JP2011060053A - Sizing support system, sizing support method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a sizing support system which estimates hardware requirements with high request accuracy in the construction of an application system. <P>SOLUTION: A sizing support server 100 is provided with measurement models, relative performance ratios of machines and processing pattern templates stored in a database. When a data quantity to be treated by an application of a construction object system, a processing pattern thereof and machine information are received from a client, the sizing support server constructs a measurement system for measuring processing performance in a client, and executes a pseudo application of an application system used in the measurement system on the client to measure machine performance (execution CPU time or usage memory amount). The sizing support server 100 corrects a value measured by the measurement system for a construction object machine by the use of measured machine information, construction object machine information and the relative performance ratio of the machine performance, and calculates a machine scale satisfying system performance requirements based on the corrected value. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a sizing support system, a sizing support method, and a program for estimating hardware requirements with high required accuracy when a system is constructed.

  As a conventional technique for supporting system construction, such as client / server system construction, there is a method for automatically creating system construction parameters by providing sizing requirements and system requirements of a server to be constructed and supporting server construction. Yes (see, for example, Patent Document 1).

  Sizing refers to estimating the assumed system load when constructing the system and preparing servers with sufficient performance and number to handle it. It also includes enhancing the server performance in response to an increase in load after starting the operation of the system.

  As a method of estimating processing performance and performing sizing, there is a method of analyzing the details of an application program and calculating the number of dynamic steps and the amount of memory required for each processing.

JP 2006-72772 A

  In the server construction support method described in Patent Document 1, a system pattern to be constructed is prepared in advance, and sizing configuration information that satisfies a system performance requirement by preparing a sizing formula for each registered system model is prepared. Ask. In order to perform high-precision sizing by this method, the system also handles the data volume and processing contents of the applications used in the system to be built, and machine specifications (CPU (Central Processing Unit) processing performance, memory, etc.). It is necessary to agree with what is assumed in the model in advance.

  The method of preparing a sizing calculation formula for each system model of Patent Document 1 is effective in the case of a system that completely matches the system model, but in an actual system construction, there is almost no such thing, If sizing is performed only with a fixed sizing calculation formula, there is a problem that an estimation error occurs and the cost of reviewing the system in a later process is high.

  Also, if you take a method of creating a prototype and measuring processing performance and calculating a sizing calculation formula every time you build a system, the applications and hardware configurations that implement the system are diversified, so all It is very difficult to prepare a sizing formula for the combination in advance, and system support cannot be performed efficiently.

  In addition, since it is necessary to perform sizing for the purpose of estimating the required machine size before building the system, there is no need for a machine that meets the system performance requirements without actually using application software (program) or hardware (machine). It is necessary to estimate the processing performance.

  On the other hand, by analyzing the details of the application program, the number of dynamic steps required for each process (the total number of machine language instructions executed by the CPU when a program that realizes a process is actually run) and the amount of memory are calculated. The method of calculating and estimating the performance is an effective means when the detailed processing method is clear, but it is difficult to analyze details such as the number of dynamic steps before system construction. In addition, in the case of a program written in Java (registered trademark), it may depend on the performance of Java VM (Virtual Machine), and the amount of memory necessary for the process is calculated only by knowing the detailed processing of the application program. It is difficult.

  An object of the present invention is to provide a sizing support system, a sizing support method, and a program for estimating hardware requirements with high required accuracy when an application system is constructed. To do. An application system refers to an information processing system that executes business applications.

  In order to achieve the above object, the sizing support system is a sizing support system that supports calculation of a machine scale that satisfies the system performance requirements before the application system is constructed, and the database is a client (for example, the user terminal 500). Is used to generate the application system model name, the processing performance model information (for example, measurement model information 300) including items of the system performance requirement and application requirement necessary for constructing the model, and the application system program. It has a processing pattern template (for example, processing pattern template 360) and machine relative performance ratio information (for example, CPU relative performance ratio 350), which is information on the relative performance ratio of the machine.

