JP-7857460-B1 - Transfer control device, transfer control method, and transfer control program

JP7857460B1JP 7857460 B1JP7857460 B1JP 7857460B1JP-7857460-B1

Abstract

[Challenge] To improve the efficiency of power consumption in data centers. [Solution] The transfer control device 10 has a determination unit 153 and a transfer control unit 154. The determination unit 153 determines which data center will execute the task from among the multiple data centers based on the power supply status to the multiple data centers and the status of the hardware resources held by the multiple data centers. The transfer control unit 154 controls the transfer of data used for the task to the data center determined by the determination unit 153. For example, the determination unit 153 determines which data center will execute the task according to the amount of power supplied to the multiple data centers. [Selection Diagram] Figure 2

Inventors

友近剛史

Assignees

ＮＴＴドコモビジネス株式会社

Dates

Publication Date: 20260512
Application Date: 20250225

Claims (8)

A decision unit that determines which data center to execute a task from among the multiple data centers based on the power supply status to the multiple data centers and the status of the hardware resources possessed by the multiple data centers, A transfer control unit that controls the transfer of data used for the task to the data center determined by the determination unit, It has, The transfer control device is characterized in that the determination unit determines that tasks with a larger size of dynamic data required for execution have a higher degree of real-time capability, and prioritizes determining which data center will execute the task with the higher degree of real-time capability among multiple tasks .
The transfer control device according to claim 1, wherein the determination unit determines which data center to execute the task from among the plurality of data centers based on the status of the hardware resources, which is the number of processors of servers provided in the plurality of data centers.
The transfer control device according to claim 1, wherein the determination unit determines which data center to execute the task from among the plurality of data centers based on the status of hardware resources, which is the ratio of unused processing capacity to the total processing capacity of the processors of the servers provided in the plurality of data centers.
The transfer control device according to claim 1, characterized in that the determination unit determines which data center to execute the task from among the plurality of data centers based on the status of hardware resources, which is the number of unused processors among the processors of servers provided in the plurality of data centers.
The transfer control device according to claim 1, characterized in that the transfer control unit instructs a plurality of devices that acquire training data for a machine learning model to transfer the training data to a data center determined by the determination unit.
The decision unit determines from among the plurality of data centers which data center will perform the task using the machine learning model. The transfer control device according to claim 1, characterized in that the transfer control unit performs control to transfer data for constituting the machine learning model to the data center determined by the determination unit.
A transfer control method performed by a computer, A decision-making process to determine which data center will perform the task from among the multiple data centers, based on the power supply status to the multiple data centers and the status of the hardware resources possessed by the multiple data centers, A transfer control step is performed to control the transfer of data used for the task to the data center determined by the determination step, Includes, The aforementioned decision process determines that tasks with a larger size of dynamic data required for execution have a higher degree of real-time capability, and then prioritizes determining which data center will execute the task with the highest degree of real-time capability among multiple tasks. A transfer control method characterized by the following.
A decision step in which a data center is selected from among the multiple data centers to perform a task, based on the power supply status to multiple data centers and the status of hardware resources possessed by the multiple data centers, A transfer control step which controls the transfer of data used for the task to the data center determined by the decision step, Have the computer run it , The aforementioned decision step determines that tasks requiring a large amount of dynamic data have a higher degree of real-time capability, and prioritizes determining which data center will execute the task with the highest degree of real-time capability among multiple tasks. A transfer control program characterized by the following:

Description

This invention relates to a transfer control device, a transfer control method, and a transfer control program. Conventionally, a method called DCI (Data Center Interconnection) for interconnecting data centers is known. Furthermore, techniques are known for resolving power supply and demand imbalances to individual servers by changing the total number of servers deployed in a data center, or the power settings of each server in operation (see, for example, Patent Document 1). International Publication No. 2022/153346 Figure 1 is a diagram illustrating the overview of the transfer control system according to the first embodiment.Figure 2 shows an example of the configuration of a transfer control device according to the first embodiment.Figure 3 shows an example of data center information.Figure 4 shows an example of server information.Figure 5 shows an example of task information.Figure 6 shows an example of weather information.Figure 7 is a flowchart showing the processing flow of the transfer control device.Figure 8 shows an example of the configuration of a computer that executes a transfer control program. The transfer control device, transfer control method, and transfer control program according to this application will be described in detail below with reference to the drawings. However, the present invention is not limited to the embodiments described below. [First Embodiment] The network system configuration will be explained using Figure 1. Figure 1 is a diagram illustrating the overview of the transfer control system according to the first embodiment. As shown in Figure 1, the network system 1 includes data centers 20a, 20b, 20c, 20d, and 20e. The network system 1 also includes a data transfer control device 10. Hereafter, data centers included in Network System 1 may be referred to as Data Center 20 or simply as "Data Center" without distinction. Furthermore, the number and arrangement of data centers included in Network System 1 are examples only and are not limited to those shown in Figure 1. A data center is a building that houses multiple network devices and multiple servers. In the first embodiment, a data center may also mean a system that includes the multiple network devices and multiple servers located within it. The servers located in the data center may be owned by the user, such as a company, or they may be rental servers leased to the user. Users can exchange data remotely with the servers located in the data center 20. As shown in Figure 1, the multiple data centers included in network system 1 are located in geographically separated locations. The data centers are connected by a network. At least a portion of the network between data centers includes an All-Photonics Network (APN). APN is a technology that enables the construction of high-speed networks by processing all network transfer functions in the optical domain. Specifically, APN is a technology that achieves low power consumption, high quality, high capacity, and low latency communication based on optical-based (photonics-based) technologies such as "optoelectronic integration technology," "high-capacity optical transmission system/device technology," "optical Ising machine," and "optical lattice clock network." For example, by using APN, network system 1 can transfer large amounts of data between data centers in a short time. Furthermore, the data centers included in network system 1 are assumed to be equipped with solar power generation systems. Each data center can use not only grid power supplied by the power company but also power supplied by the solar power generation system. From the perspective of economic efficiency and environmental impact, it is preferable that electricity supplied from the solar power generation system be consumed preferentially over grid electricity. Therefore, the transfer control device 10 controls the system so that electricity supplied from the solar power generation system is used preferentially, based on the electricity supply and demand situation. For example, the transfer control device 10 not only allocates tasks to each data center so that the power supplied from the solar power generation system is consumed efficiently, but also transfers the data necessary for those tasks between data centers. The size of data required for data center tasks can be extremely large. In contrast, the network system 1 of the first embodiment allows for high-speed transfer of large data files because the data centers are connected via APN. The transfer control device 10 may be a computer independent of each data center, or it may be a server located in one of the data centers. The configuration of the transfer control device 10 will be explained using Figure 2. Figure 2 is a diagram showing an example of the configuration of the transfer control device according to the first embodiment. As shown in Figure 2, the transfer control device 10 includes a communication unit 11, an input unit 12, an output u