WO2022202676A1

WO2022202676A1 - Information resource allocation assistance device, information resource allocation assistance method, and recording medium storing information resource allocation assistance program

Info

Publication number: WO2022202676A1
Application number: PCT/JP2022/012662
Authority: WO
Inventors: 雄一赤池
Original assignee: 日本電気株式会社
Priority date: 2021-03-25
Filing date: 2022-03-18
Publication date: 2022-09-29
Also published as: JPWO2022202676A1

Abstract

An information resource allocation assistance device 30 comprises: an acquisition unit 31 for acquiring acceptable conditions 310 relating to an execution time required for executing first AI processing and a usage fee for a hardware resource 34 used for the first AI processing, the time and fee being desired by a user; a calculation unit 32 for calculating, on the basis of calculation criteria 320 relating to the relationship between an execution time required for second AI processing and a usage fee for the hardware resource 34 used for the second AI processing, the execution time required for the first AI processing, and the usage fee for the hardware resource 34 used for the first AI processing that satisfy the acceptable conditions 310; and an output unit 33 for outputting the execution time and the usage fee calculated by the calculation unit 32. Thus, the user is provided with a resource with which it is possible to execute desired AI processing at the usage fee and within the AI processing execution time that are permissible by the user.

Description

Information resource allocation support device, information resource allocation support method, and recording medium storing information resource allocation support program

The present invention relates to an information resource allocation support device, an information resource allocation support method, and an information resource allocation support program.

In a computer system, it is important to appropriately allocate information processing resources (resources) for executing information processing in accordance with the performance requirements for the information processing to be executed. For example, if an excessive amount of information processing resources is allocated for requirements such as performance, some of the information processing resources may not be effectively used, and the usage fee for the information processing resources may become higher than necessary. Alternatively, for example, if insufficient information processing resources are allocated to requirements such as performance, the execution time of information processing may become unacceptably long, or information processing may end abnormally due to insufficient resources. Therefore, a technique for appropriately allocating information processing resources according to requirements for information processing is expected.

As a technology related to such technology, Patent Document 1 discloses an arrangement device that arranges network functions to multiple types of hardware. This device stores statistical information including performance information representing the resource consumption and performance of each of a plurality of types of hardware that are candidates for placement of functions. This device accepts input of functional description contents in a high-level language, which is a hardware description language corresponding to multiple types of hardware, and performance requirements representing required performance. This device calculates predicted performance, which is predicted performance, and predicted resource consumption, which is predicted resource consumption, for each piece of hardware using description content and a predetermined algorithm. Then, the device selects, as a placement destination, hardware for which the calculated predicted performance and performance information satisfy the performance requirements and the total value of the predicted resource consumption amount and the resource consumption amount is equal to or less than the resource capacity.

In addition, Patent Document 2 discloses a resource determination device that determines a combination of resources and resource parameters that are suitable for the needs of the service layer. The device analyzes service layer needs expressed in natural language for communication services. The device determines a combination of resources suitable for the needs of the analyzed service layer. The device then determines the parameters of the resources included in the determined combination of resources.

JP 2019-212069 A JP 2019-087105 A

For example, with cloud AI (Artificial Intelligence), it is possible to use resources on the cloud and execute desired AI processing by paying a predetermined fee. However, when using resources on the cloud, it is difficult to strike a balance between the user's desired usage fee and the user's allowed AI processing execution time.

A main object of the present invention is to provide an information resource allocation support device or the like that solves the above-mentioned problems.

Acquisition means for acquiring user-desired permissible conditions regarding the execution time required for execution of the first AI processing and the usage fee for the hardware resources used for the first AI processing; the execution time required for the first AI processing that satisfies the allowable condition, and the first and an output means for outputting the execution time and the usage fee calculated by the calculating means.

In another aspect of achieving the above object, an information resource allocation support method according to an aspect of the present invention provides a user-desired execution time required for execution of a first AI process and an execution time required for the first AI process. Acquire the permissible conditions related to the hardware resource usage fee and the usage fee of the hardware resource used for the second AI processing, based on the calculation criteria related to the relationship between the execution time required for the second AI processing and the hardware resource usage fee used for the second AI processing. An execution time required for the first AI processing that satisfies an allowable condition and a usage fee for hardware resources used in the first AI processing are calculated, and the calculated execution time and usage fee are output.

In addition, in a further aspect of achieving the above object, an information resource allocation support program according to an aspect of the present invention provides a user-desired execution time required to execute a first AI process and the first AI process. Acquisition processing for acquiring permissible conditions related to usage fees for hardware resources used, execution time required for second AI processing, and calculation criteria for the relationship between usage fees for hardware resources used in the second AI processing a calculation process for calculating an execution time required for the first AI process that satisfies the allowable condition and a usage fee for hardware resources used in the first AI process, based on the calculation process; and an output process for outputting the execution time and usage fee calculated by the computer.

