JP6940761B2 - Information processing equipment, virtual machine monitoring programs, and information processing systems - Google Patents

Description

The present invention relates to an information processing device, a virtual machine monitoring program, and an information processing system.

In a cloud computing system, for example, a service is provided using a virtual machine built on a server. When starting to provide a new service in a cloud computing system, a virtual machine with performance corresponding to the load of the service is built on the server. The performance of a virtual machine depends on the amount of hardware resources available to run the virtual machine (hereinafter, also referred to simply as resources, refers to hardware resources). Setting the resources available to run a virtual machine is sometimes referred to as allocating resources to the virtual machine. The resources allocated to the virtual machine are a CPU (Central Processing Unit), memory, a storage device, and the like. The greater the amount of these resources, the better the performance of the virtual machine.

Here, if the amount of resources allocated to the virtual machine is too small, it is not possible to provide the service with the quality required for the service provided by the virtual machine. On the other hand, if too many resources are allocated to the virtual machine, the resources in the system will be wasted.

Therefore, for example, in a cloud system, a computer resource allocation method for suppressing a decrease in performance and providing an appropriate computer resource has been considered. In this computer resource allocation method, the management computer acquires the performance of the virtual computer group that provides the service, and compares the acquired performance of the virtual computer group with the performance condition of the preset service. Then, the management computer determines the computer resource to be changed in the virtual computer group according to the comparison result.

Japanese Unexamined Patent Publication No. 2016-103179

However, if the virtual machine's resource amount is changed after confirming that the virtual machine's performance has deteriorated, there will be a delay between the deterioration of the virtual machine's performance and the change in the allocated resource amount. As a result, the quality of service deteriorates for a certain period until the process of changing the amount of resources is completed.

On the one hand, the purpose of this case is to suppress the delay in optimizing the amount of resources allocated to the virtual machine.

In one proposal, an information processing apparatus having the following processing units is provided.
The processing unit monitors the amount of requests sent to the virtual machine running by the server per unit period. The processing unit then decides whether to change the amount of hardware resources available to run the virtual machine, based on how the amount of requests changes over time. If it decides to change it, the processor tells the server to change the amount of hardware resources available to run the virtual machine.

According to one aspect, the delay in optimizing the amount of resources allocated to the virtual machine is suppressed.

It is a figure which shows an example of the configuration of the virtual machine monitoring system which concerns on 1st Embodiment. It is a figure which shows the system configuration example of the 2nd Embodiment. It is a figure which shows one configuration example of the hardware of the virtual machine monitoring apparatus used in the 2nd Embodiment. It is a block diagram which shows the function which each device has for changing the amount of resources allocated to a virtual machine. It is a figure which shows an example of the request monitoring table. It is a figure which shows an example of a resource management table. It is a figure which shows an example of a virtual machine management table. It is a figure which shows an example of a resource priority table. It is a figure which shows an example of the request management table which a load balancer stores. It is a sequence diagram which shows an example of the procedure of request monitoring processing. It is a sequence diagram which shows an example of the procedure of an optimization process. This is the first half of the flowchart showing an example of the procedure for request monitoring processing in the virtual machine monitoring device. This is the latter half of the flowchart showing an example of the procedure for request monitoring processing in the virtual machine monitoring device. It is a flowchart which shows an example of the procedure of the request amount / response time measurement processing. This is the first half of the flowchart showing an example of the cost improvement processing procedure. This is the latter half of the flowchart showing an example of the cost improvement processing procedure. This is the first half of the flowchart showing an example of the procedure for response improvement processing. This is the latter half of the flowchart showing an example of the procedure for response improvement processing.

Hereinafter, the present embodiment will be described with reference to the drawings. It should be noted that each embodiment can be implemented by combining a plurality of embodiments within a consistent range.
[First Embodiment]
First, the first embodiment will be described.

FIG. 1 is a diagram showing an example of a configuration of a virtual machine monitoring system according to the first embodiment. A plurality of terminal devices 1a and 1b are connected to the server 3 via the network 2. The server 3 has CPUs 3a-1, 3a-2, ..., Memories 3b-1, 3b-2, ... As hardware resources. The server 3 executes the virtual machine 3c using at least a part of the hardware resources. The server 3 can automatically change the amount of hardware resources that can be used to execute the virtual machine 3c. The virtual machine 3c realized by the server 3 executes processing in response to a request from the terminal devices 1a and 1b, and transmits the processing result as a response to the terminal devices 1a and 1b.

The information processing device 10 monitors the requests transmitted to the virtual machine 3c executed by the server 3, and when the change in the request amount is excessive, the hardware resource that can be used to execute the virtual machine 3c. Instruct the server 3 to change the amount of. Therefore, the information processing device 10 has a storage unit 11 and a processing unit 12. The storage unit 11 is, for example, a memory or a storage device included in the information processing device 10. The processing unit 12 is, for example, a processor or an arithmetic circuit included in the information processing device 10.

The storage unit 11 stores the amount of requests per unit period. For example, the storage unit 11 stores the request amount within the latest unit period (first unit period) and the request amount within the immediately preceding unit period (second unit period).

The processing unit 12 determines whether or not to change the amount of resources that can be used to execute the virtual machine 3c based on the request amount per unit period obtained from the monitoring result, and if so, changes the amount. Is transmitted to the server 3. Therefore, the processing unit 12 monitors the amount of requests per unit period transmitted to the virtual machine 3c executed by the server 3. For example, the processing unit 12 causes a device that relays requests transmitted from the terminal devices 1a and 1b to the server 3 in the network 2 to count the request amount per unit period, and the processing unit 12 counts the request amount per unit period from the device. Get information indicating the amount of requests. The device that relays the request is, for example, a load balancer. The processing unit 12 can also capture packets communicated via the network 2, analyze the captured packets, and measure the amount of requests transmitted to the virtual machine 3c. The processing unit 12 stores the amount of requests transmitted to the virtual machine 3c per unit period in the storage unit 11.

Whether or not the processing unit 12 changes the amount of hardware resources available for executing the virtual machine based on the degree of change in the amount of requests per unit period sent to the virtual machine 3c over time. decide. The degree of change in the amount of requests per unit period over time is expressed, for example, by the volatility. The volatility is a value obtained by dividing the request amount within the first unit period by the request amount within the second unit period.

For example, when the processing unit 12 compares the amount of requests in the most recent first unit period with the amount of requests in the second unit period immediately before the first unit period, the degree of change exceeds the threshold value. Determine if it is. When the degree of change exceeds the threshold value, the processing unit 12 determines that the amount of hardware resources available for executing the virtual machine 3c is changed.

A plurality of threshold values can be used as the threshold values used for determining whether or not to change the amount of hardware resources. The first threshold is a positive real number less than 1 (eg "0.5") and the second threshold is a real number greater than 1 (eg "1.5"). When the amount of requests in the first unit period is smaller than the amount of requests in the second unit period and the degree of decrease exceeds the first threshold value, the processing unit 12 is a virtual machine. Decide to reduce the amount of hardware resources available to run 3c. When the degree of decrease exceeds the first threshold value, for example, the volatility is less than the first threshold value. Further, when the amount of requests in the first unit period is larger than the amount of requests in the second unit period and the degree of increase exceeds the second threshold value, the processing unit 12 is virtual. Decide to increase the amount of hardware resources available to run machine 3c. When the degree of increase exceeds the second threshold value, for example, the volatility is larger than the second threshold value.

When the processing unit 12 decides to change the amount of hardware resources available for executing the virtual machine 3c, the processing unit 12 instructs the server 3 to change the amount of hardware resources available for executing the virtual machine 3c. For example, when the processing unit 12 decides to reduce the amount of hardware resources, the processing unit 12 instructs the server 3 to reduce the amount of hardware resources available for executing the virtual machine 3c. When the processing unit 12 decides to increase the amount of hardware resources, the processing unit 12 instructs the server 3 to increase the amount of hardware resources that can be used to execute the virtual machine 3c.

