JP2014174609A - Hardware configuration estimation system, hardware configuration estimation method and hardware configuration estimation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique to automatically generate estimation of hardware configuration which satisfies user requirements, and calculate specifications including hardware selection appropriate for sizing, by developing and evaluating a customer system on a virtual system.SOLUTION: A hardware configuration estimation system comprises information collection means which generates virtual machines for building a virtual system operating an application and executes the application on the virtual system to collect a performance measurement result of resources configuring the virtual system, a physical/virtual performance ratio database which stores a performance ratio of the resources in a physical system and the virtual system, a product database which stores hardware specification information, and configuration information creation means which estimates specifications of resources configuring the physical system on the basis of the performance measurement result and the performance ratio, and generates hardware configuration information on the basis of the specifications of resources and the hardware specification information.

Description

  The present invention relates to a hardware configuration estimation system, a hardware configuration estimation method, and a hardware configuration estimation program in a physical system construction using a virtual system.

  In system construction and sizing, sufficient product knowledge and skilled SI technology are required in order to design a system that satisfies user requirements and avoids over-spec for cost reduction (SI: System Integration). ). However, it is difficult to have a sufficient number of skilled SI engineers in the sales phase.

  One solution to sizing estimation is IaaS that places the virtualization system in the cloud (IaaS: Infrastructure as a Service). However, from the viewpoint of security and execution performance, a system using a physical server is still used firmly today. When a system using a physical server is virtualized, there is an overhead due to its structure. For this reason, simply replacing a system built on a virtual system with a physical system having the same resource results in overspec. Estimating overspecs is a bottleneck in sales activities, so it is necessary to create accurate estimates.

  Patent Document 1 discloses a performance evaluation system that performs performance evaluation of a virtualized system.

  In the performance evaluation system of Patent Document 1, first, a data flow in a process data flow model is converted into a data flow of a virtual server model, and further converted into a data flow on a physical server model. Next, each load of the data flow in the physical server model is calculated, and the load applied to each element of the physical server model is calculated. Furthermore, referring to the physical performance DB, the usage rate per request is calculated from the load information of each element in the physical server model (DB: Database). Finally, performance evaluation is performed based on the usage rate per request of each element in the physical server model and the arrival rate of requests input to the system, and the performance evaluation result is presented to the SE (SE: System) Engineer).

JP 2012-146015 A

  In general, when a system is constructed, it is expensive to prepare a hardware environment equivalent to the production environment from the evaluation stage. For this reason, evaluation is often performed in a small-scale environment. However, there is a problem that hardware configuration estimation must be performed in a situation where not all of the necessary performance evaluation items can be performed due to limitations of the hardware environment.

  In addition, it is necessary to size the user request without excess or deficiency, but deep knowledge about hardware is required for accurate sizing. Therefore, there is a problem that a lot of failure costs occur when an SE with little knowledge about hardware makes a mistake in sizing in estimating the hardware configuration.

  According to Patent Document 1, it is possible to easily estimate the performance of a virtualized system. However, the performance evaluation result is determined by the SE, and the specification calculation including the hardware selection optimum for sizing is not taken into consideration.

  The present invention develops and evaluates a customer system on a virtual system, calculates a specification including hardware selection optimal for sizing, and automatically creates a hardware configuration estimate that satisfies a user request. The purpose is to provide.

  The hardware configuration estimation system of the present invention creates a virtual machine for constructing a virtual system that runs an application used in a physical system, and executes the application on the virtual system to measure the performance of resources constituting the virtual system. Information collection means that implements performance measurement results, physical / virtual performance ratio database that stores resource performance ratio between physical system and virtual system, product database that stores hardware specification information, and performance measurement A configuration information generating unit configured to estimate a spec of a resource configuring the physical system from the result and the performance ratio, and to create hardware configuration information based on the resource spec and the hardware spec information;

  In the hardware configuration estimation method of the present invention, a virtual machine for constructing a virtual system for operating an application used in a physical system is created, and the performance of the resources constituting the virtual system is measured by executing the application in the virtual system. To collect the performance measurement result, and configure the physical system from the performance measurement result and the performance ratio stored in the physical / virtual performance ratio database that stores the performance ratio between the physical system and the virtual system of the resource Estimate resource specifications, and create hardware configuration information based on specifications and spec information stored in a product database that stores hardware and hardware spec information.

