JP2011113139A - Computer system, information processing method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a computer system that efficiently obtains a processing load. <P>SOLUTION: The computer system includes: a queue for storing a message in association with a storage time when the message is stored; a table for storing the number of messages stored in the queue, a total retention time that is a sum of message retention time when a message is extracted or stored the last time, and a calculation time when the first total retention time is calculated; and a calculation means for calculating the total retention time of messages stored in the queue based on a storage time for the extracted message, the number of messages stored in the table, the total retention time, and the calculation time when a message to be subjected to specific processing is extracted from a queue. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present technology relates to a technology for improving system throughput.

  For example, there is a system that performs a plurality of types of business processing on a message received via a network such as a LAN (Local Area Network) or the Internet, and transmits the processing result to, for example, a client terminal. In such a system, a message indicating completion of business processing of a specific business processing process is temporarily stored in a queue. For example, in the system shown in FIG. 1, after the business process A executes the business process A on the message, the processed message is stored in a queue. On the other hand, the business process B extracts a message stored in the queue and executes the business process B.

  In the system described above, for example, when the load is concentrated on the business process B, or when the business process B is slower than the business process A, the message stays in the queue. Sometimes. Then, since the waiting time in the queue becomes long, the time from when the system receives a message to when it is transmitted becomes long (that is, the processing efficiency of the system decreases). Therefore, it is required to monitor the status of message retention in the queue and take measures to prevent the system processing efficiency from dropping. In the example described above, for example, it is conceivable to reduce the number of business processing processes A or increase the number of business processing processes B.

  Conventionally, there is a technique for monitoring a message retention state based on the number of messages stored in a queue. That is, the process number control means of the monitoring process determines whether the number of processing target I / O requests is increasing or decreasing from the number of processing target I / O requests registered every predetermined time. It is determined whether the number of processing target I / O requests is different from the current activation I / O processing process number by a threshold value α or more. Based on the determined result, the number of startup I / O processing processes is increased, decreased, or the current state is maintained.

  However, even if the number of staying messages is small, if the staying time is long, there is a possibility that the number of processes is insufficient and the business process is delayed. Is short, it can be considered that there is no problem in the number of processes even in the current state. Therefore, it is not always appropriate to monitor the staying status using the technique as described above.

  As a technique for dealing with such a problem, a method for obtaining a total residence time, which is the total residence time of messages in a queue, is known. This method will be described with reference to FIG. In the system of FIG. 1, after the business process A executes the business process A on the message, the processed message is stored in a queue, and the business process B extracts the message and executes the business process B. In this technique, the current time is stored in a queue together with a message, and the total residence time is calculated using the current time (hereinafter referred to as storage time).

  Specifically, the total residence time is calculated by the following method. That is, the monitoring process requests (1) suspension of execution of the business process A and the business process B, (2) searches the queue, acquires the storage time for all messages, and (3) the current time and The total residence time is calculated by calculating the difference from the storage time for all messages and calculating the sum, and (4) requesting the execution of the business process A and the business process B to be resumed. In the example of FIG. 1, when the current time is 09:00, the total residence time is (09: 0-08: 30) + (09: 0-08: 20) + (09: 0-08: 10) + (09: 00-08: 00) = 180 (minutes).

  However, the method shown in FIG. 1 requires processing for searching the queue and acquiring the storage time for all messages, so the business processing must be temporarily stopped, and the efficiency of the business processing is reduced. To do.

  In addition, when monitoring by a monitoring process, it is necessary to set a monitoring cycle and periodically monitor the queue, but it is necessary to monitor frequently during times when the number of messages increases or decreases greatly On the other hand, frequent monitoring is not necessary during other times. Therefore, there is a problem that it is difficult to determine an optimal monitoring cycle. Furthermore, since inter-process communication occurs between the monitoring process and the business process, there is a problem that the solution process cannot be performed immediately.

JP 2004-348233 A

  Therefore, an object of the present technology is to provide a technology for efficiently grasping the processing load.

  The computer system according to the first aspect of the present technology includes a first queue that stores a message in association with a storage time at which the message is stored, and the number of messages stored in the first queue. A first calculation time which is a time when a first total dwell time and a first total dwell time are calculated, which is a total of message dwell times at the time when a certain first number of messages and a previous message are extracted or stored. Are stored in the first status table and the storage time for the extracted message and the first status table when a message to be processed is extracted from the first queue. Based on the first number of cases, the first total residence time, and the first calculation time, the total residence time of the messages stored in the first queue is calculated, and the first In the status table, the first number is decreased by 1, the first total residence time is updated with the total residence time of the messages stored in the first queue, and the first calculation time is updated with the current time. Calculating means.

  It becomes possible to grasp the processing load efficiently.