According to such a virtual machine monitoring system, when the amount of requests per unit period for the virtual machine 3c tends to decrease or increase more than a certain level, the hardware resources available for executing the virtual machine 3c are reduced or decreased. Can be increased. That is, the amount of hardware resources available to execute the virtual machine 3c can be optimized based on the amount of requests. As a result, the delay in optimizing the amount of resources allocated to the virtual machine 3c is suppressed as compared with the case where the amount of hardware resources is changed after confirming that the performance of the virtual machine 3c has deteriorated.

For example, the performance of a virtual machine can be measured by the response time to a request (the time from sending a request to receiving a response). Due to the nature of computers such as virtual machines, services can be provided with a sufficiently short response time up to a predetermined load, but when the load exceeds a predetermined amount, the response time may deteriorate sharply. In such a case, if the virtual machine resource amount is increased after confirming the deterioration of the response time, the response time remains deteriorated until the process of increasing the virtual machine resource amount is completed, and the service It causes deterioration of quality. Moreover, if the search for the optimum hardware resource configuration that meets the response time requirements is performed after the response time has deteriorated, the search takes time and the service quality deterioration period becomes long. The requirement of the response time is, for example, that the response time is less than a predetermined threshold value.

On the other hand, in the virtual machine monitoring system shown in FIG. 1, hardware resources that can be used for executing the virtual machine 3c are automatically added when an increasing tendency of the request amount per unit period can be detected. That is, even if the request amount continues to increase, the amount of hardware resources that can be used to execute the virtual machine 3c also increases accordingly, and it is possible to prevent the response time from deteriorating.

Moreover, in the virtual machine monitoring system shown in FIG. 1, when the request amount per unit period starts to decrease, the amount of hardware resources that can be used to execute the virtual machine 3c is automatically reduced. As a result, it is suppressed that a large amount of hardware resources are wasted for executing the virtual machine 3c. As a result, the hardware resources of the server 3 can be used efficiently.

The processing unit 12 may determine the priority of changing the amount that can be used for executing the virtual machine 3c for each type of hardware resource (CPU, memory, etc.). For example, the processing unit 12 acquires the usage rate of the virtual machine 3c for each type of hardware resource from the server 3, and the higher the usage rate of the hardware resource, the higher the priority. At this time, the information processing apparatus 10 defines the maximum value and the minimum value that can be used for executing the virtual machine 3c for each type of hardware resource. For example, the information processing device 10 stores resource management information in advance in the storage unit 11. The resource management information includes the priority of changing the amount that can be used to execute the virtual machine 3c, the minimum value of the amount that can be used to execute the virtual machine 3c, and the virtual machine for each type of hardware resource. The maximum amount available for execution is shown.

When the processing unit 12 decides to increase the amount of hardware resources that can be used to execute the virtual machine, the processing unit 12 has the highest priority type of hardware within the range that does not exceed the maximum value based on the resource management information. Decide to increase resources. When the amount of the hardware resource of the highest priority type has already reached the maximum value, the processing unit 12 increases the hardware resource of the next highest priority type within a range not exceeding the maximum value. To determine.

When the processing unit 12 decides to reduce the amount of hardware resources that can be used to execute the virtual machine, the processing unit 12 has the highest priority type of hardware within the range that does not fall below the minimum value based on the resource management information. Decide to reduce wear resources. When the amount of the hardware resource of the highest priority type has already reached the minimum value, the processing unit 12 reduces the hardware resource of the next highest priority type within a range not lower than the minimum value. To determine.

In this way, by determining the priority for changing the amount of resources based on the usage rate for each type of hardware resource, the hardware that is likely to cause deterioration of the response time of the virtual machine 3c. Resources can be increased preferentially. As a result, it is possible to efficiently change the amount of hardware resources that can be used to execute the virtual machine 3c.

That is, if the amount of hardware resources is changed without using the usage rate for each type of hardware resources, there is a possibility that the types of hardware resources that do not cause deterioration of the response time of the virtual machine 3c will be increased. There is. In this case, although the hardware resources available for executing the virtual machine 3c have been increased, it is not possible to suppress the deterioration of the response time as the amount of requests increases. On the other hand, by determining the priority for changing the amount of resources based on the usage rate of each type of hardware resource, priority is given to the amount of hardware resources of the type that is likely to cause deterioration of response time. Can be increased.

In addition to the virtual machine 3c, launch a verification virtual machine that can use the same amount of hardware resources as the hardware resources that can be used to execute the virtual machine 3c, and the verification virtual machine is appropriate. You can also search for the amount of hardware resources. However, if a search is performed using a virtual machine for verification without considering the usage rate for each type of hardware resource, a virtual machine that satisfies the requirements from among a large number of virtual machines with different hardware resource configurations. Will be specified. For example, even when the CPU usage rate of a virtual machine is high and a lack of CPU may cause deterioration of response time, a virtual machine for verification with an expanded amount of memory resources is also searched. As a result, it takes time to identify the optimum hardware resource configuration. On the other hand, when the type of hardware resource that causes the deterioration of response time is changed by determining the priority for changing the amount of hardware resource based on the usage rate for each type of hardware resource. You can search with priority. This makes it possible to identify the hardware resource configuration that meets the response time requirements at an early stage. That is, it is possible to efficiently search for the optimum hardware resource configuration that can be used to execute the virtual machine.

[Second Embodiment]
Next, the second embodiment will be described. In the second embodiment, the amount of hardware resources that can be used to execute the virtual machine is changed according to the amount of requests for the virtual machine in operation. In the second embodiment, the verification virtual machine is used to determine whether the response time requirement is met before changing the amount of hardware resources available to execute the operating virtual machine. Verify.

In the following description, hardware resources are simply referred to as resources. Setting the resources that can be used to execute a virtual machine is called allocating resources to the virtual machine.

FIG. 2 is a diagram showing a system configuration example of the second embodiment. The cloud computing system includes a virtual machine monitoring device 100, a load balancer 200, and a server 300. The virtual machine monitoring device 100, the load balancer 200, and the server 300 are connected by the management network 20. The management network 20 is a network for operation management of a cloud computing system. Further, the load balancer 200 and the server 300 are connected by the business network 41. The business network 41 is a network used by the load balancer 200 to transfer a request received from the terminal devices 31, 32, ... To the server 300, and to transmit a response from the server 300 to the request.

The virtual machine monitoring device 100 manages the amount of resources allocated to the virtual machine based on the amount of requests sent to the virtual machine in the server 300. For example, when the increase rate of the request amount per unit period transmitted to the virtual machine exceeds the threshold value of the increase rate, the virtual machine monitoring device 100 instructs the server 300 to add resources to the virtual machine. Further, when the reduction rate of the request amount per unit period transmitted to the virtual machine exceeds the threshold value of the reduction rate, the virtual machine monitoring device 100 instructs the server 300 to reduce the resources of the virtual machine.

The load balancer 200 is connected to a plurality of terminal devices 31, 32, ... Via a network 42. Each of the plurality of terminal devices 31, 32, ... Is a computer used by a user who uses the service. The load balancer 200 transfers a request sent from any of the plurality of terminal devices 31, 32, ... To any of the plurality of virtual machines in the server 300. Further, when the load balancer 200 receives the response to the request from any of the plurality of virtual machines in the server 300, the load balancer 200 transfers the response to the terminal device of the transmission source of the corresponding request. The load balancer 200 further measures the amount of requests transferred to the virtual machine per unit period and the response time to the requests based on the request from the virtual machine monitoring device 100. The load balancer 200 transmits the measurement result to the virtual machine monitoring device 100.

The server 300 is a computer that provides services. The server 300 internally creates a plurality of virtual machines. At least a part of the resources (CPU, memory, storage device, etc.) of the server 300 is allocated to each of the plurality of virtual machines. Each of the plurality of virtual machines in the server 300 uses the allocated resources to provide services to the terminal devices 31, 32, ....

