JP2006120080A - Web service request relay system, web service request relay method, relay server and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a web service management system which prevents occurrence of local heavy load status in web servers for preventing whole systems from being ceased by local service delays. <P>SOLUTION: A communication controller between services, in services in which more than one Web services work in cooperation, implements repeaters which repeat request messages and response messages communications between other services in respective computers providing web services. The repeater on a service requesting side have a feature to reduce load of a service operating side by exchanging loading information of the computers between the repeaters and making delays of requests for the service operating side in which performance degradation occurs. Also the repeaters, which operate with Web services of respective computers, are characterized to perform implementing the repeater without modifying a communication mechanism of existing Web services by acquiring and transferring the request messages and response messages for the Web services. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to execution of a business flow by cooperation of a plurality of Web service computers, and in particular, degradation of the performance of the entire system due to a transmission side computer refraining from transmitting a request for a service request to a highly loaded Web service computer. The present invention relates to a Web service request relay system, a Web service request relay method, a relay server, and a program thereof.

  Conventionally, the system integration of a company combines business applications and systems distributed by department such as an accounting system, a personnel system, and a financial accounting system in accordance with the flow of business. This system combination is realized by a function and a product called EAI (Enterprise Application Integration). As a hub between systems, EAI centrally manages data linkages on a single computer. In this system, business applications and system APIs (Application Program Interfaces) are called, and the inter-system cooperation is tightly coupled. Therefore, it is easy to control the operation of the inter-system cooperation.

  Since this system connection is performed by an adapter adapted to an interface specific to each system, the system construction is very expensive. For this reason, efforts have been made to align the APIs of individual systems into a common interface. In addition, efforts have been started to exchange data between systems according to a standardized common messaging specification.

  This standardized message passing weakens the degree of functional coupling between systems and allows loose coupling between systems. This messaging standard uses a SOAP message and Web service interface in which XML data is enveloped. An advantage of this system construction by messaging is that HTTP, which is a standard protocol for communication on the Internet and intranet, can be used.

  In this system, each system holds business logic and data and is loosely coupled. Since only the relationship between adjacent services on the business logic can be determined, the business logic cannot be centrally managed as in the conventional workflow system. A mechanism for managing such loosely coupled Web services is required.

  A conventional web service management system manages the service contents and operation statuses of a plurality of web services provided on a plurality of computers.

  The Web service includes service definition information, a communication unit, a service execution unit, and a service response message transmission unit. The service definition information defines service contents and input and output information for service execution. When a service is not completed with one service and another service is used, another web service call is defined in the service definition information. As illustrated in FIG. 13, the Web service management system includes a location information of a computer that provides the Web service and a service definition information storage device that stores the service definition information.

  This Web service management system acquires Web service definition information from a service definition information storage device. In the service composition calculation unit, each Web service repeatedly links with the adjacent Web service to link the relationship of the entire service from a plurality of Web services, and constructs a flow between the services. This flow can be regarded as a specific business processing flow using individual services.

  Since each of these Web services provides a complete service, the degree of functional coupling between the Web services is low and the service is highly independent. Therefore, each Web service continues the service request regardless of the processing capability and load status of the adjacent Web service. In addition, each Web service has a different calculation complexity of service execution, and the hardware capability of the computer itself is diverse. Therefore, when a plurality of Web services cooperate to perform one process, the overall throughput depends on a service with low throughput.

  In such an environment, service requests are likely to stay in a service with low throughput, and when the load increases, the service response performance deteriorates rapidly. Therefore, the processing performance of the entire business flow also deteriorates at the same time. Conventionally, this rapid performance deterioration has been avoided by a communication flow control using a device such as a load balancer that performs load balancing or a lower protocol of a network such as TCP.

  For example, the following content is described in Patent Document 1 as such an example. The data request transmission means accepts data transmitted from the client and relays it to the server computer. The data request transmission unit receives the comparison result by the threshold comparison unit that accumulates the number of data processing requests received from the client per unit time and compares it with a threshold value, and the measured number of data processing requests is received from the threshold output unit. If it is below the output threshold, the data request packet received by the data request receiving means is transferred to the server computer requested by the client, and the number of data processing requests measured by the server data request measuring means falls below the threshold. If it exceeds, the data request receiving means holds the received data request packet in the hold buffer.

JP 2003-224607 (6th to 8th pages, FIG. 1)

  In Patent Document 1 described above, there is a problem that logical high-load control by an upper layer executing a Web service, such as control by service contents or a service provider, cannot be performed.

  Specifically, in an environment where processing is performed by cooperation of a plurality of Web service functions as in the present invention, the service requester and the service provider may be switched depending on the service content. The conventional service requester becomes the service provider, and the service provider replaces the service requester. In Patent Document 1, there is a first problem that the number of requests from the client side to the server is managed as a threshold, and there is no content for threshold management in both directions, so that it is not possible to cope with such a situation. .

  In general, with the expansion of Web service functions, control information specific to the function occurs between the service requester and the service provider based on the HTTP protocol. The exchange of load information, etc.) has a second problem that it cannot be identified whether the message is related to a service request or control information, or is difficult to follow.

  According to the present invention, a load control unit is provided in a proxy server that performs HTTP protocol communication, thereby solving the above-described problems without remodeling the Web service function main body. A service request relay method, a relay server, and a program thereof are proposed.

  A first Web service request relay system according to the present invention includes a relay unit that transmits and receives a Web service request, a connection with the relay unit through TCP / IP communication, and an HTTP proxy setting for the relay unit. Web service execution means for receiving a Web service request received by the means and processing it to generate a new Web service request and transmitting it to the relay means, and a first server and a second server, respectively, A Web service request relay system including a network connecting the second server and the second server, wherein the relay unit of the first server is the Web service execution unit of the first server. When a new Web service request addressed to the second server is received, the new service request Recognizing the address of the second server from the TTP header, confirming whether there is a high load notification of the second server from the relay means of the second server, and if there is a high load notification, the new Web service When the request is suspended and there is no high load notification, the address is deleted from the HTTP header and transmitted to the second server. When the relay unit of the second server receives the new Web service request, When the high load state of the Web service execution unit of the second server is determined and the load is high, a high load notification is transmitted to the relay unit of the first server.

  The second Web service request relay system of the present invention is the first Web service request relay system, wherein the relay means of the second server is a unit by the Web service execution means of the second server. A threshold value defining the upper limit number of Web service request processes in time is compared with a value obtained by counting the number of new Web service requests received in the latest unit time from the first server, and the counted value is the threshold value. Is exceeded, the high load notification is transmitted to the relay means of the first server by the HTTP protocol, and the relay means of the first server receives the response message to the relay means of the second server. Is transmitted.

  According to a third Web service request relay system of the present invention, in the second Web service request relay system, the relay unit of the second server receives the Web service request or at the unit time. Each time an interrupt from the set interval timer occurs, the high load state of the second server is determined, and when a transition is made from a high load state to a non-high load state, a high load notification is transmitted to the first server. If it is, the high load release notification is transmitted to the first server by the HTTP protocol, and when the relay unit of the first server receives it, the response message is transmitted to the relay unit of the second server. It is characterized by doing.

