Server load prediction system and server load prediction method
By storing the specification information of the device and the server, predicting the load of the server in the production process, solving the problem that the server load cannot be predicted in the prior art, and achieving accurate prediction and management of the server load before the production process is changed.
Patent Information
- Application Number
- CN202110046308.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2020-03-25
- Filing Date
- 2021-01-14
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2041-01-14
AI Technical Summary
The existing server load prediction system cannot predict server load before the production process changes, resulting in server overload or inability to perform image parsing.
A server load prediction system is provided. By saving the device requirements specifications and server specifications, analog input parameters are produced based on the input calculation conditions, server load is calculated and output, and server load is predicted to predict the load of server in production process.
It can predict server load before the production process changes, help users to grasp the server overload situation, avoid bad conditions, and make appropriate preparations when predictions are insufficient, so as to improve the stability and efficiency of the system.
Smart Images

Figure CN113449899B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a server load prediction system and a server load prediction method, which are suitable for predicting the load of a server connected to a device in a production process. Background Art
[0002] In recent years, with the development of communication technology and analysis technology, technologies for obtaining various information in the production process and digitizing or visualizing various matters have been studied. In this case, in addition to the production device, a device for obtaining various information (information acquisition device) is also provided in the production process. For example, in the case of a camera provided as an information acquisition device to analyze the actions of an operator in the production process, the image captured by the camera is transmitted to a server by wire or wireless means, and the server performs image analysis. Therefore, the system including the above-mentioned server needs to be a structure suitable for the load on the server caused by image analysis.
[0003] Here, as background technology of the present invention, for example, in Patent Document 1, the following is recorded: for the purpose of "providing a cloud system, cloud server, edge server and user device that are highly robust to changes in the operating environment", in a "cloud system having cloud servers, edge servers and user devices that can communicate with each other via a network", "middleware that functions at a lower level of services that respectively operate in the local machines is provided; the middleware has: a monitoring function for monitoring the status of the network, the operating status of the cloud server, the edge server and the user device, a switching function for dynamically switching the system structure including the topological structure of the network based on the monitoring result, and a reallocation function for reallocating the working location of the service as the system structure is switched".
[0004] In addition, Patent Document 1 states that "the middleware MW automatically switches the system structure when the network connection status changes (for example, when moving to a place in the wild where the network cannot be connected) or when the device operation status changes (for example, when a cloud or edge device crashes).
[0005] Patent Document 1: Japanese Patent Application Publication No. 2019-135578
[0006] However, when an information acquisition device (e.g., a camera) is provided for the production process and the server performs image analysis as described above, the load incurred in the server by the image analysis varies not only according to the amount of image data received by the server, but also according to the content of the image and the analysis model used in the analysis even if the amount of image data is the same. Therefore, when the production device used or the number of cameras installed changes due to a change in the production process, the server load also changes. In addition, the server load also changes due to changes in the number of production devices or servers installed.
[0007] Therefore, in order to prevent the server load from exceeding the allowable range, it is desirable to predict the server load in advance before changing the production process and to study a system configuration suitable for the predicted server load.
[0008] However, although Patent Document 1 discloses monitoring the operation status and changing the system configuration based on the monitoring result, it does not disclose predicting the operation status in advance and suggesting the system configuration based on the prediction result. Therefore, in the technology disclosed in Patent Document 1, after making changes in the production process, the system configuration can only be changed based on monitoring the actual operation status. If the server is overloaded due to the above changes, it is possible that an undesirable situation such as being unable to perform image analysis processing may occur. Summary of the invention
[0009] The present invention has been made in consideration of the above problems, and an object of the present invention is to provide a server load prediction system and a server load prediction method capable of predicting the load of a server connected to a device in a production process.
[0010] Means used to solve problems
[0011] In order to solve this problem, the present invention provides a server load prediction system, which is a server load prediction system for predicting the server load of one or more servers connected to one or more devices set in a production process, and comprises: a device requirement specification storage unit, which stores the device requirement specification including server load information and operation information for each of the above-mentioned one or more devices, the above-mentioned server load information indicating the load applied to the above-mentioned server at the connection destination for the purpose of controlling the device or processing information transmitted from the device, and the above-mentioned operation information indicating information related to the operation using the device; a server specification storage unit, which stores the server specification indicating the capacity maintained by each of the above-mentioned one or more servers; an input information preparation unit, which accepts the input of calculation conditions including the designation of the process for which the above-mentioned server load is to be predicted, and prepares the input parameters required for the execution of a simulation for calculating the above-mentioned server load based on the above-mentioned calculation conditions, the above-mentioned device requirement specification and the above-mentioned server specification; a server load calculation unit, which calculates the above-mentioned server load applied to the above-mentioned server at the connection destination by the above-mentioned device used in the above-mentioned process specified by the above-mentioned calculation conditions through the above-mentioned simulation based on the above-mentioned input parameters; and an output unit, which outputs the calculation result of the above-mentioned server load calculation unit.
[0012] In order to solve the above problem, the present invention provides the following server load prediction method by a system for predicting the server load of one or more servers connected to one or more devices provided in a production process. Here, the system stores device requirement specifications including server load information and operation information for each of the one or more devices, the server load information indicating the load imposed on the server at the connection destination for the control of the device or the processing of information transmitted from the device, and the operation information indicating information related to the operation using the device; and stores server specifications indicating the capacity maintained by each of the one or more servers. Furthermore, the server load prediction method comprises: an input information preparation step of accepting input of calculation conditions including the designation of the process for which the server load is to be predicted, and preparing input parameters required for executing a simulation for calculating the server load based on the calculation conditions, the device requirement specifications and the server specifications; a server load calculation step of calculating the server load imposed on the server at the connection destination by the device used in the process specified by the calculation conditions through the simulation based on the input parameters; and an output step of outputting the calculation result of the server load calculation step.
[0013] Effects of the Invention
[0014] According to the present invention, it is possible to predict the load of a server connected to a device in a production process.
[0015] In addition, problems, structures, and effects other than those described above will become clear from the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 This is a block diagram showing a hardware configuration example of the server load prediction system 10 according to the first embodiment.
[0017] Figure 2 2 is a block diagram showing a configuration example of the production process system 20 in operation.
[0018] Figure 3 : is a block diagram showing a functional configuration example of the server load prediction system 10 .
[0019] Figure 4 : is a block diagram showing an example of the internal structure of the input information creation unit 120.
[0020] Figure 5 It is a block diagram showing an example of the internal structure of the server load calculation unit 130.
[0021] Figure 6 It is a diagram showing an example of the device requirement specification table 610 .
[0022] Figure 7 This is a diagram showing an example of the server specification table 620 .
[0023] Figure 8 This is a diagram showing an example of calculation condition 630 .
[0024] Fig. 9 It is a diagram showing an example of the process-device conversion table 640 .
[0025] Fig.10 This is a diagram showing an example of input parameter 650.
[0026] Fig.11 This is a flowchart showing an example of the processing procedure of the server load prediction process according to the first embodiment.
[0027] Fig.12 This is a flowchart showing an example of the processing procedure of the server load calculation process.
[0028] Fig.13 It is a diagram showing an example of the simulation result screen 660 .
[0029] Fig.14 This is a block diagram showing a functional configuration example of a server load prediction system 40 according to the second embodiment.
[0030] Fig.15 This is a flowchart showing an example of the processing procedure of the server load prediction process according to the second embodiment.
[0031] Fig.16 This is a diagram showing an example of the action change list 670 .
[0032] Fig.17 It is a diagram showing an example of the simulation result screen 680 .
[0033] Fig.18 1 is a block diagram showing a configuration example of a production process system 50 in operation.
[0034] Fig.19 This is a diagram showing an example of cumulative degree distribution of server load.
