JP2011145894A - Information processing system, information processor, information processing method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To estimate a load to be imposed on a resource with high accuracy when a virtual server executes application processing on a physical server. <P>SOLUTION: An information processing system 100 includes: physical servers 31 to 3m; a performance information storage part 26 for storing frequency information of a system call in association with resource information showing the amount of used resource; an acquisition part 21 for acquiring first resource information showing the amount of resource used when a virtual server executes application processing on the physical server 31; a first extraction part 22 for extracting frequency information associated with the physical server 31, and associated with the resource information which is most proximate to the first resource information; a second extraction part 23 for extracting resource information associated with the physical server 32, and associated with the frequency information which is most approximate to the extracted frequency information; and an output part 24 for outputting the extracted resource information. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an information processing system, an information processing apparatus, an information processing method, and a program.

  In order to use physical server resources efficiently, there are physical servers with various performances equipped with various hardware and software in the data center (virtual CPU, virtual storage, virtual network). In recent years, there has been a thriving technological development to support this. In the future, in such a technique, a virtual server built on a physical server can move between physical servers, and resources of the physical server will be used at the movement destination. Therefore, it is useful to manage the movement of the virtual server so that the virtual server can efficiently use resources on the physical server.

  As this type of technology, Patent Document 1 (Japanese Patent Laid-Open No. 2009-123174) describes the following technology. This technique can estimate the CPU load required for I / O emulation in a virtual environment by converting it from at least one of a disk load and a network load.

  More specifically, a plurality of physical servers are connected to a computer terminal that operates a sizing function via a LAN (Local Area Network). Each physical server has at least one CPU (Central Processing Unit) as a hardware resource, and also has an HDD (Hard Disk Drive) and a NIC (Network Interface Card). Moreover, the load measurement part with which each physical server is equipped measures the system load of an own machine, and outputs it as measurement information. The system load is, for example, a CPU load, an I / O (Input / Output) load of a disk or a network.

  When estimating the CPU load when the virtual server is operated on any of the physical servers, the computer terminal operates as follows. First, the computer terminal obtains a measured value of the CPU load of each physical server. Next, the computer terminal obtains an estimated value of the CPU load of the virtual server to be measured due to disk / network I / O from at least one of the disk load or network load of each physical server. Then, the computer terminal obtains a coefficient indicating the CPU overhead due to virtualization. Further, the computer terminal estimates the CPU load of the virtual server to be measured using the measured value, the estimated value, and the coefficient.

  The feature of this technology is to measure the CPU overhead for I / O when using each resource (disk, network) used in the virtual server, and to use each resource used by the application on the physical server. Is to measure. Then, the CPU overhead corresponding to the resource usage frequency is calculated, and the calculated CPU overhead is added to estimate the load when the virtual server executes the application process on the physical server.

JP 2009-123174 A

  However, the above technique uses a coefficient to interpolate the difference between the estimated value of the CPU load that can be obtained from the disk load and the network load and the actual measured value of the CPU load, but the CPU load is only the disk load and the network load. There is a limit to its accuracy because it does not depend on.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an information processing system and information for estimating a load applied to a resource with high accuracy when a virtual server executes application processing on a physical server. To provide a processing device, an information processing method, and a program.

  According to the present invention, a plurality of physical servers that can exchange a virtual server via a communication network, each of the physical servers, frequency information indicating a frequency of executing a system call on the physical server, Performance information storage means for associating and storing resource information representing the resource amount of the physical server used when the system call is executed at the frequency, and the virtual server first performs application processing involving the system call. Acquisition means for acquiring first resource information indicating the amount of resources used in the first physical server when executed on the physical server; and the acquisition means associated with the first physical server and acquired by the acquisition means The resource information closest to the first resource information is selected from the performance information storage means, and the selected resource information is selected. Of the frequency information associated with a second physical server that is different from the first physical server, the frequency information extracted by the first extraction means from the first extraction means that extracts the frequency information associated with the first physical server The frequency information that is most similar to the frequency information storage unit is selected from the performance information storage unit, the resource information associated with the selected frequency information is extracted, and the resource extracted by the second extraction unit An information processing system comprising an output means for outputting information is provided.

  Further, according to the present invention, there is provided an information processing apparatus capable of communicating with each of a plurality of physical servers that can exchange a virtual server via a communication network, and each of the physical servers and the physical servers Performance information storage means for storing the frequency information indicating the frequency of executing the system call and the resource information indicating the resource amount of the physical server used when the system call is executed at the frequency, An acquisition unit configured to acquire first resource information indicating a resource amount used in the first physical server when a virtual server executes an application process involving the system call on the first physical server; and the first physical server And the resource information closest to the first resource information acquired by the acquiring means is the performance information description. A first extraction unit that selects from the means and extracts the frequency information associated with the selected resource information; and the frequency information associated with a second physical server different from the first physical server, Selecting the frequency information that is most similar to the frequency information extracted by the first extraction means from the performance information storage means, and extracting the resource information associated with the selected frequency information; An information processing apparatus comprising: output means for outputting the resource information extracted by the second extraction means.

  Further, according to the present invention, each of a plurality of physical servers that can exchange a virtual server via a communication network, frequency information indicating a frequency of executing a system call on the physical server, and the system call A performance information storage step for associating and storing resource information representing the resource amount of the physical server used when executed at the frequency; and the virtual server performs application processing with the system call on the first physical server The acquisition step of acquiring the first resource information indicating the amount of resources used in the first physical server when executed in step 1, and the first resource associated with the first physical server and acquired in the acquisition step The resource information closest to the information is selected, and the frequency information associated with the selected resource information is selected. The frequency information closest to the frequency information extracted in the first extraction step among the frequency information associated with a second physical server different from the first physical server And a second extraction step for extracting the resource information associated with the selected frequency information, and an output step for outputting the resource information extracted in the second extraction step. An information processing method is provided.

  Furthermore, according to the present invention, a program stored in a computer-readable storage medium, each of a plurality of physical servers that can exchange a virtual server via a communication network, and on the physical server A performance information storage process for associating and storing frequency information indicating a frequency of executing a system call and resource information indicating a resource amount of the physical server used when the system call is executed at the frequency; An acquisition process for acquiring first resource information indicating a resource amount used in the first physical server when a virtual server executes an application process involving the system call on the first physical server; and the first physical server And the resource information that is most similar to the first resource information acquired in the acquisition process is selected. A first extraction process for extracting the frequency information associated with the selected resource information, and the first extraction process among the frequency information associated with a second physical server different from the first physical server. Selecting the frequency information that is most similar to the frequency information extracted in step (b), extracting the resource information associated with the selected frequency information, and extracting the resource information in the second extraction process. An output process for outputting resource information is executed by the computer.

