JP4640980B2 - Request processing method, request processing apparatus, and request processing program - Google Patents

Description

  The present invention relates to a priority control technique for controlling the priority of request processing in a computer.

  When the number of requests exceeding the allowable range is received by the application server, the response time is slow for all requests to handle all requests equally, or an error occurs for an unspecified request that exceeds the allowable range The page is returned and the connection is refused. Therefore, as a method for preferentially controlling requests that do not handle all requests uniformly by preferentially controlling requests, Patent Literature 1 gives priority to requests from specific users and requests to specific web pages. And how to control it. However, the method of giving priority control to a request from a specific user can be used only on the website that authenticates the user, and when request to a specific Web page is preferentially controlled, links from multiple Web pages Requests that are sent when accessing a Web page that is being used can only be handled with the same priority.

US2003 / 0009505A1

  Web services (functions realized by executing programs and objects) such as online banking and ticket reservation systems are widely provided. Since these Web sites are generally available to multiple users at the same time, they provide multiple services simultaneously, but use resources (CPU, memory, network bandwidth, etc.) for the simultaneous services. For this reason, when the number of simultaneous services exceeds the allowable range, resource shortage of the application server occurs, resulting in service stoppage due to deterioration of the quality of the provided service or, in the worst case, server down. In these cases, all requests are adversely affected to treat all requests equally.

  Web sites often have a configuration in which a plurality of Web pages are transitioned and completed when a user enjoys one service. When a request is sent from a Web page that is transitioning to allow a user to enjoy one service, if the application server is congested, the user will be unable to connect or the service will be delayed due to the slow response. May give up. As a result, there is a problem that processing of a Web page that has transitioned to receive one service is wasted.

  An object of the present invention is to solve the above problems and provide a technique capable of performing priority control of request processing according to the degree of page transition when predetermined processing is performed by request processing in a plurality of page transitions. is there.

  The present invention controls the priority of request processing in accordance with the request source page in a priority control apparatus that controls the priority of request processing in a computer.

  In the present invention, when a Web site user operates a Web page using a Web browser of a user terminal, the Web server of the user terminal communicates with the application server via the network. When a request is transmitted by a device, an application server that is a priority control device (a computer, an information processing device, or a program or object that performs such processing) of the present invention refers to the information in the request, and Reads the path name information of the requesting Web page that sent the request.

  Then, the path name-priority correspondence information indicating the priority corresponding to the path name of the request request source is read by the storage device, and the priority corresponding to the request source Web page of the read request is set to the path name-priority Obtained from degree correspondence information.

  Next, a request queue corresponding to the acquired priority is selected from the request queue group including request queues of each priority, and the request requested from the requesting Web page is stored in the selected request queue. After that, the request queues in the request queue group are referred to in order of priority, the requests are taken out and assigned to the threads, and the request processing is executed by the threads.

  In the conventional method, all requests sent to the application server are handled equally. In the present invention, the priority of each Web page is set according to the characteristics of the Web site, so that the characteristics of the Web site are improved. At the same time, priority control of request processing can be performed.

  In other words, as a characteristic of websites mainly for commercial use such as net banking and ticket reservation system, there are many configurations in which multiple web pages are transitioned and completed to enjoy one service. Each time the web page required for the transition is made, it can be said that the service is progressing toward the completion of the service. Therefore, in the present invention, for each of these websites, a request that occurs every time the web page necessary for service reception is transitioned. Set a higher priority for.

  Since the Web page located behind the Web page transitioning to service enjoyment is close to service completion, the service is completed early by processing the application server with priority given to requests sent from these Web pages. On the other hand, since the web page located in front of the transitioning web page is far from service completion immediately after the service is started, the processing priority of requests sent from these web pages is lowered.

  As an effect of this, for the application server, it is possible to preferentially process services that are likely to end early, and the number of services to be executed simultaneously is reduced and the resources used are reduced, so that stable operation can be performed. In addition, since the response to a request tends to be better every time the Web page for enjoying the service transitions, when the user progresses to the middle of the transitioning Web page, the response becomes worse due to congestion of the application server. There is less to give up.

  On the other hand, for a request from a web page located in front of the transitioning web page, the response is slow when the application server is congested. Due to the slow response, the user may notice the congestion of the application server and give up service enjoyment, and the congestion of the application server is alleviated. In addition, when it continues, the response tends to be better every time the Web page is changed, so that the user is stressed.

  According to the present invention, when predetermined processing is performed by request processing in a plurality of page transitions, priority control of request processing according to the degree of page transition can be performed.

  A priority control apparatus according to an embodiment for controlling the priority of request processing in a computer will be described below.

