JP2010170168A - Flow rate control method and system - Google Patents

Abstract

Even when a failure occurs in a system due to a hardware failure or the like, and a location that becomes a bottleneck changes within the system, an optimal flow control value is set according to the operating state of the system.
An operation monitoring server 80 monitors the operating status of a server of a system composed of a plurality of server groups, and when a non-operating server is detected, the server is based on a change in the number of operating servers. A flow control calculation value 82, which is a processing capacity for each group, is calculated. In addition, the operation monitoring server 80 sets the lowest calculated value as the flow control value 42 of the load distribution device 40. The load balancer 40 executes flow control for limiting the number of access of the system based on the flow control value 42 set by the operation monitoring server 80.
[Selection] Figure 1

Description

  The present invention relates to a technique for performing flow rate control for controlling an information processing amount to be processed in a computer system. Among these, in particular, the present invention relates to a system in which a plurality of computer systems execute predetermined processing in cooperation.

  In recent years, Web sites are used not only as information references but also as Web systems that provide various services such as payment and reservation. Under such circumstances, when access is concentrated in the Web system and an unexpectedly large amount of processing occurs, the system processing may exceed the allowable range of the system processing and the system may become unstable. Therefore, the access flow rate control is an important design element in order to keep the system in stable operation while keeping the inflow access amount within the allowable range of the system.

  Patent Document 1 calculates the number of theoretical threads that can theoretically be activated in a flow control method for a system that executes predetermined information processing in cooperation with a cooperation destination system, and calculates the number of theoretical threads and the cooperation destination system. This is a flow rate control method for calculating an appropriate number of appropriate threads from the system resource usage status and controlling to change the number of threads. When the calculated number of threads is greater than the current number of threads, the flow rate is increased by increasing the flow rate. When the calculated number of threads is less than the current number of threads, the flow rate is controlled by decreasing the flow rate. To do.

JP2007-122527A

  In the Web service system, flow control is performed so as not to accept more access than expected for the stable operation of the system. This flow control generally has the worst processing performance in configuring a system, and is designed based on a server that becomes a bottleneck first.

  However, when a failure occurs in the system, such as a hardware failure, the bottleneck location may change within the system. In such a case, it is difficult to immediately change and reflect the optimum flow rate control value according to the operating state of the system.

Japanese Patent Laid-Open No. 2004-228561 is a method of investigating the resource status in the system, calculating an appropriate flow rate based on the resource information acquired by the system, and calculating the thread.
According to this method, the CPU utilization rate and hold time that are the basis for flow rate calculation use average values, so a failure occurs in the server body or operating system that constitutes the system, and the processing capacity is suddenly significant. It is pointed out that when the decrease occurs, it may take time until the thread number control function operates. Therefore, there is a problem that excessive processing flows in before detection and flow rate control occur, resulting in processing delay and system down.

  In the present invention, an operation monitoring server for holding and managing which part of the system is a bottleneck is provided, and the operating states of all the servers constituting the system are periodically monitored. When the operating status of the system becomes abnormal, for example, when a server is stopped due to a hardware failure, the server where the failure occurred is detected and calculation is performed to determine whether there is a change in the retained bottleneck. The flow control value that is centrally managed by the change system is redesigned to reflect this result.

  When a system failure such as a hardware failure occurs and a change occurs in a location that becomes a bottleneck in the system, the optimal flow control value is set immediately to keep the access flow rate within the allowable range of the system. It can be operated.

Hereinafter, the best mode for carrying out the present invention will be described in detail with reference to the drawings.
FIG. 1 is a system configuration diagram according to the present embodiment. The Web system 20 includes a load balancer 40, a Web server group 50, an application server group 60, a DB server group 70, and an operation monitoring server 80, which are connected to each other via a network LAN 30.

  The load balancer 40 distributes the access from the user terminal 10 to the Web server group 50, and in addition to restricting the access from the user terminal 10 to a range that can be simultaneously processed by the entire Web system 20, the user Limits the number of connections that occur between the user terminal 10 and the load balancer 40 due to access from the terminal 10, and processes access from the user terminal 10 that can be processed simultaneously by the entire Web system 20 Has a function to return an error message without accepting.