  According to the above invention, there is provided an information processing system, an information processing apparatus, an information processing method, and a program for estimating a load applied to a resource with high accuracy when a virtual server executes application processing on a physical server.

1 is a configuration diagram of an information processing system according to a first embodiment of the present invention. It is a figure showing the information memorized by the performance information storage part. It is a flowchart showing operation | movement of the management server of 1st Embodiment. It is a block diagram of the information processing system which concerns on the 2nd Embodiment of this invention. It is a flowchart showing operation | movement of the performance characteristic acquisition instruction | indication part of 2nd Embodiment. It is a flowchart showing operation | movement of the virtual server for measurement of 2nd Embodiment. It is a flowchart showing operation | movement of the performance information calculation part of 2nd Embodiment. It is a flowchart showing operation | movement of the performance information calculation part of 3rd Embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In all the drawings, the same reference numerals are given to the same components, and the description will be omitted as appropriate.

[First Embodiment]
FIG. 1 is a configuration diagram of an information processing system 100 according to the first embodiment of the present invention. The information processing system 100 includes a communication network 10, a management server 20 (information processing device), and a plurality of physical servers 31 to 3m. The management server 20 and the plurality of physical servers 31 to 3 m are connected to be communicable with each other via the communication network 10. The physical servers 31 to 3 m can exchange virtual servers via the communication network 10. Further, the management server 20 can estimate the resource amount used by the application process before the virtual server executes the application process on the physical servers 31 to 3m. Hereinafter, in the present embodiment, an operation of estimating a resource amount used when a virtual server that has executed an application process on the physical server 31 is operated on the physical server 32 and the application process is executed will be described.

  The communication network 10 may be any network as long as the management server 20 and the physical servers 31 to 3m can be connected to each other, for example, the Internet or a LAN. Further, at least a part of the physical servers 31 to 3m may have different specifications with respect to the CPU, the memory mounting amount, the network bandwidth, and the hard disk. Further, the virtual server may exist anywhere in the information processing system 100. Furthermore, application processing in the present embodiment is assumed to be processing such as Web, mail, and backup.

  The management server 20 includes an acquisition unit 21, a first extraction unit 22, a second extraction unit 23, an output unit 24, and a storage unit 25. The storage unit 25 includes a performance information storage unit 26. The performance information storage unit 26 includes each physical server 31 to 3m, frequency information indicating the frequency of executing the system call on the physical servers 31 to 3m, and the information used when the system call is executed at the frequency. The resource information indicating the resource amount of the physical servers 31 to 3m is stored in association with each other. The performance information storage unit 26 may be a part or all of the storage area included in the storage unit 25. In addition to the information described below, the storage unit 25 may store any information, program, or variable that is processed in the management server 20.

  The acquisition unit 21 acquires first resource information indicating the amount of resources used in the physical server 31 when the virtual server executes an application process involving a system call on the physical server 31. The first extraction unit 22 selects, from the performance information storage unit 26, resource information that is associated with the physical server 31 and that is closest to the first resource information acquired by the acquisition unit 21, and is associated with the selected resource information. Frequency information is extracted. The second extraction unit 23 selects, from the performance information storage unit 26, the frequency information that most closely approximates the frequency information extracted by the first extraction unit 22 among the frequency information associated with the physical server 32 different from the physical server 31. The resource information associated with the selected frequency information is extracted. The output unit 24 outputs the resource information extracted by the second extraction unit 23.

  The system call described here is to call a function of the operating system. When the virtual server performs application processing, in order to use the hardware of the physical servers 31 to 3m, the virtual server needs to execute a system call to the operating system installed in the physical servers 31 to 3m. There are various types of system calls, and there are system calls (write, pwrite, fwrite, mmap, etc.) that perform file read / write processing and system calls (socket, close, etc.) that perform network processing. Different resource loads can be generated by combining multiple types of system calls to be executed. Even for the same type of system call, different resource loads can be generated by specifying a file capacity and executing it, or specifying a packet size and a communication protocol.

  Further, as the resource amount represented by the resource information or the first resource information described here, in this embodiment, the processor usage rate, the memory usage amount, the communication speed with the communication network 10 in each of the physical servers 31 to 3m, The writing speed of the HDD is used. Each of the physical servers 31 to 3m includes means for measuring the amount of resources listed here. In addition, if each of the physical servers 31 to 3m can be measured, the number of occurrences of context switches, the number of occurrences of interrupt processing, the number of waiting processes, and the like may be used as resource information.

  The frequency information described here is the number of executions of the system call executed per unit time. The size of the unit time may be fixed or may be arbitrarily changed. In the present embodiment, the unit time is 1 second.

  Subsequently, each component described above will be described in more detail. FIG. 2 is a diagram illustrating information stored in the performance information storage unit 26. Of course, the table shown here shows the concept of the information stored in the performance information storage unit 26, and the format for storing the information is not limited to this.

  As shown in FIG. 2, the performance information storage unit 26 includes single type information, frequency information corresponding to the type information, a set of plural types of resource information (resource set), and each of the physical servers 31 to 3m. Are stored in association with each other, and an identification number is assigned to each. For example, no. 1 includes type information: system call A, frequency information: 10 (times / s), CPU usage rate: 30 (%), memory usage amount: 512 (MB), HDD writing speed: 50 (KB / s) Communication speed 30 (Kbps) is associated. The performance information storage unit 26 also includes a set of multiple types of type information (type set), a set of frequency information (frequency set) corresponding to each of the type information included in the type set, and a plurality of types of resource information. A set (resource set) and each physical server 31 to 3m are stored in association with each other, and an identification number is assigned to each. For example, no. 16 includes type information: system call A and system call B, frequency information: 10 (times / s) and 1 (times / s), CPU usage rate: 35 (%), memory usage amount: 512 (MB), The HDD writing speed is associated with 60 (KB / s) and the communication speed 10 (Kbps). In addition, No. 16, the frequency information corresponding to the system call A is 10 (times / s), and the frequency information corresponding to the system call B is 1 (times / s). No. 91 includes type information: system call A, system call B, and system call C, frequency information: 10 (times / s), 1 (times / s), and 2 (times / s), CPU usage rate: 50 (% ), Memory usage: 512 (MB), HDD writing speed: 80 (KB / s), and communication speed 0 (Kbps). In addition, No. 91, the frequency information corresponding to the system call A is 10 (times / s), the frequency information corresponding to the system call B is 1 (times / s), and the frequency information corresponding to the system call C is 2 (times). / S).