  According to a fourth Web service request relay system of the present invention, in the third Web service request relay system, the relay unit of the first server receives a high load release notification from the second server. Then, if there is a new Web service request for which transmission to the second server is suspended, it is extracted, and the address of the second server is deleted from the HTTP header and transmitted to the second server. To do.

  A fifth Web service request relay system according to the present invention includes a relay unit for transmitting and receiving a Web service request, a connection with the relay unit by TCP / IP communication, and an HTTP proxy setting for the relay unit. Web service execution means for receiving a Web service request received by the means and processing it to generate a new Web service request and transmitting it to the relay means, and a first server and a second server, respectively, And a network connecting the second server, the first server receives and executes a Web service request from the second server, and the second server A Web service request relay system that receives and executes a Web service request from a first server, wherein the Web service request relay system is in front of the first server. When the relay unit receives a new Web service request destined for the second server by the Web service execution unit of the first server, the relay unit addresses the second server from the HTTP header of the new service request. And confirms whether there is a high load notification of the second server from the relay unit of the second server, and if there is a high load notification, the new Web service request is suspended and there is no high load notification In this case, the address is deleted from the HTTP header and transmitted to the second server, and when the relay unit of the second server receives the new Web service request, the Web service execution unit of the second server When the high load state is determined and the load is high, a high load notification is transmitted to the relay unit of the first server, and the relay unit of the second server When a new Web service request destined for the first server is received by the Web service execution unit of the second server, the address of the first server is recognized from the HTTP header of the new service request. The presence of the high load notification of the first server from the relay unit of the first server is confirmed, and if there is a high load notification, the new Web service request is suspended, and if there is no high load notification, an HTTP header And when the relay unit of the first server receives the new Web service request, the high load on the Web service execution unit of the first server is deleted. When the state is determined and the load is high, a high load notification is transmitted to the relay unit of the second server.

  A sixth Web service request relay system according to the present invention includes a relay unit for transmitting and receiving a Web service request, a connection with the relay unit through TCP / IP communication, and an HTTP proxy setting for the relay unit. Web service execution means for receiving a Web service request received by the means and processing it to generate a new Web service request and sending it to the relay means, respectively, and the relay means, respectively A third server comprising web service execution means for processing a received web service request and generating a new web service request, the first server, the second server, and the third server To the second server from the first server or the third server to the second server In the Web service request relay system for transmitting a new Web service request, the relay unit of the first server receives a new Web service request from the Web service execution unit of the first server. Recognizing the address of the second server from the HTTP header, confirming whether there is a high load notification of the second server from the relay means of the second server, and if there is a high load notification, the new Web service When the request is suspended and there is no high load notification, the address is deleted from the HTTP header and transmitted to the second server. When the relay unit of the second server receives the new Web service request, When the high load state of the Web service execution unit of the second server is determined and the load is high, the relay unit of the first server And transmitting a notification.

  According to a seventh Web service request relay system of the present invention, in the sixth Web service request relay system, the relay means of the second server applies a high load to the third server by an HTTP protocol. After the request message to be notified is transmitted, it is confirmed whether or not a response message is received from the third server, the third server that does not receive the response message recognizes that it does not have the relay means, The load release notification is not transmitted to the third server.

  According to an eighth Web service request relay system of the present invention, in the seventh Web service request relay system, the relay means of the second server does not have the relay means in the third server. Recognizing this, when the second server is under a heavy load, when a processing result for the new Web service request received from the third server is transmitted to the third server, a preset time The transmission is delayed.

  A ninth Web service request relay system according to the present invention includes a first server and a second server having a process for transmitting a Web service request, a process for executing a received Web service request, and a relay server for relaying the Web service request. And a network connecting each of the servers, wherein the first server has a setting to use an HTTP proxy for the relay server, and when sending a Web service request, an HTTP header The relay server has setting information for the number of Web service request messages that can be processed in the unit time of the second server and the unit time length, The second server extracted from the HTTP header of the Web service request received from the first server. The number of Web service requests for the most recent unit time addressed to the second server is counted based on the address, the counted number of Web service requests is compared with the threshold value, and if it is equal to or less than the threshold value, the Web service request is transmitted. When the threshold value is exceeded, the second server determines that the load is high and puts the Web service request on hold.

  According to a tenth Web service request relay system of the present invention, in the ninth Web service request relay system, the relay server uses the address of the second server extracted from the HTTP header of the received Web service request. Originally, the held Web service request is stored in time series according to address.

  According to an eleventh Web service request relay system of the present invention, in the ninth Web service request relay system, the relay server is an interval timer set every time a Web service request is received or based on the unit time. Each time an interrupt occurs, the high load state of the second server is determined. If the second server is not in a high load state, the presence or absence of a pending web service request is confirmed. The address of the second server is removed from the HTTP header of the Web service request and transmitted to the second server.

  In a twelfth Web service request relay system according to the present invention, in the ninth Web service request relay system, the relay server is set as a proxy server of the first server, and is connected to the second server. It is characterized by having high load control means by suppressing the number of transmissions of the Web service request.

  According to a thirteenth Web service request relay system of the present invention, in the ninth Web service request relay system, the second server transmits a Web service request to the first server, and the first server When receiving and executing it, the relay server is set as the proxy server of the second server, and the number of Web service request messages that can be processed in the unit time of the first server and the unit time length are set. By providing the information, the relay server further includes high load control means by controlling the number of transmissions of the Web service request to the second server.

  A first Web service request relay method according to the present invention includes a first server and a second server having a process for transmitting a Web service request, a process for executing a received Web service request, and a relay server for relaying the Web service request. And a Web service request relay method in a Web service system including a network connecting the servers, wherein the first server has a setting to use an HTTP proxy for the relay server, and transmits a Web service request. A step of setting the address of the second server in an HTTP header and transmitting to the relay server, the relay server including the number of Web service request messages that can be processed in a unit time of the second server, and It has setting information of unit time length and is taken from the HTTP header of the received Web service request. Counting the number of Web service requests in the most recent unit time addressed to the second server based on the address of the second server issued, and the relay server compares the counted number of Web service requests with the threshold value. And transmitting the Web service request if the threshold is less than the threshold; suspending the Web service request if the relay server exceeds the threshold; and receiving the Web service request. Each time or whenever an interrupt from the interval timer set based on the unit time occurs, the high load state of the second server is determined, and the transition from the high load state to the non-high load state is confirmed. If there is a pending web service request, the pending web service request is confirmed. Remove the address of the second server from the HTTP header of the characterized in that it comprises the steps of: transmitting to the second server.