FIG. 3 is a diagram showing a configuration example of the hardware of the virtual machine monitoring device used in the second embodiment. The entire device of the virtual machine monitoring device 100 is controlled by the processor 101. A memory 102 and a plurality of peripheral devices are connected to the processor 101 via a bus 109. The processor 101 may be a multiprocessor. The processor 101 is, for example, a CPU, an MPU (Micro Processing Unit), or a DSP (Digital Signal Processor). At least a part of the functions realized by the processor 101 executing a program may be realized by an electronic circuit such as an ASIC (Application Specific Integrated Circuit) or a PLD (Programmable Logic Device).

The memory 102 is used as the main storage device of the virtual machine monitoring device 100. At least a part of an OS (Operating System) program or an application program to be executed by the processor 101 is temporarily stored in the memory 102. Further, various data used for processing by the processor 101 are stored in the memory 102. As the memory 102, for example, a volatile semiconductor storage device such as a RAM (Random Access Memory) is used.

Peripheral devices connected to the bus 109 include a storage device 103, a graphic processing device 104, an input interface 105, an optical drive device 106, a device connection interface 107, and a network interface 108.

The storage device 103 electrically or magnetically writes and reads data from the built-in recording medium. The storage device 103 is used as an auxiliary storage device for a computer. The storage device 103 stores an OS program, an application program, and various data. As the storage device 103, for example, an HDD (Hard Disk Drive) or an SSD (Solid State Drive) can be used.

A monitor 21 is connected to the graphic processing device 104. The graphic processing device 104 causes the image to be displayed on the screen of the monitor 21 in accordance with the instruction from the processor 101. The monitor 21 includes a display device using a CRT (Cathode Ray Tube), a liquid crystal display device, and the like.

A keyboard 22 and a mouse 23 are connected to the input interface 105. The input interface 105 transmits a signal sent from the keyboard 22 and the mouse 23 to the processor 101. The mouse 23 is an example of a pointing device, and other pointing devices can also be used. Other pointing devices include touch panels, tablets, touchpads, trackballs and the like.

The optical drive device 106 reads the data recorded on the optical disk 24 by using a laser beam or the like. The optical disk 24 is a portable recording medium on which data is recorded so that it can be read by reflection of light. The optical disk 24 includes a CD-ROM (Compact Disc Read Only Memory), a CD-R (Recordable) / RW (ReWritable), and the like.

The device connection interface 107 is a communication interface for connecting peripheral devices to the virtual machine monitoring device 100. For example, a memory device 25 or a memory reader / writer 26 can be connected to the device connection interface 107. The memory device 25 is a recording medium equipped with a communication function with the device connection interface 107. The memory reader / writer 26 is a device that writes data to or reads data from the memory card 27. The memory card 27 is a card-type recording medium.

The network interface 108 is connected to the management network 20. The network interface 108 transmits / receives data to / from another computer or communication device via the management network 20.

With the hardware configuration as described above, the processing function of the virtual machine monitoring device 100 of the second embodiment can be realized. Data can be input / output to and from the virtual machine monitoring device 100 via the device connection interface 107 or the network interface 108. Therefore, it is not necessary to connect an input / output device such as a monitor 21, a keyboard 22, and a mouse 23 to the virtual machine monitoring device 100. Further, the virtual machine monitoring device 100 does not have to have the optical drive device 106.

The load balancer 200 and the server 300 can also be realized by the same hardware as the virtual machine monitoring device 100 shown in FIG. Further, the information processing device 10 shown in the first embodiment can also be realized by the same hardware as the virtual machine monitoring device 100 shown in FIG.

The virtual machine monitoring device 100 realizes the processing function of the second embodiment, for example, by executing a program recorded on a computer-readable recording medium. The program that describes the processing content to be executed by the virtual machine monitoring device 100 can be recorded on various recording media. For example, a program to be executed by the virtual machine monitoring device 100 can be stored in the storage device 103. The processor 101 loads at least a part of the program in the storage device 103 into the memory 102 and executes the program. Further, the program to be executed by the virtual machine monitoring device 100 can be recorded on a portable recording medium such as an optical disk 24, a memory device 25, and a memory card 27. The program stored in the portable recording medium can be executed after being installed in the storage device 103 under the control of the processor 101, for example. The processor 101 can also read and execute the program directly from the portable recording medium.

In the system having the above configuration, the virtual machine monitoring device 100, the load balancer 200, and the server 300 operate in cooperation with each other, so that the amount of resources allocated to the virtual machine in operation can be dynamically changed. In the second embodiment, before changing the amount of resources allocated to the virtual machine in operation, it is verified whether or not the amount of resources may be changed by using the clone of the virtual machine in operation. .. If there is no problem as a result of verification, the amount of resources allocated to the virtual machine in operation is changed.

In the following description, a virtual machine in operation may be simply referred to as an operation machine. A virtual machine for verification generated as a clone of an operation machine is sometimes called an improvement machine.

FIG. 4 is a block diagram showing a function of each device for changing the amount of resources allocated to the virtual machine. The virtual machine monitoring device 100 has a storage unit 110 and a monitoring unit 120.
The storage unit 110 stores the request monitoring table 111, the resource management table 112, the virtual machine management table 113, and the resource priority table 114. The request monitoring table 111 is a data table for managing the amount of requests sent to the virtual machine in operation and the response time for the requests. The resource management table 112 is a data table for managing the resources allocated to the virtual machine. The virtual machine management table 113 is a data table for managing the purpose of use of the virtual machine, the current resource configuration, the response time, and the like. The storage unit 110 is realized by using, for example, a part of the storage area of the memory or the storage device of the virtual machine monitoring device 100.

The monitoring unit 120 determines an increase or decrease in the resources allocated to the virtual machines 331, 332, ... Based on the amount of requests to the virtual machines 331, 332, .... Then, the monitoring unit 120 instructs the server 300 to add the resources of the virtual machines 331, 332, ..., Or delete the resources from the virtual machines 331, 332, ... According to the determination. The monitoring unit 120 automatically manages the resources of the virtual machines 331, 332, ... Based on the information preset by the operator.

The load balancer 200 has a storage unit 210, a load distribution control unit 220, and a load measurement unit 230.
The storage unit 210 stores the request management table 211 that stores the information (request information) related to the requests transferred to the virtual machines 331, 332, ... In the load balancer 200. The storage unit 210 is realized by using, for example, a part of the storage area of the memory or the storage device included in the load balancer 200.

The load balancing control unit 220 transfers the requests received from the plurality of terminal devices 31, 32, ... To the operating virtual machines among the virtual machines 331, 332, ... In the server 300. For example, when there are a plurality of virtual machines in operation, the load balancing control unit 220 determines a request transfer destination so that the loads of those virtual machines are even. When the load balancing control unit 220 receives a response to a request from an operating virtual machine in the server 300, the load balancing control unit 220 transmits the received response to the terminal device that sends the request.

The load measurement unit 230 measures the amount of requests transferred to the virtual machine in operation and the response time to the request. For example, the load measuring unit 230 measures the request amount and the response time of the virtual machine specified by the monitoring instruction in response to the monitoring instruction of the request amount and the response time from the monitoring unit 120.

Further, the load measurement unit 230 stores information about the request transferred to each of the plurality of virtual machines 331, 332, ... As request information in the request management table 211 in the storage unit 210. Then, the load measurement unit 230 transmits a request based on the request information to the corresponding virtual machine in response to the response time identification instruction of the virtual machine generated as the improvement machine from the virtual machine monitoring device 100. The load measurement unit 230 measures the response time for a request from a virtual machine, which is an improvement machine. The load measurement unit 230 transmits the measured response time to the virtual machine monitoring device 100.