[0035] Description of symbols
[0036] 10.40 Server Load Forecasting System
[0037] 11 Arithmetic Unit
[0038] 12 Memory
[0039] 13 Auxiliary storage device
[0040] 14 Communication Interface
[0041] 15 Input interface
[0042] 16 Output Interface
[0043] 17 Keyboard
[0044] 18 Mouse
[0045] 19 Display device
[0046] 20, 50 Production process system
[0047] 21(211~213) Production equipment
[0048] 22(221~223) Camera
[0049] 23(231, 232) Server
[0050] 24, 30 Network
[0051] 25 Operator
[0052] 110 Storage
[0053] 111 Device requirement specification storage unit
[0054] 112 Server Specifications Storage
[0055] 120 Input Information Production Department
[0056] 121 User Interface
[0057] 122 Process-Equipment Conversion Department
[0058] 123 Server Load Information Retrieval Department
[0059] 124 Input parameter production department
[0060] 130 Server Load Calculation Department
[0061] 131 Input interface
[0062] 132 Calculation Execution Unit
[0063] 133 Program storage unit
[0064] 134 Output Interface
[0065] 140 Output
[0066] 410 Input Information Production Department
[0067] 420 Judgment Department
[0068] 511 Production Equipment
[0069] 521~523 Camera
[0070] 531~534 Wireless
[0071] 541 Route
[0072] 542 Transport vehicle
[0073] 610 Device Requirements Specification Sheet
[0074] 620 Server Specifications
[0075] 630 Calculation Conditions
[0076] 640 Process-Device Conversion Table
[0077] 650 Input parameters
[0078] 660, 680 simulation result screen
[0079] 670 Action Change List
[0080] 690 Graph DETAILED DESCRIPTION
[0081] Hereinafter, several embodiments of the present invention will be described using the drawings.
[0082] In addition, for convenience, the following embodiments are divided into multiple parts or embodiments for description when necessary, but except for the cases specifically indicated, they are not unrelated to each other, but are in a relationship where one is a modification example, detail, supplementary explanation, etc. of a part or all of the other. Each embodiment can be implemented individually or in combination.
[0083] In addition, in the following embodiments, when referring to the number of elements (including number, numerical value, amount, range, etc.), except for cases where it is specifically stated and cases where it is obviously limited to a specific number in principle, it is not limited to the specific number and can be more or less than the specific number.
[0084] Furthermore, in the following embodiments, the constituent elements (including element steps, etc.) are of course not necessarily essential unless otherwise specified or unless otherwise considered to be essential in principle.
[0085] Likewise, in the following embodiments, when referring to the shapes, positional relationships, etc. of constituent elements, except for cases specifically indicated or cases that are considered to be obviously not the case in principle, structures similar to or similar to the shapes, etc. are substantially included. The same applies to numerical values and ranges.
[0086] In addition, in the following description, there is a case where information outputted in correspondence with input is described by expression such as "xxx table", but the information may be data of any structure. In addition, in the following description, the structure of each table is an example, and one table may be divided into two or more tables, or all or part of two or more tables may be one table. In addition, in the following description, there is a case where a function is described by expression "xxx part", but the function may also be realized by executing one or more computer programs.
[0087] In the following embodiments, a general wireless LAN device is used as a part of the system components, and at least a part of the communication required by the system is performed wirelessly. In addition, in the following embodiments, wireless methods other than wireless LAN can also be used as a part of the system components.
[0088] (1) First Embodiment
[0089] (1-1) Structure
[0090] Figure 1 1 is a block diagram showing an example of the hardware configuration of the server load prediction system 10 according to the first embodiment. Figure 1As shown, the server load prediction system 10 is a computer system including a computing unit 11, a memory 12, an auxiliary storage device 13, a communication interface 14, an input interface 15 and an output interface 16. The various functions of the server load prediction system 10 are realized by executing various programs by the computing unit 11.
[0091] The calculation unit 11 is a calculation device that executes a program stored in the memory 12. In addition, part of the processing performed by the calculation unit 11 when executing the program may be performed by another calculation device (for example, FPGA (Field Programmable Gate Array)).
[0092] The memory 12 includes a ROM (Read Only Memory) as a nonvolatile storage element and a RAM (Random Access Memory) as a volatile storage element. The ROM stores unchanging programs (such as BIOS), etc. The RAM is a high-speed and volatile storage element such as a DRAM (Dynamic Random Access Memory), which temporarily stores the program executed by the arithmetic unit 11 and the data used when the program is executed.
[0093] The auxiliary storage device 13 is a large-capacity non-volatile storage device such as a magnetic storage device (HDD (Hard Disk Drive)), a flash memory (SSD (Solid State Drive)), etc. In addition, the auxiliary storage device 13 stores data used by the operation unit 11 when executing the program, and the program executed by the operation unit 11. That is, the various functions of the server load prediction system 10 are realized by reading the program from the auxiliary storage device 13 and loading it into the memory 12 and executing it by the operation unit 11.
[0094] The communication interface 14 is a network interface device that controls communication with other devices according to a prescribed protocol. Figure 1 As shown, the communication interface 14 controls communication with a production process system 20 connected via a network 30 .
[0095] The input interface 15 is an interface to which input devices such as a keyboard 17 and a mouse 18 are connected. When the user operates the input device, the input interface 15 receives input from the user via the input device.
[0096] The output interface 16 is an interface for connecting to an output device such as a display device 19 or a printer (not shown). The output interface 16 outputs the execution result of a simulation program for calculating server load to the output device in a format that can be recognized by a user, for example.
[0097] In addition, the simulation program executed by the calculation unit 11 is provided to the server load prediction system 10 via a removable medium (CD-ROM, flash memory, etc.) or the network 30, and is stored in the non-volatile auxiliary storage device 13 as a non-temporary storage medium. Therefore, the server load prediction system 10 may also have an interface for reading data from a removable medium.
[0098] In addition, the server load prediction system 10 is not limited to the following Figure 1 Although shown as being physically configured on one computer, the computer system may be configured on a plurality of computers configured logically or physically, for example, or may operate on a virtual computer constructed on a plurality of physical computer resources, for example.
[0099] like Figure 1 As shown, the server load prediction system 10 is connected to the operating production process system 20 via the network 30. By making such a structure, it is possible to collect information on device requirement specifications (details will be described later) from the operating production process system 20.
[0100] The production process system 20 is a system constructed in the production process, and is configured by connecting one or more production devices 21 and one or more cameras 22 to one or more servers 23 via a wired or wireless network.
[0101] Figure 2 2 is a block diagram showing a configuration example of the production process system 20 in operation. Figure 2 The production process system 20 shown is a production device 211 to 213 (equivalent to Figure 1 Production device 21), cameras 221-223 (equivalent to Figure 1 Camera 22) and servers 231, 232 (equivalent to Figure 1 The server 23) is connected by a wired network 24.
[0102] The production device 21 (production devices 211 to 213) is a device for performing a predetermined production operation in a production process. Speaking individually, the production device 211 corresponds to the device A, the production device 212 corresponds to the device B, and the production device 213 corresponds to the device C.
[0103] The camera 22 (cameras 221 to 223) is an example of an information acquisition device that acquires various information in the production process, and captures the actions of the operator 25 using the production device 21. Specifically, the camera 221 (camera A) captures the actions of the operator 25 using the device A, the camera 222 (camera B) captures the actions of the operator 25 using the device B, and the camera 223 (camera C) captures the actions of the operator 25 using the device C.
[0104] The server 23 (servers 231 and 232) is a server that obtains information from the production device 21 and the camera 22 and performs processing such as image analysis. Individually speaking, the server 231 is equivalent to the server A, and the server 232 is equivalent to the server B. In addition, since the above-mentioned production device 21 and camera 22 are devices connected to the server 23 in the production process, in the following description, they are collectively referred to as "devices", "devices of the production process" or "devices used in the production process".
[0105] In addition, Figure 1 In the embodiment, the server load prediction system 10 is shown as a structure independent of the production process system 20, but as one aspect of the present invention, the server load prediction system 10 may also be configured as follows: Figure 2 The process operates in the server 231 or the server 232 of the production process system 20 shown.
[0106] Figure 3 1 is a block diagram showing an example of the functional configuration of the server load prediction system 10. Figure 1 and Figure 2 As described above, the server load prediction system 10 is a computer system that predicts the load of the server 23 in the production process system 20. Figure 3 As shown, the server load prediction system 10 includes a storage unit 110 , an input information generating unit 120 , a server load calculating unit 130 , and an output unit 140 as a functional configuration.
[0107] The storage unit 110 includes a device requirement specification storage unit 111 and a server specification storage unit 112 that stores values related to capabilities held by the server as server specifications (server specifications).
[0108] The device requirement specification storage unit 111 stores data indicating the specification (device requirement specification) required by the device for the server 23. The device requirement specification is represented by a value related to the control of the device in the production process or a value related to the capability of the server 23 required to process the information output by the device. Figure 6 , a device requirement specification table 610 is shown as an example of data of the device requirement specification stored in the device requirement specification storage unit 111.
[0109] The server specification storage unit 112 stores data indicating the specifications (server specifications) of the server 23. The server specifications are indicated by values related to the capabilities maintained by the server 23. Figure 7 , a server specification table 620 is shown as an example of data of server specifications stored in the server specification storage unit 112 .
