired by the received data management unit 211. When the response from the server device 105 corresponding to the request is cached, the response cached in the received data management unit 211 is output to the other client communication unit 209.

The received data management unit 211 has a cache area for storing a response from the server apparatus 105, and manages the cache area.
The reception data management unit 211 stores the response acquired by the representative client processing unit 210 from the server device 105 in the cache area in association with the corresponding request.
Further, the reception data management unit 211 can acquire a corresponding response with a certain request as a key.
In this embodiment, the cache area is built in the client apparatus 101. However, the cache area only needs to be accessible by the client apparatus 101 (more precisely, the representative client processing unit 210), and is built in the client apparatus 101. It does not have to be.
Hereinafter, the expression that the response is cached in the reception data management unit 211 is also used, which means that the response is cached in the cache area managed by the reception data management unit 211.

When the representative client role for the acquisition request is assigned to the own device (client device 101) and the response corresponding to the acquisition request is not cached in the received data management unit 211, the server communication unit 212 outputs the acquisition request. It transmits to the server device 105.
The server communication unit 212 provides an API for communicating with the server device 105 using the communication interface 114.
For example, when the system according to the present embodiment is constructed as a Web system and the client apparatus 101 is mounted on a Web browser, an XMLHttpRequest object, a WebSocket API, or the like corresponds to the server communication unit 212.
Note that the XMLHttpRequest object is used when the communication protocol between the server device and the client device is HTTP. However, even if the WebSocket API is used, a function equivalent to HTTP can be realized. Even if both are used together, there is no difference.

*** Explanation of operation ***
Next, an outline of a data processing method in the client device according to the present embodiment will be described.

FIG. 4 shows an operation example of the client device that is the request generation source.
Hereinafter, the description will be given using an example in which the client apparatus 101 is a request generation source.

The request generation unit 207 of the client device 101 generates a request to the server device 105 based on an instruction from the application logic unit 206 (processing 101).
Then, the request generation unit 207 outputs the generated request to the representative client determination unit 208.

The representative client determination unit 208 determines whether or not the representative client role for the request (acquisition request) is assigned to the own device (client device 101) (processing 102).
This process 102 corresponds to a representative client determination step.

If the representative client role is assigned to the own device for the request (YES in processing 102), the representative client determining unit 208 outputs the request to the representative client processing unit 210, and the representative client processing unit 210 receives the server communication unit 212. It is determined whether or not the response corresponding to the request is cached (process 103).
This processing 103 corresponds to a representative client processing step.

  When the response is cached in the reception data management unit 211 (YES in processing 103), the representative client processing unit 210 returns the response cached in the reception data management unit 211 to the request generation unit 207, The generation unit 207 returns the response to the application logic unit 206, and the application logic unit 206 performs processing using the response (processing 104).

On the other hand, if the response is not cached in the received data management unit 211 (NO in processing 103), the representative client processing unit 210 outputs a request to the server communication unit 212, and the server communication unit 212 sends a request to the server device 105. Transmit (process 105).
This process 105 corresponds to a server communication step.

The server communication unit 212 receives a response from the server device 105 (processing 106).
Thereafter, the server communication unit 212 returns a response from the server device 105 to the request generation unit 207 via the representative client processing unit 210 and the representative client determination unit 208, and the request generation unit 207 returns a response to the application logic unit 206. The request generation unit 207 performs processing using the response (processing 104).

On the other hand, when the representative client role is not assigned to the own device for the request (NO in process 102), the representative client determination unit 208 outputs the request to the other client communication unit 209, and the other client communication unit 209 transmits the request to the other device. To the client device (for example, the client device 102) (process 107).
This process 107 corresponds to another client communication step.

The other client communication unit 209 receives a response from another client device (for example, the client device 102) (process 108).
Thereafter, the other client communication unit 209 outputs the received response to the representative client processing unit 210.

  The representative client processing unit 210 determines whether or not the response output from the other client communication unit 209 is a response addressed to the own device (client device 101) (processing 109).

  If the response is for the own device (client device 101) (YES in processing 109), the representative client processing unit 210 returns the response to the request generation unit 207 via the representative client determination unit 208, and generates a request. The unit 207 returns the response to the application logic unit 206, and the application logic unit 206 performs processing using the response (processing 104).

On the other hand, if the response is not addressed to the own device (client device 101) (NO in processing 109), the representative client determination unit 208 outputs the response to the other client communication unit 209, and the other client communication unit 209 outputs the response. Transfer to another client device (for example, client device 103) (process 110).
The result of NO in the process 109 is that the request generated by the request generation unit 207 is transmitted to another client device (for example, the client device 102) and is generated by another client device (for example, the client device 103). This is a case where a request is transmitted to another client device (client device 102), and a response to a request generated by another client device (client device 103) is received from another client device (client device 102).

FIG. 5 shows an operation example of the client device to which the request is transmitted.
Here, the description will be given using an example in which the client device 102 receives a request generated by the client device 101.

In the client apparatus 102, the other client communication unit 209 receives a request transmitted from another client apparatus (client apparatus 101) (process 201).
The other client communication unit 209 outputs the received request to the representative client determination unit 208.

  The representative client determination unit 208 determines whether or not the representative client role for the request (acquisition request) is assigned to the own device (client device 102) (process 202).

  When the representative client role is assigned to the own device for the request (YES in processing 202), the representative client determining unit 208 outputs the request to the representative client processing unit 210, and the representative client processing unit 210 receives the received data. It is determined whether or not the response corresponding to the request is cached in the management unit 211 (process 203).

  If the response is cached in the received data management unit 211 (YES in processing 203), the representative client processing unit 210 sends another response cached in the received data management unit 211 via the representative client determination unit 208. The response is returned to the client communication unit 209, and the other client communication unit 209 returns the response to the other client device (client device 101) (process 204).

  On the other hand, if the response is not cached in the received data management unit 211 (NO in processing 203), the representative client processing unit 210 outputs a request to the server communication unit 212, and the server communication unit 212 sends a request to the server device 105. Transmit (process 205).

The server communication unit 212 receives a response from the server device 105 (processing 206).
Thereafter, the server communication unit 212 returns the response from the server device 105 to the other client communication unit 209 via the representative client processing unit 210 and the representative client determination unit 208, and the other client communication unit 209 returns the response to the other client device. A reply is made to (client apparatus 101) (process 204).

  On the other hand, when the representative client role is not assigned to the own device for the request (NO in process 202), the representative client determination unit 208 outputs the request to the other client communication unit 209, and the other client communication unit 209 The request is transmitted to another client device (for example, the client device 104) (process 207).

The other client communication unit 209 receives a response from another client device (client device 104) (processing 208).
Thereafter, the other client communication unit 209 returns the response to the other client device (client device 101) (processing 204).

Next, the operation of the client device according to the present embodiment will be described in detail.
First, processing performed by the request generation unit 207 after calling the API will be described with reference to FIG.

  In process 401, the request generation unit 207 generates a request that can be interpreted by the server apparatus 105 based on the processing content corresponding to the called API and the provided parameters.

FIG. 6 shows one specific example of the process 401.
FIG. 6 is a correspondence table showing an example of the generated request corresponding to the description on the code when calling the API.
Here, assuming that the service to be provided by the application logic unit 206 is plant monitoring control, an API for acquiring a plant monitoring time interval, an API for changing the monitoring time interval, and screen definition data to be displayed on the display are acquired. It is assumed that there is an API, and each API name is “getMonitorInterval”, “setMonitorInterval”, and “getPageContent”.
Further, it is assumed that the server apparatus 105 and the client apparatus communicate with each other using HTTP.
When acquiring the monitoring time interval, the application logic unit 206 calls “getMonitorInterval” without an argument.
The called request generation unit 207 generates an HTTP request configured by “http: // xxx / api / monitor_interval /” which is a URI for uniquely specifying a time interval and a method GET meaning data acquisition.
Further, when changing the monitoring time interval, the application logic unit 206 calls “setMonitorInterval” by giving the time interval set by the argument.
The called request generation unit 207 has a value designated as an argument as a query for “http: // xxx / api / monitor_interval /”, which is a URI that uniquely specifies a time interval and a method POST that means data transmission. Is generated, and an HTTP request composed of these URLs is generated.
In addition, when acquiring the screen definition data, the application logic unit 206 calls the “getPagetContent” by giving the screen name as an argument.
The called request generation unit 207 is an HTTP configured by “http: //xxx/html/Monitor.html”, which is a URI for uniquely specifying the definition data of the requested screen and the method GET meaning data acquisition. Generate a request.
In this example, it is assumed that the screen definition data is a file described in HTML (HyperText Markup Language), and is received as an argument between “http: // xxx / html /” and “.html”. A URI is constructed by inserting a screen name.

