JP6357825B2 - Congestion control system and congestion control method - Google Patents

Description

  The present invention relates to a congestion control system and a congestion control method for controlling congestion caused by an increase in processing targets.

  In addition to performing normal information processing, each functional unit that configures the information processing device monitors the congestion status that may occur in the information processing device or processor (CPU: Central Processing Unit), avoids the occurrence of congestion, and generates In such a case, it is necessary to recover from the congestion. For example, an intranet packet relay device suppresses the occurrence of a communication failure due to traffic exceeding the device processing capability by monitoring the congestion and expanding the device processing capability as the packet traffic increases. To avoid congestion.

  However, if congestion monitoring and congestion avoidance are executed in parallel in the main function unit that performs normal processing, the processing capability (performance) of the information processing apparatus itself deteriorates due to the processing load of congestion monitoring and congestion avoidance itself. In addition, when a dedicated hardware function unit that performs congestion monitoring is installed around a processor that performs normal processing to perform congestion monitoring, the cost for installing the function unit increases.

  On the other hand, if the congestion monitoring of the main function unit is a function unit located around the main function unit, for example, if it is a function unit of the active standby configuration, the standby main function unit estimates the congestion status of the active function unit Thus, a method is disclosed in which congestion monitoring can be performed without degrading the processing capability of the main function unit (Patent Document 1).

JP 2009-212862 A

  However, although the method of Patent Document 1 describes the congestion monitoring of the main function unit, it does not disclose how to monitor the congestion state of the sub function unit and how to recover when it falls into the congestion state. . Since the slave function unit also performs multitask processing by the processor, it may fall into a congestion state as with the master function unit. Therefore, in order to avoid the congestion of the entire system, it is necessary to perform the congestion monitoring and congestion recovery processing of the slave function unit.

  The present invention has been made in view of the above problems. Congestion monitoring is performed by monitoring the congestion of the subordinate function unit together with the congestion monitoring of the main function unit, and performing the congestion recovery process when the subordinate function unit falls into a congestion state. It is an object to provide a control system and a congestion control method.

  In order to achieve the above object, the first feature of the present invention is that an information processing execution unit that processes a processing message sent from the outside, and a congestion monitor that monitors the number of processing messages provided in the information processing execution unit. A congestion control system including a slave unit and a slave function unit connected to the information processing execution unit and the congestion monitoring unit for executing a task included in the processing message. There are means for estimating the processing time and determining whether the state is a congestion state or a non-congestion state based on the processing time, the number of processing messages and the threshold, and (b) means for notifying the information processing execution unit of the determined content. The subordinate function unit (c) counts the number of unfinished processes in the cycle of at least one task included in the process message, and based on the unfinished number and the threshold, Or (d) means for notifying the information processing execution unit of the determined content, and the information processing execution unit (e) when the congestion monitoring unit fails, Means for estimating the processing time for the processing message by itself and determining whether it is a congestion state or a non-congestion state based on the processing time, the number of processing messages, and a threshold; and (f) notified from the congestion monitoring unit and the subordinate function unit The gist of the present invention is that it is a congestion control system having means for suppressing processing for a processing message when the content or the content determined by itself indicates a congestion state.

  The second feature of the present invention is that an information processing execution unit that processes a processing message sent from outside, a congestion monitoring unit that is attached to the information processing execution unit and monitors the number of processing messages, and an information processing execution unit And a congestion control method used in a system having a slave unit connected to the congestion monitoring unit for executing a task included in the processing message, wherein the congestion monitoring unit estimates (a) a processing time for the processing message And a step of determining whether the state is a congestion state or a non-congestion state based on the processing time, the number of processing messages, and a threshold, and (b) a step of notifying the information processing execution unit of the determined content. (C) Counts the number of incomplete processing within a period of at least one task included in the processing message, and determines whether the state is congested or non-congested based on the number of incomplete and the threshold. And (d) a step of notifying the information processing execution unit of the determined content, and the information processing execution unit (e) itself processes messages when the congestion monitoring unit fails. And a step of determining whether it is a congestion state or a non-congestion state based on the processing time, the number of processing messages and a threshold value, and (f) content notified from the congestion monitoring unit and the subordinate function unit, or by itself The gist of the present invention is a congestion control method including a step of suppressing processing on a processing message when the content determined by indicates a congestion state.

  According to the congestion control system and the congestion control method of the present invention, the congestion monitoring of the main function unit and the congestion monitoring of the sub function unit are performed, and the congestion recovery process is performed when the congestion state occurs. Thereby, the occurrence of congestion in the entire system can be avoided, and the stability of the system can be improved.

It is a schematic block diagram which shows embodiment of the congestion control system of 1st embodiment. It is a figure which shows the internal structure of the active main function part. It is a figure which shows the internal structure of a backup main function part. It is a figure which shows the internal data of a load determination memory | storage part. It is a figure which shows the internal structure of a subordinate function part. It is a figure which shows the outline of the task process in a subordinate function part. It is a figure which shows the outline of the task process in a subordinate function part. It is a flowchart which shows operation | movement of a backup main function part. It is a flowchart which shows operation | movement of the working main function part. It is a flowchart which shows operation | movement of a subordinate function part. It is a schematic block diagram which shows embodiment of the congestion control system of 2nd embodiment. It is a figure which shows the internal structure of an information processing part. It is a figure which shows the internal structure of a congestion monitoring part. It is a flowchart which shows operation | movement of an information processing part. It is a flowchart which shows operation | movement of a congestion monitoring part.

  Next, an embodiment of the present invention will be described with reference to the drawings. In the following description of the drawings, the same or similar parts are denoted by the same or similar reference numerals. However, it should be noted that the drawings are schematic. Furthermore, it should be noted that the embodiment described below is an example, and can be appropriately changed within the scope of the same essence.

Further, in the following description of the embodiments, each component of each device indicates a functional unit block, not a hardware unit configuration. Each component of each device is an arbitrary computer CPU, memory, a program loaded in the memory, a storage medium such as a hard disk for storing the program, and an arbitrary combination of hardware and software centering on a network connection interface. Realized. There are various modifications of the implementation method and apparatus.
<First Embodiment>
(Congestion control system)
As shown in FIG. 1, the congestion control system 10 according to the first embodiment of the present invention includes a main function unit 11 and a plurality of sub function units 12, 13, 14, and 15. The main function unit 11 has a working spare configuration and includes a working main function unit 111 and a spare main function unit 112. The active main function unit 111 receives a message requesting a predetermined process (hereinafter referred to as “process message”) from an external terminal (hereinafter simply referred to as “external”) that requests the main function unit 11 to process. , And functions as an information processing execution unit that performs predetermined information processing on the processing message. The spare main function unit 112 functions as a congestion monitoring unit that monitors the number of messages processed by the active main function unit 111. The active main function unit 111 and the standby main function unit 112 operate with the same CPU (Central Processing Unit). Since the CPU is shared, the standby main function unit can monitor the congestion state of the active main function unit without providing dedicated hardware. Note that, for a specific process in the process message, the active main function unit 111 requests the sub function units 12 to 15 in the system to perform the process.

