JP5707348B2 - Event-driven processing execution apparatus and operation method thereof - Google Patents

Description

  The present invention relates to an event-driven processing execution apparatus capable of obtaining a CPU usage rate by only a single service and an operation method thereof.

  Conventionally, in order to provide a plurality of network services, servers are divided for each service, server congestion is determined based on the CPU usage rate of the entire server, and traffic has been regulated. When a plurality of services share a CPU on a single server, it is necessary to control the CPU usage by each service so that a specific service does not occupy the CPU.

  As a CPU usage rate control method in the server, there are the following methods.

  1. A desired CPU usage rate is designated for each thread, and when the CPU usage rate exceeds a specified value, processing in the thread is stopped for a certain period (see Non-Patent Document 1).

  2. The CPU time per fixed period is controlled by designating the relative distribution of CPU time of threads (see Non-Patent Document 2).

  3. In a system in which a plurality of CPUs can be used simultaneously, an upper limit of the CPU usage rate is set by specifying a CPU that executes a thread (see Non-Patent Document 3).

CPU Usage Limiter for Linux, [online], [Search January 27, 2012], Internet <URL: http://cpulimit.sourceforge.net/> Red Hat Enterprise Linux 6 Resource Management Guide, “3.2, cpu”, [online], [searched on January 27, 2012], Internet <URL: http://docs.redhat.com/docs/en- JP / Red_Hat_Enterprise_Linux / 6 / html / Resource_Management_Guide /> taskset (1)-Linux man page, [online], [searched January 27, 2012], Internet <URL: http://linux.die.net/man/1/taskset>

  However, in the above technique, there is no means for measuring the CPU usage rate of each service for a portion shared by a plurality of network services such as an operating system (hereinafter referred to as OS) and middleware, and the CPU usage rate of each service cannot be controlled. The problem is that a plurality of network services cannot be controlled by a single call processing server. In order to efficiently use the CPU, when processing a plurality of services with the same thread, there is no means for measuring the CPU usage rate for each service. Even if the processes of each service are separated and CPUs are allocated separately, when the service is executed on the same OS / middleware, the CPU usage rate of the OS / middleware part cannot be divided for each service.

  When processing a plurality of services, the same problem occurs not only in the call processing server. For example, in an event-driven server, that is, an event-driven processing execution apparatus, if the CPU usage rate cannot be obtained for each service, for example, the CPU usage rate corresponding to one event increases for a certain service. In this case, since the CPU usage rate that can be used for processing other services decreases, there is a possibility that the processing of other services cannot be executed.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide an event-driven processing execution apparatus capable of obtaining a CPU usage rate by only a single service and an operation method thereof. There is to do.

In order to solve the above-described problem, the first aspect of the present invention extracts a process execution unit that executes processing using a CPU, an event queue that stores event information for each service, and event information from the event queue, A process request unit that requests the process execution unit to perform a process corresponding to the event information, and a service event information extraction number that is the number of event information extracted per unit time for each service performed by the process request unit; A history information storage unit that stores history information including a CPU total usage rate that is a total usage rate of the CPU in the unit time for each unit time, and a usage rate by only the service of the CPU for each service. CPU usage rate meter for each service for determining the CPU usage rate for each service from the number of event information fetched for each service. CPU usage rate calculation function storage unit by service in which functions are stored, and two or more pieces of history information are read from the history information storage unit, and the number of service event information fetched by the history information and the total CPU usage rate are used. The CPU usage rate calculation function for each service is calculated for each service and stored in the CPU usage rate calculation function storage unit for each service, and the event information for each service is calculated for each service. A service-specific event information extraction number upper limit storage unit that stores an upper limit of the number of event information to be extracted, which is an upper limit of the extraction number, and the processing request unit, when extracting event information from the event queue, the history information storage unit Event information for each service from the corresponding history information The service event information retrieval number upper limit storage unit from the service event information retrieval number upper limit storage unit, and if the service event information retrieval number is less than or equal to the service event information retrieval number upper limit, An event-driven process execution device that takes out event information and requests the process execution unit to perform a process corresponding to the event information is used as a solution means.