Returning to FIG. 7, after generating the request in process 401, the request generation unit 207 transmits the request to the representative client determination unit 208 in process 402, and the response is returned from the representative client determination unit 208 in process 403. Wait until.
When the response is returned from the representative client determination unit 208, the request generation unit 207 converts the returned response into a format that can be interpreted by the application logic unit 206 in processing 404.
In step 405, the request generation unit 207 returns the conversion result to the application logic unit 206 to complete the processing.

  As an example of the response conversion process in the process 404, when the response is data in JSON (Java Script (registered trademark) Object Notation) format, it may be converted into an object that can be interpreted by Java Script (registered trademark).

  The processing of the representative client determination unit 208 that has received the request transfer will be described with reference to FIG.

  In step 501, the representative client determination unit 208 that has received the request transfer calculates the ID of the request using the distributed hash table.

The distributed hash table is a technique for managing a hash table distributed by a plurality of peers. In this embodiment, peers are interpreted in association with client devices.
The distributed hash table is functionally equivalent to the hash table except that it is managed in a distributed manner, and a certain value is projected onto a point in a certain space using a hash function, and the point is associated with the value as an ID.
By performing this association, it is possible to retrieve a value from the ID.
Input values to the hash function may be values of different scales, and can also be used to create a relationship between the values of the respective scales.
Therefore, it may be used to create a relationship between a file and a node using a distributed hash table in order to determine which file is managed by which node in the distributed file system.
In the present embodiment, the network ID serving as identification information at the time of communication between the request and the client device is treated as an input of a hash function, and for which request depending on the relationship between the ID of each request and the ID of each client device, A distributed hash table is used to determine which client device is a representative client device that is a client device issued to a server (which client device is assigned a representative client role).
A URI can be considered as a specific example of the ID of the request.
Further, as a specific example of the network ID of the client device, an IP (Internet Protocol) address or a MAC (Media Access Control) address of the client device can be considered.
In the distributed hash table, it is theoretically possible to assign a representative client role to each client device in a probabilistic and uniform manner by selecting a hash function. Therefore, a large number of representative client roles are assigned to a specific client device. It is possible to avoid uneven processing load among client devices.

Returning to FIG. 8, in processing 502, the representative client determination unit 208 uses the request ID, the ID on the distributed hash table for the client device, and the distributed hash table itself, and the client device itself sends the request to the server device. It is determined whether the client device to be issued is a representative client device.
That is, it is determined whether a representative client role is assigned to the client device for the request.
If the client device itself is a representative client device, the processing branches to processing 503. If another client device that is not the client device is a representative client device, the processing branches to processing 505.

In processing 503, since the client device itself becomes the representative client device, the representative client determination unit 208 transfers the request to the representative client processing unit 210, and a response corresponding to the request is returned from the representative client processing unit 210 in processing 504. Wait until
Thereafter, when a response is returned, the process proceeds to processing 507.

On the other hand, in process 505, since the client device other than the client device becomes the representative client device, the representative client determination unit 208 sends the request to the other client device from the other client communication unit 209 according to the algorithm of the distributed hash table. Forward.
At this time, in addition to the request, it is possible to determine to which client device the response corresponding to the request is finally returned, such as an ID calculated according to the distributed hash table of the client device itself or the network ID of the client device itself Add information and transfer.

After transferring the request, in processing 506, the representative client determination unit 208 waits until a response corresponding to the request transferred by the client device is returned from the representative client device.
If a response is returned from the representative client device, the process proceeds to processing 507.

In processing 507, the representative client determination unit 208 determines whether the client that issued the request corresponding to the response is the client device that is currently performing the processing.
As a result, if it is the client device itself, the process proceeds to processing 508, and a response from the server device 105 or the representative client is returned to the request generation unit 207.
Conversely, if the result is not the client device itself, the representative client determination unit 208 transfers a response to another client device in accordance with the distributed hash table toward the client device that issued the request in processing 509.
For example, if the client device performing the processing of FIG. 8 is the client device 101, the representative client device is the client device 103, and the request issuing client device is the client device 104, the representative client of the client device 101 The determination unit 208 transfers the response to the client device 102.

Next, processing of a client device that is a transfer destination of a request when the request is transferred to another client device in processing 505 will be described with reference to FIG.
FIG. 9 is a processing flow of the representative client determination unit 208 of the client device that is the transfer destination of the request.

The client device that has received the request transfer by the other client communication unit 209 transfers the request to the representative client determination unit 208.
In step 601, the representative client determination unit 208 that has received the request transfer calculates the ID of the request using the distributed hash table.

In process 602, the representative client determination unit 208 determines whether the client device itself becomes a representative client device by using the request ID, the ID on the distributed hash table for the client device itself, and the distributed hash table itself.
If the client apparatus itself is a representative client apparatus, the process branches to process 603, and if another client apparatus that is not the client apparatus is a representative client apparatus, the process branches to process 605.

  In process 603, since the client apparatus itself becomes the representative client apparatus for the request being processed, the representative client determination unit 208 transfers the request to the representative client processing unit 210, and in process 604, the request is sent from the representative client processing unit 210. Wait until a response corresponding to is returned.

Thereafter, when a response is returned, the process proceeds to processing 606, where the representative client determination unit 208 follows the distributed hash table toward the client device that issued the request, and transmits to the other client device via the other client communication unit 209. Forward the response.
However, if the network ID of the request issuing client device is known at this time, the response may be returned directly to the request issuing client device without following the distributed hash table.

In process 605, since the client apparatus itself is not a representative client apparatus for the request being processed, the representative client determination unit 208 transfers a response to another client apparatus according to the distributed hash table.
When the request is transferred to another client device in the processing 605, the processing of the client device that is the transfer destination of the request is performed as described in FIG. 9, and reaches the representative client device corresponding to the request in the system. Until then, the process is recursively repeated while walking around the client device.

  Next, the processing of the representative client processing unit 210 will be described with reference to FIG.

In step 701, the representative client processing unit 210 that has received the request determines whether the request is safe.
That the request is safe means that no change occurs in the server device 105 when the server device 105 processes the request.
For example, in HTTP, a request using the GET method is a method for acquiring a specified resource and can be said to be a safe request.
On the other hand, a request using a method such as POST or PUT is an unsafe request because there is a possibility that a change occurs in the server device 105.
If the request is safe, the process proceeds to process 702; otherwise, the process proceeds to process 703.

In the case of a secure request, since no change occurs in the server device 105, if a response corresponding to the request currently being processed is stored in the client device that is the representative client device, There is a possibility that the response can be diverted.
Therefore, in process 702, the representative client processing unit 210 determines whether a response corresponding to the request currently being processed is stored in the received data management unit 211.
If the corresponding response is stored in the received data management unit 211, the process proceeds to the process 706, and if not stored, the process proceeds to the process 703.

In process 706, it is confirmed that the response corresponding to the received data management unit 211 is stored.
However, it is in a state where it is not determined whether there is no problem by reusing the response. Therefore, it is determined in step 706 whether the response has expired.
If the expiration date is HTTP, it can be determined from the response header using the Expres response header or the max-age of the Cache-Control response header.
In the Expires response header, the expiration date of the response is described as a value.
If the expiration date is past compared to the current date and time, it can be determined that the response has expired.
The max-age of the Cache-Control response header describes a time when the response can be cached as a value.
The result of adding the time of max-age to the response reception time is compared with the current date and time, and if it is in the past, it can be determined that the expiration date of the response has expired.