Each of the slave function units 12 to 15 executes a predetermined process when receiving a message from the active main function unit 111, and transmits all processing results to the active main function unit 111 and the spare main function unit 112. The communication lines between the active main function unit 111 and the standby main function unit 112 and the sub function units 12 to 15 are all connected by wired OR. Further, the slave function units 12 to 15 monitor the congestion by counting the number of received messages. Furthermore, when congestion occurs, the main function unit 111 is notified of the occurrence of congestion and the received message from the main function unit 111 is discarded in order to protect the device itself.
(Current main function)
As shown in FIG. 2, the active main function unit 111 includes an external processing request receiving unit 111a, an internal processing request transmitting unit 111b, an internal processing result receiving unit 111c, an external processing result transmitting unit 111d, a congestion processing unit 111e, and a temporary congestion monitoring. Unit 111f, temporary threshold setting unit 111g, message discrimination processing unit 111h, processing execution unit 111i, temporary congestion processing storage unit 111j, importance storage unit 111k, and the like.

  The external process request receiving unit 111a receives a process message from the outside. The internal processing request transmission unit 111b transmits processing messages to be requested to the slave function units 12 to 15 to the slave function units 12 to 15 among the processing messages received from the outside. The internal processing result receiving unit 111c receives processing results from the subordinate function units 12-15. The process execution unit 111i executes the content of the process message received from the outside. Moreover, the process execution part 111i performs the content of the process message received from the outside using the process result received from the sub function parts 12-15. The external processing result transmission unit 111d transmits the execution result for the processing message to the outside.

  The temporary congestion processing storage unit 111j is a storage unit that is used to monitor its own congestion state when the standby main function unit 112 fails. The temporary running time storage unit 112j and the load determination storage unit 112k in FIG. Store the same data as. The temporary threshold value is a threshold value when the standby main function unit 112 fails to perform monitoring processing by itself, but this temporary threshold value is compared with the threshold value in the threshold storage unit 112c of the standby main function unit 112. Set lower. This is to prevent the processing capacity of the CPU from exceeding the limit in an environment where normal information processing and congestion monitoring are simultaneously performed in the active main function unit 111.

  The importance level storage unit 111k stores an importance level table for determining the importance level for each message type.

  The temporary congestion monitoring unit 111f monitors its own congestion state when the standby main function unit 112 fails. The temporary congestion monitoring unit 111f estimates a processing time for a processing message requested from the outside, and determines whether the congestion state or the non-congested state based on the estimated processing time, the number of processing messages, and a threshold value stored in the temporary threshold storage unit 111g. judge. Specifically, a reception counter unit 112a, a message type determination unit 112b, a cycle timer unit 112c, a travel time calculation unit 112d, a cumulative counter unit 112e, a counter comparison unit 112f, and a load condition calculation unit 112g of the standby main function unit 112 described later. The load determination unit 112h has the same function. Note that a failure of the standby main function unit 112 can be recognized on the active main function unit 111 side, for example, when periodic communication with the backup main function unit 112 is interrupted.

  The congestion processing unit 111e suppresses message processing to prevent or eliminate congestion. Specifically, in order to temporarily reduce the number of processes, a new process message is not accepted from the outside, or the process speed is temporarily reduced. The suppression of the processing is performed when the result of the standby main function unit 112 being determined as being in a congestion state, when the temporary congestion monitoring unit 111f determines that the congestion state is present, and in the congestion state from the slave function units 12-15. Performed when notified that there is. Furthermore, the congestion processing unit 111e checks whether it has received a congestion state notification from the standby main function unit 112 at predetermined intervals. If the notification of the congestion state at a predetermined interval has not been received, it is determined that the standby main function unit 112 has stopped, and the temporary congestion monitoring unit 111f of itself is activated.

  The temporary threshold value setting unit 111g sets a temporary threshold value lower than the threshold value considering the congestion state based on the threshold value stored in the threshold value storage unit 112c in FIG. 3 and stores the temporary threshold value in the temporary threshold value storage unit 111h.

  When notified from the standby main function unit 112 that the message discrimination processing unit 111h is in a congested state, the message discrimination processing unit 111h indicates the type of processing message received from each of the subordinate function units 12 to 15 and its importance level in the importance level storage unit 111k. The message is determined based on the importance table, and each message is processed according to the importance. For example, for processing messages with low importance, processing (congestion avoidance processing) for discarding or abnormal termination is performed in order to reduce the load on execution processing in the active main function unit 111. As the low importance message, for example, there is a message belonging to a request for establishing a new communication link. For a processing message determined to be a highly important message, normal processing is performed even in a congested state. Note that high importance messages include, for example, messages belonging to a communication link release request or maintenance monitoring request during communication. Note that the congestion state may be subdivided into a plurality of states in accordance with the degree of congestion, and the message determination unit 111h may determine the message importance based on the importance table subdivided correspondingly. Thereby, more flexible congestion avoidance control becomes possible.

The above-described units are stored in a CPU (Central Processing Unit) or the like of a computer (not shown). Alternatively, it is stored in a ROM (Read Only Memory) or a RAM (Random Access Memory) (not shown), and these processes are executed by the CPU appropriately performing arithmetic processing.
(Reserve main function part)
As shown in FIG. 3, the standby main function unit 112 includes a reception counter unit 112a, a message type identification unit 112b, a cycle timer unit 112c, a travel time calculation unit 112d, a cumulative counter unit 112e, a counter comparison unit 112f, and a load amount calculation unit. 112g, a load determination unit 112h, a determination result notification unit 112i, an estimated travel time storage unit 112j, a load determination storage unit 112k, and the like.