  FIG. 1 is a diagram showing an outline of the system of this embodiment. FIG. 1 shows an outline of the system in this embodiment and the order of request processing in the application server when a plurality of requests are generated simultaneously.

  In the Web site 101, the selection of a service is accepted from the index page 102, which is the top page, and the Web page 103 to 105 is shifted in order to complete the service. In the application server 106 of the present embodiment, the path name-priority correspondence information 107 for managing the request request Web page and the request processing priority in pairs is positioned behind to receive one service. The priority of the request from the web page is raised and the priority of the request from the web page located in the front is lowered, and the priority as the page transitions from the top page to the service provided page Indicates that it will be higher.

  In this case, when the requests 108 to 111 are simultaneously sent from the pages 102 to 105 of the Web site 101 to the application server 106, the application server 106 recognizes from which Web page the request is sent and passes the path name-priority. Processing execution priority control is performed according to the priority of the degree correspondence information 107. Therefore, the request processing is started in the order of the request 109 and the request 108 after the request 110 and the request 111 which are requests from the Web page having a high priority.

  FIG. 2 is a diagram showing details of a system configuration example of the present embodiment. The Web site user communicates with the application server 206 via the network 205 and enjoys the service by operating the Web page 204 using the Web browser 203 developed in the memory 202 of the user terminal 201. A configuration similar to the user terminal 201 exists for each Web site user.

The web server module 208 deployed in the memory 207 on the application server 206 accepts a request from the web browser 203 by the communication device.
The request storage unit 219 expanded in the memory 207 acquires a request from the Web server module 208 and stores the request in any of the request queues 210 to 212.

  The request queue group 209 includes a plurality of request queues expanded in the memory 207 on the application server 206. Each of the request queues 210 to 212 is a container that can store and retrieve requests by FIFO (First In First Out). Each of the request queues 210 to 212 stores requests having different priorities. The number of request queues represents the number of divisions according to priority. The number of request queues and the maximum number of requests that can be stored in each request queue are specified by the Web site administrator before the server is activated. In the prior art, since there is only one request queue, priority control cannot be performed for each request.

  The request queue management unit 213 manages the request queue group 209. The request queue management unit 213 knows the number of request queues and the maximum number of requests that can be stored in each request queue. Furthermore, the number of requests stored in the request queue is managed, and the request queue management unit 213 answers these information inquiries. The request queue management unit 213 generates a request queue when the application server 206 is activated.

  The request processing allocation unit 220 expanded in the memory 207 acquires a request from the request queues 210 to 212 and sends the request to any of the threads 215 to 217. The thread group 214 is expanded in the memory 202 and includes threads 215 to 217 (processes) that perform processing corresponding to the request.

  In the present embodiment, a program for causing a computer to function as each unit is recorded on a recording medium such as a CD-ROM and stored in a magnetic disk or the like, and then loaded into a memory and executed. The recording medium for recording the program may be a recording medium other than the CD-ROM. The program may be used by installing it from the recording medium into the information processing apparatus, or the program may be used by accessing the recording medium through a network.

  Further, the computer as the application server stores a program for causing the computer to function as each unit, a CPU for executing processing such as a business AP and an OS, and loading the program or temporarily storing data at the time of execution. Hardware such as a memory, a magnetic disk device for storing those programs and the information, an input device for receiving input of operation instructions and data, an output device for outputting various information, a communication device for communicating with other computers, etc. Suppose you have hardware resources.

  FIG. 3 is a diagram showing an example of the path name / priority correspondence information 218 according to this embodiment. As shown in FIG. 3, the path name / priority correspondence information 218 manages information on the path name 301 and the priority 302 of the Web page in pairs. A scheme is adopted in which the pair priority 302 can be acquired from the path name 301. The priority 302 is independent of the priority in the request queues 210 to 212, and the priority 302 and the priority of the request queues 210 to 212 do not always have to correspond one-to-one. A corresponding configuration may be used.

  The path name-priority correspondence information 218 is set by accepting input from the website administrator before and during operation of the application server, but the website administrator gives priority to all web pages. Entering the degree is a lot of work, so it is necessary to enjoy the service, focusing on the fact that many websites have the characteristic of transitioning and completing multiple web pages to enjoy one service Every time a Web page transitions, the setting to increase the priority of the request is performed mechanically, and then the priority input from the website administrator is received by the input device, and the setting contents are customized and stored in the storage device By taking the storage form, the workload of the website administrator can be reduced.

  In order to make the setting to increase the priority of the request every time the web page required for service enjoyment is changed mechanically, the website is used in the test operation to recognize the relationship between the transitioning web pages. This is done by analyzing the log at that time by the program. At that time, when the request arrives in the application server, the session ID or IP address, request time, request source URL (Uniform Resource Locator), and request destination URL are accumulated in the storage device as a log. Since this log can trace the Web page transition from the top page to the service provision, a program for calculating the link depth of each Web page from the top Web page is operated. Here, the link depth is a value indicating the link depth from the top page to the page where the service is provided. For example, the link depth increases by 1 every time one page transitions from the top page toward the service providing page. Value shall be used.