  The second Web service request relay method of the present invention comprises a relay means for transmitting / receiving a Web service request, a connection with the relay means by TCP / IP communication, and setting of an HTTP proxy for the relay means, and the relay Web service execution means for receiving a Web service request received by the means and processing it to generate a new Web service request and transmitting it to the relay means, and a first server and a second server, respectively, A web service request relay method in a web service system including a server connecting the second server and a network connecting the second server, wherein the relay unit of the first server A first step of transmitting to the Web service execution means of one server by TCP / IP communication; The Web service execution means of the first server processes the received Web service request to generate a new Web service request, sets the address of the second server as the destination of the HTTP header, and sets the first server Transmitting to the relay means of the server; and when receiving the new Web service request, the relay means of the first server, based on the address of the second server as the destination of the HTTP header, Determining whether or not the server has a high load; and if the second server has a high load, the relay means of the first server sends the new Web service request to the address of the second server. And the relay means of the first server, when the second server is not heavily loaded, to the HTTP of the new Web service Deleting the address of the second server from the server and transmitting it to the second server; and when the relay means of the second server receives the new Web service request, Determining the high load state of the Web service execution unit of the server; and when the relay unit of the second server determines that the Web service execution unit is highly loaded, the relay of the first server A step of transmitting a high load notification to the means by HTTP protocol, and the relay means of the second server confirms a transition from a high load state to a non-high load state as a result of the determination of the high load state, Sending a high load release notification to the relay means of the first server by HTTP protocol; and the relay means of the first server When a high load notification is received from the second server, the second server stores it as being in a high load state, and upon receiving a high load release notification, it is suspended based on the address of the second server Checking the presence or absence of the new Web service request, and if there is, deleting the address of the second server from the HTTP header of the new Web service request and transmitting to the second server. It is characterized by.

  The relay server of the present invention receives a Web service request in which an HTTP header is set to the address of the second server that executes the Web service request from the first server that is set to use the HTTP proxy as the relay server. Means, the number of Web service request messages that can be processed per unit time of the second server, and the setting information of the unit time length, and the address of the second server extracted from the HTTP header of the received Web service request A means for counting the number of web service requests for the most recent unit time addressed to the second server based on the above, and comparing the counted number of web service requests with the threshold value. Means for transmitting to the second Web server, and if the threshold is exceeded, the Web service request is suspended Every time a Web service request is received or every time an interrupt from an interval timer set based on the unit time occurs, the high load state of the second server is determined, and the high load state is not high. When the transition to the state is confirmed, it is confirmed whether there is a pending web service request. If there is a pending web service request, the address of the second server is removed from the HTTP header of the suspended web service request. Means for transmitting to the server.

  The program of the present invention sends a Web service request in which the address of the second server that executes the Web service request is set in the HTTP header from the first server that is set in the relay server to use the HTTP proxy. The second server having the receiving step, the number of Web service request messages that can be processed per unit time of the second server and the setting information of the unit time length, and extracted from the HTTP header of the received Web service request Counting the number of Web service requests for the most recent unit time addressed to the second server based on the address of the second server, comparing the counted number of Web service requests with the threshold, and if the threshold is less than the threshold, the Web service request To the second web server, and before the threshold is exceeded A step of suspending the web service request, and whenever the web service request is received or whenever an interrupt from the interval timer set based on the unit time occurs, the high load state of the second server is determined, When the transition from the load state to the non-high load state is confirmed, the presence / absence of the suspended web service request is confirmed. If there is a suspended web service request, the address of the second server is determined from the HTTP header of the suspended web service request. Removing and transmitting to the second server.

  The first effect is to keep the load of a plurality of Web services uniform and prevent extreme performance deterioration of local service processing execution from becoming a bottleneck of the entire service.

  For this reason, the service requester voluntarily suppresses a request for a service whose processing performance deteriorates rapidly when the load on a specific service exceeds a certain amount. Specifically, the communication delay means for avoiding this is not performed by a computer that receives a service request with high load, but by a computer that transmits a service request. It is possible to prevent the processing load due to the introduction of the apparatus from increasing.

  Next, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 1, the first embodiment of the present invention refers to a plurality of Web servers 10, a single service management device 20, a client device 30 such as a personal computer that requests a service from the Web server 10, and a Web A network 40 including the Internet that connects the server 10, the service management apparatus 20, and the client apparatus 30 is configured. The Web server 10, the service management apparatus 20, and the client 30 are all computers.

  In the present invention, the Web service management system controls the occurrence of a local high load state of the plurality of Web servers 10 and performs control so that the local service stop does not stop the entire system. For this purpose, the relay device 11 is introduced into the Web service management system, and the communication flow between the Web servers 10 is controlled.

  The Web service management system includes a relay device 11 that operates on the Web server 10 and a service management device 20, and manages each service content and operation status provided by the plurality of Web servers 10.

  The service management apparatus 20 includes a service composition calculation unit 21, an operation information storage unit 22 that stores load information of the Web server 10, and a service definition that stores location information and service definition information of the Web server 10 that provides this service. And an information storage device 23. The operation information storage unit 22 and the service definition information storage device 23 are storage devices connected to the service management device 20 including a magnetic disk device or the like.

  The Web server 10 is responsible for the business of each Web server 10 such as an order receiving system, for example, the business name, the business identifier that uniquely identifies the business, the business request source Web service name, and the business request destination Web service. Information including name, response / transfer type, etc. is stored. One business is executed by cooperation of one or more Web servers 10, but each Web server 10 stores information on a portion that it is in charge of in the entire business flow.

  The service management apparatus 20 has previously received information in charge of each Web server 10 in the business flow from the Web server 10 for each business in charge of each Web server 10. The service management device 20 stores the received partial information of the business flow in the service definition information storage device 23 for each business and for each transmitted Web server 10.

  For example, when an order receiving system is considered as one business flow, the order receiving system performs order management (identity confirmation, content validity confirmation), inventory management (inventory confirmation, order subtraction), delivery management (delivery slip creation, delivery) It consists of several functions (= Web service) such as car distribution. In this case, the message flow between the Web servers 10 such as order management → inventory management → delivery management is determined in advance in the order receiving system, and processing for transferring to the next Web server 10 in each Web server 10 is described. It will be.

  When the service management apparatus 20 finishes collecting the business flow information from the associated Web server 10, the service management apparatus 20 reads the collected definition information of the Web server 10 from the service definition information storage apparatus 23. In the service composition calculation unit 21 of the service management apparatus 20, when each Web server 10 executes one service in cooperation with the adjacent Web server 10 in terms of service, the individual Web servers 10 are repeated by repeating the association. Are associated with the entire service relationship, and the business flow between the entire Web server 10 is constructed. This flow can be regarded as a processing flow of a specific job using each Web server 10.

  The Web server 10 includes a relay device 11, a Web service 12, and service definition information 15. The web service 12 includes a communication unit 13 and a service execution unit 14. Further, as described above, the service definition information 15 defines service contents and input and output information for service execution.

  The service definition information 15 indicates whether the response process is completed by a Web service 12 with a business service or the transfer process in which the business service is not completed with the processing of one Web server 10 and continues to use another Web server 10. In the case of the transfer process, a call to another Web server 10 and its argument are defined.

  The relay device 11 is mounted on the same computer as the Web service 12, and includes a message reception unit 110, a message transmission unit 111, a service information analysis unit 112, a message storage unit 113, a transmission destination storage unit 114, The service information storage unit 115 is an application program that operates as a proxy server with a defined port number.

  The functions of the Web server 10 are executed by a plurality of software programs that operate on a computer.

  The relay device 11 that is the request source of the Web service receives the load information of the Web server 10 from the relay device 11 of the Web server 10 that is the Web service request destination. The load information includes contents in a high load state exceeding the upper limit of the number of Web service request messages that can be processed per unit time. The relay device 11 of the Web service request source that has received the high load notification from the Web service request destination by this load information suppresses transmission to the high load Web server 10. By controlling the communication flow between the Web servers 10, it is possible to operate the entire system in which the processes of the plurality of Web servers 10 are leveled.