  The estimated travel time storage unit 112j stores an estimated travel time table of the active main function unit 111 and a cumulative value of the estimated travel time. The estimated travel time table is a type of processing message received from the slave function units 12 to 15, an estimated value (hereinafter referred to as “CPU travel time”) of the time taken by the active main function unit 111 to execute this message type. Is a table in which is associated and listed. The cumulative value of the estimated travel time is a value obtained by adding all the travel times of the CPU estimated to be required for each process when a plurality of processing messages are processed.

  The load determination storage unit 112k stores a load threshold and a load determination table for determining the congestion state of the active main function unit 111.

  The reception counter unit 112a counts the number of processing messages received from each of the slave function units 12-15.

  The message type determination unit 112b and the message type of the processing message received from each of the sub function units 12 to 15 are determined. The message type determination unit 112b may have a function of selecting a message that is counted as a received message and a message that is not counted, among received processing messages, based on the determination result.

  The cycle timer unit 112c measures the processing cycle performed by the main function unit 11 on the processing message based on a preset measurement cycle.

  The total counter unit 112e counts the total number of processing messages executed by the active main function unit 111.

  The travel time calculation unit 112d acquires the CPU travel time corresponding to each message type with reference to the estimated travel time table in the estimated travel time storage unit 112j, and the acquired estimated travel time has been executed so far. The travel time cumulative value is updated by adding to the travel time cumulative value, which is the cumulative value of the travel time. Each message refers to a processing message executed within a preset measurement cycle.

  The counter comparison unit 112f increments the value of the total counter by “1” each time the acquired estimated travel time is added to the total travel time value, and compares the number of the total counter and the reception counter. Here, by determining whether or not the value of the reception counter is equal to the value of the cumulative counter, it is determined whether or not the travel time has been calculated for all processing messages counted by the reception counter and added to the cumulative counter. Do.

  The load amount calculation unit 112g calculates a load amount (hereinafter referred to as “temporary load amount”) estimated to be applied to the active main function unit 111 within the measurement cycle based on the updated travel time accumulated value. By estimating the temporary load amount, the standby main function unit 112 can quickly determine the message type of each received processing message, and can further calculate the estimated travel time quickly.

  The load determination unit 112h compares the calculated temporary load amount with the load threshold value in the load determination storage unit 112k, and determines the load level state in the active main function unit 111 in the load determination storage unit 112k in FIG. Judgment is made based on the judgment table. The load determination table can determine whether the currently calculated temporary load amount belongs to the low load, medium load, or high load of the load determination table based on the load status of the previous measurement cycle. Have been described. Here, the medium load, which is one of the determination results of the temporary load amount, is set between a low load and a high load, and indicates a load level in a non-congested state.

  For example, as shown in FIG. 4, when the active main function unit 111 is in a non-congested state in the current measurement cycle, it is a non-congested state if the load is low and a non-congested state in the subsequent measurement cycle. It is determined as a congestion state. In addition, when the active main function unit 111 has a high load, it is determined that the load is higher than a preset threshold, and it is determined as a congestion state. Determined.

  On the other hand, when the active main function unit 111 is in a congested state in the current measurement cycle, a load that is equal to or greater than a preset threshold is applied when the active main function unit 111 has a low load in the subsequent cycle. In the case, it is determined as a congestion state. If the load is less than the threshold value, it is determined as a non-congestion state. Further, when it is determined that the load is medium load or high load in the subsequent measurement cycle, it is determined that the congestion state occurs in any case.

  In addition, as described above, when the state transition between the non-congested state and the congested state is performed in continuous measurement cycles, in order to prevent the state transition from occurring excessively, a predetermined number of times are continuously performed at a predetermined time interval. When the congestion state is notified, it is preferable to determine that the congestion state is present. Moreover, it is preferable to determine that it is a non-congested state when a non-congested state is notified continuously for a predetermined number of times at a predetermined time interval. For this, a transition stage between low load, medium load, and high load, that is, a transition stage of congestion state is provided with several stages of load stages (hereinafter referred to as “protection stage number”), and a protection stage every time a state notification is received a predetermined number of times. Set to increase. This can reduce frequent updates of the congestion state and the non-congestion state.

  However, if the state transition from non-congestion to the congested state is delayed, the high load will continue for a long time and there is a risk of hindering the stable operation of the active main function unit 111. Therefore, the transition from non-congestion to the congested state The number of protection stages used for the may be set small.

  When the determination result notification unit 112i is determined to be in a congestion state or a non-congestion state, the determination result notification unit 112i notifies the active main function unit 111 of the determination result. This prompts the active main function unit 111 to quickly perform the congestion avoiding operation in response to the notification.

Note that the above-described units are stored in a CPU or the like of a computer (not shown). Alternatively, these processes are executed by being stored in a ROM or RAM (not shown) and appropriately performing arithmetic processing by the CPU. Here, it should be noted that the functional block diagrams shown in FIGS. 2 and 3 exist in the same CPU, ROM and RAM.
(Sub function part)
As shown in FIG. 5, the slave function unit includes a congestion monitoring unit 12a, a delay counter unit 12b, a message receiving unit 12c, a message processing unit 12d, a processing result transmission unit 12e, a congestion notification unit 12f, and a threshold storage unit 12g.

  The threshold storage unit 12g stores a threshold for the number of delayed tasks, which is a reference for determining a congestion state.

  The delay counter unit 12b counts the number of tasks for which processing is delayed by the delay counter. Specifically, the delay counter unit 12b counts the number of tasks other than the task D that has not been completed within one cycle of periodically generated high priority task processing (task D described later). It should be noted that what is subject to delay counting is a task that is not highly prioritized with respect to real-time performance but is susceptible to the operation of other tasks. One cycle refers to the period from the start of task D to be described later until the start of the next start.

  The congestion monitoring unit 12a monitors the delay state of its own processing. Specifically, the delay count is increased by the number of tasks being processed for each cycle. Further, by comparing the delay counter unit 12b with the threshold value, it is determined whether the device itself is in a congestion state or a non-congestion state. If it is determined that there is a congestion state, processing for avoiding the congestion is performed. Further, the congestion monitoring unit 12a urges the congestion notification unit 12f to notify the active main function unit 111 of the congestion state or the non-congestion state.

  The message receiving unit 12 c receives a processing message from the active main function unit 111.

  The message processing unit 12d executes a processing message and obtains a processing result.

  The processing result transmission unit 12e transmits the processing result to the active main function unit 111 and the spare main function unit 112.

  The congestion notification unit 12f notifies the main function unit 11 of the result of the congestion state monitored by the congestion monitoring unit 12a.