If it is determined in process 706 that the expiration date has expired, the process proceeds to process 703, and if it is determined that the expiration date has not expired, the process proceeds to process 707.
When it is determined that the expiration date has expired, the representative client processing unit 210 may cause the received data management unit 211 to delete the entry corresponding to the request.
If the data expiration date does not exist, all responses are not changed, or all responses are constantly updated and cannot be diverted, the processing 706 is not necessary. I do not care.
In this case, for example, when all responses are not changed, the process transition to the process 706 is replaced with the process transition to the process 707.
If all responses are constantly updated, the process transition to process 706 is replaced with the process transition to process 703.

The process 707 is performed in a state where it can be confirmed that the request currently being processed is safe and has not expired.
In this process, the representative client processing unit 210 acquires a response corresponding to the request currently being processed from the received data management unit 211.
After the acquisition is completed, the process proceeds to process 708.

The process 703 is performed in a state in which it is confirmed that even if the client apparatus as the representative client apparatus holds a response corresponding to the request currently being processed, it cannot be reused.
In this processing, processing for issuing a request currently being processed to the server device 105 via the server communication unit 212 is performed.
After the request issuance is completed, the process proceeds to process 704.

A process 704 is a process of waiting until a response from the server corresponding to the issued request is returned.
Transition to processing 705 is triggered by a response returned from the server apparatus 105.

In process 705, if the request corresponding to the response received from the server device 105 is safe, the representative client processing unit 210 associates the response with the request and stores it in the received data management unit 211.
Since the received data management unit 211 stores the request and its response in association with each other, it is possible to check in step 702 whether a response exists using the request as a key.
After saving is completed, the process proceeds to processing 708.

In processing 708, the representative client processing unit 210 determines whether the client device that is the source of the request corresponding to the response acquired from the server device 105 or the received data management unit 211 is itself.
As a result of the determination, if the request source is the client device itself, the process proceeds to processing 709 and a response is returned to the representative client determination unit 208.
When the request issue source is not the client device itself, the process proceeds to processing 710, and the representative client processing unit 210 follows the distributed hash table toward the client device that issued the request, and transmits the other client via the other client communication unit 209. Transfer the response to the device.

FIG. 11 shows a specific example of request and response storage in the received data management unit 211.
In this example, a request and response according to HTTP are assumed.
In the request, the start line, header, and body of the HTTP request are stored as a request, and the start line, header, and body of the HTTP response are similarly stored for the corresponding response.
In this example, the request and response on the same line are expressed as being associated with each other.
In other words, when searching for a response using a request as a key, first search for a request that matches the key request in the request column. If there is a response, the response on the same line as that request should be searched. Become.
Further, when comparing a request that is a key and a request in a request string, if all of the start line, header, and body are completely matched, it is assumed that the requests are the same.
However, for simplification, it is also possible to compare only the URIs in the start line.

*** Explanation of effects ***
By performing the client processing in the present embodiment as described above, in a situation where a plurality of client devices need to issue the same request to the server device 105, all of the plurality of client devices independently issue the same request. In this case, the representative client device corresponding to the request is uniquely determined, and only the representative client device can issue the request to the server device 105.
At this time, it is important that the server device 105 is not involved in the process of determining only the representative client device to issue a request, and the processing load for determining the representative client device does not occur as a new load on the server device 105. Is one effect.

In addition, when determining the representative client device, it is possible to implement a fair representative client device determination algorithm using a distributed hash table, so there are many depending on the timing at which the client device participates in the client / server system. Another effect is that the unfairness in which the representative client device for a request is determined to be biased to one client device does not theoretically occur.
One reason for the unfairness in which the representative client device is biased depending on the timing of participation is to search for a client device that holds a response corresponding to the request between the client devices before issuing the request, and to hold the client device In the case where there is not, there is a method in which the client device that issues the request becomes the representative client device.
In this method, when the time width between the participation time of the client device that first participates in the system and the participation time of the client device that participates next is large, the client device that first participates sends many requests to the server device 105. Inevitably, the client device that participates first is inevitably determined as the representative client device of many requests.
However, in the method according to the present embodiment using the distributed hash table, even in the same situation, when the second participating client device occurs, the function of the distributed hash table is used to re-represent the representative client device. Allocation will be done.
Therefore, even if the representative client device corresponding to many requests is determined for the first client device to join, the number of representative client devices can be leveled by the participation of the second client device. It becomes.
One of the specific distributed hash table algorithms used to obtain such an effect is Chord.
Chord can maintain and manage the distributed hash table even if the number of client devices participating in the network increases or decreases dynamically.

Embodiment 2. FIG.
Hereinafter, a specific example in the case where Chord is used as a distributed hash table in the system configuration shown in the first embodiment will be described.
In the present embodiment, a representative client role takeover operation when a new client device enters the communication system 1000 and a representative client role takeover operation when the client device leaves the communication system 1000 will be described.

*** Explanation of configuration ***
A configuration example of the server apparatus 105 according to the present embodiment is as shown in FIG. 2, and a configuration example of each client apparatus according to the present embodiment is as shown in FIG.
In the present embodiment, differences from the first embodiment will be mainly described.
Matters not described in the present embodiment are the same as those in the first embodiment.

When a new client device participates in the communication system 1000, the other client communication unit 209 operates as follows in the client device that takes over the representative client role to the new client device.
When the response corresponding to the request targeted by the representative client role assigned to the new client device is cached in the received data management unit 211, the other client communication unit 209 is cached in the received data management unit 211. The response is transmitted to the new client device.
The new client device stores the transmitted response.

When the client device leaves the communication system 1000, the other client communication unit 209 operates as follows in the remaining client device.
When the response corresponding to the request targeted by the representative client role assigned to the leaving client device leaving the communication system is cached in the leaving client device, the other client communication unit 209 displays the cached response. Received from the leaving client device.
In the remaining client device, the received data management unit 211 stores the response received from the leaving client device.

*** Explanation of operation ***
Next, an operation example of the client device according to the present embodiment will be described with reference to FIGS.

FIG. 12 shows a virtual space composed of client devices and requests in this system using Chord.
In Chord, nodes and keys are arranged in the same virtual space.
The node corresponds to the client device of the first embodiment, and the key corresponds to the request of the first embodiment.
That is, the nodes 902, 903, 904, and 905 in FIG. 12 correspond to any of the client devices 101, 102, 103, and 104 participating in the communication system 1000.
Determination of the arrangement of the nodes and keys in the virtual space uses a hash function, assigns an ID to each, and arranges the ID at the position of the ID.
FIG. 12 illustrates an example in which a plurality of nodes and keys are arranged in the virtual space.
Nodes 902, 903, 904, and 905 and keys 906, 907, 908, 909, 910, 911, and 912 are arranged on the virtual space 901 at the ID positions obtained by the hash functions, respectively.
In Chord, a node manages a key and data corresponding to the key.
A method for determining which node manages which key and data corresponding to the key is as follows.
The node manages a key existing between a node and a node adjacent to the node clockwise in the virtual space and data corresponding to the key.
For example, keys 906 and 907 exist between nodes 902 that are clockwise and adjacent to node 903.
Therefore, the node 903 manages keys 906 and 907 and data corresponding to them.
Similarly, the node 904 manages keys 908, 909, and 910 and data corresponding to them.

Based on FIG. 12, the relationship between the representative client device and the request in this embodiment will be described.
As described above, the node corresponds to the client device, and the key corresponds to the request.
In Chord, a key and data corresponding to the key are managed by the node. As data corresponding to the key, at least the following is supported in the present embodiment.
-Response acquired by request-List of client devices from which requests are issued It means that the client device corresponding to the node 904 is the representative client device of the request corresponding to the keys 908, 909, 910.
Therefore, in order for a client device corresponding to a node other than the node 904 to obtain a response to a request corresponding to the key 908, for example, the request is transferred to the client device corresponding to the node 904, and the client device corresponding to the node 904 Get a response from

  Next, the behavior when a new client device joins the communication system 1000 will be described with reference to FIG.