It is a figure for demonstrating a prior art. 1 is a configuration diagram of a system according to an embodiment of the present technology. It is a figure which shows an example of the data stored in a queue. It is a figure which shows an example of the data stored in a common status table. It is a figure which shows an example of the case where a message stays. It is a figure which shows the processing flow of the process number adjustment process 1. FIG. (A) And (b) is a figure which shows an example of the data stored in a queue and a common status table. It is a figure which shows the processing flow of the process number adjustment process 1. FIG. It is a figure which shows an example of the data stored in a common status table. It is a figure which shows the processing flow of the process number adjustment process. It is a figure for demonstrating the calculation method of total residence time. It is a figure for demonstrating the calculation method of total residence time. It is a figure for demonstrating the calculation method of total residence time. It is a figure for demonstrating the calculation method of total residence time. It is a figure for demonstrating the calculation method of total residence time. It is a figure for demonstrating the calculation method of total residence time. It is a figure for demonstrating the method to adjust the total residence time of a queue to an appropriate state. It is a figure for demonstrating the method to adjust the total residence time of a queue to an appropriate state. It is a functional block diagram of a computer.

  FIG. 2 shows a functional block diagram of the business system 1 according to the present embodiment. The business system 1 according to the present embodiment includes a reception unit 3, a queue 7A, a common status table 9A, one or more business processing processes 11A, a queue 7B, a common status table 9B, and one or more It has a business process 11B, a queue 7C, a common status table 9C, one or a plurality of business processes 11C, and a transmission unit 5.

  The business process 11A is a process that is executed to perform the business process A on the message. The business process unit 101A, the total residence time calculation unit 103A, and the process number adjustment processing unit 105A include Have. Similarly, the business process 11B is a process executed to execute the business process B on the message, and includes a business process unit 101B, a total residence time calculation unit 103B, and a process number adjustment processing unit 105B. Have Further, the business process 11C is a process that is executed to execute the business process C on the message, and includes a business process unit 101C, a total residence time calculation unit 103C, and a process number adjustment processing unit 105C. Have.

  The receiving unit 3 receives a message via a network such as a LAN (Local Area Network) or the Internet and stores it in the queue 7A. The business processing unit 101A of the business processing process 11A extracts a message from the queue 7A according to a FIFO (First In First Out) method, performs the business processing A, and then stores the processed message in the queue 7B. The total dwell time calculation unit 103A calculates the total dwell time of the messages stored in the queue 7A based on the storage time when the messages are stored in the queue 7A and the data stored in the common status table 9A. The total residence time of the messages stored in the queue 7B is calculated based on the data stored in the status table 9B. Based on the total residence time for the queue 7A and the total residence time for the queue 7B, the process number adjustment processing unit 105A stops the business processing process 11A including the process number adjustment processing unit 105A or newly starts the business processing process 11A. to start.

  In addition, the business processing unit 101B of the business processing process 11B extracts a message from the queue 7B according to the FIFO method, performs the business processing B, and then stores the processed message in the queue 7C. The total residence time calculation unit 103B calculates the total residence time of the messages stored in the queue 7B based on the storage time when the message is stored in the queue 7B and the data stored in the common status table 9B. The total residence time of messages stored in the queue 7C is calculated based on the data stored in the status table 9C. Based on the total residence time for the queue 7B and the total residence time for the queue 7C, the process number adjustment processing unit 105B stops the business processing process 11B including the process number adjustment processing unit 105B or newly starts the business processing process 11B. to start.

  The business processing unit 101C of the business processing process 11C extracts a message from the queue 7C according to the FIFO method, performs the business processing C, and then outputs the processed message to the transmission unit 5. The total residence time calculation unit 103C calculates the total residence time of the message stored in the queue 7C based on the storage time when the message is stored in the queue 7C and the data stored in the common status table 9C. Based on the total residence time of the queue 7C, the process number adjustment processing unit 105C stops the business processing process 11C including the process number adjustment processing unit 105C or newly starts the business processing process 11C. Then, the transmission unit 5 transmits a message in which all necessary processes have been executed to a terminal or the like that should receive the message via the network.

  FIG. 3 shows an example of data stored in the queue 7A. In the example of FIG. 3, a message and a storage time when the message is stored are stored in the queue 7A. In the example of FIG. 3, since the message whose storage time is 11:00 is the message stored first, the message is extracted from the queue 7A first. On the other hand, since the message whose storage time is 12:00 is the message stored last, the message is finally extracted from the queue 7A. Similar data is stored in the queue 7B and the queue 7C.

  FIG. 4 shows an example of data stored in the common status table 9A for the queue 7A. In the example of FIG. 4, the common status table 9 </ b> A includes the previous total residence time that is the sum of the residence times of messages at the time when the previous message was extracted or stored, and the previous time that is the time at which the previous total residence time was calculated. The number of messages stored in the queue 7A is stored. The previous total residence time is calculated by a process described later. The common status table 9B and the common status table 9C store data in the same format, but the common status table 9B stores data about the queue 7B, and the common status table 9C stores data about the queue 7C. Is stored.

  Here, an example of a case where a message stored in the queue stays will be described with reference to FIG. FIG. 5 shows that in the system according to the present embodiment shown in FIG. 2, the number of business processing processes 11A is 2, the number of business processing processes 11B is 3, and the number of business processing processes 11C is 3. 3 is shown. Here, it is assumed that the business process A of the business process 11A and the business process C of the business process 11C do not take time, but the business process B takes time.

