JP2011095869A - Request information processing method and computer system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve a problem wherein an application server cannot dynamically change a time period before a time-out of request processing, and a temporary load on a server, etc. may wrongly lead to the time-out, while the request processing continues on the application server, and so a response cannot be returned even when the request processing is completed, which results in the useless processing. <P>SOLUTION: An application server processing program 210 notifies a time-out management server processing program 410 of a request processing result received from a client device 40 and stores the request processing result, as statistical information, in a request information table 420. Based on the statistical information, a Web server processing program 610 determines whether to change a time-out value, and dynamically extends the time-out value. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a technique for controlling the function of a computer that processes an accepted request.

Conventionally, in a system connected from a client device via a network, a timeout occurs when the processing is not completed within a predetermined time after the Web server processing program on the server device transfers the request to the application server processing program. There is a technique for returning an error to a client device.
Also, in the storage process in the image reading system, the image reading method, and the document management server, a timeout value can be set for each server. If the storage process takes a long time, such as when the load is high, it is normal. There is a technique for determining whether or not the storage process is being performed, and extending the processing time if it is operating normally.

JP 2006-166100 A

  However, in the conventional technology, once the timeout period can be set, it cannot be changed dynamically. Therefore, a request that should have been completed a little later, such as when there is a high load, also results in an error, and the request that resulted in the error Is also a useless process because the process continues in the application server processing program of the server device.

  Further, the same request is often transmitted again from the client device that has timed out in this way, and those requests also time out. By repeating this operation, the load on the server device side is further increased, and not only is the request more likely to be timed out, but also the load on the server device is increased. Requests that have not timed out until now are likely to be timed out, and may increase the load on the server device. On the other hand, in the request information processing system, However, since it is difficult to determine whether the request is operating normally, the processing time cannot be extended by determining whether the request processing is operating normally. Therefore, an object of the present invention is to eliminate these dangers in the timeout technique by dynamically changing the time until timeout occurs.

  It is another object of the present invention to prevent the load on the server device from increasing meaninglessly by reducing unnecessary retries from the client device by appropriately extending the time until timeout.

  In order to achieve the above object, in the present invention, when a timeout occurs when a request received from a client device is processed by a server device, the time required for the processing of the request finally from the client device is determined. When a timeout does not occur, a request URI (Uniform Resource Identifier) of the client device is held as information on the memory of another server device capable of mutual communication with the server device, and the client device Before the Web server processing program on the server device transfers the request from the server to the application server processing program, the time until timeout is changed based on the held information.

  By dynamically changing the timeout, it is possible to save the request that the processing has been completed in a short time.

It is a figure which shows the structure of the request processing system to which the server apparatus in this embodiment was connected. It is a figure which shows the flow of a process when timeout does not generate | occur | produce in the flow of a process of the request processing system in this embodiment. It is a figure which shows the flow of a process when timeout occurs in the flow of a process of the request processing system in this embodiment. It is a figure which shows the structure of the statistical information of the request which the application server process part in the timeout management server process part in this embodiment processed. It is a figure which shows the structure of the file which defines the operation | movement in the Web server process part in this embodiment. It is a flowchart figure which shows the flow of a process of a Web server processing program in the server apparatus in this embodiment.

  Hereinafter, modes for carrying out the present invention (hereinafter referred to as “embodiments”) will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating a request processing system to which a server device according to the present embodiment is connected. As shown in FIG. 1, the request processing system 1 in this embodiment includes a plurality of server devices 10, a server device 20, and a server device 30. The client device 40 and the server device 30 are configured to be able to communicate with each other via the network 50. Here, one server device 10 and one server device 30 are shown, but the number is not limited even if there are a plurality of each, and the server device 10 and the server device 30 should be combined into one server device. Is also possible. Further, although two client devices 40 are shown, the number is not particularly limited as long as there are a plurality of client devices 40.

  The client device 40 has a function of transmitting a request to the server device 30 via the network 50 and a function of receiving a response to the request from the server device 30.