  The operation monitoring server 80 periodically checks the operation status of the Web server group 50, the application server group 60, and the DB server group 70, and also includes an information collection table 85, a server number management table 86, and a single processing capacity management table 87. Using the data stored in the overall flow rate management table 88, the number of connections that can be simultaneously processed in the entire Web system 20 is calculated as the flow rate control calculated value 82. In addition, the operation monitoring server 80 transmits the calculated flow rate control calculated value 82 to the load balancer 40 and updates the number of connections that can be simultaneously processed in the entire Web system 20 as the flow rate control value 42.

Next, the information collection table 300 in FIG. 2 will be described.
The information collection table 300 is one of the data tables stored in the operation monitoring server 80 shown in FIG. 1, and the operation monitoring server 80 is provided for the Web server group 50, the application server group 60, and the DB server group 70. Data indicating the results of checking the operating states of the Web server group 50, the application server group 60, and the DB server group 70 by periodically accessing is stored.

  The server 301 of the information collection table 300 stores data preset by an operator, that is, the server names of the Web server group 50, application server group 60, and DB server group 70 of FIG. The group 303 also stores data preset by the person in charge of operation, that is, data grouped in each of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. The operation status 302 stores data as a result of confirming the operation status of all the servers of the Web server group 50, the application server group 60, and the DB server group 70 by the operation monitoring server 80 of FIG. As a result of the operation monitoring server 80 of FIG. 1 confirming the operation state for all the servers of the Web server group 50, the application server group 60, and the DB server group 70, if the server is operating, the data “active” If the server is not operating, the operation monitoring server 80 of FIG.

Next, update of the information collection table (update of the number of operating units) when the Web server No. 3 in FIG. 3 becomes non-operation will be described.
The information collection tables 400 and 500 in FIG. 3 are the information collection tables shown in FIG. 2. From the state of the information collection table 400, the operation monitoring server 80 in FIG. 1 has the Web server group 50, the application server group 60, and the DB server group. As a result of confirming the operating state with respect to all 70 servers, the information collection table 500 is shown to be updated, and the information collection table is updated.

  The information collection table 400 indicates that all the servers of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. 1 are operating. In addition, the information collection table 500 indicates that only the third Web server is not operating among all the servers of the Web server group 50, the application server group 60, and the DB server group 70 in FIG.

Next, update of the information collection table when the application server No. 4 in FIG. 4 becomes non-operation (update of the number of operating units) will be described.
The information collection tables 600 and 700 are the information collection tables shown in FIG. 2, and have the same description as the update of the information collection table of FIG.

  The information collection table 600 indicates that all the servers of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. 1 are operating. The information collection table 700 indicates that only the application server No. 4 is inactive among all the servers of the Web server group 50, the application server group 60, and the DB server group 70 of FIG.

Next, the server number management table 310 in FIG. 5 will be described.
The number-of-servers management table 310 is one of the data tables stored in the operation monitoring server 80 of FIG. 1, and the server number management table 310 is a group of servers in the Web server group 50, the application server group 60, and the DB server group 70. The number of data is stored.

  The group 311 of the server number management table 310 stores data set in advance by an operator, that is, data grouped in each of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. ing. The number 312 also stores data preset by the person in charge of operation, that is, data on the number of servers in each of the Web server group 50, application server group 60, and DB server group 70 in FIG.

Next, the server single processing capacity management table 320 of FIG. 6 will be described.
The single server processing capacity management table 320 is one of the data tables stored in the operation monitoring server 80 of FIG. 1, and the server 1 in each of the Web server group 50, the application server group 60, and the DB server group 70. The data of the number of connections within a range that can be processed simultaneously by one server for the access from the user terminal 10 is stored.

  In the group 321 of the single server processing capacity management table 320, data set in advance by an operator, that is, data grouped in each of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. Stored. In addition, the user terminal is also connected to the data set in advance by the person in charge of operation for the maximum flow rate 322, that is, one server in each of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. Stores data for the number of connections that can be processed simultaneously for accesses from 10.