The server 300 has a storage unit 310 and a hypervisor 320.
The storage unit 310 stores snapshot 311 of any of the plurality of virtual machines 321, 322, .... A snapshot is information that can reproduce the state of a virtual machine in operation at a specific point in time. For example, the snapshot includes information in the memory of the virtual machine in operation, information on the contents of the storage device, and information in the registers of the CPU.

The hypervisor 320 manages a plurality of virtual machines 331, 332, .... Specifically, the hypervisor 320 determines the resources to be allocated to each of the plurality of virtual machines 331, 332, ..., And operates the plurality of virtual machines 331, 332, ... Using the allocated resources. For example, when the hypervisor 320 generates the virtual machine 331, the hypervisor 320 determines the amount of resources to be allocated to the virtual machine 331 from the resources of the server 300. Next, the hypervisor 320 generates the virtual CPU 331a, the virtual memory 331b, and the virtual storage device 331c with the determined amount of resources. The hypervisor 320 creates a virtual machine 331 having these virtual resources. The virtual machine 331 executes, for example, the business application software 331d. The virtual machine 331 functions as, for example, an application server by executing the business application software 331d. Then, the virtual machine 331 executes the process for the received request based on the business application software 331d, and transmits a response indicating the process result.

The plurality of virtual machines 331, 332, ... Some are in operation and some are used as improvement machines. For example, a virtual machine in operation provides a service in response to requests from a plurality of terminal devices 31, 32, .... The improved machine is allocated the same amount of resources as the changed resource of the operating virtual machine whose allocated resource amount is scheduled to be changed. Then, the improvement machine is used to verify whether the required quality of service can be satisfied with the changed amount of resources.

In such a system, the monitoring unit 120 manages the addition or deletion of the resource amount of the plurality of virtual machines 331, 332, .... For example, the monitoring unit 120 acquires the measurement results of the request amount and the response time from the load balancer 200, and records the acquired measurement results in the request monitoring table 111. Then, the monitoring unit 120 calculates the volatility of the request amount for the virtual machine from the measurement results acquired in the past and the latest measurement results.

Further, the monitoring unit 120 acquires the resource usage information indicating the resource usage rate for each type of the resource of the virtual machine from the hypervisor 320 operating on the server 300. The monitoring unit 120 records the acquired resource usage information in the resource priority table 114. Then, the monitoring unit 120 determines the priority of resource change for each resource type from the comparison result of the resource usage rate between the resource types.

The monitoring unit 120 performs a process of reducing the allocated resources to the virtual machine (cost improvement) or adding the allocated resources (response improvement) based on the fluctuation rate of the request amount. The monitoring unit 120 acquires the resource configuration information of the virtual machine from the hypervisor 320 when performing each process of cost improvement or response improvement. The monitoring unit 120 records the acquired resource configuration information in the virtual machine management table 113. The monitoring unit 120 selects the resource to be allocated to the virtual machine to be changed and the amount of change by using the resource configuration information. The selection of the resource to be changed is determined based on the priority (order) of the resource change. Based on the decision, the monitoring unit 120 instructs the hypervisor 320 in the server 300 to change the amount of resources allocated to the virtual machine.

The function of each element shown in FIG. 4 can be realized, for example, by causing a computer to execute a program module corresponding to the element.
Hereinafter, the information stored in each device will be described in detail with reference to FIGS. 5 to 9.

FIG. 5 is a diagram showing an example of a request monitoring table. The request monitoring table 111 is provided with columns for interval period, request amount, response time, volatility, cost improvement, and response improvement.

In the interval period column, the interval period, which is the unit interval for measuring the request amount, is set. In the example of FIG. 5, the interval period is "24 hours", and the request amount is measured every 24 hours. The request amount column is divided into the previous column and the new column. In the previous column of the request amount, the number of requests measured at the time of measuring the previous request amount is set. In the new field of request amount, the number of requests measured at the time of the last measurement of the request amount is set. In the response time column, a representative value (for example, an average value) of the response time within the measurement period of the new request amount is set. In the volatility column, the volatility R of the request amount is set. The volatility R is a value obtained by dividing the new request amount by the previous request amount. In the cost improvement column, whether or not to execute the cost improvement process is set. For example, if the volatility is less than the cost improvement threshold (“0.5” in the example of FIG. 5), the cost improvement process is executed. In the response improvement column, whether or not to execute the response improvement process is set. For example, when the volatility is larger than the response improvement threshold value (“1.5” in the example of FIG. 5), the response improvement process is executed.

FIG. 6 is a diagram showing an example of a resource management table. The resource management table 112 is provided with columns for resource, Min, Max, increase / decrease unit, resource unit price, and priority.

The resource type is set in the resource column. In the Min column, the minimum value of the amount of resources that can be allocated to the virtual machine is set. In the Max column, the maximum value of the amount of resources that can be allocated to the virtual machine is set. In the case of a storage device, a rank for each type of storage device is set instead of the amount of resources. As the storage type, for example, five types of "type 1" to "type 5" are provided. The number to the right of the type indicates the rank, and the smaller the number indicating the rank, the smaller the amount of resources. For example, a "type 1" storage device is a near-line storage system of RAID (Redundant Arrays of Inexpensive Disks) 5 using a hard disk. The "type 2" storage device is a RAID 5 storage system using a hard disk with a disk rotation speed of 10,000 rotations. The "type 3" storage device is a RAID 1 storage system using a hard disk with a disk rotation speed of 10,000 rotations. The "type 4" storage device is a RAID 5 storage system using SSDs. The "type 5" storage device is a RAID 1 storage system using SSDs.

In the increase / decrease unit column, the minimum unit that can be increased / decreased when increasing / decreasing the amount of resources is set. For example, in the case of memory, it can be increased or decreased in units of 2 GB. If it is a CPU, it can be increased or decreased in units of two CPU cores (2 cores). If it is a storage device, the rank can be changed step by step.

In the resource unit price column, the monthly usage amount of resources for the increase / decrease unit is set. For example, if the memory allocated to a virtual machine is increased by 2 GB, the extra usage fee paid by the user using the virtual machine is $ 80. If the number of CPU cores allocated to a virtual machine is increased by two, the extra usage fee paid by the user using the virtual machine is $ 100. Further, if the rank of the storage device assigned to the virtual machine is increased by 1, the extra usage fee paid by the user using the virtual machine is $ 40.

The priority column is the priority of the resource to be added or deleted when adding or deleting a resource to the virtual machine. "High" is set in the priority column of the resource with the highest priority, and "Mid (medium)" is set in the priority column of the resource with the next highest priority. "Low" is set in the priority column of the resource with low. In the example of FIG. 6, the memory has the highest priority, the CPU has the highest priority, and the storage device has the lowest priority.

FIG. 7 is a diagram showing an example of a virtual machine management table. The virtual machine management table 113 is provided with columns for virtual machine, purpose, resource configuration, usage cost, and response time.

In the virtual machine column, the name of the virtual machine operating in the server 300 is set. In the purpose column, the purpose of using the virtual machine is set. The objectives are "operation", "cost improvement", and "response improvement". The purpose "operation" indicates that it is a virtual machine used for the operation of the service. The purpose "cost improvement" indicates that a virtual machine is used as an improvement machine used for performance evaluation for cost improvement. The purpose "response improvement" indicates that a virtual machine is used as an improvement machine used for performance evaluation for response improvement.

In the resource configuration column, the amount of resources allocated to the virtual machine is set. For example, in the resource configuration column, the amount of resources for each of the CPU, memory, and storage device is set.

In the usage cost column, the monthly usage fee of the virtual machine is set. The usage cost is calculated based on the amount of resources allocated to the virtual machine.
Information about the response time of the virtual machine is set in the response time column. Information on response time includes reference values, safety factors, measured values, and determination results. The reference value (Cr) is the response time required for the virtual machine. The safety factor (k) is a constant of 1 or less used for calculating the response time threshold value so that the response time does not exceed the reference value. The value obtained by multiplying the reference value by the safety factor is the response time threshold value (k × Cr). In the example of FIG. 7, since the reference value (Cr) is “5 ms” and the safety factor is “0.85”, “4.25 ms (= 0.85 × 5)” is the threshold value.