  For example, the event-driven processing execution device includes, for each service, a service-specific CPU usage upper limit storage unit that stores a service-specific CPU usage upper limit that is an upper limit of the usage rate of only the CPU concerned service, and for each service, A service-specific CPU usage rate calculation function is obtained from the service-specific CPU usage rate calculation function storage unit, and a service-specific CPU usage rate upper limit unit is obtained from the service-specific CPU usage rate upper limit storage unit. A service-specific event information extraction number upper limit calculation unit that calculates a service-specific event information extraction number upper limit of the service using the service-specific CPU usage rate upper limit and stores the service-specific event information extraction number upper limit storage unit;

The second aspect of the present invention is an operation method of an event-driven process execution device, wherein the event-driven process execution device stores a process execution unit that executes a process using a CPU and event information for each service. An event queue, a process request unit that extracts event information from the event queue and requests the process execution unit to perform processing corresponding to the event information, and an event per unit time for each service performed by the process request unit A history information storage unit that stores, for each unit time, history information that includes the number of service-related event information extractions that are information extractions and the CPU total usage rate that is the total usage rate of the CPU in the unit time, and the service Separately, the CPU usage rate for each service, which is the usage rate of the CPU only for that service, is used as the event information for each service for that service. And service-specific CPU usage calculation function storage unit that service by CPU usage rate calculation function for obtaining from the take-out number is stored, by the service, the an upper limit of each service event information number extraction by service event information extraction speed limit And an event information retrieval number upper limit storage unit for each service that stores the event information , when the processing request unit retrieves event information from the event queue, the operation method extracts the event from the corresponding history information in the history information storage unit. The number of event information fetched for each service of the information service is acquired, the service information event information fetch count upper limit for the service is obtained from the service event count fetch count upper limit storage unit, and the service event information fetch count is the service event count. Less than the upper limit of information retrieval Retrieves the event information, and requests processing corresponding to the event information to the processing execution unit, the service-specific CPU usage calculation function calculating unit of the event-driven processing execution apparatus, two or more from the history information storage unit , The CPU usage rate calculation function for each service is calculated for each service, and the CPU usage rate calculation for each service is calculated using the number of event information fetched for each service and the total CPU usage rate in the history information. The solving means is an operation method of an event-driven process execution device characterized by being stored in a function storage unit.

  For example, the event-driven processing execution device includes a service-specific CPU usage upper limit storage unit that stores a service-specific CPU usage upper limit that is an upper limit of the usage rate of only the CPU for each service, and the operation The method is such that the service-specific event information retrieval number upper limit calculation unit of the event-driven processing execution device is configured to store, for each service, the service-specific CPU usage rate calculation function from the service-specific CPU usage rate calculation function storage unit, and the service-specific CPU usage. Each service CPU usage upper limit is obtained from the rate upper limit storage unit, and the service-specific CPU usage rate calculation function and the service-specific CPU usage upper limit are calculated. The service-specific event information extraction number upper limit storage unit is stored.

  According to the present invention, it is possible to obtain the CPU usage rate by only a single service.

It is a functional block diagram which shows the structure of the event drive type | mold process execution apparatus which concerns on this Embodiment. It is a figure which shows an example of a structure and data of the log | history information storage part. It is a figure which shows an example of a structure and data of the CPU usage rate calculation function memory | storage part 108 classified by service. It is a figure which shows an example of a structure and data of the event information extraction number upper limit storage part 110 according to service. It is a figure which shows an example of a structure and data of a storage area comprised in the CPU usage rate upper limit storage unit 111 by service. It is a flowchart which shows the initial operation | movement of the CPU usage rate calculation function calculation part 109 classified by service. It is a flowchart which shows operation | movement of the CPU utilization rate upper limit memory | storage part 111 classified by service. It is a flowchart which shows operation | movement of the event information extraction number upper limit calculation part 112 according to service. 4 is a flowchart showing an operation of a process request unit 105. It is a flowchart which shows the operation | movement after the initial operation | movement of the CPU usage rate calculation function calculation part 109 classified by service.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a functional block diagram showing the configuration of the event-driven process execution device according to the present embodiment.

