JP2006338264A - Task assignment device and task assignment method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technology for assigning a task by considering time to be spent until a processing result is acquired, in a distributed processing system. <P>SOLUTION: In the distributed processing system in which a plurality of information processors are connected through a communication network, a task assignment device for assigning a task to the information processor is provided with: a computing resource storage means for storing the computing power, communication overhead time and data transfer speed of the plurality of information processors; a task information acquisition part for acquiring a computing amount necessary for execution of a requested task and a transfer data amount generated when the task is assigned to the information processor; and a task assignment destination deciding means for deciding the information processor to which the task should be assigned based on information about an information processor stored in the computing resource storage means, and the computing amount for the task and the data transfer amount acquired by the task information acquisition means. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a technique for assigning tasks in a distributed processing system connected by a communication network.

In a distributed processing system in which a plurality of computers are connected by a network, a process to be executed is divided into a plurality of parts (tasks), and these computers perform processing in parallel. At this time, by appropriately assigning the tasks to the respective computers, it is possible to efficiently use the computing resources, and therefore, it is possible to distribute the load of the entire system and increase the processing speed. As described in Patent Document 1, this is realized by grasping the load status of each computer and causing a computer having the lowest load to execute a task.
Japanese Patent Laid-Open No. 11-3323 JP 2003-298599 A

  However, in the case of the prior art as described above, since the task assignment destination is determined based only on the load status (CPU processing capacity, free memory capacity, etc.) of each computer, the computer that completes the process most quickly. There was a problem that tasks could not be assigned to.

  That is, each computer is connected via a network, and in order to execute a task, it is necessary to transfer data, a program, etc. through this network as needed. Therefore, for example, when a large amount of data has to be communicated, if a task is assigned to a computer connected to a communication path having a high processor processing speed but a low data transfer speed, it takes time to obtain a processing result. There was a problem that it took.

  As described above, in the conventional technique, tasks are allocated so that the processing time in the allocation destination computer is shortened, and the time until the processing result can be acquired from the allocation is not considered.

  This is particularly a problem when the distributed processing system is used for real-time applications where the demand for processing time guarantee is severe.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a technique for assigning tasks in consideration of the time until a processing result is acquired in a distributed processing system.

  In order to achieve the above object, the present invention has the following configuration.

  A task assignment device according to the present invention is a task assignment device that assigns a task requested to be executed to an information processing device in a distributed processing system in which a plurality of information processing devices are connected via a communication network. The task assignment device receives a task execution request, determines an information processing device to execute the task, and assigns the task. The task assignment device transfers the task to the information processing device to which the task is assigned, and receives a processing result from the information processing device. Then, after obtaining the processing result, the processing result is transmitted to the request source.

  The task allocation device according to the present invention includes a computing resource storage unit, a task information acquisition unit, and a task allocation destination determination unit.

  The computing resource storage means stores computing capacity, communication overhead time, and data transfer rate of a plurality of information processing devices (computing resources) in the system.

  The computing power of the information processing device is, for example, information indicating the amount of calculation that can be performed by the CPU (central processing unit) per unit time, the CPU usage rate, the usable memory capacity, and other information that affects computation processing. Or a combination of these can be used. Further, as a calculation amount that can be performed by the CPU per unit time, a numerical value such as MIPS (Million Instructions Per Second), a type of CPU, an operating frequency thereof, and the like can be used.

  The communication overhead time is the time taken from the execution of a command to start communication to the actual start of communication, and is irrelevant to the amount of transfer data to be transmitted.

  The data transfer rate is the amount of data that can be transferred per unit time.

  The task information acquisition means acquires a calculation amount necessary for executing the requested task and a transfer data amount generated when the task is assigned to the information processing apparatus. The amount of computation required for executing a task is the amount of computation performed by the information processing apparatus to which the task is assigned. The amount of transfer data generated when a task is assigned to an information processing device is the amount of data or program data transmitted to the information processing device when a task is assigned to the information processing device. Further, it may include a data amount of a processing result transmitted from the information processing apparatus to the task assignment apparatus after completion of the task execution.

  The task calculation amount and the transfer data amount may be notified at the same time when the task allocation device is requested to execute the task, or may be recorded in the task itself.

  The task allocation destination determining means, when requested to execute the task, includes information related to the information processing device stored in the computing resource storage means, and the task computation amount and transfer data amount obtained by the task information obtaining means. Based on this, an information processing apparatus to which a task is assigned is determined. At this time, the task may be assigned to the information processing apparatus that can obtain the processing result earliest, and if the time limit (deadline) until the task execution is completed is determined, the processing should be completed within the time. A task may be assigned to an appropriate information processing apparatus capable of processing.