  The server device 10 includes a CPU (Central Processing Unit) 100 and a memory 200. The memory 200 as a storage unit has a function of storing including the application server processing unit 210. The CPU 100 as a processing unit has a function of executing a program that constitutes the processing unit stored in the memory 200.

  The server device 20 includes a CPU 300 and a memory 400. The memory 400 as a storage unit has a function of storing a timeout management server processing unit 410 and a request information table 420. The CPU 300 as a processing unit has a function of executing a program that constitutes the processing unit stored in the memory 400.

  The server device 30 includes a CPU 500 and a memory 600. The memory 600 as a storage unit has a function of storing a Web server processing unit 610 and a server operation definition file 620. The CPU 500 as a processing unit has a function of executing a program constituting the processing unit stored in the memory 600.

  When a request is transmitted from the Web server processing unit 610, the application server processing unit 210 executes the processing of the request, returns a response as the execution result to the Web server processing unit 610, and the Web server processing unit 610 times out. If a timeout has not occurred, the URI is notified to the timeout management server processing unit 410, and a timeout has occurred in the Web server processing program 610, and the response is not returned to the Web server processing unit 610 because the connection has been disconnected. In this case, the program describes a code for notifying the timeout management server processing unit 410 of the URI and the processing time required for processing the request.

  The timeout management server processing unit 410 acquires a timeout value set in the server operation definition file 620 from the Web server processing processing unit 610 at the time of startup, and when the Web server processing unit 610 receives a request from the client device 40 The code for acquiring information related to the request from the request information table 420 and notifying the web server processing unit 610 and the code stored in the request information table 420 when the information related to the request is notified from the application server processing unit 210 are described. Program.

  The request information table 420 holds, as statistical information, the URI, the total number of requests, the number of times that the timeout setting time is exceeded, and the average timeout setting time.

  When a request is transmitted from the client device 40, the Web server processing unit 610 transmits information on the URI of the request to the timeout management server 410, and based on the information notified from the timeout management server 410, After calculating the time until timeout and transferring the request to the application server processing unit 210, the timer monitoring process is performed until the reply from the application server processing unit 210 is obtained within the calculated time until timeout, and the application server processing unit When a response is returned from 210, the response is returned to the client device 40, and until the timeout defined in the server operation definition file 620 before the response is returned from the application server processing unit 210. When the time has elapsed, disconnects the connection to the application server processing unit 210 as a time-out processing, a program code for returning the error is described in the client device 40.

  The server operation definition file 620 includes information for controlling the operation of the Web server 610, transfers the request transmitted from the client device 40 to the application server processing unit 210, and returns a response from the application server processing program 210. Defines the waiting time (time to timeout), the threshold of how many times to exceed the total number of requests, and the upper limit of how long the time to timeout is extended at maximum, the threshold is As a percentage, the upper limit of the extension time is defined in seconds. The threshold and the upper limit of the extension time are values common to all requests, but can be defined by changing the values for each URI.

  FIG. 2 is a diagram showing processing when the request processing is completed without extending the time until timeout in the processing flow of the request processing system at the normal time in the present embodiment. As shown in FIG. 2, here, the case where the request processing system 1 receives a request from the client 40 will be described as an example of the change in time until timeout and the request processing. Further, the web server processing unit 610, timeout management server processing unit 410, and application server processing unit 210 are respectively the application server processing unit 210, timeout management server processing unit 410, and web server processing unit 610 in FIG. It is realized by being executed by the CPU 300 and the CPU 500, respectively.

  When a request (http: //...index.jsp) is transmitted from the client 40 (step S1001), the Web server processing unit 610 requests statistical information regarding the URI of the request from the timeout management server processing unit 410 (step S1001). Then, the timeout management server processing unit 410 acquires statistical information (step S1003) and notifies the Web server (step S1004). Thereafter, the Web server processing unit 610 performs timeout extension determination processing for determining whether or not to change the time until timeout in the request (step S1005), determines the time until timeout, and then requests the application server processing unit 210 to make a request. Is transferred (step S1006), and the timer is monitored for the time until the determined timeout. The application server 210 processes the request transferred from the web server processing unit 610 (step S1007), returns the processing result to the web server 610 as a response (step S1008), and then sends the request to the timeout management server 410 processing unit. The timeout management server processing unit 410 stores the notified information as statistical information (step S1010). When the Web server 730 receives a response from the application server 710 (step S1008), it returns a response to the client 40 (step S1011), and the client 40 receives it (step S1012).