  Note that the above-described units are stored in a CPU or the like of a computer (not shown). Alternatively, these processes are executed by being stored in a ROM or RAM (not shown) and appropriately performing arithmetic processing by the CPU.

  Next, the mechanism of occurrence of processing delay in the slave function units 12 to 15 will be described. The sub-function units 12 to 15 process messages received from the active main function unit 111 by multitasking, but in multitasking processing, real-time performance is important in each one process (hereinafter referred to as “task”). Non-tasks are interrupted by higher priority tasks. For this reason, when the load increases, the real-time property is not high, but a delay occurs in a frequently used task.

  The operation state of the processor when there is a task priority with respect to real-time property and there are few messages received will be described with reference to FIG. In FIG. 6, the horizontal axis represents the type of task to be executed, the vertical axis represents time elapsed, and the processor operation graph at the right end represents the time during which the processor is operating for processing for each task. The priority of the task with respect to the real-time property is highest in the task D, and then in the order of the task C, the task B, and the task A. Since tasks B and C have different processing operations, the processor operating time also differs. On the other hand, the task D is a process that is triggered by a periodic timer within the slave function units 12 to 15 irrespective of message reception, and uses a processor in the same manner as other tasks. Message a executes tasks A and B, and message b executes tasks A and C. In the section (a) of FIG. 6, the task A is generated when the message a from the main function unit 11 is received, and after the processing of the task A is completed, the process proceeds to the task B for executing the predetermined function processing. After sending the result to the main function unit 11 as a message, the processing is completed. Similarly, in section (b) of FIG. 6, task A is executed after message b is received, and the processing result is sent to main function unit 11 after completion of task C execution. The section (c) in FIG. 6 performs the processing of the message a in the same manner as the section (a) in FIG. In both figures, tasks A and B are processed. In section (a), task D occurs soon after task B is completed, whereas in section (c), task D occurs after task B is completed. doing. In the example of FIG. 6, as is apparent from the rightmost processor operation graph, each task is processed smoothly regardless of the priority of the task.

Next, the operation state of the processor when there is a task priority with respect to real-time property and the number of received messages is large and the processes compete with each other will be described with reference to FIG. The task priority is the same as in FIG. In the section (a) in FIG. 7, the tasks A and B are finished after the message a1 is received. Since another message b1 is received during the processing of task B of message a1, task A of message b1 is executed immediately after the processing of message a1 is completed. However, since the task D having a higher priority is activated during the processing of the task A, the task A being processed enters a standby state, and after waiting for the task D to complete, the task A returns from the standby state to the execution state to complete the processing. Following this, task C is executed, and the processing for message b1 is completed. However, at this time, the entire processing time of the message b1 is delayed. In the section (b) of FIG. 7, the message b2 is received during the processing of the task B of the message a2. In this case, the task A of the message b2 is in a standby state until the task B of the message a2 is completed, and the processing completion of the message b2 is delayed by the waiting time. Further, in this state, when the messages b3 and a3 are continuously received in the section (c) of FIG. 7, the task D is activated during the processing of the task C of the message b2, and the processing of the task C becomes a standby state. The processing completion time is delayed. Immediately after the completion of task C of message b2, task A and task C of message b3 are completed, and then task A and task C of message a3 are completed. As a result, the processing completion time of message a3 is significantly larger than planned. Be late to. In this way, in the example of FIG. 7, due to the priority of the task, a delay occurs in the completion of other task processing.
(Operation of congestion control system)
Next, the congestion state monitoring operation in the standby main function unit 112 of the congestion control system 10, the congestion avoiding operation in the active main function unit 111, and the congestion avoiding operation in the sub function units 12 to 15 will be described. First, the congestion monitoring operation in the standby main function unit 112 will be described with reference to the flowchart of FIG.
(A) First, in step S101, the standby main function unit 112 in FIG. 3 activates the cycle timer unit 112c so that the CPU running time can be measured at a preset measurement cycle. In step S102, the estimated traveling time accumulated value stored in the reception counter unit 112a, the cumulative counter unit 112e, and the estimated traveling time storage unit 112j is initialized.
(B) In step S103, the message type determination unit 112b registers the number of processing messages to be received in the reception counter unit 112a and receives each processing message from the outside or the subordinate function units 12 to 15 to identify the message type. To do.
(C) In step S104, the traveling time calculation unit 112d acquires the estimated traveling time value of the processing message in which the message type is specified based on the estimated traveling time table stored in the estimated traveling time storage unit 112j. In step S105, the traveling time calculation unit 112d adds the acquired estimated traveling time value to the estimated traveling time accumulated value held in the cumulative counter unit 112e. In step S106, the traveling time calculation unit 112d increases the value of the cumulative counter by “1”.
(D) In step S107, the counter comparison unit 112f determines whether or not the values of the reception counter and the cumulative counter are equal. If they are equal, it is determined that the estimated travel time values of all the received processing messages have been set. Then, the process proceeds to step S108. If they are not equal, the process returns to step S103, and the setting of the estimated travel time value of the other received processing message is continued.
03).
(E) In step S108, the load calculation unit 112g calculates a temporary load amount based on the estimated running time cumulative value. The load determination unit 112h determines whether the temporary load amount corresponds to a low load, a medium load, or a high load based on the load threshold value in the load determination storage unit 112k. In step S109, the load determination unit 112h sets the active main function unit 111 to the congestion state based on the determination result and the previous state of the active main function unit 111 described in the load determination table in the load determination storage unit 112k. It is determined whether there is a non-congested state. In step S110, when the state of the active main function unit 111 changes from a non-congested state to a congested state, or when it is determined that it is a congested state regardless of the previous state, the determination result notifying unit 112i displays the active main function unit 111. A congestion state notification indicating that the state is a congestion state is transmitted to the server. On the other hand, when the state of the active main function unit 111 changes from a congested state to a non-congested state, or when it is determined as a non-congested state regardless of the previous state, the determination result notifying unit 112i A congestion state notification indicating that the state is a non-congested state is transmitted to the terminal.

  With the operation described above, the congestion avoiding operation of the active main function unit 111 can be promoted. Here, if the transition from the non-congested state to the congested state is delayed, the high load will continue for a long time, and the stable operation of the main function unit 111 may be hindered. Therefore, the transition from the non-congested state to the congested state It is desirable to keep the load threshold value used for.

