KR102304782B1 - Apparatus for scale calculation for constructing artificial intelligence and big data center method thereof - Google Patents

Abstract

Disclosed are a scale calculation apparatus for artificial intelligence and big data center construction and a method thereof. The present invention defines an architecture reference model, derives input factors, defines a calculation formula, then selects an architecture reference model based on an input value in accordance with a user input, calculates server specifications and quantities for each asset purpose, optimizes the configuration of a network, security, control, etc., calculates costs and priorities, adjusts influence factors in accordance with a priority definition, and designates priorities and detailed information to systematize a scale calculation method to eliminate the need for waiting for design results of field experts in an initial plan step, shorten the time required for design gathering and decision making, and provide a priority of a target model through repeated calculations and a plurality of cases in accordance with the priority to selectively proceed with tasks in accordance with acceptable budgets and space ranges and easily and quickly modify an initial plan through the adjustment of the input value if a problem occurs in the plan and a change is needed.

Description

Apparatus for scale calculation for constructing artificial intelligence and big data center method thereof

The present invention relates to an apparatus and method for calculating a scale for building an artificial intelligence and big data center, and in particular, defining an architecture reference model, deriving input factors, defining a calculation formula, and then based on input values according to user input Select an architectural reference model, calculate server specifications and quantity for each asset use, optimize the configuration of network, security, control, etc., calculate cost and priority, adjust influence factors according to priority definition, and prioritize And it relates to a sizing apparatus and method for building an artificial intelligence and big data center that stores detailed information.

An information system is one of the main systems constituting a management system. It is a system that quickly collects, processes, stores, retrieves, and presents information to create accurate information to make a confirmed business decision. The establishment of a comprehensive information processing system centered on

Despite the urgency of constructing the artificial intelligence and big data centers that make up these information systems, considerable time delays occur in the initial planning stage, impeding the implementation, and budgets must be secured in the initial planning stage. In order to secure the budget, it is necessary to know the hardware and software construction cost, network and security construction cost, and infrastructure construction cost, but it is difficult to design an integrated initial plan because experts who can design these are divided by field.

In addition, in addition to the delay time that occurs while assembling the designs in each field, the design changes several times in the face of many cost and spatial constraints, such as exceeding the acceptable budget or the number of facilities exceeding the available space. is being repeated.

In other words, in order to secure an agile response in the present age, the initial plan of artificial intelligence and big data centers must be quickly planned in consideration of productivity and efficiency, but it is difficult to establish an initial plan because it requires the input of many experts and complicated procedures. .

Korean Patent Registration No. 10-2081727 [Title: Service providing system and method for service optimization operation management and decision support]

An object of the present invention is to define an architecture reference model, derive input factors, define a calculation formula, select an architecture reference model based on input values according to user input, calculate server specifications and quantities for each asset use, and network A sizing device for building an artificial intelligence and big data center that optimizes the configuration of , security, control, etc., calculates cost and priority, adjusts influence factors according to priority definition, and stores priority and detailed information and to provide a method therefor.

Another object of the present invention is to provide an apparatus and method for estimating a scale for building an artificial intelligence and big data center that provides an initial draft when constructing an artificial intelligence and big data center based on an open source.

An apparatus for calculating a scale for building an artificial intelligence and big data center according to an embodiment of the present invention includes a display unit for displaying an information input screen; and one or more input values corresponding to one or more input items requiring user setting among a plurality of input values corresponding to a plurality of input items required for size calculation for artificial intelligence and big data center construction according to a user input on the information input screen any one of a plurality of architecture reference models set in advance based on at least one of whether network redundancy configuration, network network separation, and server virtualization included in the input value related to the network configuration among the received input values. Select a reference model, calculate the tpmC for each asset use based on the input value related to the hardware capacity and the input value related to the operating period among the received input values, and use the calculated tpmC to determine the CPU core type, number of cores and Calculate the CPU specifications for each asset use based on the clock speed, the calculated tpmC for each asset use, the calculated CPU specifications for each asset use, an input value related to the number of software users among the received input values, an input value related to the number of affiliated organizations, and a control unit for calculating server capacity based on an input value related to data capacity and the selected architecture reference model, and selecting server specifications for each asset use based on the calculated server capacity.

As an example related to the present invention, the plurality of input items are items necessary for size calculation for artificial intelligence and big data center construction, and are preset according to a designer's design, and the user setting among the plurality of input values is The remaining input values excluding one or more necessary input values may have default values set in advance according to a designer's design.

As an example related to the present invention, the one or more input values that require user setting include an input value related to hardware capacity, an input value related to the number of software users, an input value related to the number of linked organizations, an input value related to data capacity, and network configuration. It may include at least one of an input value related to , an input value related to quality characteristics, and an input value related to an operating period.

As an example related to the present invention, the input values related to the hardware capacity include the number of concurrent users of the external portal, the number of concurrent users of the internal portal, the annual increase rate of the external portal concurrent users, the annual increase rate of the internal portal concurrent users, the number of tasks per external portal user, and the internal at least one of the number of tasks per portal user, the number of transactions per external portal task, the number of transactions per internal portal task, the number of operations per external portal user, the number of operations per user of the internal portal, an external portal application load, and an internal portal application load; The input value related to the number of people using the software includes the number of people using the business software, the input value related to the number of linked organizations includes at least one of the number of linked external organizations and the number of linked internal organizations, and the input related to the data capacity The value includes at least one of daily data collection and processing capacity and data storage capacity, and the input value related to the network configuration includes at least one of whether network redundancy is configured and whether or not network is divided, and the quality characteristics and The related input value may include at least one of performance importance, availability importance, and security importance, and the input value related to the operation period may include an operation period.

As an example related to the present invention, the control unit applies a quality characteristic weight to the selected server specification for each asset use based on an input value related to a quality characteristic among the selected server specification for each asset use and the received input value, Calculating the quantity for each asset use based on the server specification for each asset use to which the quality characteristic weight is applied, and managing the network configuration, security configuration and control configuration based on the server specification for each selected asset use and the calculated quantity for each asset use, For configuration, security configuration, and control configuration, calculate the cost according to the server specifications for each asset use and the quantity by asset use, and calculate the size calculation results based on information on changes in the importance of quality characteristics and changes in the configuration of big data clusters. Calculating a priority, defining a priority for a sizing result based on the priority of the calculated sizing result, and inputting at least one of the input values according to the priority for the defined sizing result. Adjusting the value, reflecting the adjusted at least one input value, repeating the process of adjusting the influence factor until the priority list corresponds to the preset number of priority lists, and the server specification for each selected asset use Based on the priority list reflecting the server configuration for the network, security and control configuration, and the adjusted input value, it is possible to store the sizing result including the priority and detailed information in the storage unit.

The method for calculating the scale for building an artificial intelligence and big data center according to an embodiment of the present invention is a plurality of necessary for estimating the scale for building an artificial intelligence and big data center according to a user input on an information input screen displayed on a display unit by a control unit. receiving one or more input values corresponding to one or more input items requiring user setting from among a plurality of input values corresponding to the input items of ; Any one of a plurality of architectural reference models set in advance based on at least one of whether network redundancy configuration, network network separation, and server virtualization is included in the input value related to the network configuration among the received input values selecting an architecture reference model; calculating, by the control unit, tpmC for each asset use based on an input value related to a hardware capacity and an input value related to an operating period among the received input values; calculating, by the controller, the CPU specifications for each asset use by inferring the core type, number of cores, and clock speed of the CPU using the calculated tpmC; By the control unit, the calculated tpmC for each asset use, the calculated CPU specification for each asset use, an input value related to the number of software users among the received input values, an input value related to the number of linked organizations, an input value related to data capacity, and the calculating server capacity based on the selected architecture reference model; and selecting, by the control unit, server specifications for each asset use based on the calculated server capacity.

As an example related to the present invention, the step of applying, by the control unit, a quality characteristic weight to the selected server specification for each asset use based on an input value related to a quality characteristic among the selected server specification for each asset use and the received input value ; calculating, by the control unit, the quantity for each asset use based on the server specification for each asset use to which the quality characteristic weight is applied; managing, by the control unit, a network configuration, a security configuration, and a control configuration based on the server specification for each selected asset use and the calculated quantity for each asset use; calculating, by the control unit, a cost according to a specification of a server for each asset use and a quantity for each asset use for a network configuration, a security configuration, and a control configuration; calculating, by the control unit, the priority of the scale calculation result based on the information on the change in the importance of quality characteristics and the information on the big data cluster configuration change; defining, by the control unit, a priority for a scale calculation result based on the priority for the calculated scale calculation result; The control unit adjusts at least one input value from among the input values according to the priority for the defined scale calculation result, and reflects the adjusted at least one input value, so that the priority list is set in advance repeating the process of adjusting the influence factor until it corresponds to the number of lists; And, by the control unit, the server specification for each asset use, the network by asset use, the server configuration for the security and control configuration, and the scale calculation including the priority and detailed information based on the priority list in which the adjusted input value is reflected The method may further include storing the result in a storage unit.

As an example related to the present invention, the step of calculating the server capacity may include calculating the big data platform server capacity, the artificial intelligence service server capacity, the portal service server capacity, the linked service server capacity, and the service server capacity related to security/control, respectively. have.

As an example related to the present invention, the step of calculating the priority for the scale calculation result may include: setting the first derived scale calculation result as the first priority; a process of calculating a priority score by reflecting the information on the change in the importance of quality characteristics; a process of calculating a priority score by reflecting information related to a change in the configuration of a big data cluster; a process of collecting the priority score reflecting the information on the change in the importance of quality characteristics calculated and the priority score reflecting the information related to the calculated big data cluster configuration change, and storing the T/F value; And it may include a process of determining the priority by reflecting the total score and removing duplicates.

As an example related to the present invention, when a priority menu is selected from among a plurality of menus displayed on the display unit, a scale calculation result calculated based on an input value previously received in response to the selected priority menu by the control unit displaying a list of priorities according to the above on the display unit; When a resource list menu is selected from among the plurality of menus displayed on the display unit, the display unit displays a resource list according to a scale calculation result calculated based on an input value previously received in response to the selected resource list menu by the control unit. to mark on; When a server specification menu is selected from among the plurality of menus displayed on the display unit, the display unit displays, by the control unit, a server specification according to a scale calculation result calculated based on an input value previously received in response to the selected server specification menu. to mark on; and when the infrastructure configuration menu is selected from among the plurality of menus displayed on the display unit, the infrastructure configuration according to the scale calculation result calculated based on the previously received input value in response to the selected infrastructure configuration menu by the control unit. The method may further include displaying on the display unit.

The present invention defines an architecture reference model, derives input factors, defines a calculation formula, selects an architecture reference model based on input values according to user input, calculates server specifications and quantity for each asset use, network, security By optimizing the configuration of , control, etc., calculating costs and priorities, adjusting influence factors according to priority definition, and storing priorities and detailed information, the sizing method is systematized in each field in the initial planning stage. There is no need to wait for design results from experts, shortening the time required for design collection and decision-making, and providing a number of cases according to the priority and priorities of the target model through iterative calculations so that the budget and space can be accommodated. It has the effect of selectively promoting the task according to the schedule, and if a change is required due to a setback in the plan, the initial draft can be easily and quickly modified by adjusting the input value.

In addition, the present invention provides an initial draft when constructing an artificial intelligence and big data center based on open source, so that there is no separate software purchase cost, and the architecture can be handled in a controllable range, so that the budget calculation does not significantly deviate from the error range This is possible, and it has the effect of composing an initial plan close to the goal of building artificial intelligence and big data centers by appropriately deriving input variables related to business strategy.

1 is a block diagram showing the configuration of a scale calculation device for building an artificial intelligence and big data center according to an embodiment of the present invention.
2 is a diagram illustrating an example of a classification system of a scale calculation target according to an embodiment of the present invention.
3 is a diagram illustrating an example of a correspondence relationship between a classification system and hardware, software, and infrastructure according to an embodiment of the present invention.
4 is a diagram illustrating an example of a research range of a scale calculating device according to an embodiment of the present invention.
5 is a diagram showing an example of hardware classification of artificial intelligence and big data center according to an embodiment of the present invention.
6 is a diagram illustrating an example of a commercial software list for building artificial intelligence and big data centers according to an embodiment of the present invention.
7 is a diagram illustrating an example of classification of infrastructure of an artificial intelligence and big data center according to an embodiment of the present invention.
8 is a diagram illustrating an example of architectural reference model types and relationships according to an embodiment of the present invention.
9 is a diagram illustrating an example of a business architecture reference model according to an embodiment of the present invention.
10 is a diagram illustrating an example of artificial intelligence and big data center work types and roles according to an embodiment of the present invention.
11 is a diagram illustrating an example of an application architecture reference model according to an embodiment of the present invention.
12 is a diagram illustrating an example of a data architecture reference model according to an embodiment of the present invention.
13 to 20 are diagrams illustrating examples of infrastructure architecture reference models #1 to #8 according to an embodiment of the present invention.
21 is a diagram illustrating an example of an infrastructure architecture reference model according to an embodiment of the present invention.
22 is a view showing an example of a technical standard detailed item for a standard rack according to an embodiment of the present invention.
23 is a diagram illustrating an example of an influence factor identification result for an architecture reference model according to an embodiment of the present invention.
24 is a diagram illustrating an example of a method for calculating CPU/memory specifications for each server purpose according to an embodiment of the present invention.
25 is a diagram illustrating an example of dividing a fixed factor and a variable factor for CPU/memory calculation according to an embodiment of the present invention.
26 is a diagram illustrating an example of classification of variable factor items for CPU/memory calculation according to an embodiment of the present invention.
27 is a diagram illustrating a comparison example of a difference between a maximum value and a minimum value for each correction item according to an embodiment of the present invention.
28 is a diagram illustrating an example of tpmC change according to an application load input value according to an embodiment of the present invention.
29 is a diagram illustrating a relationship between input items, calculation expressions, and outputs according to an embodiment of the present invention.
30 is a diagram illustrating an example of a tpmC calculation process for an input item according to an embodiment of the present invention.
31 and 32 are diagrams illustrating examples of the relationship between tpmC and CPU specifications according to an embodiment of the present invention.
33 is a diagram illustrating a scatter diagram of CPU specifications according to tpmC according to an embodiment of the present invention.
34 is a diagram illustrating an example of a clock speed of a CPU to which an actual value, a linear equation, and a Newton interpolation method are applied, respectively, according to an embodiment of the present invention.
35 is a diagram illustrating an example of a priority score calculated according to a change in the importance of quality characteristics according to an embodiment of the present invention.
36 is a diagram illustrating an example of a use case diagram design result expressing user-side requirements according to an embodiment of the present invention.
37 is a diagram illustrating an example of a sequence diagram design result according to an embodiment of the present invention.
38 is a diagram illustrating an example of a communication diagram design result according to an embodiment of the present invention.
39 is a diagram illustrating an example of a relationship between an input item and an output screen according to an embodiment of the present invention.
40 is a diagram illustrating an example of an activity diagram design result according to an embodiment of the present invention.
41 is a diagram illustrating an example of a state machine diagram design result according to an embodiment of the present invention.
42 is a diagram illustrating an example of logical data modeling (logical ER diagram) according to an embodiment of the present invention.
43 is a diagram illustrating an example of physical data modeling (physical ER diagram) according to an embodiment of the present invention.
44 is a diagram illustrating an example of a verification range of a scale calculating device according to an embodiment of the present invention.
45 to 46 are flowcharts illustrating a scale calculation method for constructing an artificial intelligence and big data center according to an embodiment of the present invention.
47 is a diagram illustrating an example of an information input screen of a scale calculating device according to an embodiment of the present invention.
48 is a diagram illustrating an example of a priority list screen of a scale calculating apparatus according to an embodiment of the present invention.
49 is a diagram illustrating an example of a resource list screen of an apparatus for calculating a scale according to an embodiment of the present invention.
50 is a diagram illustrating an example of a server specification screen of a scale calculating apparatus according to an embodiment of the present invention.
51 is a diagram illustrating an example of an infrastructure configuration screen of a scale calculation apparatus according to an embodiment of the present invention.