Furthermore, the present invention can also be implemented by a computer-readable, non-volatile recording medium storing such an information resource allocation support program (computer program).

According to the present invention, it is possible to provide the user with a resource capable of executing desired AI processing at the usage fee and execution time allowed by the user.

1 is a block diagram showing the configuration of an information resource allocation support device 10 according to a first embodiment of the present invention; FIG. When the acquisition unit 11 according to the first embodiment of the present invention acquires the permissible condition information 141 via the chatbot, an input instruction for the permissible condition information 141 by the chatbot and a user input in response to the input instruction 14 is a diagram exemplifying permissible condition information 141. FIG. 4 is a diagram illustrating execution history information 143 according to the first embodiment of the present invention; FIG. 4 is a flow chart showing the overall operation of the information resource allocation support device 10 according to the first embodiment of the present invention; Based on the calculation criteria 142 generated or updated by the information resource allocation support apparatus 10 according to the first embodiment of the present invention while performing a regression analysis on the evaluation performance of the execution time of the script and the usage fee of the

hardware resource

15, 4 is a flow chart showing detailed operations for calculating an execution time and usage fee; It is a block diagram which shows the structure of the information resource allocation assistance apparatus 30 based on the 2nd Embodiment of this invention. 1 is a block diagram showing the configuration of an information processing device 900 capable of realizing an information resource allocation support device according to each embodiment of the present invention; FIG.

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

<First embodiment>
FIG. 1 is a block diagram showing the configuration of an information resource allocation support device 10 according to the first embodiment of the present invention. When executing a script (an example of the first AI processing) using the hardware resource 15, the information resource allocation support device 10 balances the execution time of the script and the usage fee of the hardware resource 15 based on the user's allowable conditions. It is a device for calculating the value obtained by Note that the script according to the present embodiment is, for example, AI processing that analyzes an event to be analyzed by machine learning. AI processing is roughly divided into “learning processing” for generating a learned model and “analysis processing” executed using the learned model.

The hardware resource 15 has a CPU (Central Processing Unit) 151 and a memory 152 . The hardware resource 15 may comprise multiple CPUs 151, and the CPUs 151 may be processors having multiple cores. The memory 152 is a main memory that stores a program representing a script executed by the CPU 151 and data used when the CPU 151 executes the script. The hardware resource 15 may include a storage device such as a hard disk that is used as extended storage when the CPU 151 executes the script.

The hardware resource 15 may be provided in the information resource allocation support device 10, as illustrated in FIG. 1, or may be a device such as a server different from the information resource allocation support device 10. It may be provided in a device communicably connected to the support device 10 .

The information resource allocation support device 10 is communicably connected to the terminal device 20 . The terminal device 20 is an information processing device including a display screen 200, such as a personal computer, tablet, or smart phone. The terminal device 20 accepts an input operation by the user when the user inputs information to the information resource allocation support device 10 . The terminal device 20 also displays information output from the information resource allocation support device 10 on the display screen 200 .

The information resource allocation support device 10 includes an acquisition unit 11, a calculation unit 12, an output unit 13, a storage unit 14, and a generation unit 16. The acquisition unit 11, the calculation unit 12, the output unit 13, and the generation unit 16 are examples of acquisition means, calculation means, output means, and generation means, respectively.

The storage unit 14 is a storage device such as a RAM (Random Access Memory) 903 or a hard disk 904, which will be described later with reference to FIG. The storage unit 14 stores permissible condition information 141, calculation criteria 142, and execution history information 143, which will be described later.

The acquisition unit 11 acquires the permissible condition information 141 regarding the execution time required to execute the AI processing and the usage fee for the hardware resources used for the AI processing. In addition to the above, the allowable condition information 141 may include the name (identification information) of the script to be executed by the hardware resource 15 and the size of the input data used by the script. Also, the permissible condition information 141 is not limited to the above as long as it is information related to the execution of AI processing.

FIG. 2 shows an input instruction of the permissible condition information 141 by the chatbot when the acquisition unit 11 according to the present embodiment acquires the permissible condition information 141 via the chatbot, and a user input in response to the input instruction. 3 is a diagram exemplifying permissible condition information 141. FIG. The acquisition unit 11 displays an input instruction for the permissible condition information 141 illustrated in FIG. 2 on the display screen 200 of the terminal device 20, and acquires the permissible condition information 141 input by the user in response to the input instruction. Note that the information resource allocation support device 10 performs login processing (authentication processing) using a user ID (Identification) before displaying the screen illustrated in FIG. 2 on the display screen 200 .

In the example of FIG. 2, the acquisition unit 11 displays an input instruction for the name of the script, "Please enter the script to be executed." /home/hoge/script1.py” to get the name of the script.