  The acquisition unit 21 acquires from the physical server 31 a first resource set that is a set of multiple types of first resource information. For example, the acquisition unit 21 sets the CPU usage rate: 40 (%), the memory usage: 512 (MB), the HDD writing speed: 50 (KB / s), and the communication speed 10 (Kbps) as the first resource set. get.

  In the present embodiment, the timing at which the acquisition unit 21 acquires the first resource set may be a predetermined time, or when an instruction input by an operator of the management server 20 is received. Or other conditions may be used as a trigger.

  In the present embodiment, the acquisition unit 21 acquires the first resource set from the physical server 31, but the acquisition destination is not limited to this. For example, if the management server 20 has previously received a first resource set (first resource information) from the physical server 31, the storage unit 25 stores the first resource set, and the acquisition unit 21 stores the first resource set. The first resource set may be acquired from the unit 25. Alternatively, if the amount of resources used in the physical server 31 can be measured by another device not shown in FIG. 1 and the first resource set (first resource information) is stored, the acquiring unit 21 The first resource set may be acquired from the device.

  The first extraction unit 22 selects, from the performance information storage unit 26, resource information that has a minimum deviation from the first resource information acquired by the acquisition unit 21 among the resource information associated with the physical server 31. Extract frequency information associated with the selected resource information. More specifically, the first extraction unit 22 associates each resource set that is associated with the physical server 31 and is the same type of resource information as the first resource set acquired by the acquisition unit 21, and the first resource set. Compare each other and calculate the degree of divergence between each other. Then, the first extraction unit 22 selects a resource set that minimizes the calculated degree of divergence, and extracts a type set and a frequency set associated with the selected resource set. Note that the comparison with the resource information of the same type here means comparing the CPU usage rates included in the resource sets or comparing the memory usage amounts. Here, the degree of divergence is a value obtained by obtaining the square of a value obtained by making the difference between the compared resource information non-dimensional for each resource type and adding them up. In the present embodiment, the sum of the squares of the values obtained by making dimensionless is used, but it may be the sum of the absolute values of the values obtained by making dimensionless.

The calculation formula by which the first extraction unit 22 calculates the degree of divergence can be expressed by the following formula (1).

However, the meaning of each symbol in Formula (1) is as follows.
n: Number of types of resource information (Up ₁ ,... Up _n ): Resource set to be calculated (Ut ₁ ,... Ut _n ): First resource set (Umax ₁ ,... Umax _n ) : A set of maximum values of resource amount used per unit time Vu: Degree of divergence between the resource set to be calculated and the first resource set

  Note that the maximum value of the resource amount used per unit time may be the maximum value among the resource information stored in the performance information storage unit 26, or may be previously set in each physical server 31 to 3m. It may be a maximum value determined by actual measurement or a maximum value set by an operator. However, when the maximum value is set by the operator, it is necessary that each physical server 31 to 3m be restricted so that a resource amount equal to or larger than the set maximum value cannot be used in unit time.

  The second extraction unit 23 selects, from the performance information storage unit 26, frequency information that has a minimum deviation from the frequency information extracted by the first extraction unit 22 among the frequency information associated with the physical server 32. Resource information associated with the selected frequency information is extracted. More specifically, the second extraction unit 23 selects at least one type set that matches the type set extracted by the first extraction unit 22 and is associated with the physical server 32, and sets the selected type set to the selected type set. Each of the associated frequency sets and the frequency set extracted by the first extraction unit 22 are compared, the degrees of divergence are calculated, the frequency set that minimizes the divergence is selected, and the selected frequency A resource set (resource information) associated with the set is extracted. Here, “matching” may be, for example, that all of the type information included in each of the type sets to be compared may match each other, or may partially match each other. Here, the degree of divergence is a value obtained by calculating the square of values obtained by making the difference between the compared frequency information non-dimensional for each type of system call and adding them up. In the present embodiment, the sum of the squares of the values obtained by making dimensionless is used, but it may be the sum of the absolute values of the values obtained by making dimensionless.

The calculation formula used by the second extraction unit 23 to calculate the degree of divergence can be expressed by the following formula (2).

However, the meaning of each symbol in Formula (2) is as follows.
n: number of types of system calls (Ep ₁ ,... Ep _n ): frequency set to be calculated (Et ₁ ,... Et _n ): frequency set (Emax ₁ ) extracted by the first extraction unit 22 ,... Emax _n ): set of maximum values of execution frequency of system calls executed in unit time Ve: degree of divergence between the resource set to be calculated and the first resource set

  Note that the maximum value of the frequency of system calls executed in unit time may be the maximum value in the frequency information stored in the performance information storage unit 26, or may be set in advance for each physical server 31. It may be a maximum value determined by actual measurement at ˜3 m, or may be a maximum value set by an operator. However, when the maximum value is set by the operator, it is necessary that each physical server 31 to 3m be restricted so that each system call cannot be executed with a frequency equal to or higher than the set maximum value. is there.

  The output unit 24 uses the resource set (resource information) extracted by the second extraction unit 23 as an estimated value of the amount of resources used in the physical server 32 when the virtual server executes application processing on the physical server 32. Output. The resource amount indicated by the resource set output by the output unit 24 is an estimated value of the resource amount of the physical server 32 used by the application process. The output destination and the format of the resource information output by the output unit 24 are not particularly limited. For example, the output unit 24 may output the resource information to other components included in the management server 20 or may output the resource information to a device different from the management server 20. The output unit 24 may output the resource information in a data format readable by the processor, may display and output the resource information so that the operator can recognize it, or may print it out.

  All or some of the components included in the management server 20 may be realized by hardware, or a program (or program code) that causes a processor (not shown) included in the management server 20 to execute processing. ).