  The relay device 11 of the Web service request source performs transmission delay and communication by disassembling, storing, reconstructing, and transmitting the Web service request message to the Web server 10 of the Web service request destination that has become a heavy load. Control the flow.

  The relay device 11 relays an HTTP (Hyper Text Transfer Protocol) request from one Web server 10 to another Web server 10. At this time, since the communication of the HTTP protocol is established on a TCP (Transmission Control Protocol) connection, the header information of the normal HTTP request does not include the address information of the destination computer. However, if the communication unit 13 of the Web service 12 is set to use a standard HTTP proxy prepared in a program using HTTP, the HTTP of the message passing from the Web service 12 to the relay device 11 on the same computer. Destination information can be embedded in the request header. The relay device 11 can refer to the destination information to identify the destination Web server 10 and determine whether or not the destination Web server 10 has a high load.

  The above is organized as shown in FIG.

In FIG. 12, it is assumed that a Web service request message is transmitted from the server (B) to the server (A).
After the introduction of the relay device, the address of the server (A) is included in the protocol header of the message transmitted from the Web service of the server (B) to the relay device of the same server (B).
HTTP header: described as POST http: // abc / ... Thereby, the relay apparatus can know the address of the destination server. The relay device of the server (B) makes a TCP connection to the abc of the server (A) based on http: // abc / ... of the HTTP header and transmits a normal POST.

  As a result, when the relay apparatus 11 receives a high-load notification from the Web server 10 that is the Web service request destination and the request destination is a high load, the relay apparatus 11 decomposes, temporarily stores, and loads the transmission message about the Web service request destination. It is possible to reconstruct and send the disassembled message at the time of release. In order to ensure consistency between the existing Web service and the contents of the present invention, the port number of the relay apparatus 11 uses the port number defined in the existing Web service, and the existing port number of the Web service 12 is used. Change to a different number. Transmission / reception of messages between the relay apparatus 11 and the Web service 12 on the same Web server 10 is executed by TCP / IP communication using port numbers defined externally.

  By implementing as described above, the relay device 11 is a mechanism that is externally attached to the Web service 12, so that the existing Web service and the communication protocol between the Web services can be used transparently. The Web service 12 has an effect that the service can be transparently provided even after the introduction of the inter-service communication control device, without modifying and modifying the function of the existing Web service 12.

  The relay device 11 disassembles the Web service request message according to the analysis result of the destination computer address of the header of the HTTP request, and is linked to the destination storage unit 114 that stores the header information for each destination of the Web server 10. By having a message storage unit 113 that stores the body of the message, a plurality of Web service request messages are queued and accumulated by destination.

  When the high load on the destination computer is released, the relay device 11 extracts the IP address, URL, and message body of the destination computer from the information stored in the transmission destination storage unit 114 and the message storage unit 113 temporarily stored, and again complete HTTP Requests can be assembled and communication can be resumed. Thereby, the relay apparatus 11 can control the service request to the high load relay apparatus 11 so as not to exceed a certain number of service requests.

  Further, the service management device 20 that manages the entire Web server 10 can precisely manage the load status at a specific time point by acquiring and recording the load information held by the relay device 11.

  Note that the client device 30 can access the service management device 20 from a browser and check the load state of each Web server 10 in advance, and can transmit a Web service request message when the load state is not high.

  Next, the operation of the best mode for carrying out the present invention will be described with reference to the drawings.

  A case where a plurality of Web services are configured will be described with reference to FIG. 7 and a flowchart of FIG. In this case, since the message is processed between a plurality of services, the response between services is asynchronous unless otherwise noted.

  First, settings on the service providing side of the Web server 10 will be described.

  The Web service provider sets the upper limit value of the number of service requests that can be processed per unit time by the service execution unit 14 of the Web server 10 to the relay device 11 as the first threshold value, and the number of service requests exceeding the first threshold value. When the number of service requests drops below a certain value after receiving, a second threshold value for returning from the high load state to the normal state is set. The time length of the unit time is set using units such as seconds and minutes. From the above, there is a relationship of first threshold> second threshold.

  Further, the Web service provider also sets the IP address and port number of the service management apparatus 20 (step S1). For the setting operation, an input device such as a keyboard and a display (not shown) connected to the Web server 10 and a display device are used. Information such as the threshold value and the unit time length set in this way is stored in the service information storage unit 115 such as a magnetic disk device connected to the Web server 10.

  Further, the Web service provider registers the content of the service provided by the Web server 10 with respect to the service definition information 15 such as a magnetic disk device (step S2).

  Further, the Web service provider uses the input device and the display device to set the use of the HTTP proxy for the Web service 12, and sets the IP address and port number in the execution environment (Step S3).

  As a result, all messages transmitted and received by the Web server 10 according to the HTTP protocol pass through the relay device 11. The port number that the relay apparatus 11 has is set to the port number used so far of the Web service 12, and another port number is assigned to the Web service 12.

  The relay device 11 takes in the service content of its own Web service 12 from the service definition information 15 and transmits it to the service management device 20 as an HTTP message using the IP address and port number of the service management device 20 (step). S4).

  In the service management device 20, the received service definition information 15 is stored in the service definition information storage device 23. When the service composition calculation unit 21 of the service management device 20 receives the service definition information 15 from the related Web server 10, the service composition calculation unit 21 synthesizes the service definition information 15 provided from each Web server 10 for each business, The entire business flow spanning the server 10 is created and stored in the service definition information storage device 23.

  Since the entire business flow is displayed on the display device of the service management device 20 as required by the administrator or the like, the contents can be referred to.

  The operation on the service providing side after the above settings are performed will be described with reference to FIG.

  Briefly describing FIG. 7, the Web server 10C includes the relay device 11C, and transmits a Web service request to the Web server 10A including the relay device 11A. The Web server 10D that does not have the relay device 11 also transmits a Web service request to the Web server 10A. When the Web server 10A receives the service request from the Web server 10C and terminates the service execution that it is in charge of, the Web server 10A continues to issue the next service execution request to the Web server 10B including the relay device 11B. The Web server 10B receives a service execution request from the Web server 10A and executes a service process defined for itself.

  Hereinafter, description will be given with reference to the flowchart of FIG.

  When the relay apparatus 11A of the Web server 10A is started up, the setting information such as the first threshold value, the second threshold value, and the unit time set previously is extracted from the service information storage unit 115 and the relay apparatus 11A itself Store in the storage memory (step S5).

  The relay device 11A transmits to the service management device 20 information including the first and second threshold values and the unit time by using the HTTP message as the event type with the start of the service by the Web server 10 (step S6).

  When the service management apparatus 20 receives the information, the service management apparatus 20 writes the information together with the received time in the storage unit related to the Web server 10A of the operation information storage unit 22 (step S7).

  The service information analysis unit 112 of the relay device 11A clears the received message storage area of the main storage memory and the high load determination flag for counting the number of service request messages received per unit time (step S8).