In the measurement value (RT) column, the measurement result of the response time of the virtual machine (for example, the average value of the response times measured within a predetermined period) is set. In the determination result column, the comparison result between the response time of the virtual machine and the threshold value is set. In the example of FIG. 7, since the measured value of the virtual machine "VM A" is smaller than the threshold value, the determination result is "OK". Since the measured value of the virtual machine "VM B" is equal to or greater than the threshold value, the determination result is "NG". Since the measured value of the virtual machine "VM C" is smaller than the threshold value, the determination result is "OK".

FIG. 8 is a diagram showing an example of a resource priority table. The resource priority table 114 is provided with columns for resources, resource usage rate: average value, resource usage rate: maximum value, and resource change priority determination value.

The resource type is set in the resource column.
Resource usage rate: In the column of average value, columns of usage rate, rank (PA), and coefficient (kA) are further provided. Resource usage rate: In the average value usage rate column, the average value of the usage rate of the corresponding resource in the virtual machine within a predetermined period is set. Resource usage rate: In the column of the ranking of the average value, the ranking (PA) of the corresponding resource when a plurality of resources are arranged in descending order of the average value of the usage rate is set. Resource usage rate: In the column of the coefficient of the average value, a coefficient (kA) indicating a weight with respect to the rank of the average value of the usage rate when calculating the priority determination value of the change of the corresponding resource is set.

Resource usage rate: In the column of maximum value, columns of usage rate, rank (PM), and coefficient (km) are further provided. Resource usage rate: In the maximum usage rate column, the maximum value of the usage rate of the corresponding resource in the virtual machine within a predetermined period is set. Resource usage rate: In the column of the ranking of the maximum value, the ranking (PM) of the corresponding resource when a plurality of resources are arranged in descending order of the maximum value of the usage rate is set. Resource utilization: In the coefficient column of the maximum value, a coefficient (km) indicating a weight with respect to the rank of the maximum value of the utilization rate when calculating the priority determination value of the change of the corresponding resource is set.

In the resource change priority determination value column, the priority determination value for changing the amount of the corresponding resource is set. The priority determination value is, for example, a value calculated by "PA × kA + PM × kM". The smaller the priority determination value, the higher the priority for changing the resource amount. For example, when there are three types of resources, when the resources are arranged in ascending order of priority judgment value, the priority of the first resource is "High" and the priority of the second resource is "Mid (medium)". The priority of the third resource is "Low".

The coefficient of the average value or the maximum value of each resource is predetermined by the user based on the nature of the resource. For example, even if the data stored in the memory exceeds the memory capacity, the memory can be dealt with by a technique such as swapping, so that the memory is often used to the maximum extent. That is, the maximum value of the memory usage rate is often close to 100%, and even in such a state, it cannot be determined that the virtual machine is overloaded. Therefore, in the example of FIG. 8, the coefficient with respect to the maximum value of the memory usage rate is set higher than that of other resources. As a result, even if the memory usage rate ranks higher with respect to the maximum resource usage rate, the priority determination value is overestimated.

In addition, the storage device will change the type of storage in order to realize cost improvement or response improvement. For example, the storage device assigned to the virtual machine is changed from the storage device of RAID 5 to the storage device of RAID 1. Such a storage type change process involves the movement of data and takes time. Therefore, in the example of FIG. 8, the coefficient logarithmic to the average value of the resource usage rate of the storage device is set higher than that of other resources. As a result, even if the ranking of the storage device usage rate with respect to the average value of the resource usage rate is high, the priority determination value is overestimated.

FIG. 9 is a diagram showing an example of a request management table stored in the load balancer. The request management table 211 is provided with columns for reception time, request, transfer destination virtual machine, and response time.

In the reception time column, the reception time of the request received from the terminal devices 31, 32, ... Is set. The content of the received request is set in the request field. The name of the virtual machine to which the request is transferred is set in the transfer destination virtual machine field. In the response time column, the time (response time) from when the request is transferred to the virtual machine until the response is returned from the virtual machine is set.

Next, the automatic change processing of the resource amount will be described in detail. The automatic resource amount change process is roughly divided into request monitoring process and optimization process (cost improvement or response improvement process). Hereinafter, the outline of the request monitoring process and the optimization process will be described with reference to FIGS. 10 and 11.

FIG. 10 is a sequence diagram showing an example of the procedure of the request monitoring process. In the request monitoring process, the virtual machine monitoring device 100 first transmits a snapshot generation instruction of the virtual machine in operation to the server 300 (step S101). Upon receiving the snapshot generation instruction, the server 300 generates a snapshot of the virtual machine (step S102).

Next, the virtual machine monitoring device 100 transmits a request amount / response time measurement instruction to the load balancer 200 (step S103). The load balancer 200 counts the amount of requests transferred to the operating virtual machine in the server 300 in response to the request amount / response time measurement instruction, and calculates the average value of the response times for each request (step S104). Then, the load balancer 200 transmits the request amount and the average value of the response times to the virtual machine monitoring device 100 as the measurement result (step S105).

Next, the virtual machine monitoring device 100 transmits a request for the resource usage rate of the virtual machine in operation to the server 300 (step S106). The server 300 responds to the information on the resource utilization rate of the virtual machine in response to the request from the virtual machine monitoring device 100 (step S107). For example, in the server 300, each OS of the virtual machines 331, 332, ... Measures the resource usage rate. Then, the hypervisor 320 acquires the resource usage information from each OS of the virtual machines 331, 332, ... In response to the request from the virtual machine monitoring device 100, and transmits the information to the virtual machine monitoring device 100.

The virtual machine monitoring device 100 calculates a priority determination value for each resource based on the information on the resource usage rate of the virtual machine, and determines the priority of resource change (step S108). Next, the virtual machine monitoring device 100 calculates the volatility of the request amount based on the request amount (step S109). Then, the virtual machine monitoring device 100 analyzes the acquired information and selects an optimization process (cost improvement process or response improvement process).

FIG. 11 is a sequence diagram showing an example of the procedure of the optimization process. In the optimization process, the virtual machine monitoring device 100 requests the server 300 for information on the configuration of resources allocated to the virtual machine (operating machine) being operated (step S121). Then, the server 300 transmits the information indicating the configuration of the resources allocated to the operation machine to the virtual machine monitoring device 100 (step S122).

The virtual machine monitoring device 100 stores the amount of resources allocated to the operation machine (step S123). Next, the virtual machine monitoring device 100 determines the optimum amount of resources of the operating machine (step S124). Then, the virtual machine monitoring device 100 instructs the server 300 to generate an improved machine (step S125). The server 300 creates a virtual machine to be used as an improvement machine according to the instruction (step S126).

Next, the virtual machine monitoring device 100 transmits a snapshot application instruction to the improvement machine to the server 300 (step S127). The server 300 restores the snapshot of the operation machine on the improvement machine according to the instruction (step S128). Subsequently, the virtual machine monitoring device 100 transmits a resource amount change instruction of the improvement machine to the server 300 so as to allocate the optimum resource amount determined in step S124 to the improvement machine (step S129). The server 300 changes the amount of resources allocated to the improvement machine according to the instruction (step S130).

The virtual machine monitoring device 100 transmits a response time measurement instruction of the improved machine to the load balancer 200 (step S131). The load balancer 200 records a log of requests sent to the operation server in the past, and sends the recorded request to the improvement machine in response to a response time measurement instruction from the virtual machine monitoring device 100 ( Step S132). On the server 300, the improvement machine executes processing in response to the request. Then, the improvement machine in the server 300 sends a response to the request to the load balancer 200 (step S133). The load balancer 200 measures the time from the transmission of the request to the response, and uses the measurement result as the response time. The load balancer 200 transmits the measurement result of the response time to the virtual machine monitoring device 100 (step S134).