[0110] In addition, the "device required specifications" stored in the server load prediction system of the present invention can be measured and collected by operating each device in the production process for the purpose of data collection, or can be collected from each device when it is actually operating in the production process. In addition, the required specifications when operating in the production process can be continuously measured and the data can be updated as needed. In addition, it can also be configured to continuously retain past information until a specified period has passed.
[0111] In addition, regarding the above-mentioned "device requirement specifications", based on the requirement specifications under a certain condition, the requirement specifications under other conditions can be calculated by calculation. Specifically, for example, the requirement specifications required for the image processing of the camera 22 vary depending on the analytical model of the image, but there are also cases where they vary depending on the number of people in the image. In such a case, if the change in the requirement specifications corresponding to the change in the number of people in the image when the same analytical model is used is known, the requirement specifications for multiple people can also be calculated based on the requirement specifications when the number of people is one.
[0112] The input information creation unit 120 has a function of creating input parameters required for server load calculation by the server load calculation unit 130 and outputting the created input parameters to the server load calculation unit 130 .
[0113] Figure 4 1 is a block diagram showing an example of the internal structure of the input information creation unit 120. Figure 4 As shown, the input information generating unit 120 includes a user interface 121, a process-device conversion unit 122, a server load information search unit 123, and an input parameter generating unit 124. The specific functions of each unit constituting the input information generating unit 120 are described in detail in Fig.11 Details are described in the description of server load prediction processing, etc.
[0114] The server load calculation unit 130 has a function of calculating the server load. Figure 5 , Fig.12 This will be described later.
[0115] Figure 5 1 is a block diagram showing an example of the internal structure of the server load calculation unit 130. Figure 5 As shown in FIG. 1 , the server load calculation unit 130 includes an input interface 131, a calculation execution unit 132, a program storage unit 133, and an output interface 134. Fig.11 Details are described in the description of server load prediction processing, etc.
[0116] The output unit 140 has a function of outputting the calculation result of the server load calculation unit 130. Fig.13 An example of a display screen output by the output unit 140 is shown in FIG.
[0117] (1-2) Data
[0118] Hereinafter, the data used in the server load prediction system 10 will be described with reference to the accompanying drawings. Fig.11 Details are described in the description of server load prediction processing, etc.
[0119] Figure 6 2 is a diagram showing an example of the device requirement specification table 610. The device requirement specification table 610 is an example of the device requirement specification stored in the device requirement specification storage unit 111, and is data for managing the specification required by the device used in the production process for the server 23.
[0120] exist Figure 6 In the case of , the device requirement specification table 610 is composed of data items including data ID 611, device 612, server load information 613, and operation information 614. In addition, the server load information 613 includes small items including amount of computation 6131, memory 6132, and storage 6133, and the operation information 614 includes small items including operation ID 6141, operator ID 6142, associated device 6143, analytical model 6144, and acquisition time 6145.
[0121] The data ID 611 is an identifier of the device requirement specification. The data ID 611 may be assigned by the user or may be automatically assigned by the system.
[0122] Device 612 is a device used in the production process. The "device" that can be recorded in device 612 includes not only devices such as excavators or drilling machines used directly to produce products (such as production device 21), but also devices such as cameras or sensors that are not directly used to produce products (such as camera 22).
[0123] The server load information 613 is information indicating the resources required by the server 23 in order to operate the device 612. Figure 6 In the case of, in the server load information 613, as an example, the amount of calculation 6131, memory 6132, and storage 6133 required for the action are recorded.
[0124] In the job information 614 , information related to the job performed by the device 612 is recorded as various information that may become factors that affect the load of the server 23 (server load information 613 ).
[0125] Specifically, since the server load information 613 changes according to the work performed in the production process and the operator 25 performing the work, Figure 6 In the job information 614, a job ID 6141 as an identifier indicating the job and an operator ID 6142 as an identifier indicating the operator are recorded.
[0126] In addition, the associated device that operates in cooperation with the device 612 corresponding to the server load information 613 may also affect the server load information 613. Figure 6 In the operation information 614, if there is an associated device, the information of the associated device is recorded in the associated device 6143.
[0127] In addition, since the analytical model information used for analyzing the information may also have an impact on the server load information 613, Figure 6 In the operation information 614 of the system, the information representing the analytical model is recorded in the analytical model 6144. It is supplemented as follows: "Analytical model" is equivalent to, for example, an analytical model used when extracting the operator 25 from the images obtained by the cameras A, B, and C in the production process and analyzing the actions of the operator 25. In the case of such an analytical model for image analysis, it is assumed that the information of the image input from each camera 22 to the server 23 is the same data volume. If the type of analytical model used changes, at least one of the required amount of calculation, memory, and storage will change. In addition, depending on the analytical model, there are also cases where the server load changes according to the number of operators 25 in the image. That is, there are cases where the analytical model has an impact on the server load information 613.
[0128] In addition, considering the situation that the server load changes over time and the situation of comparing new and old information, Figure 6 In the operation information 614, the time when the information was obtained is recorded in the acquisition time 6145.
[0129] In addition, regarding the device requirement specification table 610, Figure 6 The data structure shown is only an example, and the device requirement specification table 610 related to this embodiment may also be composed of other data items. Specifically, for example, in the server load information 613, information related to the parameter that has the greatest impact on the server load (for example, only the amount of calculation 6131 when the amount of calculation has the greatest impact) may be stored, and information related to the parameters that have the greatest impact on the server load may also be stored. Figure 6 In addition, for example, in the operation information 614, only the items that have the greatest impact on the server load can be recorded, or only the items that have the greatest impact on the server load can be kept. Figure 6 Items not shown in the example.
[0130] In addition, the value of each item of the device requirement specification table 610 can be recorded or updated at any timing. For example, the server load information 613 and the acquisition time 6145 can also be obtained from the server 23 of the production process system 20 in operation. In addition, for example, regarding the job (job ID 6141), the operator (operator ID 6142), the associated device 6143, and the analytical model 6144, in the case where a management device for managing information about these is separately provided, it can also be obtained from the management device. In addition, for example, the user of the server load prediction system 10 can operate the input device such as the keyboard 17 and the mouse 18 to input the required information to record or update the value of the corresponding item.
[0131] Figure 7 2 is a diagram showing an example of the server specification table 620. The server specification table 620 is an example of the server specification stored in the server specification storage unit 112, and is data for managing information on the capabilities held by the server 23.
[0132] The server specification table 620 is a table that summarizes the catalog values and values obtained through measurement based on the performance of the server 23 used or likely to be used in the production process, and the resource values set by the user as the upper limit when using the server 23. Figure 7 In the case of , the server specification table 620 is composed of data items including a server 621 , a computing amount 622 , a memory 623 , and a storage 624 .
[0133] In the computing amount 622, memory 623, and storage 624 of the server specification table 620, values that take into account the performance of each resource, etc., may be preset, or the user may input values each time when using the server load prediction system 10 in consideration of the production process. In addition, the data items in the server specification table 620 are not limited to Figure 7 Examples include information related to other resources, or only information that has a greater impact on the capabilities of the server required by the equipment in the production process.
[0134] Figure 8 630 is a diagram showing an example of a calculation condition 630. The calculation condition 630 is information that specifies the conditions when the server load prediction system 10 predicts the server load, and can be roughly divided into information 631 on the calculation method and information 632 on the process. The user inputs the desired content of the calculation condition 630 by the user performing an input operation on the user interface 121 of the input information creation unit 120.
[0135] like Figure 8As shown, in the information 631 about the calculation method, a calculation method number 633 indicating the number of the calculation method used in the prediction of the server load is input. In this embodiment, as the calculation method used in the prediction of the server load, for example, Figure 5 The program storage unit 133 stores three calculation methods, namely, calculation method 1 to calculation method 3, and stores simulation programs corresponding to the respective calculation methods in the program storage unit 133. In addition, numbers 1 to 3 are assigned as calculation method numbers corresponding to calculation methods 1 to 3. That is, the program storage unit 133 stores a program numbered 1, a program numbered 2, and a program numbered 3. The characteristics of each calculation method will be described in the server load calculation process described later.
[0136] like Figure 8 As shown, in the information 632 about the process, a process name 634 indicating the name of the process for which the server load is to be predicted, an operator ID 635 indicating the information of the operator responsible for each device (for example, the operator 25 who operates the production device 21), a server name 636 indicating the information of the server 23 planned to be used, and a connection relationship 637 indicating the relationship between each server 23 and the device whose control and information analysis are performed by the server 23 are input.