  The hardware configuration estimation program according to the present invention includes a process for creating a virtual machine for constructing a virtual system that operates an application used in a physical system, and a performance of a resource that configures the virtual system by executing the application in the virtual system. The physical system from the process of collecting the performance measurement result by performing measurement, the performance measurement result, and the performance ratio stored in the physical / virtual performance ratio database storing the performance ratio between the physical system and the virtual system of the resource Processing for estimating the specs of the resources constituting the hardware, processing for creating hardware configuration information based on the specs and spec information stored in the product database storing the hardware and hardware spec information, and Is executed on the computer.

  According to the present invention, specifications including hardware selection optimal for sizing can be calculated, and the hardware configuration of a customer system evaluated and developed on a virtual system can be automatically estimated so as to satisfy a user request.

It is a block diagram which shows the structure of the product structure estimation system which concerns on embodiment of this invention. It is a figure which shows the cooperation of each process part to the whole system structure and hardware structure production | generation. It is a flowchart of the operation | movement of hardware configuration estimation in the physical system construction using a virtual system. It is a flowchart of the preparation procedure of the virtualization platform implemented with an information collection means. It is a flowchart of information collection operation | movement which implements performance evaluation and is required for structure information generation. It is a flowchart of physical system configuration information generation. It is a figure for demonstrating the difference in the overhead by the difference in each mode in CPU emulation (CPU: Central Processing Unit). It is an example of the physical / virtual performance ratio table regarding the performance ratio of a physical machine and a virtual machine.

  EMBODIMENT OF THE INVENTION Below, the form for implementing this invention is demonstrated using drawing. However, the preferred embodiments described below are technically preferable for carrying out the present invention, but the scope of the invention is not limited to the following.

(Constitution)
FIG. 1 is a block diagram for explaining a hardware configuration estimation system according to an embodiment of the present invention. FIG. 2 is a conceptual diagram for explaining a configuration of a virtual system including a virtual machine in a virtualization platform created by the hardware configuration estimation system according to the present embodiment.

  The hardware configuration estimation system 1 in FIG. 1 is constructed on the environment of the host OS 3 operating on the physical server 2 as shown in FIG. Although not shown, a guest OS environment different from the host OS 3 may be arranged inside the virtual machine 21. FIG. 2 shows an example of host OS type virtualization in which the host OS 3 operates on the physical server 2, but hypervisor type virtualization that does not use the host OS 3 may be used.

(Hardware configuration estimation system)
As illustrated in FIG. 1, the hardware configuration estimation system 1 according to the present embodiment includes an information collection unit 10, a configuration information generation unit 40, a physical / virtual performance ratio DB 41, and a product DB 42.

(Information collection means)
First, the information collecting unit 10 will be described in detail.

  The information collecting unit 10 mainly executes a virtualization platform creation process and an information collection process for performing a performance measurement of an application operating in the customer system and acquiring a performance measurement result.

  The information collection unit 10 includes an input unit 101, a virtualization platform creation unit 102, a storage unit 105, a collection unit 106, a determination unit 107, an editing unit 108, and a data output unit 109.

  The input means 101 inputs user requests to the hardware such as CPU / memory usage rate, network / input / output flow rate, and the like. The user request may be input each time by SE or the like, or only necessary data may be automatically input from information entered in accordance with a previously prepared format. Note that the user request input from the input unit 101 is stored in the storage unit 105.

  As shown in the conceptual diagram of FIG. 2, the virtualization platform creating unit 102 creates a virtual machine 21 for verifying an application that operates on the customer system in the host OS 3 environment, and creates a virtualization platform. A configuration using a hypervisor instead of the host OS 3 may be used.

  The virtualization platform creation unit 102 includes a virtual machine management unit 103 and an information collection tool creation unit 104.

  The virtual machine management unit 103 creates a virtual machine 21 in the target system 20. For creating a virtual machine, libvirt or the like can be used. The virtual machine management means 103 can create a plurality of virtual machines 21 according to the configuration of the customer system.