  In such a case, since the business process A does not take time, the message is stored in the queue 7B at a relatively high frequency. On the other hand, since the business process B takes time, the frequency of extracting messages from the queue 7B is not high. Therefore, messages tend to stay in the queue 7B. Further, since the business process B takes time, the frequency of storing messages in the business process C queue is not high. On the other hand, since the business process C does not take time, the business process 11C tends to be in a waiting state. This is not preferable because not only the waiting time in the queue 7B becomes long and the throughput of the system decreases, but also the resources of the business process 11C cannot be effectively used.

  Therefore, in the present embodiment, by performing the processing described below, it is possible to monitor the message retention status and take measures to prevent a reduction in system throughput. The processing contents of the system shown in FIG. 2 will be described with reference to FIGS. First, a description will be given of a business process process (process number adjustment process 1) that performs both a process of extracting a message from a queue and a process of storing a message in the queue as in the business process 11A and 11B. Here, the business process 11B will be described as an example. It is assumed that data has already been stored in the queue 7A, the queue 7B, and the queue 7C.

  First, the business processing unit 101B of the business processing process 11B extracts one message and the storage time of the message from the extraction queue (that is, the queue 7B) according to the FIFO method, and stores it in a storage device such as a main memory. (FIG. 6: Step S1). For example, when the data shown in FIG. 7A is stored in the queue 7B, the message in the uppermost row and the storage time 11:10 are extracted.

  Then, the total residence time calculation unit 103B acquires the current time and stores it in a storage device such as a main memory (step S3). The current time is acquired from, for example, a system clock (not shown) of the business system 1. Here, it is assumed that the current time is 12:30. The total residence time calculation unit 103B reads the previous time, the previous total residence time, and the number of messages from the common status table 9B, and stores them in a storage device such as a main memory (step S5). For example, when the data shown in FIG. 7B is stored in the common status table 9B, the previous time 12:10, the previous total residence time 200 minutes, and the number of messages “7” are read.

The total residence time calculating unit 103B is the total residence time T ₁ of the the queue 7B, T 1 ₌ the previous total residence time + (current time - last time) × number of messages - (current time - storing time) was calculated as And stored in a storage device such as a main memory (step S7). Here, since the total residence time T ₁ = 200 minutes + (12: 30-12: 10) × 7− (12: 30-11: 10), T ₁ is 260 minutes. In addition, when calculating by the method demonstrated using FIG. 1, (12: 30-11: 30) + (12: 30-11: 35) + (12: 30-11: 40) + (12: 30-11: 45) + (12: 30-12: 00) + (12: 30-12: 10) = 260 minutes. In any method, the total residence time of the messages stored in the queue 7B is 260 minutes.

  The total residence time calculation unit 103B then updates the previous time with the current time in the common status table 9B, updates the previous total residence time with the total residence time calculated in step S7, and reduces the number of messages by 1 (step S9). ). Here, in the common status table 9B, 12:30 is stored as the previous time, 260 minutes is stored as the previous total residence time, and “6” is stored as the number of messages.

  Then, the business processing unit 101B performs business processing B on the message extracted in step S1 (step S11). After step S11 is completed, the task processing unit 101B acquires the current time and adds it to the message (step S13). In step S13, the current time is stored in a storage device such as a main memory. Here, it is assumed that the current time is 12:50. The process proceeds to step S15 in FIG.

  Shifting to the description of FIG. 8, the total residence time calculation unit 103B reads the previous time, the previous total residence time, and the number of messages from the common status table 9C, and stores them in a storage device such as a main memory (step S15). For example, when the data shown in FIG. 9 is stored in the common status table 9C, the previous time 12:30, the previous total residence time 200 minutes, and the number of messages “6” are read.

The total residence time calculating unit 103B, the total residence time T ₂ of the the queue 7C, T 2 ₌ previous total residence time + - calculated as (current time last time) × number of messages, a storage device such as a main memory Store (step S17). Here, since the total residence time T ₂ = 200 minutes + (12: 50−12: 30) × 6, T ₂ is 320 minutes. In addition, when calculating by the conventional method demonstrated using FIG. 1, (12: 50-11: 30) + (12: 50-11: 35) + (12: 50-11: 50) + ( 12: 50-11: 55) + (12: 50-12: 20) + (12: 50-12: 30) = 320 minutes. In any method, the total residence time of messages stored in the queue 7C is 320 minutes.

  Then, the business processing unit 101B stores the message to which the current time is added in step S13 in the queue 7C according to the FIFO method (step S19). The storage time in the queue 7C is the current time stored in the queue 7C in step S19. Further, the total residence time calculation unit 103B updates the previous time with the current time in the common status table 9C, updates the previous total residence time with the total residence time calculated in step S17, and increases the number of messages by 1 (step S21). ). Here, in the common status table 9C, 12:50 is stored as the previous time, 320 minutes is stored as the previous total dwell time, and “7” is stored as the number of messages.