  In the received message storage area, the time according to the hour, minute and second when the service request message is received and the IP address of the message transmission source are set as information of one entry, and each time the service request message is received, the reception request storage area sequentially It will be remembered. When the message is stored up to the final position of the received message storage area, the message is returned to the start position of the received message storage area and stored in a circular manner. The received message storage area has a size capable of storing all entries for the maximum number of messages expected to be received per unit time.

  The Web server 10C transmits a Web service request message to the Web server 10A in a timely manner (Step S9). Further, the Web server 10D having no relay device function also transmits a Web service request message to the Web server 10A in a timely manner (step S9).

  The message receiving unit 110 of the relay device 11A receives a service request message from the Web server 10C or the Web server 10D (Step S10). The message reception unit 110 passes the received service request message to the message transmission unit 111. The message transmission unit 111 transmits the passed message to the Web service 12 using TCP / IP using the port number of the Web service 12 (step S11). The message transmission unit 111 stores the time when the message is transmitted to the Web service 12 as the final transmission time.

  At the same time, when receiving a message from the Web server 10C, the message receiving unit 110 of the relay device 11A obtains the received time, generates an entry together with the IP address of the source of the message extracted from the TCP / IP function, and receives the received message. The entry is stored at a position next to the position where the message received immediately before in the storage area is stored (step S12).

  Next, the number of messages received per unit time is counted by referring to the time of each entry stored in the received message storage area, and the number of messages received between the current time and the time that has gone back by the unit time. Is obtained (step S13).

  Next, it is determined by the high load determination flag whether the current state is a normal state or a high load state (step S14). If the load is high, the number of received messages per unit time is compared with the second threshold (step S15). If the result of the comparison is less than or equal to the second threshold, it is determined that the normal state has been restored from the high load state, and the process jumps to (Step S40). If the second threshold value is exceeded, it is determined that the high load state is still continuing, and the process jumps to step S31.

  If the result of determination in step S14 is normal, the counted value is compared with the first threshold value (step S16). As a result of this comparison, if the counted value is equal to or smaller than the first threshold, the message reception waiting state is entered as a normal state.

  When the counted value exceeds the first threshold, the high load determination flag is turned on (step S17), and the process jumps to (step S31).

  Hereinafter, description will be made with reference to the flowchart of FIG.

  The message transmitted from the relay device 11A in (Step S11) is received by the Web service 12A. The service execution unit 14 of the Web service 12 </ b> A extracts the business identifier described in the head of the service request message received from the communication unit 13, for example, and compares it with the business identifier previously extracted from the service definition information 15.

  FIG. 8A illustrates the definition information of the service definition information 15, which is composed of a business identifier, the number of arguments, definition information on the data format of each argument, and the like. As a result of the comparison, the service execution unit 14 extracts the number of arguments and the data format information of each argument from the service definition information 15 having the matched business identifier, and analyzes the body information of the Web service request message based on this. Processing for the Web service request message is executed (step S18).

  During service processing, the service execution unit 14 searches the service definition information 15 at the time of response or transfer illustrated in FIG. 8B for a service message that matches the business identifier in the service message. The service definition information 15 in the case of transfer includes a transfer / response type (transfer in this case), a transfer destination Web service identifier, the number of arguments, and the data format of each argument. The service definition information in the case of a response includes a transfer / response type (response in this case), a normal / abnormal type, the number of response values, and a data format of each response value.

  The service execution unit 14 searches for definition information that matches the business identifier, and generates a result according to the format defined by the searched service definition information 15 (step S19). The service execution unit 14 passes the generated result to the communication unit 13. The communication unit 13 determines whether to respond to the service requester according to the searched service definition information 15 or to transfer to another Web server 10 (step S20), and the communication unit 13 accordingly determines the transfer destination and the response destination. The destination information of the HTTP header is generated (in this case, the transfer destination Web server 10 is assumed to be the Web server 10B), and this is added to the message body created by the service execution unit 14 (step S21). The information generated in this way is transmitted to the relay device 11A by TCP / IP communication (step S22).

  The message transmission unit 111 of the relay apparatus 11A obtains the transmission destination Web server 10 from the destination information of the HTTP header created by the Web service 12 (step S23).

  Based on the determined transmission destination, the high load notification destination table is referred to and it is confirmed whether the transmission destination is registered (step S24). If registered, it is checked whether or not the transmission destination has a relay device function (step S25). If the Web server 10D does not have the relay device function, the set transmission delay time is delayed (step S26). This transmission delay time can also be registered in the service information storage unit 115 by the Web service provider.

  Next, the high load state of the response destination or the transfer destination is searched from the high load notification source table according to the destination information of the header to check whether or not it is registered (step S27). If there is no registration and the load is not high, the relay device 11A deletes the destination information from the HTTP header to generate a new HTTP header, specifies the retrieved destination information as an argument at the time of TCP / IP call, and sends the destination Web It transmits to the server 10 (step S28). And we wait for various events.

  For example, when the Web server 10B is registered in the high load notification source table and the transmission destination is in a high load state, the HTTP request message to be transmitted is disassembled, and the destination Web server of the HTTP header is sent to the transmission destination storage unit 114 The address, URL, and the like of 10A are stored, the message body is stored in the message storage unit 113, and the transmission of the message is suspended (step S29).

Then, various event waiting states are entered. (In the flowchart, it is described so as to start receiving processing of the next Web service execution request.)
When receiving the message transferred from the Web server 10A, the Web server 10B refers to the service definition information 15 and executes a service for the transferred message (step S30).

  Hereinafter, description will be made using the flowchart of FIG.

  When the number of Web service request messages received within a unit time exceeds the first threshold value or the second threshold value as a result of comparison with the threshold value by the relay device 11A in (Step S16), the relay device 11A A request message for notifying a high load is transmitted to the transmission source Web server 10C or Web server 10D as another message of the HTTP protocol (step S31).

  Further, the relay device 11A also transmits a message including the number of received messages per unit time and having a high event type to the service management device 20 (step S32). The service management apparatus 20 stores this content in the operation information storage unit 22 together with the identification name, reception time, etc. of the Web server 10 (step S33).

  The service information analysis unit 112 of the relay device 11A, as illustrated in FIG. 9A, the IP address and port of the relay device 11C or the relay device 11D that made the notification in the high load notification destination table of the main memory. The number, the notification time, and the count value of the received message at that time are stored (step S34).

  When the relay device 11C receives a request message notifying the high load from the relay device 11A, the main memory stores the IP address and port number of the Web server 10A and information in the request message as illustrated in FIG. 9B. The high load notification source table on the memory is stored (step S35). The relay device 11C transmits a response message in response to the high load notification to the relay device 11A (step S36).

  The relay device 11A confirms whether or not a response message to the high load notification message has been received (step S37). When the response message is received from the Web server 10C, the relay device 11A recognizes that the Web server 10C has the relay device function of the present invention, and the Web server 10C stores the relay function in the high load notification destination table (step S38). ). And we wait for various events.

  When the transmission source is the Web server 10D that does not have the relay device function of the present invention, even if a high-load notification request message is received, the content of the request message cannot be read. Do not send to 10A.