  When the components included in the management server 20 are implemented by a program, the program is stored in a storage medium such as a non-volatile memory that can be read by a processor (computer). The program includes each of a plurality of physical servers 31 to 3m that can send and receive a virtual server via the communication network 10, and frequency information indicating the frequency of executing a system call on the physical servers 31 to 3m. The processor is caused to execute a performance information storage process for associating and storing resource information representing the resource amount of the physical servers 31 to 3m used when the system call is executed at the frequency. In addition, the program executes, in the processor, an acquisition process for acquiring first resource information indicating the amount of resources used in the physical server 31 when the virtual server executes an application process involving a system call on the physical server 31. Let In addition, the program selects resource information that is associated with the physical server 31 and that is closest to the first resource information acquired in the acquisition process, and extracts frequency information associated with the selected resource information. The processor executes the extraction process. In addition, the program selects frequency information closest to the frequency information extracted in the first extraction process from the frequency information associated with the physical server 32 different from the physical server 31, and sets the selected frequency information. The processor executes a second extraction process for extracting the associated resource information. Further, the program causes the processor to execute an output process for outputting the resource information extracted in the second extraction process.

  FIG. 3 is a flowchart showing the operation of the management server 20 of this embodiment. First, the management server 20 includes each of a plurality of physical servers 31 to 3m that can send and receive virtual servers via the communication network 10, and frequency information indicating the frequency of executing system calls on the physical servers 31 to 3m. The resource information indicating the resource amount of the physical servers 31 to 3m used when the system call is executed at the frequency is stored in association with each other (step S1). Subsequently, the management server 20 acquires first resource information indicating the amount of resources used in the physical server 31 when the virtual server executes an application process involving a system call on the physical server 31 (step S2). Then, the management server 20 selects resource information that is associated with the physical server 31 and that is closest to the first resource information acquired in step S2, and extracts frequency information associated with the selected resource information ( Step S3). Also, the management server 20 selects the frequency information that is most similar to the frequency information extracted in step S3 from the frequency information associated with the physical server 32 that is different from the physical server 31, and associates it with the selected frequency information. The obtained resource information is extracted (step S4). Furthermore, the management server 20 outputs the resource information extracted in step S4 as an estimated value of the amount of resources used in the physical server 32 when the virtual server executes application processing on the physical server 32 (step S5). ).

  Here, the effect of this embodiment will be described. The present embodiment is characterized by paying attention to a system call accompanying the application process in order to estimate the resource amount of the physical server used by the application process by the virtual server. When the virtual server uses hardware resources on the physical server, a system call is executed. Therefore, if the relationship between the type and frequency of system calls and the amount of resources used in the physical server at that time is covered in advance as much as possible, the application can be recorded with very high accuracy. It is possible to infer the type and frequency of the system call associated with. In addition, if the above relationships are stored in advance for each physical server to be measured, similar application processing is executed on other physical servers based on the estimated type and frequency of the system call. Sometimes the load on resources can be estimated with very high accuracy. Note that, when the method of the present embodiment is used, the accuracy of estimation is further improved by storing in advance more system call types and more various system call execution frequencies.

  In the present embodiment, the resource information (resource set) output as an estimated value of the resource amount used in the physical server 32 indicates a plurality of types of resource amounts. Therefore, as a result, the amount of information obtained is larger than the method in which only a single type of resource amount is required, and it can be used for general purposes.

[Second Embodiment]
FIG. 4 is a configuration diagram of an information processing system 200 according to the second embodiment of the present invention. A management server 40, physical servers 31 to 3m, and a management client 50 are connected to the communication network 10. The communication network 10 and the physical servers 31 to 3m of the present embodiment are equivalent to the communication network 10 and the physical servers 31 to 3m of the first embodiment. In addition, two types of business virtual servers and measurement virtual servers appear in this embodiment. The business virtual server can execute application processing on the physical server, and is equivalent to the virtual server of the first embodiment. The measurement virtual server can measure the amount of resources used on the physical server when the system call is executed in accordance with a command from the performance characteristic acquisition instruction unit 41 at a plurality of frequencies. . Then, the measurement virtual server can use the frequency of executing the system call as frequency information, can use the measured resource amount as resource information, and can send the frequency information and resource information to the performance information storage unit 46. Details of the operation of the measurement virtual server will be described later.

  The management server 40 includes a performance characteristic acquisition instruction unit 41, a virtual server operating unit 42, a performance information calculation unit 43, and a storage unit 44. The storage unit 44 includes a virtual server image storage unit 45, a performance information storage unit 46, a time-series performance information storage unit 47, and a load information storage unit 48. The virtual server image storage unit 45, the performance information storage unit 46, the time-series performance information storage unit 47, and the load information storage unit 48 are part of the storage area included in the storage unit 44. In addition to the information described below, the storage unit 44 may store information, programs, variables, and the like that are processed in the management server 40.

  In the virtual server image storage unit 45, an image of a measurement virtual server and an image of a business virtual server are registered. The virtual server operating unit 42 operates the measurement virtual server on each of the physical servers 31 to 3m. More specifically, the virtual server operating unit 42 copies the image of the measurement virtual server registered in the virtual server image storage unit 45 onto each of the physical servers 31 to 3m, and stores the copied measurement virtual server. Instruct to start. Note that the virtual server operating unit 42 may operate a plurality of measurement virtual servers on a single physical server.

  Then, the virtual server operating unit 42 operates the business virtual server on each of the physical servers 31 to 3m. More specifically, the virtual server operating unit 42 copies the business virtual server image registered in the virtual server image storage unit 45 onto the physical servers 31 to 3m, and starts the copied business virtual server. To instruct. The virtual server operating unit 42 may operate a plurality of business virtual servers on a single physical server. In addition, it is assumed that the business virtual server generates a load and consumes physical server resources by client processing or job scheduling processing.

  The physical servers 31 to 3m are at least one of the resource usage of the entire own device, the resource usage of the measurement virtual server running on the own device, and the resource usage of the business virtual server running on the own device Get on a regular basis.

  The management client 50 is a terminal operated by an operator. Although not illustrated in FIG. 4, the management client 50 may include a keyboard, a pointer, and the like that can be input by an operator. In addition, the management client 50 may have a function of displaying and printing out information to be processed in its own device so that the operator can recognize it, or printing it out.

  In the following description, the virtual server operating unit 42 has already operated at least one measurement virtual server from the virtual server image storage unit 45 to each of the physical servers 31 to 3m. The explanation will be made on the assumption that the system call can be executed in response to the instruction. The virtual server operating unit 42 has already operated the business virtual server from the virtual server image storage unit 45 to the physical server 31, and the business virtual server is after executing application processing on the physical server 31. It will be explained on the assumption. Note that when the measurement virtual server is on standby, a load that affects resources on the physical server is not generated.