It should be noted that the technical terms used in the present invention are only used to describe specific embodiments, and are not intended to limit the present invention. In addition, the technical terms used in the present invention should be interpreted as meanings generally understood by those of ordinary skill in the art to which the present invention belongs, unless otherwise defined in particular in the present invention, and excessively comprehensive It should not be construed in the meaning of a human being or in an excessively reduced meaning. In addition, when the technical term used in the present invention is an incorrect technical term that does not accurately express the spirit of the present invention, it should be understood by being replaced with a technical term that can be correctly understood by those skilled in the art. In addition, general terms used in the present invention should be interpreted as defined in advance or according to the context before and after, and should not be interpreted in an excessively reduced meaning.

Also, as used herein, the singular expression includes the plural expression unless the context clearly dictates otherwise. In the present invention, terms such as "consisting of" or "comprising" should not be construed as necessarily including all of the various elements or several steps described in the invention, and some of the elements or some steps may not be included. It should be construed that it may further include additional components or steps.

In addition, terms including ordinal numbers such as first, second, etc. used in the present invention may be used to describe the components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, a first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the accompanying drawings, but the same or similar components are assigned the same reference numerals regardless of reference numerals, and redundant description thereof will be omitted.

In addition, in the description of the present invention, if it is determined that a detailed description of a related known technology may obscure the gist of the present invention, the detailed description thereof will be omitted. In addition, it should be noted that the accompanying drawings are only for easy understanding of the spirit of the present invention, and should not be construed as limiting the spirit of the present invention by the accompanying drawings.

1 is a block diagram showing the configuration of an apparatus 100 for calculating a scale for constructing an artificial intelligence and big data center according to an embodiment of the present invention.

As shown in FIG. 1 , the apparatus 100 for calculating the scale for building an artificial intelligence and big data center includes a communication unit 110 , a storage unit 120 , a display unit 130 , a voice output unit 140 , and a control unit 150 . ) is composed of Not all of the components of the sizing apparatus 100 shown in FIG. 1 are essential components, and the sizing apparatus 100 may be implemented by more components than the components shown in FIG. 1 , and less The scale estimating device 100 may also be implemented by the components.

The scale calculating device 100 is a smart phone, a portable terminal, a mobile terminal, a foldable terminal, a personal digital assistant (PDA), a PMP ( Portable Multimedia Player) terminal, telematics terminal, navigation terminal, personal computer, notebook computer, slate PC, tablet PC, ultrabook, wearable device (Including Wearable Device, for example, watch-type terminal (Smartwatch), glass-type terminal (Smart Glass), HMD (Head Mounted Display), etc.), Wibro (Wibro) terminal, IPTV (Internet Protocol Television) terminal, smart TV, It can be applied to various terminals such as a digital broadcasting terminal, an AVN (Audio Video Navigation) terminal, an A/V (Audio/Video) system, a flexible terminal, and a digital signage device.

The communication unit 110 communicates with any internal component or at least one external terminal through a wired/wireless communication network. In this case, the external arbitrary terminal may include a server (not shown), another terminal (not shown), and the like. Here, as wireless Internet technologies, wireless LAN (WLAN), DLNA (Digital Living Network Alliance), WiBro (Wireless Broadband: Wibro), Wimax (World Interoperability for Microwave Access: Wimax), HSDPA (High Speed Downlink Packet Access) ), High Speed Uplink Packet Access (HSUPA), IEEE 802.16, Long Term Evolution (LTE), Long Term Evolution-Advanced (LTE-A), Wireless Mobile Broadband Service (WMBS), etc. In this case, the communication unit 110 transmits and receives data according to at least one wireless Internet technology within a range including Internet technologies not listed above. In addition, short-range communication technologies include Bluetooth, Radio Frequency Identification (RFID), Infrared Data Association (IrDA), Ultra Wideband (UWB), ZigBee, and Near Field Communication (NFC). , Ultra Sound Communication (USC), Visible Light Communication (VLC), Wi-Fi (Wi-Fi), Wi-Fi Direct (Wi-Fi Direct), etc. may be included. In addition, the wired communication technology may include Power Line Communication (PLC), USB communication, Ethernet, serial communication, optical/coaxial cable, and the like.

Also, the communication unit 110 may mutually transmit information with an arbitrary terminal through a Universal Serial Bus (USB).

In addition, the communication unit 110 is a technology standard or communication method for mobile communication (eg, GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), CDMA2000 (Code Division Multi Access 2000), EV -DO (Enhanced Voice-Data Optimized or Enhanced Voice-Data Only), WCDMA (Wideband CDMA), HSDPA (High Speed Downlink Packet Access), HSUPA (High Speed Uplink Packet Access), LTE (Long Term Evolution), LTE-A (Long Term Evolution-Advanced, etc.) transmits and receives radio signals to and from the base station, the server, and the other terminal on a mobile communication network constructed according to the same.

In addition, the communication unit 110 transmits, under the control of the control unit 150, the size calculation result, etc. to the server, the other terminal, and the like.

The storage unit 120 stores various user interfaces (UIs), graphic user interfaces (GUIs), and the like.

In addition, the storage unit 120 stores data and programs necessary for the operation of the scale calculating device 100 .

That is, the storage unit 120 may store a plurality of application programs (application programs or applications) driven in the scale estimating device 100 , data for the operation of the scale estimating device 100 , and commands. have. At least some of these applications may be downloaded from an external server via wireless communication. In addition, at least some of these application programs may exist on the scale estimator 100 from the time of shipment for the basic function of the scale estimator 100 . On the other hand, the application program may be stored in the storage unit 120 , installed in the scale calculating device 100 , and driven to perform the operation (or function) of the scale calculating device 100 by the control unit 150 . have.

In addition, the storage unit 120 is a flash memory type (Flash Memory Type), a hard disk type (Hard Disk Type), a multimedia card micro type (Multimedia Card Micro Type), a card type memory (eg, SD or XD) memory, etc.), magnetic memory, magnetic disk, optical disk, RAM (Random Access Memory: RAM), SRAM (Static Random Access Memory), ROM (Read-Only Memory: ROM), EEPROM (Electrically Erasable Programmable Read-Only Memory), It may include at least one storage medium among Programmable Read-Only Memory (PROM). In addition, the scale calculating apparatus 100 may operate a web storage that performs a storage function of the storage unit 120 on the Internet, or may operate in connection with the web storage.

In addition, the storage unit 120 stores the scale calculation result and the like under the control of the control unit 150 .

The display unit (or display unit) 130 may display various contents such as various menu screens using a user interface and/or a graphic user interface stored in the storage unit 120 under the control of the control unit 150 . have. Here, the content displayed on the display unit 130 includes various text or image data (including various information data) and a menu screen including data such as icons, list menus, and combo boxes. Also, the display unit 130 may be a touch screen.

In addition, the display unit 130 includes a liquid crystal display (LCD), a thin film transistor liquid crystal display (TFT LCD), an organic light-emitting diode (OLED), and a flexible display. It may include at least one of a flexible display, a 3D display, an e-ink display, and a Light Emitting Diode (LED).

In addition, the display unit 130 displays the scale calculation result and the like under the control of the control unit 150 .

The audio output unit 140 outputs audio information included in a signal processed by the control unit 150 . Here, the audio output unit 140 may include a receiver, a speaker, a buzzer, and the like.

In addition, the voice output unit 140 outputs a guide voice generated by the control unit 150 .

In addition, the audio output unit 140 outputs audio information (or sound information) corresponding to the scale calculation result displayed on the display unit 130 under the control of the controller 150 .

The controller or microcontroller unit (MCU) 150 executes an overall control function of the scale estimating device 100 for building the artificial intelligence and big data center.

In addition, the control unit 150 executes the overall control function of the scale calculating apparatus 100 by using the program and data stored in the storage unit 120 . The controller 150 may include a RAM, ROM, CPU, GPU, and bus, and the RAM, ROM, CPU, GPU, etc. may be connected to each other through a bus. The CPU may access the storage unit 120 and perform booting using the O/S stored in the storage unit 120 , and use various programs, contents, data, etc. stored in the storage unit 120 . to perform various operations.

The informatization paradigm following the 4th industrial revolution emphasizes hyper-connectivity, super-intelligence, and super-convergence through the convergence of information and communication technologies. In each industry, intelligent services and technology infrastructure are attracting attention, and with the advent of the term ICBMA (Internet of Things & Cloud Computing & Bigdata & Mobile & Artificial Intelligence), the importance of Internet of Things, Cloud, Big Data, Mobile, and Artificial Intelligence is further increasing.

The ICBMA competency has become an important measure for the nation's operation, provision of services to the public, and global competitiveness of companies. As the necessity of the center is increasing day by day, the continuous construction of the center and related investment are rapidly increasing.

The AI/big data center (or artificial intelligence and big data sensor) is a data center built to utilize both artificial intelligence technology and big data technology, and is different from previous big data centers in terms of configuration. Whereas the previous big data centers often consisted of connecting low-spec servers, AI/big data centers are characterized by requiring higher-spec hardware for big data learning and execution of artificial intelligence algorithms. Recently, the urgency of using artificial intelligence and big data has been highlighted due to the aftermath of COVID-19, and the construction of AI and big data centers in various industries is emerging as a task in order to respond to the post-corona era.

Accordingly, the government is promoting the construction of data centers that can process and store big data and implement artificial intelligence services for major industries such as automobiles and energy. is planning

The scale calculation apparatus 100 according to an embodiment of the present invention provides a scale calculation system that presents an appropriate initial plan when constructing an AI/big data center limited to an open source basis.

Since the open source base does not incur a separate software purchase cost and the architecture can be handled within a controllable range, budget calculations that do not deviate significantly from the error range are possible. It is possible to configure an initial plan that is close to the data center construction goal.

Open source is a term that refers to source code or software that is released free of charge, and is called open software in Korea.

The open source has a characteristic that anyone can improve and redistribute the source code by releasing the source code, and in the Open Source Initiative (OSI), free distribution (for example, some or all of the software is freely irrespective of economic compensation). must be redistributable), open source code (e.g., the source code must be provided in a path accessible to those who wish to use it), allow derivative works (e.g., allow derivative work, and the output must be under the same license characteristics), restriction on source code modification (for example, it should be possible to add a restriction on source code modification within the license), and non-discrimination against individuals or groups (for example, the use of the software is not limited to individuals or groups) not to discriminate against), no restrictions on the field of use (e.g., no restrictions on the field of use of the software source code), distribution of licenses (e.g., it can be distributed to anyone without a separate license approval or transfer process) maintain uniformity in license application (e.g., it must have the same effect at all stages of distribution as it was originally distributed), and comprehensive acceptance of other licenses (e.g., do not restrict software distributed with open source) must not), and the technical neutrality of the license (for example, an open software license must be technically neutral).

Hadoop is an open source framework that started as a sub-project of Apache Lucene and was promoted to the main project as a solution for storing and processing big data.

The scale calculation apparatus 100 according to an embodiment of the present invention applies the Hadoop Ecosystem.

In addition, the scale calculation apparatus 100 according to the embodiment of the present invention does not have to wait for the design results of experts in each field in the initial planning stage as the scale calculation method is systemized, and the time required for design collection and decision-making can be shortened. have. In addition, through iterative calculations, the priority of the target model and the number of cases according to the priority can be provided, so that the task can be selectively pursued according to the acceptable budget and space range If this is necessary, the initial draft can be easily and quickly modified by adjusting the input value.

In addition, as shown in Figure 2, the control unit 150 is a big data platform, artificial intelligence service, portal service, linked service, security / control system construction, hardware and software infrastructure construction, commercial software introduction, infrastructure purchase and Eight systems of construction are used to classify the objects of size calculation.

That is, as shown in FIG. 2 , the control unit 150 collects, processes, stores, analyzes, and visualizes structured, semi-structured, and unstructured data. , the scale of artificial intelligence services that provide prediction and recommendation services, the scale of portal services to enable access and convenience to internal and external users, the scale of linked services that integrate and link internal and external organizations, and the scale of data transmission and reception The scale of building a security/control system for network construction and information protection, the scale of hardware that can install and operate software, the scale of commercial software for services that are difficult to use open source, and facilities for integrating and operating physical equipment Sizing targets are classified according to purchase and data center construction scale.

In addition, as shown in FIG. 3 , the control unit 150 manages eight types of scale calculation targets as hardware, software, and infrastructure.

The hardware includes servers, storage, network and security equipment necessary for AI/big data platform and services, and the software includes an operating system, open source software including Hadoop ecosystem, and other commercial software. The software development, service operation organization composition, and personal information protection construction are excluded from the scope due to significant deviations depending on business characteristics, and resource composition for providing internal/external portal services as a basic specification is exceptionally included. Here, the artificial intelligence (AI) refers to a technology in which human learning ability, reasoning ability, perception ability, etc. are realized by a computer program. In addition, the big data is data generated in a digital environment, and represents data that has a large scale, a short generation cycle, and can handle various types of data.