  In addition, the virtual machine management unit 103 causes the virtual machine 21 to operate the application 22 and activates an information collection tool described later.

  Further, the virtual machine management means 103 constructs a virtual system corresponding to the customer system on the virtual machine 21 in accordance with the user requirements stored in the storage means 105. User requirements include system requirements including performance. The user requirements include information about specifications required for each resource when executing an application to be operated in the customer system. The virtual machine management unit 103 can construct, for example, resources included in the hardware configuration of the physical system on the virtual system.

  Further, the virtual machine management unit 103 can reconstruct the virtual system by redistributing resources based on the system construction information stored in the storage unit 105. The system construction information includes information related to resource allocation that is helpful in constructing a customer system.

  The information collection tool creation unit 104 installs the information collection tool 23 in each virtual machine 21. In addition, the information collection tool creation unit 104 installs the application 23 in each virtual machine 21.

  By installing the information collection tool 23 and the application 22 by the information collection tool creation unit 104 in the virtual machine 21 created by the virtual machine management unit 103, a virtualization platform is formed.

  The storage unit 105 stores an information collection tool 23 installation file, information on user requests (user requirements), information on resource allocation (system construction information), and the like. The storage unit 105 stores performance measurement results and collected data 30 described later. In FIG. 1, the storage unit 105 is depicted as being disposed outside the virtualization platform creation unit 102, but may be included within the virtualization platform creation unit 102.

  The collection unit 106 is connected to the virtual machine 21 in FIG. 2 so as to be able to perform data communication, and acquires the performance measurement result obtained in each virtual machine 21 from the information collection tool 23 in FIG.

  The determination unit 107 refers to the information related to the user request stored in the storage unit 105 and checks whether or not the performance measurement result acquired by the collection unit 106 from the information collection tool 23 satisfies the user requirement.

  The editing unit 108 edits the performance measurement result collected by the collecting unit 106 into the collected data 30 having a format that can be processed by the configuration information generating unit 40.

  The data output unit 109 transmits the collected data 30 edited by the editing unit 108 to the configuration information generation unit 40.

  The above is the description of the configuration of the information collecting unit 10 and the function of each component.

(Configuration information generation means)
Next, the configuration information generating unit 40 will be described in detail.

  The configuration information generation means 40 mainly executes a hardware configuration list creation process for configuring the physical system.

  The configuration information generation unit 40 includes a data acquisition unit 401, a reference unit 402, a system configuration construction unit 403, and a list output unit 404.

  The data acquisition unit 401 acquires the collection data 30 transmitted from the data output unit 109 of the information collection unit 10.

  The reference unit 402 refers to the physical / virtual performance ratio DB 41 and the product DB 42 and presents the data necessary for the system configuration construction unit 403 to construct the physical system configuration to the system configuration construction unit 403.

  The system configuration construction unit 403 corrects the performance information for each CPU and instruction based on the collected data 30 acquired by the data acquisition unit 401 and the data in the DB presented from the reference unit 402, and the hardware of the physical system A hardware configuration list 50 is created.

  Further, the system configuration construction unit 403 calculates the specifications required in the physical system from the collected data 30 obtained from the performance evaluation of the application in the virtual system and the information on the user requirements included in the physical / performance virtual ratio DB 41. .

  The list output unit 404 outputs the hardware configuration list created by the system configuration construction unit 403.

  The above is the description of the configuration of the configuration information generating unit 40 and the operation of each component.

(Database)
Next, a DB (database) referred to by the reference unit 402 of the configuration information generation unit 40 will be described.

  The physical / virtual performance ratio DB 41 stores the performance ratio between a virtual machine and a physical machine corresponding to each resource. In other words, the physical / virtual performance ratio DB 41 stores the performance ratio between the physical system and the virtual system with respect to each resource (CPU, memory, network, input / output). Further, since the performance ratio of the network and the input / output varies depending on the NIC driver, the physical / performance virtual ratio DB 41 stores information corresponding to each of the network and the input / output (NIC: Network Interface Card).