  A processing unit (for example, CPU 120) of the sizing support system presents a plurality of models based on the processing performance model information to the client, information on the model selected from the client, and a system necessary for constructing the selected model When the performance requirement, the application requirement, and the construction target machine information are received (for example, step S1209), the measurement system construction program necessary for constructing the measurement system that measures the processing performance when the selected model is executed by the client And a pseudo application of the application system used in the measurement system based on the processing pattern template (for example, step S1210), and the measurement system construction program and the pseudo application are deployed to the client to measure the measurement system. When the processing performance information is received from the measurement system constructed and constructed at the client, the value measured by the measurement system is calculated based on the client machine information, construction target machine information, and machine relative performance ratio information included in the processing performance information. It is corrected for the machine to be constructed, a machine scale that satisfies the system performance requirements is calculated based on the corrected value (for example, step S1408), and the calculated machine scale is presented to the client (for example, step S1213). And

  The database further includes system construction history information (for example, system construction history information 450) including system performance requirements, application requirements, system builder information, and past system construction information input in the past. When receiving a history search request from the client, the unit refers to the system construction history information and presents it to the client as a reference value of an input item for building the application system (for example, step S1218).

  According to the present invention, it is possible to estimate hardware requirements with high required accuracy when building an application system.

It is a figure showing composition of a sizing support system concerning an embodiment. It is a figure which shows the structure of a sizing assistance server. It is a figure which shows the structure of a sizing assistance database server. It is a figure which shows the structure of a system construction history database server. It is a figure which shows the structure of the user terminal after measurement system construction. It is a figure which shows the example of a screen which selects the performance measurement model in a user terminal. It is a figure which shows the example of measurement model information. It is a figure which shows the example of measurement model specific information. It is a figure which shows the example of system construction history information. It is a figure which shows the example of a table produced in the database of the measurement system. It is a figure which shows the example of a screen which inputs a system performance requirement, an application requirement, and the machine information of construction object. It is a flowchart which shows the procedure of the process which performs sizing assistance. It is a flowchart which shows the procedure of the process which performs pseudo | simulation application automatic generation. It is a flowchart which shows the procedure of the process which performs a performance measurement process. It is a figure which shows the example of a display of the result which a sizing assistance system outputs.

DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the drawings.
FIG. 1 is a diagram showing a schematic configuration of a sizing support system according to an embodiment of the present invention. The sizing support system 1 includes a sizing support server 100, a sizing support database server 200, a system construction history database server 400, and a user terminal 500 (client), which are connected to each other via a network 600.

  The system constructed by the sizing support system 1 is a system patterned by the system construction pattern provided by the sizing support database server 200, and the application to be used is a business application that does not include scientific calculation logic. In particular, the present invention relates to a technique that is effective when using an application program (hereinafter referred to as an application) written in Java that operates on a J2EE (Java 2 Enterprise Edition) server such as a Web system. The sizing support system 1 calculates a machine scale (the number of machines and the number of CPU cores) (see FIG. 15) that satisfies the estimated system performance requirements without creating an actual application with respect to the processing load of the application.

  Specifically, the system performance requirement and application requirement are input from the user terminal 500, and based on the information stored in the sizing support database server 200, the sizing support server 100 includes a performance measurement system and a performance measurement system. Automatically generate a pseudo-application to be used in. The sizing support server 100 executes the pseudo application generated in the user terminal 500 by the performance measurement system, collects processing performance, and performs sizing.

  The performance measurement system refers to a performance measurement system constructed in the user terminal 500 based on data acquired from the user terminal 500. Specifically, a measurement system construction program, which is a program necessary for constructing a performance measurement system (hereinafter referred to as a measurement system), includes a definition file 541, a construction script 542, and an SQL for table creation shown in FIG. A sentence 543 and a processing performance acquisition application 142 are included. Details will be described later with reference to FIG. As a program used when executing the measurement system, there is a pseudo application 141 of an application system desired to be constructed by the user.

  The sizing support server 100 receives sizing data (the number of machines, CPU) based on the selected performance measurement model (see FIG. 6) and the input system performance requirement and application requirement (see FIG. 11) from the user terminal 500. Core number). The sizing support server 100 will be described in detail with reference to FIG.

  The sizing support database server 200 stores data necessary for the sizing support server 100 to generate sizing data. The sizing support database server 200 will be described in detail with reference to FIG.

  The system construction history database server 400 stores data input to the sizing support system 1, processing performance measured using the sizing support system 1, calculated sizing data, and system information actually constructed. The system construction history database server 400 will be described in detail with reference to FIG.