Next, the acquisition unit 11 displays on the display screen 200 an input instruction for the size of the number of samples of the input data used by the script, "Please enter the number of samples (number of data lines)." acquires the value of the number of samples "300000" input by the user in response to .

Next, the acquisition unit 11 displays on the display screen 200 an input instruction for the size of the number of attributes of the input data used by the script, such as "Please enter the number of attributes (the number of data columns)." Acquire the value of the attribute number "50" input by the user in response to .

Next, the acquisition unit 11 displays an instruction to input an allowable script execution time (maximum value of script execution time allowed by the user) such as “Please enter an allowable execution time (seconds)” on the display screen 200 . , and acquires the allowable execution time (seconds) value of "2000" input by the user in response to the input instruction.

Next, the acquisition unit 11 receives the server (hardware resource 15 ) is displayed on the display screen 200 to prompt the user for confirmation. However, the server usage fee is an amount based on the amount of the hardware resource 15 used for executing the script, which is calculated by the calculation unit 12 to be described later. In the example of FIG. 2, the user has entered "N" to indicate that the allowable execution time is to be changed. This is because the user judged that the server usage fee of 1000 yen was too high.

Next, in response to the user inputting "N", the acquisition unit 11 accepts an input by the user to change the allowable execution time, such as "Please enter the allowable execution time (seconds)." is displayed on the display screen 200. In the example of FIG. 2, a new allowable execution time (seconds) value of "3000" has been entered by the user.

Next, the acquisition unit 11 asks the user about the server usage fee, "The server usage fee is expected to be 700 yen. Enter Y to continue execution, and enter N to change the allowable execution time." An instruction to request confirmation again is displayed on the display screen 200 . However, the server usage fee of 700 yen is an amount based on the amount of hardware resources 15 used to execute the script calculated by the calculator 12 based on the new allowable execution time of 3000 seconds. In the example of FIG. 2, the user has entered "Y" indicating that the script is to be executed without changing the allowable execution time. This is because the user judged that the server usage fee of 700 yen was appropriate.

After that, the acquisition unit 11 displays on the display screen 200 "Preparing the execution environment. Please wait for a while." Display a button that the user can press on the screen, "Access environment". After the button is pressed by the user, the hardware resource 15 starts executing the script "/home/hoge/script1.py".

The acquisition unit 11 changes the input format by the user regarding execution requirements such as the size of the input data used by the script according to the name of the script in the UI (User Interface) via the chatbot described above.

Note that the input instruction of the permissible condition information 141 by the chatbot illustrated in FIG. 2 and the permissible condition information 141 input by the user in response to the input instruction are examples. may differ from the example of FIG.

In the example of FIG. 2, it is possible to select whether or not to accept the amount of the usage fee presented to the user from the server side, but the user may input the desired usage fee amount. . For example, the acquisition unit 11 may acquire an amount (permissible usage fee) such as "1000 yen to 1500 yen" from the user as the usage fee for the server desired by the user.

In addition, the acquisition unit 11 may acquire the permissible condition information 141 using a method other than the UI such as the chatbot described above. The acquisition unit 11 may acquire the permissible condition information 141 input via an API (Application Programming Interface), for example.

The acquisition unit 11 stores the permissible condition information 141 acquired as described above in the storage unit 14 .

The calculation unit 12 calculates the execution time of the script and the usage fee of the hardware resource 15 based on the permissible condition information 141 stored in the storage unit 14 and the calculation criteria 142 .

Calculation standard 142 calculates the execution time of the script and the usage fee of the hardware resource 15 to be used from the permissible condition information 141 regarding the script to be executed, from the viewpoint of the balance between performance (execution time) and cost (usage fee). , are criteria for suitable calculation based on the user's allowable conditions (execution requirements). The calculation criteria 142 are criteria given by the user, and are generated or updated by the generation unit 16 using the execution history information 143 as will be described later. However, the execution history information 143 represents the execution requirements, the execution time, the usage fee, and the amount of the hardware resource 15 that executed the script (an example of the second AI processing) regarding the script executed in the past, It is information generated or updated by the information resource allocation support device 10 .

FIG. 3 is a diagram illustrating the execution history information 143 according to this embodiment. The execution history information 143 may be information managed for each user who uses the hardware resource 15 . In the execution history information 143 illustrated in FIG. 3, one row individually numbered represents an execution history of a script executed on a certain date and time. In the execution history information 143 illustrated in FIG. 3, lines numbered 1 to 6 represent the execution history related to the script "script1.py", and lines numbered 7 to 10 represent the execution history related to the script "script2.py". represents.

The execution history information 143 includes the data size (MB: megabytes), the number of data attributes (number of data columns), and the number of data samples (number of data rows) as the size of the input data used by the executed script. The number of data attributes and the number of data samples are values input by the user when the script is executed, as illustrated in FIG. 2, and the data size is a value determined from the number of data attributes and the number of data samples.