In addition, the infrastructure includes building equipment, communication/power equipment, air conditioning equipment, firefighting equipment, and security/control equipment, and covers the space in which hardware and software are installed and operated. Therefore, the infrastructure includes access control facilities, CCTV, etc. installed inside the operating space, but excludes access control for the entire building, communication room, machine room, etc., and is physically separated into two or more centers However, it does not include the composition of a Disaster Recovery Center.

As shown in FIG. 4 , the control unit 150 may consider (or set) the research scope of the scale calculating apparatus 100 .

As shown in FIG. 5 , the control unit 150 defines hardware classification of artificial intelligence and big data centers.

In the case of the above software, since the scope of using open source is used, in principle, a separate cost for software introduction is not required. However, special purpose software such as security solutions and linked solutions has no choice but to use commercial software due to the characteristics of the product, and as shown in FIG. 6 , the control unit 150 applies commercial software to the corresponding item. .

As shown in FIG. 7 , the control unit 150 defines the classification of infrastructure including construction, communication, power, air conditioning, firefighting, control, and security facilities.

When the structure, function, and performance requirements of the center are presented, the scale calculation means converting them into equipment requirements, and it can be estimated through methods such as formula calculation, reference method, and simulation method. Here, the formula calculation method is a method of presenting a calculation result using a calculation formula for calculating the scale, the reference method is a method of comparative calculation based on data such as similar cases, and the simulation method applies a load to the target and performs simulation This is a method of estimating the size.

In addition, the control unit 150 defines an architecture reference model according to the hardware, the software, and the infrastructure.

That is, as shown in FIG. 8 , the control unit 150 is a business architecture reference model from the business side, application side, data side, infrastructure side, and infrastructure side based on the open source-based AI/big data center construction case. , application architecture reference model, data architecture reference model, infrastructure architecture reference model, and infrastructure architecture reference model are defined respectively. Here, the infrastructure architecture reference model may correspond to a technical architecture of an enterprise architecture, and the technical architecture is a structure affected by the business architecture reference model, the application architecture reference model, and the data architecture reference model.

As shown in FIG. 9 , the business architecture reference model is a reference model in which the organization and tasks of the center are classified and defined as a whole.

As shown in FIG. 10 , the business architecture reference model may assign roles and responsibilities according to tasks in detail.

11, the application architecture reference model is a reference model that classifies and defines the functions of the application program. The application programs required in the AI/big data center can be largely divided into portal services, artificial intelligence services, big data services, linked services, and information protection services. Construct the application architecture reference model.

The detailed list of open source software included in the application system may be changed, and the basic configuration of the overall system may be configured similarly to the configuration shown in FIG. 11 .

However, since there are products that are provided for a fee among open source software, the control unit 150 may calculate the scale by reflecting the relevant matters when planning product introduction.

As shown in FIG. 12 , the data architecture reference model is a reference model for building an open source-based big data platform.

In addition, the data architecture reference model has a lifecycle of big data collection, storage, processing, analysis, and visualization, and a big data cluster that can be configured in individual cluster units and a big data cluster support server group that centrally provides processing and support services can be divided into

The big data cluster refers to a cluster of Hadoop ecosystem resources, such as name nodes, data nodes, and real-time collection nodes, which are highly variable according to data collection and storage capacity, and the big data cluster support server is It means the metadata/quality management server, data mart, visualization server, etc. that were used.

In the case of AI and big data centers, data for AI learning should be considered. Such data is stored in the big data cluster, and the machine learning server for analysis also exists in the big data cluster.

The infrastructure architecture reference model is a reference model that identifies hardware and software resources for building an application architecture and data architecture, and presents a network and security configuration.

In addition, the infrastructure architecture reference model is influenced by other reference models and has various structures for each business.

Therefore, it is very important to define how to design and configure the infrastructure architecture, and the control unit 150 considers the number of possible cases in the following [Table 1], and considers 8 types of infrastructure architecture reference models reflecting major specific situations. is defined as in FIGS. 13 to 20 .

Reference model number characteristic duplication network separation Server virtualization Reference Model #1 simplest configuration Reference model #2 Simplest configuration + server virtualization ○ Reference Model #3 Redundancy of network and security equipment ○ Reference model #4 Redundancy configuration + server virtualization ○ ○ Reference model #5 Separation of internal and external networks ○ Reference model #6 Network Separation + Server Virtualization ○ ○ Reference model #7 Redundancy configuration + network separation ○ ○ Reference model #8 Redundancy configuration + network separation + server virtualization ○ ○ ○

The reference models #1 and #2 shown in FIGS. 13 and 14 have the simplest and lowest cost structure, the reference model #1 is a structure in which all servers are physically configured, and the reference model #2 is a structure in which virtualization is applied to a server capable of server virtualization.

In addition, the reference model #3 shown in FIG. 15 is a network-duplexed structure of the reference model #1, and the reference model #5 shown in FIG. 17 is a network structure obtained by dividing the network of the reference model #1. am.

In addition, the reference models #4 and #6 shown in FIGS. 16 and 18 are structures in which the reference models #3 and #5 are implemented as virtual servers, respectively.

In addition, the reference models #7 and #8 shown in FIGS. 19 and 20 are structures in which both network redundancy and network separation are applied to the reference models #1 and #2, which are adopted when both availability and security are important. is a model Here, the reference models #7 and #8 are mostly applied when constructing the center.

As shown in FIG. 21 , the infrastructure architecture reference model is an architecture reference model for providing an environment in which hardware and software can be installed and operated.

The infrastructure architecture reference model according to an embodiment of the present invention is TIA-942 Telecommunications Infrastructure Standard for Data Centers (American Telecommunications Association IDC Standard), Green Data Center Construction Guidelines (TTA, 2010. 12), Integrated Information and Communication Facility Protection Guidelines (Korea Communications Commission, No. 2008-15), premises communication facility standards (Korea Information and Communication Corporation), etc., refer to guidelines and legal matters to define Tier 3 as the standard.

In addition, as shown in FIG. 22, the standard rack to be placed in the server room in the AI/big data center is defined to comply with the relevant technical standards.

In this way, the control unit 150 may analyze the requirements required for the open source-based AI/big data center from a plurality of construction cases, and define each architecture reference model according to the analysis result.

In order to estimate the construction scale of the AI/big data center using the architectural reference model according to the embodiment of the present invention, it is necessary to identify and adjust the factors affecting the scale.

As shown in FIG. 23 , the architecture reference model that is changed according to an input value among the business architecture reference model, the application architecture reference model, the data architecture reference model, the infrastructure architecture reference model, and the infrastructure architecture reference model is the These are the infrastructure architecture reference model and the infrastructure architecture reference model, and these two architecture reference models can be derived by knowing the hardware specification and quantity, the software specification and quantity, and the network/security configuration.

Therefore, the control unit 150 grasps CPU/memory capacity and data capacity, and additionally calculates the importance of quality characteristics such as performance, availability, and security and the quantity of commercial software, and is suitable for business goals among various target models. Close models can be printed.

That is, the control unit 150 selects (or sets) an input factor for calculating CPU/memory capacity through the following process.

In addition, the control unit 150 calculates the scale (or CPU/memory capacity) for CPU and memory by applying (or based on) the 'information system hardware sizing guideline' established by the Korea Information and Communication Technology Association (TTA). do. At this time, as shown in FIG. 24 , in order to provide AI and big data services, the control unit 150 needs to consider the core type, the number of cores, and the clock speed in addition to the traditional scale calculation method, so it depends on the server purpose. Distinguish the calculation method.

In addition, as shown in FIG. 25 , the control unit 150 includes a variable factor (or variable factor) and a fixed factor ( or fixed factors).

In addition, in the case of a fixed factor that is not classified as a variable factor, the control unit 150 sets it as a preset fixed value (or default value/constant), or classifies it as a fixed value according to the characteristics of the AI/big data center (or set).

In addition, as shown in FIG. 26 , the control unit 150 classifies the items classified as variable factors into items requiring user input again and items that can be automatically calculated by the corresponding scale calculating device 100 . .

As shown in FIG. 26 , items that require user input among the variable factors are the number of concurrent users, the number of transactions per minute, database size correction, application load correction, interface load correction, and the like.

In this case, it is not easy for the user to determine the input value of the database size correction value and the interface load correction value. This is because the database size correction value should be weighted by analyzing the actual business system in detail, and the interface load correction value should be arbitrarily selected from 1.1 and 1.2 based on the ratio value for communication with other servers.

Accordingly, the controller 150 sets the database size correction value to 1.7 and the interface load correction value to 1.1 before use.

Also, in the case of the application load compensation, the guidelines suggest a general value of 1.7, but to classify it as a fixed factor such as the database size compensation value and the interface load compensation value, more things need to be considered.

That is, as shown in FIG. 27 , the application load correction has a large deviation between the maximum and minimum values compared to other correction items, and when the maximum and minimum values are reflected in the tpmC calculation, the difference is about 2 times as shown in FIG. 28 . , so that the application load correction does not apply a general value, but is defined as including a value input from the user in the input item.

In addition, the control unit 150 summarizes the input items and input methods by synthesizing the defined contents as shown in [Table 2] below.

division item Input method






CPU



OLTP Server number of concurrent users - If it is difficult to know the number of concurrent users,
Enter the number of service users and the ratio of concurrent users
- Enter the annual increase rate of concurrent users and operating period Transactions per minute - Separately input the number of tasks per user and the number of transactions per task
- If the value is difficult to know, apply the default values of '2' and '5' respectively Application load compensation - Apply within the range of 1.3~2.2 by inputting high, medium, and low
WAS Server number of concurrent users - If it is difficult to know the number of concurrent users,
Enter the number of service users and the ratio of concurrent users
- Enter the annual increase rate of concurrent users and operating period Number of operations per user - If the value is difficult to know, enter '2'
web server number of concurrent users - If it is difficult to know the number of concurrent users,
Enter the number of service users and the ratio of concurrent users
- Enter the annual increase rate of concurrent users and operating period Number of operations per user - If the value is difficult to know, enter '2'
Memory number of concurrent users - If it is difficult to know the number of concurrent users,
Enter the number of service users and the ratio of concurrent users
- Enter the annual increase rate of concurrent users and operating period

The input items in [Table 2] are for tpmC calculation, and characteristics by server use are not considered, and as the calculation method was classified in FIG. 24, for a specific server, the CPU core type, number of cores, clock speed, etc. This should be taken into account.

Accordingly, the control unit 150 classifies the items to be additionally considered according to the server purpose as shown in Table 3 below.

Server use Additional considerations Example Minimum CPU Requirements
(There is a difference by specific use) core clock speed memory, etc.
All specifications
portal WAS - - - None (calculated according to tpmC) WEB - - - None (calculated according to tpmC) DB - ● - 2.5 GHz or higher A.I AI service ● - ● 8Core*2 or higher machine learning ● ● ● 2.5GHz or higher 16Core*2 or higher big data
cluster collection node ● ● ● 2.5GHz or higher 8Core*2 or higher management node ● ● ● 2.5GHz or higher 8Core*2 or higher data node ● ● ● 2.5GHz or higher 16Core*2 or higher Linkage, security, etc. - - - Apply manufacturer's recommended specifications

In addition, the control unit 150 derives the processing throughput, performance importance, and availability importance of the big data cluster as additional input items to reflect additional considerations depending on the server purpose when calculating the CPU/memory scale, and input items The final input items for CPU/memory capacity calculation are divided (or defined) as shown in [Table 4] through the rearrangement of

division input input range



hardware capacity Number of simultaneous internal and external users It can be replaced by inputting the number of service users and the ratio of concurrent users Annual growth rate of simultaneous internal and external users Enter as '%' Operating period Enter in 'years' Number of tasks per user Present default value '2' Transactions per job Present default value of '5' Number of operations per user in WAS/WEB Present default value of '5' application load Enter as high, medium, or low data capacity Big Data Cluster Throughput Per Minute Enter in 'TB'
quality characteristics performance criticality Enter as high, medium, or low Availability Critical Enter as high, medium, or low

In addition, the control unit 150 selects (or sets) an input factor for data capacity calculation through the following process.

Data capacity is a very important factor for estimating the size of a big data cluster. Servers other than the big data cluster each have disks of sufficient capacity and are connected to the integrated storage, so they are not significantly affected by the data capacity. However, in the case of a big data cluster, big data collection, processing, storage, analysis, and visualization Data capacity is very important because it must provide the foundation for

As defined in the application architecture reference model and the data architecture reference model, the resources included in the big data cluster include a real-time collection server, a name node server, a management node server, a data node server, and a machine learning server, and the control unit ( 150) defines the necessary items as in the following [Table 5] for estimating their size.

division resource type What you need to know for sizing collection Real-time collection server - Daily data collection and throughput

processing/management namenode server - Capacity of data node
- Availability importance Managed node server - Capacity of data node
- Availability importance
- Number of users using the software Save datanode server - Total amount of data to be stored annually
- Annual data growth rate
- Data retention period
- Operating period Analysis/Visualization machine learning server - Number of users using the software

In addition, the control unit 150 defines input items necessary for data capacity calculation based on the items for each resource type presented in [Table 5] as shown in [Table 6] below.

division input input range data
Volume Daily data collection and throughput Enter in 'TB' Total amount to be stored during the operating period Enter in 'TB' etc Operating period Enter the entire operating period Availability Critical Enter as high, medium, or low

In addition, the control unit 150 selects (or sets) an input factor for calculating the importance of quality characteristics through the following process.

The importance of quality characteristics is a part that can be referred to when calculating the specification and quantity of all resources and configuring the network and security.

In the embodiment of the present invention, performance, availability, and security were selected as key quality characteristics for building AI/big data center, and each input value was set to 'high', 'medium', 'as shown in [Table 7] below. It is configured so that you can receive it as 'ha'. In addition, network redundancy and network separation are configured so that they can be separately input regardless of the importance of quality characteristics.

division input input range network
Configuration Whether to configure network redundancy Enter 'yes' or 'not applicable' Whether the network is split Enter 'yes' or 'not applicable'
quality characteristics performance criticality Enter as high, medium, or low Availability Critical Enter as high, medium, or low Security importance Enter as high, medium, or low

In addition, the control unit 150 selects (or sets) an input factor for calculating the quantity of commercial software through the following process.