  The user terminal 500 is a terminal of a user (for example, a system engineer) who constructs a system using the sizing support system 1. A user selects a performance measurement model via the user terminal 500, inputs items such as system performance requirements and application requirements, and transmits them to the sizing support server 100. For example, a general-purpose Web browser is used as an interface for setting items. Any terminal connected to the network can be used as the user terminal 500.

  FIG. 2 is a diagram illustrating the configuration of the sizing support server. Reference is made to FIG. 1 as appropriate. The sizing support server 100 includes a main storage device 110, a CPU 120 (processing unit), and an interface 130, and is connected with a disk drive 140. The main storage device 110 stores a sizing support processing unit 111, a pseudo application automatic generation processing unit 112, and a performance measurement processing unit 113.

  The sizing support processing unit 111 executes a pseudo application automatic generation processing unit 112 that generates a pseudo application 141 necessary for sizing, and builds a performance measurement system and calculates a sizing data from the measured processing performance data. 113 is a program that executes step 113 and outputs sizing data obtained as a result.

  The pseudo application automatic generation processing unit 112 is a program that generates the pseudo application 141 according to the performance measurement model selected by the user terminal 500 and the application requirements. The generated pseudo application 141 is stored in the disk drive 140.

  The performance measurement processing unit 113 constructs a measurement system in the user terminal 500 based on the performance measurement model selected by the user terminal 500, and acquires the pseudo application 141 and processing performance data generated by the pseudo application automatic generation processing unit 112. This is a program that is deployed to the measurement system in which the processing performance acquisition application 142 is constructed. Furthermore, it has a function of executing the deployed pseudo application 141 on the user terminal 500, acquiring processing performance data from the processing performance acquisition application 142, and calculating sizing data.

  The CPU 120 executes programs stored in the main storage device 110 and executes various processes. The interface 130 receives input from the user terminal 500 and connects to the network 600 to access the sizing support database server 200 and the system construction history database server 400.

  FIG. 3 is a diagram showing the configuration of the sizing support database server. Reference is made to FIG. 1 as appropriate. The sizing support database server 200 includes a main storage device 210, a CPU 220, and an interface 230, and is connected to a disk drive 240. The disk drive 240 stores data provided to the sizing support server 100.

  The main storage device 210 stores a program executed by the CPU 220 and data necessary for executing the program. Specifically, a database management system 211 for providing data to the sizing support server 100 is stored. The CPU 220 executes a program stored in the main storage device 210 and executes various processes. The interface 230 is an interface connected to the network 600 in order to access the sizing support server 100.

  The disk drive 240 includes measurement model information 300 (see FIG. 7), measurement model specific information 310 (see FIG. 8) including a definition file 320 and a construction script 330, a database table creation template 340, a CPU relative performance ratio 350 (machine relative). Performance ratio), a processing pattern template 360 is stored.

  When performing sizing based on the embodiment of the present invention, first, the sizing support server 100 presents a performance measurement model to the user terminal 500 based on the measurement model information 300 (see FIG. 6, in the figure). In this case, the user selects a performance measurement model. The performance measurement model registration information (measurement model information 300, measurement model specific information 310) will be described with reference to FIGS.

  FIG. 7 is a diagram illustrating an example of measurement model information. The measurement model information 300 stores an ID (Identification) 301, a measurement model name 302 of a performance measurement model, model image data 303 illustrating the model configuration, a system performance requirement 304, and an application requirement 305. ID 301 is uniquely held for each measurement model. FIG. 6 shows an example in which a list of measurement models (performance measurement models) is displayed on the user terminal 500 based on the measurement model information 300 in accordance with a processing request from the sizing support server 100. Details of FIG. 6 will be described later.

  FIG. 8 is a diagram illustrating an example of measurement model specific information. The measurement model specific information 310 includes an ID 301 necessary for identifying the measurement model, a definition file 320 necessary for constructing the measurement system, and a construction script 330. The measurement model specific information 310 corresponding to the performance measurement model selected by the user is stored as data necessary for constructing a measurement system corresponding to the selected processing performance measurement model.

  Returning to FIG. 3, the database table creation template 340 stores a SQL sentence template necessary for creating a table on the server based on the application requirements input from the user terminal 500.

  The CPU relative performance ratio 350 stores the CPU relative performance ratio based on public information such as SPECint (Standard Performance Evaluation Corporation Integer benchmark) that evaluates the performance of the CPU.