  Since the Web server 10A does not receive the response message from the Web server 10D even after a certain period of time, the Web server 10D recognizes that the Web server 10D does not have the relay function in the present invention, and stores it in the high load notification destination table. The Web server 10D stores that there is no relay function (step S39). 11 A of relay apparatuses will be in the message reception waiting state, after finishing the above process.

  Hereinafter, a description will be given using the flowchart of FIG.

  When the service information analysis unit 112 of the Web server 10A confirms the transition from the high load state to the normal state, the service information analysis unit 112 turns off the high load determination flag (step S40), and the event type is event information that returns to the normal state. Is transmitted to the service management apparatus 20 (step S41). The service management apparatus 20 that has received this information stores it in the operation information storage unit 22 together with the transmitted identification name and reception time of the Web server 10 (step S42). The web service management unit 20 edits a plurality of events received from each web server 10 retroactively from the present time. The editing result is responded to an inquiry from the browser of the client device 30 as the load status of each Web server 10.

  Further, it is checked whether there is information registered in the high load notification destination table (step S43). If there is no information registered in the high load notification destination table, it enters a waiting state for a received message.

  If there is information registered in the high load notification destination table, the relay device 11A performs the following processing on all the relay devices 11 having the relay device function based on the contents of the high load notification destination table.

  Based on the IP address and port number of the registered high load notification destination table, a high load release notification is transmitted to each notification destination (relay device 11C, etc.) by an HTTP message (step S44). The relay device 11C or the like that has received the cancellation notification responds with a response message to the relay device 11A (step S45).

  When the relay device 11A receives a response message from the relay device 11C or the like of the Web server 10 that has transmitted the release notification, the relay device 11A searches the high load notification destination table based on the IP address of the Web server 10 that is the response message transmission destination, and applies. The entry is deleted (step S46). The Web server 10 that is registered in the high load notification destination table and does not have the relay device function continues to be registered without being deleted.

  The Web server 10C that has received the high load release notification from the relay device 11A searches the high load notification source table with the IP address of the Web server 10A that has transmitted the release notification, and deletes the corresponding entry (step S47).

  The relay device 11C confirms whether there is a message for suspending transmission related to the Web server 10A that has transmitted the release notification (step S48). If there is a pending message, the event type is canceled, and the name of the service request destination web server 10 to be suspended and the number to be suspended are transmitted to the service management apparatus 20 (step S49). Receiving this, the service management apparatus 20 stores it in the operation information storage unit 22 (step S50).

  The relay device 11C compares the destination address of the HTTP header stored in the temporarily stored destination storage unit 114 with the address of the Web server 10A that has notified the cancellation of the high load for the message held in the destination storage unit 114. .

  If they match, the address and URL of the destination Web server 10A are extracted from the destination storage unit 114, the message body is extracted from the information in the message storage unit 113, a complete HTTP request is assembled again, and the destination address in the HTTP header is deleted. To the Web server 10A (step S51).

  Note that even if a high load state occurs in the Web server 10A, a state in which the Web server 10A does not receive a message for more than a unit time may continue thereafter. In this case, the processing of the Web service request message is stagnated even though the high load state is released. In order to prevent this, the following processing is executed.

  The relay apparatus 11A sets an appropriate time equal to or less than the unit time in the interval timer and periodically generates a timer interrupt. When a timer interrupt occurs, the high load determination flag is checked, and if it is in a high load state, it is determined whether the time difference between the last transmission time stored previously and the current time exceeds the unit time. If the unit time is exceeded, the process jumps to (Step S40) to execute the release processing of the high load state.

  In the above description, it is assumed that the web server 10A receives and executes the web service request of the web server 10C. At the same time, the web server 10C receives and executes the web service request from the web server 10A. In some cases. Therefore, the first threshold value, which is the number of processes that can be processed per unit time of the Web service function of the Web server 10C, and the second threshold value described above are set in the relay device 11C. As a result, the Web server 10C and the Web server 10A can perform mutual high load control in both directions.

  Next, a second embodiment of the present invention will be described.

  FIG. 10 is a block diagram for explaining the configuration of the second embodiment of the present invention.

  In the first embodiment, the Web service 12 and the relay device 11 coexist on the same Web server 10, but in the second embodiment, each operates on an independent computer, and each computer is a network including the Internet. It is assumed that the configuration is connected.

  In FIG. 10, Web server 1 to Web server L transmit a Web service request to one of Web servers N from Web server M, which is a computer that executes the received Web service request, via relay device 11. . The relay device 11 is a computer and operates by a software program.

  At this time, it is assumed that the relay device 11 is set as a proxy server of the Web server L from the Web server 1 that transmits the Web service request. Further, as in the first embodiment, the setting to use the HTTP proxy prepared as a standard in the program using HTTP is performed from the Web server 1 to the communication unit 13 of the Web service 12 of the Web server L. The destination information can be embedded in the HTTP request header of the Web service request message transmitted from the Web service 12 of the Web server L to the relay device 11 from the Web server 1. The relay device 11 refers to the destination information, identifies the destination Web servers M to N, determines the destination, determines whether the destination Web servers L to M have a high load, and performs transmission control. Is possible.

  The relay device 11 is preset with a unit time for each of the Web servers M to N having the same contents as the first embodiment and a threshold that the Web servers M to N can process in the unit time. And The Web servers M to N do not receive Web service requests from the Web servers 10 other than the Web servers 1 to L. However, if a message is received from another Web server 10, past results, etc. To determine thresholds corresponding to the Web servers 1 to L. It is assumed that the relay device 11 has a received message storage area prepared for each of the Web servers M to N as in the first embodiment.

  Next, the operation of this embodiment will be described with reference to the flowchart of FIG.

  For example, assume that the Web server L transmits a Web service request message to the Web server M. When receiving this message, the relay device 11 recognizes that the transmission destination is the Web server M from the destination address of the HTTP header (step S60).

  The relay device 11 registers an entry including the IP address of the Web server L that transmitted the received message, the time when the message was received, and the transmitted determination flag in the received message storage area prepared in the Web server M. (Step S61). The transmitted determination flag is a flag for determining whether the Web service request has been transmitted to the transmission destination Web servers M to N or not, and is registered as untransmitted when the entry is registered.

  Next, the unit time set in the Web server M is taken out, and the number of messages received between the current time and the time before the current unit time is calculated as the reception time of each entry in the received message storage area of the Web server M. Counting is originally performed (step S62). The number of received messages in this unit time is compared with the threshold value of the Web server M (step S63).

  If the number of received messages per unit time is less than or equal to the threshold, it is checked whether there is a transmission determination flag that is not yet transmitted among the counted received messages (step S64). If there is a message in an unsent state, corresponding messages are sequentially extracted from the heads of the transmission destination storage unit 114 and the message storage unit 113 according to the time series, and both pieces of information are assembled and transmitted to the Web server M (step S65). Then, the previously received message is transmitted to the Web server M (step S66). At this time, the relay apparatus 11 removes the destination address of the HTTP header and transmits it.

  It is assumed that the entry of the received message storage area where the above operation has been performed and the transmitted determination flag of the latest received message have been transmitted. Further, the time of transmission to the Web server M is stored in the corresponding entry in the received message storage area (step S67).