[0137] In addition, in the server name 636, not only the name of the server 23 to be used can be specified, but also the server capacity allowed for the server 23 during operation can be specified. Specifically, for example, when "70% of the performance upper limit value" is specified, it is specified that 70% of the server capacity (e.g., computing capacity) maintained by the server 23 is used during operation.
[0138] Fig. 9 6 is a diagram showing an example of a process-device conversion table 640. The process-device conversion table 640 is information showing the correspondence between each process of the production process and the device used in each process, and has data items including a process 641 showing each process, a device 642 showing the device used in the process 641, and a job ID 643 showing the identifier of the job performed in the process 641 using the device 642. The process-device conversion table 640 is stored in advance in a storage unit (not shown) in the process-device conversion unit 122.
[0139] Fig.10 6 is a diagram showing an example of the input parameter 650 . The input parameter 650 is a parameter required for the server load calculation unit 130 to calculate the server load, and is created by the input parameter creation unit 124 in the server load prediction process executed by the server load prediction system 10 .
[0140] like Fig.10As shown, input parameter 650 includes device requirement specification 651, server specification 652, calculation method number 653 and connection relationship 654. Since the structure of device requirement specification 651 is based on the structure of device 612 and server load information 613 of device requirement specification table 610, and the structure of server specification 652 is based on the structure of server specification table 620, separate descriptions are omitted. In addition, calculation method number 653 corresponds to calculation method number 633 of calculation condition 630, and connection relationship 654 corresponds to connection relationship 637 of calculation condition 630.
[0141] (1-3) Processing
[0142] The following describes in detail the server load prediction process performed by the server load prediction system 10. The server load prediction process is a process for predicting the server load under conditions specified by the user for the server 23 connected to the production process devices (production device 21, camera 22) in the production process system 20.
[0143] Fig.11 FIG. 1 is a flowchart showing an example of a processing procedure of a server load prediction process according to the first embodiment. Fig.11 In the processing sequence shown, the processing of steps S101 to S108 is mainly performed by the input information creation unit 120 , the processing of step S109 is performed by the server load calculation unit 130 , and the processing of step S110 is mainly performed by the output unit 140 .
[0144] according to Fig.11 First, the calculation condition 630 is input to the input information generating unit 120 by the user inputting the calculation condition to the user interface 121 of the input information generating unit 120 (step S101).
[0145] Next, the process-device conversion unit 122 extracts the process name 634 from the calculation condition 630 input in step S101 (step S102), and further refers to the process-device conversion table 640 to extract the device 642 and operation ID 643 used in the process 641 corresponding to the process name 634 extracted in step S202 (step S103).
[0146] Next, the process-device conversion unit 122 extracts the operator ID 635 that performs work on each device 642 from the calculation condition 630 input in step S101 for each device 642 extracted in step S103 (step S104). Furthermore, the process-device conversion unit 122 outputs the information of the device, work ID, and operator ID extracted in steps S102 to S103 to the server load information search unit 123.
[0147] Next, the server load information search unit 123 uses the device, operation ID, and operator ID input from the process-device conversion unit 122 to search for the corresponding server load information 613 from the device requirement specification table 610 stored in the device requirement specification storage unit 111 (step S105). Furthermore, the server load information search unit 123 outputs the server load information 613 of each device as the search result of step S105 to the input parameter creation unit 124 as the requirement specification (device requirement specification) of each device (step S106).
[0148] In addition, in the present embodiment, in step S103, when the process 641 matching the process name 634 extracted from the calculation condition 630 in step S102 is not recorded in the process-device conversion table 640, or in step S105, when the combination of the job ID 6141 and the operator ID 6142 matching the combination of the job ID and the operator ID input from the process-device conversion unit 122 is not recorded in the device requirement specification table 610, a message notifying the user that "no match" may be displayed, and the user may be requested to input the server load information. Alternatively, the process-device conversion unit 122 or the server load information search unit 123 may search for information on similar conditions and present it.
[0149] Next, the input parameter creation unit 124 creates the input parameter 650 based on the device requirement specification input from the server load information search unit 123 in step S106 and the calculation condition 630 input to the input information creation unit 120 in step S101 (step S107 ).
[0150] Here, the specific method of creating the input parameter 650 by the input parameter creation unit 124 is supplemented. Fig.10 As shown, the input parameter 650 is composed of a device requirement specification 651, a server specification 652, a calculation method number 653, and a connection relationship 654. In step S107, the above-mentioned information constituting the input parameter 650 is created as follows.
[0151] The device requirement specification 651 is created using the device requirement specification input from the server load information search unit 123. The server specification 652 is created by extracting the corresponding server specification (e.g., amount of computation 622, memory 623, storage 624) from the server specification table 620 stored in the server specification storage unit 112 based on the server name 636 recorded in the calculation condition 630. The calculation method number 653 is created using the calculation method number 633 recorded in the calculation condition 630. The connection relationship 654 is created using the connection relationship 637 recorded in the calculation condition 630.
[0152] Next, the input parameter creation unit 124 outputs the input parameter 650 created in step S107 to the server load calculation unit 130 (step S108 ).
[0153] Next, the server load calculation unit 130 performs a process of calculating the server load (server load calculation process) based on the input parameter 650 input in step S108 (step S109). Fig.12 Although described later, in the server load calculation process, a simulation for calculating the server load is executed based on the input parameter 650 , and the calculation result of the simulation is output to the output unit 140 .
[0154] Finally, the output unit 140 outputs the above-mentioned simulation calculation results to the outside (step S110), and the server load prediction process is terminated. In addition, in step S109, the output unit 140 preferably outputs not only the simulation calculation results but also other information related to the simulation (see the following description). Fig.13 ).
[0155] Fig.12 1 is a flowchart showing an example of the processing sequence of server load calculation processing. Fig.12 In Fig.11 The server load calculation process executed in step S109 is an example of a processing procedure showing its details.
[0156] In the server load calculation process, the server load calculation unit 130 starts operating when the input parameter 650 is input from the input parameter creation unit 124 of the input information creation unit 120. At this time, the input interface 131 of the server load calculation unit 130 reads the calculation method number 653 from the input parameter 650 (step S201).
[0157] Next, the calculation execution unit 132 selects a program with a number corresponding to the calculation method number 653 read in step S201 from the programs stored in the program storage unit 133 (step S202), and executes the selected program according to the input parameter 650, thereby performing a simulation of calculating the server load (step S203).
[0158] When the simulation is finished, the calculation execution unit 132 outputs the calculation results of the simulation to the output interface 134 , and the output interface 134 outputs the calculation results of the simulation to the output unit 140 (step S204 ).
[0159] As described above, the server load calculation unit 130 can predict and calculate the server load applied to the server 23 of the production process system 20 under various conditions using various simulation programs according to the calculation conditions 630 specified by the user by executing the server load calculation processing.
[0160] In this embodiment, as the server load, specifically, for example, the ratio of the amount of computation required by the device to the upper limit of the server's computational capacity (server capacity) can be calculated. The object of the server load that can be calculated in this embodiment is not limited to the "computational capacity", and the same applies to other parameters recorded in the server load information 613 of the device requirement specification table 610, that is, the memory and storage. In the following, as the server load, the "computational capacity" is cited as an example as the parameter (item) with the greatest impact, and the amount of computation required for the server for the connection relationship between the device and the server is called the "server load".
[0161] The server load calculation unit 130 stores multiple simulation programs for calculating the server load (here, the amount of computation required by the server) in the program storage unit 133. The user can calculate the server load in a desired calculation method by specifying a number in the calculation method number 631 of the calculation condition 630. Figure 5 In the case of the server load calculation unit 130 shown, simulation programs (programs numbered 1 to 3) corresponding to three calculation methods (calculation methods 1 to 3) are stored in the program storage unit 133. The user can calculate the server load in the desired manner by specifying one of "number 1" to "number 3" in the calculation method number 631 of the calculation condition 630.
[0162] Hereinafter, the features and effects of each program corresponding to each calculation method will be described.
[0163] The program of number 1 simulates the server load based on the connection relationship specified by the user. Therefore, when number 1 is specified, the user needs to input the connection relationship 637 to the calculation condition 630. On the other hand, the programs of number 2 and 3 calculate the server load by changing the connection relationship between the device and the server in various ways during the simulation, so the input of the connection relationship 637 in the calculation condition 630 is not required.