Then, the process number adjustment processing unit 105B determines whether T ₁ ≧ TH ₁ and T ₂ <TH ₂ (step S23). TH ₁ and TH ₂ are preset threshold values. For example TH ₁ may be set to relatively large value to determines that message to a queue 7B is retained sufficiently in the case of satisfying the T ₁ ≧ TH _1. For example, a small value is set for TH ₂ to some extent, and when T ₂ <TH ₂ is satisfied, it is determined that the message is hardly retained in the queue 7C.

When T ₁ ≧ TH ₁ and T ₂ <TH ₂ (step S23: Yes route), the process number adjustment processing unit 105B starts one new business process 11B (step S25). As described with reference to FIG. 5, when T ₁ ≧ TH ₁ and T ₂ <TH ₂ , it is considered that the waiting time in the queue 7B is long, so the waiting time is reduced and the throughput of the system is reduced. Improve.

On the other hand, when T ₁ ≧ TH ₁ and T ₂ <TH ₂ are not satisfied (step S23: No route), the process number adjustment processing unit 105B determines whether T ₁ <TH ₃ and T ₂ ≧ TH ₄ (step S27). ). TH ₃ and TH ₄ are preset threshold values. For example, a small value is set to TH ₃ to some extent, and when T ₁ <TH ₃ is satisfied, it is determined that the message is hardly retained in the queue 7B. Further, for example, TH ₄ may be set to relatively large value to determines that message to a queue 7C is retained sufficiently in the case satisfying T ₂ ≧ TH _4.

When T ₁ <TH ₃ and T ₂ ≧ TH ₄ (step S27: Yes route), the process number adjustment processing unit 105B stops the business process 11B including the process number adjustment processing unit 105B (step S29). . That is, it stops itself. As described with reference to FIG. 5, when T ₁ <TH ₃ and T ₂ ≧ TH ₄ , there is a possibility that the resources of the business process 11A may not be effectively used. It reduces the resources allocated.

On the other hand, if T ₁ <TH ₃ and T ₂ ≧ TH ₄ are not satisfied (step S27: No route), the process is terminated. This is because it is considered that there is no problem with the number of the business process 11B as it is.

  As described above, the total residence time is calculated without obtaining the storage time for all messages stored in the queue, so there is no need to stop the business process when calculating the total residence time. Processing efficiency is improved. In addition, since the number of business processing processes can be adjusted autonomously, there is no need to prepare a separate monitoring process and monitor the queue. In the case where a separate monitoring process is provided, inter-process communication occurs between the monitoring process etc. and the processing unit, so there is a problem that the number of processes cannot be controlled immediately. Such processing does not cause such a problem.

  Next, with reference to FIG. 10, a description will be given of a business process process (process number adjustment process 2) that does not perform a process of storing a message in a queue. Here, the business process 11C will be described as an example. Note that the contents of the processes in steps S41 to S51 are the same as those in steps S1 to S11, and thus description using a specific example is omitted.

  First, the business processing unit 101C of the business processing process 11C extracts one message and the storage time of the message from the queue 7C according to the FIFO method, and stores them in a storage device such as a main memory (FIG. 10: Step S41). . Then, the total residence time calculation unit 103C acquires the current time and stores it in a storage device such as a main memory (step S43). The current time is acquired from, for example, a system clock (not shown) of the business system 1. The total residence time calculation unit 103C reads the previous time, the previous total residence time, and the number of messages from the common status table 9C, and stores them in a storage device such as a main memory (step S45).

The total residence time calculating unit 103C, the total residence time T ₃ in the queue 7C, T 3 ₌ previous total residence time + (current time - last time) × number of messages - (current time - storing time) was calculated as Then, it is stored in a storage device such as a main memory (step S47). Further, the total residence time calculation unit 103C updates the previous time with the current time in the common status table 9C, updates the previous total residence time with the total residence time calculated in step S47, and reduces the number of messages by 1 (step S49). ).

Then, the business processing unit 101C performs business processing C on the message extracted in step S41 (step S51). Further, the process number adjustment processing unit 105C determines whether TH ₅ ≦ T ₃ is satisfied (step S53). TH ₅ is a preset threshold value. For example TH to ₅ have set up relatively large value, it is determined that the message queue 7C is retained sufficiently in the case of satisfying the TH ₅ ≦ T _3.

When TH ₅ ≦ T ₃ (step S53: Yes route), the process number adjustment processing unit 105C newly starts one business process 11C (step S55). This is because the waiting time in the queue 7C is considered to be long, so that the waiting time is reduced and the throughput of the system is improved.

On the other hand, when TH ₅ ≦ T ₃ is not satisfied (step S53: No route), the process number adjustment processing unit 105C determines whether T ₃ <TH ₆ is satisfied (step S57). TH ₆ is a preset threshold value. For example may be set a lower value enough in the TH _6, when satisfying T 3 _{<TH 6} is in the queue 7C determines that the message is hardly retained.

When T ₃ <TH ₆ (step S57: Yes route), the process number adjustment processing unit 105C stops the business process 11C including the process number adjustment processing unit 105C (step S59). That is, it stops itself. Since it is considered that the resources of the business process 11C cannot be effectively used, the resources allocated to the business process 11C are reduced.