  When the number of received messages per unit time exceeds the threshold value of the Web server M, the relay apparatus 11 determines that the Web server M is in a high load state, and the transmission destination storage unit that classifies the message according to the destination with the Web server M as a destination. It is stored at the end of 114 together with the contents of the HTTP header. The message body information is stored at the end of the message storage unit 113 (step S68).

  When receiving a message exceeding the threshold value over a plurality of unit times, it is determined that the high load state continues over a plurality of unit times. For example, there is a possibility that it will not be executed for more than one unit time.

  Therefore, for example, when the difference between the time when the relay device 11 was last transmitted to the Web server M and the current time has exceeded a unit time set in the Web server M, the relay device 11 stores the received message storage area of the Web server M. An unsent message is taken out, and the corresponding message is transmitted to the Web server M within a range equal to or less than the threshold. For this reason, the relay apparatus 11 sets the minimum unit time among the unit times of the Web servers M to N as the interval timer value, and executes the above-described processing for the Web services M to N every time a notification by a timer interrupt is received. .

  Furthermore, in the above description, the web servers 1 to L on the web service message request side and the servers M to N on the web service request execution side are separated. By setting the received message storage area, there is an effect that high load control in any combination of Web servers 10 becomes possible.

  According to the present invention, the Web service integrated management system can be applied to the integration of enterprise systems. Further, since the response performance of the integrated system can be guaranteed, it can be applied to a system that guarantees service quality.

It is a block diagram explaining the structure of the Example of this invention. It is a flowchart explaining operation | movement of the Example of this invention. It is a flowchart explaining operation | movement of the Example of this invention. It is a flowchart explaining operation | movement of the Example of this invention. It is a flowchart explaining operation | movement of the Example of this invention. It is a flowchart explaining operation | movement of the Example of this invention. It is a block diagram of one structural example used for description of the Example of this invention. It is explanatory drawing of the service definition information of the Example of this invention. It is explanatory drawing of the high load transmission destination / transmission source table of the Example of this invention. It is a block diagram which shows the structure at the time of making the relay apparatus of 2nd Example of this invention into a different apparatus from a web server. It is a flowchart explaining the operation | movement of the 2nd Example of this invention. It is explanatory drawing of operation | movement after performing proxy setting to the relay apparatus of the Example of this invention. It is a block diagram explaining the structure of the conventional web service management system.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Web server 11 Relay apparatus 110 Message receiving part 111 Message transmission part 112 Service information analysis part 113 Message storage part 114 Transmission destination storage part 115 Service information storage part 12 Web service 13 Communication part 14 Service execution part 15 Service definition information 20 Service management Device 21 Service composition calculation unit 22 Operation information storage unit 23 Service definition information storage device 30 Client device 40 Network

Claims (17)