The virtual machine monitoring device 100 determines whether or not the response time of the improved machine satisfies the response time requirement required for the operating machine (step S135). When the virtual machine monitoring device 100 determines that the response time requirement is satisfied, the virtual machine monitoring device 100 instructs the server 300 to change the resource amount of the operating machine to the optimum resource amount determined in step S124 (step). S136). The server 300 changes the resource amount of the operation machine according to the instruction (step S137).

In this way, the amount of resources of the production machine is automatically changed.
Next, the processing executed by each device will be described in detail.
FIG. 12 is the first half of a flowchart showing an example of the procedure of request monitoring processing in the virtual machine monitoring device. Hereinafter, the process shown in FIG. 12 will be described along with the step numbers.

[Step S201] The monitoring unit 120 generates various data tables. Specifically, the monitoring unit 120 newly generates a request monitoring table 111, a resource management table 112, and a resource priority table 114. At this point, only the interval period is set in the request monitoring table 111. The interval period is a period for counting the request amount. The interval period is predetermined by the user.

If the request monitoring table 111, the resource management table 112, and the resource priority table 114 have already been generated by the request monitoring process executed in the past, the process in step S201 is omitted.

[Step S202] The monitoring unit 120 receives input of data to be set in various data tables. For example, the monitoring unit 120 receives input of data to be set in the resource, Min, Max, increase / decrease unit, and resource unit price columns in each column (see FIG. 6) of the resource management table 112. Then, the monitoring unit 120 sets the data input by the user in the resource management table 112. Further, the monitoring unit 120 inputs a value to be set in each column of the resource priority table 114 (see FIG. 8) in the resource usage rate: average value coefficient column and the resource usage rate: maximum value coefficient column. Accept. Then, the monitoring unit 120 sets the data input by the user in the resource priority table 114.

[Step S203] The monitoring unit 120 connects communication between the load measuring unit 230 in the load balancer 200 and the hypervisor 320 in the server 300 via the management network 20.

[Step S204] The monitoring unit 120 instructs the hypervisor 320 in the server 300 to create a snapshot of the virtual machine in operation.
[Step S205] The monitoring unit 120 transmits a measurement instruction of the request amount and the response time to the load measuring unit 230 in the load balancer 200. In response to this measurement instruction, the load measurement unit 230 monitors the request to the virtual machine in operation, and starts measuring the request amount and the response time.

[Step S206] The monitoring unit 120 waits for a predetermined time. The predetermined time to wait is preset by the user.
[Step S207] The monitoring unit 120 instructs the load measuring unit 230 in the load balancer 200 to end the measurement of the request information at the same time as the end of monitoring the request amount. In response to the measurement end instruction, the load measurement unit 230 responds with the measurement result. The measurement result includes, for example, the amount of requests sent to the virtual machine within a predetermined period and the average response time during that period.

[Step S208] The monitoring unit 120 updates the request monitoring table 111 based on the returned measurement result. For example, the monitoring unit 120 sets the request amount included in the latest measurement result in a new column of the request amount in the request monitoring table 111. At this time, if the request amount shown in the previous measurement result is set in the new column, after copying the request amount to the previous column of the request amount, the latest request amount is set in the new column. To write. Further, the monitoring unit 120 sets the response time included in the latest measurement result in the response time column of the request monitoring table 111. Further, the monitoring unit 120 calculates the volatility by dividing the new request amount by the previous request amount, and sets it in the volatility column of the request monitoring table 111.

[Step S209] The monitoring unit 120 determines whether or not the previous request amount is set in the request monitoring table 111. If the previous request amount is set, the monitoring unit 120 advances the process to step S210. If the previous request amount is not set, the monitoring unit 120 advances the process to step S205.

[Step S210] The monitoring unit 120 requests the hypervisor 320 in the server 300 for information on the resource usage rate of the virtual machine. In response to this request, the hypervisor 320 responds with information about resource utilization. The information regarding the resource usage rate includes the average value and the maximum value of the usage rate of each of the CPU, the memory, and the storage device.

[Step S211] The monitoring unit 120 updates the resource priority table 114 based on the information regarding the resource usage rate. For example, the monitoring unit 120 sets the average value of the usage rate of each resource in the resource usage rate: average value usage rate column of the resource priority table 114, and the resource usage rate of the resource priority table 114: maximum value. Set the maximum value of the usage rate of each resource in the usage rate column of. The monitoring unit 120 then proceeds to step S221 (see FIG. 13).

FIG. 13 is the latter half of the flowchart showing an example of the procedure of the request monitoring process in the virtual machine monitoring device. Hereinafter, the process shown in FIG. 13 will be described along with the step numbers.
[Step S221] The monitoring unit 120 determines the order of the resource usage rate of each resource. For example, the monitoring unit 120 refers to the resource priority table 114, first compares the usage rate of each resource set in the resource usage rate: average value usage rate column, and determines the ranking in descending order of usage rate. do. Then, the monitoring unit 120 sets the rank of the average value of the resource usage rate in the column of the resource usage rate: rank of the average value of the resource priority table 114 in association with each resource. Next, the monitoring unit 120 compares the usage rates of each resource set in the resource usage rate: maximum value column of the resource priority table 114, and determines the order in descending order of usage rate. Then, the monitoring unit 120 sets the order of the maximum value of the resource usage rate in the column of the resource usage rate: maximum value order of the resource priority table 114 in association with each resource.

[Step S222] The monitoring unit 120 determines the priority of resource change for each resource. For example, the monitoring unit 120 adds the value obtained by multiplying the rank (PA) of the average value of the usage rate by the coefficient (kA) and the value obtained by multiplying the rank (PM) of the maximum value of the usage rate by the coefficient (km). Is the priority judgment value for resource change. The monitoring unit 120 sets the resource change priority determination value calculated for each resource in the resource change priority determination value column of the resource priority table 114.

Further, the monitoring unit 120 determines the priority of each resource based on the priority determination value of the resource change of each resource, and sets the priority in the priority column of the resource management table 112. For example, the monitoring unit 120 sets the priority of the resource having the highest resource change priority judgment value as "High", the priority of the resource having the second highest resource change priority judgment value as "Mid", and the resource change priority. The priority of the resource with the lowest degree judgment value is set to "Low".

[Step S223] The monitoring unit 120 determines whether or not the response time is equal to or greater than the threshold value. The threshold value is the reference value × safety factor set in the virtual machine management table 113. When the response time is equal to or greater than the threshold value, the monitoring unit 120 advances the process to step S228. If the response time is less than the threshold value, the monitoring unit 120 advances the process to step S224.

[Step S224] The monitoring unit 120 determines the tendency of increase / decrease in the request amount based on the volatility set in the request monitoring table 111. For example, when the volatility is less than the cost improvement threshold value (for example, "0.5"), the monitoring unit 120 determines that the request amount tends to decrease, and "Yes" in the "cost improvement column" of the request monitoring table 111. "Written. Further, when the volatility is equal to or higher than the cost improvement threshold value, the monitoring unit 120 determines that the request amount does not tend to decrease, and writes "No" in the cost improvement column of the request monitoring table 111. Further, the monitoring unit 120 determines that the request amount tends to increase when the volatility is larger than the response improvement threshold value (for example, "1.5"), and "Yes" in the response improvement column of the request monitoring table 111. To write. Further, the monitoring unit 120 determines that the request amount does not tend to increase when the volatility is equal to or less than the response improvement threshold value, and writes "No" in the response improvement column of the request monitoring table 111.

