Information processing device, prediction method, and program
Patent Information
- Application Number
- PCT/JP2024/008362
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-03-05
- Publication Date
- 2025-10-02
AI Technical Summary
Existing technologies face challenges in predicting the optimal timing for transitioning from air conditioning to immersion or liquid cooling methods in data centers due to the difficulty in anticipating the introduction of high heat density ICT equipment, necessitating unplanned layout and equipment changes.
An information processing device that counts communication packets from high heat density ICT devices and predicts the introduction timing of a new cooling method using a prediction unit based on packet number changes, employing a configuration with a filter, counter, storage, and user interface.
Enables accurate prediction of the need for new cooling methods by analyzing packet data, allowing for planned introduction of cooling systems to accommodate increasing demand for high heat density ICT devices.
Smart Images
Figure JP2024008362_02102025_PF_FP_ABST
Abstract
Description
Information processing device, prediction method, and program
[0001] The present invention relates to a technique for predicting when a given cooling method should be implemented.
[0002] In typical data centers, the entire server room is cooled by air conditioning to keep ICT (Information and Communication Technology) equipment at an appropriate temperature. In recent years, due to the spread of services using generative AI, the number of ICT devices installed in data centers that generate high amounts of heat, such as GPU (Graphics Processing Unit) servers, has been increasing, and new cooling methods to replace air conditioning are being investigated.
[0003] An example of a new cooling method that can replace air conditioning is the technology disclosed in Non-Patent Document 1. In the technology disclosed in Non-Patent Document 1, ICT devices with low heat generation density are cooled by air conditioning as usual, and for ICT devices with high heat generation density, two types of cooling methods are applied: a method in which the ICT device is immersed in a tank filled with cooling liquid (liquid immersion method), and a method in which cooling liquid is dripped onto the ICT device (liquid droplet method), thereby efficiently cooling the ICT devices.
[0004] However, to use the immersion or droplet cooling methods, it is necessary to prepare a tank to immerse the ICT equipment and a pump to pump up the cooling liquid that falls to the bottom due to droplets.In addition to the tank and pump, additional equipment such as piping to carry the cooling liquid is also required.For this reason, when introducing new cooling methods such as immersion or liquid cooling, it is best to proceed with the introduction in a planned manner.
[0005] NTT Technical Journal, 2017 Vol. 29 No. 7 pp. 40-43 "Initiatives to Improve Power Consumption Efficiency in Data Centers"
[0006] As mentioned above, when changing the cooling method for ICT equipment in a data center from air conditioning to a new cooling method such as immersion or liquid cooling, it may be necessary to make extensive revisions to the layout and equipment, so it is advisable to proceed in a planned manner.
[0007] However, there is a problem in that it is difficult to predict when a new cooling method should be introduced. For example, the introduction of ICT equipment with high heat density may necessitate the introduction of a new cooling method, but it is difficult to predict the timing of the introduction of such ICT equipment. As a result, the decision to introduce such ICT equipment is made suddenly, and in such cases, it may not be possible to immediately apply the new cooling method to the introduced ICT equipment. Note that this problem can arise not only when switching from an air-conditioning system to an immersion system or liquid-cooling system, etc.
[0008] The present invention has been made in view of the above points, and an object of the present invention is to provide a technique for predicting when a predetermined cooling method should be introduced.
[0009] According to the disclosed technology, an information processing device is provided that includes a counting unit that counts the number of packets whose source or destination is a specific type of device, and a prediction unit that predicts when a specified cooling method should be introduced based on changes in the number of packets over time.
[0010] The disclosed technology provides a technique for predicting when a given cooling method should be implemented.
[0011] Fig. 1 is a diagram showing a schematic configuration when an immersion method is used. Fig. 2 is a diagram showing an example of the configuration of an information processing apparatus 100. Fig. 3 is a flowchart for explaining the operation of the information processing apparatus 100. Fig. 4 is a diagram for explaining prediction processing. Fig. 5 is a diagram showing Example 1. Fig. 6 is a diagram showing Example 2. Fig. 7 is a diagram showing an example of the hardware configuration of the information processing apparatus 100.
[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The embodiment described below is merely an example, and the embodiment to which the present invention is applied is not limited to the following embodiment.
[0013] A technique for predicting the timing at which a predetermined cooling method should be introduced will be described below.