Next, the overall flow rate management table 330 in FIG. 7 will be described.
The overall flow rate management table 330 is one of the data tables stored in the operation monitoring server 80 of FIG. 1, and the group processing capability in each of the Web server group 50, the application server group 60, and the DB server group 70. That is, the data of the number of connections within the range that can be processed simultaneously by the group for the access from the user terminal 10 is stored, and the group of each of the Web server group 50, the application server group 60, and the DB server group 70 is stored. Among them, data having the smallest value of the number of connections in a range that can be simultaneously processed for access from the user terminal 10 is stored.

  Further, in the group in each of the Web server group 50, the application server group 60, and the DB server group 70 shown in FIG. The data of the number of connections of the information in FIG. 2 in which the operation monitoring server 80 stores the data obtained as a result of confirming the operation status of all the servers of the Web server group 50, the application server group 60, and the DB server group 70. Data of the number of connections within the range that can be processed simultaneously for access from the user terminal 10 by one server in each of the collection table 300 and the Web server group 50, the application server group 60, and the DB server group 70 in FIG. 2 is stored in the server single processing capacity management table of FIG. Each time data of the operating status 302 of the collecting table 300 is updated, connection number of data stored in the entire flow rate management table is also updated.

  In addition, the operation monitoring server 80 in FIG. 1 confirms the operation status of all the servers of the Web server group 50, the application server group 60, and the DB server group 70 for the number of connections data stored in the overall flow rate management table. Two types of data are stored when the operation monitoring server 80 confirms the operation state for all the servers of the Web server group 50, the application server group 60, and the DB server group 70 immediately before and after that.

  The group 331 of the overall flow rate management table 330 includes data set in advance by an operator, that is, data grouped in the Web server group 50, the application server group 60, and the DB server group 70 in FIG. Stores the data with the smallest number of connections within the range that can be processed simultaneously for access from the user terminal 10 among the groups in one Web server group 50, application server group 60, and DB server group 70. Data to store is stored. As the current value 333, the Web server group 50, the application server when the operation monitoring server 80 in FIG. 1 confirms the operating state of all the servers of the Web server group 50, the application server group 60, and the DB server group 70. The operation monitoring server 80 stores data on the number of connections within a range that can be simultaneously processed for access from the user terminal 10 in each of the group 60 and the DB server group 70.

  Further, the immediately preceding value 332 is the value when the operation monitoring server 80 in FIG. 1 confirms the operating state for all the servers of the Web server group 50, the application server group 60, and the DB server group 70 immediately before the current value 333. The operation monitoring server 80 stores data on the number of connections in a range that can be simultaneously processed for access from the user terminal 10 in each of the Web server group 50, the application server group 60, and the DB server group 70. ing.

Next, update of the entire flow rate management table when the Web server No. 3 in FIG. 8 becomes non-operational (flow rate change due to update of the number of operating units—minimum value change) will be described.
The overall flow rate management tables 430 and 530 in FIG. 8 are the overall flow rate management tables shown in FIG. 7. From the state of the overall flow rate management table 430, the operation monitoring server 80 in FIG. This shows that the state of the overall flow rate management table 530 is updated when the operating state is confirmed for all the servers in the DB server group 70, and the update of the overall flow rate management table is shown.

  The overall flow rate management table 430 is the state of the information collection table 400 in FIG. 3, that is, the Web server when all the servers of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. Data on the number of connections in a range that can be simultaneously processed for access from the user terminal 10 in each of the group 50, the application server group 60, and the DB server group 70 is shown. The overall flow rate management table 530 is the state of the information collection table 500 in FIG. 3, that is, only the third Web server among all the servers of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. Indicates the number of connections within the range that can be processed simultaneously for access from the user terminal 10 in each group of the Web server group 50, the application server group 60, and the DB server group 70 when not in operation. ing.

Next, the update of the entire flow rate management table when the application server No. 4 in FIG. 9 becomes inactive (flow rate change due to update of operating units-no change in minimum value) will be described.
The overall flow rate management tables 630 and 730 are the overall flow rate management table shown in FIG. 7, and have the same description as the update of the overall flow rate management table in FIG.