The quantity of commercial software may need to increase or decrease depending on the service purpose. Typically, in the case of an ESB (Enterprise service bus) connection solution, the number of adapter modules varies according to the number of connection organizations, and in the case of network connection software, the cost varies according to the number of people using the software.

The control unit 150 reflects these matters and defines input items for calculating the quantity of commercial software as shown in Table 8 below.

division input input range Number of SW users Number of people using business software Direct input Number of affiliated organizations Number of affiliated external organizations Direct input (without input, except ESB) Number of affiliated internal organizations Direct input (without input, except ESB)

In addition, the controller 150 synthesizes the input items defined above and performs a method of integrating similar parts except for overlapping, and then defines the final input items as shown in Table 9 below.

division input input range








hardware
Volume 1. [External Portal] Number of concurrent users Direct input of the number of people (persons) 1.1 Number of service users For the public,
Enter 'unspecified majority' 1.2 Concurrent User Ratio Ratio (%) Direct input 2. [External portal] Annual growth rate of concurrent users Ratio (%) Direct input 3. [External portal] Number of tasks per user If you don't know the value, enter the default value of '2' 4. [External portal] Number of transactions per task If you don't know the value, enter the default value of '5' 5. [External portal] Number of operations per user If you don't know the value, enter the default value of '5' 6. [External portal] Application load Enter 'upper', 'medium' or 'lower' 7. [Internal Portal] Number of Concurrent Users Direct input of the number of people (persons) 7.1 Number of service users For the public,
Enter 'unspecified majority' 7.2 Concurrent User Ratio Ratio (%) Direct input 8. [Internal Portal] Annual Growth Rate of Concurrent Users Ratio (%) Direct input 9. [Internal portal] Number of tasks per user If you don't know the value, enter the default value of '2' 10. [Internal portal] Number of transactions per task If you don't know the value, enter the default value of '5' 11. [Internal portal] Number of operations per user If you don't know the value, enter the default value of '5' 12. [Internal Portal] Application Load Enter 'upper', 'medium' or 'lower' Number of SW users 13. Number of people using business software (analysis, etc.) Direct input of the number of people (persons) Number of affiliated organizations 14. Number of affiliated external organizations Direct input (when not input, except ESB) 15. Number of affiliated internal organizations Direct input (when not input, except ESB) data
Volume 16. Daily data collection and processing capacity Capacity (TB) Direct input 17. Data storage capacity Capacity (TB) Direct input network
Configuration 18. Whether network redundancy is configured Enter either 'yes' or 'not applicable' 19. Whether the network is split Enter either 'yes' or 'not applicable'
quality characteristics 20. Performance Importance Enter 'upper', 'medium' or 'lower' 21. Availability Critical Enter 'upper', 'medium' or 'lower' 22. Security Critical Enter 'upper', 'medium' or 'lower' Operating period 23. Operating Period Period (year) Direct input

Accordingly, in the embodiment of the present invention, for a plurality of input items necessary for building artificial intelligence and big data center, one or more input items defined in [Table 9] are configured to receive an input value according to the manager input, , A fixed value (or default value/constant) may be preset for the remaining input items excluding the one or more input items among a plurality of input items necessary for building the artificial intelligence and big data center according to the administrator setting.

In this way, the control unit 150 may identify an influence factor in order to select a suitable model from among the defined architectural reference models and set (or derive) it as an input item for receiving a user input.

In addition, the control unit 150 defines a calculation formula for calculating a scale calculation result from one or more input values according to a manager input and a preset fixed value (or default value).

Here, as shown in FIG. 29 , the calculation formula includes calculation of tpmC, calculation of CPU specifications using tpmC, calculation of server capacity, calculation of quality importance weight, calculation of priority, and the like.

In addition, the control unit 150 may generate a scale calculation result including a hardware introduction list and quantity, a software introduction list and quantity, an infrastructure introduction list and quantity, and network security/configuration through this calculation.

In addition, the control unit 150 defines a calculation equation (or equation) related to tpmC through the following process.

One server has a workload that can be processed according to its specifications. Since each manufacturer has different standards, it is not easy to compare the performance between products. Therefore, it is the same performance standards as the Transaction Processing Performance Council (TPC) and the Standard Performance Evaluation Corporation (SPEC) to unify the standards, and in Korea, the TPC-C (tpmC) ), SPEC's SPECjbb2015 (max-jOPS) and SPC-1 (IOPS) are applied.

Here, the tpmC is a unit of measurement used when estimating the size of an OLTP server (On-Line Transaction Processing Server). At this time, in the case of the WAS server and the WEB server, the OPS (Open Profiling Standard) of SPECjbb2015 is the standard, and in the embodiment of the present invention, after calculating the OPS, a preset conversion ratio (for example, 2.02) It can be converted to the tpmC by multiplication.

30 , the controller 150 defines a tpmC calculation process according to an input item. Here, among the plurality of input values necessary for calculating the tpmC, the remaining input values excluding one or more input values according to the manager input may be in a state in which a fixed value (or a default value) is preset according to a manager setting.

In addition, the control unit 150 calculates the CPU specification through the following process.

That is, the reason for obtaining the tpmC is to select and introduce a server product satisfying the calculated tpmC from among many server products.

Server specifications are largely influenced by CPU specifications rather than memory and disk, and CPU specifications are estimated by mapping CPU core type, core quantity, and clock speed to tpmC through estimated data collected from global server manufacturers and distributors. configure to do

In addition, the existing 'information system hardware sizing guideline' guides only the process of calculating tpmC, and does not include the process of mapping tpmC and equipment to be introduced.

In an embodiment of the present invention, a calculation process for inferring the type, number of cores, and clock speed of the CPU is defined using the tpmC, and through this, the CPU specification and server capacity can be calculated based on the tpmC.

In addition, in the embodiment of the present invention, the relationship between tpmC and the CPU specification is defined as shown in FIGS. 31 and 32 through the analysis of the relationship between tpmC and the CPU specification. Here, the model name is de-identified.

33 is a diagram illustrating a scatter plot of CPU specifications according to the tpmC.

In addition, the control unit 150 generates (or derives) linear equations such as the following [Equation 1] to [Equation 3] based on the specification of a scatterplot having a positive correlation.

Here, f(x) represents the core type of the CPU, g(x) represents the number of cores of the CPU, and h(x) represents the clock speed of the CPU.

[Table 10] shows the difference between the applied value of the linear equation and the actual CPU specification.

turn tpmC core type number of cores clock speed Apply the equation difference Apply the equation difference Apply the equation difference One 798,491 4.0339074875 1.97 8.0678149750 3.93 1.7014128120 0.00 2 1,024,380 4.9359600666 3.06 9.8719201331 6.13 1.7389983361 0.04 3 1,050,646 5.0408492512 1.04 10.0816985025 2.08 1.7433687188 0.86 4 1,468,278 6.7085976705 2.71 13.4171953411 5.42 1.812852363 1.79 5 1,657,394 7.4638029950 0.54 14.9276059900 1.07 1.8443251248 0.04 6 1,825,498 8.1351001664 0.14 16.2702003328 0.27 1.8722958403 0.23 7 1,954,202 8.6490595674 2.65 17.2981191348 5.30 1.8937108153 1.51 8 2,342,941 10.2014282862 0.20 20.4028565724 0.40 1.9583928453 0.24 9 2,508,418 10.8622349418 1.14 21.7244698835 2.28 1.9859264559 0.11 10 2,560,950 11.0720133111 1.07 22.1440266223 2.14 1.9946672213 0.41 11 2,708,040 11.6593943428 3.66 23.3187886855 7.32 2.0191414309 1.18 12 2,968,075 12.6978036606 0.70 25.3956073211 1.40 2.0624084859 0.24 13 3,335,802 14.1662642263 0.17 28.3325284526 0.33 2.1235943428 0.08 14 3,388,334 14.3760425957 2.38 28.7520851913 4.75 2.1323351082 0.47 15 3,606,343 15.2466276206 0.75 30.4932552413 1.51 2.1686094842 0.17 16 3,698,274 15.6137397671 3.61 31.2274795341 7.23 2.1839058236 0.82 17 3,834,858 16.1591667221 0.16 32.3183334443 0.32 2.2066319468 0.11 18 3,887,391 16.3689490849 2.37 32.7378981697 4.74 2.2153728785 0.38 19 4,491,512 18.7814123128 1.22 37.5628246256 2.44 2.3158921797 0.32 20 4,570,311 19.0960838602 1.10 38.1921677205 2.19 2.3290034942 0.03 21 4,780,440 19.9352013311 1.94 39.8704026622 3.87 2.3639667221 0.34 22 5,016,835 20.8792079867 0.88 41.7584159734 1.76 2.4033003328 0.00 23 5,043,102 20.9841011647 1.02 41.9682023295 2.03 2.4076708819 0.31 24 5,253,231 21.8232186356 3.82 43.6464372712 7.65 2.4426341098 0.56 25 5,332,029 22.1378861897 1.86 44.2757723794 3.72 2.4557452579 0.36 26 5,550,038 23.0084712146 2.99 46.0169424293 5.98 2.4920196339 0.39 27 5,647,223 23.3965643927 2.60 46.7931287854 5.21 2.5081901830 0.51 28 6,198,812 25.5992492512 2.40 51.1984985025 4.80 2.5999687188 0.50 29 6,303,877 26.0188099834 2.02 52.0376199667 4.04 2.6174504160 0.08 30 6,776,668 27.9068272879 0.09 55.8136545757 0.19 2.6961178037 0.20 Consider the expansion unit
Tolerance of error: 4 Consider the expansion unit
Tolerance of error: 8 Consider the expansion unit
Tolerance of error: 0.5

In addition, the control unit 150 sets (or defines) 4 Cores, 8 Cores, and 0.5 GHz in consideration of the physical expansion units for the respective error tolerance ranges for the core type, the number of cores, and the clock speed. Here, since the core type increases by 4 cores and the number of cores increases by 2 times, 4 and 8 (4*2), which are the minimum units, are defined as the error acceptance range of the core type and the number of cores, respectively. Also, in the case of clock speed, since the recommended specifications to be presented by the target model are classified in units of 0.5 GHz, the error acceptance range of the clock speed is defined as 0.5 GHz.

In addition, in the case of the clock speed, as shown in [Table 10], it can be confirmed that there are about 25% (7) values that are above the error acceptance range. This is because, unlike the core type or number of cores, the clock speed may decrease even if the CPU specifications improve.

Accordingly, the controller 150 defines a non-linear equation such as the following [Equation 4] rather than a linear equation by using the Newtonian interpolation method.

Here, h(x) represents the clock speed of the CPU using the Newton interpolation method.

34 is a diagram illustrating an example of a clock speed of a CPU to which an actual value, a linear equation, and a Newton interpolation method are respectively applied.

As shown in [Table 11], it can be confirmed that the error rate of the CPU clock speed equation to which Newton interpolation is applied is reduced from 27% (7 pieces) to 16% (5 pieces). Although the number of values out of error is not significantly reduced, the improved equation is very good in that the remaining 5 are all products with a clock speed of 3.0 GHz or higher (here, products with 3.0 GHz or higher are only applied for special purpose purposes such as machine learning). Reliability is high.

turn tpmC real
clock speed Applying linear equations Newton interpolation method applied Apply the equation difference Apply the equation difference One 798,491 1.7 1.7014128120 0.0014 2.1032490760 0.4032 2 1,024,380 1.7 1.7389983361 0.0390 2.1013866235 0.4014 3 1,050,646 2.6 1.7433687188 0.8566 2.1031693745 0.4968 4 1,468,278 3.6 1.812852363 1.7871 2.1489926657 1.4510 5 1,657,394 1.8 1.8443251248 0.0443 2.1696762896 0.3697 6 1,825,498 2.1 1.8722958403 0.227 2.1855453126 0.0855 7 1,954,202 3.4 1.8937108153 1.5063 2.1963471914 1.2037 8 2,342,941 2.2 1.9583928453 0.2416 2.2270840441 0.0271 9 2,508,418 2.1 1.9859264559 0.1141 2.2418573552 0.1419 10 2,560,950 2.4 1.9946672213 0.4053 2.2469882264 0.1530 11 2,708,040 3.2 2.0191414309 1.1809 2.2626916768 0.9373 12 2,968,075 2.3 2.0624084859 0.2376 2.2955405757 0.0045 13 3,335,802 2.2 2.1235943428 0.0764 2.3509943381 0.1510 14 3,388,334 2.6 2.1323351082 0.4677 2.3593853269 0.2406 15 3,606,343 2 2.1686094842 0.1686 2.3940908946 0.3941 16 3,698,274 3 2.1839058236 0.8161 2.4082563877 0.5917 17 3,834,858 2.1 2.2066319468 0.1066 2.4282397140 0.3282 18 3,887,391 2.6 2.2153728785 0.3846 2.4354843751 0.1645 19 4,491,512 2 2.3158921797 0.3159 2.4926967160 0.4927 20 4,570,311 2.3 2.3290034942 0.0290 2.4960869221 0.1961 21 4,780,440 2.7 2.3639667221 0.3360 2.5007451034 0.1993 22 5,016,835 2.4 2.4033003328 0.0033 2.4997665018 0.0998 23 5,043,102 2.1 2.4076708819 0.3077 2.4993642548 0.3994 24 5,253,231 3 2.4426341098 0.5574 2.4950714301 0.5049 25 5,332,029 2.1 2.4557452579 0.3557 2.4933258846 0.3933 26 5,550,038 2.1 2.4920196339 0.3920 2.4898294805 0.3898 27 5,647,223 2 2.5081901830 0.5082 2.4895361137 0.4895 28 6,198,812 2.1 2.5999687188 0.5000 2.5162368856 0.4162 29 6,303,877 2.7 2.6174504160 0.0825 2.5281567451 0.1718 30 6,776,668 2.5 2.6961178037 0.1961 2.6061806210 0.1062 Number of error items: 7 Number of error items: 5
(However, all 5 are 3.0GHz or higher,
used in special cases)

In this way, the controller 150 may calculate the CPU specification from the tpmC through [Equation 1], [Equation 2], and [Equation 4].

In addition, the control unit 150 calculates the server capacity through the following process.

In the case of a CPU with a clock speed of 3.0 GHz or higher, it is used in a server for a rather special purpose.

For example, in the case of machine learning and real-time collection, it is recommended to use a CPU with a specification of 2.5 GHz or higher, and set to use 3.0 GHz or higher in consideration of business characteristics.