  In the processing pattern template 360, a pseudo application 141 (see FIG. 2) is created based on the application requirements input from the user terminal 500. Therefore, the screen display (input to the application) is performed in the processing contents of the application. And a program template for database access processing.

  FIG. 4 is a diagram showing the configuration of the system construction history database server. Reference is made to FIG. 1 as appropriate. The system construction history database server 400 includes a main storage device 410, a CPU 420, and an interface 430, to which a disk drive 440 is connected. The disk drive 440 stores data provided to the sizing support server 100, for example, system construction history information 450 (see FIG. 9).

  The main storage device 410 stores a program executed by the CPU 420 and data necessary for executing the program. Specifically, a database management system 411 for providing data to the sizing support server 100 is stored. The CPU 420 executes a program stored in the main storage device 410 and executes various processes. The interface 430 is an interface connected to the network 600 in order to access the sizing support server 100.

  FIG. 9 is a diagram illustrating an example of system construction history information. The system construction history information 450 stores a user name 455, a system builder skill 451, a construction system configuration 452, a system performance requirement 453, and an application requirement 454.

  The skill 451 stores the number of years in which the user has been engaged in system construction and the levels of advanced, intermediate, and elementary levels based on self-evaluation. The construction system configuration 452 stores machine information (machine specifications), operating OS (Operating System), used software information, and processing performance information of a system constructed in the past. The system performance requirement 453 stores the system performance requirement input from the user terminal 500 and received by the sizing support server 100. The application requirement 454 stores the application requirement input from the user terminal 500 and received by the sizing support server 100.

  The system construction history information 450 shown in FIG. 9 includes past system construction information (for example, construction system configuration 452), system performance requirement information (for example, system performance requirement 453) input to the sizing support system 1, and application requirement information. (For example, application requirements 454), when the information of the system builder (for example, skill 451) is accumulated, and what value should be input by the system user for the items to be input to the sizing support system 1 It is used when providing historical information as a reference.

  FIG. 5 is a diagram showing the configuration of the user terminal after the construction of the measurement system. Reference is made to FIG. 1 as appropriate. The configuration illustrated in FIG. 5 illustrates a configuration after the measurement system is constructed in the user terminal 500 and the pseudo application 141 and the processing performance acquisition application 142 are deployed.

  The user terminal 500 includes a main storage device 510, a CPU 520, and an interface 530, to which a disk drive 540 and an input / output device 550 are connected. The disk drive 540 stores data distributed from the sizing support server 100, and specifically stores a definition file 541, a construction script 542, and an SQL statement 543.

  The main storage device 510 stores a program executed by the CPU 520 and data necessary for executing the program. Specifically, a pseudo application 141, a processing performance acquisition application 142, and a J2EE server 511 for providing data to the sizing support server 100 are stored. The CPU 520 executes a program stored in the main storage device 510 and executes various processes. The interface 530 is an interface connected to the network 600 in order to access the sizing support server.

  The pseudo application 141 and the processing performance acquisition application 142 are deployed with a pseudo application created by the sizing support server 100. The J2EE server 511 may download a setup program from the sizing support server 100 or store it in the user terminal 500 in advance.

  When the user gives an instruction to execute the construction script 542 in the user terminal 500 via the input / output device 550, the measurement system defined by the definition file 541 is constructed. Furthermore, when the SQL statement 543 is executed, a table (see FIG. 10) that is accessed from the pseudo application 141 is created in the database of the constructed measurement system. The execution of the construction script 542 and the execution of the SQL statement 543 may provide a GUI (Graphical User Interface) so that it can be executed from a Web browser displayed on the screen of the input / output device 550, or a command prompt screen. It is also possible to use a command from the command.

  FIG. 10 is a diagram illustrating an example of a table created in the database of the measurement system. The example shown in FIG. 10 is a case where the number of items handled in one record is 3, and is a table prepared with three columns. The table 560 includes, for example, a column A, a column B, and a column C. The table 560 may be registered with the maximum number of reference records.

Next, a screen example of the input / output device 550 of the user terminal 500 will be described.
FIG. 6 is a diagram illustrating an example of a screen for selecting a measurement model. The user examines and selects which model is close to the system to be built from the list of performance measurement models (left column) corresponding to the system pattern to be built displayed on the screen 551. In the example shown in FIG. 6, a JSP / Servlet-DB model is specified. An outline of the designated model is shown in the right column of the screen 551, and it can be confirmed whether or not the designated model is a model close to the system to be constructed. JSP is an abbreviation for JavaServer Pages, and DB is an abbreviation for Data Base.