FIG. 13 shows a situation where a client device corresponding to the node 1001 newly participates in the virtual space in the state of FIG.
In this case, the client device corresponding to the node 1001 must be the representative client device of the request corresponding to the key 906.
In Chord, when a new participating node joins the virtual space, it communicates with nodes before and after the new participating node to update the state.
At that time, the client device corresponding to the newly participating node receives the request to be responsible for the representative client device and its related data (response, list of client devices that issued the request) from the representative client device before participation.
The representative client device before participation of the client device corresponding to the node 1001 is a client device corresponding to the node 903.
For this reason, the client device corresponding to the node 1001 needs to take over the request and related data corresponding to the key 906 from the client device corresponding to the node 903.
At this time, when taking over the request and related data from the takeover source client device to the takeover destination client device, if the takeover source client device is receiving a response from the server device 105, the takeover processing is performed after the reception is completed. To start.
By waiting for reception completion, it is possible to avoid taking over an incomplete response.

Next, the processing when the client device is detached will be described by taking as an example the behavior when the client device corresponding to the node 1001 leaves the communication system 1000 from the state of FIG.
Note that when this separation is completed, it means returning to the state of FIG.

When the client device corresponding to the node 1001 leaves, it is necessary to continue the representative client of the request corresponding to the key 906 for which the client device was responsible for the representative client device to the client device remaining in the system.
When Chord is used, the client device that takes over the representative client is a client device corresponding to the node 903.
Therefore, the client corresponding to the node 1001 takes over the request and related data, which is responsible for the representative client device, to the client device corresponding to the node 903 before leaving.
At this time, if a response is received from the server device 105 as in the case of new participation, the client device corresponding to the node 903 is taken over after waiting for the completion of reception of the response.

*** Explanation of effects ***
According to the present embodiment, even when a client device joins or leaves the system, an appropriate representative client device can be taken over.

Embodiment 3 FIG.
In the present embodiment, a procedure related to detachment of a client device from the communication system 1000 will be further described.

The system detachment of a certain client device can be broadly divided into a detachment planned as a client device and an unscheduled detachment mainly due to forced termination due to unauthorized processing.
In the case of planned departure, it is possible to appropriately take over the representative client role by using the distributed hash table in the procedure described above in the second embodiment, but in the case of unplanned departure, the representative client role Cannot be taken over by the above procedure.
In the present embodiment, a procedure corresponding to unscheduled detachment of a client device will be described.

*** Explanation of configuration ***
A configuration example of the server apparatus 105 according to the present embodiment is as shown in FIG. 2, and a configuration example of each client apparatus according to the present embodiment is as shown in FIG.
In the present embodiment, differences from the first and second embodiments will be mainly described.
Matters not described in the present embodiment are the same as those in the first and second embodiments.

In the present embodiment, the representative client processing unit 210 replaces a spare representative client device (hereinafter also simply referred to as a spare representative client) that operates on behalf of the own device after the own device leaves the communication system 1000 with another client device. The identifier of the backup representative client device is added to the request transmitted to the server device 105.
Then, the server communication unit 212 transmits a request to which the identifier of the backup representative client device has been added by the representative client processing unit 210 to the server device 105.
In the server device 105, the communication unit 213 transmits a response to the backup representative client device when the response cannot be transmitted to the request issuing client device.

*** Explanation of operation ***
Next, an operation example of the representative client processing unit 210 according to the present embodiment will be described with reference to FIG.
FIG. 14 is a processing flow diagram in which the processing of the representative client processing unit 210 shown in FIG. 10 is expanded, and processing 1101 and 1102 are added to FIG.
Below, only the part relevant to an additional process is demonstrated.

A process 1101 is a process that is first processed when it is determined that a request is transmitted to the server apparatus 105 as a representative client apparatus.
For this reason, it is found that the response is not stored in the received data management unit 211 in the process 702, and that the corresponding response stored in the received data management unit 211 is expired in the process 706. In this case, the processing transits to processing 1101.

In process 1101, when the client apparatus performing this process is forced to leave without waiting for a response reply from the server apparatus 105 in process 704, the client apparatus takes over the process of the representative client being processed. The preliminary representative client device is searched.
Specifically, in the distributed hash table at that time, when the representative client device leaves, the client device that takes over the processing of all requests and responses handled by the representative client device is determined as the backup representative client device.
If Chord is used as the distributed hash table, the client device corresponds to the node adjacent to the representative client in the clockwise direction.

  Next, in processing 1102, the representative client processing unit 210 combines the information of the spare representative client determined in processing 1101 with the request.

FIG. 15 shows an example in which the information of the backup representative client is combined with the HTTP request.
In FIG. 15, information on the backup representative client is combined as an HTTP header.
Specifically, an X-Representative-Client-Address header and an X-Requester-Client-Address-List header.
The network ID of the backup representative client is set as a value in the X-Representative-Client-Address header, and the network ID list of the request source client of this request is set as the value in the X-Requester-Client-Address-List header. doing.
Note that the header name is an example, and even if the header name is different, there is no essential difference.

FIG. 16 shows another example in which the information of the backup representative client is combined with the HTTP request.
In FIG. 16, the information of the backup representative client is combined as an HTTP request body.
Specifically, information on the backup representative client is expressed in the JSON format, the network ID of the backup representative client is set as a value in the Representative property, and a list of the network IDs of the request source client of this request is set in the Requests property. It is set as.
The spare representative client need not be unique, and assuming that the spare representative client is also disconnected, a plurality of spare representative clients including the spare representative client ahead may be transmitted to the server device 105. I do not care.
In that case, the network ID of the backup representative client is transmitted to the server device 105 as a list.

Next, server processing using information of the backup representative client will be described with reference to FIG.
In order to simplify the description, it is assumed that the processing flow of this server operates with a single thread, but a server that operates with a multi-thread operation may be used.

In process 1401, the server apparatus 105 waits for a request to be transmitted from the client apparatus.
When a request is transmitted from the client device, the process proceeds to processing 1402.

  In processing 1402, the communication unit 213 receives the transmitted request and expands it on the memory.

  In processing 1403, the request processing unit 214 analyzes the request developed on the memory and makes the contents interpretable.

In processing 1404, the request processing unit 214 performs processing according to the contents described in the request.
Since the processing content depends on the request, it is not mentioned here.

  In processing 1405, the request processing unit 214 generates, as a response, contents to be notified to the client as a result of processing in processing 1404.

  In processing 1406, the communication unit 213 transmits the response generated in processing 1405 to the client device that has transmitted the request.

In processing 1407, the request processing unit 214 determines whether or not the transmission processing in processing 1406 has been normally completed. If the transmission processing is completed normally, the request processing unit 214 proceeds to processing 1401 and prepares for request transmission from another client device. Wait.
On the other hand, if it is not completed normally, the process proceeds to process 1408.

  In processing 1408, since the response could not be normally returned to the client device, the request processing unit 214 in the server device 105 obtains one of the information of the unprocessed backup representative client from the result of analyzing the request in processing 1403. Try.

In processing 1409, the request processing unit 214 determines whether or not the information of the unprocessed backup representative client has been successfully acquired as a result of processing 1408.
If the acquisition is successful, the process proceeds to process 1410. If the acquisition is unsuccessful, there is no information on the spare representative client any more, so the current process is terminated as an error and the result is discarded, and the process proceeds to process 1401.

In processing 1410, the communication unit 213 acquires the network ID of the standby representative client described in the information of the unprocessed standby representative client acquired in processing 1408, and transmits a response to the partner of the network ID. Try.
In this transmission, since a protocol starting from a request from a client cannot be used like HTTP, a protocol that can start from a server such as WebSocket or Server-Sent Events is used.

In processing 1411, the request processing unit 214 determines whether or not the transmission processing in processing 1410 has been completed normally. If the processing is completed normally, the request processing unit 214 transitions to processing 1401 to prepare for request transmission from another client device. Wait.
On the other hand, if not completed normally, the process proceeds to process 1408, and the transmission process using information of the remaining unprocessed backup representative client is repeated.

*** Explanation of effects ***
According to the present embodiment, even if the unplanned departure occurs in the representative client device, the spare representative client device can reliably return a response to the request issuer.