  As described above, the task assignment device according to the present invention stores the calculation capability and information related to communication of the communication overhead time and the data transfer rate for each information processing device, and the task calculation amount and the transfer data amount. In order to determine the information processing device to which tasks are allocated based on the task, the task allocation destination is determined in consideration of the time from task allocation including not only the time required for computation but also communication time to acquisition of processing results. Is possible.

  Estimate the time required for communication processing by dividing it into communication overhead time that is constant for each information processing device regardless of the amount of transfer data and time required for actual data transfer (transfer data amount and data transfer speed). Therefore, it is possible to make a more accurate evaluation.

Further, the task assignment destination determining means in the present invention uses the computing capacity, communication overhead time and data transfer speed of the information processing apparatus stored in the computing resource storage means, and the task computation amount and the transfer data amount as variables. The task processing capability of each information processing device may be evaluated using an evaluation function, and an information processing device to which a task is assigned may be determined based on the task processing capability.

  Task processing capability is processing capability that includes not only the processing capability of the information processing device but also the capability of communication processing with the information processing device. From task assignment to obtaining processing results It expresses how quickly all the processes during the period can be performed.

  By using the evaluation function in this way, it becomes easy to determine the task assignment destination.

  The task assignment destination determining means in the present invention estimates the calculation time from the calculation capability and the calculation amount, estimates the communication time from the communication overhead time, the data transfer rate and the transfer data amount, and calculates the sum of the calculation time and the communication time. It is also preferable to use an evaluation function for task processing capability.

  The calculation time can be estimated from the calculation capability and the calculation amount. Further, the time required for data transfer can be estimated from the transfer data amount and the data transfer rate, and the communication time can be estimated by adding the communication overhead time.

  An evaluation function that uses the sum of these as the task processing capability of the information processing apparatus is used. That is, the task processing capability is higher as the sum is smaller.

  As described above, since the information processing device to which the task is assigned is determined by predicting the time from the task assignment to the processing result being obtained, the task can be assigned to the information processing device that completes the processing earliest.

  In addition, the task assignment destination determination unit according to the present invention determines which of the task calculation amount and the transfer data amount acquired by the task information acquisition unit has a greater influence on the time until the task processing result is obtained. That is, it is determined whether the requested task has a large calculation amount and a long time required for calculation, or a large amount of transfer data and a long time required for communication. When it is determined that the degree of influence of the calculation amount is large, an evaluation function with a large calculation capacity weight is used. When it is determined that the amount of influence of the transfer data amount is large, the communication overhead time and the data transfer rate are determined. It is preferable to use an evaluation function in which the weighting is increased.

  Here, the above determination can be made based on, for example, whether or not the transfer data amount is equal to or greater than a predetermined threshold. That is, if the transfer data amount is equal to or greater than the threshold value, it is determined that the influence level of the transfer data amount is large. If the transfer data amount is less than the threshold value, it is determined that the influence amount of the calculation amount is large.

  As a result of this determination, when it is determined that the degree of influence of the transfer data amount is large, the allocation overhead may be determined by weighting the communication overhead time and the data transfer speed with a large weight and the calculation capacity with a small weight. The information processing apparatus to be assigned may be determined based only on the transfer speed (that is, the calculation capacity is weighted to 0). When it is determined that the degree of influence of the calculation amount is large, the calculation capacity may be heavily weighted, and the allocation destination may be determined by weighting the communication overhead time and the data transfer speed smallly. In other words, the information processing apparatus to which the data is assigned may be determined (with the data transfer rate weighted to 0).

  In this way, it is possible to easily determine the assignment destination according to the type of task by determining whether the task has a large amount of computation or a task having a large amount of communication and uses a different evaluation function.

The present invention is more effective when applied to a distributed processing system in which an information processing apparatus that is a mobile terminal is connected by a wireless communication network. This is because in wireless communication, the communication overhead time and data transfer speed vary depending on the radio wave condition, so that the importance of using these pieces of information for determining task assignment destinations increases.

  In addition, the task assignment device according to the present invention preferably includes a computing resource monitoring unit that periodically measures a communication overhead time and a data transfer rate when communicating with the information processing device. The computing resource monitoring unit updates the communication overhead time and the data transfer rate stored in the computing resource storage unit.