As such, the recommended specifications of the core type, number of cores, and clock speed are variously set according to the purpose of each server. can be defined

tpmC threshold clock speed core type number of mounts note less than 500,000 - - - irrelevant

over 500,000
less than 1,000,000 2.0 GHz or higher 4 or more cores One basic 2.0 GHz or higher 8 cores or more One basic 2.0 GHz or higher 12 cores or more 2 When high core count is required 2.0 GHz or higher 16 cores or more 2 When high core count is required 2.5 GHz or higher 4 or more cores One When high clock is required

1,000,000 or more
less than 1,500,000 2.0 GHz or higher 4 or more cores One basic 2.0 GHz or higher 8 cores or more One basic 2.0 GHz or higher 12 cores or more 2 When high core count is required 2.0 GHz or higher 16 cores or more 2 When high core count is required 2.5 GHz or higher 4 or more cores 2 When high clock is required

over 1,500,000
less than 2,000,000 2.0 GHz or higher 4 or more cores 2 basic 2.0 GHz or higher 8 cores or more 2 basic 2.0 GHz or higher 12 cores or more 2 When high core count is required 2.0 GHz or higher 16 cores or more 2 When high core count is required 2.5 GHz or higher 4 or more cores 2 When high clock is required
over 2,000,000
less than 2,500,000 2.0 GHz or higher 8 cores or more 2 basic 2.0 GHz or higher 12 cores or more 2 When high core count is required 2.0 GHz or higher 16 cores or more 2 When high core count is required 2.5 GHz or higher 8 cores or more 2 When high clock is required
Over 2,500,000
less than 3,000,000 2.0 GHz or higher 8 cores or more 2 basic 2.0 GHz or higher 12 cores or more 2 When high core count is required 2.0 GHz or higher 16 cores or more 2 When high core count is required 2.5 GHz or higher 8 cores or more 2 When high clock is required
3,000,000 or more
less than 3,500,000 2.0 GHz or higher 12 cores or more 2 basic 2.0 GHz or higher 16 cores or more 2 When high core count is required 2.5 GHz or higher 12 cores or more 2 When high clock is required
over 3,500,000
less than 4,000,000 2.0 GHz or higher 12 cores or more 2 basic 2.0 GHz or higher 16 cores or more 2 When high core count is required 2.5 GHz or higher 12 cores or more 2 When high clock is required 4,000,000 or more
less than 4,500,000 2.0 GHz or higher 16 cores or more 2 basic 2.5 GHz or higher 16 cores or more 2 When high clock is required 4,500,000 or more
less than 5,000,000 2.5 GHz or higher 16 cores or more 2 basic 5,000,000 or more
less than 5,500,000 2.5 GHz or higher 20 or more cores 2 basic 5,500,000 or more
less than 6,000,000 2.5 GHz or higher 20 or more cores 2 basic 6,000,000 or more 2.5 GHz or higher 20 or more cores 2 basic

In addition, the control unit 150 defines a server having a special purpose in which the core type, number of cores, and clock speed should be considered as shown in Table 13 below.

Server use standard note core type number of mounts clock speed
portal WAS - - - Apply tpmC WEB - - - max-jOPS
apply DB - - 2.5 GHz or higher A.I AI service 8 cores or more 2 or more -


Apply tpmC machine learning 16 cores or more 2 or more 2.5 GHz or higher big data
cluster collection node 8 cores or more 2 or more 2.5 GHz or higher management node 8 cores or more 2 or more 2.5 GHz or higher data node 16 cores or more 2 or more 2.5 GHz or higher Linkage, security, etc. - - - manufacturer
Recommended specification applied

As a result, the server to be introduced must satisfy the minimum recommended specification while satisfying tpmC. However, since the value of tpmC is not determined only by the specifications of the CPU, the controller 150 finally calculates the capacity of the server by summing all of the memory, disk, battery, network adapter, and the like. At this time, in the embodiment of the present invention, the part number, standard, cost, etc. of each detailed device are made into a database so that the corresponding control unit 150 can combine and sum them.

In addition, the control unit 150 calculates the quality weight through the following process.

That is, as input items for the importance of quality characteristics, performance importance, availability importance, and security importance were configured to receive any one of 'high', 'medium' and 'low'.

In an embodiment of the present invention, in order to establish application standards according to 'upper', 'middle' and 'lower', informatization planning, application/database development, hardware/software introduction and construction, network and security configuration, and infrastructure construction fields Conduct a survey on industry experts, and finally define criteria that can reflect importance as shown in [Table 14] below.

division initial value Performance availability security Prize middle under Prize middle under Prize middle under portal duplication maintain maintain single server maintain maintain single server maintain maintain maintain ESB
Link single maintain maintain maintain duplication maintain maintain maintain maintain maintain AI
service single maintain maintain maintain duplication maintain maintain maintain maintain maintain Load
Dispersion apply maintain maintain Unapplied maintain maintain maintain maintain maintain maintain NW
duplication Requirements
reflection maintain maintain maintain duplication maintain Unapplied maintain maintain maintain network separation Requirements
reflection maintain maintain maintain maintain maintain maintain maintain maintain maintain control all
introduction maintain maintain APM
NMS
except maintain maintain maintain maintain maintain ESM
except security
solution all
introduction maintain maintain maintain maintain maintain maintain maintain maintain all
except security
equipment all
introduction maintain maintain maintain maintain maintain maintain maintain maintain DDoS
except

In addition, the control unit 150 calculates the priority through the following process.

That is, the output obtained through input and calculation may or may not be the optimal target model. Since many environmental factors may act as variables, in the embodiment of the present invention, the number of cases is derived by sequentially changing the parts that need to be adjusted.

In addition, the control unit 150 calculates (or calculates) the priority of the scale calculation result based on the information on the change in the importance of quality characteristics and the information on the big data cluster configuration change. Here, the information on the change in the quality characteristic importance includes change information on the performance importance, the availability importance, and the security importance. In addition, the information related to the big data cluster configuration change includes information set according to the infrastructure (or infrastructure configuration) based on the data storage capacity included in the received input value.

That is, the controller 150 sets the first derived scale calculation result as the first priority.

Also, the control unit 150 calculates a priority score by reflecting the information on the change in the importance of quality characteristics. Here, the control unit 150 gives a score (for example, high: 5 points, medium: 3 points, bottom: 2 points) for the importance of quality characteristics for each high, medium, and low. In addition, the control unit 150 derives the number of possible changes.

In addition, the control unit 150 derives a difference by digitizing before and after the change.

Also, as shown in FIG. 35 , the control unit 150 sorts in ascending order from the number of cases in which the difference is small. Here, the collected information considers the scope of change from the front, portal, AI service, ESB linkage, load balancing, APM (Application Performance Management: Application Performance Management), (Network Management System: Network Management System), ESM (Enterprise Security) Management: Indicates T(true)/F(false) values for Integrated Security Management System), DB Encryption, DB Access Control, Server Access Control, Distributed Denial of Service, and Cluster Support.

In addition, the control unit 150 calculates a priority score by reflecting the information related to the big data cluster configuration change. Here, the control unit 150 differentially assigns preset 0.5 points in ascending order from the number of cases having the least influence when changing the cluster configuration. In this case, the cluster configuration change includes name nodes, utility nodes, data nodes, collection nodes, and adjustment of the quantity and configuration of machine learning resources.

In addition, the control unit 150 collects the priority score reflecting the information on the change in the importance of the quality characteristic and the priority score reflecting the information related to the calculated big data cluster configuration change, and the T/F value ( or True/False values).

In addition, the control unit 150 reflects the total score and calculates (or calculates) the priority by removing duplicates.

In this way, the control unit 150 may define a plurality of calculation formulas for calculating the scale.

In addition, the sizing apparatus 100 for building the artificial intelligence and big data center configures (or designs) the entire system based on input items, architecture reference models, calculation formulas, etc. to implement the sizing requirements. Here, the detailed design of the software is designed (or composed) using the use case diagram, sequence diagram, communication diagram, state machine diagram, and interaction diagram of UML (Unified Modeling Language), and the detailed data design is CRUD (Create, Read , Update, Delete) and ERD (Entity Relationship Diagram) are used to design. In addition, development was programmed using Windows Server, Javascript, Active Server Pages, and MS-SQL.

In addition, the scale calculating apparatus 100 derives a user-side function, and designs a sequential process procedure for the input items.

Fig. 36 is a use case diagram expressing user-side requirements.

Since the use case diagram does not include detailed content, the use case diagram is detailed, and a sequence diagram and a communication diagram are shown as shown in FIGS. 37 and 38 .

Activity diagram and state machine diagram were used for the design of interaction and state transition.

In addition, in order to clarify the relationship between the input and the output, the relationship between the input item and the output screen is defined as shown in FIG. 39 and reflected in the activity diagram shown in FIG. 40 and the state machine diagram shown in FIG. 41 .

Input items 1 to 12 and 23 in [Table 9] above are input factors affecting tpmC and CPU specifications, and input items 13 to 17 are input factors affecting server capacity calculation. In addition, input items 18 to 19 are input factors for selecting an architecture reference model, and input items 20 to 22 are input factors for adjusting priorities through quality characteristic weights.

In addition, input factors input to each activity influence other activities by interaction as shown in FIG. 41, and finally provide the output of priority list, resource list, server SPEC, and infrastructure configuration.

While the activity diagram is a diagram of a process or the interaction between each activity, the state machine diagram is a diagram of dynamic states and state changes.

In addition, by designing (or configuring) the state machine diagram by classifying the input items, the control unit 150 makes it easy to identify the influence of the input factors on the results, and it is possible to trace the input items from each state configure to do

In addition, the control unit 150 analyzes the correlation between the process and data as shown in Table 15 below for designing the scale estimation database, and performs database modeling as shown in FIGS. 42 and 43 . Here, C represents Create, R represents Read, U represents Update, and D represents Delete.

process asset
classification
DB input
DB CPU-
tpmC
mapping
standard
DB by asset
tpmC
-CPU
mapping
DB per server
fix
part
DB server
part
expense
DB by asset
Calculation
result
DB Initial baseline data entry C C C C C Input items C Calculate tpmC and CPU specs R R C Server capacity calculation R R R R Calculation of quality weight and resource quantity R R R R/U HW, SW list and quantity output R R Server detailed specification output R R R R

The database is included in the storage unit 120 and is classified into input, calculation, and output systems, and the data mapping operation is configured to operate in the calculation system.

In this way, the control unit 150 may configure (or develop) business logic to meet the functions, processes, and database design requirements of the scale estimation apparatus 100 .

In addition, the control unit 150 performs tpmC calculation, CPU specification calculation, server capacity calculation, quality weight calculation, and priority calculation centering on input and output functions, and finally configures business logic for resource specification and quantity calculation do.

In addition, the controller 150 installs a dedicated app (or application/application program/specific app) provided from a server (not shown) in the corresponding scale calculating device 100 . In this case, the dedicated app may be an app for performing a function of generating a scale calculation result based on a default value according to a user setting and an input value according to a user input for building an artificial intelligence and big data center.

In addition, the control unit 150 executes a dedicated app pre-installed in the corresponding scale calculating device 100, and displays an app execution result screen according to the execution of the dedicated app. At this time, the app execution result screen receives one or more input values corresponding to one or more input items that require user (or designer/administrator) setting among a plurality of input values required for size calculation for artificial intelligence and big data center construction. Information input menu (or item/button) for , Priority menu to provide a priority list according to the scale calculation result calculated according to the input value, Resource list menu to provide a resource list according to the sizing result, Scale Calculation It includes a server specification menu for providing a server specification (or server specification) according to the result, an infrastructure configuration menu for providing an infrastructure configuration according to the scale calculation result, and the like. At this time, only when an input value according to a user input is received after the information input menu is selected among the plurality of menus, the priority menu, the resource list menu, the server specification menu, and the infrastructure configuration menu may be activated and displayed, When the input value is not received, the priority menu, resource list menu, server specification menu, and infrastructure configuration menu may be inactivated and displayed. Here, the scale calculation device 100 is a member of the server providing the dedicated app, and when executing the dedicated app using an ID and password according to membership registration, a barcode or QR code including the ID, etc. By performing a login procedure, one or more functions (eg, including a sizing function, etc.) of the dedicated app may be performed.

In addition, when the information input menu is selected on the app execution result screen displayed on the display unit 130, the control unit 150 controls the user (or An information input screen for receiving one or more input values corresponding to one or more input items that require setting (designer/administrator) is displayed on the display unit 130 .

In addition, the control unit 150 according to the user input (or user/administrator selection/touch/control) of the artificial intelligence and big data center construction apparatus 100 on the information input screen displayed on the display unit 130, artificial intelligence and One or more input values corresponding to one or more input items requiring user (or designer/administrator) setting from among a plurality of input values corresponding to a plurality of input items required for size calculation for building a big data center are received. Here, the plurality of input items are items necessary for calculating the scale for building artificial intelligence and big data center, and are preset according to the designer's design, and among the plurality of input values, one or more input values requiring the user setting are selected. The remaining input values (or at least one remaining input value) may have fixed values (or default values/constants) set in advance according to a designer's design. In addition, the at least one input value requiring user setting is a value that a user must directly input according to an environment among the plurality of input values, and is an input value related to hardware capacity (eg, the number of simultaneous users of external portal/internal portal, external Annual growth rate of simultaneous portal/internal portal users, number of tasks per external portal/internal portal user, number of transactions per external portal/internal portal task, number of operations per external portal/internal portal user, external portal/internal portal application load, etc.); Input values related to the number of people using the software (eg, including the number of users using business software (including analysis, etc.) Input values related to (eg, daily data collection and processing capacity, data storage capacity, etc.), input values related to network configuration (eg, whether network redundancy is configured, whether network is divided, etc.), input related to quality characteristics Include values (including, for example, performance criticality, availability criticality, security criticality, etc.) and input values related to the operational period (including operational period, etc.).

In addition, the control unit 150 selects (or selects) an architecture reference model corresponding to the received input value from among a plurality of preset architecture reference models based on the received input value.

That is, the control unit 150 determines whether the network is configured to be redundant, whether or not to divide the network, and whether or not to configure the network included in the input values related to the network configuration among the received input values (eg, whether to configure the redundant network, whether to separate the network, etc.). Based on at least one of virtualization, any one architecture reference model is selected from among eight preset architecture reference models. Here, the eight architecture reference models (or infrastructure architecture reference models) are the first to eighth architecture reference models shown in FIGS. 13 to 20 according to the number and characteristics of the infrastructure architecture selection of Table 1 above. Includes architectural reference models.