[0164] Therefore, when number 1 is specified, the load of each server can be calculated in the connection relationship between the device and the server specified by the user, so the required calculation amount is less than the simulation performed by the program of numbers 2 and 3, and the calculation result can be expected to be obtained in a relatively short time. However, in the simulation of number 1, if the content of the connection relationship 637 specified by the user is inappropriate, it is possible that a large bias in server load will occur between servers, or a server with a server load exceeding the upper limit of the specification will occur.
[0165] In the simulation performed by the program of number 2, the number of servers required to meet the required specifications of the device (device required specifications 651) is calculated based on the device, and the load of each server is calculated after the connection between each server and the device is determined. In the simulation of number 2, it is also possible to configure the calculation using an existing algorithm (for example, an algorithm for solving the bin packing problem) for the combination of the server and the device so as not to exceed the upper limit of the server specification (server specification 652) and to reduce the number of required servers.
[0166] Therefore, when number 2 is specified, the load of each server can be calculated based on the consideration of the appropriate connection relationship between the server and the device, with the required specifications of the device being given priority. In addition, when the above-mentioned algorithm is used, the load of each server can be known based on the connection relationship between the server and the device being close to the optimal solution. However, in the simulation of number 2, since the required specifications of the device are given priority, there may be a case where the result of the calculation outputs a result that more servers are required than the number of servers assumed by the user.
[0167] Furthermore, when number 2 is specified, the user can obtain an understanding of how many servers with what performance are needed to meet the required specifications of the device, or that the required specifications of the device can be met with fewer servers than expected, based on the results of the simulation. In addition, by comparing the results of the simulation with the connection relationship assumed by the user, the degree of surplus or shortage of the server load can be known. Therefore, if it is known that there are insufficient servers, the system structure can be improved before an adverse condition actually occurs by studying the addition of servers. In addition, if the surplus of servers is known, the use of the remaining servers can be stopped by studying the cessation of the use of the remaining servers, which helps to save energy in the system. Furthermore, by understanding the surplus degree based on the calculation results of the load of each server, for example, it is also possible to study in advance which server the device connected to the server can be connected to in the event of a server failure.
[0168] In the simulation performed by the program of No. 3, the connection between the server and the device is determined within the range that does not exceed the upper limit of the server specifications (server capacity) based on the specifications and number of servers, and the load of each server is calculated. In the simulation of No. 3, it can also be configured so that in order to determine the efficient connection relationship between the server and the device, an existing algorithm (for example, an algorithm for solving the knapsack problem) is used for calculation. However, in the simulation of No. 3, since the specifications and number of servers are fixed and the simulation is performed, when the total of the required specifications of each device exceeds the server capacity, there may be a situation where the device cannot be connected to the server.
[0169] Therefore, when number 3 is specified, the load of each server can be calculated based on the study of the appropriate connection relationship between the server and the device, with the server specifications being given priority. Based on the results of this simulation, the user can know the structure of the device that a certain server can currently accommodate (process). In addition, when the above-mentioned algorithm is used, the load of each server can be known based on making the connection relationship between the server and the device close to the optimal solution.
[0170] In addition, the above three calculation methods 1 to 3 (programs numbered 1 to 3) are just examples, and the present embodiment is obviously not limited to these methods. In addition, it is also possible to use a calculation method that determines the connection relationship between the server and the device in such a way that the number of devices connected to the server becomes equal, and it is also possible to use a calculation method that specifies the connection relationship in such a way that the device is connected to the physically closest server.
[0171] Fig.13 660 is a diagram showing an example of a simulation result screen 660. The simulation result screen 660 is Fig.11 An example of the calculation result of the simulation outputted by the output unit 140 in step S110 is displayed on the screen of the display device 19 .
[0172] Fig.13 The simulation result screen 660 is composed of a calculation method number 661 , a server load 662 , and a connection relationship 663 between the server and the device.
[0173] The calculation method number 661 indicates the number corresponding to the calculation method used in the simulation. Fig.13 In the case of , since it is displayed as "No. 1", it can be seen that the server load calculation was performed using calculation method 1.
[0174] The server load 662 shows the calculation result based on the simulation of the server load calculation process. For example, when the server load is set as the ratio of the amount of computation required by the device to the upper limit of the amount of computation of the server (server capacity), Fig.13 The server load 662 of the simulation shows "Server A: 30% of the upper limit specified by the user", which means that the simulation calculation result is that the load (computation amount) predicted by server A is 30% of the server capacity at the time of operation specified by the calculation condition 630. By displaying the server load 662 in this way, it can be indicated whether the server load calculated by the simulation is more or less than the allowable server capacity specified by the user.
[0175] The server-device connection relationship 663 shows the server-device connection relationship when the calculation result of the server load 662 is calculated by simulation. Fig.13 In this case, since the calculation method number 661 is "number 1", the server load is calculated based on the connection relationship specified by the user, and the display content of the connection relationship 663 between the server and the device corresponds to the connection relationship 637 of the calculation condition 630.
[0176] in addition, Fig.13 The display example is merely an example, and the output unit 140 according to the present embodiment may be configured to output arbitrary information of the server load prediction system 10 .
[0177] As described above, the server load prediction system 10 of the first embodiment can predict the server load of the server 23 connected to the device of the production process and prompt the user before the production process (or before the structure, etc. are changed in the production process) by executing the server load prediction process, by calculating and outputting the server load based on the calculation condition 630 specified by the user. As a result, the user can understand the excess and shortage of the server, and when the shortage of the server is predicted, the user can study whether to prepare the necessary server before the production process. In addition, when the surplus of the server is predicted, the user can study whether the server in the production process can continue to be used, and by not using the server, energy saving in the production process can be achieved.
[0178] In addition, in the present embodiment, in the server load prediction processing, not only the name of the device connected to the server is used, but also the information of the operation and operator using the device, and the data of the analysis model that analyzes the data from the device (for example, the camera 22) are also used (see Figure 6 Device requirement specification table 610), so the server load can be calculated with high accuracy.
[0179] Furthermore, in this embodiment, since the calculation method for calculating the server load in the server load prediction process can be selected from a plurality of calculation methods (for example, calculation methods 1 to 3), the server load can be predicted using a calculation method that meets the user's wishes.
[0180] In addition, in this embodiment, as a derivative example, when "No. 2" or "No. 3" is specified in the calculation method number 633 of the calculation condition 630, the system configuration assumed by the user may be input into the calculation condition 630. In such a case, by providing a function of comparing the system configuration assumed by the user with the calculation result of the simulation, and outputting the comparison result by the output unit 140, the excess and deficiency of the server obtained by the simulation can be clearly indicated on the simulation result screen 660 displayed on the display device 19. As a specific display example, when the result of insufficient servers is obtained, "Insufficient servers: xx servers with performance xx are still required" or the like can be displayed when the result of insufficient servers is obtained, or "Servers remaining: xx servers with performance xx can be stopped" or the like can be displayed when the result of insufficient servers is obtained.
[0181] In addition, in this embodiment, for example, it can also be configured so that machine learning is performed by the server load calculation unit 130 to predict the optimal (or close to the optimal) action change so as to use as many devices and servers as possible within the range in which the server load does not become overloaded in the structure.
[0182] (2) Second Embodiment
[0183] A server load prediction system 40 according to a second embodiment of the present invention will be described. In the second embodiment, the differences from the server load prediction system 10 according to the first embodiment will be mainly described, and description of the configuration and processing common to the first embodiment will be omitted.
[0184] The server load prediction system 10 related to the first embodiment described above is a structure that calculates the server load according to the input conditions (calculation conditions 630) and outputs the calculation results, while the server load prediction system 40 related to the second embodiment performs calculations to change the action conditions or system structure of the device or server so that the server load can be calculated below a threshold, thereby being able to find the action conditions and system structures that make the calculation results of the server load below the threshold from a pre-prepared list (the action change list 670 described later).
[0185] In the second embodiment, the user pre-sets the upper limit of the permissible server load as a threshold. In the second embodiment, the following situations are expressed as "the server load exceeds the threshold": in the calculation method 1 described in the first embodiment, the result of calculating the server load with the combination specified by the user is that the server load exceeds the upper limit; in the calculation method 2, the result of calculating the performance and number of servers that meet the device requirement specifications is that at least one of the performance or number of servers is greater than the performance or number of servers assumed by the user; in the calculation method 3, the result of selecting the device to be connected in such a way that the server load does not exceed the threshold is that a device that cannot connect to the server is generated.