  The measurement of the communication overhead time and the data transfer rate may be performed by actually transmitting data, or may be estimated by measuring the radio wave intensity or the error rate.

  As described above, by measuring the communication overhead time and the data transfer rate, it is possible to appropriately grasp the change in the communication overhead time and the data transfer rate according to the change in the radio wave condition. Therefore, a more accurate determination can be made in determining the task assignment destination.

  It is also preferable to determine the task assignment destination in consideration of the task priority. The priority of the task is considered when determining the assignment destination as follows. Each task in the present invention is given a priority, and the computing resource storage means stores the priority of the task already assigned to the information processing apparatus. Then, the task assignment destination determination means determines the information processing apparatus to which the requested task is assigned in consideration of the priority of the requested task and the priority of the task already assigned to the information processing apparatus.

  By using the priority in this way, priority control for quickly processing important tasks becomes possible. At this time, by assigning a task to an information processing device to which a task having a lower priority than the requested task is assigned or to an information processing device to which no task is assigned, the task is immediately executed in the information processing device to which the task is assigned. Guaranteed to be processed.

  The task assignment method according to the present invention is a task assignment method for assigning a task to an information processing device in a distributed processing system in which a plurality of information processing devices are connected via a communication network. The task allocation method according to the present invention includes a calculation resource required for execution of a requested task by a task allocation device having calculation resource storage means for storing the calculation capability, communication overhead time, and data transfer rate of a plurality of information processing devices. And a step of acquiring a transfer data amount generated when the task is assigned to the information processing device, information on the information processing device stored in the calculation resource storage means, and a calculation amount of the task acquired by the acquiring step And determining an information processing apparatus to which a task is assigned based on the transfer data amount.

  According to the present invention, it is possible to assign tasks in consideration of the time until a processing result is acquired in a distributed processing system.

  Exemplary embodiments of the present invention will be described in detail below with reference to the drawings.

  Here, as a specific example, an example in which a task is requested from a vehicle-mounted terminal mounted on a vehicle using road-to-vehicle communication to a task allocation device installed on the base station side will be described.

<System configuration>
FIG. 1 is a diagram showing a system configuration of the present embodiment. The base station 1 installed on the roadside has LA
Via N, the task assignment device 2 and servers D and E for executing the requested task are connected. Vehicles A, B, and C exist within the communicable area of the base station 1, and vehicle-mounted terminals 3a, 3b, and 3c are mounted on the vehicles, respectively. The in-vehicle terminal of each vehicle requests the task assignment device 2 to execute the task via the base station 1. The in-vehicle terminals 3a, 3b, and 3c are also assigned tasks from the task assignment device 2 and execute the tasks.

  FIG. 2 is a schematic diagram showing the configuration of the task assignment device 2. The task assignment device 2 includes a computation resource monitoring unit 11, a computation resource storage unit 14, a task management unit 15, a task information acquisition unit 16, and a task assignment destination determination unit 17.

  The computing resource monitoring unit 11 monitors information about computing resources in the system (in this case, the in-vehicle terminals 3a, 3b, 3c and the servers D, E). The information monitored by the computing resource monitoring unit 11 is information related to the computing capability of each computing resource and information related to communication between the task allocation device 2 and the computing resource. These pieces of information may take a configuration in which each computing resource periodically notifies the computing resource monitoring unit 11 or may make a configuration in which an inquiry is made to each computing resource from the computing resource monitoring unit 11 side.

  Information acquired by the computing resource monitoring unit 11 is stored in the computing resource storage unit 14. FIG. 3 is a diagram showing a table storing information of each computing resource that the computing resource storage unit 14 has.

  In the present embodiment, the computing capacity of the computing resource is managed by the CPU ability 21 and the load on the CPU (CPU availability 22). The CPU capability 21 is managed by a score (score) that is proportional to the amount of calculation that the CPU of each calculation resource can perform per unit time. Since the CPU capability 21 is a calculation capability in a state where no load is applied to the CPU, the substantial calculation capability of each arithmetic device at the time of task allocation can be obtained from the CPU capability 21 and the CPU vacancy rate 22.

  The communication overhead time 23 is a time required from when a transmission command is instructed to when communication is actually started when communication is performed between the task assignment device 2 and the computing resource. The communication overhead time 23 is a value that is not affected by the amount of data to be transmitted. The communication overhead time 23 can be obtained, for example, as a difference between the time when the transmission source terminal instructs transmission and the time when the reception destination terminal starts reception processing.