Next, referring to the flowchart of FIG. 9 for the congestion avoiding operation in the active main function unit 111 when the active main function unit 111 is notified of the congestion or non-congestion state from the standby main function unit 112 To explain.
(A) First, in step S200, the congestion processing unit 111e of the active main function unit 111 in FIG. 2 determines whether a congestion state notification is received from the standby main function unit 112 at predetermined intervals. When the congestion state notification is received, the process proceeds to step S201, and when the congestion state notification is not received, the process proceeds to step S206. In step S201, the process execution unit 111i receives a process message from the outside or the subordinate function units 12 to 15, and starts processing the received process message. In step S202, the congestion processing unit 111e determines whether the active main function unit 111 is in a congestion state or a non-congestion state based on the congestion state notification received from the standby main function unit 112. If it is congested, the process proceeds to step S203, and if not congested, normal processing is performed in step S205.
(B) In step S203, the message determination unit 111h determines the level of importance of the received processing message according to the importance of the message type set in the importance storage unit 111k in advance. If the degree of importance is low as a result of the determination, the process proceeds to step S204. If the degree of importance is low, normal processing is performed in step S205.
(C) In step S204, the congestion processing unit 111e executes a discard process or an abnormal termination process (congestion avoidance process) for a process message that is determined to be low in importance. For example, a request for establishing a new communication link is discarded.
(D) In step S206, it is determined that the standby main function unit 112 has failed, and the temporary congestion monitoring unit 111f of the active main function unit 111 performs the processing from steps S101 to S109. At this time, in step S108, the temporary threshold value setting unit 111g sets a temporary threshold value lower than this threshold value considering the congestion state based on the threshold value stored in the load determination storage unit 112k in FIG. The load determination unit determines that the active main function unit 111 is in a congested state based on the determination result based on the temporary threshold and the previous state of the active main function unit 111 described in the load determination table in the load determination storage unit 112k. Or whether it is in a non-congested state.

Next, the congestion avoiding operation in the slave function units 12 to 15 will be described with reference to the flowchart of FIG.
(A) In step S301, the congestion monitoring unit 12a in FIG. 5 initializes the delay counter unit 12b to 0. In step S302, when a processing message is received from the active main function unit 111, the message processing unit 12d performs task processing for one cycle. In step S <b> 303, the congestion monitoring unit 12 a determines whether there is an incomplete task that has not been processed before the current task D starts from the previous task D. Specifically, the task D is periodically started to count the number of incomplete tasks, and the congestion monitoring unit 12a receives the count result and determines whether there are any incomplete tasks.
(B) If there is an incomplete task, in step S304, the congestion monitoring unit 12a determines whether this task is a delay target task, that is, a task that has a low priority for real-time performance but is easily affected by other task operations. If the target task is determined, the delay counter unit 12b is incremented by “+1” by the number of target tasks in step S305. If it is not the target task, the process returns to step S302. In step S306, the congestion monitoring unit 12a compares the value of the delay counter unit 12b with a threshold value stored in the determination threshold value storage unit 112e. If the value of the delay counter unit 12b is greater than the threshold value as a result of the comparison, the congestion monitoring unit 12a determines that it is in a congested state in step S307 and changes the state from a non-congested state to a congested state. In the case of a congestion state, the congestion notification unit 12f notifies the main function unit 11 of the congestion state in order to reduce the load, and suppresses new process reception from the active main function unit 111. Further, when a new process is accepted, it is discarded. By these processes, it is possible to easily avoid or eliminate the congestion of the slave function units 12 to 15. If the value of the delay counter unit 12b is lower than the threshold value as a result of the comparison in step S306, the congestion monitoring unit 12a determines that the congestion state is not present, and returns to step S302 as it is.
(C) If all the task processes are completed within one period in step S303, the congestion monitoring unit 12a determines in step S308 whether these tasks are target tasks. If it is a target task, in step S309, the delay counter unit 12b decreases the delay count value by “1” by the number of target tasks. If it is not the target task, the process directly returns to step S302. In step S310, the congestion monitoring unit 12a compares the value of the delay counter with the threshold value of the determination threshold value storage unit 112e. If it is smaller than the threshold value, in step 311, the own state is changed from the congested state to the non-congested state. In the non-congested state, the congestion notifying unit 12f notifies the main function unit 11 of the non-congested state and resumes accepting a new process from the active main function unit 111. If the result of comparison in step S310 is greater than the threshold, the congestion monitoring unit 12a determines that the congestion state is still present, and returns to step 302 as it is.

  Since the task operating time varies depending on the type and load status, it is preferable to perform protection stage number determination processing that suppresses erroneous determination due to temporary variation.

  As described above, according to the first embodiment of the present invention, all the processing messages addressed to the active main function unit 111 are also transmitted to the standby main function unit 112, so that the backup main function unit 112 has the current main function. The congestion state of the unit 111 can be estimated. For this reason, it is possible to monitor the processing state of the active function unit 111 without degrading the processing capability of the active function unit 111 that executes an execution process on the processing message.

  Further, by estimating the congestion state of the active function unit 111, the congestion state can be notified more quickly. For this reason, the active function unit 111 can quickly perform the process of reducing the congestion state under heavy load, so that the active main function unit 111 can maintain a stable operation.

  In the slave function units 12 to 15, it is possible to prevent a congestion state by performing its own congestion monitoring. In addition, since only tasks that cause processing delay are counted, it is possible to easily and accurately determine processing delay. In addition, the secondary function units 12 to 15 perform congestion monitoring in its own apparatus, and perform congestion recovery processing when the state falls into a congestion state, thereby further improving the stability of the system.

  In addition, the congestion avoiding process of the active main function unit 111, the congestion monitoring process of the standby main function unit 112, and the congestion monitoring process of the slave function units 12 to 15 are realized by software control, thereby installing the information processing apparatus. Cost can be reduced.

Note that the congestion control method in the first embodiment captures each operation of the congestion control system 10 as a method invention. The congestion control program of the first embodiment is a program for causing a computer to function each unit included in the congestion control system 10.
<Second Embodiment>
(Congestion control system)
In the first embodiment, the information processing execution unit and the congestion monitoring unit of the main function unit 11 have been described for the current standby congestion control system 10, but the information processing execution unit and the congestion monitoring unit may exist independently. Absent. Therefore, in the second embodiment, the information processing execution unit and the congestion monitoring unit of the main function unit exist in different information processing apparatuses and are connected to be communicable with each other with reference to FIG. explain.