  The overall flow rate management table 630 is the state of the information collection table 600 in FIG. 4, that is, the Web server when all the servers of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. Data on the number of connections in a range that can be simultaneously processed for access from the user terminal 10 in each of the group 50, the application server group 60, and the DB server group 70 is shown. The overall flow rate management table 730 is the state of the information collection table 700 of FIG. 4, that is, only the application server No. 4 among all the servers of the Web server group 50, the application server group 60, and the DB server group 70 of FIG. Indicates the number of connections within the range that can be processed simultaneously for access from the user terminal 10 in each group of the Web server group 50, the application server group 60, and the DB server group 70 when not in operation. ing.

Next, the function of the flow rate control value calculation process in FIG. 10 will be described.
The flow control value calculation process in FIG. 10 is a process executed by the flow control calculation processing unit 81 of the operation monitoring server 80 in FIG. 1, and the operation monitoring server 80 includes the Web server group 50, the application server group 60, and the DB server group. The operating status of 70 is confirmed, and data on the number of connections that can be simultaneously processed by the entire Web system 20 for the access from the user terminal 10 is calculated.

  In step 1000, when the program of the flow rate control value calculation process is started, a preparation process for initializing the entire flow rate management table 330 in FIG. 7 is executed. Data of the number of connections within the range that can be processed simultaneously for access from the user terminal 10 when all the servers of the Web server group 50, the application server group 60, and the DB server group 70 in FIG. 1 are operating. Is calculated. Therefore, information on the number of servers in each group of the Web server group 50, the application server group 60, and the DB server group 70 in FIG.

  The operation monitoring server 80 uses the group values (A to C in this example) as keys to extract the number of servers from the server number management table 310 in FIG. In the example, the maximum flow rate 322, which is the average processing capacity of one server, is extracted from the single server processing capacity management table 320 of FIG. Data of the number of connections that can be simultaneously processed for access from the user terminal 10 in each of the Web server group 50, the application server group 60, and the DB server group 70 is calculated.

  In this example, when the number management table 310 in FIG. 5 and the server single processing capacity management table 320 in FIG. 6 are searched using the A group as a key, data 4 and 200 are extracted, respectively, and the extracted 4 and 200 are multiplied. The data is 800. Therefore, if all four servers in the A group are operating, the maximum number of connections in the range that can be simultaneously processed for access from the user terminal 10 in FIG. Similarly to the A group, when the B group and the C group are processed, data of 900 and 750 are obtained, respectively.

  In step 1010, processing for initializing the entire flow rate management table 330 in FIG. 7 is executed. The operation monitoring server 80 stores the data calculated in step 1000 in both the immediately preceding value 332 and the current value 333 of the overall flow rate management table 330 in FIG. The data having the smallest value among the data extracted from the immediately preceding value 332 and the current value 333 using the A group, the B group, and the C group as keys is stored as the minimum value of the group 331.

  In this example, when the data calculated in step 1000 is searched using the A group as a key, data 800 is extracted. Next, the extracted 800 data is used as the key for the A group as a whole flow management table 330 in FIG. Are stored in the immediately preceding value 332 and the current value 333, respectively. Similar to the A group, when the B group and the C group are processed, data 900 and 750 are stored in the previous value 332 and the current value 333, respectively.

  In addition, when step 1000 is searched using A group, B group, and C group as keys, data of 800, 900, and 750 are extracted, and the data of the smallest value among the extracted data of 800, 900, and 750 is searched. Then 750 data is extracted. The extracted data 750 is stored in the previous value 332 and the current value 333 of the overall flow rate management table 330 in FIG. 7 using the minimum value as a key.

  In step 1020, the operation monitoring server 80 in FIG. 1 executes processing for initializing the operation state 302 of the information collection table 300 in FIG. The operation monitoring server 80 stores data in an operating state with respect to all servers of the server 301, that is, data that is all active.

  In this example, data “active” is stored in the operation state 302 of the information collection table 300 of FIG.