Embodiment 4 FIG.
In a client / server system, a server device does not always immediately return a response to a request from a client device.
For example, when server processing corresponding to a request takes a very long time, it is assumed that the response return timing is delayed depending on the convenience of the server device. In this case, a response is returned to the client device over time. It will be.
In a system where multicast delivery is not possible when the response from the server device is delayed in the situation where the same request is generated from multiple client devices, the representative client device of the request receives all the request issuers after receiving the response. Responses must be sent back to the client devices in sequence.
In such a case, the reply load of the representative client device is concentrated at one time.
Also, there is a difference in the timing of receiving a response between the client device that issued the response first and the client device that issued the request last.
This difference increases as the number of client devices that issue requests increases.
The present embodiment has been made in view of these circumstances.

*** Explanation of configuration ***
A configuration example of the server apparatus 105 according to the present embodiment is as shown in FIG. 2, and a configuration example of each client apparatus according to the present embodiment is as shown in FIG.
In the present embodiment, differences from the first embodiment will be mainly described.
Matters not described in the present embodiment are the same as those in the first embodiment.

In the present embodiment, the representative client processing unit 210 acquires the same request generated by two or more client devices as an acquisition request, and the representative client role for the acquisition request is assigned to the own device. Any one of the two or more client devices of the request generation source is designated as a representative client clone device (hereinafter also simply referred to as a representative client clone) cooperating with the own device.
More specifically, the representative client processing unit 210 designates a representative client clone device when the number of request generation client devices of the acquisition request exceeds a threshold value.
The other client communication unit 209 notifies the client device designated as the representative client clone device that the designated client clone device has been designated, and is designated as the representative client clone device among the two or more client devices. A notification message for notifying the client device that has not been received as a response destination is transmitted.
Further, the other client communication unit 209 transmits a response corresponding to the acquisition request to the representative client clone device, and causes the representative client clone device to transmit a response to the destination client device notified by the notification message.
The other client communication unit 209 does not transmit a response to the client device to which the representative client clone device transmits a response.

In the client device designated as the representative client clone device, the other client communication unit 209 receives the notification message from the other client device.
As described above, the notification message notifies that the representative client clone device has been designated, and also notifies the client device to which the response is transmitted.
When the other client communication unit 209 receives a response from another client device after receiving the notification message, the other client communication unit 209 transmits the received response to the destination client device notified by the notification message.

Further, in the client device designated as the representative client clone device, the representative client processing unit 210 determines that the number of destination client devices notified by the notification message is 2 or more and the number is equal to or more than a threshold. One of the two or more client devices is designated as a new representative client clone device.
Then, the other client communication unit 209 notifies the client device designated as the new representative client clone device that it has been designated as the representative client clone device, and the new representative client among the two or more client devices. A notification message for notifying a client device not designated as a clone device as a response transmission destination is transmitted.
When the other client communication unit 209 receives a response from another client device, the other client communication unit 209 transmits the received response to the new representative client clone device, and the new representative client clone device sends a new representative client clone device. A response is transmitted to the destination client device notified by the notification message.
The other client communication unit 209 does not send a response to the client device to which the new representative client clone device sends a response.

*** Explanation of operation ***
The outline of the operation of the client device according to the present embodiment will be described with reference to FIGS.
22 to 28, the representative client device for a certain request is represented by client 0, and the representative client device is represented by a circle.
In FIG. 22 to FIG. 28, the client device that is the source of the request for the representative client device is represented by a square.
In FIG. 22 to FIG. 28, the client device that is the representative client clone device is represented by a hexagon.
22 to 28, numbers 0 to 7 are client device ID numbers.
In addition, the representative client and the representative client clone are represented by an arrow extending with respect to the figure of the client that is in charge of returning a response.
For example, FIG. 22 means that the representative client 0 is in charge of returning a response to the request issuer client 1.

In the following description, the threshold for designating the representative client clone is assumed to be N = 3.
That is, when the same request is issued by three clients, the representative client processing unit 210 of the representative client device designates a representative client clone from the three clients.
Similarly, the representative client processing unit 210 of the representative client clone designates a new representative client clone from the three clients when the representative client requests transmission of responses to the three clients.

As shown in FIG. 22, when the client 1 first issues a request, the representative client 0 is in charge of processing the request.
Next, when the client 2 issues a request, the representative client 0 is also in charge of processing the request, as shown in FIG.
Next, when the client 3 issues a request, since the threshold value N = 3 is reached, the representative client processing unit 210 of the representative client 0 designates the client 1 as a representative client clone for the client 3, for example.
Then, the other client communication unit 209 of the representative client 0 transmits a notification message (hereinafter referred to as a representative client clone request) to the client 1.
In the representative client clone request, it is notified that the client 1 is designated as the representative client clone and that a response is transmitted to the client 3.
As a result, the client 1 becomes a representative client clone in charge of the client 3 as shown in FIG.

Next, when the client 4 issues a request, since the threshold value N = 3 is reached, the representative client processing unit 210 of the representative client 0 designates the client 2 as a representative client clone for the client 4, for example.
Then, the other client communication unit 209 of the representative client 0 transmits a representative client clone request to the client 2.
In this representative client clone request, it is notified that the client 2 is designated as the representative client clone and that a response is transmitted to the client 4.
As a result, the client 2 becomes the representative client clone in charge of the client 4 as shown in FIG.

Next, when the client 5 issues a request, since the threshold value N = 3 is reached, the representative client processing unit 210 of the representative client 0 designates the client 1 as a representative client clone for the client 5, for example.
Then, the other client communication unit 209 of the representative client 0 transmits a representative client clone request to the client 1.
In this representative client clone request, it is notified that the client 1 is designated as a representative client clone and that a response is transmitted to the client 5.
As a result, as shown in FIG. 26, the client 1 becomes a representative client clone in charge of the client 3 and the client 5.

Next, when the client 6 issues a request, since the threshold value N = 3 is reached, the representative client processing unit 210 of the representative client 0 designates the client 2 as a representative client clone for the client 6, for example.
Then, the other client communication unit 209 of the representative client 0 transmits a representative client clone request to the client 2.
In this representative client clone request, it is notified that the client 2 is designated as a representative client clone and that a response is transmitted to the client 6.
As a result, as shown in FIG. 27, the client 2 becomes a representative client clone in charge of the client 4 and the client 6.

Next, when the client 7 issues a request, since the threshold value N = 3 has been reached, the representative client processing unit 210 of the representative client 0 designates the client 1 as a representative client clone for the client 7, for example.
Then, the other client communication unit 209 of the representative client 0 transmits a representative client clone request to the client 1.
However, even the representative client clone 1 reaches the threshold value N = 3.
Therefore, the representative client processing unit 210 of the representative client clone 1 newly designates the client 3 as a representative client clone for the client 7, for example.
Then, the other client communication unit 209 of the representative client clone 1 transmits a representative client clone request to the client 3.
In this representative client clone request, it is notified that the client 3 is designated as the representative client clone and that a response is transmitted to the client 7.
As a result, as shown in FIG. 27, the client 3 becomes a new representative client clone in charge of the client 7.

Next, the processing flow of the representative client processing unit 210 in the present embodiment will be described with reference to FIGS. 18, 19, and 20.
18, 19, and 20 are processing flows obtained by extending the processing flow of the representative client processing unit 210 illustrated in FIG. 10, and processes 1501 to 1509 are added to FIG. 10.
Below, only the part relevant to an additional process is demonstrated.

In this embodiment, the representative client processing unit 210 manages a request and a list of request issuers.
FIG. 21 shows an example of request and request issuer list management.

In the list shown in FIG. 21, it is possible to manage information of a plurality of request issuers for one request.
The request issuer information includes information that enables at least a corresponding response such as the request issuer network ID to be transferred to each request issuer by communication.
In this example, for the “GET http: // xxx / api / monitor_interval /” request, the request issue sources are “192.168.0.1”, “192.168.0.2”, “192.168 .. A total of three "0.3" are registered.
In addition, for the “GET http: //xxx/html/Monitor.html” request, the request issue sources are “192.168.0.5”, “192.168.0.2”, “192.168.8.0”. .6 ”and“ 192.168.0.8 ”are registered in total.