  The congestion control system 20 includes information processing devices 1, 2, and 3 as main function units 21, the information processing device 1 includes an interface function unit 17 as an information processing execution unit, and the information processing devices 2 and 3 as a congestion monitoring unit. Interface function units 27 and 37. The interface function unit 17 and the interface function unit 27, and the interface function unit 17 and the interface function unit 37 are connected via a communication line. When the interface function unit 27 breaks down, the interface function unit 17 shifts at least a part of the congestion monitoring process to the processing message sent from the outside to another information processing apparatus 3. In this case, the interface function unit 37 of the information processing apparatus 3 functions as a new congestion monitoring unit.

  Although not shown in FIG. 11, a plurality of secondary function units are connected to the information processing apparatuses 1 to 3 in a communicable manner as in FIG. 1. The sub functional unit has the same configuration as that of the first embodiment. Description is omitted.

  As shown in FIG. 12, the interface function unit 17 includes a processing time storage unit 17a, a processing message reception unit 17b, a processing message execution processing unit 17c, a message determination unit 17d, a congestion state confirmation unit 17e, a message processing execution unit 17f, and the like. Prepare.

  The processing time storage unit 17a stores the CPU processing time required for receiving and transmitting a single received message in the interface function unit 17.

  The processing message receiving unit 17b receives a processing message sent from the outside or a sub function unit.

  The processing message execution processing unit 17c executes processing of the content of the processing message.

  The message determination unit 17d determines the type of the received processing message and its importance. The processing method is determined according to the determined importance of the processing message.

  The congestion state confirmation unit 17 e confirms the congestion state of the interface function unit 17 based on the congestion state notification received from the interface function unit 27. Furthermore, when the congestion state confirmation unit 17e does not receive the congestion state notification from the interface function unit 27 for a certain period, the congestion state confirmation unit 17e determines that the interface function unit 27 has failed and requests the spare interface function unit 37 to monitor the congestion state. .

  When notified that the message processing execution unit 17f is in a congested state, the message processing execution unit 17f performs different processing depending on the importance. For example, a processing message with low importance is discarded or abnormally terminated (congestion avoidance processing). This reduces the load on the execution process of the interface function unit 17 in the congestion state. As the low importance message, for example, a message belonging to a request for establishing a new communication link is shown. For messages with high importance, normal processing is performed. A message having a high degree of importance indicates, for example, a message belonging to a request to release a communication link during communication or a maintenance monitoring request.

  The notified congestion state may be further subdivided into a plurality of states according to the degree of congestion, and the message importance may be subdivided and set accordingly. Thereby, more flexible congestion avoidance control can be performed.

  As shown in FIG. 13, the interface function unit 27 includes a message transmission / reception counter unit 27a, a transmission / reception processing estimated time storage unit 27b, a cycle timer unit 27c, a travel time value calculation unit 27d, a load amount calculation unit 27e, a congestion state determination unit 27f, and A congestion state notification unit 27g and the like are provided.

  The message transmission / reception counter unit 27a counts the number of processes each time a processing message transmission or reception event executed between the interface function unit 17 and the interface function unit 27 or 37 is processed.

  The transmission / reception processing estimated time storage unit 27b stores a transmission / reception processing estimated time table in which processing times for transmission / reception processing for each processing message in the interface function unit 17 are set in advance. Based on the transmission / reception processing estimated time table, the interface function unit 27 can quickly read the processing time of the transmission / reception event in the interface function unit 17.

  The cycle timer unit 27c measures a preset measurement cycle.

  The travel time value calculation unit 27d multiplies the number of transmission events or reception events of processing messages executed within a preset measurement cycle by the processing time read from the transmission / reception processing estimated time table, An estimated CPU running time value in the CPU of the interface function unit 17 is calculated.

  The load amount calculation unit 27e calculates a load amount estimated to be applied to the interface function unit 17 within a cycle (hereinafter referred to as “temporary load amount”) based on the calculated estimated CPU travel time value.

  The congestion state determination unit 27f determines whether the congestion state of the interface function unit 27 or 37 is low load, medium load, or high load based on the calculated temporary load amount. Here, the medium load that is one of the determination results of the temporary load amount is set between a low load and a high load, and indicates a load level in a non-congested state. As in the first embodiment, the determination of whether the interface function unit 17 is in a congested state or a non-congested state is performed based on the load status of the previous measurement cycle, as shown in FIG. Here, when the interface function unit 17 is in a non-congested state in the previous measurement cycle, it is determined that the load is low if the load is low in the subsequent measurement cycle, and the non-congested state if the load is medium. Further, when the interface function unit 17 is heavily loaded in the previous measurement cycle, it is determined that the load is higher than a preset threshold value, the congestion state is determined. Determined.

  On the other hand, if the interface function unit 17 is in a congested state in the previous measurement cycle, if the interface function unit 17 has a low load in the subsequent cycle, or if a load greater than a preset threshold is applied, If the load is less than the threshold value, it is determined as a non-congested state. Further, when it is determined that the load is medium load or high load in the subsequent measurement cycle, it is determined that the congestion state occurs in any case.

  As described above, when a state transition from a non-congested state to a congested state is performed in successive measurement cycles, the congestion state and the non-congested state are frequently updated by providing the number of protection stages for the state transition. Can be reduced. However, if the state transition from the non-congestion state to the congested state is delayed, the high load continues for a long time, and there is a possibility that the stable operation of the interface function unit 17 may be hindered. The number of protection stages used may be set small.

  When the interface function unit 17 is determined to be in a congestion state, or when it is determined that the state of the interface function unit 17 has transitioned from a non-congestion state to a congestion state, the congestion state notification unit 27g The unit 17 is notified that it is in a congested state. This prompts the interface function unit 17 to execute a congestion avoiding operation in response to the notification.

The interface function unit 37 has the same configuration as the interface function unit 27. Description is omitted.
(Operation of congestion control system)
Next, the operation of the congestion control system 20 will be described. The operation of the congestion control system 20 includes a congestion monitoring operation of the interface function unit 27 of the information processing device 2 and a congestion avoiding operation of the interface function unit 17 of the information processing device 1. The congestion avoiding operation of the slave function unit is the same as that of the first embodiment, and thus the description thereof is omitted.