  FIG. 3 is a diagram illustrating a process when a time-out occurs after the time until the time-out is extended in the processing flow of the request processing system 1 in the present embodiment. As shown in FIG. 3, here, the request processing system 1 receives a request from the client device 40 and executes the process. However, the process takes time and the timeout occurs. Processing when it occurs will be described. In addition, the Web server 610, the timeout management server 410, and the application server processing unit 210 are respectively the application server processing processing unit 210, the timeout management server processing unit 410, and the Web server processing program 610 in FIG. It is realized by being executed by.

When a request (http: //.../index.jsp) is transmitted from the client 40 (step S2001), the Web server processing unit 610 requests statistical information regarding the URI of the request from the timeout management server processing unit 410 ( In step S2002), the timeout management server processing unit 410 acquires statistical information (step S2003) and notifies the Web server (step S2004). Thereafter, the Web server processing unit 610 performs processing to determine whether or not to change the time until timeout in the request (step S2005), determines the time until timeout, and transfers the request to the application server processing unit 210. (Step S2006), the timer is monitored for the time until the determined time-out, but the time until the time-out determined until the response is returned after the application server processing unit 210 starts request processing (Step S2007) In such a case, the web server processing unit 610 disconnects the connection with the application server processing unit 210 as a timeout process (step S2008), returns an error to the client 40 (step S2009), and the client 40 To receive (step 2010).
Thereafter, when the request processing in the application server processing unit 210 (step S2007) is completed, the processing time required for the request is notified to the timeout management server processing unit 410 together with the URI (step S2011), and the timeout management server The processing unit 410 stores the processing time and URI of the request received from the application server processing unit 210 in the request information table (step S2012).

  FIG. 4 shows an example of the data structure of the request information table 420 in this embodiment. As illustrated in FIG. 4, the request information table 420 includes a URI 801, an excess count 802 that is the number of times a timeout has occurred, a total request count 803, and an average excess time 804. In the request information table 420, after the request processing in the application server processing unit 210 (see FIG. 2) is completed, information on the processed requests is organized and managed. In step 1009, registration and information update are performed. .

  The URI 801 indicates the identifier of the request transmitted from the client device 40 (see FIG. 2), and is notified from the application server processing unit 210 (see FIG. 2) when the processing of the request is completed. If registered in 420, 1 is added to the total number of requests 803 corresponding to the URI, and if not registered in the request information table 420, the total number of requests 801 is registered in the request information table 420.

The number of times of excess 802 indicates the number of times that the timeout value set in the notified URI has exceeded and timed out.
The total number of requests 803 indicates the total number of requests for the notified URI.

  The average excess time 804 is {(average excess time 804 registered in the request information table at the time of calculation × total number of requests 802 registered in the request information table at the time of calculation + request processing notified from the application server processing unit 210 (Time) ÷ total number of requests including the request} − {timeout setting value defined in Web server operation definition file 620 (see FIG. 1)}, and is overwritten each time a timeout occurs.

  FIG. 5 shows an example of the data structure of the Web server operation definition file 620 in this embodiment. As shown in FIG. 5, the Web server action definition file 620 transfers the request received from the client to the application server processing unit 210 (see FIG. 2) until receiving a response from the application server processing unit 210 (see FIG. 2). A timeout value 901 that defines the timeout value of the request, a threshold value 902 that sets how many times the total number of requests is allowed to be exceeded, and an upper limit on how much the maximum time can be extended when changing the timeout period A timeout extension upper limit value 903 for setting a value is included. Data in the Web server operation definition file 620 is set by an administrator of the request processing system 1 or the like.