  The product DB 42 stores specific hardware configuration information. The hardware configuration information stored in the product DB 42 includes information regarding a specific hardware product and specification information regarding the specifications of the product as information regarding hardware configuring the physical system.

  The system configuration construction unit 403 selects a product that satisfies the specifications calculated based on the collected data 30 and the information in the physical / performance virtual ratio DB 41 based on the hardware product specification information stored in the product DB 42.

  The above is the description of the DB referred to by the reference unit 402 of the configuration information generation unit 40. The physical / virtual performance ratio DB 41 and the product DB 42 can be configured by a storage device such as a storage or a hard disk.

  The above is the description of the configuration of the hardware configuration estimation device 1 according to the present embodiment. In addition, the above-mentioned embodiment is an example, Comprising: Various changes and additions to this embodiment are also included in the scope of the present invention.

(Target system)
Here, a virtual system constructed by the hardware configuration estimation system 1 of FIG. 1 will be described with reference to FIG.

  The target system 20 includes a plurality of virtual machines 21 including an application 22 and an information collection tool 23. The target system 20 includes a virtual machine 21 created by the virtualization platform creating unit 102 of the information collecting unit 10. The virtual machine 21 is managed by the virtual machine management unit 103.

  The target system 20 operates on an emulator or virtualization software (not shown). In the case of host OS type virtualization, an emulator or virtualization software operating on the host OS 3 is used, and the virtual machine 21 is constructed on the emulator or virtualization software. In the case of hypervisor type virtualization, the host OS 3 is omitted, and the virtual machine 21 can be constructed on the hypervisor (virtualization software).

  In the example of FIG. 2, as with the target system 20, the information collection unit 10 also operates on the host OS 3. The information collecting unit 10 can be directly operated on the host OS 3 without using an emulator or virtualization software. Further, in the case of a hypervisor type that does not use the host OS 3, the information collecting unit 10 can be operated on the same hypervisor (virtualization software) as the virtual machine 21.

  The virtual machine 21 is created by the virtual machine management means 103 in the virtualization platform creation means 102 of the information collection means 10.

  The virtual machine 21 is a system for constructing, developing, and evaluating a customer system, and a plurality of systems can be created on the physical server 2 as necessary. The virtual machine 21 causes the application 22 that operates in the customer system to operate on the virtual system, measures the performance of the virtual system, and collects the performance measurement results of the virtual system by the information collection tool 23.

  The application 22 is an application that operates on the customer system. The application 22 is installed in the virtual machine 21 by the information collection tool 104 and executed by the virtual machine management unit. In the present embodiment, the performance of the virtual system is evaluated by operating the application 22 on the virtual system, and the amount of resources required in the customer system is estimated.

  The information collection tool 23 is created by the information collection tool creation means 104 in the virtualization platform creation means 102 of the information collection means 10. The information collection tool 23 is installed with an installation file necessary for information collection by the information collection tool creation means 104.

  The information collection tool 23 performs performance measurement of the virtual system constructed in the virtual machine 21 and transmits the performance measurement result to the information collection unit 10.

  The above is the description regarding FIG. FIG. 2 is a conceptual diagram that visualizes the configuration of the target system 20, the arrangement of the virtual machines 21 in the target system 20, and the flow of data from the information collection unit 10 to the configuration information generation unit 40.

(Operation)
Next, an operation example of hardware configuration estimation in a physical system construction using a virtual system will be described using the flowchart of FIG. Note that some operations will be described briefly, and details will be described later.

  First, in FIG. 3, the virtual machine management means 103 of the information collection means 10 creates a virtualization platform in order to construct a virtual system corresponding to the customer system in the virtual machine 21 (step S31). The virtualization platform creation process will be described in detail later with reference to FIG.

  Next, environment setting of the application 22 executed in the customer system is performed by system integration (step S32).

  In system integration, software and technologies that exist within computers and networks are selected and combined to build virtual systems that meet customer system usage objectives. For example, when the virtual machine 21 is created, the virtual machine management unit 103 allocates resources on the virtual machine 21 based on information stored in the storage unit 105 as part of user requirements required for each resource. Allocate and build a virtual system. Further, for example, a virtual system may be constructed by allocating resources on the virtual machine 21 by the input unit 101 based on the system construction information inputted through the input unit 101. In the virtual system constructed in this way, the environment setting of the application 22 may be performed.