  In step 1030, the operation monitoring server 80 of FIG. 1 accesses all the servers of the Web server group 50, the application server group 60, and the DB server group 70, and the Web server group 50, the application server group 60, and the DB server are accessed. Check the operating status for all servers in group 70. The operation monitoring server 80 shown in FIG. 1 confirms the operating state of all the servers of the Web server group 50, the application server group 60, and the DB server group 70. If the server is in a non-operating state, the data is inactive.

  In this example, as indicated by the information collection table 500 in FIG. 3, data is inactive when the Web server No. 3 is inactive, and data is active when all servers other than the Web server No. 3 are in operation. It becomes. Further, as shown in the information collection table 700 of FIG. 4, the data is inactive when the application server No. 4 is inactive, and the data is active when all servers other than the application server No. 4 are in operation. .

  In step 1040, in order to detect whether the number of operating servers in the Web server group 50, the application server group 60, and the DB server group 70 in FIG. The server operating status and the operating status 302 of the information collection table 300 in FIG. 2 are searched, and the server operating status is extracted and compared using all servers as keys, and whether the server operating status is the same. Or determine whether they are different.

  In this example, as shown in the information collection table 500 of FIG. 3, when the server operation state confirmed in step 1030 and the operation state 402 of the information collection table 400 of FIG. , Active data is extracted and determined to be different data. Similarly to Web server No. 3, data other than Web server No. 3 is extracted as active and active data and determined to be the same data.

  As shown in the information collection table 700 of FIG. 4, when the server operating state confirmed in step 1030 and the operation state 602 of the information collecting table 600 of FIG. Is extracted and determined as different data. Similarly to the application server No. 4, the servers other than the application server No. 4 extract the data active and active, respectively, and determine that the data is the same.

  In step 1100, the server operation status checked in step 1030 and the data extracted from the operation status 302 of the information collection table 300 in FIG. Compare and check if there is data that the server operating status is different, such as inactive and active, or active and inactive, or active and active, or inactive and inactive. Determine whether the data indicates that the server operating status is the same.

  That is, it is possible to detect whether the number of servers in operation in the Web server group 50, the application server group 60, and the DB server group 70 in FIG. 1 has increased or decreased. If the result data processed in step 1040 is the same data in all servers, the processing returns to step 1030. Also, if the data processed as a result of step 1040 includes data that is different in one server among all servers, the process proceeds to step 1200.

  In this example, as shown in the update of the information collection table in FIG. 3, the data processed in step 1040 is different from inactive and active in the Web server No. 3, and the process proceeds to step 1200. . The same applies to FIG. 4, and the process proceeds to step 1200.

  In step 1200, since it is determined in step 1100 that the number of active servers in the Web server group 50, application server group 60, and DB server group 70 in FIG. 1 has increased or decreased, step 1030 is performed using all servers as keys. In the operating status of the servers checked in step 1, the data “active” is extracted if it is a server that is in operation, and the data “inactive” is extracted if it is a server that is not operating, and all servers are keyed. Is stored in the operating state 302 of the information collection table 300 of FIG.

  In this example, as shown in the update of the information collection table 500 in FIG. 3, the server operating state checked in step 1030 is searched using the Web server No. 3 as a key, and data called inactive is extracted. Is stored in the operating state 502 of the information collection table 500 of FIG. Similarly to the Web server No. 3, the server operating state checked in step 1030 is searched by using a key other than the Web server No. 3 as a key, and the data “active” is extracted, and the information of FIG. Stored in the operation state 502 of the collection table 500.

  Further, as shown in the information collection table 700 of FIG. 4, the server operating state checked in step 1030 is searched using the application server No. 4 as a key to extract data called inactive, and the application server No. 4 is used as a key. The information is stored in the operating state 702 of the information collection table 700 of FIG. Similarly to application server No. 4, data other than application server No. 4 is used as a key to extract active data from the server operating state determined in step 1030, and the information collection of FIG. 4 is performed using each server name as a key. Stored in the operating state 702 of the table 700.