  FIG. 6 is a flowchart showing a detailed operation of the timeout value extension determination process (step 1004 in FIG. 2) of the Web server processing processing unit 610 in the Web server in the server device according to the present embodiment.

  When the Web server processing unit 730 (see FIG. 2) receives the request, the timeout management server processing unit 410 (see FIG. 2) acquires statistical information corresponding to the URI of the request from the request information table 420 (step S101). Based on the excess count 802 (see FIG. 4) and the total request count 803 (see FIG. 4), the rate at which the timeout setting value is exceeded in the URI is calculated, and the calculated result is the Web server operation definition file. It is compared with a threshold value 902 (see FIG. 5) set to 620 (see FIG. 5) (step S102). If the threshold value 902 is not exceeded (less than n% in step S102), the timeout value is not changed, and the set value of the timeout value 901 in the server operation definition file 620 is used as the timeout value (step S106), and the process is terminated. . On the other hand, if the calculated result is equal to or greater than the threshold value 902 (see FIG. 5) (n% or more in step S102), the average excess time 804 (see FIG. 4) acquired from the timeout management server processing unit 410 (see FIG. 2) Comparison with the timeout extension upper limit 903 (see FIG. 5) set in the Web server operation definition file 620 (see FIG. 5) is performed (step S103). If the average excess time 804 (see FIG. 4) is longer than the timeout extension upper limit 903 (see FIG. 5) set in the Web server operation definition file 620 (see FIG. 5) (longer than the upper limit in step S103). The timeout value is extended by the timeout extension upper limit value 903 (see FIG. 5) (step S104), and if the timeout value is shorter than the timeout extension upper limit value 903 (see FIG. 5) (the upper limit value in step S103). Hereinafter, the timeout value is extended by the average excess time 804 (see FIG. 4) and the timeout value 901 (see FIG. 5) is extended (step S105), and the process is terminated.

  As an effect attached to the present embodiment, when a time-out occurs, the time required for a request that is continuously processed by the application server is accumulated in the request table 420 as statistical information, thereby increasing the accuracy of the time to be extended. It is possible to raise it.

  In addition, since appropriate timer monitoring can be performed, unnecessary retries can be eliminated, and the burden on the server side as well as the client side can be reduced.

  As mentioned above, although embodiment of this invention was described, this embodiment is only the illustration for description of this invention, and is not the meaning which limits the scope of the present invention only to that embodiment. The present invention can be implemented in various other modes without departing from the gist thereof.

DESCRIPTION OF SYMBOLS 1 Request processing system 10 Server apparatus 20 Server apparatus 30 Server apparatus 40 Client apparatus 50 Network 100 CPU
200 Memory 210 Application server processing unit 300 CPU
400 Memory 410 Timeout management server processing unit 420 Request information table 500 CPU
600 Memory 610 Web server processing unit 620 Server operation definition file

Claims (5)

In a communication control method in a computer system including a Web server processing device and a request received from a client device and including a communication control device including a storage device,
The timeout management processing unit stores the statistical information of the request in a storage device,
When the time-out management server processing unit receives the request, it reads out the statistical information of the request accumulated in the time-out management,
The request information processing method, wherein the Web server processing apparatus changes a time until timeout based on statistical information of the timeout information. The request information processing method according to claim 1, wherein the application server processing unit processes the request transferred from the Web server processing device, and notifies the timeout management server of a processing result.   The time-out management server processing unit holds a time-out excess rate threshold value as a reference for changing a time-out time in the operation definition file of the Web server processing device. Request information processing method.   2. The request information according to claim 1, wherein the timeout management server processing unit holds a timeout extension upper limit value that is an upper limit time for extending a timeout in the operation definition file of the Web server processing apparatus. Processing method. In a computer system including a Web server processing device and a communication control device that processes a request received from a client device and includes a storage device,
The timeout management processing unit stores the statistical information of the request in a storage device,
When the time-out management server processing unit receives the request, it reads out the statistical information of the request accumulated in the time-out management,
The computer system, wherein the Web server processing device changes a time until timeout based on statistical information of the timeout information.

Images