A relay unit for transmitting and receiving a Web service request; a connection with the relay unit by TCP / IP communication; and an HTTP proxy setting for the relay unit; and a Web service request received by the relay unit is received and processed Connecting the first and second servers, the first server, and the second server, each having a Web service execution unit that generates a new Web service request and transmits the request to the relay unit A web service request relay system including:
When the relay unit of the first server receives a new Web service request addressed to the second server by the Web service execution unit of the first server, the relay unit starts from the HTTP header of the new service request. Recognizes the address of the second server, checks whether there is a high load notification of the second server from the relay means of the second server, and if there is a high load notification, holds the new Web service request If there is no high load notification, the address is deleted from the HTTP header and sent to the second server,
When the relay unit of the second server receives the new Web service request, the relay unit of the second server determines the high load state of the Web service execution unit of the second server, and when the load is high, the relay unit of the first server A Web service request relay system, wherein a high load notification is transmitted to the relay means. The relay unit of the second server has a threshold value defining an upper limit number of Web service request processing per unit time by the Web service execution unit of the second server, and a unit time closest to the first server. The received value is compared with a value obtained by counting the number of new Web service requests, and when the counted value exceeds the threshold, a high load notification is transmitted to the relay unit of the first server using the HTTP protocol. 2. The Web service request relay system according to claim 1, wherein the relay unit of the first server transmits a response message to the relay unit of the second server when receiving the request. The relay means of the second server determines the high load state of the second server every time a Web service request is received or every time an interrupt from an interval timer set based on the unit time occurs. When a transition from a high load state to a non-high load state is made, when a high load notification is transmitted to the first server, a high load release notification is transmitted to the first server using an HTTP protocol, and the first server 3. The Web service request relay system according to claim 2, wherein when the relay unit of the server receives the response, a response message is transmitted to the relay unit of the second server. When the relay means of the first server receives a high load release notification from the second server, it takes out a new Web service request for which transmission to the second server is suspended and extracts it. 4. The Web service request relay system according to claim 3, wherein the address of the second server is deleted from the header and transmitted to the second server. A relay unit for transmitting and receiving a Web service request; a connection with the relay unit by TCP / IP communication; and an HTTP proxy setting for the relay unit; and a Web service request received by the relay unit is received and processed Connecting the first and second servers, the first server, and the second server, each having a Web service execution unit that generates a new Web service request and transmits the request to the relay unit The first server receives and executes a Web service request from the second server, and the second server receives a Web service request from the first server. A Web service request relay system to execute,
When the relay unit of the first server receives a new Web service request addressed to the second server by the Web service execution unit of the first server, the relay unit starts from the HTTP header of the new service request. Recognizes the address of the second server, checks whether there is a high load notification of the second server from the relay means of the second server, and if there is a high load notification, holds the new Web service request If there is no high load notification, the address is deleted from the HTTP header and sent to the second server,
When the relay unit of the second server receives the new Web service request, the relay unit of the second server determines the high load state of the Web service execution unit of the second server, and when the load is high, the relay unit of the first server A high load notification is sent to the relay means;
When the relay unit of the second server receives a new Web service request addressed to the first server by the Web service execution unit of the second server, the relay unit starts from the HTTP header of the new service request. Recognizes the address of the first server, checks whether there is a high load notification of the first server from the relay means of the first server, and if there is a high load notification, holds the new Web service request If there is no high load notification, the address is deleted from the HTTP header and sent to the first server,
When the relay unit of the first server receives the new Web service request, the relay unit of the first server determines the high load state of the Web service execution unit of the first server, and in the case of a high load, the relay unit of the second server A Web service request relay system, wherein a high load notification is transmitted to the relay means. A relay unit for transmitting and receiving a Web service request; a connection with the relay unit by TCP / IP communication; and an HTTP proxy setting for the relay unit; and a Web service request received by the relay unit is received and processed The first and second servers each having a new web service request generated and transmitted to the relay unit, and the received web service request without the relay unit are newly processed. A third server provided with a web service execution means for generating a simple web service request, a network connecting the first server, the second server, and the third server, Web server that transmits the new Web service request from the first server or the third server to the second server A scan request relay system,
When the relay unit of the first server receives a new Web service request from the Web service execution unit of the first server, the relay unit recognizes the address of the second server from an HTTP header, and The presence of the high load notification of the second server from the relay means is confirmed, and if there is a high load notification, the new Web service request is suspended, and if there is no high load notification, the address is deleted from the HTTP header To the second server,
When the relay unit of the second server receives the new Web service request, the relay unit of the second server determines the high load state of the Web service execution unit of the second server, and when the load is high, the relay unit of the first server A Web service request relay system, wherein a high load notification is transmitted to the relay means. The relay means of the second server confirms whether or not a response message has been received from the third server after transmitting a request message for notifying the third server of a high load based on the HTTP protocol. 7. The Web service request according to claim 6, wherein the third server that does not receive the notification recognizes that the relay unit is not provided, and does not transmit a subsequent high load release notification to the third server. Relay system. When the relay means of the second server recognizes that the third server does not have the relay means, the new means received from the third server when the second server is heavily loaded 8. The Web service request relay system according to claim 7, wherein when a processing result for a Web service request is transmitted to the third server, the processing result is transmitted with a preset time delay. 1st and 2nd server provided with the process which transmits a web service request, and the process which performs the received web service request, The relay server which relays a web service request, The network which connects each said server is included In the Web service request relay system,
The first server has a setting to use an HTTP proxy with respect to the relay server, and when transmitting a Web service request, sets the address of the second server in an HTTP header and performs transmission.
The relay server has setting information for the number of Web service request messages that can be processed per unit time of the second server and the unit time length, and is extracted from the HTTP header of the Web service request received from the first server. Based on the address of the second server, the number of Web service requests for the most recent unit time addressed to the second server is counted, and the counted number of Web service requests is compared with the threshold value. A Web service request relay system that transmits the Web service request, and when the threshold value is exceeded, the second server determines that the load is high and suspends the Web service request. 10. The relay server stores the suspended web service request according to time based on the address of the second server extracted from the HTTP header of the received web service request according to a time series. Web service request relay system. The relay server determines the high load state of the second server every time a Web service request is received or every time an interrupt from an interval timer set based on the unit time occurs, and is not in a high load state If there is a pending Web service request, the second server address is removed from the HTTP header of the pending Web service request and transmitted to the second server. The Web service request relay system according to claim 9. The said relay server is set as a proxy server of the said 1st server, The high load control means by suppression of the transmission number of the said web service request to the said 2nd server is provided, The said load server is characterized by the above-mentioned. Web service request relay system. When the second server sends a Web service request to the first server and the first server receives and executes it, the relay server is set as a proxy server of the second server, By providing the number of Web service request messages that can be processed per unit time of the first server and the setting information of the unit time length, the relay server can determine the number of transmissions of the Web service request to the second server. 10. The Web service request relay system according to claim 9, further comprising high-load control means by control. Web service including first and second servers having processing for transmitting a Web service request and processing for executing a received Web service request, a relay server for relaying the Web service request, and a network connecting the servers A web service request relay method in a system,
The first server has a setting to use an HTTP proxy for the relay server, and sets the address of the second server in an HTTP header when transmitting a Web service request, and transmits to the relay server When,
The relay server has the number of Web service request messages that can be processed per unit time of the second server and setting information of the unit time length, and the second server extracted from the HTTP header of the received Web service request Counting the number of web service requests in the most recent unit time addressed to the second server based on the address of
The relay server compares the counted number of Web service requests with the threshold value, and transmits the Web service request when the threshold value is less than or equal to the threshold value;
The relay server suspending the web service request if the threshold is exceeded;
The relay server determines the high load state of the second server every time a Web service request is received or every time an interrupt from an interval timer set based on the unit time occurs, When the transition to a non-load state is confirmed, the presence or absence of a pending web service request is confirmed. If there is a pending web service request, the address of the second server is removed from the HTTP header of the suspended web service request. Transmitting to a second server;
A Web service request relay method comprising: A relay unit for transmitting and receiving a Web service request; a connection with the relay unit by TCP / IP communication; and an HTTP proxy setting for the relay unit; and a Web service request received by the relay unit is received and processed Connecting the first and second servers, the first server, and the second server, each having a Web service execution unit that generates a new Web service request and transmits the request to the relay unit A web service request relay method in a web service system including:
The relay unit of the first server transmits a received Web service request to the Web service execution unit of the first server by TCP / IP communication;
The Web service execution unit of the first server processes the received Web service request to generate a new Web service request, sets the address of the second server as the destination of the HTTP header, and sets the first server Transmitting to the relay means of the server of
When the relay unit of the first server receives the new Web service request, the relay unit determines whether or not the second server has a high load based on the address of the second server as the destination of the HTTP header. When,
The relay means of the first server holds the new Web service request based on the address of the second server when the second server is heavily loaded;
The relay unit of the first server deletes the address of the second server from the HTTP header of the new Web service and transmits it to the second server when the second server is not heavily loaded. And steps to
The relay means of the second server, upon receiving the new Web service request, determining a high load state of the Web service execution means of the second server;
When the relay unit of the second server determines that the Web service execution unit has a high load, the relay unit transmits a high load notification to the relay unit of the first server by an HTTP protocol;
When the relay unit of the second server confirms the transition from the high load state to the non-high load state as a result of the determination of the high load state, the relay unit of the first server is notified of the release of the high load. Transmitting by the HTTP protocol;
When the relay means of the first server receives a high load notification from the second server, the relay means stores that the second server is in a high load state, and receives a high load release notification, The presence / absence of the new Web service request suspended based on the address of the second server is confirmed. If there is, the second server address is deleted from the HTTP header of the new Web service request. Sending to the server of
A Web service request relay method comprising: Means for receiving a web service request in which an address of a second server that executes a web service request is set in an HTTP header from a first server that is set in the relay server to use an HTTP proxy;
Based on the address of the second server extracted from the HTTP header of the received Web service request, which has setting information for the number of Web service request messages that can be processed per unit time of the second server and the unit time length. Means for counting the number of Web service requests for the most recent unit time addressed to the second server;
Means for comparing the counted number of web service requests with the threshold value and, if the number is equal to or smaller than the threshold value, transmitting the web service request to the second web server;
Means for deferring the web service request if the threshold is exceeded;
Each time a Web service request is received or every time an interrupt from an interval timer set based on the unit time occurs, the high load state of the second server is determined, and the high load state is changed to a non-high load state. When the transition is confirmed, it is confirmed whether or not there is a pending web service request. If there is a pending web service request, the address of the second server is removed from the HTTP header of the suspended web service request to the second server. A relay server. On the computer,
Receiving a Web service request in which the address of the second server that executes the Web service request is set in the HTTP header from the first server that sets the setting to use the HTTP proxy in the relay server;
Based on the address of the second server extracted from the HTTP header of the received Web service request, which has setting information for the number of Web service request messages that can be processed per unit time of the second server and the unit time length. Counting the number of web service requests in the most recent unit time addressed to the second server;
Comparing the counted number of Web service requests with the threshold value, and transmitting the Web service request to the second Web server if the threshold value is less than or equal to the threshold value;
Suspending the web service request if the threshold is exceeded;
Each time a Web service request is received or every time an interrupt from an interval timer set based on the unit time occurs, the high load state of the second server is determined, and the high load state is changed to a non-high load state. When the transition is confirmed, it is confirmed whether or not there is a pending web service request. If there is a pending web service request, the address of the second server is removed from the HTTP header of the suspended web service request to the second server. And a step of transmitting the program.

Images