[0186] When the server load exceeds the threshold, one countermeasure method is to install an additional server as described in the first embodiment to reduce the load of the overloaded server. However, when it is difficult to add a server or it takes time to add a server, the server load prediction system 40 of the second embodiment is particularly effective. In addition, when it is desired to study a method of changing the device operation as a countermeasure in the case of a sudden server failure, the server load prediction system 40 of the second embodiment is also particularly effective.
[0187] Fig.14 4 is a block diagram showing an example of the functional configuration of the server load prediction system 40 according to the second embodiment. Fig.14 As shown, the server load prediction system 40 is similar to the server load prediction system 10 of the first embodiment (see FIG. 1 ), except that the input information creation unit 120 is replaced by an input information creation unit 410 and a determination unit 420 is added. Figure 3 )different.
[0188] The input information generating unit 410 has the following functions, similar to the input information generating unit 120 of the first embodiment: generating input parameters required for the server load calculation of the server load calculating unit 130, and outputting the generated input parameters to the server load calculating unit 130. Furthermore, the input information generating unit 410 has the following functions: when the determining unit 420 determines that the server load exceeds the threshold, the server load is calculated based on the pre-generated list ( Fig.16 After changing the action conditions and system configuration of the server or device, the input parameters are re-created. In addition, the internal structure of the input information creation unit 410 is the same as that described in the first embodiment. Figure 4 The structures are the same and are omitted from illustration.
[0189] The determination unit 420 has a function of determining whether the calculation result of the simulation performed by the server load calculation unit 130 can be output as the final calculation result of the server load prediction process. Fig.15 Described in the description of the server load prediction processing.
[0190] In addition, although Fig.14 Although not shown in the figure, in the server load prediction system 40 (for example, the storage unit 110 or the input information generation unit 410), an action change list 670 is pre-created and stored as information for managing changes in the action conditions of the device or server in the production process. Fig.15 The server load prediction process shown in FIG. Fig.16 A specific example thereof is shown in FIG.
[0191] Fig.15 This is a flowchart showing an example of the processing sequence of the server load prediction process according to the second embodiment. Fig.15 The server load prediction processing performed by the server load prediction system 40 will be described in detail. Fig.11 The same processing as the server load prediction processing represented in is assigned the same step number and its description is omitted.
[0192] First, in Fig.15 In the process from when the calculation conditions specified by the user are input into the input information generating unit 410 to when the server load calculating unit 130 performs the initial simulation (steps S101 to S109), the first embodiment ( Fig.11 In the server load calculation process of step S109, the server load calculation unit 130 outputs the calculation result including the simulation of the server load to the determination unit 420.
[0193] Next, the determination unit 420 compares the server load calculated in the server load calculation process of step S109 with the upper limit (threshold value) of the server load set in advance by the user, and determines whether the server load is below the threshold value (step S301 ).
[0194] In step S301, when the server load is below the threshold value (Yes in step S301), it is determined that the server load does not need to be recalculated, so after the determination unit 420 outputs the simulation calculation result to the output unit 140, the process proceeds to step S110. In step S110, as described in the first embodiment, the output unit 140 outputs the simulation calculation result to the outside, and the server load prediction process ends.
[0195] On the other hand, when the server load exceeds the threshold in step S301 (No in step S301 ), the determination unit 420 refers to the action change list 670 to determine whether there is any unimplemented action condition among the action conditions recorded in the action change list 670 (step S302 ).
[0196] Back to Fig.15 In step S302, if there are no unimplemented action conditions in the action change list 670 (Yes in step S302), since no further action condition changes can be made, the determination unit 420 outputs the simulation calculation results to the output unit 140 and then proceeds to step S110. On the other hand, if there are unimplemented action conditions in the action change list 670 in step S302 (No in step S302), the determination unit 420 notifies the input information creation unit 120 of the change in the action condition and then proceeds to step S303.
[0197] In step S303 , the input information creation unit 120 refers to the action change list 670 and creates input parameters for the next simulation based on the record with the highest priority in the order of calculation among the records where the action conditions have not been executed.
[0198] After the processing of step S303, the server load calculation unit 130 executes the simulation again based on the newly created input parameters by returning to step S108 and performing the subsequent processing. The above processing is repeatedly executed until the server load becomes below the threshold (yes in step S301) or there is no action condition 674 for which the simulation is not performed (yes in step S302).
[0199] Fig.16 6 is a diagram showing an example of an action change list 670. The action change list 670 is a list for managing changes in action conditions of a device or server in a production process, and collects a plurality of action conditions assigned with priority. Fig.16 , the action change list 670 is composed of data items including a calculation order 671, a device 672, an action change content 673, and an action condition 674. The calculation order 671 indicates the priority order when the action condition 674 is changed to simulate when the server load exceeds the threshold, and the action conditions are selected in order from the lowest priority. The device 672 indicates the device for changing the action. The action change content 673 indicates the content of the action change. The action condition 674 indicates the content of the change in the specification required from the device to the server.
[0200] use Fig.16 Action change list 670, specifically confirm Fig.15 The process of server load prediction processing.
[0201] In the calculation results of the initial simulation, when the load of at least one server among the multiple servers 23 of the production process system 20 exceeds the threshold (No in step S301), after the determination of step S302 performed by the determination unit 420, the input parameters are re-produced by the input information production unit 410 (step S303), and the simulation (server load calculation processing) is performed by the server load calculation unit 130 (step S109).
[0202] The simulation based on the above-mentioned recreated input parameters is performed using the action condition 674 recorded with "1" in the calculation order 671 in the action change list 670. More specifically, the simulation is performed by changing the transmission interval of the image transmitted from the camera A to the server from the normal 100 millisecond interval to the 200 millisecond interval, so that the required specification for the server becomes 0.5 times the normal one.
[0203] Furthermore, in the calculation results of the above simulation, if the load of at least one server exceeds the threshold value (No in step S301), the input parameter creation and simulation are performed again using the action condition 674 of the record in which "2" is recorded in the calculation order 671 in the action change list 670. In this case, specifically, the transmission interval of the images transmitted from camera A and camera D to the server is changed from the normal 100 millisecond interval to the 200 millisecond interval, so that the required specifications for the servers of the two cameras become 0.5 times the normal ones, and the simulation is performed.
[0204] In this way, Fig.15 In the server load prediction processing, the action condition 674 is changed according to the calculation order 671 of the action change list 670, and the simulation is repeatedly performed until the server load of all the multiple servers 23 of the production process system 20 becomes below the threshold (step S301 is yes) or there are no unimplemented action conditions 674 in the action change list 670 (step S302 is yes).
[0205] In addition, the changes of the operating conditions that can be performed in the second embodiment are not limited to Fig.16 The contents of the action change list 670 shown can be changed in various ways.
[0206] For example, as an example of the action change content 673, a connection destination server of a change device and a related device can be cited. Figure 6In the device requirement specification table 610, although "camera A" is an associated device of "device A", calculation is performed under the constraint that the device and the associated device are connected to the same server in the initial simulation. As a result, if the server load exceeds the threshold, calculations may be performed next and subsequent times under a relaxed condition that the associated device (camera A) can be connected to a server different from the device (device A).
[0207] In addition, not only the conditions of the device can be changed, for example, the server specifications can also be changed, that is, the calculation can be performed under the condition that the upper limit of the server's computing capacity or memory or storage is relaxed, and the relaxation of the threshold can be prompted to the user. In this case, it is only necessary to set the conditions of how much the threshold is relaxed and the order of the calculation for each server, and the calculation can be performed according to the settings. As a result, for example, the user can be prompted with information such as "If 1.1 times the threshold initially set is used as the threshold, the server load converges below the threshold."
[0208] Alternatively, the conditions for achieving a condition below a threshold value may be calculated by changing the analytical model used in analyzing the data to a model that reduces the analytical accuracy but requires less computation, and the user may be presented with the calculation results of the action method and server load.
[0209] Furthermore, in the above description, the changes in the operation of the device, the changes in the restrictions on the connection between the server and the device, the changes in the specifications of the server, and the changes in the analysis model are explained separately, but changes in the operation conditions that combine them may also be suggested.