  FIG. 11 is a diagram showing an example of a screen for inputting system performance requirements, application requirements, and machine information to be constructed. A screen 552 is displayed on the screen of the input / output device 550. First, the user terminal 500 receives a measurement system model designation, and a “measurement system” is generated in the user terminal 500 based on the selected measurement model. For example, sizing is performed so as to satisfy the system performance requirement from the processing performance acquired by the measurement system created using the sizing support server 100 based on information such as the target response time and the number of requests per second as the system performance requirement. .

  The application requirements are, for example, the maximum number of reference records, the number of items handled in one record, the average data size of each item, and the maximum number of simultaneous accesses. Based on the information, a table 560 (see FIG. 10) is created by executing an SQL statement in the database, and the sizing support server 100 generates a pseudo application 141 and a processing performance acquisition application 142 including processing for accessing the table 560. Then, the processing performance is measured by deploying and operating the pseudo application 141 and the processing performance acquisition application 142 created in the user terminal 500.

  FIG. 12 is a flowchart illustrating a processing procedure for performing sizing support. The process shown in FIG. 12 shows the relationship between the processes performed by the sizing support server 100 and the user terminal 500. The processing on the sizing support server 100 side is mainly performed by the sizing support processing unit 111.

  In order for the user who constructs the system to size the system, the sizing support server 100 displays a login screen on the user terminal 500 (step S1201), and the user selects the sizing support system from the login screen of the user terminal 500. 1 is logged in (step S1202).

  The sizing support server 100 acquires the measurement model information 300 from the sizing support database server 200, and based on the measurement model information 300, a list of performance measurement models is displayed on the screen of the input / output device 550 of the user terminal 500 (step) S1203). An example of the displayed screen is shown in FIG. 6 (see screen 551). The user terminal 500 receives the selection of the measurement model from the performance measurement model list from the user, and transmits it to the sizing support server 100 (step S1204).

  The sizing support server 100 acquires the system performance requirement 304 and application requirement 305 items (requirement definition) stored in the measurement model information 300 corresponding to the received model from the sizing support database server 200 (step S1205). Based on the acquired data, a screen (requirements definition input screen) for receiving the requirement items is displayed on the user terminal 500 (step S1206). The user inputs the requirements via a screen that accepts the system performance requirements, application requirements, construction target machine information, and the like received by the user terminal 500 (step S1207). An example of the displayed screen is shown in FIG.

  The user determines whether there is an item that is difficult to input (step S1208). If there is no item that is difficult to input, that is, if all items can be input (No in step S1208), the user inputs the input item and presses the completion button (see FIG. 11). When the user terminal 500 receives the pressing of the completion button (step S1209), the user terminal 500 transmits the information input to the sizing support server 100.

  The sizing support server 100 receives each requirement definition (system performance requirement, application requirement, DB access method, etc.) and machine information from the user terminal 500. The sizing support server 100 automatically generates the pseudo application 141 based on the application requirements and the DB access method input from the user terminal 500 (step S1210). A flowchart of the pseudo-application creation processing procedure will be described with reference to FIG.

  The sizing support server 100 generates a corresponding definition file 320, construction script 330, and database table creation template 340 from the measurement model specific information 310 based on the ID that uniquely identifies the performance measurement model selected in step S1204. Obtained from the server 200 (step S1211).

  Then, the sizing support server 100 performs a performance measurement process (see FIG. 14) (step S1212). Specifically, a program necessary for constructing a performance measurement system is transmitted to the user terminal 500, and the measurement system is constructed in the user terminal 500. The sizing support server 100 calculates sizing data based on the processing performance data acquired by the user terminal 500. A flowchart of the performance measurement process will be described with reference to FIG.

  The sizing support server 100 displays the performance measurement processing result (sizing data) on the user terminal 500 (step S1213). The user confirms the sizing data displayed on the user terminal 500 (step S1214). A display example of sizing data is shown in FIG.