First, the congestion monitoring operation of the interface function unit 27 of the information processing apparatus 2 will be described with reference to the flowchart of FIG.
(A) First, in step S401, the interface function unit 27 activates the cycle timer unit 27c and measures a preset measurement cycle. In step S402, the travel time calculation unit 27d reads processing time information indicating the transmission / reception event processing time of the interface function unit 17 from the transmission / reception processing estimated time table of the transmission / reception processing estimated time storage unit 27b. In step S403, the running time value calculation unit 27d multiplies the read processing time information by the number of processing message transmission and reception events stored in advance in the message transmission / reception counter unit 27a, thereby calculating an estimated CPU running time value. Is calculated.
(B) In step S404, the load amount calculation unit 27e calculates a temporary load amount from the estimated CPU travel time value, and whether the calculated temporary load amount corresponds to a low load, a medium load, or a high load. Determine.
(C) In step S405, the congestion state determination unit 27f determines whether the interface function unit 17 is in a congestion state or a non-congestion state based on the determination result. In step S406, when the congestion state notification unit 27g determines that the state of the interface function unit 17 has changed from a non-congestion state to a congestion state, or is determined to be a congestion state (high load state) regardless of the previous state, An opposite node congestion state notification indicating the congestion state is transmitted to the interface function unit 17. Further, when the interface function unit 17 is in a non-congested state, it is notified that it is in a non-congested state.

  By this notification, the congestion avoiding operation of the interface function unit 17 can be promoted. Here, when the transition from the non-congestion state to the congestion state is delayed, the high load continues for a long time, and there is a possibility that the stable operation of the interface function unit 17 may be disturbed. Therefore, the transition from the non-congestion state to the congestion state It is desirable to keep the number of protection stages used for the above.

  Here, the case where the interface function unit 27 performs the congestion monitoring process on the interface function unit 17 has been described. However, the interface function unit 37 can also perform the congestion monitoring process.

Next, the congestion avoiding operation of the interface function unit 17 when it is notified from the interface function unit 27 to the interface function unit 17 will be described based on the flowchart of FIG.
(A) In step S500, the congestion state monitoring unit 17e of the interface function unit 17 in FIG. 12 determines whether to receive a processing message or a congestion state notification from the interface function unit 27 within a predetermined period. When a processing message or a congestion state notification is received within a predetermined period, processing is started in step S501. If no processing message or congestion state notification is received within the predetermined period, the process proceeds to step S506.
(B) In step S502, the congestion state confirmation unit 17e determines whether the interface function unit 17 itself is in a congestion state or a non-congestion state based on the received congestion state notification. If it is determined as a non-congested state, the message processing execution unit 17f performs normal processing in step S505. If it is determined that there is a congestion state, the process proceeds to step S503.
(C) In step S503, the message determination unit 17d determines the level of importance of the received processing message in accordance with the message type described in the importance table in the preset importance storage unit 17g. If it is determined that the importance level is high, the message processing execution unit 17f performs normal processing in step S505. If it is determined that the importance level is low, the process proceeds to step S504.
(D) In step S504, the message process execution unit 17f discards or processes the process message determined to be less important as an abnormal end. Further, reception of processing messages is suppressed.
(E) If no processing message is received from the interface function unit 27 for a predetermined period, the congestion state confirmation unit 17e determines in step S506 that the interface function unit 27 has failed, and the spare interface function unit 37 monitors the congestion state. Request. The interface function unit 37 that has received the request performs the congestion monitoring process in steps S401 to S406 instead.

  As described above, in the second embodiment, the interface function unit 17 causes the interface function unit 27 or 37 to monitor the congestion state. As a result, the congestion state can be monitored without degrading the execution processing capability in the interface function unit 17.

  Further, by estimating the congestion state of the interface function unit 17, the congestion state can be notified more quickly. For this reason, since the interface function part 17 can perform the process which reduces the congestion state to which high load was applied rapidly, it can maintain stable operation | movement.

  According to the second embodiment, in addition to the effects of the first embodiment, when the interface function unit 27 as the congestion monitoring unit fails, a part of the processing for the processing message is transferred to the information processing device 3. Thus, the processing load of the interface function unit 17 as the information processing execution unit can be reduced.

  Further, by realizing the congestion avoiding process of the interface function unit 17 and the congestion monitoring process of the interface function unit 27 or 37 by software control, the cost for installing the information processing apparatus can be reduced.

  The congestion control method according to the second embodiment captures each operation of the congestion control system 20 as a method invention. The congestion control program of the second embodiment is for causing a computer to function each means included in the congestion control system 20.