  The event-driven processing execution device 1 is connected to the input device 2 and has a CPU (Central Processing Unit) 101, a processing execution unit 102 that executes processing using the CPU 101, and an event queue 103 that stores event information for each service. An event queue update unit 104 that stores event information in the event queue 103, a process request unit 105 that extracts event information from the event queue 103, and requests the process execution unit 102 to perform processing corresponding to the event information, and a processing request The history information including the number of event information extracted by service, which is the number of event information extracted per unit time by service performed by the unit 105, and the CPU total usage rate, which is the total usage rate of the CPU 101 in the unit time, is a unit. History information storage unit 106 that stores every hour and CPU total usage rate The CPU total usage rate acquisition unit 107 stores, for each service, a CPU usage rate calculation function for each service for obtaining the CPU usage rate for each service, which is the usage rate of only the service of the CPU 101, from the number of event information fetched for each service. Two or more pieces of history information are read out from the service-specific CPU usage rate calculation function storage unit 108 and the history information storage unit 106, and the service-specific event information extraction number and the CPU total usage rate in the history information are used to Separately, the service-specific CPU usage rate calculation function calculation unit 109 that calculates the service-specific CPU usage rate calculation function and stores it in the service-specific CPU usage rate calculation function storage unit 108, and the service-specific event information retrieval number for each service A service-specific event that stores the upper limit of the number of event information fetched by a service Output information retrieval number upper limit storage unit 110, service-specific CPU usage rate upper limit storage unit 111 for storing the service-specific CPU usage rate upper limit, which is the upper limit of the usage rate of only the service of the CPU 101, for each service, service-specific, service-specific The CPU usage rate calculation function for each service is obtained from the CPU usage rate calculation function storage unit 108, and the CPU usage rate upper limit for each service is obtained from the CPU usage rate upper limit storage unit 111 for each service. A service-specific event information extraction number upper limit calculation unit 112 that calculates the service-specific event information extraction number upper limit of the service using the rate upper limit and stores the service-specific event information extraction number upper limit storage unit 110 is provided. The event-driven process execution device 1 is, for example, a call processing server used for call processing.

  FIG. 2 is a diagram illustrating an example of the configuration and data of the history information storage unit 106.

  For example, history information is generated in the history information storage unit 106 every time a time zone of a unit time of 1 second newly starts. In the history information, the number of service-specific event information fetched in the time zone is recorded. In the history information, for example, an average CPU total usage rate in the time zone is recorded.

  For example, in the history information of the time period T1, the number of event information extracted by service “280” for service A, the number of event information extracted for service B of service B “560”, and the average CPU total usage rate “80%” are included. It is recorded.

  For example, in the history information of the time period T2, the number of event information retrieved by service “310” for service A, the number of event information retrieved by service “480” for service B, and the average CPU total usage rate “83%” are included. It is recorded.

  FIG. 3 is a diagram illustrating an example of the configuration and data of the service-specific CPU usage rate calculation function storage unit 108.

  The service-specific CPU usage rate calculation function storage unit 108 stores a service-specific CPU usage rate calculation function for each service. For example, when the service CPU usage rate is proportional to the number of service event information fetched, the service CPU usage rate calculation function is actually a proportional coefficient.

  The service-specific CPU usage rate calculation function storage unit 108 stores, for example, the service-specific CPU usage rate calculation function (proportional coefficient) f_A of the service A and the service-specific CPU usage rate calculation function (proportional coefficient) f_B of the service B. The

  FIG. 4 is a diagram illustrating an example of the configuration and data of the service-specific event information extraction number upper limit storage unit 110.

  The service-specific event information retrieval number upper limit storage unit 110 stores the service-specific event information retrieval number upper limit for each service.

  The service-specific event information extraction number upper limit storage unit 110 stores, for example, a service A event-specific event information extraction number upper limit E_Amax and a service B service-specific event information extraction number upper limit E_Bmax.

  FIG. 5 is a diagram illustrating an example of the configuration and data of the storage area configured in the service-specific CPU usage rate upper limit storage unit 111.

  In the storage area of the CPU usage rate upper limit storage unit 111 for each service, the CPU usage rate upper limit for each service is stored for each service.

  In the storage area of the service-specific CPU usage rate upper limit storage unit 111, for example, the service-specific CPU usage rate upper limit “50%” of the service A and the service-specific CPU usage rate upper limit “50%” are stored.

  FIG. 6 is a flowchart showing the initial operation of the service-specific CPU usage rate calculation function calculation unit 109.

  The CPU usage rate calculation function calculation unit 109 for each service, when the event-driven process execution apparatus 1 is intentionally processing only one service, when a certain time period is over, the process request unit 105, the number of service-specific event information fetched in that time zone is acquired (S1), and the CPU total usage rate acquisition unit 107 obtains the CPU total usage rate in that time zone (only one service is processed). Therefore, the CPU usage rate for each service of the service is acquired (S3). That is, the process request unit 105 calculates the number of event information extracted from the event queue 103 in the time period (the number of event information extracted by service), and provides the history information storage unit 107 (S1). Further, the CPU total usage rate acquisition unit 107 continues to measure the usage rate of the CPU 101 in the time period, and gives the average (CPU total usage rate) to the history information function unit 107 (S3). Then, the history information function unit 107 provides the service-specific event information extraction number and the CPU total usage rate to the service-specific CPU usage rate calculation function calculation unit 109.