  In step 1210, since it was determined in step 1100 that the number of operating servers in the Web server group 50, application server group 60, and DB server group 70 in FIG. On the other hand, there is a possibility that data of the number of connections that can be processed simultaneously in each group of the Web server group 50, the application server group 60, and the DB server group 70 may increase or decrease. In order to calculate the data of the number of connections that can be processed simultaneously in each group of the Web server group 50, the application server group 60, and the DB server group 70 for access from the user terminal 10 of FIG. 50, information on the number of servers that are operating in each of the application server group 60 and the DB server group 70, and information on the single processing capacity of the servers is required.

  The number of data servers in which the operation state 302 of the information collection table 300 in FIG. 2 is in operation, that is, servers whose operation state is active, using each group value (A to C in this example) as a key. The maximum flow rate 322 that is the average processing capacity of one server from the server single processing capacity management table 320 of FIG. 6 using the data obtained by extracting the number of units and the group values (A to C in this example) as keys. The data obtained by multiplying the extracted data is processed and processed simultaneously in each group of the Web server group 50, the application server group 60, and the DB server group 70 for the access from the user terminal 10 of FIG. Recalculate the data for the number of possible connections.

  In this example, as shown in the update of the information collection table of FIG. 3, the number of actives is extracted from the operating state 502 of the information collection table 500 of FIG. The maximum flow rate 322 of the server single processing capacity management table 320 in FIG. 6 is searched using the A group as a key, and data of 200 is extracted. When this extracted data is multiplied by 3 and 200, data of 600 is calculated.

  Similarly to the A group, when the B group and the C group are processed, data of 900 and 750 are calculated, respectively. Further, as shown in the update of the information collection table of FIG. 4, the number of active is retrieved from the operating state 702 of the information collection table 700 of FIG. Using the group as a key, the maximum flow rate 322 of the single server processing capacity management table 320 in FIG. 6 is searched to extract data 150. When this extracted data is multiplied by 5 and 150, data of 750 is calculated. Similarly to the B group, when the A group and the C group are processed, data of 800 and 750 are calculated, respectively.

  In step 1220, the Web server group 50, the application server group 60, and the DB server group 70 are simultaneously controlled for the access from the user terminal 10 of FIG. In order to update the data of the number of connections that can be processed to the latest data, the data calculated in step 1210 is extracted using each group value (A to C in this example) as a key, and each group value (main In the example, the current value 333 of the overall flow rate management table 330 is stored using A to C) as keys.

  In this example, as shown in the update of the overall flow rate management table 530 in FIG. 8, when the step 1210 is searched using the A group as a key, data 600 is extracted, and then the extracted data 600 is used as the A group. Is stored in the current value 533 of the overall flow rate management table 530 in FIG. Similar to the A group, when the B group and the C group are processed, data 900 and 750 are stored in the current value 533 of the overall flow rate management table 530 in FIG. 8 using the B group and the C group as keys, respectively.

  Further, as shown in the update of the overall flow rate management table 530 in FIG. 9, when the step 1210 is searched using the B group as a key, data 750 is extracted, and then the extracted data 750 is used as the key for the B group. Is stored in the current value 733 of the overall flow rate management table 730 in FIG. When the A group and the C group are processed in the same manner as the B group, data 800 and 750 are stored in the current value 733 of the overall flow rate management table 730 in FIG. 9 using the A group and the C group as keys, respectively.

  In step 1230, for the access from the user terminal 10 of FIG. 1 managed by the overall flow rate management table 330 of FIG. 7, the data of the number of connections that can be simultaneously processed by the entire Web system 20 is updated to the latest data. Therefore, the current value 333 of the overall flow rate management table 330 updated in step 1220 is searched using each value of the group (A to C in this example) as a key, and the smallest value data is extracted. Store as the minimum value of the current value 333 as a key.

  In this example, as shown in the update of the overall flow rate management table 530 in FIG. 8, when the current value 533 of the overall flow rate management table 530 in FIG. Data of 750 is extracted, and when the smallest value is searched among the extracted data of 600, 900 and 750, data of 600 is extracted. The extracted data 600 is stored in the current value 533 of the overall flow rate management table 530 in FIG. 8 using the minimum value as a key.