  FIG. 5 is a flowchart showing the operation of the performance characteristic acquisition instruction unit 41 of the present embodiment. First, the performance characteristic acquisition instruction unit 41 receives a performance characteristic acquisition instruction from the management client 50 (step A1). The performance characteristic acquisition instruction unit 41 transmits a command in accordance with the performance characteristic instruction received in step A1 to each of the measurement virtual servers on the physical servers 31 to 3m (step A2). Here, the performance characteristic acquisition instruction specifically refers to which type of system call is to be executed by the measurement virtual server or how often the system call is to be executed by the measurement virtual server (sampling point). Number) etc. is information set by the operator.

  Further, the performance characteristic acquisition instructing unit 41 instructs the physical servers 31 to 3m to periodically output the resource usage acquired by the physical servers 31 to 3m to the measurement virtual server on the own machine. (Step A3). Then, the performance characteristic acquisition instructing unit 41 receives the measurement result from each measurement virtual server as a response to the command transmitted in step A2, and stores the received measurement result in the performance information storage unit 46 and the load information storage unit 48. Store (step A4). Then, the performance characteristic acquisition instruction unit 41 instructs the physical server that is the transmission source of the measurement result received in Step A4 to stop outputting the resource usage to the measurement virtual server on the own device (Step S4). A5). If all the measurement results of the physical servers 31 to 3m are stored in the performance information storage unit 46 (YES in step A6), the operation of the performance characteristic acquisition instruction unit 41 is finished. If some of the measurement results of the physical servers 31 to 3m are not yet stored in the performance information storage unit 46 (NO in step A6), the processes in step A4 and step A5 are repeated again.

  The resource usage output from the physical servers 31 to 3m to the measurement virtual server in accordance with the instruction in step A3 is stored in the measurement virtual server each time. The measurement result received in step A4 is the measurement result transmitted in step B13, which will be described later. The type of system call, the number of executions, and the resource usage stored in the performance information storage unit 46 in association with step B11. The average value of the quantity is stored. The load information storage unit 48 stores the system call type, the number of executions, and the resource usage stored in association with each other in step B4.

  FIG. 6 is a flowchart showing the operation of the measurement virtual server of this embodiment. The measurement virtual server receives and stores a command (command transmitted in step A2) in accordance with the performance characteristic instruction from the performance characteristic acquisition instruction unit 41 (step B1). Next, the measurement virtual server reads the contents of the performance characteristic instruction stored in step B1 (step B2). Then, the measurement virtual server keeps calling system calls of the type specified by the performance characteristic instruction one by one (step B3). Then, the measurement virtual server measures the number of executions of the system call that can be generated per unit time, and further, the latest value of the resource usage input from the physical server, the measured number of executions, and the type of the system call Are stored in association with each other (step B4). If all types of system calls designated by the performance characteristic instruction have not been called (NO in step B5), the processing in step B3 and step B4 is repeated again. The process from step B3 to step B5 is referred to as “prior measurement”.

  If all types of system calls of the type specified in the performance characteristic instruction have been called (YES in step B5), the measurement virtual server determines the number of executions of the system call stored in step B4 in the performance characteristic instruction. A value obtained by dividing by the designated sampling point number is stored as a reference execution number, and a value obtained by multiplying the reference execution number by an integer is obtained by the number represented by the sampling point number (step B6). Then, the measurement virtual server combines all the types of system calls designated by the performance characteristic instruction with all the values obtained by multiplying the reference execution number obtained in step B6 by an integer, and obtains all possible measurement patterns. Obtain (step B7).

If there are three types of system calls designated by the performance characteristic instruction, that is, system call A, system call B, and system call C, and the number of sampling points designated by the performance characteristic instruction is 5, it is obtained in step B7. The total number of measurement patterns is as follows. First, there are 15 measurement patterns (= combinations of system calls: 3 combinations x combinations of execution times: 5 patterns) for executing one type of system call. Also, two types of measurement pattern to execute a system call types 75 (= a combination of system calls: a combination of three types × execution count: 5 ² types) exist. Furthermore, three types of system calls the measurement pattern for the run ways 125 (= combination of system calls: 1 patterns × number of executions combination: 5 ^triplicate) exists. Therefore, under the above conditions, the maximum number of measurement patterns is 15 + 75 + 125 = 215.

  The measurement virtual server sequentially executes the measurement pattern obtained in step B7 and calls a system call (step B8). Then, the measurement virtual server checks whether any resource usage amount is the maximum value during execution of the measurement pattern. When it is the maximum value (YES in step B9), the process proceeds to step B8, and the measurement virtual server executes a new measurement pattern. If the maximum value is not reached (NO in step B9), the measurement virtual server continues to execute the measurement pattern currently being executed. The maximum value in step B9 may be the maximum value in the resource usage stored in step B4, or may be the maximum value set by the operator. However, if the maximum value is set by the operator, the physical server must be restricted so that a resource amount equal to or greater than the set maximum value cannot be used in unit time.

  In addition, the measurement virtual server calls the system call according to the new measurement pattern in step B8, and then confirms the number of times the resource usage is stored. The measurement virtual server is in a standby state until the number of confirmed times reaches a predetermined number (NO in Step B10), and repeats the check in Step B9. Then, when the number of confirmed times reaches the predetermined number (YES in step B10), the measurement virtual server obtains an average value of the latest resource usage and the previous resource usage (step B11). The predetermined number of times in step B10 may be a fixed value determined by default in the measurement virtual server, or may be a value that can be arbitrarily changed by the operator. Further, the number of resource usages that are averaged in step B11 is not limited.

  The measurement virtual server associates and stores the type of the system call executed in step B8, the number of executions, and the average value of the resource usage obtained in step B11 (step B12). If all the measurement patterns obtained in step B7 have not been executed (NO in step B13), the measurement virtual server repeatedly executes the processes in steps B8 to B12. The process from step B8 to step B12 is referred to as “combination measurement”. In addition, when the measurement result measured by “combination measurement” is executed in combination with a plurality of types of system calls at a plurality of frequencies as shown in FIG. 2 of the first embodiment, on the physical server. The amount of resources used can be obtained.

  If all the measurement patterns obtained in step B7 have been executed (YES in step B13), the measurement virtual server stores the type of system call, the number of executions, the resource usage, and the step stored in association with step B4. The type of system call, the number of executions, and the average value of the resource usage stored in association with B12 are transmitted as measurement results to the management server 40 (step B14).