The execution history information 143 also includes the actual execution time (seconds) of the script, as illustrated in FIG.

The execution history information 143 also includes the number of used cores for the CPU 151 and the maximum used memory (MB) and reserved memory (MB) for the memory 152 as the amount of the hardware resource 15 that executed the script. The reserved memory indicates the size of the area reserved in the memory 152 as a storage area for data that can be used by the script, and the maximum used memory is the area actually used by the script in the reserved area. represents the size of

The execution history information 143 also includes the usage fee (yen) of the hardware resource 15 that executed the script, calculated by the calculation unit 12 .

The execution history information 143 also includes the user's evaluation of the actual execution time of the script and the usage fee of the hardware resource 15 . For example, the execution history of number 1 in the execution history information 143 illustrated in FIG. 3 includes an evaluation by the user that "the execution time may be longer, so I would like the usage fee to be lower." This means that the execution of the script indicated by the execution history of number 1 used an excessive amount of the hardware resource 15 for the execution requirements originally desired by the user.

Alternatively, for example, the execution history numbered 2 in the execution history information 143 illustrated in FIG. This means that an insufficient amount of the hardware resource 15 was used for the execution requirements originally desired by the user in executing the script indicated by the execution history of number 2 .

Alternatively, for example, the execution history number 4 in the execution history information 143 illustrated in FIG. 3 includes the user's evaluation of "appropriate." This means that an appropriate amount of hardware resources 15 that meet the execution requirements originally desired by the user were used in executing the script indicated by the execution history of number 4 .

Note that the execution history information 143 illustrated in FIG. 3 is an example, and the execution history information 143 may include items different from the items illustrated in FIG.

The calculation unit 12 refers to the execution history information 143 when calculating the execution time of the script and the usage fee of the hardware resource 15 .

The calculation unit 12 first determines in the execution history information 143 whether or not the number of executions of the script to be executed is equal to or greater than a predetermined threshold. That is, the calculation unit 12 determines whether or not a sufficient amount of execution history has been accumulated for the script to be executed. When the number of executions of the script is less than a predetermined threshold (that is, the execution history is not sufficiently accumulated yet), the calculation unit 12 uses the calculation formula given in advance by the user as the calculation standard 142. . The calculation unit 12 calculates the execution time of the script and the usage fee of the hardware resource 15 based on this calculation formula and the permissible condition information 141 . The calculation formula is, for example, a calculation formula that uses the script name (identification information), the size of the input data used by the script, and the allowable execution time of the script, which are included in the allowable condition information 141, as variables. be.

When the execution history information 143 determines that the number of executions of the script to be executed is equal to or greater than a predetermined threshold value (that is, the execution history is sufficiently accumulated), the calculation unit 12 calculates based on the calculation criteria 142. , the execution time of the script and the usage fee of the hardware resource 15 are calculated. At this time, the generation unit 16 generates or updates the calculation criteria 142 while performing regression analysis on the user's evaluation indicated by the execution history information 143 .

In this case, the calculation unit 12 sets the amount of the hardware resource 15 to be used as a variable X, and sets the value of the variable X to a predetermined initial value. The calculation unit 12 sets, for example, the number of cores of the CPU 151 as a variable X, and sets “1” as its initial value.

The calculation unit 12 finds in the execution history of the script to be executed that the size of the input data is the same and the amount of the hardware resource 15 to be used is the same as X (that is, the initial value “1”). Search for what is For example, the execution history information 143 is as exemplified in FIG. 12000". In this case, the execution history of number 2 in the execution history information 143 substantially matches the search condition described above. Then, since the evaluation by the user in the execution history of number 2 is "I want to make the execution time shorter because the usage fee may be higher", the generation unit 16 updates the calculation criteria 142, for example, the variable Set X to "2" which is twice the current value.

Next, the calculation unit 12 finds, in the execution history related to the script to be executed, that the size of the input data is the same and the amount of the hardware resource 15 to be used is the same as X (that is, "2"). search for something. In this case, the execution history of number 3 in the execution history information 143 substantially matches the search condition described above. Then, since the evaluation by the user in the execution history of number 3 is "I want to make the execution time shorter because the usage fee may be higher", the generation unit 16 updates the calculation criteria 142, for example, the variable Set X to "4" which is double the current value.

Further, the calculation unit 12 further determines, in the execution history of the script to be executed, that the size of the input data is the same and the amount of the hardware resource 15 to be used is the same as X (that is, "4"). Search for what is In this case, the execution history of number 4 in the execution history information 143 substantially matches the search condition described above. Since the user's evaluation of the execution history of number 4 is "appropriate", the generation unit 16 determines the value of the desired variable X (that is, the number of cores of the CPU 151) to be "4".

The calculation unit 12 and the generation unit 16 also calculate the execution time of the script and the usage fee of the hardware resource 15 based on the permissible condition information 141 and the calculation criteria 142 by a procedure different from the procedure described above. good too.