A part or all of the above-described embodiment can be described as in the following supplementary notes, but is not limited thereto.
[Appendix 1]
An information processing execution unit that performs processing on a processing message sent from the outside, a congestion monitoring unit that is attached to the information processing execution unit and monitors the number of processing messages, the information processing execution unit, and the congestion monitoring unit A congestion control system comprising a connected subordinate function unit for executing a task included in the processing message,
The congestion monitoring unit
Means for estimating a processing time for the processing message and determining whether the processing message is in a congested state or a non-congested state based on the processing time, the number of processing messages and a threshold
Means for notifying the information processing execution unit of the determined content, and the slave function unit includes:
Means for counting the number of unfinished processes within a period of at least one task included in the process message, and determining whether the state is a congestion state or a non-congestion state based on the number of unfinished tasks and a threshold;
Means for notifying the information processing execution unit of the determined content, the information processing execution unit,
Means for estimating the processing time for the processing message by itself when the congestion monitoring unit fails, and determining whether the processing state is the congestion state or the non-congestion state based on the processing time, the number of processing messages and a threshold;
A congestion control system comprising: means for suppressing processing for the processing message when content notified from the congestion monitoring unit and the subordinate function unit or content determined by itself indicates a congestion state.
[Appendix 2]
The congestion control system according to appendix 1, wherein the information processing execution unit further includes means for resetting the threshold value when the congestion monitoring unit fails.
[Appendix 3]
The information processing execution unit and the congestion monitoring unit are separately provided in different information processing apparatuses that are connected via a communication line and can communicate with each other.
The congestion control system according to appendix 1 or 2.
[Appendix 4]
The information processing execution unit performs processing for the processing message sent from the outside to the other information processing apparatuses connected via the communication line and capable of communicating with each other when the congestion monitoring unit fails. The congestion control system according to supplementary note 3, further comprising means for shifting a section.
[Appendix 5]
The information processing execution unit and the congestion monitoring unit are functional units that operate by program control in the same central processing unit.
The congestion control system according to appendix 1 or 2.
[Appendix 6]
The means for determining the slave function unit counts tasks that are not highly prioritized for real-time performance among the uncompleted tasks, but are susceptible to the operation of other tasks. The congestion control system described in 1.
[Appendix 7]
At least one of the determination unit of the congestion monitoring unit and the determination unit of the information processing execution unit changes the state when the congestion state or the non-congestion state is notified more than a predetermined number of times in a predetermined interval. The congestion control system according to appendix 1.
[Appendix 8]
An information processing execution unit that performs processing on a processing message sent from the outside, a congestion monitoring unit that is attached to the information processing execution unit and monitors the number of processing messages, the information processing execution unit, and the congestion monitoring unit A congestion control method used in a connected system including a slave function unit that executes a task included in the processing message,
The congestion monitoring unit
Estimating a processing time for the processing message, and determining, based on the processing time, the number of processing messages and a threshold value, whether it is a congestion state or a non-congestion state;
The step of notifying the information processing execution unit of the determined content, the slave function unit,
Counting the number of incomplete processing within a period of at least one task included in the processing message, and determining whether the state is a congested state or a non-congested state based on the incomplete number and a threshold;
A step of notifying the information processing execution unit of the determined content, the information processing execution unit,
When the congestion monitoring unit fails, the process itself estimates the processing time for the processing message, and based on this processing time, the number of processing messages and a threshold, determines whether it is a congestion state or a non-congestion state;
And a step of suppressing processing for the processing message when the content notified from the congestion monitoring unit and the subordinate function unit or the content determined by itself indicates a congestion state.
[Appendix 9]
9. The congestion control method according to appendix 8, wherein the information processing execution unit further includes a step of resetting the threshold value when the congestion monitoring unit fails.
[Appendix 10]
Supplementary note 8 is characterized in that the step of determining by the subordinate function unit counts tasks that are not highly prioritized for real-time performance among the uncompleted tasks, but are susceptible to the operation of other tasks. Or the congestion control method according to 9.
[Appendix 11]
The congestion control method according to appendix 10, wherein the information processing execution unit and the congestion monitoring unit are separately provided in different information processing apparatuses that are connected via a communication line and can communicate with each other.
[Appendix 12]
The information processing execution unit performs processing for the processing message sent from the outside to the other information processing apparatuses connected via the communication line and capable of communicating with each other when the congestion monitoring unit fails. The congestion control method according to appendix 8 or 9, further comprising a step of migrating parts.
[Appendix 13]
The information processing execution unit and the congestion monitoring unit are functional units that operate by program control in the same central processing unit.
The congestion control method according to appendix 8 or 9.
[Appendix 14]
At least one of the determination step of the congestion monitoring unit and the determination step of the information processing execution unit changes the state when the congestion state or the non-congestion state is notified a predetermined number of times or more in a predetermined interval. The congestion control method according to appendix 8.

1, 2, 3 Information processing device 10, 20 Congestion control system 11 Main function unit 111 Current main function unit 112 Spare main function unit 12, 13, 14, 15 Sub function unit 17, 27, 37 Interface function unit

Claims (10)

An information processing execution unit that receives processing messages sent from the outside, a congestion monitoring unit that is provided alongside the information processing execution unit and monitors the number of processing messages, is connected to the information processing execution unit and the congestion monitoring unit , A congestion control system including a slave function unit that receives a process message for a specific process from the information processing execution unit and executes a task included in the process message for the specific process among the process messages. ,
The congestion monitoring unit
Determination processing means for determining the type and importance of the processing message received from the slave function unit based on the importance table, and processing each processing message according to the determined importance The slave function unit is
The number of incomplete processing within a period of at least one task included in the processing message is counted, and the number of incomplete processing is compared with a threshold value that is used as a criterion for determining a congestion state. Means for determining whether it is in a state;
Means for notifying the information processing execution unit of the determined content, the information processing execution unit,
When the congestion monitoring unit fails, it counts the number of unfinished processes within a period of at least one task included in the processing message, and becomes a reference for judging the number of unfinished and the congestion state. Means for comparing the threshold value to determine whether the congestion state or the non-congestion state;
A congestion control system comprising: means for suppressing processing for the processing message when content notified from the congestion monitoring unit and the subordinate function unit or content determined by itself indicates a congestion state. Congestion control system according to claim 1 wherein the information processing execution unit, characterized by further comprising means for setting a temporary threshold value lower than the threshold when the congestion monitor fails. The information processing execution unit and the congestion monitoring unit are separately provided in different information processing apparatuses that are connected via a communication line and can communicate with each other.
The congestion control system according to claim 1 or 2. The information processing execution unit performs processing for the processing message sent from the outside to the other information processing apparatuses connected via the communication line and capable of communicating with each other when the congestion monitoring unit fails. The congestion control system according to claim 3, further comprising means for shifting the unit. The information processing execution unit and the congestion monitoring unit are functional units that operate by program control in the same central processing unit.
The congestion control system according to claim 1 or 2. An information processing execution unit that receives processing messages sent from the outside, a congestion monitoring unit that is provided alongside the information processing execution unit and monitors the number of processing messages, is connected to the information processing execution unit and the congestion monitoring unit , among the processing messages, is the processing messages for a particular processing received from the information processing execution unit, the congestion control for use in systems with a slave function unit to perform a task included in the processing messages for the specific treatment A method,
The congestion monitoring unit,
Determining the type and importance of the processing message received from the slave function unit based on the importance table, and processing each processing message according to the determined importance ;
The slave function unit is
The number of incomplete processing within a period of at least one task included in the processing message is counted, and the number of incomplete processing is compared with a threshold value that is used as a criterion for determining a congestion state. Determining whether it is in a state;
Notifying the information processing execution unit of the determined content,
The information processing execution unit
When the congestion monitoring unit fails, it counts the number of unfinished processes within a period of at least one task included in the processing message, and becomes a reference for judging the number of unfinished and the congestion state. Comparing the threshold to determine whether it is congested or non-congested;
And a step of suppressing processing for the processing message when the content notified from the congestion monitoring unit and the subordinate function unit or the content determined by itself indicates a congestion state. Wherein the information processing execution unit, a congestion control method according to claim 6, further comprising the step of setting a temporary threshold lower than the threshold value when the congestion monitor fails.
The step of determining by the subordinate function unit counts, among the unfinished tasks, a task that is not high in priority for real-time property but is easily affected by operation of other tasks. 8. The congestion control method according to 6 or 7. The congestion control method according to claim 8, wherein the information processing execution unit and the congestion monitoring unit are separately provided in different information processing apparatuses that are connected via a communication line and can communicate with each other. The information processing executing unit, when the congestion monitor fails, the information processing apparatus is connected to the other that can communicate with each other via a communication line, one process for the processing messages sent from outside The congestion control method according to claim 6, further comprising a step of transferring a part.

Images