On the other hand, if T ₃ <TH ₆ is not satisfied (step S57: No route), the process is terminated. This is because it is considered that there is no problem with the number of the business process 11C as it is.

  By performing the processing as described above, the number of processes can be appropriately adjusted even for a business process that does not store a message after the business process in the queue, such as the business process 11C. The point that it is not necessary to stop the business process in calculating the total residence time and the point that the business process can autonomously adjust the number of processes are the same as the process number adjustment process 1.

  Here, the calculation method of the total residence time according to the present embodiment will be described using the specific examples shown in FIGS. In the example shown in FIGS. 11 to 16, the business process A performs the business process A on the message and then stores it in the queue. Then, the business process B extracts the message from the queue and executes the business process B.

  First, at the time of FIG. 11 (current time is 9:00), the business process A stores the message 1 in the queue. At this time, the total residence time is calculated as 0 minutes + (9: 00-8: 50) × 0 = 0 minutes. In the common status table, the previous total dwell time is updated to “0 minutes”, the previous time is updated to “9:00”, and the number of messages is updated to “1”.

  Also, at the time of FIG. 12 (current time is 9:10), the business process A stores the message 2 in the queue. At this time, the total residence time is calculated as 0 minute + (9: 10−9: 00) × 1 = 10 minutes. In the common status table, the previous total dwell time is updated to “10 minutes”, the previous time is updated to “9:10”, and the number of messages is updated to “2”.

  Further, at the time of FIG. 13 (current time is 9:30), the business process A stores the message 3 in the queue. At this time, the total residence time is calculated as 10 minutes + (9: 30-9: 10) × 2 = 50 minutes. In the common status table, the previous total residence time is updated to “50 minutes”, the previous time is updated to “9:30”, and the number of messages is updated to “3”.

  On the other hand, at the time of FIG. 14 (current time is 10:00), the business process B extracts the message 1 from the queue. Since the message stored first is message 1, message 1 is extracted. At this time, the total residence time is calculated as 50 minutes + (10: 00-9: 30) × 3- (10: 00−9: 00) = 80 minutes. In the common status table, the previous total residence time is updated to “80 minutes”, the previous time is updated to “10:00”, and the number of messages is updated to “2”.

  Further, at the time of FIG. 15 (current time is 10:10), the business process B extracts the message 2 from the queue. Since message 2 is stored first in message 2 and message 3, message 2 is extracted. At this time, the total residence time is calculated as 80 minutes + (10: 10-10: 00) × 2- (10: 10-9: 10) = 40 minutes. In the common status table, the previous total dwell time is updated to “40 minutes”, the previous time is updated to “10:10”, and the number of messages is updated to “1”.

  Then, at the time of FIG. 16 (current time is 10:30), the business process B extracts the message 3 from the queue. At this time, the total residence time is calculated as 40 minutes + (10: 30−10: 10) × 1− (10: 30−9: 30) = 0 minutes. Then, in the common status table, the previous total residence time is updated to “0 minutes”, the previous time is updated to “10:30”, and the number of messages is updated to “0”.

  In this way, the common status table is updated during message extraction and storage, and the total residence time for the queue is calculated.

  Next, using the specific examples shown in FIGS. 17 and 18, an example of a method for adjusting the total residence time of the queue to an appropriate state by the process number adjustment processing according to the present embodiment will be described. Note that the systems shown in FIGS. 17 and 18 are simplified from the system shown in FIG.

  First, an example of a method for adjusting the total residence time of all the queues to an appropriate state in the case where the total residence times of the queue 7A, the queue 7B, and the queue 7C are all long will be described with reference to FIG. For example, the business process 11C determines that the total residence time of the queue 7C is long, and starts the business process 11C newly to increase the number of processes (first stage in FIG. 17). This shortens the total residence time of the queue 7C. Then, the business process 11B determines that the total residence time of the queue 7B is long and the total residence time of the queue 7C is short, and the business process 11B is newly activated to increase the number of processes (FIG. 17). Second stage). As a result, the frequency at which messages are stored in the queue 7C is increased, so that the total residence time of the queue 7C becomes a little longer and becomes an appropriate state, while the total residence time of the queue 7B becomes shorter. Then, the business process 11A determines that the total residence time of the queue 7A is long and the total residence time of the queue 7B is short, and the business process 11A is newly activated to increase the number of processes (FIG. 17). (3rd stage). As a result, the frequency at which messages are stored in the queue 7B increases, so that the total residence time of the queue 7B becomes a little longer and becomes an appropriate state, while the total residence time of the queue 7A is also adjusted to an appropriate state ( (4th stage of FIG. 17). That is, the total residence time of all the queues is adjusted to an appropriate state.