[0014] In this embodiment, the target (predetermined cooling method) for which the introduction time is predicted is a cooling method using the liquid immersion method or the liquid droplet method, but the predetermined cooling method is not limited to the liquid immersion method or the liquid droplet method. Furthermore, the technology according to this embodiment is not limited to data centers but can be applied to all facilities that use ICT devices.
[0015] In addition, in this embodiment, the packets to be counted are packets from ICT devices with high heat density (an example of a specific type of device), but the packets to be counted are not limited to packets from ICT devices with high heat density.
[0016] (Cooling ICT Devices by Liquid Immersion) As described above, this embodiment assumes that the liquid immersion method or the droplet method is used as a cooling method for ICT devices with high heat generation density. Figure 1 shows a schematic configuration when the liquid immersion method is used. As shown in Figure 1, the ICT device (e.g., a motherboard) is immersed in a bath filled with cooling liquid.
[0017] As mentioned above, when switching from air conditioning to immersion or droplet cooling for ICT equipment in a data center, it may be necessary to make extensive changes to the layout and facilities, so it is desirable to proceed with the introduction of immersion / liquid cooling in a planned manner.However, with conventional technology, it was difficult to predict when to introduce immersion / liquid cooling.
[0018] In this embodiment, the information processing device 100 (described later) solves the above problem by predicting the timing based on the number of communication packets handled by the ICT devices in the data center. The configuration and operation of the information processing device 100 will be described in detail below.
[0019] (Device Configuration Example) Fig. 2 shows a configuration example of an information processing device 100 that performs the above-described prediction. As shown in Fig. 2, the information processing device 100 has a filter unit 110, a counter unit 120, a storage unit 130, a prediction unit 140, and a user interface 150. Note that the "filter unit 110 + counter unit 120" may be referred to as a counter unit.
[0020] Furthermore, a cable 20 with a connector 10 extends from the information processing device 100, and the connector 10 is connected to a port of the communication device 200, thereby connecting the communication device 200 to the information processing device 100. One or more ICT devices are connected to the communication device 200 directly or indirectly.
[0021] For example, if the communication device 200 is a repeater hub, the connector 10 of the information processing device 100 is connected to a port of the communication device 200, similar to an ICT device. If the communication device 200 has a mirror port, the connector 10 is connected to the mirror port.
[0022] (Operation of Information Processing Apparatus 100) The operation of the information processing apparatus 100 having the above-described configuration will be described with reference to the flowchart shown in FIG.
[0023] In S1 (step 1), the information processing device 100 captures (acquires) from the communication device 200 packets being transmitted and received at the communication device 200. This capture may be performed by the filter unit 110. The packets captured by the information processing device 100 are packets whose source or destination is an ICT device connected to the communication device 200.
[0024] In S2, the filter unit 110 filters only packets from the packets acquired in S1 that are sent from or destined for ICT devices with high heat density. Note that which ICT devices have high heat density is determined in advance. Specifically, for example, the IP addresses of the ICT devices with high heat density are set in the filter unit 110 in advance.
[0025] The “packets whose source or destination is an ICT device with a high heat density” obtained by filtering by the filter unit 110 are sent to the counter unit 120 .
[0026] The reason for focusing on packets that are either source or destination of ICT devices with high heat density is as follows.
[0027] If the number of packets whose source or destination is an ICT device with a high heat density is on the rise, it is believed that the demand for such ICT devices is on the rise. If demand is increasing, it is expected that ICT devices with even higher heat densities will be introduced into data centers. Therefore, this embodiment focuses on packets whose source or destination is an ICT device with a high heat density.
[0028] In S3, the counter unit 120 counts the number of packets filtered by the filter unit 110 and transmits data on the number of packets to the storage unit 130. The counter unit 120 counts the number of packets per unit time, for example.
[0029] In S4, the storage unit 130 stores the counted number of packets as a list.
[0030] In S5, the prediction unit 140 predicts the planned date for introducing the new cooling method based on the list of packet counts read from the storage unit 130. The prediction method will be described in detail later.
[0031] In S6, the user interface 150 such as a monitor displays (presents) the prediction results obtained by the prediction unit 140 to the user.
[0032] The process of the flowchart in FIG. 3 is executed, for example, periodically, and each time the process is executed, the prediction result displayed in S6 is updated.