Returning to FIG. 18, the representative client processing unit 210 confirms whether there is another client waiting to acquire a response corresponding to the transferred request by the processing 1501.
For example, if there is a list currently being processed in the request and the request issuer list, it can be confirmed that there is another client.
If there is no other client, the process proceeds to process 701, and if it exists, the process proceeds to process 1502.

When the request is not secure in the process 701, the process for transitioning to the process 1507 instead of the process 703 is different.
Similarly, the transition destination when not stored in the process 702 is changed to the process 1507 instead of the process 703.
In addition, about the process when the request made in process 701 is not safe, it is determined when the request source client processes the request. If it is not safe, this process is not performed and the request is issued. The original client may issue a request directly to the server.
By adopting such behavior, it is possible to reduce the communication processing for the request transfer.

In processing 1507, the representative client processing unit 210 creates the entry and issues a request for the entry because the entry corresponding to the request currently being processed does not exist in the request and request issuer list (FIG. 21). Register the request issuer currently being processed as a source.
After registration is completed, the process proceeds to process 703.

Further, the transition destination of processing 705 and processing 707 is changed from processing 708 to processing 1508.
In processing 1508, the representative client processing unit 210 determines whether or not an unprocessed request issuer client serving as a transfer destination of a response corresponding to the request being processed exists in the request issuer list (FIG. 21).
If it does not exist, the process is completed. If it exists, the process proceeds to process 1509, and the representative client processing unit 210 acquires the unprocessed request issue source, and transitions to process 708.
As the subsequent processing, the transition destination of processing 709 and processing 710 is not completed, but processing 1508 is performed, and the processing is repeated until responses are returned to all unprocessed request issuers.

  In processing 1502, there is a request issuer for the request currently being processed, and the representative client processing unit 210 issues a request for an entry corresponding to the request in the request and request issuer list (FIG. 21). Register the request issuer that is currently being processed.

In processing 1503, the representative client processing unit 210 determines whether there are N or more request issuers for the same request, which is a threshold value.
As a result of the determination, if it is less than N, the process is terminated.
On the other hand, if there are N or more, the process proceeds to processing 1504.
Note that the threshold value N may be given as a fixed value, or may be given as a value that is dynamically changed in consideration of the processing load status of the network or the client, but (N−1) is a single representative. It means the maximum number of repetitions of processing for transferring a response to a single request to the client that issued the request.

In processing 1504, the representative client processing unit 210 determines a representative client clone from all request issue sources for the request currently being processed.
At this time, the representative client clone may be determined uniquely or a plurality of client clones may be determined.
However, when the number is determined to be plural, the processing 1505 and the processing 1506 are repeated by that number.

  In processing 1505, the representative client processing unit 210 determines a request issuing client that each representative client clone determined in processing 1504 is responsible for transferring a response.

  In processing 1506, the other client communication unit 209 sends a representative client clone request having information and request information related to the request issuing client in charge determined in processing 1505 to the representative client clone, and the processing is completed.

  Next, a processing flow of the representative client processing unit 210 when a representative client clone request is received will be described with reference to FIGS. 29 and 30.

  In processing 2401, the representative client processing unit 210 acquires information of a corresponding request that is requested to become a representative client clone from the representative client clone request.

In process 2402, the representative client processing unit 210 determines whether an entry corresponding to the request acquired in process 2401 exists in the request and request issuer list (FIG. 21).
If there is an entry, the process proceeds to process 2404; otherwise, the process proceeds to process 2403.

  In processing 2403, the representative client processing unit 210 newly creates an entry corresponding to the request currently being processed in the request and request issuer list (FIG. 21), and transitions to processing 2404.

  In processing 2404, the representative client processing unit 210 performs information on the request issuer client in charge described in the representative client clone request in the entry corresponding to the request currently being processed in the request and request issuer list (FIG. 21). Add all request issuers described in.

In processing 2405, the representative client processing unit 210 determines whether there are N or more request issuers for the same request, which is a threshold value.
As a result of the determination, if the number is less than N, the process proceeds to processing 2409.
On the other hand, if there are N or more, the process proceeds to process 2406.
Note that the threshold value N may be given as a fixed value, or may be given as a value that is dynamically changed in consideration of the processing load status of the network or the client, but (N−1) is a single representative. This means the maximum number of repetitions of processing for transferring a response to a single request with a client clone to the client that issued the request.

In processing 2406, the representative client processing unit 210 determines a representative client clone from all request issue sources for the request currently being processed.
At this time, the representative client clone may be determined uniquely or a plurality of client clones may be determined.
However, when the number is determined to be plural, the processing 2407 and the processing 2408 are repeated by that number.

  In processing 2407, the representative client processing unit 210 determines a request issuing client that each representative client clone determined in processing 2406 is responsible for responding to a response.

  In process 2408, the other client communication unit 209 sends the representative client clone request having the request issuer information and request information determined in process 2407 to the representative client clone, and transitions to process 2409.

In process 2409, the process waits until a response corresponding to the request being processed is transmitted from the representative client or another representative client clone.
When a response corresponding to the request being processed is transmitted, the process proceeds to process 1508.
Subsequent processing is the same as the processing described with reference to FIGS.

*** Explanation of effects ***
By performing the above processing, when a large number of request issuers are waiting for a single request, it is possible to distribute the response transfer processing to the request issuer in a representative client clone. Thus, it is possible to reduce the processing time until transfer to all request issuers without increasing the processing load of a single representative client.

Specific effects of the present embodiment will be described with reference to FIGS.
As shown in FIGS. 22 to 28, it is assumed that the same request is generated from the client 1 to the client 7.
At this time, when the representative client clone is not set, the representative client 0 sequentially returns responses to all the request issuing clients, and thus the reply process is performed seven times.
If the reply processing time is assumed to be constant and the time is expressed as T, the client response that is returned last is delayed by 7T.
On the other hand, when the representative client clone is set with the threshold value N = 3, the number of response return processes for the representative client and each representative client clone is at most two.
That is, the representative client 0 only needs to send a response back to the representative client clone 1 and the representative client clone 2.
The representative client clone 1 may return a response only to the representative client clone 3 and the client 5.
The representative client clone 2 needs to send a response only to the client 4 and the client 6.
The representative client clone 3 only needs to send a response back to the client 7.
Thus, it can be seen that there is no bias in the load of the representative client compared to the case where no representative client clone is set.
Further, in FIG. 28, when the reply process is performed with the left arrow before the right arrow, the client 6 that is replied last is the client 6.
Responses are returned in the order of client 0 → client 1, client 0 → client 2, client 2 → client 4, client 2 → client 6.
For this reason, the delay time in the client 6 is 4T, and it can be seen that the delay time is shortened by 3T compared to the case where no representative client clone is set.

Embodiment 5 FIG.
Generally, a parameter may be added to a request issued from the client device to the server device 105.
For example, in the case of a URI, a parameter may be added in such a way that a “=” symbol is inserted between the name and value of the parameter after the “?” Symbol.
For example, “val = 5” in “POST http: // xxx / api / monitor_interval /? Val = 5” which is one of the request examples shown in FIG. 6 is a parameter, and the parameter whose name is “val” “5” is designated as the value.
In addition, when a plurality of parameters are added, each parameter may be separated by an “&” symbol and connected.
For example, in the case of “http: // xxx / api / monitor_interval /? Val = 5 & target = temperature”, “5” is set as the value of the parameter whose name is “val”, and “5” is set as the value of the parameter whose name is “target”. “temperature” is specified.
The character string after the “?” Symbol describing these parameters is called a query part.
In the query part, a search condition expression to be searched by the server device 105 is defined.