  Next, the service-specific CPU usage rate calculation function calculation unit 109, for example, divides the total CPU usage rate by the number of service-specific event information fetches, and calculates the proportional coefficient that is the calculation result, that is, the service-specific CPU usage rate calculation function. The data is stored in the separate CPU usage rate calculation function storage unit 108 (S5).

  If there is another service (S7: YES), the service-specific CPU usage rate calculation function calculation unit 109 executes steps S1 to S5 in the same manner in a situation where only the service is being processed. To do.

  If the service-specific CPU usage rate calculation function calculation unit 109 executes steps S1 to S5 for all the services (S7: NO), the process ends.

  FIG. 7 is a flowchart showing the operation of the service-specific CPU utilization upper limit storage unit 111.

  If the CPU usage rate upper limit for each service is input from the input device 2 to the service-specific CPU usage upper limit storage unit 111, the service-specific CPU usage upper limit storage unit 111 stores the input service-specific CPU usage upper limit for its own storage area. If another CPU usage rate upper limit already exists, it is stored instead (S11). If the service-specific CPU usage rate upper limit storage unit 111 receives another service-specific CPU usage rate upper limit (S13: YES), it also acquires and stores it (S11). If the CPU usage rate upper limit storage unit 111 for each service executes Steps S1 to S3 for all services (S13: NO), the process ends.

  FIG. 8 is a flowchart showing the operation of the service-specific event information extraction number upper limit calculation unit 112.

  As an initial operation, the service-specific event information retrieval number upper limit calculation unit 112 stores the service-specific CPU usage rate calculation function for all services in the service-specific CPU usage rate calculation function storage unit 108, and uses the service-specific CPU usage rate. If the service upper limit CPU usage rate upper limit is stored for all services in the rate upper limit storage unit 111 (S21: YES), or the subsequent operation is as follows: If at least one of the CPU usage rate upper limits is changed (S21: YES), a service-specific CPU usage rate calculation function is acquired from the service-specific CPU usage rate calculation function storage unit 108 for the corresponding service (S23).

  Next, the service-specific event information retrieval number upper limit calculation unit 112 is an inverse function of the service-specific CPU usage rate calculation function, that is, a function for obtaining the service-specific event information retrieval number from the service-specific CPU usage rate (hereinafter referred to as service-specific CPU usage rate). The event information extraction number calculation function is obtained (S25). When the service-specific CPU usage rate calculation function is a proportional coefficient, the service-specific event information extraction number calculation function is the inverse of the proportional coefficient.

  The service-specific event information retrieval number upper limit calculation unit 112 performs this for all services as an initial operation, and performs the corresponding service for subsequent operations.

  Next, the service-specific event information retrieval number upper limit calculation unit 112 acquires the service-specific CPU usage rate upper limit from the service-specific CPU usage rate upper limit storage unit 111 and applies the service-specific event information retrieval number upper limit calculation function to this, for example, Multiplying the reciprocal, the calculation result, that is, the upper limit of the number of service-specific event information extraction (the upper limit of the number of service-specific event information extraction) is calculated. Then, the service-specific event information retrieval number upper limit calculation unit 112 sets the calculated event-specific event information retrieval number upper limit to the service-specific event information retrieval number upper limit storage unit 110, particularly when the service-specific event information retrieval number upper limit already exists. Instead, it is stored (S27).

  The service-specific event information retrieval number upper limit calculation unit 112 performs this for all services as an initial operation, and performs the corresponding service for subsequent operations.

  After that, when at least one of the service-specific CPU usage rate calculation function and the service-specific CPU usage rate upper limit is changed for a certain service (S21: YES), the service-specific event information retrieval number upper limit calculation unit 112 performs the processing of steps S23 to S27. I do.

  FIG. 9 is a flowchart showing the operation of the processing request unit 105.

  First, the process request unit 105 inquires of the process execution unit 102 whether or not the process can be requested (S31). If the process request unit 105 obtains an answer that the request is not possible (S31: NO), the process is terminated.