  In step S1208, if there is an item that is difficult to input in the displayed requirement definition (step S1208, Yes), the user presses the history search button (see FIG. 11). When the user terminal 500 accepts the press of the history search button (history search request) (step S1215), the system construction history has many items that match the items already input when the history search button accepts the press. The information 450 is acquired from the system construction history database server 400, and a system construction example having a high coincidence rate with the input items is searched (step S1216). The data that can be acquired is the system construction history information 450 shown in FIG.

  For example, when system performance requirements and construction target machine information are confirmed, history information whose values match or are close to each other is acquired. If there are many examples of similar system construction history information 450, the scope is narrowed down to system construction history information 450 constructed by an advanced person based on the skills of the stem builder (step S1217). Then, an example of the system construction history information is displayed on the user terminal 500 (step S1218), and the user can input with reference to the example of the system construction history information (step S1219). When the input is completed, the process proceeds to step S1209.

  FIG. 13 is a flowchart showing a procedure of processing for automatically generating a pseudo application. The process illustrated in FIG. 13 is a process performed by the pseudo application automatic generation processing unit 112 of the sizing support server 100.

  As the processing outline, a pseudo application is created in accordance with application requirements input from the user terminal 500 using a program for creating a screen and a template for a data access program from the processing pattern template 360 stored in the sizing support database server 200. 141 is a processing procedure for generating 141.

  The sizing support server 100 acquires a processing pattern template from the sizing support database server 200 (step S1301). Based on the application requirements received from the user terminal 500, the sizing support server 100 uses a processing pattern template for creating a GUI input text box, and repeats the data items handled by one record several times (step S1302). A text box that accepts an input value is created by repeating the number of data items handled in one record (step S1303).

  Next, the sizing support server 100 uses the database table creation template 340 to create a program that implements database access processing. The database table creation template 340 is prepared in advance, for example, a program for searching a database and acquiring the search result or a program for updating a record in the database.

  The sizing support server 100 determines whether the DB access is a reference (step S1304). If the DB access is a reference (step S1304, Yes), a program for issuing an SQL statement to be referred to the DB is created (step S1307), and the process proceeds to step S1306. When the DB access is not a reference (step S1304, No), the sizing support server 100 determines whether the DB access is an update (step S1305). When the DB access is update (step S1305, Yes), a program for issuing an SQL statement for updating the DB is created (step S1308), and the process proceeds to step S1306. If the DB access is not an update (step S1305, No), the process proceeds to step S1306. Finally, the sizing support server 100 merges the programs realized by the respective processes to generate the pseudo application 141 (step S1306).

  Specifically, it is a program that accesses the database using the DriverManager class of Java, issues an SQL statement using the Statement class, and acquires the result. The SQL statement part is used as a variable, and the program is generated by appropriately changing the template based on the application requirements.

  FIG. 14 is a flowchart illustrating a processing procedure for performing the performance measurement processing. The process shown in FIG. 14 shows the relationship between the processes performed by the sizing support server 100 and the user terminal 500. The processing on the sizing support server 100 side is processing performed by the performance measurement processing unit 113.

  The sizing support server 100 uses the database table creation template 340 acquired from the sizing support database server 200 in step S1211 of FIG. 12 to create a table creation SQL statement (step S1401). Specifically, a SQL statement for creating a table and creating data in the database is created. For example, if the number of items handled in one record input to the user terminal 500 is 3, this SQL sentence is created with an arbitrary character string corresponding to the average data size of each item in a table having three columns. In this example, a process for realizing the registered data in the table for the maximum number of reference records is described. An example of the table creation is shown in FIG.

  The sizing support server 100 distributes the definition file 320, the construction script 330, and the created SQL sentence acquired from the sizing support database server 200 to the user terminal 500 (step S1402).

  The user terminal 500 executes the received construction script 330 (step S1403), constructs the measurement system, and operates the measurement system.

  Also, the sizing support server 100 distributes the pseudo application 141 and the processing performance acquisition application 142 to the user terminal 500 (step S1404), and the user terminal 500 deploys the received pseudo application 141 and the processing performance acquisition application 142. (Step S1405). Then, the user terminal 500 executes the pseudo application 141 (step S1406). Note that the execution of applications (such as the pseudo application 141 and the processing performance acquisition application 142) may be performed by distributing an automatic execution program to the user terminal 500 or, for example, the pseudo application 141 may be executed manually.