For example, according to the user's evaluation indicated by the execution history of number 1 in the execution history information 143 shown in FIG. That is, the allowable execution time (not shown) in the execution history of number 1 is longer than the actual execution time of 130 seconds. Regarding the execution of the script "script1.py", if the execution requirements regarding the size of the input data and the allowable execution time are as shown in number 1 of the execution history information 143, the generation unit 16 determines that the execution time of the script is Create or update the calculation criteria 142 so that they are longer and cost less to use the hardware resource 15 .

Also, according to the user's evaluation indicated by the execution history of number 2 in the execution history information 143 illustrated in FIG. That is, the permissible execution time (not shown) in the execution history of number 2 is a value shorter than the actual execution time of 1350 seconds. Regarding the execution of the script "script1.py", if the execution requirements regarding the size of the input data and the allowable execution time are as shown in number 2 of the execution history information 143, the generation unit 16 determines that the execution time of the script is Create or update the calculation criteria 142 so that it is shorter and charges more for using the hardware resource 15 .

Also, according to the user's evaluation indicated by the execution history of number 4 in the execution history information 143 illustrated in FIG. . That is, the permissible execution time (not shown) in the execution history of number 2 is approximately the same value as the actual execution time of 658 seconds. Regarding the execution of the script "script1.py", if the execution requirements regarding the size of the input data and the allowable execution time are as shown in number 4 of the execution history information 143, the generation unit 16 determines the execution time and The calculation criteria 142 are generated or updated so that the usage fee for the hardware resource 15 becomes the value indicated by number 4 .

The generation unit 16 generates the calculation criteria 142 by learning the relationship between the permissible condition information 141 acquired in the past and the execution history information 143 corresponding to the permissible condition information 141 .

For example, the generation unit 16 generates or updates a model representing the calculation criteria 142 by performing learning using the execution history information 143 as learning data. In this case, the generating unit 16, for example, for each script to be executed, determines the relationship between the size of the input data and the allowable execution time (not shown), the execution time of the script, and the usage fee of the hardware resource 15. , learning may be performed using the user's evaluation as a label.

The output unit 13 outputs the usage fee and the execution time calculated by the calculation unit 12.

For example, the output unit 13 outputs the usage fee and the execution time calculated by the calculation unit 12 to the terminal device 20 .

Further, for example, after obtaining approval from the user, the output unit 13 outputs the execution time of the script and the usage fee of the hardware resource 15 calculated as described above based on the allowable condition information 141 and the calculation criteria 142. You may output to the allocation part (not shown) with which the resource allocation assistance apparatus 10 is equipped.

The allocation unit determines the amount of hardware resources 15 to be used according to the execution time of the script output from the calculation unit 12 and the usage fee for the hardware resources 15 . The allocation unit then allocates the determined amount of hardware resources 15 to execute the script.

It should be noted that the hardware resource 15 may include a plurality of execution environments in which scripts can be executed, with different configurations of the CPU 151 and memory 152 . The multiple execution environments with different configurations are, for example, an execution environment including one core in the CPU 151 and a 1 GB (gigabyte) storage area in the memory 152, and two cores in the CPU 151 and a 2 GB storage area in the memory 152. It is an execution environment including In this case, the information resource allocation support device 10 selects one of the plurality of execution environments described above based on the execution time of the script output from the output unit 13 and the usage fee of the hardware resource 15. All you have to do is assign execution. If there is no execution environment with a configuration that matches the amount of hardware resources 15 determined by the allocation unit, the information resource allocation support device 10 instructs the hardware resources 15 to newly start up an execution environment that matches the amount. may be controlled.

The information resource allocation support device 10 controls the hardware resource 15 allocated to execute the script so as to start executing the script. After the execution of the script by the hardware resource 15 is completed, the information resource allocation support device 10 generates or updates the execution history information 143 representing the execution history of the script.

The information resource allocation support device 10 is provided with library functions for reading data (files) from the storage unit 14 . When data is read by the library function, the information resource allocation support device 10 stores information such as the size of the read data, the number of data attributes, the number of data samples, etc. in the storage unit 14 as the execution history information 143 . By reading the library function, the information resource allocation support device 10 stores the script execution start date and time and the execution time obtained from the difference between the execution end date and time and the execution start date and time in the storage unit 14 as the execution history information 143. Store.

Next, with reference to the flowcharts of FIGS. 4 and 5, the operation (processing) of the information resource allocation support device 10 according to this embodiment will be described in detail.

FIG. 4 is a flowchart showing the overall operation of the information resource allocation support device 10 according to this embodiment.