  FIG. 7 is a flowchart showing the operation of the performance information calculation unit 43 of this embodiment. First, the performance information calculation unit 43 receives a load reproduction verification instruction from the management client 50 (step C1). The performance information calculation unit 43 issues an instruction to extract the resource usage from a plurality of pieces of information including a pair of the number of system call executions and the resource usage measured on the physical server 31 existing in the performance information storage unit 46. . Furthermore, the performance information calculation unit 43 is instructed to extract the resource usage of the business virtual server that was running on the physical server 31 stored in the load information storage unit 48. And the performance information calculation part 43 calculates the deviation degree with those resource usage. The performance information calculation unit 43 compares the degree of divergence calculated above and causes the time series performance information storage unit 47 to store the one having the smallest divergence degree. Under this condition, it is possible to extract the estimated value of the number of executions of the closest system call in order to output the resource usage of the business virtual server on the physical server 31 (step C2). The degree of divergence calculated in step C2 is calculated using the equation (1) described in the first embodiment. That is, the performance information calculation unit 43 performs a process equivalent to the first extraction unit 22 described in the first embodiment.

  The performance information calculation unit 43 extracts the number of executions and the number of executions in step C2 from a plurality of pieces of information consisting of system call execution times and resource usage pairs measured on the physical server 32 existing in the performance information storage unit 46. Calculate the divergence of the number of executions. The performance information calculation unit 43 compares the degree of divergence calculated above and causes the time series performance information storage unit 47 to store the one having the smallest divergence degree. Under this condition, the estimated value of the resource usage on the physical server 32 can be extracted (step C3). The divergence degree calculated in step C3 is calculated using the equation (2) described in the first embodiment. That is, the performance information calculation unit 43 performs processing equivalent to that of the second extraction unit 23 described in the first embodiment.

  The performance information calculation unit 43 transmits the data of the number of executions of each system call for each time series in the time series performance information storage unit 47 to the measurement virtual server on the physical server 32. That is, the performance information calculation unit 43 performs a process equivalent to that of the output unit 24 described in the first embodiment. The measurement virtual server executes a system call at the frequency indicated by the received frequency information, measures the resource amount used on the physical server 32 when the system call is executed, and measures the measured resource amount (first 2 resource information) is returned to the performance information calculation unit 43 (step C4).

  The performance information calculation unit 43 calculates the divergence between the estimated value of the resource usage of the business virtual server on the physical server 32 extracted in Step C3 and the resource usage returned in Step C4 (Step C5). ). The degree of divergence calculated in step C5 is calculated using the equation (1) described in the first embodiment. If the calculation result of step C5 is transmitted to the management client 50 and the management client 50 notifies the operator by displaying the calculation result, etc., the operator confirms the consistency of whether these estimates are correct. can do.

  As a result of step C5, the consistency of the estimated value of the resource usage on the physical server 32 extracted in step C3 is confirmed, and the estimated value of the resource usage is the maximum value of each resource usage of the physical server 32. If not reached (NO in step C6), the virtual server operating unit 42 operates the business virtual server existing on the physical server 31 on the physical server 32 (step C7). If any of the resource usage amounts has reached the maximum value (YES in step C6), the process ends at this point.

  The performance information calculation unit 43 uses the estimated value of the resource usage of the business virtual server on the physical server 32 extracted in step C3, and when the business virtual server actually executes the application process on the physical server 32. The degree of deviation from the resource usage is calculated (step C8). The degree of divergence calculated in step C8 is calculated using the equation (1) described in the first embodiment. If the calculation result of step C8 is sent to the management client 50 and the management client 50 notifies the operator by displaying the calculation result, etc., the operator confirms the consistency of whether these estimates are correct. can do.

  Here, the effect of this embodiment will be described. In the present embodiment, information necessary for estimating resource usage (information stored in the performance information storage unit 43) can be collected using the measurement virtual servers operated on the physical servers 31 to 3m. it can.

  Also, in this embodiment, the business virtual server is operated when the estimated value of the obtained resource usage does not reach the maximum value, so the business virtual server reliably executes application processing on the physical server be able to. Furthermore, in this embodiment, since the consistency of the estimated value of the resource usage obtained is verified, a physical server that does not reach the maximum value of the resource usage is selected, and the business virtual server is operated on the physical server. Can be made. Therefore, the business virtual server can more reliably execute application processing on the physical server.

[Third Embodiment]
The third embodiment has a configuration equivalent to the information processing system 200 of the second embodiment. However, it is assumed that the first measurement server and the second measurement server are operating on the physical server 31. The operations of the management server 40 and the measurement management server on the physical servers 31 to 3m are the same as the operations described with reference to FIGS. 5 and 6, and thus detailed description thereof is omitted.

  FIG. 8 is a flowchart showing the operation of the performance information calculation unit 43 of this embodiment. The performance information calculation unit 43 receives the free resource verification instruction of the physical server 31 stored in the load information storage unit 48 from the management client 50 (step E1). Here, it is assumed that the resource usage of the business virtual server operating on the physical server 31 is stored in the load information storage unit 48.

  The performance information calculation unit 43 issues an instruction to extract the resource usage from a plurality of pieces of information including a pair of the number of system call executions and the resource usage measured on the physical server 31 existing in the performance information storage unit 46. . Furthermore, the performance information calculation unit 43 subtracts the resource usage of the business virtual server that was running on the physical server 31 stored in the load information storage unit 48 from the maximum resource usage of the physical server 31 for each resource. Instructs to extract resource usage. And the performance information calculation part 43 calculates the deviation degree with those resource usage. The performance information calculation unit 43 compares the degree of divergence calculated above and causes the time series performance information storage unit 47 to store the one having the smallest degree of divergence. Under this condition, an estimated value of the nearest execution count is extracted in order to output the resource usage for the empty capacity on the physical server 31 (step E2). Note that the degree of divergence calculated in step E2 is calculated using the equation (1) described in the first embodiment.

  The performance information calculation unit 43 transmits estimated value data of the number of system call executions of the business virtual server for each time series in the time series performance information storage unit 47 to the first measurement virtual server on the physical server 31. To do. Further, the performance information calculation unit 43 estimates the number of execution times of system calls for the empty capacity for each time series in the time series performance information storage unit 47 for the second measurement virtual server on the physical server 31. Is transmitted (step E3). Based on the estimated value data of the number of executions of the system call transmitted in step E3, the first measurement virtual server and the second measurement virtual server generate a load and measure the resource usage. The measured resource usage is sent to the management server 40 and stored in the load information storage unit 48.