  When there are a plurality of requirements for performing performance measurement, a plurality of virtual machines 21 are generated, and performance measurement is performed in each virtual machine 21.

  The information collection tool 23 of each virtualization machine 21 measures the performance of the virtual system in each virtualization machine 21 (step S33). The information collection process will be described later in detail with reference to FIG.

  The collection unit 106 of the information collection unit 10 acquires the performance measurement result from the information collection tool 23, and the determination unit 107 checks whether the performance measurement result satisfies a system requirement (performance requirement) including performance. (Step S34).

  If the performance measurement result does not satisfy the system requirements including performance (No in step S34), tuning is performed (step S35), and performance measurement is performed again (step S33).

  In the tuning, it is checked whether there is a portion where the performance performance of the application 22 operating in the system or the customer system is bad, and if there is a portion where the performance is bad, it is improved. Specifically, the performance measurement of resources such as the I / O amount, CPU load, memory usage amount, etc. when executing the application 22 is performed, and the performance measurement results are improved so as to satisfy the user requirements. For example, if the performance measurement result does not satisfy the user requirements for all resources, the virtual machine management unit 103 may redistribute the resources based on the system construction information stored in the storage unit 105. In step S35, for example, the virtual system on the virtual machine 21 may be tuned by the input unit 101 based on information input via the input unit 101.

  Note that if tuning is performed with the maximum load as requested by the user, it is possible to prevent a situation in which the hardware capability is insufficient when the customer system is constructed as a physical system based on the tuning. Also, if the physical system is constructed with a configuration that satisfies the maximum load requested by the user, it is not necessary to incorporate excessive performance exceeding the user request into the physical system, so it is possible to prevent the physical system from becoming overspec.

  Here, when the performance measurement result satisfies the system requirements including performance (Yes in step S34), the data output unit 109 of the information collection unit 10 sends the performance measurement result (collected data 30) to the configuration information generation unit 40. Is transmitted (step S36).

  The configuration information generation unit 40 generates physical system configuration information based on the information acquired from the information collection unit 10 and the information of the physical / virtual performance ratio DB 41 and the product DB 42, and outputs a hardware configuration list (step S37). ). The configuration information generation process will be described in detail with reference to FIG.

  The above is the description of the operation example of hardware configuration estimation in the physical system construction using the virtual system.

(Virtualization platform creation)
Next, an example of details of the virtual platform creation processing will be described with reference to the flowchart of FIG.

  First, information related to resources such as CPU, memory, network, and input / output necessary for creating the virtual machine 21 is input via the input unit 101 of the information collecting unit 10 (step S41). This input operation can be performed by the operator, but data registered in the storage unit 105 as a user request in advance may be used.

  Next, the virtualization platform creation means 102 of the information collection means 10 creates the required number of virtual machines 21 in the environment of the host OS 3 using the virtual machine management means 103 (step S42). The number of virtual machines 21 to be created may be matched with the number of applications 22 to be verified, for example.

  The information collection tool creation unit 104 of the virtualization platform creation unit 102 installs the information collection tool 23 and the application 22 in each virtual machine 21, and creates a virtualization platform (step S43).

  This completes the description of an example of the details of the virtual platform creation process.

(Information collection process)
Next, an example of details of the information collection process for editing the performance measurement of the virtual system and the performance measurement result into the collection data 30 will be described using the flowchart of FIG.

  In FIG. 5, the virtual machine management means 103 of the information collection means 10 activates the information collection tool 23 in each virtual machine 21 (step S51).

  Each virtual machine 21 starts execution of the application 22 that performs performance measurement (step S52).

  The information collection tool 23 collects the measurement result regarding the memory amount used by the virtual system in each virtual machine 21 (step S53).

  Next, the information collection tool 23 collects measurement results regarding the CPU usage rate, network transaction, and input / output transaction by the virtual system in each virtual machine 21 (step S54). Note that the measurement in step S54 need not be performed simultaneously for each measurement item, and may be set so that each measurement item is performed sequentially.