  Then, the user terminal 500 acquires a log (processing performance information) using the processing performance acquisition application 142 (step S1407). Specifically, to obtain the processing performance, measure the memory consumed by JavaVM (Virtual Machine) using JMX (Java Management Extensions), etc., and calculate the memory used per request, OS command, etc. It collects the time from the start of the program to the end of use and the actual processing time. Note that the processing performance information in step S1407 includes machine information of the user terminal 500 (client).

  The sizing support server 100 calculates the number of CPU cores and the used memory size based on the execution CPU time from the log acquired at the user terminal 500, acquires the CPU relative performance ratio from the sizing support database server 200, and executes the process of the construction target machine Sizing data is calculated by calculating the number of machines and the number of CPU cores that satisfy the system performance requirements in terms of performance (step S1408). Then, the calculation result is returned to the sizing support processing unit 111 as sizing data (step S1409).

  Here, a specific example of the method for calculating the number of CPU cores as an example of a machine scale that satisfies the system performance requirements in step S1408 will be described. The sizing support server 100 acquires the processing time (execution CPU time) of the program per request, and calculates the number of requests that can be processed per second. For example, if the execution CPU time is 0.1 second, the number of requests that can be processed per second is 10. However, if the upper limit value of the CPU usage rate is 70% for each measurement model so that the CPU usage rate that is a resource does not become a bottleneck, the number of requests that can be processed in 1 second by one CPU is 7 (the current CPU The number of processing requests at is A).

Assuming that the measured machine information is X (for example, the number of CPU clocks), the construction target machine information is Y, and the relative performance ratio of the machine performance based on the CPU relative performance ratio 350 is Z, the value measured by the measurement system (for example, A) is corrected for the machine to be constructed. Based on this corrected value (A × (Y / X) × Z)), a machine scale that satisfies the system performance requirements can be calculated. That is, assuming that the number of requests per second of the system performance requirement input by the user in FIG. 11 is B, the number of CPU cores satisfying the requested request number is:
Number of CPU cores = B / (A × (Y / X) × Z))
It can be calculated from

  According to the present embodiment, before purchasing the software of the application that is actually used in the system, only the minimum information about the application (information about the processing pattern and data to be processed) and the information about the hardware are input. This makes it possible to estimate hardware requirements that meet system performance requirements.

  Further, reference information can be obtained from past system construction information even when there are items that are difficult to input before system construction in the minimum information. By searching the system construction history using the amount of data handled by the application, the processing method, and the specifications of the machine to be constructed as keys, it is possible to retrieve similar system construction examples from the viewpoint of system performance requirements. Further, by presenting it linked with the skill of the system builder, it is possible to obtain a criterion for narrowing down more effective information from the presented system construction history.

  As shown in FIG. 1, the sizing support system 1 according to the present embodiment includes a sizing support server 100, a sizing support database server 200, a system construction history database server 400, and a user terminal 500. However, the present invention is not limited to this. Is not to be done. For example, a configuration including the sizing support server 100 and the user terminal 500 may be adopted. In this case, information (see FIG. 3) of the disk drive 240 in the sizing support database server 200 and the system construction history database server 400 in the disk drive 140 (database) (see FIG. 2) of the sizing support server 100 Information on the disk drive 440 (see FIG. 4) may be stored.

DESCRIPTION OF SYMBOLS 1 Sizing support system 100 Sizing support server 110,210,410,510 Main storage device 111 Sizing support processing part 112 Pseudo application automatic generation processing part 113 Performance measurement processing part 120,220,420,520 CPU (processing part)
130, 230, 430, 530 Interface 140, 240, 440, 540 Disk drive 141 Pseudo application 142 Processing performance acquisition application 200 Sizing support database server 211, 411 Database management system 300 Measurement model information 310 Measurement model specific information 320 Definition file 330 Construction Script 340 Database table creation template 350 CPU relative performance ratio (machine relative performance ratio)
360 processing pattern template 400 system construction history database server 450 system construction history information 451 skill 452 construction system configuration 453 system performance requirement 454 application requirement 500 user terminal (client)
511 J2EE server 541 definition file 542 construction script 543 SQL statement 550 input / output device 600 network

Claims (7)