  The performance information calculation unit 43 uses the actual resource usage amount of the physical server 31 stored in the time-series performance information storage unit 47 and the resource usage amount of the second measurement virtual server stored in the load information storage unit 48. The degree of deviation from the calculated estimated value of free resource usage is calculated (step E4). The degree of divergence calculated in step E4 is calculated using the equation (1) described in the first embodiment. If the calculation result of step E4 is transmitted to the management client 50 and the management client 50 notifies the operator by displaying and outputting the calculation result, the operator confirms whether or not these estimates are correct. can do.

  Here, the effect of this embodiment will be described. By executing “preliminary measurement” and “combination measurement”, the correspondence between the resource usage and the number of executions of the system call is found. This process is performed so as to cover the pattern of the generated load by executing the system call on the measurement program assuming the execution process of the system call in the application. As a result of this processing, there is a pattern in the measured data that shows resource usage that is almost equal to the available free resources other than the business virtual server, so the corresponding one is extracted and converted into the number of executions . Free resources can be compared on each server by the number of executions converted. From this, it is possible to select a pattern closer to the load generated by the application by estimating the load including overhead between system calls at various layers, which was not considered in the conventional method. Load accuracy is improved. Furthermore, by including a verification process, it is possible to confirm the consistency of whether the load estimate is correct.

  As mentioned above, although embodiment of this invention was described with reference to drawings, these are the illustrations of this invention, Various structures other than the above are also employable.

  For example, although the above embodiment has been described on the assumption that there are a plurality of types of resource information, the present invention may be implemented using only one type of resource information.

  In the third embodiment, the configuration of the information processing system has been described using a configuration in which a plurality of physical servers, a management server, and a management client are connected via a communication network. However, the configuration is not limited to this. For example, a configuration in which a single physical server and a management server are connected via a communication network may be used, or a single device in which functions equivalent to those of the management server are incorporated in a single physical server. It doesn't matter.

  In the first embodiment, the first extraction unit 22 or the second extraction unit 23 has been described so as to select the resource information or the frequency information with the smallest calculated divergence, but the compared values match. Depending on whether or not, resource information or frequency information to be selected may be determined. In addition, the first extraction unit 22 or the second extraction unit 23 may extract resource information or frequency information that minimizes the difference between each calculated degree of deviation (index value) and a predetermined determination value. At this time, the predetermined determination value may be 0 or another numerical value. Further, the first extraction unit 22 and the second extraction unit 23 may use different determination values.

  In the description of the above embodiment, a plurality of flowcharts are used, and a plurality of steps are described in order. The order of description is the order in which the steps of the information processing method of the present invention are executed. It is not intended to limit. For this reason, when the information processing method of the present invention is executed, the order of the plurality of steps can be changed within a range that does not hinder the contents.

  Needless to say, the above-described embodiment and a plurality of modifications can be combined within a range in which the contents do not conflict with each other. Moreover, although the functions and the like of each component have been specifically described in the above-described embodiments and modifications, the functions and the like can be changed in various ways within a range that satisfies the present invention.

(Appendix 1) A plurality of physical servers connected via a communication network, each of the physical servers, frequency information indicating the frequency of executing a system call on the physical server, and executing the system call at the frequency The resource information indicating the resource amount of the physical server used at the time of performance is stored in association with the performance information storage means, and the resource amount used when the first physical server executes the application process involving the system call. A degree of approximation of each of the resource information and the first resource information stored in the performance information storage means and associated with the first physical server; The first extraction that extracts the first frequency information associated with the resource information that minimizes the difference between each index value indicating the first determination value Each index value indicating a degree of approximation between each frequency information and the first frequency information stored in the performance information storage unit and associated with a second physical server different from the first physical server, and a second index value A second extraction unit that extracts resource information associated with frequency information that minimizes a difference from a determination value; and an output unit that outputs the resource information extracted by the second extraction unit. Information processing system.

(Supplementary note 2) Frequency information indicating the frequency of executing system calls on a physical server and resource information indicating the amount of resources of the physical server used when the system calls are executed at the frequency are stored in association with each other. Performance information storage means, acquisition means for obtaining first resource information indicating a resource amount used in the physical server when the virtual server executes the application process involving the system call on the physical server, A difference between the maximum value of the resource amount represented by the resource information associated with the first physical server and the resource amount represented by the first resource information acquired by the acquisition unit is calculated and associated with the first physical server The resource representing the resource amount having the smallest deviation from the calculated difference among the obtained resource information Information is selected from the performance information storage means, and the third extraction means for extracting the frequency information associated with the selected resource information, and when the virtual server executes the application process on the physical server Output means for outputting the frequency information extracted by the third extraction means as an estimated value of the frequency of the system call that can be executed using the amount of free resources on the physical server, Information processing system.

(Supplementary Note 3) The performance information storage unit stores a resource set, which is a set of a plurality of types of resource information, in association with the frequency information, and the acquisition unit is a set of a plurality of types of the first resource information. The first resource set is acquired, and the third extracting means extracts and extracts the maximum value of the resource amount represented by the resource information included in the resource set associated with the first physical server for each type. The difference between the maximum value of the resource amount and the first resource information included in the first resource set is calculated for each type, is associated with the first physical server, and is the same type of resource as the first resource set The degree of divergence between each of the resource sets that are information and each of the calculated differences is calculated, and the resource set that minimizes the degree of divergence is selected and selected. Information processing system and extracting the frequency information associated with the resource set was.

(Supplementary Note 4) The virtual server can execute a plurality of types of the system calls on each of the physical servers, and the performance information storage unit includes a type set that is a set of a plurality of types of type information, A frequency set that is a set of frequency information corresponding to each of the type information included in the type set, the resource information, and each of the physical servers are stored in association with each other, and the third extracting unit stores the performance information The type set and the frequency set associated with the resource information selected from the information storage unit are extracted, and the output unit outputs the type set and the frequency set extracted by the third extraction unit. An information processing system characterized by

DESCRIPTION OF SYMBOLS 100 Information processing system 200 Information processing system 10 Communication network 20 Management server 21 Acquisition part 22 Extraction part 23 Extraction part 24 Output part 25 Storage part 26 Performance information storage part 31-3m Physical server 40 Management server 41 Performance characteristic acquisition instruction part 42 Virtual Server operation unit 43 Performance information calculation unit 44 Storage unit 45 Virtual server image storage unit 46 Performance information storage unit 47 Time-series performance information storage unit 48 Load information storage unit 50 Management client

Claims (10)