  Further, as shown in the update of the overall flow rate management table 730 in FIG. 9, when the current value 733 of the overall flow rate management table 730 in FIG. 9 is searched using the A group, the B group, and the C group as keys, they are 800, 750, and 750. Data is extracted, and when the smallest value is searched among the extracted data of 800, 750, and 750, data of 750 is extracted. The extracted data 750 is stored in the current value 733 of the overall flow rate management table 730 in FIG. 9 using the minimum value as a key.

  In step 1240, since the number of operating servers in the Web server group 50, application server group 60, and DB server group 70 in FIG. 1 has increased or decreased in step 1100, access from the user terminal 10 in FIG. On the other hand, there is a possibility that data of the number of connections that can be simultaneously processed in the entire Web system 20 may increase or decrease. Therefore, using the minimum value as a key, the previous value 332 and the current value 333 of the overall flow management table 330 in FIG. Compare whether the value data is the same data or different data.

  In this example, as shown in the update of the overall flow rate management table 530 of FIG. 8, the previous value 532 and the current value 533 of the overall flow rate management table 530 of FIG. When the data is extracted, it is determined that the minimum value data is different data. Further, as shown in the update of the overall flow rate management table 730 in FIG. 9, the previous value 732 and the current value 733 of the overall flow rate management table 730 in FIG. When extracted, it is determined that the minimum value data is the same data.

  In step 1300, it is determined whether or not the minimum value data of the immediately preceding value 332 and the current value 333 of the overall flow rate management table 330 compared in step 1240 are different. That is, in step 1100, the entire Web system 20 is controlled against access from the user terminal 10 due to the increase / decrease in the number of operating servers in the Web server group 50, the application server group 60, and the DB server group 70 in FIG. It can be determined whether or not the data of the number of connections that can be processed simultaneously has increased or decreased. If the minimum value data of the previous value 332 and the current value 333 of the overall flow rate management table 330 compared in step 1240 are the same data, the process proceeds to step 1500. If the minimum value data of the immediately preceding value 332 and the current value 333 of the overall flow management table 330 compared in step 1240 are different data, the process proceeds to step 1400.

  In this example, as shown in the update of the overall flow rate management table 530 in FIG. 8, when searching the data as a result of comparison in step 1240 using the minimum value as a key, data 750 and 600 are extracted, and the minimum value data is Since it is determined that the data is different, the process proceeds to step 1400. Also, as shown in the update of the overall flow rate management table 730 in FIG. 9, when searching for data as a result of comparison in step 1240 using the minimum value as a key, data 750 and 750 are extracted, and the minimum value data is the same data Therefore, the process proceeds to step 1500.

  In step 1400, the number of connections that can be processed simultaneously in the entire web system 20 for the access from the user terminal 10 in FIG. Since the flow control value 42 in which the data of the number of connections is set for the load balancer 40 that controls the number of data needs to be updated, the minimum value is used as a key in the overall flow management table 330 in FIG. The current value 333 is searched and the minimum value data is extracted and stored in the flow rate control calculated value 82 managed by the operation monitoring server 80.

  In this example, as shown in the update of the overall flow management table 530 of FIG. 8, the current value 533 of the overall flow management table 530 of FIG. 8 is searched using the minimum value as a key, and data of 600 is extracted and extracted. Data 600 is stored in the flow control calculated value 82 managed by the operation monitoring server 80 in FIG.

In step 1410, the flow rate control calculated value 82 of FIG. 1 stored in step 1400 is transmitted to the load balancer 40.
In this example, as shown in the update of the overall flow rate management table 530 in FIG. 8, the flow rate control calculated value 82 storing the data of 600 in step 1400 is transmitted to the load balancer 40 in FIG.

  In Step 1500, each group value (A to C in this example) is used as a key to search and extract the current value 333 of the overall flow rate management table 330 in FIG. A to C) are stored in the previous value 332 as keys. That is, the data of the immediately preceding value 332 and the current value 333 in each group 331 are the same data. After this process ends, the process proceeds to step 1030.