A sizing support system that supports the calculation of a machine scale that satisfies the system performance requirements before building an application system,
The database
A name of the model of the application system to be presented to the client, processing performance model information including items of system performance requirements and application requirements necessary for constructing the model, and a processing pattern template used for generating a program of the application system; And machine relative performance ratio information which is information of the relative performance ratio of the machine,
The processing unit
Presenting a plurality of models based on the processing performance model information to the client,
When receiving the information of the model selected from the client, the system performance requirement necessary for the construction of the selected model, the application requirement, and the construction target machine information,
A measurement system construction program necessary for construction of a measurement system for measuring processing performance when the selected model is executed by the client, and a pseudo application of the application system used in the measurement system based on the processing pattern template And generate
Deploying the generated measurement system construction program and the pseudo application to the client to construct the measurement system,
When processing performance information is received from the measurement system constructed in the client, measurement is performed in the measurement system based on the client machine information, the construction target machine information, and the machine relative performance ratio information included in the processing performance information. A sizing support system characterized in that a machine size that satisfies the system performance requirement is calculated based on the corrected value for a machine to be constructed, and the calculated machine size is presented to the client. The database further comprises:
System construction history information including system performance requirements, application requirements, system builder information, and past system construction information input in the past,
The processor is
The sizing support according to claim 1, wherein when a history search request is received from the client, the system construction history information is referred to and presented to the client as a reference value of an input item for construction of the application system. system. The measurement system construction program includes an SQL statement necessary for creating a table on the client based on the application requirements and a processing performance acquisition application for acquiring processing performance when the pseudo application is executed in the measurement system. The sizing support system according to claim 1, further comprising: A sizing support method for a sizing support system that supports calculation of a machine scale that satisfies system performance requirements before building an application system,
The database of the sizing support system includes
A name of the model of the application system to be presented to the client, processing performance model information including items of system performance requirements and application requirements necessary for constructing the model, and a processing pattern template used for generating a program of the application system; , Machine relative performance ratio information that is information of the machine relative performance ratio is stored,
The processing unit
Presenting a plurality of models based on the processing performance model information to the client,
When receiving the information of the model selected from the client, the system performance requirement necessary for the construction of the selected model, the application requirement, and the construction target machine information,
A measurement system construction program necessary for construction of a measurement system for measuring processing performance when the selected model is executed by the client, and a pseudo application of the application system used in the measurement system based on the processing pattern template And generate
Deploying the generated measurement system construction program and the pseudo application to the client to construct the measurement system,
When the machine performance information measured by the measurement system constructed in the client is received, the measurement is performed based on the client machine information, the construction target machine information, and the CPU relative performance ratio information included in the processing performance information. Sizing support characterized by correcting the value measured by the system for the machine to be built, calculating the machine size that satisfies the system performance requirements based on the corrected value, and presenting the calculated machine size to the client Method. The database includes
Stores system construction history information including system performance requirements, application requirements, system builder information, and past system construction information entered in the past,
The processor is
5. The sizing support according to claim 4, wherein when a history search request is received from the client, the system construction history information is referred to and presented to the client as a reference value of an input item for construction of the application system. Method. The measurement system construction program includes an SQL statement necessary for creating a table on the client based on the application requirements and a processing performance acquisition application for acquiring processing performance when the pseudo application is executed in the measurement system. The sizing support method according to claim 4, further comprising: To the sizing support computer that supports the calculation of the machine scale that meets the system performance requirements before building the application system,
A name of the model of the application system to be presented to the client, processing performance model information including items of system performance requirements and application requirements necessary for constructing the model, and a processing pattern template used for generating a program of the application system; , A function of storing machine relative performance ratio information, which is information on the relative performance ratio of the machine, as a database;
A function of presenting a plurality of models based on the processing performance model information to the client;
When receiving the information of the model selected from the client, the system performance requirement necessary for the construction of the selected model, the application requirement, and the construction target machine information,
A measurement system construction program necessary for construction of a measurement system for measuring processing performance when the selected model is executed by the client, and a pseudo application of the application system used in the measurement system based on the processing pattern template A function to generate and
The client deploys the generated measurement system construction program and the pseudo application to construct the measurement system,
When the machine performance information measured by the measurement system constructed in the client is received, the measurement is performed based on the client machine information, the construction target machine information, and the CPU relative performance ratio information included in the processing performance information. A function for correcting a value measured by the system for a machine to be constructed, calculating a machine scale that satisfies the system performance requirement based on the corrected value, and presenting the calculated machine scale to the client;
A program to realize

Images