In addition, the control unit 150 calculates (or calculates) the tpmC for each asset use and the CPU specification for each asset use based on the received input value.

That is, the controller 150 controls the input value related to the hardware capacity among the received input values according to the calculation items in the database server (not shown) (eg, the number of simultaneous users of the external portal/internal portal, the external portal/internal portal). Concurrent user annual growth rate, number of tasks per external portal/internal portal user, number of transactions per external portal/internal portal task, number of operations per external portal/internal portal user, external portal/internal portal application load, etc.) Calculate tpmC for each asset use based on input values (eg, including operating period). Here, the tpmC is a unit of measurement used when estimating the size of an OLTP server (On-Line Transaction Processing Server). At this time, in the case of the WAS server and the WEB server, the OPS (Open Profiling Standard) of SPECjbb2015 is the standard, and in the embodiment of the present invention, after calculating the OPS, a preset conversion ratio (for example, 2.02) It can be converted to the tpmC by multiplication.

In addition, the controller 150 calculates (or calculates) CPU specifications for each asset use by inferring the core type, number of cores, and clock speed of the CPU using the calculated (or calculated) tpmC.

That is, the control unit 150 applies the calculated tpmC to the variables of [Equation 1], [Equation 2] and [Equation 4] related to the core type of the CPU, the number of cores of the CPU and the clock speed of the CPU. Thus, the CPU specifications for each asset use are calculated. Here, [Equation 1] and [Equation 2] may be in the form of linear equations related to the tpmC, and [Equation 4] may be in the form of applying Newtonian interpolation in which the number of error items is reduced compared to the linear equation.

In addition, the control unit 150 controls the calculated (or calculated) tpmC for each asset use, the calculated CPU specification for each asset use, and an input value related to the number of software users among the received input values (for example, business software (analysis, etc.) Including) number of users, etc.), input values related to the number of affiliated organizations (including, for example, the number of connected external organizations, the number of connected internal organizations, etc.), and input values related to data capacity (eg, daily data collection and processing capacity, data storage capacity, etc.), calculates (or calculates) the server capacity based on the selected architecture reference model, etc.

That is, the control unit 150 controls the calculated (or calculated) tpmC for each asset use, the calculated CPU specification for each asset use, and an input value related to the number of software users among the received input values (for example, business software (analysis, etc.) Including) number of users, etc.), input values related to the number of affiliated organizations (eg, the number of connected external organizations, number of connected internal organizations, etc.), input values related to data capacity (eg, daily data collection and processing capacity, data storage capacity, etc.), big data platform server capacity, artificial intelligence service server capacity, portal service server capacity, linked service server capacity, and other service server capacity such as security/control, respectively, based on the selected architecture reference model, etc. Calculate.

In addition, the control unit 150 selects a server specification for each asset use based on the calculated (or calculated) server capacity.

That is, the control unit 150 controls big data by asset use based on the calculated big data platform server capacity, artificial intelligence service server capacity, portal service server capacity, linked service server capacity, and other service server capacity such as security/control. Platform server specifications, artificial intelligence service server specifications by asset use, portal service server specifications by asset use, linked service server specifications by asset use, and other service server specifications by asset use are selected, respectively.

In addition, the control unit 150 is the selected asset based on the server specification for each asset use and input values related to quality characteristics among the received input values (including, for example, performance importance, availability importance, security importance, etc.) Apply (or calculate) the quality characteristic weight to the server specification for each purpose.

That is, the control unit 150 controls the selected big data platform server specifications for each asset use, artificial intelligence service server specifications for each asset use, portal service server specifications for each asset use, linked service server specifications for each asset use, and other service server specifications for each asset use. Each quality characteristic weight is applied according to the input value related to the characteristic (including, for example, performance importance, availability importance, security importance, etc.).

In addition, the control unit 150 calculates (or calculates) the quantity for each asset use based on the server specification for each asset use to which the quality characteristic weight is applied (or calculated).

That is, the control unit 150 controls the big data platform server specification for each asset use to which the quality characteristic weight is applied, the artificial intelligence service server specification for each asset use, the portal service server specification for each asset use, the linked service server specification for each asset use, and other service servers for each asset use. Based on specifications, etc., the quantity for each asset use is calculated.

In addition, the control unit 150 optimizes the network configuration, security configuration, and control configuration based on the selected server specification for each asset use and the calculated quantity for each asset use.

That is, the control unit 150 optimizes the network configuration, the security configuration, and the control configuration according to the selected server specification for each asset use and the quantity for each asset use.

In addition, the control unit 150 calculates (or calculates) the cost according to the specification of the server for each asset use and the quantity for each asset use in which the network configuration, the security configuration, and the control configuration are optimized.

That is, the control unit 150 calculates the cost for each asset use with respect to the specification of the server for each asset use and the quantity for each asset use in a state where the network configuration, the security configuration, and the control configuration are optimized.

In addition, the control unit 150 calculates (or calculates) the priority of the scale calculation result based on the information on the change in the importance of quality characteristics and the information on the big data cluster configuration change. Here, the information on the change in the quality characteristic importance includes change information on the performance importance, the availability importance, and the security importance. In addition, the information related to the big data cluster configuration change includes information set according to the infrastructure (or infrastructure configuration) based on the data storage capacity included in the received input value.

Also, the control unit 150 defines a priority for the scale calculation result based on the priority for the calculated (or calculated/set) scale calculation result.

Also, the controller 150 adjusts at least one input value (or an influence factor) among the input values according to the priority of the defined scale calculation result.

Also, the controller 150 reflects (or applies) the adjusted at least one input value.

In addition, the control unit 150 repeats the process of adjusting the influence factor until the priority list corresponds to the preset number of priority lists (eg, 30).

In addition, the control unit 150 includes the selected server specification for each asset use, the network for each asset use, the configuration of the server optimized for security and control configuration (or the specification of the server for each asset use with the optimized network, security and control configuration), adjusted The scale calculation result including the priority and detailed information is stored (or adjusted) in the storage unit 120 based on the priority list in which the input value is reflected. Here, the detailed information includes hardware introduction list and quantity, software introduction list and quantity, infrastructure introduction list and quantity, resource list by asset use such as network security/configuration, server specifications, infrastructure configuration, and the like.

In addition, when a priority menu is selected from among a plurality of menus displayed on the display unit 130 , the control unit 150 responds to the scale calculation result calculated based on the previously received input value in response to the selected priority menu. A priority list (or a priority list included in the scale calculation result) is displayed (or output) on the display unit 130 . Here, the plurality of menus are for receiving one or more input values corresponding to one or more input items that require user (or designer/administrator) setting among a plurality of input values required for size calculation for artificial intelligence and big data center construction. Information input menu, priority menu to provide a priority list according to the result of sizing calculated according to the input value, resource list menu to provide a list of resources according to the result of sizing, server specification according to the result of sizing (or server specification) for providing the server specification menu, and the infrastructure configuration menu for providing the infrastructure configuration according to the size calculation result. At this time, only when an input value according to a user input is received after the information input menu is selected among the plurality of menus, the priority menu, the resource list menu, the server specification menu, and the infrastructure configuration menu may be activated and displayed, When the input value is not received, the priority menu, resource list menu, server specification menu, and infrastructure configuration menu may be inactivated and displayed.

In addition, when a resource list menu is selected from among a plurality of menus displayed on the display unit 130, the control unit 150 is based on the scale calculation result calculated based on the previously received input value in response to the selected resource list menu. A corresponding resource list (or a resource list included in the scale calculation result) is displayed (or output) on the display unit 130 . At this time, only when the priority menu is selected from among the plurality of menus, the resource list menu may be activated and displayed, and when the priority menu is not selected, the resource list menu, server specification menu, and infrastructure configuration menu may be displayed as inactive.

In addition, when a server specification menu (or server specification menu) is selected from among the plurality of menus displayed on the display unit 130 , the control unit 150 is configured based on the previously received input value in response to the selected server specification menu. The server specification (or server specification/server specification included in the scale calculation result) according to the calculated scale calculation result is displayed (or output) on the display unit 130 . At this time, only when the resource list menu is selected among the plurality of menus, the server specification menu may be activated and displayed, and when the resource list menu is not selected, the server specification menu and the infrastructure configuration menu are inactivated and displayed can be

In addition, when the infrastructure configuration menu is selected from among the plurality of menus displayed on the display unit 130, the control unit 150 responds to the selected infrastructure configuration menu based on the previously received input value. The corresponding infrastructure configuration (or the infrastructure configuration included in the scale calculation result) is displayed (or output) on the display unit 130 . At this time, only when the server specification menu is selected from among the plurality of menus, the infrastructure configuration menu may be activated and displayed, and when the server specification menu is not selected, the infrastructure configuration menu may be displayed in inactive.

In an embodiment of the present invention, in order to verify the validity and effectiveness of the scale estimating device 100 , the corresponding scale estimating device 100 is simulated and verified based on an actual design case or a construction example.

The case used for verification is the design/construction case of government ministries and city/provincial offices conducted from 2019 to 2020. In order to prevent the leakage of unnecessary information, institutional information and detailed contents were excluded. As shown in -1>, by checking whether there is a model with the same hardware quantity as the actual model within the sizing priority and whether the cost difference with the actual model excluding the infrastructure cost satisfies the preset error range within 6.49%. allowed to judge. Here, the criterion for determining the error range for the 6.49% cost is the maximum for the average error (for example, 12.98%=(2.72%+19.39%+16.84%)/3) according to the estimate comparison result of three or more preset companies. It may be an intermediate value taking into account the difference between the value and the minimum value.

[Table 16] shows the criteria for simulation verification.

division Verification items Criteria

One - Whether a model with the same hardware quantity as the actual model exists within the priority output as a result of the sizing. - Fit: the same model exists within the priority
(However, as a difference according to business characteristics,
Appropriate if there is a good reason)
- Non-conformity: The same model does not exist within the priority.

2 - Whether the cost difference from the actual model satisfies the margin of error within 6.49%
(Excluding infrastructure) - Appropriate: The difference between the cost according to the scale calculation and the actual cost is within 6.49% of the actual cost.
- Nonconformity: The difference between the cost and the actual cost according to the sizing exceeds the range of 6.49% of the actual cost

As shown in FIG. 44 , the verification target is limited to the intersection of the actual case and the result of the sizing device 100 , and in the case of infrastructure, depending on the hardware quantity, a rack, an uninterruptible power supply (UPS), etc. It is excluded from the scope of verification because the capacity of

A case example for verification is a case of government department A, an example of building an open AI/big data center for public institutions and private companies, and the information input screen 4700 shown in FIG. 47 according to the example described below. ) using one or more input values received from and preset default values.

In the case of the above example, many services are included in addition to the simple portal service, and the actual design result based on the verification target is summarized in [Table 17] below.

division Quantity total cost

server portal 18


8,903,286,700 won
(tax included) A.I 9 big data cluster 61 Linkage, security, etc. 32 storage 2 network equipment 24 security equipment 10 equipment 156

In addition, as a result of inputting an input value to the scale calculating device 100 and executing it, a result similar to the actual design result was obtained in the first priority model, and details are shown in [Table 18] below.

purpose real
Quantity Result of sizing device run result adjustment HW cost (KRW) SW cost (KRW) Internal portal WAS 2 2 2 17,748,000 - Internal portal WEB 2 2 2 17,748,000 - Internal portal DB 2 2 2 22,534,000 - Internal portal WAS (development) One One One 8,874,000 - Internal portal WEB (development) One One One 8,874,000 - Internal portal DB (development) One One One 11,267,000 - External portal WAS 2 2 2 17,748,000 - External portal WEB 2 2 2 17,748,000 - External portal DB 2 2 2 25,438,000 - External portal WAS (development) One One One 8,874,000 - External portal WEB (development) One One One 8,874,000 - External portal DB (development) One One One 12,719,000 - AI service 4 4 4 58,980,000 - machine learning 5 5 5 312,335,000 - Real-time collection 3 3 3 98,310,000 - collection inside 3 3 3 101,304,000 - outside collection One One One 9,869,000 - SQL/NoSQL engine 2 2 2 19,738,000 - master node 3 3 3 52,518,000 - utility node 2 2 2 36,420,000 - data node 50 50 50 2,900,900,000 - Metadata/Quality Control One One One 9,869,000 - data mart 2 2 2 19,738,000 - Advanced analysis 2 2 2 19,738,000 - text analysis 2 2 2 19,738,000 - visualization 2 2 2 19,738,000 - Internal connection ESB 2 2 2 19,738,000 630,000,000 Externally linked ESB 2 2 2 19,738,000 630,000,000 DB access control 2 2 2 19,308,000 140,000,000 Server access control 2 2 2 19,308,000 69,500,000 DB encryption 2 2 2 19,308,000 196,000,000 NMS 2 2 2 20,474,000 45,000,000 ESM 2 2 2 20,474,000 176,000,000 APM 2 2 2 20,474,000 330,000,000 material linkage 2 2 2 26,176,000 20,400,000 stream association 2 2 2 26,176,000 20,000,000 internal network storage One One One 126,904,000 - external network storage One One One 126,904,000 - internal network backbone 2 2 2 198,924,000 - external network backbone 2 2 2 160,824,000 - L4 switch 4 4 4 212.8 million - L3 switch 2 2 2 43,780,000 - cluster L2 switch 6 6 6 303,738,000 - Internal network L2 switch 2 2 2 43,780,000 - External network L2 switch 2 2 2 31,434,000 - SAN switch 4 4 4 109,976,000 - firewall 4 4 4 174,000,000 - Anti-DDos 2 2 2 86,000,000 - IPS 2 2 2 102,000,000 - web firewall 2 2 2 160,000,000 - total 156 156 156 5,949,879,000 2,256,900,000

The actual design result is a total number of 156 units and a total cost of 8,903,286,700 won (including tax). about 1.39%).

Since it is smaller than the 6.49% preset error range, which is the verification standard, the verification result through case examples can be judged to be appropriate.

In the embodiment of the present invention, the cost incurred during function development and data collection/operation from the user's side is not considered, but it may be configured to consider the corresponding cost according to the design of the designer.

In addition, the product list and cost data managed by the scale calculating device 100 may be configured by reflecting changes in prices and cost increase/decrease according to market prices in real time.