[Step S225] The monitoring unit 120 inquires about the cost improvement status of the request monitoring table 111. For example, when "Yes" is set in the cost improvement column, the monitoring unit 120 advances the process to step S227. When "No" is set in the cost improvement column, the monitoring unit 120 advances the process to step S226.

[Step S226] The monitoring unit 120 inquires about the response improvement status of the request monitoring table 111. If "Yes" is set in the response improvement column, the monitoring unit 120 proceeds to step S228. When "No" is set in the response improvement column, the monitoring unit 120 proceeds to step S205 (see FIG. 12).

[Step S227] The monitoring unit 120 performs cost improvement processing. Details of the cost improvement process will be described later (see FIGS. 15 and 16). When the cost improvement process is completed, the monitoring unit 120 ends the request monitoring process.

[Step S228] The monitoring unit 120 performs a response improvement process. Details of the response improvement process will be described later (see FIGS. 17 and 18). When the cost improvement process is completed, the monitoring unit 120 ends the request monitoring process.

In this way, the request is monitored, and if the request amount suddenly increases, the response improvement process is performed, and if the request amount suddenly decreases, the cost improvement process is performed.
Here, before explaining the details of the cost improvement process and the response improvement process, the request amount / response time measurement process by the load balancer 200 will be described in detail.

FIG. 14 is a flowchart showing an example of the procedure for measuring the request amount and response time. The request amount monitoring process is executed when a request amount / response time measurement instruction is received from the monitoring unit 120 in the virtual machine monitoring device 100. Hereinafter, the process shown in FIG. 14 will be described along with the step numbers.

[Step S301] The load measuring unit 230 of the load balancer 200 monitors the load distribution process by the load distribution control unit 220, and sends requests from the terminal devices 31, 32, ... To the virtual machines 331, 332, .... Determine if it has been transferred. When the request is transferred, the load measurement unit 230 advances the process to step S302. If the load measurement unit 230 does not transfer the request, the load measurement unit 230 proceeds to step S304.

[Step S302] The load measurement unit 230 registers the request information regarding the transferred request in the request management table 211.
[Step S303] The load measurement unit 230 waits for a response from the virtual machine corresponding to the transferred request, and measures the response time for the request. Then, the load measurement unit 230 associates the measured response time with the request information registered in step S302, and sets it in the response time column of the request management table 211.

[Step S304] The load measuring unit 230 determines whether or not a measurement end instruction has been received from the monitoring unit 120 in the virtual machine monitoring device 100. When the load measurement unit 230 receives the measurement end instruction, the load measurement unit 230 proceeds to step S305. If the load measurement unit 230 has not received the measurement end instruction, the load measurement unit 230 proceeds to step S301.

[Step S305] The load measurement unit 230 counts the request amount based on the request information stored in the request management table 211. For example, in the load measurement unit 230, the time within the period from receiving the request amount / response time measurement instruction to receiving the measurement end instruction is set as the reception time, and the transfer destination virtual machine is an operating virtual machine. Identify the request information for a request. Then, the load measurement unit 230 counts the number of the specified request information and sets it as the request amount to the virtual machine in operation.

[Step S306] The load measuring unit 230 calculates the average value of the response times. For example, the response times of the request information specified in step S305 are totaled, and the result of dividing the total by the request amount is used as the average value of the response times.

[Step S307] The load measurement unit 230 transmits the request amount and the average value of the response time to the monitoring unit 120 of the virtual machine monitoring device 100 as the measurement result.
In this way, based on the request amount and the response time (average value) measured by the load balancer 200, it is determined and determined whether or not to execute the cost improvement process or the response improvement process in the virtual machine monitoring device 100. Processing is executed.

FIG. 15 is the first half of a flowchart showing an example of the cost improvement processing procedure. Assuming that the cost improvement process of the virtual machine 331 in operation is executed, the process shown in FIG. 15 will be described below along with the step numbers.

[Step S401] The monitoring unit 120 of the virtual machine monitoring device 100 generates a virtual machine management table 113 that manages information of the virtual machine 331. If the virtual machine management table 113 has already been generated, the monitoring unit 120 does not generate a new virtual machine management table 113.

For example, the monitoring unit 120 registers the record of the virtual machine 331 in the virtual machine management table 113. At this point, the purpose, usage cost, response time reference value, safety factor, and determination result are set in the registered record in association with the name "VMA" of the virtual machine 331. The purpose set at this time is "operation". The reference value / safety factor of the response time is a value specified in advance by the user. The measured value of the response time is the average value of the response times calculated by the load balancer 200 in step S306 (see FIG. 14). The average value of the response time can be obtained from the response time column of the request monitoring table 111. The determination of the response time of the virtual machine 331, which is the operating machine, is performed in step S223 (see FIG. 13). The monitoring unit 120 sets the determination result in step S223 in the response time determination result column of the virtual machine management table 113.

If the virtual machine management table 113 has already been generated by the request monitoring process executed in the past, the process in step S401 is omitted.
[Step S402] The monitoring unit 120 acquires the resource configuration (allocation amount of each resource) of the virtual machine 331 from the hypervisor 320. Then, the monitoring unit 120 sets the acquired resource configuration in the virtual machine management table 113.

[Step S403] The monitoring unit 120 adds a copy of the record of the virtual machine 331 registered in the virtual machine management table 113 to the virtual machine management table 113 as a record of the virtual machine used as the improvement machine. Then, the monitoring unit 120 changes the name of the virtual machine in the newly added record, and sets "cost improvement" in the target column.

Next, the monitoring unit 120 sets the resource configuration of the low specifications as compared with the virtual machine 331 in operation to the virtual machine used as the improvement machine by the processing of steps S404 to S406.

[Step S404] The monitoring unit 120 subtracts one unit from the allocation amount of the resource having the higher priority among the allocation amounts of each resource shown in the resource configuration for the record of the improvement machine in the virtual machine management table 113. .. The priority of the resource can be determined by referring to the priority column of the resource management table 112. The monitoring unit 120 sets the allocated amount of the high-priority resource after subtraction in the resource configuration column of the record of the improved machine in the virtual machine management table 113. If the high-priority resource allocated to the improvement machine is the minimum available value of the resource, the monitoring unit 120 subtracts one unit from the allocation amount of the next highest-priority resource. The minimum available value for each resource is set in the Min column of the resource management table 112.

In the example of the virtual machine management table 113 of FIG. 7, the virtual machine “VMA” is a virtual machine in operation. In the resource management table 112 shown in FIG. 6, the memory has the highest priority. The amount of memory allocated to the virtual machine "VM A" is 4 GB. Therefore, the monitoring unit 120 sets the value "2GB" obtained by subtracting one unit (2GB) from 4GB in the memory column of the resource configuration of the record of the virtual machine "VM B" whose purpose is "cost improvement" in the virtual machine management table 113. To write.

Here, if the memory capacity allocated to the operating virtual machine "VM A" is 2 GB (minimum value), the allocated amount of the CPU with the next highest priority is subtracted by 1 unit (2 cores). .. At this time, the monitoring unit 120 may add only one unit for the resource (memory) whose allocated resource amount cannot be subtracted at the minimum value.

[Step S405] The monitoring unit 120 calculates the usage cost of the virtual machine in operation and the usage cost of the improvement machine. For example, the monitoring unit 120 acquires the increase / decrease unit price for each resource from the resource management table 112. Next, the monitoring unit 120 multiplies the resource allocation amount to the virtual machine by the increase / decrease unit price shown in the virtual machine management table 113 to obtain the usage cost for each resource of the virtual machine. Then, the monitoring unit 120 totals the usage costs for each resource and sets it as the usage cost of the corresponding virtual machine. The monitoring unit 120 sets the usage costs calculated for each of the usage cost of the virtual machine in operation and the usage cost of the improvement machine in the usage cost column of the virtual machine management table 113.