  Then, the information collection tool 23 collects measurement results regarding the count of instructions executed in each virtual machine 21 (step S55).

  The processing from step S53 to step S55 is repeated until the performance measurement is completed (step S56). Note that the order of processing from step S53 to step S55 may be arbitrarily changed.

  When the measurement is completed, the information obtained by each information collecting tool 23 is passed to the information collecting means 10 (step S57). Note that the performance measurement is set in advance, such as being terminated when the user requirements are satisfied for all resources.

  The collection unit 106 of the information collection unit 10 takes out the performance measurement result connected to the system configuration information from each virtual machine 21 (step S58).

  The editing unit 108 of the information collecting unit 10 edits the performance measurement result connected to the extracted configuration information of each system, and outputs it as the collected data 30 (step S59).

  The above is an explanation of an example of details of the information collection processing necessary for the performance measurement and configuration information generation means 40.

(Physical system configuration information generation)
An example of details of the physical system configuration information creation processing will be described with reference to the flowchart of FIG.

  The system configuration construction unit 403 of the configuration information generation unit 40 reads the collection data 30 acquired by the data acquisition unit 401 from the information collection tool 23 (step S61).

  The system configuration construction unit 403 refers to the physical / virtual performance ratio DB 41 and performs network correction for correcting the performance evaluation result regarding the network flow rate obtained in the virtual system to the specifications required in the physical system (step S62). ). For example, if the performance value related to the network flow rate obtained by the virtual system is multiplied by the performance ratio related to the network flow rate stored in the physical / virtual performance ratio DB 41, the specifications necessary for the physical system can be calculated.

  The system configuration construction unit 403 refers to the physical / virtual performance ratio DB 41 and performs input / output correction for correcting the performance evaluation result regarding the input / output flow rate obtained in the virtual system to the specifications required in the physical system ( Step S63). For example, if the performance value related to the input / output flow rate obtained in the virtual system is multiplied by the performance ratio related to the input / output flow rate stored in the physical / virtual performance ratio DB 41, the specifications required in the physical system can be calculated.

  The system configuration construction unit 403 refers to the physical / virtual performance ratio DB 41, and performs correction for each instruction that corrects the performance evaluation result regarding the instruction count obtained in the virtual system to the specifications required in the physical system. (Step S64). For example, if the performance value related to the instruction count obtained in the virtual system is multiplied by the performance ratio related to the instruction count stored in the physical / virtual performance ratio DB 41, the specifications required in the physical system can be calculated.

  The system configuration construction unit 403 refers to the physical / virtual performance ratio DB 41 and performs CPU correction for correcting the performance evaluation result regarding the CPU obtained in the virtual system to the specifications required in the physical system (step S65). . For example, if the performance ratio related to the CPU stored in the physical / virtual performance ratio DB 41 is multiplied by the performance value related to the CPU obtained in the virtual system, the specifications required in the physical system can be calculated.

  The list output unit 404 outputs the hardware configuration list 50 created by the system configuration construction unit 403 (step S66). The hardware configuration list 50 is created based on the specifications of each resource required in the physical system and the specifications information of the hardware product stored in the product DB 42. For example, it is only necessary to list hardware products having specification information exceeding the minimum specifications of each resource required in the physical system.

  The above is an example of the details of the creation process related to the physical system configuration information of FIG.

(Emulation)
With reference to FIG. 7, the overhead due to the difference in each mode in CPU emulation will be described.

  The way of emulation differs depending on the software instruction.

  According to FIG. 7, the emulation is performed natively in guest mode if it is neither an input / output nor a sensitive instruction. In this case, there is little overhead.

  In the case of a sensitive instruction such as a privileged instruction, the emulation is executed via the KVM module in the host kernel mode (KVM: Kernel-based Virtual Machine). In this case, the overhead is small.

  In the case of input / output, emulation is performed via QEMU in host user mode. In this case, the overhead becomes large.

  Thus, there are emulation methods such as those executed natively and those executed by QEMU, and the overhead differs depending on how the instructions are executed. That is, since correction for each instruction is necessary, information on the performance ratio is stored in the physical / virtual performance ratio DB 41 in advance.