  Conversely, FIG. 18 shows an example of a method for adjusting the total residence time of all the queues to an appropriate state in a case where the total residence times of the queues 7A, 7B, and 7C are all short. . For example, since the business process 11C determines that the total residence time of the queue 7C is short, it stops itself and reduces the number of processes (first stage in FIG. 18). As a result, the total residence time of the queue 7C becomes longer. Then, since the business process 11B determines that the total residence time of the queue 7B is short and the total residence time of the queue 7C is long, it stops itself and reduces the number of processes (second stage in FIG. 18). . As a result, the frequency at which messages are stored in the queue 7C is reduced, so that the total residence time of the queue 7C is slightly shortened to an appropriate state, while the total residence time of the queue 7B is increased. Then, since the business process 11A determines that the total residence time of the queue 7A is short and the total residence time of the queue 7B is long, it stops itself and reduces the number of processes (third stage in FIG. 18). . As a result, the frequency at which messages are stored in the queue 7B is reduced, so that the total residence time of the queue 7B is shortened to an appropriate state, while the total residence time of the queue 7A is adjusted to an appropriate state ( (Fourth stage in FIG. 18). That is, the total residence time of all the queues is adjusted to an appropriate state.

  Although one embodiment of the present technology has been described above, the present technology is not limited to this. For example, the functional block diagram of the business system 1 described above does not necessarily correspond to an actual program module configuration.

  Further, the configuration of each table described above is an example, and the configuration as described above is not necessarily required. Further, in the processing flow, the processing order can be changed if the processing result does not change. Further, it may be executed in parallel.

  In steps S25 and S55, the number of processes to be activated is not one but may be plural.

  In steps S29 and S59, the process number adjustment processing unit stops the business process including itself. However, the process number adjustment processing unit confirms the status of other business processing processes and can stop other business processing processes. The business process may be stopped.

  As shown in FIG. 19, the business system 1 includes a memory 2501 (storage unit), a CPU 2503 (processing unit), a hard disk drive (HDD) 2505, a display control unit 2507 connected to a display device 2509, and a removable disk. A drive device 2513 for 2511, an input device 2515, and a communication control unit 2517 for connecting to a network are connected by a bus 2519. Application programs including the OS and the Web browser are stored in the HDD 2505, and are read from the HDD 2505 to the memory 2501 when executed by the CPU 2503. If necessary, the CPU 2503 controls the display control unit 2507, the communication control unit 2517, and the drive device 2513 to perform necessary operations. Further, data in the middle of processing is stored in the memory 2501 and stored in the HDD 2505 if necessary. Such a computer realizes various functions as described above by organically cooperating hardware such as the CPU 2503 and the memory 2501 described above with the OS and necessary application programs.

  The embodiments of the present technology described above are summarized as follows.

  The computer system according to the first aspect includes a first queue that stores a message in association with a storage time at which the message is stored, and a first number that is the number of messages stored in the first queue. And the first total residence time that is the sum of the residence times of the messages at the time when the previous message was extracted or stored, and the first calculation time that is the time at which the first total residence time was calculated. When a message to be processed is extracted from the first status table and the first queue, the storage time for the extracted message and the first status table stored in the first status table are stored. Based on the number of cases, the first total residence time, and the first calculation time, the total residence time of the messages stored in the first queue is calculated, and the first stay is calculated. In the stable, the first number of cases is decreased by 1, the first total residence time is updated with the total residence time of the messages stored in the first queue, and the first calculation time is updated with the current time. Calculating means.

  With such a configuration, when calculating the total residence time, the process of searching the first queue and acquiring the storage time for all messages is unnecessary, and the specific process is temporarily stopped. Since it is not necessary, the efficiency of a specific process is improved. In addition, since the total residence time is calculated at the time of message extraction, there is no need to periodically calculate the total residence time and monitor the queue, and it is difficult to determine the monitoring cycle (for example, This is also effective for systems in which the number varies greatly depending on the time zone.

  In addition, the calculation means described above determines the difference between the current time and the first calculation time stored in the first status table in the first total residence time stored in the first status table. The message stored in the first queue is obtained by adding a value obtained by multiplying the first number stored in the first status table and subtracting the difference between the current time and the storage time for the extracted message. The total residence time may be calculated. In this way, it is not necessary to search the first queue and acquire the storage time for all messages.

  In addition, a second queue that stores a message after a specific process is associated with a storage time at which the message is stored, and a second queue that is the number of messages stored in the second queue. The number of cases and the second total residence time that is the sum of the residence times of the messages at the time when the previous message was extracted or stored and the second calculation time that is the time at which the second total residence time was calculated are stored. You may make it have a 2nd status table further. Then, after the calculation means described above performs a specific process on the extracted message, the second number stored in the second status table, the second total residence time, and the second Based on the calculated time, the total residence time of messages stored in the second queue is calculated, the second number is increased by 1 in the second status table, and the messages stored in the second queue are stored. The second total residence time may be updated with the sum of the residence times, and the second calculation time may be updated with the current time. As a result, the total residence time can be calculated for the second queue that stores the message after the specific processing is performed without temporarily stopping the specific processing.

  In addition, the calculation means described above determines the difference between the current time and the second calculation time stored in the second status table in the second total residence time stored in the second status table. You may make it calculate the sum total of the residence time of the message stored in the 2nd queue by adding the value which multiplied the 2nd number stored in the 2nd status table. In this way, it is not necessary to search the second queue and acquire the storage time for all messages.