That is, through a web crawling method, etc., it is possible to provide more intelligent information by reflecting the cash circulation market and checking, predicting, and learning in real time changes such as specifications and/or prices of constituent resources.

Although the embodiment of the present invention is configured to receive input values according to the manager input for about 23 input items (or input factors) requiring manager input, the present invention is not limited thereto. Since it may be difficult to input accurate input values, it may be configured to provide a recommendation function for some input values based on analysis of business requirements, laws, systems, and stakeholders by industry and/or specific industry have.

Through this, the effectiveness of the use of the scale calculating device 100 is further increased, and it can be configured so that anyone can use it easily.

In this way, after defining the architecture reference model, deriving input factors, and defining the calculation formula, the architecture reference model is selected based on the input value according to the user input, the server specifications and quantity for each asset use are calculated, network, security It is possible to optimize the configuration of , control, etc., calculate cost and priority, adjust influence factors according to priority definition, and store priority and detailed information.

In addition, in this way, it is possible to provide an initial draft when constructing an artificial intelligence and big data center based on open source.

Hereinafter, a scale calculation method for building an artificial intelligence and big data center according to the present invention will be described in detail with reference to FIGS. 1 to 51 .

45 to 46 are flowcharts illustrating a scale calculation method for constructing an artificial intelligence and big data center according to an embodiment of the present invention.

First, the control unit 150 according to the user input (or user/administrator selection/touch/control) of the artificial intelligence and big data center construction apparatus 100 on the information input screen displayed on the display unit 130, artificial intelligence and big data One or more input values corresponding to one or more input items requiring user (or designer/administrator) setting are received from among a plurality of input values corresponding to a plurality of input items required for size calculation for building a center. Here, the plurality of input items are items necessary for calculating the scale for building artificial intelligence and big data center, and are preset according to the designer's design, and among the plurality of input values, one or more input values requiring the user setting are selected. The remaining input values (or at least one remaining input value) may have fixed values (or default values/constants) set in advance according to a designer's design. In addition, the at least one input value requiring user setting is a value that a user must directly input according to an environment among the plurality of input values, and is an input value related to hardware capacity (eg, the number of simultaneous users of external portal/internal portal, external Annual growth rate of simultaneous portal/internal portal users, number of tasks per external portal/internal portal user, number of transactions per external portal/internal portal task, number of operations per external portal/internal portal user, external portal/internal portal application load, etc.); Input values related to the number of people using the software (eg, including the number of users using business software (including analysis, etc.) Input values related to (eg, daily data collection and processing capacity, data storage capacity, etc.), input values related to network configuration (eg, whether network redundancy is configured, whether network is divided, etc.), input related to quality characteristics Include values (including, for example, performance criticality, availability criticality, security criticality, etc.) and input values related to the operational period (including operational period, etc.).

For example, as shown in FIG. 47 , the first control unit 150 controls the number of users (eg, unspecified) of the first external portal service according to the manager input on the information input screen 4700 displayed on the first display unit 130 . 1216 people), the ratio of concurrent users of the first external portal (eg 10%), the annual increase rate of the concurrent users of the first external portal (eg 5%), the number of tasks per user of the first external portal (eg, 10%) 2), the number of transactions per task of the first external portal (eg 5), the number of operations per user of the first external portal (eg 4), the load of the first external portal application (eg, 2.2 preset to 'top') ), the number of concurrent users of the first internal portal (eg, 137), the annual increase rate of concurrent users of the first internal portal (eg, 5%), the number of tasks per user of the first internal portal (eg, 2), the first The number of transactions per internal portal task (eg 5), the number of operations per user of the first internal portal (eg 4), the first internal portal application load (eg 1.5 preset to 'low'), the first task Number of software users (eg 20), number of first affiliated external organizations (eg 15), number of first affiliated internal organizations (eg 15), first daily data processing capacity (eg 7 TB), 1 data storage capacity (eg 1250TB), whether the first network redundancy is configured (eg, corresponding), whether the first network is divided (eg, corresponding), the first performance importance (eg, phase), the first Availability priority (eg, high), first security importance (eg, medium), and a first operating period (eg, 5 years) are received (S4510).

Thereafter, the control unit 150 selects (or selects) an architecture reference model corresponding to the received input value from among a plurality of preset architecture reference models based on the received input value.

That is, the control unit 150 determines whether the network is configured to be redundant, whether or not to divide the network, and whether or not to configure the network included in the input values related to the network configuration among the received input values (eg, whether to configure the redundant network, whether to separate the network, etc.). Based on at least one of virtualization, any one architecture reference model is selected from among eight preset architecture reference models. Here, the eight architecture reference models (or infrastructure architecture reference models) are the first to eighth architecture reference models shown in FIGS. 13 to 20 according to the number and characteristics of the infrastructure architecture selection of Table 1 above. Includes architectural reference models.

For example, the first control unit is based on the received information on whether the first network redundancy configuration (eg, corresponding) and the information on whether the first network is divided (eg, corresponding) in [Table 1] The seventh architecture reference model shown in FIG. 19 is selected according to the number and characteristics of the case for selecting the infrastructure architecture of (S4520).

Thereafter, the controller 150 calculates (or calculates) the tpmC for each asset use and the CPU specification for each asset use based on the received input value.

That is, the controller 150 controls the input value related to the hardware capacity among the received input values according to the calculation items in the database server (not shown) (eg, the number of simultaneous users of the external portal/internal portal, the external portal/internal portal). Concurrent user annual growth rate, number of tasks per external portal/internal portal user, number of transactions per external portal/internal portal task, number of operations per external portal/internal portal user, external portal/internal portal application load, etc.) Calculate tpmC for each asset use based on input values (eg, including operating period). Here, the tpmC is a unit of measurement used when estimating the size of an OLTP server (On-Line Transaction Processing Server). At this time, in the case of the WAS server and the WEB server, the OPS (Open Profiling Standard) of SPECjbb2015 is the standard, and in the embodiment of the present invention, after calculating the OPS, a preset conversion ratio (for example, 2.02) It can be converted to the tpmC by multiplication.

In addition, the controller 150 calculates (or calculates) CPU specifications for each asset use by inferring the core type, number of cores, and clock speed of the CPU using the calculated (or calculated) tpmC.

That is, the control unit 150 applies the calculated tpmC to the variables of [Equation 1], [Equation 2] and [Equation 4] related to the core type of the CPU, the number of cores of the CPU and the clock speed of the CPU. Thus, the CPU specifications for each asset use are calculated. Here, [Equation 1] and [Equation 2] may be in the form of linear equations related to the tpmC, and [Equation 4] may be in the form of applying Newtonian interpolation in which the number of error items is reduced compared to the linear equation.

For example, the first control unit may control '(transactions per minute * basic tpmC correction value * peak type load correction value * DB size correction value * application structure correction value * application load correction value * cluster according to the calculation item in the database server. Calculate tpmC for each asset use according to 'correction value * system margin) / system target utilization rate'. Here, the number of transactions per minute and the application load correction value are included in the received input value, and the basic tpmC correction value, the peak type load correction value, the DB size correction value, the application structure correction value, and the The cluster correction value, the system margin ratio, and the system target utilization ratio may be included in the plurality of input values, and different values may be set as defaults, respectively.

In addition, the first control unit calculates the CPU specification for each asset use by calculating the core type, the number of cores, and the clock speed of the CPU based on the calculated tpmC for each asset use ( S4530 ).

Thereafter, the control unit 150 controls the calculated (or calculated) tpmC for each asset use, the calculated CPU specification for each asset use, and an input value related to the number of software users among the received input values (for example, business software (analysis, etc.) Including) number of users, etc.), input values related to the number of affiliated organizations (including, for example, the number of connected external organizations, the number of connected internal organizations, etc.), and input values related to data capacity (eg, daily data collection and processing capacity, data storage capacity, etc.), calculates (or calculates) the server capacity based on the selected architecture reference model, etc.

That is, the control unit 150 controls the calculated (or calculated) tpmC for each asset use, the calculated CPU specification for each asset use, and an input value related to the number of software users among the received input values (for example, business software (analysis, etc.) Including) number of users, etc.), input values related to the number of affiliated organizations (eg, the number of connected external organizations, number of connected internal organizations, etc.), input values related to data capacity (eg, daily data collection and processing capacity, data storage capacity, etc.), big data platform server capacity, artificial intelligence service server capacity, portal service server capacity, linked service server capacity, and other service server capacity such as security/control, respectively, based on the selected architecture reference model, etc. Calculate.

For example, the first control unit may include the calculated tpmC for each asset use, the calculated CPU specification for each asset use, the number of people using the received first business software (eg 20), and the received number of first linked external organizations (eg, 15), the received first number of associated internal organs (eg 15), the received first daily data processing capacity (eg 7TB), the received first data storage capacity (eg 1250TB) ), the first big data platform server capacity, the first artificial intelligence service server capacity, the first portal service server capacity, the first linked service server capacity, the first other service server capacity, based on the selected seventh architecture reference model, etc. etc. are calculated respectively (S4540).

Thereafter, the control unit 150 selects a server specification for each asset use based on the calculated (or calculated) server capacity.

That is, the control unit 150 controls big data by asset use based on the calculated big data platform server capacity, artificial intelligence service server capacity, portal service server capacity, linked service server capacity, and other service server capacity such as security/control. Platform server specifications, artificial intelligence service server specifications by asset use, portal service server specifications by asset use, linked service server specifications by asset use, and other service server specifications by asset use are selected, respectively.

For example, the first control unit is based on the calculated first big data platform server capacity, first artificial intelligence service server capacity, first portal service server capacity, first linked service server capacity, first other service server capacity, etc. The first big data platform server specifications, the first artificial intelligence service server specifications, the first portal service server specifications, the first linked service server specifications, and the first other service server specifications are respectively selected by asset use (S4550).

Thereafter, the control unit 150 determines the selected asset based on the server specification for each selected asset use and input values related to quality characteristics among the received input values (eg, including performance importance, availability importance, security importance, etc.) Apply (or calculate) the quality characteristic weight to the server specification for each purpose.

That is, the control unit 150 controls the selected big data platform server specifications for each asset use, artificial intelligence service server specifications for each asset use, portal service server specifications for each asset use, linked service server specifications for each asset use, and other service server specifications for each asset use. Each quality characteristic weight is applied according to the input value related to the characteristic (including, for example, performance importance, availability importance, security importance, etc.).

For example, the first control unit receives the selected first big data platform server specification, first artificial intelligence service server specification, first portal service server specification, first linked service server specification, first other service server specification, etc. The quality characteristic weights according to the first performance importance (eg, high), the first availability importance (eg, high), and the first security importance (eg, medium) are applied (S4560).

Thereafter, the control unit 150 calculates (or calculates) the quantity for each asset use based on the server specification for each asset use to which the quality characteristic weight is applied (or calculated).

That is, the control unit 150 controls the big data platform server specification for each asset use to which the quality characteristic weight is applied, the artificial intelligence service server specification for each asset use, the portal service server specification for each asset use, the linked service server specification for each asset use, and other service servers for each asset use. Based on specifications, etc., the quantity for each asset use is calculated.

For example, the first control unit may include a first big data platform server specification, a first artificial intelligence service server specification, a first portal service server specification, a first linked service server specification, and a first other service server to which the quality characteristic weight is applied, respectively. Quantities for each asset use are calculated based on specifications, etc. (S4570).

Thereafter, the control unit 150 optimizes the network configuration, security configuration, and control configuration based on the selected server specification for each asset use and the calculated quantity for each asset use.

That is, the control unit 150 optimizes the network configuration, the security configuration, and the control configuration according to the selected server specification for each asset use and the quantity for each asset use.

For example, the first control unit may include the selected first big data platform server specification, the first artificial intelligence service server specification, the first portal service server specification, the first linked service server specification, the first other service server specification, and the like; Based on the calculated quantity for each asset use, the network configuration is optimized according to the server specification and quantity for each asset use, the security configuration is optimized, and the control configuration is optimized (S4580).

Thereafter, the control unit 150 calculates (or calculates) the cost according to the specification of the server for each asset use and the quantity for each asset use in which the network configuration, the security configuration, and the control configuration are optimized.

That is, the control unit 150 calculates the cost for each asset use with respect to the specification of the server for each asset use and the quantity for each asset use in a state where the network configuration, the security configuration, and the control configuration are optimized.

For example, the first control unit calculates the cost for each asset use in a state in which the network configuration, the security configuration, and the control configuration are optimized (S4590).

Thereafter, the control unit 150 calculates (or calculates) the priority of the scale calculation result based on the information on the change in the importance of quality characteristics and the information on the big data cluster configuration change. Here, the information on the change in the quality characteristic importance includes change information on the performance importance, the availability importance, and the security importance. In addition, the information related to the big data cluster configuration change includes information set according to the infrastructure (or infrastructure configuration) based on the data storage capacity included in the received input value.

That is, the controller 150 sets the first derived scale calculation result as the first priority.

Also, the control unit 150 calculates a priority score by reflecting the information on the change in the importance of quality characteristics. Here, the control unit 150 gives a score (for example, high: 5 points, medium: 3 points, bottom: 2 points) for the importance of quality characteristics for each high, medium, and low. In addition, the control unit 150 derives the number of possible changes. In addition, the control unit 150 derives a difference by digitizing before and after the change. In addition, the control unit 150 sorts in ascending order from the number of cases in which the difference is small.

In addition, the control unit 150 calculates a priority score by reflecting the information related to the big data cluster configuration change. Here, the control unit 150 differentially assigns preset 0.5 points in ascending order from the number of cases having the least influence when changing the cluster configuration. In this case, the cluster configuration change includes name nodes, utility nodes, data nodes, collection nodes, and adjustment of the quantity and configuration of machine learning resources.

In addition, the control unit 150 collects the priority score reflecting the information on the change in the importance of the quality characteristic and the priority score reflecting the information related to the calculated big data cluster configuration change, and the T/F value ( or True/False values).

In addition, the control unit 150 reflects the total score and calculates (or calculates) the priority by removing duplicates.

Also, the control unit 150 defines a priority for the scale calculation result based on the priority for the calculated (or calculated/set) scale calculation result.

For example, the first control unit calculates the priority of each of the scale calculation results based on the information on the change in the importance of the quality characteristic and the information on the big data cluster configuration change.

Also, the first controller defines priorities for the scale calculation results based on the calculated priorities for the scale calculation results ( S4600 ).