  The physical / virtual performance ratio table 80 in FIG. 8 is an example of a table stored in the physical / virtual performance ratio DB 41, and is a physical / virtual performance ratio table 80 showing performance in process generation. The physical / virtual performance ratio table 80 only needs to summarize the performance ratios of the physical system and virtual system of each resource in any processing executed in the virtual system.

  In FIG. 8, “fork” is a process for generating a copy of the process. System_awk is a process in which, in the virtual system, it is checked whether or not the pattern specified for all the input lines is matched, and if it matches, the corresponding action is executed. System_exit is a normal termination process of the system. System_grep is a process of searching for a line that matches the regular expression from the virtual system and outputting it.

  As can be seen from FIG. 8, the performance degradation of the virtual system with respect to the physical system of each instruction is generally about 30 to 50%. For this reason, when migrating from a virtual system to a physical system, it is necessary to perform performance adjustment by correcting each instruction. Detailed information regarding the physical / virtual performance ratio measured in advance is stored in the physical / virtual performance ratio DB 41.

  The system configuration construction unit 403 of the configuration information generation unit 40 satisfies the user request based on the collection data 30 obtained by editing the performance measurement result collected from the virtual system and the physical / virtual performance ratio of each command measured in advance. Resource specifications can be calculated. Further, according to the product DB 42, product information in which identification information such as a product name or product ID of a specific hardware product is associated with specifications of the product is included (ID: Identification). Therefore, if a hardware specification satisfying the user request is calculated, a hardware configuration list in which specific product names having the specification are described can be obtained.

  The hardware configuration estimation method according to the embodiment of the present invention will be summarized below. The following hardware configuration estimation method is complementary to the hardware configuration estimation system described so far.

  In the hardware configuration estimation method according to the embodiment of the present invention, a virtual system is constructed in a virtualization system such as a cloud before actually constructing a physical system.

  Thereafter, tuning is performed on the virtual system so that the performance required by the user can be obtained. This tuning is preferably performed at the maximum load at the user's request.

  For example, on the hypervisor, a resource measurement tool that can measure and record the ratio of each instruction type, CPU / memory usage, and network / input / output flow rate can be launched to monitor virtual machines. Set and evaluate the maximum load test and scenario test.

  From the evaluation result, the configuration information of the virtual machine and the resource measurement result are output.

  A physical / virtual performance ratio DB that stores the performance ratio between the virtual machine and the physical machine for each resource is prepared in advance. Since the network and input / output have different performance ratios depending on the NIC driver, the physical / virtual performance ratio DB stores information corresponding to each of the network and the input / output. A product DB in which hardware configuration information is stored is also prepared in advance.

  Based on the physical / virtual performance ratio DB and the resource measurement result obtained by the evaluation, the system configuration and the specifications necessary for the physical server are calculated.

  The system configuration and the specifications required for the physical server are compared with the hardware product specification information stored in the product DB, and the configuration estimation information adapted to the user requirements is output as a list.

  The above is a summary of the hardware configuration estimation method according to the embodiment of the present invention. Note that the hardware configuration estimation method according to the present embodiment includes all of the methods described in the description of the hardware configuration estimation system. A program using the hardware configuration estimation method according to the present embodiment is also included in the scope of the present invention.

  According to the embodiment of the present invention, a resource required for a system created on a virtual system is measured by a measurement tool, converted to a resource required by a physical machine (physical server), and an appropriate product is selected from the product DB. Thus, a hardware configuration estimate can be created.

  In addition, according to the embodiment of the present invention, by developing and evaluating a system on a virtual system, it is possible to prevent mistakes related to platforms such as memory and CPU, and to automatically estimate an accurate hardware configuration of a physical system. Can be done.

  Furthermore, according to the embodiment of the present invention, it is possible to calculate specifications including hardware selection optimal for sizing by holding performance information of a plurality of hardware.