  The data transfer rate 24 is the amount of data that can be transmitted per unit time when communication is performed between the task assignment device 2 and the computing resource. For example, the data transfer rate 24 is obtained from the amount of data per unit time transmitted in the past, or is obtained by periodically transmitting data of a certain length and measuring the time required to transmit this data. It can be obtained by transmitting a certain amount of data in the idle time and measuring the time required to transmit this data.

  The priority 25 is the priority of the task currently assigned to the computing resource. Here, the smaller the number, the higher the priority. Information about priority is not stored in a computing device to which no task is assigned.

  The task management unit 15 receives a task execution request from the request source, receives the task, transfers the task to the computation resource, receives the task processing result from the computation resource, and transmits the processing result to the request source. .

The task information acquisition unit 16 acquires a calculation amount necessary for executing the requested task and a transfer data amount generated between the calculation resources. The time required for calculation can be estimated from the calculation amount and the calculation capability of the calculation resource stored in the calculation resource storage unit 14.

  The amount of transfer data generated between the computing resources is the total of the communication amount generated when transferring a task and the communication amount generated when receiving a processing result. It is possible to estimate the time required for communication from the transfer data amount and the transfer speed between the calculation resources stored in the calculation resource storage unit 12.

  The task assignment destination determination unit 17 determines the assignment destination of the task for which the task management unit 15 has accepted the request, based on the information stored in the computing resource storage unit 14 and the information on the task acquired by the task information acquisition unit 16. . The allocation destination determination process will be described later.

<Task>
Examples of tasks requested to be executed in the present embodiment include a route search process. The route search processing can be divided into a plurality of tasks of general road priority and required time priority, or can be divided into tasks for performing route search for each section by designating an intermediate point. The division of tasks may be performed on the in-vehicle terminal (navigation system) side, or may be performed by the task management unit 15 in the task assignment device 2. In this embodiment, it is assumed that tasks are divided on the in-vehicle terminal side and the task allocation device 2 is requested to execute each task.

<Processing flow>
The flow of processing from when a task is requested to be executed by the task assignment device 2 until the task is assigned to a computing resource will be described with reference to the flowchart of FIG.

  First, in step S101, the task management unit 15 receives a task request. In step S102, the task information acquisition unit 16 acquires the requested task calculation amount and transfer data amount.

  In step S103, the task assignment destination determination unit 17 assigns a task based on the information on each computation resource stored in the computation resource storage unit 14, and the computation amount and transfer data amount obtained by the task information acquisition unit 16. To decide.

  In step S104, the task management unit 15 transfers the task to the calculation resource to which the task is assigned.

<Assignee determination process>
A task assignment destination determination process performed by the task assignment destination determination unit 17 in step S103 will be described.

  In the present embodiment, the computing capacity, communication overhead time, and data transfer rate of the computing resource stored in the computing resource storage unit 14, and the task computation amount and transfer data amount acquired by the task information acquisition unit 16 are used as variables. The task processing capability of each computing resource is evaluated using the evaluation function to be performed. In the evaluation function in this embodiment, the time from when a task is assigned to a computing resource until a processing result is obtained is defined as the task processing capability of the computing resource.

In the present embodiment, the task processing capability of the computing resource is evaluated using the following evaluation function.
(Task processing capacity) = (computation amount) / (computation capacity) + (communication overhead time) + (transfer data amount) / (data transfer speed)

FIG. 5 conceptually illustrates the above evaluation function. First, it is possible to estimate the time required for calculation from the calculation amount of the task and the calculation capability of the calculation resource. In this embodiment, since the computing power of computing resources is grasped by the CPU processing capacity and the CPU load (free CPU ratio), the CPU processing is calculated as the computing capacity of each computing resource at the time of task allocation. Use capacity multiplied by free CPU rate. In addition, the processing capacity of the CPU is grasped by the score, and the amount of computation that can be processed by the CPU having a predetermined score is known in advance, so it is possible to estimate the time required for task computation from the amount of computation required for task execution. is there.

  Further, the time required for data transfer is estimated based on the transfer data amount and the transfer speed. The communication time is estimated by adding the transfer time and the communication overhead time generated when performing communication.

  By adding the calculation time and communication time calculated in this way, it is possible to obtain the processing time required from task transfer to result acquisition. In the evaluation function according to the present embodiment, this processing time is the task processing capability of each computing resource, and the task processing capability is greater as the processing time is shorter.

  The computing resource storage unit 14 stores the priority of the task assigned to each computing resource, and task assignment is assigned to a computing resource to which a priority higher than the priority of the requested task is assigned. Do not do. This is because even if a task is assigned, it is not executed immediately.