  Moreover, you may make it further have the 1 or several process part which performs a specific process. Then, each processing unit calculates the sum of the residence times of the messages stored in the second queue and the sum of the residence times of the messages stored in the first queue is equal to or greater than the first threshold. Is less than the second threshold value, the processing unit is newly activated, and the total residence time of the messages stored in the first queue is less than the first threshold value and stored in the second queue. If the total stay time of the messages being received is equal to or greater than the second threshold value, an adjusting means for stopping at least one of the processing units may be provided. For example, even if the total residence time for the first queue is long, it is not always appropriate to increase the number of processing units if the total residence time for the second queue is long. With the configuration as described above, the number of processing units can be adjusted appropriately in consideration of the total residence time for the first queue and the total residence time for the second queue. . Further, there is no need to prepare a separate monitoring unit or the like to monitor the queue, and the number of processing units that perform specific processing can be autonomously adjusted by the processing unit. Note that the processing unit is, for example, a program (that is, a process) that is executed in a computer system in order to perform specific processing. The calculation unit and the adjustment unit described above are activated when the processing unit is activated, and realize the functions described above.

  The information processing method according to the second aspect includes a first queue that stores a message and a storage time at which the message is stored in association with each other, and a first queue that is the number of messages stored in the first queue. And the first total residence time that is the sum of the residence times of the messages at the time when the previous message was extracted or stored, and the first calculation time that is the time at which the first total residence time was calculated. Executed by a computer having access to the first status table. In the information processing method, when a message to be processed is extracted from the first queue, the storage time for the extracted message and the first status table stored in the first status table are stored. Based on the number of cases 1, the first total residence time, and the first calculation time, the total residence time of the messages stored in the first queue is calculated. In the first status table, the first number of cases 1 is updated, the first total residence time is updated with the total residence time of the messages stored in the first queue, and the first calculation time is updated with the current time.

  A program for causing a computer to perform the processing according to the above method can be created. The program can be a computer-readable storage medium such as a flexible disk, a CD-ROM, a magneto-optical disk, a semiconductor memory, or a hard disk. It is stored in a storage device. The intermediate processing result is temporarily stored in a storage device such as a main memory.

  The following supplementary notes are further disclosed with respect to the embodiments including the above examples.

(Appendix 1)
A first queue that stores a message in association with a storage time at which the message is stored;
The first total residence time that is the total number of messages stored in the first queue and the total residence time of messages at the time when the previous message was extracted or stored, and the first total A first status table storing a first calculation time which is a time at which the residence time is calculated;
When a message to be processed is extracted from the first queue, the storage time for the extracted message and the first number stored in the first status table, Based on the first total residence time and the first calculation time, a total residence time of messages stored in the first queue is calculated, and in the first status table, the first Calculating means for reducing the number of cases by 1, updating the first total residence time with the total residence time of messages stored in the first queue, and updating the first calculation time with the current time;
A computer system.

(Appendix 2)
The calculating means is
In the first total residence time stored in the first status table, the difference between the current time and the first calculation time stored in the first status table is the first status table. The value stored in the first queue is added, and the difference between the current time and the storage time for the extracted message is subtracted to thereby hold the message stored in the first queue. The computer system according to appendix 1, wherein a total of time is calculated.

(Appendix 3)
A second queue for storing the message after the specific processing and the storage time when the message is stored in association with each other;
A second total residence time that is the total number of messages stored in the second queue and a total residence time of messages at the time when the previous message was extracted or stored, and the second total A second status table storing a second calculation time which is a time at which the residence time is calculated;
Further comprising
The calculating means is
After the specific processing is performed on the extracted message, based on the second number, the second total residence time, and the second calculation time stored in the second status table Calculating the total residence time of the messages stored in the second queue, incrementing the second number by 1 in the second status table, and storing the messages stored in the second queue The computer system according to claim 1 or 2, wherein the second total residence time is updated with a total of the residence times, and the second calculation time is updated with the current time.

(Appendix 4)
The calculating means is
In the second total residence time stored in the second status table, the difference between the current time and the second calculation time stored in the second status table is the second status table. The computer system according to claim 3, wherein the sum of the residence times of the messages stored in the second queue is calculated by adding a value obtained by multiplying the second number stored in the second queue. .

(Appendix 5)
And further comprising one or more processing units for performing the specific processing,
Each of the processing units
The calculating means;
When the total residence time of messages stored in the first queue is greater than or equal to a first threshold and the total residence time of messages stored in the second queue is less than a second threshold The processing unit is newly activated, and the total retention time of messages stored in the first queue is less than the first threshold and the retention of messages stored in the second queue Adjusting means for stopping at least one of the processing units when the total time is equal to or greater than the second threshold;
The computer system according to appendix 3 or 4, which has the following.

(Appendix 6)
A program for grasping the processing load,
When a message to be subjected to a specific process is extracted from the first queue that stores the message and the storage time at which the message is stored, the storage time for the extracted message, The first total residence time that is the total number of messages stored in the first queue and the total residence time of messages at the time when the previous message was extracted or stored, and the first total residence time Based on the first number, the first total residence time, and the first calculation time stored in the first status table that stores the first calculation time that is the time at which the time is calculated. , The total residence time of the messages stored in the first queue is calculated, and the first number is decreased by 1 in the first status table, and the first A program for updating the first total residence time by the total residence time of the message stored in a queue to execute a step of updating the first calculated time at the current time to the computer.