Although specific values are specified in the parameters exemplified so far, a set or range of values can also be specified as an interpretation of the parameters themselves.
For example, “http: // xxx / api / user? Height = 160 | 170 | 180” can be interpreted as the value of the parameter “height” specifying “160”, “170”, or “180”. is there.
Also, for example, “http: // xxx / api / user? Height_gt = 160” means that the value of the parameter “height_gt” is “160”, but the parameter “height_gt” has the value of the parameter “height” is “height_gt” It is also possible to specify a range in which the value of the parameter “height” is finally greater than “160” by interpreting it as a parameter indicating that the value is greater than “” (gt = Greater Than).
The query part of the URI is not limited to the expression as in the above example, and the same contents may be expressed in different formats.
Further, in order to simplify the explanation, in the present embodiment, only the URI query part will be described as an example, but the parameters are similarly expressed in the request body part of the HTTP POST method without expressing the parameters in the URI query part. You can specify the method.

As described above, when it is assumed that a parameter is given to a request, it is conceivable that a request except for the query part of the URI from the request is the same, but only the query part is different.
For example,
http: // xxx / api / user? height_gt = 160
http: // xxx / api / user? height_gt = 170
Considering these two URIs as an example, each request will be processed as a separate request.
http: // xxx / api / user? height_gt = 160
Only a request related to the server is issued to the server, and from the obtained response, http: // xxx / api / user? height_gt = 170
In some cases, the representative client can extract only responses that conform to.
Specifically, there is a case where each response data group has a value of the property “height”, and this becomes clear when the response data structure is determined during system development.
In that case, it is possible to obtain responses to the two types of requests by simply issuing one type of request to the server device 105 for the two types of requests.

  An embodiment for realizing further server load reduction in consideration of the above will be described.

*** Explanation of configuration ***
A configuration example of the server apparatus 105 according to the present embodiment is as shown in FIG. 2, and a configuration example of each client apparatus according to the present embodiment is as shown in FIG.
In the present embodiment, differences from the first embodiment will be mainly described.
Matters not described in the present embodiment are the same as those in the first embodiment.

In this embodiment, the representative client role for the acquisition request is assigned to the own device, the response corresponding to the acquisition request is not cached in the received data management unit 211, and the search request includes a search condition expression. When the search condition formula can be converted into a broad search condition formula that provides a wider range of search results than the search result obtained with the search condition formula, the representative client processing unit 210 converts the search condition formula into the broad search condition formula. Convert to an expression.
For example, when the acquisition request is “http: // xxx / api / user? Height_gt = 170” described above, the search condition expression “height_gt = 170” included in the query part can obtain a wider range of search results. It can be converted into a search condition expression “height_gt = 160”.
Therefore, the representative client processing unit 210 converts the search condition expression “height_gt = 170” in the query part into a broad search condition expression “height_gt = 160”.
The server communication unit 212 transmits a request including the broad search conditional expression converted by the representative client processing unit 210 to the server device 105.
In the above example, the server communication unit 212 transmits “http: // xxx / api / user? Height_gt = 160” to the server device 105.

*** Explanation of operation ***
The processing flow of the representative client determination unit 208 of the client device that is the request generation source in this embodiment will be described with reference to FIG.
FIG. 31 is a processing flow obtained by extending the processing flow of the representative client determination unit 208 shown in FIG. 8, and processes 2501 to 2503 are added to FIG. 8.
Only the part related to the additional processing will be described.

In processing 2501, the representative client determination unit 208 determines whether there is a query part in the transferred request.
The determination can be made on the condition that “?” Exists in the URI.
If there is a query part, the process proceeds to process 2502, and if there is no query part, the process can proceed to process 501 because the same processing as in the previous embodiments is possible.

In processing 2502, when the representative client determination unit 208 obtains a response with a request in which the query part of the request being processed differs from the same request that differs only in the contents of the query part, the representative client determination part 208 converts the response to the corresponding query part. Determine whether separation is possible.
As described above, for this determination, for example, it is determined whether or not a logical sum can be configured for a request that does not include a query at the time of system development. .
If the logical sum configuration is possible, the process proceeds to a process 2503. If the logical sum configuration is not possible, the process proceeds to the process 501.

In process 2503, a request ID calculation process is performed for a request that can be ORed with the query part.
As a key for calculating the ID, a result obtained by removing at least the query part information from the request is calculated as a key.
After completing the ID calculation, the process proceeds to processing 502.
Subsequent processing is as described above, and a description thereof is omitted.

Next, processing of the client device that is the transfer destination of the request when the request is transferred to another client device in processing 505 will be described with reference to FIG.
FIG. 32 is a processing flow of the representative client determination unit 208 of the client that is the transfer destination of the request.

FIG. 32 is a processing flow obtained by extending the processing flow of the representative client determination unit 208 shown in FIG. 9, and processing 2601 to 2603 is added to FIG.
Only the part related to the additional processing will be described.

In process 2601, the representative client determination unit 208 determines whether there is a query part in the transferred request.
The determination can be made on the condition that “?” Exists in the URI.
If there is a query part, the process proceeds to process 2602, and if there is no query part, the process can proceed to process 601 because the same process as in the previous embodiments is possible.

In process 2602, when the representative client determination unit 208 obtains a response by using a logical OR-structured request for the query part of the request being processed and the same request that differs only in the contents of the query part, Determine whether separation is possible.
As described above, for this determination, for example, it is determined whether or not a logical sum can be configured for a request that does not include a query at the time of system development. .
If the logical sum configuration is possible, the process proceeds to process 2603, and if the logical sum configuration is not possible, the process proceeds to process 601.

In a process 2603, a request ID calculation process is performed for a request that can be ORed with the query part.
As a key for calculating the ID, a result obtained by removing at least the query part information from the request is calculated as a key.
After completing the ID calculation, the process proceeds to process 602.
Subsequent processing is as described above, and a description thereof is omitted.

Next, FIG. 33 and FIG. 34 are processing flow diagrams in which the processing of the representative client processing unit 210 shown in FIG. 10 is extended, and processing 2701, 2702, 2703, 2704 are added to FIG.
Only the part related to the additional processing will be described.

A process 2701 is a process that is reached when the corresponding response is not stored in the received data management unit 211 in the process 702 or when the corresponding response has expired in the process 706.
In this processing, the representative client processing unit 210 determines whether a response corresponding to the currently processed request is stored in the received data management unit 211 with the query unit excluded from comparison.
If the request is not stored, the current processing request needs to be issued to the server device as it is, and the processing transitions to processing 703.
If stored, the process proceeds to process 2702.

In processing 2702, the representative client processing unit 210 generates a logical sum of queries between the query part of the request currently being processed and the request query part corresponding to the response hit in the process 2701, and the generated result is used as a new query. A logical sum request is generated as a request and is issued to the server device 105 in processing 703.
After the completion of the generation of the logical sum request, the process proceeds to process 2705.

The process 2703 is a process that transitions from the process 705 or the process 707.
In this processing, the representative client processing unit 210 determines whether the request corresponding to the response received from the server device 105 is based on the logical sum request generated in processing 2702.
If it is not based on the logical sum request, the process proceeds to process 708. If it is based, the process proceeds to process 2704.

In the process 2704, since the response based on the logical sum request is received from the server apparatus 105, it is necessary to generate a response composed only of data corresponding to the query part of the request currently being processed. I do.
The generation processing result is transitioned to processing 708 as a response corresponding to the request being processed.

A process 2705 is a process transitioning from the process 2702.
In this process, it is determined whether or not the response corresponding to the request generated in process 2702 has been saved in the received data management unit 211.
If the response has been stored in the received data management unit 211, the process proceeds to process 706, and if not stored, the process proceeds to process 703.

  In this embodiment, after the process 707, the process shifts to the process 2703 instead of the process 708.

FIG. 35 shows an example of a table storing logical sum requests and corresponding responses.
In FIG. 35, a request having a search condition expression “min = 3” and a request having a search condition expression “min = 4” in the query part for a URI “http: //xxx/data.dat”. Two requests are assumed.
Here, the property “min” means that the value of the property “val” obtained by the response has the value of the property “min” as a lower limit.
According to this meaning, the logical sum of the query parts of the two requests is “min = 3”, and the response obtained by the logical sum request is stored in the response entry.
When a response composed only of data corresponding to each request is generated in the process 2704, the request part matches with the logical sum request for the request with “min = 3”, so the content of the response entry becomes the response as it is. .
On the other hand, for the request with “min = 4” in the query part, processing is performed to extract only those whose value of the property “val” of each data is not less than 4 that is the lower limit and does not match the logical sum request. 2704 will be performed.
As a result, the obtained response is as shown in FIG.