  On the other hand, if the processing request unit 105 obtains an answer that the request can be made (S31: YES), next, the history request storage unit 106 searches the history information storage unit 106 for history information about the time zone including the current time, and from the history information. In addition to obtaining the number of service-specific event information retrievals for the same service as the event information service to be retrieved, the service-specific event information retrieval number upper limit storage unit 110 obtains the service-specific event information retrieval number upper limit, It is determined whether or not the number of event information fetched is greater than the service event fetch count upper limit (S33).

  As a method of determining event information to be extracted, for example, a method of extracting event information of a service determined by a round robin method is conceivable.

  Alternatively, when event information is stored in the event queue 103, a method of taking out event information of the same service as that event information service is also conceivable.

  Also, a method of taking out the event information stored in the event queue 103 earliestly from the event information of the service determined in this way can be considered.

  If the number of service-specific event information extraction number is less than or equal to the service-specific event information extraction number upper limit (S33: NO), the process request unit 105 extracts the event information from the event queue 103 and performs processing corresponding to the event information on the process execution unit 102. Request (S35).

  Next, the processing request unit 105 adds 1 to the number of event information fetched by service for the same service as the service of the stored event information in the retrieved history information (S37), and ends the processing.

  On the other hand, if the number of service-specific event information extraction numbers exceeds the service-specific event information extraction number upper limit (S33: YES), the processing request unit 105 performs the processing of S33 to S37 for another service (S34: YES, S33: NO). ). For all services, if the number of service-specific event information extraction numbers exceeds the upper limit of service-specific event information extraction numbers, the process ends.

  FIG. 10 is a flowchart showing the operation after the initial operation of the service-specific CPU usage rate calculation function calculation unit 109.

  The CPU usage rate calculation function calculation unit 109 for each service, in the event-driven processing execution device 1, performs normal operation, that is, if a certain time period ends in a situation where all services can be processed if necessary, the history information Two or more pieces of history information are read from the storage unit 106 (S41).

  Next, the service-specific CPU usage rate calculation function calculation unit 109 calculates a service-specific CPU usage rate calculation function for each service using the number of service-specific event information fetched numbers in the history information and the total CPU usage rate. Instead of the CPU usage rate calculation function for each service in the separate CPU usage rate calculation function storage unit 108, the calculated CPU usage rate calculation function for each service is stored (S43), and the process is terminated.

  Specifically, for example, in a situation where only the services A and B can be processed, the service-specific CPU usage rate calculation function calculation unit 109 reads two pieces of history information (S41).

  Next, the service-specific CPU usage rate calculation function calculation unit 109 includes the service-related event information extraction number E_A1 of the service A, the service-specific event information extraction number E_B1 of the service B, and the CPU total usage rate R_1 in one history information. On the other hand, if the other history information includes the number of event information fetched by service E_A2 of service A, the number of event information fetched by service E_B2 of service B, and the total CPU usage rate R_2, the following simultaneous equations (1) Solve. That is, the service-specific CPU usage rate calculation function calculation unit 109 obtains the service-specific CPU usage rate calculation function f_A of the service A and the service-specific CPU usage rate calculation function f_B of the service B, and stores the service-specific CPU usage rate calculation function f_B. The service-specific CPU usage rate calculation functions f_A and f_B are stored in place of the service-specific CPU usage rate calculation functions stored in the unit 108 (S43).

f_A × E_A1 + f_B × E_B1 = R_1
f_A × E_A2 + f_B × E_B2 = R_2 (1)
Also, for example, the service-specific CPU usage rate calculation function calculation unit 109 obtains the service-specific CPU usage rate calculation function using the nonlinear least square method (S43).

  For example, for the service A, when the CPU usage rate corresponding to one event information is the same as before, the service-specific CPU usage rate calculation function of the service A in the service-specific CPU usage rate calculation function storage unit 108 is consequently obtained. Will not be changed.

  On the other hand, for example, for the service B, when the CPU usage rate corresponding to one event information is higher than before, for example, the CPU usage rate calculation for each service of the service B in the service-specific CPU usage rate calculation function storage unit 108 is calculated. The function will eventually change.

  Here, since the CPU usage rate corresponding to one event information is higher than before, the service-specific CPU usage rate calculation function (proportional coefficient) is higher than before.

  In this case, in FIG. 8, since it is determined that the service-specific CPU usage rate calculation function has been changed for service B (S21: YES), in the subsequent step S25, the number of new event-specific event information extractions for service B A calculation function is obtained.