  The obtained link depth is set in the path name-priority correspondence information 218 as the priority of each web page (the higher the link depth value, the higher the priority). As a result, the higher the link depth from the top page, the higher the priority.

  The path name / priority correspondence information 218 is created so that it can be exported / imported, and the path name / priority correspondence information 218 generated by one application server can be used by another application server. And This is used when a load balancer is installed in the system for load balancing, and the same Web site is placed and operated on multiple application servers.

  That is, the path name-priority correspondence information collected by one application server is exported and recorded on a storage medium such as a hard disk, and the other path name-priority correspondence information is imported by another application server. Read path name-priority correspondence information from the storage medium. Thus, if the path name / priority correspondence information is generated by one application server, the generation of path name / priority correspondence information by another application server becomes unnecessary.

  Here, in order to consider the migration of the path name / priority correspondence information 218 between application servers having different numbers of request queues, the priorities of the path name / priority correspondence information 218 and the request queues 210 to 212 are independent. Shall.

  FIG. 4 is a diagram showing a processing flow of the request storage unit 219 of the present embodiment. When the web server module 208 receives a request from the user terminal via the communication device, the request storage unit 219 acquires the request from the web server module 208 via the memory 207 in step 401.

  In step 402, the acquired request is examined, and the path name of the URL of the Web page that made the request is acquired. That is, the URL is acquired by reading the content of the HTTP (Hyper Text Transfer Protocol) Referer request header held by the acquired request from the storage device, and the path name is extracted from the acquired URL. In step 403, the path name-priority correspondence information 218 is read from the memory 207, and the path name-priority correspondence information 218 is searched using the path name acquired above as a key to obtain the priority corresponding to the path name. To do.

  Here, in steps 402 and 403, when the priority cannot be acquired from the path name-priority correspondence information 218 because the Referer request header is not added to the request or the acquired path name does not exist. The priority to be used is determined in advance. If the priority cannot be acquired in step 403, the priority determined in advance is read from the storage device and used. Also, the path name of the page transitioned in each request is stored in the storage device for each session, and the path name of the request source page is read by reading the path name stored immediately before the next request from the storage device. It may be obtained.

  In step 404, it is determined from the priority acquired in step 403 which priority is stored in the request queue. Since the priority in the path name-priority correspondence information 218 and the priority of the request queue are independent, the number of divisions may be different. Therefore, processing for evenly mapping the priority in the path name / priority correspondence information 218 to the priority of the request queue is performed, and which request queue is stored is determined. In step 405, the request is stored in the request queue determined above. After storing the request, the request storage unit 219 returns to step 401 to store the next request.

  Depending on the Web site, in order to reduce the load on the application server even when it is crowded, it is possible to configure so that requests from Web pages with low priority are not stored in the request queue. In this case, the priority is acquired in step 403, and the request queue management unit 213 is inquired whether the priority can be stored in the request queue. If possible, store the request in the request queue, otherwise return an error page. Here, as a material for determining whether to store the request in the request queue, the priority for rejecting the request, the number of stored requests for all the request queues that become the threshold for rejecting the storage, or the stored requests for the maximum number of stored requests If the ratio of the number is specified in advance and the request priority is to be rejected, the request in the request queue is read from the storage device, the number of stored requests is counted, and the number of stored requests or the maximum stored requests When the ratio of the number of stored requests to the number exceeds the specified value, processing for stopping the request in the request queue is performed.

  FIG. 5 is a diagram showing a processing flow of the request processing allocation unit 220 of the present embodiment. In step 501, the thread on the application server becomes ready for request processing and waits until a request exists in any of the request queues.

  In step 502, the request queue having the highest priority is set as a candidate queue from which requests are extracted. In step 503, it is confirmed whether there is a request in the request queue of the candidate queue. If there is no request, in step 504, the request queue having the next highest priority after the candidate queue is set as the candidate queue, and the process proceeds to step 503. If there is a request in the candidate queue, in step 505, the request is extracted from the request queue of the candidate queue, and in step 506, the extracted request is sent to the thread. Then, the request process allocation unit 220 proceeds to step 501 and continues the process, and the thread to which the request is sent performs a process according to the request.

  As described above, according to the priority control apparatus of the present embodiment, the priority of request processing is controlled according to the request source page, so when performing predetermined processing by request processing in a plurality of page transitions. It is possible to perform priority control of request processing according to the degree of page transition.