Thereafter, the controller 150 adjusts at least one input value (or an influence factor) among the input values according to the priority of the defined scale calculation result.

Also, the controller 150 reflects (or applies) the adjusted at least one input value.

In addition, the control unit 150 repeats the process of adjusting the influence factor until the priority list corresponds to the preset number of priority lists (eg, 30).

For example, the first control unit adjusts at least one input value among the one or more received input values according to the priority of the defined scale calculation result, and reflects the adjusted at least one input value.

In addition, the first control unit generates the priority included in the priority list by the preset number of priority lists (for example, 30), according to the priority definition, the at least one input value (or The process of adjusting the influence factor) is repeatedly performed (S4610).

Thereafter, the control unit 150 determines the server specification for each asset use, the network for each asset use, the configuration of the server optimized for security and control configuration (or the specification of the server for each asset use in which the network, security and control configuration is optimized), adjusted The scale calculation result including the priority and detailed information is stored (or adjusted) in the storage unit 120 based on the priority list in which the input value is reflected. Here, the detailed information includes hardware introduction list and quantity, software introduction list and quantity, infrastructure introduction list and quantity, resource list by asset use such as network security/configuration, server specifications, infrastructure configuration, and the like.

For example, the first control unit includes the selected first big data platform server specification, first artificial intelligence service server specification, first portal service server specification, first linked service server specification, first other service server specification, etc., assets Based on the server configuration with optimized network, security and control configuration for each purpose and the priority list that reflects the input values adjusted according to the number of priority lists (for example, 30), The first scale calculation result is stored in the first storage unit 120 (S4620).

Thereafter, when a priority menu is selected from among the plurality of menus displayed on the display unit 130, the control unit 150 responds to the scale calculation result calculated based on the previously received input value in response to the selected priority menu. A priority list (or a priority list included in the scale calculation result) is displayed (or output) on the display unit 130 . Here, the plurality of menus are for receiving one or more input values corresponding to one or more input items that require user (or designer/administrator) setting among a plurality of input values required for size calculation for artificial intelligence and big data center construction. Information input menu, priority menu to provide a priority list according to the result of sizing calculated according to the input value, resource list menu to provide a list of resources according to the result of sizing, server specification according to the result of sizing (or server specification) for providing the server specification menu, and the infrastructure configuration menu for providing the infrastructure configuration according to the size calculation result. At this time, only when an input value according to a user input is received after the information input menu is selected among the plurality of menus, the priority menu, the resource list menu, the server specification menu, and the infrastructure configuration menu may be activated and displayed, When the input value is not received, the priority menu, resource list menu, server specification menu, and infrastructure configuration menu may be inactivated and displayed.

For example, as shown in FIG. 48 , when a priority menu is selected from among a plurality of menus displayed on the first display unit, the first control unit displays a priority list 4800 according to the scale calculation result ( S4630).

In addition, when a resource list menu is selected from among a plurality of menus displayed on the display unit 130, the control unit 150 is based on the scale calculation result calculated based on the previously received input value in response to the selected resource list menu. A corresponding resource list (or a resource list included in the scale calculation result) is displayed (or output) on the display unit 130 . At this time, only when the priority menu is selected from among the plurality of menus, the resource list menu may be activated and displayed, and when the priority menu is not selected, the resource list menu, server specification menu, and infrastructure configuration menu may be displayed as inactive.

For example, as shown in FIG. 49 , when a resource list menu is selected from among a plurality of menus displayed on the first display unit, the first control unit displays a resource list 4900 according to the scale calculation result on the first display unit. is displayed (S4640).

In addition, when a server specification menu (or server specification menu) is selected from among the plurality of menus displayed on the display unit 130 , the control unit 150 is configured based on the previously received input value in response to the selected server specification menu. The server specification (or server specification/server specification included in the scale calculation result) according to the calculated scale calculation result is displayed (or output) on the display unit 130 . At this time, only when the resource list menu is selected among the plurality of menus, the server specification menu may be activated and displayed, and when the resource list menu is not selected, the server specification menu and the infrastructure configuration menu are inactivated and displayed can be

For example, as shown in FIG. 50 , when a server specification menu is selected from among a plurality of menus displayed on the first display unit, the first control unit displays the server specification 5000 according to the scale calculation result on the first display unit. is displayed (S4650).

In addition, when the infrastructure configuration menu is selected from among the plurality of menus displayed on the display unit 130, the control unit 150 responds to the selected infrastructure configuration menu based on the previously received input value. The corresponding infrastructure configuration (or the infrastructure configuration included in the scale calculation result) is displayed (or output) on the display unit 130 . At this time, only when the server specification menu is selected from among the plurality of menus, the infrastructure configuration menu may be activated and displayed, and when the server specification menu is not selected, the infrastructure configuration menu may be displayed in inactive.

For example, as shown in FIG. 51 , when an infrastructure configuration menu is selected from among a plurality of menus displayed on the first display unit, the first control unit displays the infrastructure configuration 5100 according to the scale calculation result on the first display unit is displayed (S4660).

As described above, the embodiment of the present invention defines an architecture reference model, derives an input factor, defines a calculation formula, selects an architecture reference model based on an input value according to a user input, and selects an architecture reference model for each asset use. and quantity, optimize the configuration of network, security, control, etc., calculate cost and priority, adjust influence factors according to priority definition, save priority and detailed information, and adjust the sizing method Systematization eliminates the need to wait for design results from experts in each field at the initial planning stage, shortens the time required for design collection and decision-making, and reduces the number of cases according to the priority and priority of the target model through repeated calculations. The project is selectively promoted according to the budget and space range that can be accommodated by providing it.

In addition, as described above, the embodiment of the present invention provides an initial plan when constructing an artificial intelligence and big data center based on open source, so that there is no separate software purchase cost, and the architecture can be handled in a controllable range. Therefore, it is possible to calculate a budget that does not deviate significantly from the margin of error, and by appropriately deriving input variables related to business strategy, it is possible to compose an initial plan close to the goal of building artificial intelligence and big data centers.

Those of ordinary skill in the art to which the present invention pertains may modify and modify the above-described contents without departing from the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention, but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. The protection scope of the present invention should be construed by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

100: Sizing device for building artificial intelligence and big data centers
110: communication unit 120: storage unit
130: display unit 140: audio output unit
150: control unit

Claims (10)

a display unit for displaying an information input screen; and
According to the user input on the information input screen, one or more input values corresponding to one or more input items requiring user setting among a plurality of input values corresponding to a plurality of input items required for scale calculation for artificial intelligence and big data center construction are selected. Receive, and select any one architecture reference model from among a plurality of preset architecture reference models based on whether network redundancy configuration, network network separation, and server virtualization included in the input value related to the network configuration among the received input values and calculates the tpmC for each asset use based on the input value related to the hardware capacity and the input value related to the operating period among the received input values, and uses the calculated tpmC to infer the core type, number of cores, and clock speed of the CPU to calculate the CPU specification for each asset use, the calculated tpmC for each asset use, the calculated CPU specification for each asset use, the input value related to the number of software users among the received input values, the input value related to the number of connected organizations, and the data capacity A control unit that calculates server capacity based on the input value and the selected architecture reference model, and selects server specifications for each asset use based on the calculated server capacity,
The control unit is
Scale calculation target including big data platform size, artificial intelligence service size, portal service size, linked service size, security/control system construction size, hardware introduction size, commercial software introduction size, and facility purchase and data center construction size A sizing device for building artificial intelligence and big data centers, characterized in that classifying and managing the sizing target with hardware, software and infrastructure. The method of claim 1,
The plurality of input items are
It is an item necessary for estimating the scale for building artificial intelligence and big data center, which is set in advance according to the designer's design. A sizing device for building artificial intelligence and big data centers, characterized in that the default values are preset according to the The method of claim 1,
One or more input values that require the user setting are
Input values related to hardware capacity, input values related to the number of software users, input values related to the number of connected organizations, input values related to data capacity, input values related to network configuration, input values related to quality characteristics, and input values related to operating period A sizing device for building artificial intelligence and big data centers, comprising at least one of. 4. The method of claim 3,
The input value related to the hardware capacity is
Number of concurrent users of external portal, number of concurrent users of internal portal, annual growth rate of concurrent users of external portal, annual growth rate of concurrent users of internal portal, number of tasks per user of external portal, number of tasks per user of internal portal, number of transactions per task of external portal, number of transactions per task of internal portal at least one of the number of transactions per transaction, the number of operations per user of the external portal, the number of operations per user of the internal portal, an external portal application load, and an internal portal application load;
The input value related to the number of software users is,
Includes personnel using business software;
The input value related to the number of affiliated organizations is
It includes at least one of the number of connected external organizations and the number of connected internal organizations,
The input value related to the data capacity is
at least one of daily data collection and processing capacity and data storage capacity;
The input value related to the network configuration is
It includes at least one of whether network redundancy is configured and whether or not network is divided,
The input value related to the quality characteristic is
includes at least one of performance criticality, availability criticality, and security criticality;
The input value related to the operating period is,
A sizing device for building artificial intelligence and big data centers, characterized in that it includes an operating period. The method of claim 1,
The control unit is
Based on the server specification for each asset use selected and the input value related to quality characteristics among the received input values, a quality characteristic weight is applied to the server specification for each asset use selected, and based on the server specification for each asset use to which the quality characteristic weight is applied. to calculate the quantity for each asset use, and manage the network configuration, security configuration, and control configuration based on the server specification for each asset use selected above and the calculated quantity for each asset use, and for the network configuration, security configuration and control configuration by asset use Calculate the cost according to the quantity of each server specification and asset use, calculate the priority of the sizing result based on the information on the change in the importance of quality characteristics and the information on the big data cluster configuration change, and calculate the calculated size Define a priority for the scale calculation result based on the priority of the result, adjust at least one input value among the input values according to the priority for the defined scale calculation result, and the adjusted at least one By reflecting the input value, the process of adjusting the influence factor is repeatedly performed until the priority list corresponds to the preset number of priority lists, A sizing device for building an artificial intelligence and big data center, characterized in that it stores the sizing result including the priority and detailed information in the storage unit based on the priority list in which the server configuration and the adjusted input value are reflected. According to the user input on the information input screen displayed on the display unit, by the control unit, one or more input items requiring user setting from among a plurality of input values corresponding to a plurality of input items required for size calculation for building an artificial intelligence and big data center receiving one or more corresponding input values;
By the controller, the received input values of the network if the configuration and the network redundancy in the input values related to, data network separation, and whether the server virtualization whether any one of the architecture reference in basis plurality of architecture reference model predetermined as a model selecting a;
calculating, by the controller, tpmC for each asset use based on an input value related to a hardware capacity and an input value related to an operating period among the received input values;
calculating, by the controller, the CPU specifications for each asset use by inferring the core type, number of cores, and clock speed of the CPU using the calculated tpmC;
By the control unit, the calculated tpmC for each asset use, the calculated CPU specification for each asset use, an input value related to the number of software users among the received input values, an input value related to the number of linked organizations, an input value related to data capacity, and the calculating server capacity based on the selected architecture reference model; and
Selecting, by the control unit, a server specification for each asset use based on the calculated server capacity,
The control unit is
Scale calculation target including big data platform size, artificial intelligence service size, portal service size, linked service size, security/control system construction size, hardware introduction size, commercial software introduction size, and facility purchase and data center construction size A method of estimating a scale for building an artificial intelligence and big data center, characterized in that classifying and managing the scale target as hardware, software and infrastructure. 7. The method of claim 6,
applying, by the control unit, a quality characteristic weight to the selected server specification for each asset use and an input value related to a quality characteristic among the received input values, to the selected server specification for each asset use;
calculating, by the control unit, the quantity for each asset use based on the server specification for each asset use to which the quality characteristic weight is applied;
managing, by the control unit, a network configuration, a security configuration, and a control configuration based on the server specification for each selected asset use and the calculated quantity for each asset use;
calculating, by the control unit, a cost according to a specification of a server for each asset use and a quantity for each asset use for a network configuration, a security configuration, and a control configuration;
calculating, by the control unit, the priority of the scale calculation result based on the information on the change in the importance of quality characteristics and the information on the big data cluster configuration change;
defining, by the control unit, a priority for a scale calculation result based on the priority for the calculated scale calculation result;
The control unit adjusts at least one input value from among the input values according to the priority for the defined scale calculation result, and reflects the adjusted at least one input value, so that the priority list is set in advance repeating the process of adjusting the influence factor until it corresponds to the number of lists; and
Scale calculation result including priority and detailed information based on the priority list in which the server specification for each asset use selected by the control unit, the network for each asset use, the server configuration for the security and control configuration, and the adjusted input value are reflected Scale calculation method for building artificial intelligence and big data center, characterized in that it further comprises the step of storing in the storage unit. 7. The method of claim 6,
Calculating the server capacity includes:
Big data platform server capacity, artificial intelligence service server capacity, portal service server capacity, linked service server capacity, and service server capacity related to security/control. . 8. The method of claim 7,
The step of calculating the priority for the scale calculation result is,
The process of setting the first derived scale calculation result as the first priority;
a process of calculating a priority score by reflecting the information on the change in the importance of quality characteristics;
a process of calculating a priority score by reflecting information related to a change in the configuration of a big data cluster;
a process of collecting the priority score reflecting the information on the change in the importance of quality characteristics calculated and the priority score reflecting the information related to the calculated big data cluster configuration change, and storing the T/F value; and
A sizing method for building artificial intelligence and big data centers, characterized in that it includes the process of determining the priority by reflecting the total score and removing duplicates. 8. The method of claim 7,
When a priority menu is selected from among the plurality of menus displayed on the display unit, a priority list according to a scale calculation result calculated based on an input value previously received in response to the selected priority menu is displayed by the control unit. displaying on the display unit;
When a resource list menu is selected from among the plurality of menus displayed on the display unit, the display unit displays a resource list according to a scale calculation result calculated based on an input value previously received in response to the selected resource list menu by the control unit. to mark on;
When a server specification menu is selected from among the plurality of menus displayed on the display unit, the display unit displays, by the control unit, a server specification according to a scale calculation result calculated based on an input value previously received in response to the selected server specification menu. to mark on; and
When the infrastructure configuration menu is selected from among the plurality of menus displayed on the display unit, the display unit displays the infrastructure configuration according to the scale calculation result calculated based on the previously received input value in response to the selected infrastructure configuration menu by the control unit. Scale calculation method for building artificial intelligence and big data center, characterized in that it further comprises the step of displaying in the.

Images