  Here, since the service-specific CPU usage rate calculation function (proportional coefficient) is larger than before, the new service-specific event information extraction number calculation function (reciprocal of the proportional coefficient) becomes smaller than before.

  In the subsequent step S27, a new upper limit on the number of event information by service for service B is obtained. For example, if the service-specific CPU usage upper limit of service B has not been changed, the service-specific event information retrieval number calculation function (reciprocal of the proportionality coefficient) is smaller than before, so the new event-specific event information retrieval number upper limit is the previous one. Lower.

  In FIG. 9, the processing request unit 105 controls the number of service-specific event information extractions to be less than or equal to the new service-specific event information extraction number upper limit lower than before (S33). Maintain the value of.

  That is, for service B, even if the CPU usage rate corresponding to one event information is higher than before, the CPU usage rate by service of service B (for example, the average CPU usage rate of service B in the same time zone) is Since there is no change, for example, there is no adverse effect on the other service A, and the other service A can be processed within the range of the remaining CPU usage rate.

  Further, for example, when it is desired to use a larger number of CPUs for the processing of the service B, the input device 2 provides a lower CPU usage rate upper limit for each service for the service A and a higher service usage for the service B than the previous service. The CPU usage rate upper limit is input.

  For example, if the service-specific CPU usage upper limit for services A and B is 50%, the new service-specific CPU usage upper limits for services A and B are 30% and 70%, respectively.

  In this case, in FIG. 8, since it is determined that the service-specific CPU usage upper limit has been changed for services A and B (S21: YES), in the subsequent step S27, new service-specific events for services A and B are performed. An upper limit on the number of information retrieval is obtained.

  Since the service-specific CPU usage upper limit of service A is lower than before, the upper limit of the number of new service-specific event information extraction for service A is also lower than before.

  On the other hand, the service-specific CPU utilization upper limit for service B is higher than before, so the new service-specific event information retrieval number upper limit for service B is also higher than before.

  In FIG. 9, the processing request unit 105 controls the number of event information fetched by service A to be lower than the upper limit of the number of new event information fetched by service lower than before (S33). The processing amount per hit decreases. Further, the processing request unit 105 controls the number of service information extracted by service B to be lower than the upper limit of the number of new event information extracted by service (S33). The number of processes can be increased.

  As described above, according to the event-driven processing execution apparatus 1 according to the present embodiment, the CPU usage for each service is classified according to service using the number of event information extracted for each service in the history information and the total CPU usage rate. Since the CPU usage rate calculation function calculation unit for each service for calculating the rate calculation function is provided, the CPU usage rate for each service can be calculated by applying the CPU usage rate calculation function for each service to the number of event information fetched by service. Can be sought.

  In other words, the CPU usage rate for each service can be obtained including a portion shared by a plurality of network services such as OS and middleware. Thereby, for example, when the CPU usage rate for each service exceeds a preset threshold, the corresponding service can be interrupted. In addition, as described above, the service-specific CPU usage rate can be controlled by changing the upper limit of the number of service-specific event information extractions. For example, if the call processing server is configured like the event-driven process execution device 1, a plurality of network services can be freely controlled.

  A computer program for causing a computer to function as the event-driven processing execution apparatus 1 can be recorded on a computer-readable recording medium such as a semiconductor memory, a magnetic disk, an optical disk, a magneto-optical disk, a magnetic tape, and the Internet. It is possible to distribute widely through the communication network.

1 Event Driven Processing Execution Device 2 Input Device 101 CPU (Central Processing Unit)
102 process execution unit 103 event queue 104 event queue update unit 105 process request unit 106 history information storage unit 107 CPU total usage rate acquisition unit 108 CPU usage rate calculation function storage unit by service 109 CPU usage rate calculation function calculation unit by service 110 service Separate event information retrieval number upper limit storage unit 111 Service-specific CPU utilization upper limit storage unit 112 Service-specific event information retrieval number upper limit calculation unit

Claims (5)