  The task assignment destination determination unit 17 selects the computation resource having the largest task processing capacity calculated above from among the computation resources to which a task having a priority lower than the requested task is assigned or to which the task is not assigned. Decide to assign tasks. However, it is not always necessary to assign a task to a computing resource with the maximum task processing capacity. If a time limit (deadline) until task processing is completed is determined, processing that can complete processing within this time limit A task may be assigned to any computing resource as long as it is a resource.

  As described above, the task assignment device 2 according to the present embodiment determines the assignment destination based on the total time of the time required for calculation and the time required for communication when determining the calculation resource (information processing device) to which the task is assigned. Therefore, it is possible to assign a task to a computing resource that takes the shortest time from assigning a task to obtaining a processing result.

  Further, since the communication time is calculated separately for the communication overhead time that does not depend on the transfer data amount and the data transfer rate that is proportional to the transfer data amount, a more accurate communication time can be calculated. For example, when the amount of transferred data is small, the communication overhead time is relatively important, and when the amount of transferred data is large, the data transfer rate is relatively important.

  In addition, since the distributed processing system according to the present embodiment is configured by a mobile terminal (on-vehicle terminal) using wireless communication, the radio wave condition constantly changes as the terminal moves. Since the computing resource monitoring unit 11 monitors the communication overhead time and the data transfer rate, the allocation according to the communication status becomes possible.

<Modification>
(Modification of assignment destination determination process)
In addition to the assignment destination determination method as described above, the task assignment destination determination unit 17 in this embodiment determines whether the requested task is a task with a large amount of computation or a task with a large amount of communication, and determines a different evaluation function. The method used may be used.

That is, the following functions are used as evaluation functions, and different values are used for the weighting parameters p and q depending on the type of task.
(Task processing capacity) = pC (computation amount, computation ability) + qT (transfer data amount, communication overhead time, data transfer rate)
Here, C is a function for evaluating the capability (calculation time) related to calculation from the calculation amount and calculation capability, and T is a function for evaluating the capability (communication time) related to communication from the transfer data amount, communication overhead time, and data transfer speed. , P and q are weighting parameters.

  Task characteristic analysis is performed, for example, by determining whether or not the communication amount is equal to or greater than a predetermined threshold value. Can be considered.

  In the case of a task with a large amount of computation, an evaluation function is used that increases the weighting of the CPU processing capacity and the CPU vacancy rate and decreases the weighting of the transfer speed and the communication overhead time. In the case of a task with a large amount of communication, an evaluation function is used that increases the weighting of the transfer speed and the communication overhead time and decreases the weighting of the CPU processing capacity and the CPU idle rate.

  In the case of a task with a large amount of calculation, an evaluation function may be used in which the weighting of the transfer rate is eliminated and the task processing capability is a value obtained by multiplying the CPU processing capability and the CPU vacancy rate. In the case of a task with a large amount of communication, an evaluation function may be used in which the CPU processing capacity is not weighted and the task processing capacity is a value based on the transfer speed and communication overhead time.

  By determining the assignment destination by such processing, it is possible to perform evaluation according to the characteristics of the task by simple processing.

(Variation of monitoring method for computing resources)
In this embodiment, only the processing power and load of the CPU are used as computing power. However, for example, the available memory amount of computing resources and the memory amount required for task execution are acquired, and these information is also taken into account for allocation. The destination may be determined. Further, the allocation destination may be determined in consideration of other resources.

  In the present embodiment, the communication resource monitoring unit 11 measures the communication overhead time and the data transfer rate by actual measurement. However, the communication overhead time and the data transfer rate are determined using predetermined values for each OS and each device used. Also good. This has the advantage of being efficient because the measurement can be omitted when the communication situation is stable.

(Modification of system configuration)
In this embodiment, there is one base station 1 and the vehicles A, B, and C exist within the communicable area of the base station 1, but a plurality of base stations 1a, 1b, and 1c exist as shown in FIG. However, the system configuration may be such that they exist via a network.

  By connecting a plurality of base stations through the network in this way, the communicable area of the base station is substantially expanded, and a system covering a wide range can be constructed.

  In the present embodiment, only one task allocation device 2 exists in the system, but a plurality of task allocation devices 2 may exist. For example, as shown in FIG. 7, in an ad hoc network in which each terminal autonomously forms a network without a base station, each terminal (vehicles A, B, C) has task allocation devices 2a, 2b, 2c, respectively. Is preferred.