[0033] (Details of Prediction Process) An example of the prediction process will be described in detail with reference to Figures 4 and 5. Figure 4(a) shows an example of a packet captured from the communication device 200 by the information processing device 100. In Figure 4(a), it is assumed that the underlined IP addresses A and IP addresses B are IP addresses of ICT devices with high heat density.
[0034] FIG. 4(b) shows the packets that have been filtered and extracted from the captured packets, with the source or destination being an ICT device with a high heat density.
[0035] The counter unit 120 counts the number of packets within a certain unit time among the filtered packets, and stores the relationship between the unit time and the number of packets as a list in the storage unit 130.
[0036] Fig. 5A shows an example of a list stored in the storage unit 130. Fig. 5A shows an example in which the unit time is one hour, and there is one packet in the first unit time and two packets in the second unit time.
[0037] The prediction unit 140 uses the relationship stored in the storage unit 130 to predict the predicted date and time when the number of packets will exceed the threshold (i.e., the timing when a new cooling method should be introduced). Figure 5(b) shows an image of this prediction method. For example, if the number of packets does not exceed the threshold at time A of a certain unit of time, but the number of packets exceeds the threshold (or is equal to the threshold) at time B of the next unit of time, time B is the predicted timing when a new cooling method should be introduced.
[0038] The method of making a prediction using the relationship stored in the storage unit 130 is not limited to a specific method, but may be, for example, a method using the least squares method.
[0039] When the least squares method is used, if the time from the start of measurement is x and the number of packets is y, the following straight line y=ax+b is used, with a and b as coefficients.
[0040]
[0041] Note that n is the number of stored data, and x ｉ is the time from the start of measurement for each data, y ｉ is the number of packets for each data.
[0042] The prediction unit 140 calculates a and b using the above formula, and calculates x, the time when y exceeds the threshold, as the predicted date and time based on y = ax + b. For example, if the threshold is c, the prediction unit 140 calculates the date and time by adding (c - b) / a to the measurement start time.
[0043] Even after a prediction has been made once, the actual data on the number of packets increases as the unit time passes, so the predicted date and time is calculated again using the above method and is constantly updated.
[0044] Hereinafter, examples of how the information processing device 100 is used will be described as Example 1 and Example 2.
[0045] 6 is a system configuration diagram according to Example 1. In Example 1, prediction is performed by the information processing device 100 for each rack in a data center.
[0046] As shown in Fig. 6, an information processing device 100 is connected to a communication device 200 in a rack. The rack shown in Fig. 6 is equipped with ICT devices 210 and 230 with high heat density, such as GPU servers, and other ICT devices 220 and 240.
[0047] The information processing device 100 predicts the timing to introduce a new cooling method for the ICT devices in the rack based on the number of packets whose source or destination is the ICT devices 210, 230 with high heat density in the rack.
[0048] (Example 2) Fig. 7 is a system configuration diagram in Example 2. In Example 2, prediction is performed by the information processing device 100 for each floor in a data center.
[0049] As shown in Fig. 7, an information processing device 100 is connected to a communication device 200 that manages multiple racks on a floor. Two racks are shown in Fig. 7, and are connected to the communication device 200 via communication devices 201 and 202, respectively. Each rack is equipped with an ICT device with a high heat density and other ICT devices.
[0050] The information processing device 100 predicts the timing to introduce a new cooling method for the ICT devices on the floor based on the number of packets whose source or destination is the ICT devices 210, 230 with high heat density in the rack.
[0051] (Hardware Configuration Example) The information processing device 100 described in this embodiment can be realized, for example, by causing a computer to execute a program. This computer may be a physical computer or a virtual machine on the cloud.
[0052] That is, the information processing device 100 can be realized by using hardware resources such as a CPU and memory built into a computer to execute a program corresponding to the processing performed by the information processing device 100. The program can be recorded on a computer-readable recording medium (such as a portable memory) and can be saved or distributed. The program can also be provided via a network such as the Internet or email.
[0053] Fig. 8 is a diagram showing an example of the hardware configuration of the computer. The computer in Fig. 8 includes a drive device 1000, an auxiliary storage device 1002, a memory device 1003, a CPU 1004, an interface device 1005, a display device 1006, an input device 1007, an output device 1008, and the like, all of which are interconnected via a bus B. The computer may further include a GPU.