The acquisition unit 11 acquires the permissible condition information 141 regarding the script to be executed via the chatbot, and stores it in the storage unit 14 (step S101). The calculation unit 12 refers to the execution history regarding the script to be executed, which is included in the execution history information 143 stored in the storage unit 14 (step S102). The calculation unit 12 determines whether or not the number of times the script has been executed is greater than or equal to a threshold (step S103).

If the number of times the script has been executed is greater than or equal to the threshold (Yes in step S104), the information resource allocation support device 10 performs the process of step S200, which will be described later with reference to FIG. 5, and the process proceeds to step S106. When the number of executions of the script is not equal to or greater than the threshold (No in step S104), the calculation unit 12 uses a calculation formula given in advance as the calculation criterion 142, and based on the calculation criterion 142 and the permissible condition information 141, , the execution time of the script and the usage fee of the hardware resource 15 are calculated (step S105). The output unit 13 outputs the calculated execution time of the script and the usage fee of the hardware resource 15 to the terminal device 20, and after the usage fee and the execution time are approved by the user in the terminal device 20, the usage fee is calculated. And the execution time is output to the allocation unit (step S106).

Based on the execution time and usage fee calculated by the calculation unit 12, the allocation unit selects a suitable execution environment for executing the script from among a plurality of execution environments with different configurations included in the hardware resource 15. (step S107). The information resource allocation support device 10 controls the selected execution environment so as to execute the script (step S108). After the execution of the script is completed, the information resource allocation support device 10 generates or updates the execution history information 143 including the execution history and the user's evaluation of the execution time and usage fee (step S109). Processing ends.

FIG. 5 shows that the information resource allocation support apparatus 10 according to the present embodiment executes based on the calculation criteria 142 generated or updated while performing the regression analysis on the evaluation performance of the execution time of the script and the usage fee of the hardware resource 15. 4 is a flowchart showing detailed operations for calculating time and usage fees; However, the flowchart shown in FIG. 5 corresponds to the process of step S200 described above with reference to FIG.

The calculation unit 12 sets the value of the variable X to a predetermined initial value related to the amount of hardware resources 15 to be used (step S201). The calculation unit 12 searches the execution histories related to the script to be executed for those having the same size of input data and the same amount of hardware resources 15 to be used as X (step S202). .

If there is no unconfirmed execution history that satisfies the conditions (No in step S203), the calculation unit 12 performs the process of step S105 described above with reference to FIG. 4, and the process of step S200 ends. If there is an unconfirmed execution history that satisfies the conditions (Yes in step S203), the calculation unit 12 confirms the execution time of the script in the execution history that satisfies the conditions and the evaluation of the usage fee of the hardware resource 15 (step S204). .

If the evaluation is "appropriate" (Yes in step S205), the calculation unit 12 determines the amount of the hardware resource 15 to be used to be the value of X (step S207), and the process of step S200 ends. If the evaluation is not "appropriate" (No in step S205), the calculation unit 12 sets the value of X to a predetermined criterion If the evaluation is "I want to shorten the execution time even if the usage fee is high", the value of X is increased according to a predetermined standard (step S206), and the process returns to step S202.

The information resource allocation support device 10 according to the present embodiment can provide the user with resources that can execute desired AI processing at the usage fee and AI processing execution time allowed by the user. This is because the information resource allocation support device 10 uses the calculation criteria 142 to calculate the execution time and usage fee that satisfy the user's allowable conditions for the hardware resource 15 that executes information processing.

The effects realized by the information resource allocation support device 10 according to this embodiment will be described below.

For example, with cloud AI, by paying a predetermined fee, it is possible to use hardware resources such as data centers of other companies and execute desired AI processing. However, when using the hardware resources of other companies, it is difficult to strike a balance between the user's desired usage fee and the user's allowable AI processing execution time.

The information resource allocation support device 10 according to this embodiment includes an acquisition unit 11, a calculation unit 12, and an output unit 13, and operates as described above with reference to FIGS. 1 to 5, for example. That is, the acquisition unit 11 acquires the user-desired permissible condition information 141 regarding the execution time required to execute the first AI processing and the usage fee of the hardware resource 15 used for the first AI processing. . The calculation unit 12 calculates the first data that satisfies the permissible condition based on the calculation standard 142 regarding the relationship between the execution time required for the second AI processing and the usage fee of the hardware resource 15 used for the second AI processing. The execution time required for the AI processing and the usage fee for the hardware resource 15 used for the first AI processing are calculated. Then, the output unit 13 outputs the execution time and usage fee calculated by the calculation unit 12 .

That is, the information resource allocation support apparatus 10 according to the present embodiment calculates the execution time and usage fee that satisfy the user's allowable conditions for the hardware resource 15 for executing the script using the calculation criteria 142. It is possible to provide the user with a resource capable of executing a desired AI process at the usage fee and the execution time of the AI process that are allowed by the user.