  In this example, as shown in the update of the overall flow rate management table 530 in FIG. 8, when the current value 533 of the overall flow rate management table 530 is searched using the A group as a key, data “600” is extracted. Is stored in the immediately preceding value 532 of the overall flow rate management table 530 using the A group as a key. Similar to the A group, when the B group, the C group, and the minimum value are also processed, data 900, 750, and 600 are stored in the previous value 532 of the overall flow rate management table 530, respectively. After this process is completed, the process proceeds to step 1030. Further, as shown in the update of the overall flow rate management table 730 in FIG. 9, when the current value 733 of the overall flow rate management table 730 is searched using the B group as a key, data 750 is extracted. Stored in the immediately preceding value 732 of the overall flow rate management table 730 using the B group as a key. Similar to the B group, when the A group, the C group, and the minimum value are processed, data of 800, 750, and 750 are stored in the immediately preceding value 732 of the overall flow rate management table 730, respectively. After the completion of this process, the process proceeds to step 1030.

Next, the function of the flow rate control value update process in FIG. 11 will be described.
In step 2000, the flow rate control value update process in FIG. 11 is a process executed by the flow rate control value update processing unit 41 of the load distribution apparatus 40 in FIG. 1, and the operation in FIG. 1 processed in step 1410 in FIG. The flow rate received from the operation monitoring server 80 by receiving the flow control calculated value 82 storing the data of the number of connections that can be processed simultaneously by the entire Web system 20 for the access from the user terminal 10 transmitted from the monitoring server 80 The data of the control calculation value 82 is stored in the flow rate control value 42 managed by the load balancer 40.

  In this example, the flow control calculated value 82 stored in the data of 600 transmitted from the operation monitoring server 80 of FIG. 1 processed in step 1410 of FIG. 10 is received, and the received data of 600 is converted to the flow control value. Store in 42. Therefore, the number of connections that can be simultaneously processed in the entire Web system 20 for access from the user terminal 10 in FIG. 1 is updated from 750 to 600.

It is a system configuration figure of a flow control system in this embodiment. It is a data block diagram of the information collection table in this embodiment. It is a figure which shows the update state of an information collection table when the Web server 3 machine in this embodiment becomes non-operation. It is a figure which shows the update state of an information collection table when the application server 4 machine in this embodiment becomes non-operation. It is a data block diagram of the server number management table in this embodiment. It is a data block diagram of the server single-unit capacity table in this embodiment. It is a data block diagram of the whole flow volume management table in this embodiment. It is a figure which shows the update state of the whole flow volume management table when the Web server 3 machine in this embodiment becomes non-operation. It is a figure which shows the update state of the whole flow volume management table when the application server No. 4 machine in this embodiment becomes non-operation. It is a flowchart which shows the process sequence of the flow control value calculation process in this embodiment. It is a flowchart which shows the process sequence of the flow control value update process in this embodiment.

10 User terminal 11 Network 20 Web system 30 Network LAN
40 Load Distribution Device 41 Flow Control Value Update Processing Unit 42 Flow Control Value 50 Web Server Group 60 Application Server Group 70 DB Server Group 80 Operation Monitoring Server 81 Flow Control Calculation Processing Unit 82 Flow Control Calculation Value 85 Information Collection Table 86 Server Number Management Table 87 Single processing capacity management table 88 Overall flow management table

Claims (1)

A first server group composed of a plurality of first servers, a second server group composed of a plurality of second servers, a third server group composed of a plurality of third servers, and settings In an operation monitoring server connected via a network to a load balancer that limits the number of accesses to the first to third server groups based on values,
Storage means for storing the number of first to third servers constituting each of the first to third server groups and each processing capacity;
Based on the number of the first to third servers stored in the storage means and the processing capacity of each of the first to third servers, the total flow calculation for calculating the total flow indicating the processing capacity for each of the first to third server groups. Means,
Flow rate control calculation value setting for setting the lowest value among the total flow rates for each of the first to third server groups calculated by the flow rate control value as a flow rate control calculation value that is a setting value of the load distribution device Means,
Operating status monitoring means for monitoring the operating status of the servers constituting the first to third server groups;
And a means for updating the number of the first to third servers stored in the storage means when the operating status monitoring means detects that the server is not in operation. Operation monitoring server.

Images