Method for deploying task to node based on execution completion point, task deployment apparatus and storage medium

1. A non-transitory computer-readable storage medium storing a program that causes a computer to execute a process, the process comprising:

acquiring, when a new process occurred in a management node, a number of processes in a queue included in a node coupled to the management node;

acquiring a process execution rate of the node;

calculating an execution time for the processes by dividing the acquired number of processes by the acquired process execution rate;

calculating an execution completion point for the new process in a case where execution of the new process is started by the node based on the calculated execution time and a communication delay between the management node and the node; and

transmitting the calculated execution completion point to the management node,

the management node is configured to determine whether to cause the node to execute the new process based on the calculated execution completion point.

2. The storage medium according to claim 1 , wherein the process further comprising

adding a predetermined number of dummy processes to the queue in the node based on process time corresponding to the communication delay when the number of processes in the queue storing the process executed by the node is acquired; and

replacing any of the dummy processes added to the queue with the new task in the node when an execution request for the new process is received from the management node.

3. The storage medium according to claim 2 , wherein the adding includes:

when the number of processes in the queue storing the process executed by the node is acquired, causing the node to add a first dummy process to the queue until process time corresponding to the queue becomes process time corresponding to the communication delay, and

causing the node to add a second dummy process which is able to be replaced with the new process, and

the process further comprising

when a new second process is generated based on predetermined data in response to reception of the predetermined data, causing the node to replace the first dummy process added to the queue with the generated second process.

4. The storage medium according to claim 1 , wherein the process further comprising:

receiving, from a second management node managing one or more nodes, second information for specifying the execution completion point for a new second process in a case where a node among the one or more nodes is to execute the second process which is generated based on information regarding a communication delay between the computer and any node and is generated by the computer; and

when the new second process is generated based on predetermined data in response to reception of the predetermined data, determining whether the generated second process is to be executed by the node based on the received second information.

5. The storage medium according to claim 1 , wherein the process further comprising:

generating information for specifying the execution completion point for a second process in a case where the node is to execute the second process based on the acquired number of processes in the node, information regarding a process execution rate in the node, and information regarding a communication delay between the computer and the node; and

when the new second process is generated based on predetermined data in response to reception of the predetermined data, determining whether the generated second process is to be executed by the node based on the generated information for specifying the execution completion point for the generated second process.

6. The storage medium according to claim 1 , wherein the process further comprising

selecting a management node having a smallest communication delay with the node from among one or more management nodes,

wherein the generating includes generating information for specifying the execution completion point for the new process in a case where the node is to execute the new process is generated based on the acquired number of processes in the node, information regarding a process execution rate in the node, and information regarding a communication delay between the selected management node and the node, and

wherein the transmitting includes transmitting the generated information for specifying the execution completion point to the selected management node.

7. The storage medium according to claim 1 , wherein the process further comprising:

selecting a management node having a largest process load among one or more management nodes,

wherein the generating includes generating information for specifying the execution completion point for the new process when the node is to execute the new process is generated based on the acquired number of processes in the node, information regarding a process execution rate in the node, and information regarding a communication delay between the selected management node and the node, and

wherein the transmitting includes transmitting the generated information for specifying the execution completion point to the selected management node.

8. The storage medium according to claim 1 , wherein the acquiring includes acquiring the number of processes in a queue storing a process executed by a node which is randomly selected from among one or more nodes is acquired.

9. The storage medium according to claim 1 , wherein

the acquiring includes acquiring the number of processes in a queue storing a process executed by each of one or more nodes is acquired, and

the generating includes generating information for specifying the execution completion point for the new process in a case where a node which is selected from among the one or more nodes and has a smallest number of processes is to execute the new process is generated based on the number of processes in the node, information regarding a process execution rate in the node, and information regarding a communication delay between the management node and the node.

10. A non-transitory computer-readable storage medium storing a program that causes a computer coupled to a task deployment apparatus to execute a process, the process comprising:

receiving an execution completion point for a new process from the task deployment apparatus, the execution completion point being calculated by a processor included in the task deployment apparatus by using a method that includes acquiring a number of processes in a queue included in a node of the one or more nodes when the new process occurred in the management node, acquiring a process execution rate of the node, calculating an execution time for the processes by dividing the acquired number of processes by the acquired process execution rate, calculating an execution completion point for the new process in a case where execution of the new process is started by the node based on the calculated execution time and a communication delay between the management node and the node; and

determining whether to cause the node to execute the new process based on the calculated execution completion point.

11. The storage medium according to claim 10 , wherein the process further comprising

acquiring a transmission time limit for an execution result of the new process,

wherein the acquiring includes acquiring an execution completion point for the new process in a case where a node managed by the management node is to execute the new process is received from each of a plurality of management nodes, and

wherein the determining includes determining a node in which the execution completion point is earlier than the acquired transmission time limit as a node which is to execute the new process.

12. The storage medium according to claim 10 ,

wherein the acquiring includes acquiring the execution completion point for the new process in a case where a node managed by the management node is to execute the new process is received from each of a plurality of management nodes, and

wherein the determining includes determining a node in which the execution completion point is closest to the present time is determined as a node which is to execute the new process.

13. A task deployment method comprising:

acquiring, by a computer, a number of processes in a queue included in a node coupled to a management node when a new process occurred in the management node;

acquiring a process execution rate of the node;

calculating an execution time for the processes by dividing the acquired number of processes by the acquired process execution rate;

calculating an execution completion point for the new process in a case where execution of the new process is started by the node based on the calculated execution time and a communication delay between the management node and the node;

transmitting the calculated execution completion point to the management node;

receiving, by the management node, the calculated execution completion point for the new process from the computer, and

determining whether to cause the node to execute new process based on the calculated execution completion point.

14. A task deployment apparatus comprising:

a memory; and

a processor coupled to the memory and configured to:

acquire a number of processes in a queue included in a node coupled to a management node when a new process occurred in the management node;

acquire a process execution rate of the node;

calculate an execution time for the processes by dividing the acquired number of processes by the acquired process execution rate;

calculate an execution completion point for the new process in a case where execution of the new process is started by the node based on the calculated execution time and a communication delay between the management node and the node; and

transmitting the calculated execution completion point to the management node,

the management node is configured to determine whether to cause the node to execute the new process based on the calculated execution completion point.

15. A task deployment apparatus comprising:

a memory; and

a processor coupled to the memory and configured to:

receive an execution completion point for a new process from the computer, the execution completion point being calculated by using a process that includes acquiring a number of processes in a queue included in a node of the one or more nodes when the new process occurred in the management node, acquiring a process execution rate of the node, calculating an execution time for the processes by dividing the acquired number of processes by the acquired process execution rate, calculating an execution completion point for the new process in a case where execution of the new process is started by the node based on the calculated execution time and a communication delay between the management node and the node; and

determine whether to cause the node to execute the new process based on the calculated execution completion point.