  Vehicles A, B, and C notify each other of information such as the computing capabilities of in-vehicle terminals 3a, 3b, and 3c possessed by each vehicle through inter-vehicle communication. And when requesting a task from the in-vehicle terminal, the request is made to the task assignment device of the own vehicle.

  At this time, it is preferable that this system not only performs inter-vehicle communication but also has a server capable of performing communication with the base station and performing task calculations on the base station side. In this case, the in-vehicle terminal is generally a computing resource with low computing capacity but low communication cost, and the server is a computing resource with high computing capacity but high communication cost. Therefore, in such an environment where various computing resources are mixed, it is important that the task assignment device of each vehicle determines an appropriate assignment destination according to the task characteristics.

It is a schematic diagram which shows the system configuration | structure in this embodiment. It is a figure which shows the structure of the task assignment apparatus in this embodiment. It is a figure which shows the example of the table which memorize | stored the information about the calculation resource in this embodiment. It is a flowchart figure which shows the process which determines the allocation destination of the task in this embodiment. It is a figure explaining the calculation method of the processing capacity of each computing resource in this embodiment. It is a schematic diagram which shows the system configuration | structure in this embodiment. It is a schematic diagram which shows the system configuration | structure in this embodiment.

Explanation of symbols

A, B, C Vehicle D, E Server 1 Base station 2, 2a, 2b, 2c Task allocation device 3a, 3b, 3c In-vehicle terminal 11 Computing resource monitoring unit 14 Computing resource storage unit 15 Task management unit 16 Task information obtaining unit 17 Task assignment destination determination section

Claims (8)

In a distributed processing system in which a plurality of information processing devices are connected via a communication network, a task assignment device that assigns a task to the information processing device,
Computing resource storage means for storing computing power, communication overhead time and data transfer rate of the plurality of information processing devices;
Task information acquisition means for acquiring the amount of calculation required for execution of the requested task and the amount of transfer data generated when the task is assigned to the information processing device;
The information processing device to which the task is assigned is determined based on the information related to the information processing device stored in the computing resource storage unit and the calculation amount and transfer data amount of the task acquired by the task information acquisition unit. A task assignment destination determination means;
A task allocation device having: The task assignment destination determining means is configured to perform each information processing by an evaluation function using the calculation capability of the information processing apparatus, the communication overhead time and the data transfer rate, and the calculation amount and the transfer data amount of the task as variables. The task assignment device according to claim 1, wherein the task processing capability of the device is evaluated, and an information processing device to which the task is assigned is determined based on the task processing capability. The task allocation destination determining means estimates a calculation time from the calculation capability and the calculation amount, estimates a communication time from the communication overhead time, the data transfer rate, and the transfer data amount, and calculates a sum of the calculation time and the communication time. The task allocation apparatus according to claim 2, wherein an evaluation function having the task processing capability as an evaluation function is used. The task allocation destination determining means determines which of the calculation amount and transfer data amount of the task has a large influence on the time until the task processing result is obtained, and determines that the influence amount of the calculation amount is large. When the evaluation function with the increased weight of the computing capacity is used, and when it is determined that the degree of influence of the transfer data amount is large, the evaluation function with the increased weight of the communication overhead time and the data transfer rate is The task assignment device according to claim 2, wherein the task assignment device is used. The task assignment apparatus according to claim 1, wherein the communication network is a wireless communication network. Computation resource monitoring means for periodically measuring communication overhead time and data transfer speed when communicating with the information processing apparatus and updating the communication overhead time and data transfer speed stored in the computation resource storage means The task assignment device according to claim 1, wherein: Each task has a priority,
The computing resource storage means stores the priority of a task already assigned to the information processing apparatus,
The task assignment destination determination means considers the priority of the requested task and the priority of the task already assigned to the information processing device when determining the information processing device to which the requested task is assigned. The task assignment device according to claim 1, wherein the task assignment device is a device. In a distributed processing system in which a plurality of information processing devices are connected via a communication network, a task assignment method for assigning a task to the information processing device,
A task allocation device having computing resource storage means for storing computing power, communication overhead time and data transfer rate of the plurality of information processing devices,
Obtaining a calculation amount necessary for executing the requested task and a transfer data amount generated when the task is allocated to the information processing device;
The information processing device to which the task is allocated is determined based on the information regarding the information processing device stored in the computing resource storage unit and the calculation amount and the transfer data amount of the task acquired in the acquiring step. And steps to
Task assignment method including

Images