  Although the present invention has been described with reference to the embodiments, the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

1 Hardware configuration estimation system 2 Physical server 3 Host OS
DESCRIPTION OF SYMBOLS 10 Information collection means 20 Target system 21 Virtual machine 22 Application 23 Information collection tool 30 Collection data 40 Configuration information generation means 41 Physical / virtual performance ratio DB
42 Product DB
50 Hardware Configuration List 80 Physical / Virtual Performance Ratio Table 101 Input Means 102 Virtualization Platform Creation Means 103 Virtual Machine Management Means 104 Information Collection Tool Creation Means 105 Storage Means 106 Collection Means 107 Determination Means 108 Editing Means 109 Data Output Means 401 Data Acquisition means 402 Reference means 403 System configuration construction means 404 List output means

Claims (10)

A virtual machine for constructing a virtual system for operating an application used in a physical system is created, and the performance measurement result is obtained by executing the application on the virtual system and measuring the performance of the resources constituting the virtual system. Information collecting means to collect;
A physical / virtual performance ratio database storing performance ratios of the resources in the physical system and the virtual system;
A product database that stores hardware spec information;
Configuration information generation for estimating the specifications of the resources constituting the physical system from the performance measurement result and the performance ratio, and creating the hardware configuration information based on the resource specifications and the hardware specification information And a hardware configuration estimation system. The information collecting means includes
The hardware configuration estimation system according to claim 1, wherein an information collection tool that performs performance measurement of the virtual system and collects the performance measurement result is installed in the virtual machine. The information collecting means includes
Virtual machine management means for creating the virtual machine and operating the application and starting the information collection tool;
An information collection tool creating means for installing the information collection tool for performing the performance measurement of the virtual system and collecting the performance measurement result on the virtual machine,
The virtual machine operates the application when the information collection tool is activated by the virtual machine management means,
The hardware configuration estimation system according to claim 2, wherein the information collection tool collects the performance measurement result obtained by the operation of the application. The information collection tool is:
4. The hardware according to claim 3, wherein as the performance measurement, measurement is performed on a memory amount used by the virtual system, a usage rate of a central processing unit, and a flow rate of a network transaction and an input / output transaction. Hardware configuration estimation system. The information collection tool is:
5. The hardware configuration estimation system according to claim 3, wherein instructions executed in the virtual machine are counted. The physical / virtual performance ratio database is:
6. The hardware configuration estimation system according to claim 1, further comprising a physical / virtual performance ratio table including a performance ratio between the physical machine and the virtual machine. The configuration information generating means
Referring to the physical / virtual performance ratio table,
Performing network correction based on the performance measurement results for the network transaction;
Perform I / O correction based on the performance measurement results for the I / O transaction
Performing correction for each instruction based on the performance measurement result regarding the count of the instruction;
Performing correction of the central processing unit based on the performance measurement results relating to the central processing unit;
The hardware configuration estimation system according to claim 6, wherein the configuration of the physical system is estimated. The configuration information generating means
The resource specifications estimated from the performance measurement result and the performance ratio are associated with the hardware specification information stored in the product database,
The hardware configuration estimation system according to any one of claims 1 to 7, wherein a hardware configuration information list describing the configuration information of the hardware that satisfies the user request is created. Create a virtual machine to build a virtual system that runs applications used in the physical system,
Collect the performance measurement results by executing the application in the virtual system and measuring the performance of the resources constituting the virtual system,
From the performance measurement result and the performance ratio stored in the physical / virtual performance ratio database storing the performance ratio of the resource between the physical system and the virtual system, a specification of the resource constituting the physical system is obtained. Estimate,
A hardware configuration estimation method that creates the hardware configuration information based on the specifications and the specification information stored in a product database that stores the hardware and the specification information of the hardware. . A process of creating a virtual machine for building a virtual system that runs an application used in a physical system;
Processing to collect the performance measurement results by executing the application on the virtual system and measuring the performance of the resources constituting the virtual system;
From the performance measurement result and the performance ratio stored in the physical / virtual performance ratio database storing the performance ratio of the resource between the physical system and the virtual system, a specification of the resource constituting the physical system is obtained. Processing to estimate,
Causing the computer to execute a process of creating the hardware configuration information based on the specifications and the specification information stored in a product database storing the hardware and the hardware specification information. Feature hardware configuration estimation program.

Images