A plurality of physical servers capable of sending and receiving virtual servers via a communication network;
Each of the physical servers, frequency information representing the frequency of executing the system call on the physical server, resource information representing the resource amount of the physical server used when the system call is executed at the frequency, and Performance information storage means for storing
Obtaining means for obtaining first resource information indicating a resource amount used in the first physical server when the virtual server executes an application process involving the system call on the first physical server;
The resource information associated with the first physical server and closest to the first resource information obtained by the obtaining unit is selected from the performance information storage unit, and the resource information associated with the selected resource information is selected. First extraction means for extracting frequency information;
Of the frequency information associated with a second physical server different from the first physical server, the frequency information closest to the frequency information extracted by the first extraction unit is selected from the performance information storage unit. Second extraction means for extracting the resource information associated with the selected frequency information;
Output means for outputting the resource information extracted by the second extraction means;
An information processing system comprising: The information processing system according to claim 1,
The first extraction unit stores the resource information having a minimum deviation from the first resource information acquired by the acquisition unit among the resource information associated with the first physical server as the performance information storage. Selecting from the means, extracting the frequency information associated with the selected resource information;
The second extraction unit stores the frequency information having a minimum deviation from the frequency information extracted by the first extraction unit among the frequency information associated with the second physical server in the performance information storage. An information processing system characterized by selecting from the means and extracting the resource information associated with the selected frequency information. The information processing system according to claim 1 or 2,
The virtual server can execute a plurality of types of the system calls on each physical server,
The performance information storage means includes a type set that is a set of type information indicating the type of each system call, and a frequency set that is a set of frequency information corresponding to each of the type information included in the type set, Storing the resource information in association with each physical server;
The first extraction unit extracts the type set and the frequency set associated with the resource information selected from the performance information storage unit,
The second extraction unit selects at least one set of the type set that matches the type set extracted by the first extraction unit and is associated with the second physical server, and sets the selected type set to the selected type set. Each of the associated frequency sets is compared with the frequency set extracted by the first extraction unit, and the degree of divergence is calculated, and the frequency set that minimizes the degree of divergence is selected and selected. An information processing system, wherein the resource information associated with the frequency set is extracted. An information processing system according to any one of claims 1 to 3,
The performance information storage means stores a resource set that is a set of a plurality of types of the resource information in association with the frequency information,
The acquisition means acquires a first resource set that is a set of a plurality of types of the first resource information,
Each of the resource sets that is associated with the first physical server and that is the resource information of the same type as the first resource set acquired by the acquiring unit, and the first resource set, , Calculating the degree of divergence from each other, selecting the resource set that minimizes the degree of divergence, and extracting the frequency information associated with the selected resource set . An information processing system according to claim 1,
Virtual server operation means for operating a measurement virtual server on each of the physical servers,
The measurement virtual server operated on the physical server by the virtual server operating unit executes the system call at a plurality of frequencies, and resources used on the physical server when the system call is executed Measuring the amount, the frequency of executing the system call as the frequency information, the measured resource amount as the resource information, further sending the frequency information and the resource information to the performance information storage means,
The information processing system, wherein the performance information storage means stores the frequency information and the resource information that have arrived from each of the measurement virtual servers. The information processing system according to claim 5,
The output means outputs the resource information extracted by the second extraction means and the frequency information associated with the resource information to the measurement virtual server operated on the second physical server. ,
The measurement virtual server executes the system call at a frequency indicated by the frequency information received from the output unit, and measures the amount of resources used on the second physical server when the system call is executed. The measured resource amount is sent to the first extraction means as second resource information,
The information processing system, wherein the first extraction unit calculates a divergence degree between the second resource information that has arrived from the measurement virtual server and the resource information received from the output unit. The information processing system according to claim 5 or 6,
When the resource amount represented by the resource information output from the output unit is smaller than the maximum value of the resource amount represented by the resource information associated with the second physical server, the virtual server operating unit 2. An information processing system in which the virtual server is operated on two physical servers. An information processing apparatus capable of communicating with each of a plurality of physical servers that can exchange a virtual server via a communication network,
Each of the physical servers, frequency information representing the frequency of executing the system call on the physical server, resource information representing the resource amount of the physical server used when the system call is executed at the frequency, and Performance information storage means for storing
Obtaining means for obtaining first resource information indicating a resource amount used in the first physical server when the virtual server executes an application process involving the system call on the first physical server;
The resource information associated with the first physical server and closest to the first resource information obtained by the obtaining unit is selected from the performance information storage unit, and the resource information associated with the selected resource information is selected. First extraction means for extracting frequency information;
Of the frequency information associated with a second physical server different from the first physical server, the frequency information closest to the frequency information extracted by the first extraction unit is selected from the performance information storage unit. Second extraction means for extracting the resource information associated with the selected frequency information;
Output means for outputting the resource information extracted by the second extraction means;
An information processing apparatus comprising: Each of a plurality of physical servers that can send and receive a virtual server via a communication network, frequency information indicating the frequency of executing a system call on the physical server, and used when the system call is executed at the frequency A resource information indicating the resource amount of the physical server that has been performed, and a performance information storage step for storing the information in association with each other,
An acquisition step of acquiring first resource information indicating a resource amount used in the first physical server when the virtual server executes an application process involving the system call on the first physical server;
The resource information associated with the first physical server and closest to the first resource information obtained in the obtaining step is selected, and the frequency information associated with the selected resource information is extracted. One extraction step;
The frequency information closest to the frequency information extracted in the first extraction step is selected from the frequency information associated with a second physical server different from the first physical server, and the selected frequency A second extraction step for extracting the resource information associated with the information;
An output step of outputting the resource information extracted in the second extraction step;
An information processing method comprising: A program stored in a computer-readable storage medium,
Each of a plurality of physical servers that can send and receive a virtual server via a communication network, frequency information indicating the frequency of executing a system call on the physical server, and used when the system call is executed at the frequency Resource information representing the resource amount of the physical server that has been performed, and a performance information storage process for storing the information in association with each other,
An acquisition process for acquiring first resource information indicating a resource amount used in the first physical server when the virtual server executes an application process involving the system call on the first physical server;
The resource information associated with the first physical server and closest to the first resource information acquired in the acquisition process is selected, and the frequency information associated with the selected resource information is extracted. 1 extraction process,
The frequency information closest to the frequency information extracted in the first extraction process is selected from the frequency information associated with a second physical server different from the first physical server, and the selected frequency A second extraction process for extracting the resource information associated with the information;
An output process for outputting the resource information extracted in the second extraction process;
That causes the computer to execute the program.

Images