*** Explanation of effects ***
By performing the above processing, if the query part can be logically ORed for multiple requests that differ only in the query part, these multiple requests are combined into a single OR request, and each request is In this case, only the logical sum request can be issued to the server, and the server load can be reduced.

As mentioned above, although embodiment of this invention was described, you may implement in combination of 2 or more among these embodiment.
Alternatively, one of these embodiments may be partially implemented.
Alternatively, two or more of these embodiments may be partially combined.
In addition, this invention is not limited to these embodiment, A various change is possible as needed.

101 Client Device, 102 Client Device, 103 Client Device, 104 Client Device, 105 Server Device, 110 CPU, 111 Memory, 112 Input Interface, 113 Output Interface, 114 Communication Interface, 115 Persistent Storage Interface, 116 Disk, 120 CPU, 121 Memory, 122 input interface, 123 output interface, 124 communication interface, 125 persistent storage interface, 126 disk, 206 application logic unit, 207 request generation unit, 208 representative client determination unit, 209 other client communication unit, 210 representative client processing unit, 211 reception data management unit, 212 server communication unit, 213 communication unit, 214 request processing unit

Claims (17)

A client device included in a communication system including a plurality of client devices,
If a request to the server device generated by any client device is acquired, whether or not a representative client role that is a role for processing the request on behalf of the acquired acquisition request is assigned to the own device A representative client determination unit for determining;
When a representative client role for the acquisition request is assigned to the own device, it is determined whether a response from the server device corresponding to the acquisition request is cached in a cache area accessible by the own device. A representative client processing unit;
A server communication unit that transmits the acquisition request to the server device when a representative client role for the acquisition request is assigned to the own device and a response corresponding to the acquisition request is not cached in the cache area; ,
A client apparatus comprising: another client communication unit that transmits the acquisition request to another client apparatus when a representative client role for the acquisition request is not assigned to the own apparatus. The other client communication unit
The acquisition request is a request generated by another client device, a representative client role for the acquisition request is assigned to the own device, and a response corresponding to the acquisition request is cached in the cache area The client device according to claim 1, wherein the response cached in the cache area is transmitted to the client device that has generated the acquisition request. The client device further includes:
A request generation unit for generating a request to the server device;
The representative client processing unit
When the acquisition request is a request generated by the request generation unit, a representative client role for the acquisition request is assigned to the own device, and a response corresponding to the acquisition request is cached in the cache area The client device according to claim 1, wherein a response cached in the cache area is output to the request generation unit. The representative client determination unit
The client device according to claim 1, wherein it is determined whether or not a representative client role for the acquisition request is assigned to the own device, using a distributed hash table shared among the plurality of client devices. The representative client processing unit
The client device according to claim 1, wherein it is determined whether a response corresponding to the acquisition request and within a valid period is cached in the cache area. The other client communication unit
When a response corresponding to a request targeted by a representative client role assigned to a new client device entering the communication system is cached in the cache area, the response cached in the cache area is changed to the new client apparatus. The client device according to claim 1, wherein the client device transmits to a client device. The other client communication unit
When the response corresponding to the request targeted by the representative client role assigned to the leaving client device leaving the communication system is cached in the leaving client device, the response cached in the leaving client device is The client device according to claim 1, receiving from the leaving client device. The representative client processing unit
A backup representative client device that operates on behalf of the own device after the device has left the communication system is selected from other client devices, and the identifier of the backup representative client device is added to the request transmitted to the server device. ,
The server communication unit
The client device according to claim 1, wherein the request to which the identifier of the backup representative client device is added by the representative client processing unit is transmitted to the server device. The representative client processing unit
When the same request generated by two or more client devices is acquired as an acquisition request, and the representative client role for the acquisition request is assigned to the own device, Designate one of the client devices as a representative client clone device that cooperates with its own device,
The other client communication unit
A client device designated as the representative client clone device is notified that it has been designated as the representative client clone device, and of the two or more client devices, the client device not designated as the representative client clone device Send a notification message to notify
The client device according to claim 1, wherein a response corresponding to the acquisition request is transmitted to the representative client clone device, and the response is transmitted from the representative client clone device to a destination client device notified by the notification message. . The other client communication unit
The client device according to claim 9, wherein a response is not transmitted to a client device to which the representative client clone device transmits a response. The representative client processing unit
The client device according to claim 9, wherein the representative client clone device is designated when the number of client devices that generate the request for the acquisition request exceeds a threshold value. The other client communication unit
In addition to notifying that the representative client clone device has been designated from another client device, there may be received a notification message notifying the client device of the response transmission destination,
The client apparatus according to claim 9, wherein when a response is received from the other client apparatus after receiving the notification message, the received response is transmitted to the destination client apparatus notified by the notification message. The representative client processing unit
If the number of destination client devices notified by the notification message is two or more, designate any one of the two or more client devices as a new representative client clone device;
The other client communication unit
Notifying the client device designated as the new representative client clone device that it has been designated as the representative client clone device, and not being designated as the new representative client clone device among the two or more client devices. Send a notification message to notify the client device as the response destination,
When a response is received from the other client device, the received response is transmitted to the new representative client clone device, and a notification message is sent from the new representative client clone device to the new representative client clone device. The client device according to claim 12, wherein the response is transmitted to the notified client device of the transmission destination. The representative client processing unit
A representative client role for the acquisition request is assigned to the own device, a response corresponding to the acquisition request is not cached in the cache area, the search request includes a search condition expression, and the search condition expression If the search condition expression can be converted into a broad search condition expression that provides a broader search result than the search result obtained in
Converting the search condition formula into the broad search condition formula;
The server communication unit
The client apparatus according to claim 1, wherein a request including the broad search condition expression converted by the representative client processing unit is transmitted to the server apparatus. A communication system including a plurality of client devices,
Each client device
If a request to the server device generated by any client device is acquired, whether or not a representative client role that is a role for processing the request on behalf of the acquired acquisition request is assigned to the own device Judgment,
When a representative client role for the acquisition request is assigned to the own device, it is determined whether a response from the server device corresponding to the acquisition request is cached in a cache area accessible by the own device. ,
When a representative client role for the acquisition request is assigned to the own device and a response corresponding to the acquisition request is not cached in the cache area, the acquisition request is transmitted to the server device,
A communication system that transmits the acquisition request to another client device when a representative client role for the acquisition request is not assigned to the own device. A data processing method performed by a client device which is a computer included in a communication system including a plurality of client devices,
When the client device acquires a request to the server device generated by any one of the client devices, a representative client role that is a role of processing the request on behalf of the acquired acquisition request is assigned to the own device. A representative client determination step of determining whether or not
Whether a response from the server device corresponding to the acquisition request is cached in a cache area accessible by the client device when a representative client role for the acquisition request is assigned to the client device Representative client processing step for determining whether or not,
When a representative client role for the acquisition request is assigned to the own device and a response corresponding to the acquisition request is not cached in the cache area, the client device transmits the acquisition request to the server device. Server communication step to
A data processing method comprising: another client communication step in which the client device transmits the acquisition request to another client device when a representative client role for the acquisition request is not assigned to the own device. In a client device which is a computer included in a communication system including a plurality of client devices,
If a request to the server device generated by any client device is acquired, whether or not a representative client role that is a role for processing the request on behalf of the acquired acquisition request is assigned to the own device A representative client determination step for determining;
When a representative client role for the acquisition request is assigned to the own device, it is determined whether a response from the server device corresponding to the acquisition request is cached in a cache area accessible by the own device. A representative client processing step;
A server communication step of transmitting the acquisition request to the server device when a representative client role for the acquisition request is assigned to the own device and a response corresponding to the acquisition request is not cached in the cache area; ,
A program for executing another client communication step of transmitting the acquisition request to another client device when a representative client role for the acquisition request is not assigned to the own device.

Images