Also, the information resource allocation support device 10 according to this embodiment acquires the execution history information 143 including the execution time required to execute the script and the usage fee of the hardware resource 15 in executing the script. Then, the information resource allocation support device 10 refers to the evaluation result as to whether or not the execution time and usage fee are appropriate. As a result, the information resource allocation support device 10 can calculate a well-balanced usage fee and execution time for script execution.

Also, the acquisition unit 11 according to the present embodiment may further acquire the data size of the program code related to the script to be executed. In this case, the calculation unit 12 further uses the data size, and based on the calculation criteria 142, the execution time required to execute the script that satisfies the allowable condition indicated by the allowable condition information 141, and the hardware to be used. A usage fee for the wear resource 15 may also be calculated.

Also, the information resource allocation support device 10 according to this embodiment acquires the permissible condition information 141 input by the user via the chatbot. Then, the information resource allocation support device 10 changes the input format by the user regarding the execution requirements such as the size of the input data in the UI via the chatbot, according to the script to be executed. As a result, the information resource allocation support device 10 can acquire the permissible condition information 141 that more accurately reflects the user's request through a simple input operation by the user.

Further, the information resource allocation support device 10 according to this embodiment includes a plurality of hardware resources 15 with different configurations that are ready to execute scripts, and the hardware resources 15 according to the calculated execution time and usage fee. , the script execution is assigned to one of the plurality of hardware resources 15 . As a result, the information resource allocation support device 10 can quickly start execution of the script by the hardware resource 15 .

<Second embodiment>
FIG. 6 is a block diagram showing the configuration of the information resource allocation support device 30 according to the second embodiment of the present invention. However, for convenience of explanation, the hardware resources 34 are shown in FIG.

The information resource allocation support device 30 according to this embodiment includes an acquisition unit 31 , a calculation unit 32 and an output unit 33 . The acquisition unit 31, the calculation unit 32, and the output unit 33 are examples of acquisition means, calculation means, and allocation means, respectively.

The acquisition unit 31 acquires user-desired permissible conditions 310 regarding the execution time required to execute the first AI process and the usage fee for the hardware resource 34 used in the first AI process. The hardware resource 34 is, for example, a resource similar to the hardware resource 15 according to the first embodiment. The first AI process is, for example, a script-like process according to the first embodiment. The permissible condition 310 is, for example, information similar to the permissible condition information 141 according to the first embodiment. The acquisition unit 31 operates, for example, in the same manner as the acquisition unit 11 according to the first embodiment.

The calculation unit 32 calculates the first AI processing that satisfies the permissible condition 310 based on the calculation criteria 320 regarding the relationship between the execution time required for the second AI processing and the usage fee of the hardware resource 34 used for the second information processing. and the usage fee for the hardware resources used for the first AI processing. The calculation criterion 320 is, for example, the same criterion as the calculation criterion 142 according to the first embodiment. The calculator 32 operates, for example, in the same manner as the calculator 12 according to the first embodiment.

The output unit 33 outputs the execution time and usage fee calculated by the calculation unit 32 . The output unit 33 operates, for example, in the same manner as the output unit 13 according to the first embodiment.

The information resource allocation support device 30 according to the present embodiment can provide the user with resources that can execute desired AI processing at the usage fee and AI processing execution time allowed by the user. The reason for this is that the information resource allocation support device 30 can present to the user the execution time and usage fee that satisfy the user's allowable conditions regarding the hardware resource 34 for executing information processing.

<Hardware configuration example>
Each unit in the information resource allocation support apparatus shown in FIG. 1 and FIG. 6 in each of the above-described embodiments can be realized by a dedicated HW (Hardware) (electronic circuit). In addition, in FIGS. 1 and 6, at least the following configuration can be regarded as a functional (processing) unit (software module) of a software program including instructions to be executed by a processor.
-

Acquisition units

11 and 31,
-

calculators

12 and 32,
-

output units

13 and 33,
- A storage control function in the storage unit 14,
- Generation unit 16 .

However, the division of each part shown in these drawings is a configuration for convenience of explanation, and various configurations can be assumed upon implementation. An example of the hardware environment in this case will be described with reference to FIG.

FIG. 7 is a diagram illustrating the configuration of an information processing device 900 (computer) capable of realizing an information resource allocation support device according to each embodiment of the present invention. That is, FIG. 7 shows the configuration of a computer (information processing device) capable of realizing the information resource allocation support device shown in FIGS. show.

The information processing apparatus 900 shown in FIG. 7 has the following components as components.
CPU (Central_Processing_Unit) 901,
ROM (Read_Only_Memory) 902,
RAM (Random_Access_Memory) 903,
- Hard disk (storage device) 904,
a communication interface 905;
- Bus 906 (communication line),
A reader/writer 908 capable of reading and writing data stored in a recording medium 907 such as a CD-ROM (Compact_Disc_Read_Only_Memory);
- An input/output interface 909 such as a monitor, a speaker, and a keyboard.