(Appendix 7)
A first queue that stores a message in association with a storage time at which the message is stored, a first number that is the number of messages stored in the first queue, and a previous message are extracted or stored. The first status table that stores the first total residence time that is the sum of the residence times of the messages at the time and the first calculation time that is the time at which the first total residence time is calculated. An information processing method executed by a simple computer,
When a message to be processed is extracted from the first queue, the storage time for the extracted message and the first number stored in the first status table, Based on the first total residence time and the first calculation time, a total residence time of messages stored in the first queue is calculated, and in the first status table, the first Information including a step of reducing the number of cases by 1, updating the first total residence time with the total residence time of messages stored in the first queue, and updating the first calculation time with the current time. Processing method.

DESCRIPTION OF SYMBOLS 1 Business system 3 Reception part 5 Transmission part 7A, 7B, 7C Queue 9A, 9B, 9C Common status table 11A, 11B, 11C Business processing process 101A, 101B, 101C Business processing part 103A, 103B, 103C Total residence time calculation part 105A , 105B, 105C Process number adjustment processing unit

Claims (7)

A first queue that stores a message in association with a storage time at which the message is stored;
The first total residence time that is the total number of messages stored in the first queue and the total residence time of messages at the time when the previous message was extracted or stored, and the first total A first status table storing a first calculation time which is a time at which the residence time is calculated;
When a message to be processed is extracted from the first queue, the storage time for the extracted message and the first number stored in the first status table, Based on the first total residence time and the first calculation time, a total residence time of messages stored in the first queue is calculated, and in the first status table, the first Calculating means for reducing the number of cases by 1, updating the first total residence time with the total residence time of messages stored in the first queue, and updating the first calculation time with the current time;
A computer system. The calculating means is
In the first total residence time stored in the first status table, the difference between the current time and the first calculation time stored in the first status table is the first status table. The value stored in the first queue is added, and the difference between the current time and the storage time for the extracted message is subtracted to thereby hold the message stored in the first queue. The computer system according to claim 1, wherein a total time is calculated. A second queue for storing the message after the specific processing and the storage time when the message is stored in association with each other;
A second total residence time that is the total number of messages stored in the second queue and a total residence time of messages at the time when the previous message was extracted or stored, and the second total A second status table storing a second calculation time which is a time at which the residence time is calculated;
Further comprising
The calculating means is
After the specific processing is performed on the extracted message, based on the second number, the second total residence time, and the second calculation time stored in the second status table Calculating the total residence time of the messages stored in the second queue, incrementing the second number by 1 in the second status table, and storing the messages stored in the second queue 3. The computer system according to claim 1, wherein the second total residence time is updated with a total of the residence times, and the second calculation time is updated with the current time. The calculating means is
In the second total residence time stored in the second status table, the difference between the current time and the second calculation time stored in the second status table is the second status table. 4. The computer according to claim 3, wherein the sum of the residence times of the messages stored in the second queue is calculated by adding a value obtained by multiplying the second number stored in the second queue. system. And further comprising one or more processing units for performing the specific processing,
Each of the processing units
The calculating means;
When the total residence time of messages stored in the first queue is greater than or equal to a first threshold and the total residence time of messages stored in the second queue is less than a second threshold The processing unit is newly activated, and the total retention time of messages stored in the first queue is less than the first threshold and the retention of messages stored in the second queue Adjusting means for stopping at least one of the processing units when the total time is equal to or greater than the second threshold;
5. The computer system according to claim 3 or 4, further comprising: A program for grasping the processing load,
When a message to be subjected to a specific process is extracted from the first queue that stores the message and the storage time at which the message is stored, the storage time for the extracted message, The first total residence time that is the total number of messages stored in the first queue and the total residence time of messages at the time when the previous message was extracted or stored, and the first total residence time Based on the first number, the first total residence time, and the first calculation time stored in the first status table that stores the first calculation time that is the time at which the time is calculated. , The total residence time of the messages stored in the first queue is calculated, and the first number is decreased by 1 in the first status table, and the first A program for updating the first total residence time by the total residence time of the message stored in a queue to execute a step of updating the first calculated time at the current time to the computer. A first queue that stores a message in association with a storage time at which the message is stored, a first number that is the number of messages stored in the first queue, and a previous message are extracted or stored. The first status table that stores the first total residence time that is the sum of the residence times of the messages at the time and the first calculation time that is the time at which the first total residence time is calculated. An information processing method executed by a simple computer,
When a message to be processed is extracted from the first queue, the storage time for the extracted message and the first number stored in the first status table, Based on the first total residence time and the first calculation time, a total residence time of messages stored in the first queue is calculated, and in the first status table, the first Information including a step of reducing the number of cases by 1, updating the first total residence time with the total residence time of messages stored in the first queue, and updating the first calculation time with the current time. Processing method.

Images