A process execution unit that executes processes using a CPU;
An event queue that stores event information for each service;
A process request unit that retrieves event information from the event queue and requests the process execution unit to perform a process corresponding to the event information;
Including the number of event information extracted by service, which is the number of event information extracted per unit time for each service performed by the processing request unit, and the CPU total usage rate, which is the total usage rate of the CPU in the unit time. A history information storage unit for storing history information per unit time;
For each service, a CPU usage rate for each service that stores a CPU usage rate for each service for calculating the CPU usage rate for each service, which is the usage rate of the CPU only for that service, from the number of event information fetched for each service. A calculation function storage unit;
Two or more pieces of history information are read from the history information storage unit, and the CPU usage rate calculation function for each service is calculated for each service using the number of event information fetched for each service in the history information and the total CPU usage rate. A CPU usage rate calculation function calculation unit for each service to be stored in the CPU usage rate calculation function storage unit for each service ;
A service-specific event information retrieval number upper limit storage unit that stores an upper limit of service-related event information retrieval number, which is an upper limit of the service-specific event information retrieval number, for each service;
The processing request unit
In retrieving event information from the event queue, the event information retrieval number for each service of the event information is obtained from the corresponding history information in the history information storage unit, and the event information retrieval number upper limit storage unit for each service is retrieved from the upper limit storage unit. Obtain the upper limit of the number of event information to be retrieved for each service, and if the number of event information to be retrieved for each service is less than the upper limit of the number of event information to be retrieved for each service, retrieve the event information and request the process execution unit to perform processing corresponding to the event information An event-driven process execution device. For each service, a service-specific CPU usage rate upper limit storage unit that stores a service-specific CPU usage rate upper limit that is an upper limit of the usage rate of the CPU only by the service;
For each service, a CPU usage rate calculation function for each service is obtained from the CPU usage rate calculation function storage unit for each service, and a CPU usage rate upper limit for each service is obtained from the CPU usage rate upper limit storage unit for each service. By using the calculation function and the service-specific CPU usage upper limit, the service-specific event information retrieval number upper limit is calculated and stored in the service-specific event information retrieval number upper limit storage unit. The event-driven process execution device according to claim 1, further comprising: An event-driven process execution device operating method comprising:
The event-driven process execution device includes:
A process execution unit that executes processes using a CPU;
An event queue that stores event information for each service;
A process request unit that retrieves event information from the event queue and requests the process execution unit to perform a process corresponding to the event information;
Including the number of event information extracted by service, which is the number of event information extracted per unit time for each service performed by the processing request unit, and the CPU total usage rate, which is the total usage rate of the CPU in the unit time. A history information storage unit for storing history information per unit time;
For each service, a CPU usage rate for each service that stores a CPU usage rate for each service for calculating the CPU usage rate for each service, which is the usage rate of the CPU only for that service, from the number of event information fetched for each service. A calculation function storage unit ;
A service-specific event information retrieval number upper limit storage unit that stores an upper limit of service-related event information retrieval number, which is an upper limit of the service-specific event information retrieval number, for each service ;
The operation method is as follows:
The processing request unit
In retrieving event information from the event queue, the event information retrieval number for each service of the event information is obtained from the corresponding history information in the history information storage unit, and the event information retrieval number upper limit storage unit for each service is retrieved from the upper limit storage unit. Obtain the upper limit of the number of event information to be retrieved for each service, and if the number of event information to be retrieved for each service is less than the upper limit of the number of event information to be retrieved for each service, retrieve the event information and request the process execution unit to perform processing corresponding to the event information And
CPU usage rate calculation function calculation unit for each service of the event-driven processing execution device,
Two or more pieces of history information are read from the history information storage unit, and the CPU usage rate calculation function for each service is calculated for each service using the number of event information fetched for each service in the history information and the total CPU usage rate. An operation method of an event-driven process execution device, wherein the event-driven process execution device is stored in the service-specific CPU usage rate calculation function storage unit. The event-driven process execution device includes:
A service-specific CPU usage rate upper limit storage unit that stores a service-specific CPU usage rate upper limit that is an upper limit of the usage rate by only the service of the CPU for each service,
The operation method is as follows:
The event-specific process information take-out number upper limit calculation unit of the event-driven process execution device,
For each service, a CPU usage rate calculation function for each service is obtained from the CPU usage rate calculation function storage unit for each service, and a CPU usage rate upper limit for each service is obtained from the CPU usage rate upper limit storage unit for each service. calculation functions and by using the service-specific CPU utilization limits, the service-specific event information extraction maximum number of the service to calculate, according to claim 3, wherein the to be stored in the service-specific event information extraction speed limit storage unit Method of event-driven process execution apparatus. Computer program for causing a computer to function as an event-driven process execution apparatus according to claim 1 or 2 wherein.

Images