That is, the information processing device 900 having the above components is a general computer in which these components are connected via a bus 906 . The information processing apparatus 900 may include a plurality of CPUs 901 or may include CPUs 901 configured by multi-cores.

Further, in the above-described embodiment, a computer program capable of realizing the following functions may be supplied to the information processing apparatus 900 shown in FIG. For example, the function is the above-described configuration in the block configuration diagrams (FIGS. 1 and 6) referred to in the description of the embodiment, or the function of the flow charts (FIGS. 4 and 5). The functions of the information resource allocation support apparatus according to this embodiment are achieved by reading the computer program to the CPU 901 of the hardware, interpreting it, and executing it. Further, the computer program supplied to the apparatus may be stored in a readable/writable volatile memory (RAM 903) or a nonvolatile storage device such as ROM 902 or hard disk 904.

Also, in the above case, the method of supplying the computer program to the hardware concerned can adopt a general procedure at present. The procedure includes, for example, a method of installing in the device via various recording media 907 such as a CD-ROM, and a method of downloading from the outside via a communication line such as the Internet. In such a case, the computer program supplied to the information processing apparatus according to this embodiment can be considered to be composed of the code that constitutes the program or the recording medium 907 that stores the code. .

The present invention has been described above using the above-described embodiments as exemplary examples. However, the invention is not limited to the embodiments described above. That is, within the scope of the present invention, various aspects that can be understood by those skilled in the art can be applied to the present invention.

This application claims priority based on Japanese Patent Application No. 2021-051807 filed on March 25, 2021, and the entire disclosure thereof is incorporated herein.

REFERENCE SIGNS LIST 10 information resource allocation support device 11 acquisition unit 12 calculation unit 13 output unit 14 storage unit 141 permissible condition information 142 calculation criteria 143 execution history information 15 hardware resources 151 CPU
152 memory 16 generation unit 20 terminal device 200 display screen 30 information resource allocation support device 31 acquisition unit 310 permissible condition 32 calculation unit 320 calculation criteria 33 output unit 34 hardware resources 900 information processing device 901 CPU
902 ROMs
903 RAM
904 hard disk (storage device)
905 communication interface 906 bus 907 recording medium 908 reader/writer 909 input/output interface

Claims

Acquisition means for acquiring user-desired allowable conditions relating to the execution time required to execute a first AI (Artificial Intelligence) process and the usage fee for hardware resources used in the first AI process;
Execution required for the first AI processing that satisfies the permissible condition based on a calculation standard relating to the relationship between the execution time required for the second AI processing and the usage fee for the hardware resources used for the second AI processing. calculation means for calculating time and usage fees for hardware resources used in the first AI processing;
an output means for outputting the execution time and usage fee calculated by the calculation means;
An information resource allocation support device comprising:
The acquisition means acquires the permissible condition input by the user via a chatbot.
2. The information resource allocation support device according to claim 1.
Further comprising generating means for generating the calculation criteria by learning the relationship between the execution time required for the second AI processing and the usage fee of the hardware resource used for the second AI processing,
3. The information resource allocation support device according to claim 1 or 2.
the acquisition means acquires the user's evaluation of the execution time and the usage fee output by the output means;
4. The information resource allocation support apparatus according to claim 3, wherein said generating means updates said calculation criteria based on said evaluation by said user and said execution time and said usage fee output by said output means.
the acquisition means acquires the permissible condition including the size of input data used by the first AI process;
5. The information resource allocation support device according to any one of claims 1 to 4.
The acquisition means further acquires the data size of the program code related to the first AI processing,
The calculation means further uses the data size, and based on the calculation criteria, the execution time required for the first AI processing that satisfies the allowable condition, and the hardware used for the first AI processing. 6. The information resource allocation support device according to any one of claims 1 to 5, further comprising: calculating a usage fee for the ware resource.
Information processing equipment
Acquiring user-desired permissible conditions regarding the execution time required for execution of the first AI processing and usage fees for hardware resources used in the first AI processing;
Execution required for the first AI processing that satisfies the permissible condition based on a calculation standard relating to the relationship between the execution time required for the second AI processing and the usage fee for the hardware resources used for the second AI processing. calculating the time and the usage fee of the hardware resources used for the first AI processing;
Output the calculated execution time and usage fee,
Information resource allocation support method.
Acquisition processing for acquiring user-desired permissible conditions relating to the execution time required to execute the first AI processing and usage fees for hardware resources used in the first AI processing;
Execution required for the first AI processing that satisfies the permissible condition based on a calculation standard relating to the relationship between the execution time required for the second AI processing and the usage fee for the hardware resources used for the second AI processing. a calculation process for calculating time and a usage fee for hardware resources used in the first AI process;
an output process for outputting the execution time and usage fee calculated by the calculation process;
A recording medium storing an information resource allocation support program for causing a computer to execute