It is a figure which shows the outline of the system of this embodiment. It is a figure which shows the detail of the system configuration example of this embodiment. It is a figure which shows the example of the path name-priority corresponding | compatible information 218 of this embodiment. It is a figure which shows the processing flow of the request storage part 219 of this embodiment. It is a figure which shows the processing flow of the request process allocation part 220 of this embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 101 ... Web site, 102 ... Index page, 103-105 ... Web page, 106 ... Application server, 107 ... Path name-priority correspondence information, 108-111 ... Request, 201 ... User terminal, 202 ... Memory, 203 ... Web browser, 204 ... Web page, 205 ... Network, 206 ... Application server, 207 ... Memory, 208 ... Web server module, 209 ... Request queue group, 210-212 ... Request queue, 213 ... Request queue manager, 214 ... Thread Group, 215 to 217 ... thread, 218 ... path name-priority correspondence information, 219 ... request storage section, 220 ... request processing allocation section, 301 ... path name, 302 ... priority.

Claims (6)

In a request processing method in a request processing device that processes a request from each of a plurality of request source Web pages that are sequentially transmitted and that is transmitted by operating a Web browser of a user terminal.
The CPU of the request processing device executes a program, assigns a priority to each of the URLs of the plurality of request source Web pages, and stores in advance in the storage device of the request processing device as path name-priority correspondence information. ,
When the CPU of the request processing apparatus receives a request including the URL of the request source Web page sent from the user terminal by executing the Web server module,
The request storage unit of the request processing device extracts the URL of the request source Web page from the request , refers to the path name-priority correspondence information, and sets the priority corresponding to the URL of the request source Web page. Get and also
The received request is stored in a request queue prepared in advance in the storage device corresponding to the acquired priority.
The request processing allocation unit of the request processing device refers to the request queue, extracts the requests in order from the request queue having a high priority, and sends the requests to a thread executed by the CPU in the request processing device. The request processing method characterized by processing said taken-out request. The request storage unit stores log data of the request source web page that changes in response to the request in a storage device,
The request storage unit reads and specifies the order in which the request source Web page transitions from the log data from the storage device,
The request storage unit sets the priority corresponding to the URL of the request source Web page according to the order in the path name-priority correspondence information and stores the priority in the storage device. Request processing method described in. The request storage unit stores, in a storage device, a URL of the request source web page that transitions according to the request ,
The request processing according to claim 1, wherein when the request is received, the request storage unit identifies the stored URL as the request source web page that is an issuer of the received request. Method. The request queue management unit of the request processing device reads the requests in the request queue from the storage device, counts the number of stored requests indicating the total number of the requests stored in the request queue,
The request storage unit stores the request in the request queue when the counted number of stored requests exceeds a predetermined value, or when the ratio of the number of stored requests to the maximum number of stored requests exceeds a predetermined value. The request processing method according to claim 1, wherein the request processing method is stopped. In a request processing device that processes a request from each of a plurality of request source Web pages that are sequentially transmitted, which are transmitted by operating a Web browser of a user terminal.
Storage means for assigning a priority to each of the URLs of the plurality of request source web pages and pre-storing the path name-priority correspondence information in a storage device;
A web server module that accepts a request including a URL of the request source web page sent from a user terminal;
The URL of the request source Web page is extracted from the request , the priority corresponding to the URL of the request source Web page is acquired with reference to the path name-priority correspondence information, and the received request is A request storage unit for storing in a request queue provided in advance in the storage device corresponding to the acquired priority;
Referring to the request queue provided in advance in the storage device, requests are taken out in order from the request queue having the highest priority, sent to a thread executed by the CPU in the request processing device, and taken out by execution of the thread A request processing apparatus comprising: a request processing allocation unit that processes a request. In a request processing program to be executed by a request processing device that processes a request from each of a plurality of request source Web pages that are sequentially transferred, which is transmitted by operating a Web browser of a user terminal.
The CPU of the request processing device executes a program, assigns a priority to each of the URLs of the plurality of request source Web pages, and stores in advance in the storage device of the request processing device as path name-priority correspondence information. Steps,
The CPU of the request processing device receives a request including a URL of the request source Web page sent from a user terminal by executing a Web server module;
The request storage unit of the request processing device extracts the URL of the request source Web page from the request , refers to the path name-priority correspondence information, and sets the priority corresponding to the URL of the request source Web page. Get and also
Storing the accepted request in a request queue provided in advance in the storage device corresponding to the acquired priority;
The request processing allocation unit of the request processing device refers to the request queue, extracts the requests in order from the request queue having a high priority, and sends the requests to a thread executed by the CPU in the request processing device. A request processing program that causes the request processing device to execute the step of processing the fetched request.

Images