[0210] In addition, Fig.15In the server load prediction process shown, the calculation is repeated until the end condition is satisfied, so it takes time to output the final calculation result of the simulation. Generally, in the algorithm used to determine the connection relationship between the server and the device, it can be expected that it takes more calculation time when a calculation method that obtains a result closer to the optimal solution is used. Therefore, it can also be configured that, regarding the calculation method (calculation method 2 or calculation method 3 described in the first embodiment) for determining the connection between the server and the device by the algorithm, a plurality of calculation algorithms are prepared, such as an algorithm that can output a result in a shorter calculation time for the connection between the server and the device although the optimality is worse than the optimal solution, and an algorithm that can output a result closer to the optimal solution for the connection between the server and the device although the calculation time becomes longer, and the user can select the algorithm according to the length of time that can be waited until the result is obtained. In this case, the user can also specify it with the calculation method number, and it can also be configured that the user inputs the allowable time as a calculation condition, and the server load prediction system 40 selects the method used in the calculation according to the input allowable time and information about the process.
[0211] Fig.17 680 is a diagram showing an example of a simulation result screen 680. The simulation result screen 680 is Fig.15 An example of the calculation result of the simulation outputted by the output unit 140 in step S110 is displayed on the screen of the display device 19 .
[0212] Fig.17 The simulation result screen 680 includes a calculation method number 681, a server load 682, a connection relationship between the server and the device 683, and a device and an action condition 684 for changing the action. In the connection relationship between the server and the device 683, the correspondence between the server and the device in the connection relationship is displayed. In the device and the action condition 684 for changing the action, if there is a device for changing the action, the device and the action condition are displayed.
[0213] As described above, the server load prediction system 40 of the second embodiment not only calculates and outputs the server load based on the calculation condition 630 specified by the user through the execution of the server load prediction process, but also changes the action condition and system structure of the device or server and recalculates the server load when the calculation result of the server load exceeds a preset threshold value, thereby being able to find an action condition and system structure that can make the server load below the threshold value from the action change list 670. According to such a server load prediction system 40 of the second embodiment, not only can the server load of the server 23 connected to the device of the production process be predicted by simulation before the production process is started (or before the structure is changed in the production process) and the user is prompted, but also when there is a server whose server load exceeds the threshold value through simulation, the action condition and system structure for avoiding this situation can be prompted to the user.
[0214] In addition, in this embodiment, Figure 2 The production process system 20 shown is explained as an object for predicting server load, but for example, in the case where other servers (upper-level servers) are set up above server A (server 231) and server B (server 232) in the production process system 20, the server load prediction system 40 can also be configured to have the following simulation program, which implements a calculation method for making the server load of each server below a threshold by causing the upper-level server to bear part of the processing performed by servers A and B.
[0215] (3) Third Embodiment
[0216] In the first and second embodiments described above, the server load prediction systems 10 and 40 are described as being able to use the production process system 20 (see FIG. 1 ) in which the production process devices and servers are connected via the wired network 24. Figure 2 ) is used as an object to predict the server load. However, the production process system that the server load prediction system of the present invention can predict the server load is not limited to the production process system 20. Therefore, in the third embodiment, the production process system ( Fig.18 The production process system 50 shown is a server load prediction system for predicting server load.
[0217] Fig.18 2 is a block diagram showing an example of the structure of a working production process system 50. The production process system 50 is a system constructed for a production process assuming that wireless communication is used in at least some of the devices and the server. The same structures as the production process system 20 are given the same reference numerals and their descriptions are omitted.
[0218] like Fig.18As shown, the production process system 50 includes a server 231 (server A), a server 232 (server B), a production device 211 (device A), a production device 511 (device D), a camera 221 (camera A), a camera 521 (camera Da), a camera 522 (camera Db), a camera 523 (camera E), a wireless device 531 (wireless device Wa), a wireless device 532 (wireless device Wb), a wireless device 533 (wireless device Wc), a wireless device 534 (wireless device Wd), and a transport vehicle 542 that moves along a set route 541.
[0219] In the production process system 50, the cameras Da and Db capture the actions of the operator 25 using the device D, and the camera E is mounted on the transport vehicle 542 to capture the surroundings of the transport vehicle 542. The image captured by the camera Db is transmitted from the wireless device Wb to the wireless device Wa by wireless communication, and then transmitted from the wireless device Wa to the server A or the server B by wire. In addition, the image captured by the camera E is transmitted from the wireless device Wd to the wireless device Wc by wireless communication, for example, and then transmitted from the wireless device Wc to the server A or the server B by wire. In addition, in the communication between the transport vehicle 542 and the server A or the server B (for example, the transmission of the control signal to the transport vehicle 542), the wireless communication between the wireless devices Wd and the wireless devices Wc is also used.
[0220] In the production process system 50 in which at least a part of the devices use wireless communication in communication with the server as described above, the communication speed varies according to the state of the radio wave. In addition, as the transport vehicle 542 moves, the wireless device Wd mounted on the transport vehicle 542 also moves, so the communication speed varies according to the relative relationship with the wireless device Wc as the communication partner, and the communication may be interrupted by other devices. As described above, in wireless communication, the data reaching the server from the device changes over time, so it can be imagined that the server load imposed on the server A or server B by the device (e.g., camera Db, camera E, transport vehicle 542) that transmits data to server A or server B using wireless communication also changes over time.
[0221] Therefore, in this embodiment, when the server load prediction system (for example, it can be either the server load prediction system 10 or the server load prediction system 40) predicts the server load with the production process system 50 as the object, the following processing is performed.
[0222] As a first example, the server load prediction system simulates the change in communication volume generated by wireless communication when the transport vehicle 542 moves along the route 541, uses the calculated communication volume to calculate the change in server load, and uses the calculation result to apply the device requirement specifications of the transport vehicle 542 to the server.
[0223] In addition, as a second example, a network monitoring and analysis function for monitoring and analyzing the state of the network is installed in the server A or other devices. And, the above-mentioned network monitoring and analysis function is used to analyze the communication between the wireless device Wd and the server A when the transport vehicle 542 is traveling under similar conditions, such as when the transport vehicle 542 is traveling along the same route 541 at the same speed, or when the transport vehicle 542 is traveling along the same route as the route 541 (not shown) at the same speed. Based on the change in the amount of communication obtained as a result of the analysis, the change in the server load is calculated and saved in the server A or other devices. And, the server load prediction system uses the above-mentioned saved change in the server load for the device requirement specification.
[0224] Furthermore, in this embodiment, based on the changes in server load obtained through the simulation of the first example or the monitoring and analysis of the network in the second example, statistical processing is performed to obtain the cumulative degree distribution of the server load, thereby enabling the server load to be predicted based on the cumulative relative degree.
[0225] Fig.19 This is a diagram showing an example of cumulative degree distribution of server load. Fig.19 The graph 690 shown is a graph showing an example of changes in server load, with the cumulative relative degree as the vertical axis and the server load as the horizontal axis. According to such a graph 690, by specifying the cumulative relative degree, it is possible to predict the server load. Specifically, for example, the dotted line in the graph 690 indicates that the server load becomes less than the value indicated as the 75% value with a probability of 75%. Therefore, when calculating with a probability of 75% to meet such a condition, the server load prediction system only needs to calculate a value equivalent to the 75% value as the device requirement specification.
[0226] In addition, when the communication time varies according to the process and the route of the transport vehicle 542, or varies according to time zones such as morning, afternoon, and night, it is only necessary to find the cumulative degree distribution under each condition and calculate the server load value obtained from the cumulative degree distribution as the device requirement specification.
[0227] In addition, when the server load changes as described above, the user may input the value of the device requirement specification itself to the server load prediction system as the calculation condition 630. In addition, the server load prediction system may be configured to prepare in advance a table that summarizes the values that may be used for calculation, such as the 75% value, the 95% value, the average value, the median value, etc., and save the table in the server load prediction system, and by the user inputting "X% value" to the calculation condition 630, the value of the server load equivalent to the X% value recorded in the above table is calculated as the device requirement specification.
[0228] In addition, as described in the first embodiment, the camera (e.g., camera 22) installed in the production process is installed so that the server 23 can analyze the actions of the worker 25 in the image data taken by the camera. Depending on the analysis model used, the required server calculation amount, that is, the server load, may vary depending on the number of workers in the image data. Therefore, even if the same amount of data is regularly transmitted from the same device to the server 23 (231, 232), the server load may vary.
[0229] Therefore, in this embodiment, the above-mentioned method of calculating the device requirement specification based on the change of the server load can be applied to the device requirement specification of the camera, similarly to the case of the transport vehicle 542 connected to the server via wireless communication. In this case, the camera whose number of people in the image changes can also use the X% value, mean value, median value, maximum value and other values of the cumulative frequency distribution as the device requirement specification, similarly to wireless communication, and predict the server load based on these device requirement specifications.