[0054] The program that realizes the processing on the computer is provided by a recording medium 1001, such as a CD-ROM or a memory card. When the recording medium 1001 storing the program is set in the drive device 1000, the program is installed from the recording medium 1001 to the auxiliary storage device 1002 via the drive device 1000. However, the program does not necessarily have to be installed from the recording medium 1001, but may be downloaded from another computer via a network. The auxiliary storage device 1002 stores the installed program as well as necessary files, data, etc.
[0055] The memory device 1003 reads and stores a program from the auxiliary storage device 1002 when an instruction to start the program is received. The CPU 1004 realizes functions related to the information processing device 100 in accordance with the program stored in the memory device 1003. The interface device 1005 is used as an interface for connecting to a network, etc. The display device 1006 displays a GUI (Graphical User Interface) or the like according to the program. The input device 1007 is composed of a keyboard, mouse, buttons, a touch panel, etc., and is used to input various operation instructions. The output device 1008 outputs the results of calculations.
[0056] (Effects of the technology related to the embodiment) As explained above, the technology described in this embodiment makes it possible to predict future demand for ICT devices with high heat generation density from the number of packets whose source or destination is an ICT device with high heat generation density, thereby making it possible to predict when a new cooling method should be introduced.
[0057] The following additional notes are provided regarding the above-described embodiments.
[0058] <Additional Notes> (Additional Item 1) An information processing device comprising: a counting unit that counts the number of packets sent from or destined for a specific type of device; and a prediction unit that predicts when a predetermined cooling method should be introduced based on changes in the number of packets over time. (Additional Item 2) The information processing device according to Additional Item 1, wherein the specific type of device is a device with a high heat density. (Additional Item 3) The information processing device according to Additional Item 1 or 2, further comprising: a filter unit that receives packets from a communication device connected to a plurality of devices including the specific type of device, and extracts packets sent from or destined for the specific type of device from the received plurality of packets. (Additional Item 4) The information processing device according to any one of Additional Items 1 to 3, wherein the prediction unit predicts when the number of packets per unit time will exceed a threshold as the time to introduce the predetermined cooling method. (Additional Item 5) The information processing device according to any one of Additional Items 1 to 4, wherein the predetermined cooling method is a liquid immersion or droplet cooling method used in data centers. (Supplementary Item 6) A prediction method executed by an information processing device, comprising the steps of counting the number of packets whose source or destination is a specific type of device, and predicting when to introduce a predetermined cooling method based on changes in the number of packets over time. (Supplementary Item 7) A non-transitory storage medium storing a program for causing a computer to function as each unit in the information processing device described in any one of Supplementary Items 1 to 5.
[0059] Although the present embodiment has been described above, the present invention is not limited to such a specific embodiment, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.
[0060] REFERENCE SIGNS LIST 10 Connector 20 Cable 100 Information processing device 110 Filter unit 120 Counter unit 130 Storage unit 140 Prediction unit 150 User interface 200, 201, 202 Communication device 1000 Drive device 1001 Recording medium 1002 Auxiliary storage device 1003 Memory device 1004 CPU 1005 Interface device 1006 Display device 1007 Input device 1008 Output device
Claims
1. An information processing device comprising: a counting unit that counts the number of packets whose source or destination is a specific type of device; and a prediction unit that predicts when a specified cooling method should be introduced based on changes in the number of packets over time.
2. The information processing device according to claim 1, wherein the specific type of device is a device with a high heat density.
3. The information processing device according to claim 1, further comprising a filter unit that receives packets from a communication device connected to a plurality of devices including the specific type of device, and extracts packets whose source or destination is the specific type of device from the plurality of received packets.
4. The information processing device according to claim 1, wherein the prediction unit predicts the time when the number of packets per unit time will exceed a threshold as the time when the predetermined cooling method should be introduced.
5. The information processing device according to claim 1, wherein the predetermined cooling method is a liquid immersion method or a liquid droplet method used in data centers.
6. A prediction method executed by an information processing device, comprising the steps of: counting the number of packets whose source or destination is a specific type of device; and predicting when a predetermined cooling method should be introduced based on changes in the number of packets over time.
7. A program for causing a computer to function as each unit in the information processing device according to any one of claims 1 to 5.