[0230] Furthermore, in the present embodiment, when the server load varies according to the calculation conditions, it is preferable to configure so that the simulated conditions are also presented to the user together with the simulation results in the output of the output unit 140. For example, in the case of analyzing an image from a camera, since the server load varies according to the number of people in the image, the user may be presented with "In this calculation, the server load that is satisfied with a probability of 95% when the number of people in the image is 1 person with a probability of 80%, 2 people with a probability of 15%, and 3 people with a probability of 5% was calculated as the device requirement specification". In the present embodiment, by displaying the simulated conditions together as described above, the user can confirm the calculation conditions of the server load, which can help specify the calculation conditions later.
[0231] As described above, the present invention, by constructing a server load prediction system related to each embodiment, can predict the server load after the process is changed, and then make proposals for the system structure, the operating conditions of each device and the server so that the server load does not exceed the threshold. As a result, when the server load exceeds the threshold, the user can study the addition of servers, and when it is difficult to add servers or it takes time to add servers, the user can study reducing the server load by changing the operation of the device. In addition, as a result of predicting the server load, if it is known that there is a server with a smaller computing power than the server originally planned to be used, but better from the perspective of energy consumption, and the server with smaller computing power can also meet the required specifications of the device, the user can study replacement with a server with smaller computing power.
[0232] In addition, the present invention is not limited to the first to third embodiments described above, but includes various modifications and equivalent structures within the subject matter of the attached claims. For example, the above-mentioned embodiments are described in detail in order to explain the present invention in an easy-to-understand manner, and the present invention is not limited to necessarily having all the structures described. In addition, a part of the structure of a certain embodiment may be replaced with the structure of other embodiments. In addition, the structure of other embodiments may be added to the structure of a certain embodiment. In addition, with respect to a part of the structure of each embodiment, other structures may be added, deleted, or replaced. For example, in each embodiment, a structure in which there are multiple servers in a production process is described, but the present invention can also be applied when there is only one server.
[0233] In addition, the various structures, functions, processing units, processing mechanisms, etc. of the above-mentioned embodiments may also be implemented by hardware by designing a part or all of them, for example, using integrated circuits. In addition, the various structures, functions, etc. described above may also be implemented by software by interpreting and executing programs that implement the various functions by a processor. Information on programs, tables, files, etc. that implement the various functions may be saved in a recording device such as a memory, a hard disk, an SSD, or a recording medium such as an IC card, an SD card, or a DVD.
[0234] In the drawings, only those control lines and information lines that are considered necessary for explanation are shown, and not all control lines and information lines are necessarily shown in the product. In reality, it is considered that almost all structures are connected to each other.
Claims
1. A server load prediction system for predicting server loads of one or more servers connected to one or more devices installed in a production process, characterized in that: have: a device requirement specification storage unit storing, for each of the one or more devices, device requirement specifications including server load information and operation information, wherein the server load information indicates a load imposed on the server at a connection destination for controlling the device or processing information transmitted from the device, and the operation information indicates information related to an operation using the device; A server specification storage unit storing server specifications indicating capabilities maintained by each of the one or more servers; an input information preparation unit that receives input of calculation conditions including designation of a process for predicting the server load, and prepares input parameters required for executing a simulation for calculating the server load based on the calculation conditions, the device requirement specifications, and the server specifications; A server load calculation unit that calculates, based on the input parameters, by the simulation, the server load applied to the server at the connection destination by the device used in the step specified by the calculation condition; as well as an output unit, outputting the calculation result of the server load calculation unit, In the above input information production section, storing process device conversion information indicating the correspondence between each process of the production process and the device used in each process, The device used in the process specified by the calculation condition is determined based on the operation information and the process device conversion information, and the input parameter is created using the server load information corresponding to the device.
2. The server load prediction system according to claim 1, characterized in that: In the calculation conditions, in addition to specifying the process for predicting the server load, the connection relationship between the device and the server in the process is also specified. The server load calculation unit calculates the server load of the server based on the connection relationship specified by the calculation condition in the simulation.
3. The server load prediction system according to claim 1, characterized in that: In the calculation conditions, in addition to specifying the process for predicting the server load, when the device used in the process is specified, In the simulation, the server load calculation unit calculates the number of servers required to meet the device requirement specifications of the device based on the device specified by the calculation conditions, and calculates the server load of the server after determining the server connected to the device.
4. The server load prediction system according to claim 1, characterized in that: In the calculation conditions, in addition to specifying the process for predicting the server load, the specifications and number of the servers connected to the device used in the process are specified. In the above simulation, the server load calculation unit Based on the specifications and number of the server specified by the calculation conditions, the device connected to the server is determined within a range that does not exceed the capacity of the server, and then the server load of the server is calculated.
5. The server load prediction system according to claim 1, characterized in that: In the above device requirement specifications, The operation information includes information indicating a parsing model used by the server when parsing information transmitted from the device. The server load information indicates the load applied to the server by the analysis using the analysis model.
6. The server load prediction system according to claim 1, characterized in that: The device requirement specifications include the server load information or the operation information based on actual measurement data measured during the operation of each device.
7. The server load prediction system according to claim 1, characterized in that: In the above calculation conditions, a determination condition for the above server load is also set. The server load prediction system further includes a determination unit that determines whether the server load calculated by the server load calculation unit satisfies the determination condition. When the determination unit determines that the calculation result of the server load does not satisfy the determination condition, The input information production unit changes the input parameters. The server load calculation unit repeatedly performs a process of recalculating the server load based on the changed input parameters. When the determination unit determines that the calculation result of the server load satisfies the determination condition, The output unit outputs the calculation result of the server load.
8. The server load prediction system according to claim 7, characterized in that: In the above changes to the input parameters, For the device and the server whose connection relationship is assumed when calculating the server load determined not to satisfy the determination condition, at least one of the information included in the device requirement specification corresponding to the device or the server specification corresponding to the server is changed.
9. The server load prediction system according to claim 8, characterized in that: The output unit outputs information on the changed device requirement specification or the server specification together with the server load determined to satisfy the determination condition.
10. The server load prediction system according to claim 1, characterized in that: If the specifications of the above-mentioned device are changed during the above-mentioned production process, The numerical value obtained by statistically processing the above-mentioned changes is stored in the above-mentioned device requirement specification storage unit, The server load calculation unit calculates the server load using the numerical value stored in the device requirement specification storage unit.
11. The server load prediction system according to claim 10, characterized in that: In the case where at least a part of the above-mentioned devices in the above-mentioned production process are connected to the above-mentioned server via wireless communication, The server load calculation unit calculates the server load using a numerical value obtained by statistically processing a communication volume that varies through the wireless communication.
12. The server load prediction system according to claim 1, wherein: When the above calculation condition specifies the capacity allowed for the above server at runtime, The output unit outputs the server load calculated by the server load calculation unit in a manner that allows determination of excess or deficiency with respect to the capacity allowed for the server during operation specified by the calculation condition.
13. The server load prediction system according to claim 1, wherein: The output unit outputs, together with the server load calculated by the server load calculation unit, a connection relationship between the device and the server assumed in the calculation of the server load.
14. A server load prediction method, a system for predicting server loads of one or more servers connected to one or more devices installed in a production process, characterized in that: In the above system, For each of the one or more devices, device requirement specifications including server load information and operation information are stored, wherein the server load information indicates a load imposed on the server at a connection destination for controlling the device or processing information transmitted from the device, and the operation information indicates information related to an operation using the device. storing server specifications indicating capabilities maintained by each of the one or more servers; The above server load prediction method has the following features: An input information preparation step of accepting input of calculation conditions including designation of a process for predicting the server load, and preparing input parameters required for executing a simulation for calculating the server load based on the calculation conditions, the device requirement specifications, and the server specifications; A server load calculation step of calculating, based on the input parameters, the server load applied to the server at the connection destination by the device used in the step specified by the calculation condition through the simulation; as well as Output step, output the calculation result of the above server load calculation step, In the above input information preparation step, Based on the above-mentioned operation information and the process device transformation information representing the correspondence between each process of the above-mentioned production process and the device used in each process, the above-mentioned device used in the above-mentioned process specified by the above-mentioned calculation conditions is judged, and the above-mentioned input parameters are prepared using the above-mentioned server load information corresponding to the device.
Citation Information
Patent Citations
Cloud system, cloud server, edge server and user device
JP2019135578A
Load distributing method, load distributing program, and load distributing device
CN101611381A
Server control program, control server, virtual server distribution method
CN102449603A
Image forming apparatus, management apparatus and load control method
CN106470287A
Procedure management system
CN1517925A