[Step S406] The monitoring unit 120 determines whether or not the improvement machine has a lower utilization cost. If the improvement machine has a lower usage cost, the monitoring unit 120 advances the process to step S411. If the operating machine has a lower utilization cost or the utilization cost does not change, the monitoring unit 120 proceeds to step S404 to further reduce the resource amount of the improvement machine. Although not shown in the flowchart of FIG. 15, if there are no resources that can be reduced, the cost improvement process ends.

FIG. 16 is the latter half of the flowchart showing an example of the procedure of the cost improvement process. Hereinafter, the process shown in FIG. 16 will be described along with the step numbers.
[Step S411] The monitoring unit 120 instructs the hypervisor 320 of the server 300 to clone the virtual machine 331 in operation. The hypervisor 320 creates a virtual machine that is a clone of the virtual machine 331 according to the instructions. Hereinafter, it is assumed that the virtual machine 332 is generated as a clone.

[Step S412] The monitoring unit 120 instructs the hypervisor 320 to apply the snapshot to the generated virtual machine 332. In response to this instruction, the hypervisor 320 applies a snapshot of the virtual machine 331 to the virtual machine 332. As a result, the operating state of the virtual machine 331 is reproduced in the virtual machine 332.

[Step S413] The monitoring unit 120 instructs the hypervisor 320 to reduce the amount of resources allocated to the virtual machine 332, which is a clone of the virtual machine 331 in operation. For example, the monitoring unit 120 notifies the hypervisor 320 of the resource configuration of the virtual machine for the purpose “cost improvement” of the virtual machine management table 113, and the hypervisor 320 so that the resource configuration of the virtual machine 332 is the notified resource configuration. Instruct.

[Step S414] The monitoring unit 120 specifies to the hypervisor 320 the name of the virtual machine 332 that is a clone of the virtual machine 331 in operation (for example, "VM B"), and instructs the hypervisor 320 to start the virtual machine 332. The hypervisor 320 starts the virtual machine 332 according to the instruction.

[Step S415] The monitoring unit 120 specifies the name "VM B" of the virtual machine 332 to the load measuring unit 230 of the load balancer 200, and instructs the load measuring unit 230 to measure the response time of the virtual machine 332. The load measurement unit 230 transmits a request to the virtual machine 332 based on the request information stored in the storage unit 210 in response to the measurement instruction, and measures the response time for the request. For example, the load measurement unit 230 transmits a plurality of requests to the virtual machine 332, and sets the average value of the response times for each request as the response time of the virtual machine 332.

[Step S416] The monitoring unit 120 acquires the measurement result of the response time of the virtual machine 332 from the load measuring unit 230. The monitoring unit 120 sets the acquired response time in the measured value column of the response time of the record of the purpose "cost improvement" in the virtual machine management table 113.

[Step S417] The monitoring unit 120 determines whether or not the response time of the virtual machine 332 is less than the response time threshold value (reference value × safety factor) required for the virtual machine 331 in operation. When the response time is less than the threshold value, the monitoring unit 120 sets "OK" in the response time determination result column of the record of the purpose "cost improvement" in the virtual machine management table 113, and proceeds to the process in step S418. If the response time is equal to or greater than the threshold value, the monitoring unit 120 sets "NG" in the response time determination result column of the record of the purpose "cost improvement" in the virtual machine management table 113, and ends the cost improvement process. ..

[Step S418] The monitoring unit 120 instructs the hypervisor 320 in the server 300 to reduce the amount of resources allocated to the virtual machine 331 in operation. For example, the monitoring unit 120 notifies the hypervisor 320 of the resource configuration of the virtual machine for the purpose “cost improvement” of the virtual machine management table 113, and sets the resource configuration of the virtual machine 331 in operation as the notified resource configuration. Instruct the hypervisor 320.

[Step S419] The monitoring unit 120 updates the resource configuration of the virtual machine 331 in operation in the virtual machine management table 113. For example, the monitoring unit 120 writes a copy of the resource configuration of the record of the purpose "cost improvement" by overwriting the resource configuration of the record of the purpose "operation". After that, the monitoring unit 120 advances the process to step S402, and searches for whether or not the resource amount can be further reduced.

In this way, the amount of resources allocated to the virtual machine 331 in operation is reduced in order from the resource having the highest priority. Ultimately, the amount of resources allocated to the virtual machine 331 in operation is the minimum amount of resources that satisfies the response time condition.

Next, the response improvement process will be described in detail.
FIG. 17 is the first half of a flowchart showing an example of the procedure of the response improvement process. Assuming that the cost improvement process of the virtual machine 331 in operation is executed, the process shown in FIG. 17 will be described below along with the step numbers.

[Step S501] The monitoring unit 120 of the virtual machine monitoring device 100 generates a virtual machine management table 113 that manages information of the virtual machine 331. If the virtual machine management table 113 has already been generated by the request monitoring process executed in the past, the process in step S501 is omitted. The details of the generation process of the virtual machine management table 113 are the same as those in step S401 of FIG.

[Step S502] The resource configuration (allocation amount of each resource) of the virtual machine 331 is acquired from the hypervisor 320. Then, the monitoring unit 120 sets the acquired resource configuration in the virtual machine management table 113.

[Step S503] The monitoring unit 120 adds a copy of the record of the virtual machine 331 registered in the virtual machine management table 113 to the virtual machine management table 113 as a record of the virtual machine used as the improvement machine. Then, the monitoring unit 120 changes the name of the virtual machine in the newly added record, and sets "response improvement" in the target column.

[Step S504] The monitoring unit 120 adds one unit to the allocation amount of the resource having the higher priority among the allocation amounts of each resource shown in the resource configuration for the record of the improvement machine in the virtual machine management table 113. .. The monitoring unit 120 sets the allocated amount of the high-priority resource after addition in the resource configuration column of the record of the improved machine in the virtual machine management table 113. If the high-priority resource allocated to the improvement machine is the maximum available value of the resource, the monitoring unit 120 adds one unit to the allocation amount of the next highest-priority resource. The maximum available value of each resource is set in the Max column of the resource management table 112.

For example, if the resource with the highest priority is memory and the amount of memory allocated to the virtual machine in operation is 2 GB, 1 unit (2 GB) is added to the amount of memory of the virtual machine for the purpose "response improvement". It becomes 4GB.

When the amount of memory allocated to the virtual machine in operation is 64 GB (maximum value), the amount allocated to the CPU, which is the next highest priority resource, is added by 1 unit (2 cores). At this time, the allocation amount for the resource (memory) whose resource usage amount is the maximum value may be subtracted by one unit.

After the resource amount increase processing, the monitoring unit 120 proceeds to the processing in step S511 (see FIG. 18).
FIG. 18 is the latter half of the flowchart showing an example of the procedure of the response improvement process. Hereinafter, the process shown in FIG. 18 will be described along with the step numbers.

[Step S511] The monitoring unit 120 instructs the hypervisor 320 of the server 300 to clone the virtual machine 331 in operation. The hypervisor 320 creates a virtual machine that is a clone of the virtual machine 331 according to the instructions. Hereinafter, it is assumed that the virtual machine 332 is generated as a clone.

[Step S512] The monitoring unit 120 instructs the hypervisor 320 to apply the snapshot to the generated virtual machine 332. In response to this instruction, the hypervisor 320 applies a snapshot of the virtual machine 331 to the virtual machine 332. As a result, the operating state of the virtual machine 331 is reproduced in the virtual machine 332.

[Step S513] The monitoring unit 120 instructs the hypervisor 320 to increase the amount of resources allocated to the virtual machine 332, which is a clone of the virtual machine 331 in operation. For example, the monitoring unit 120 notifies the hypervisor 320 of the resource configuration of the virtual machine for the purpose “response improvement” of the virtual machine management table 113, and the hypervisor 320 so that the resource configuration of the virtual machine 332 is the notified resource configuration. Instruct.

[Step S514] The monitoring unit 120 specifies to the hypervisor 320 the name of the virtual machine 332 (for example, "VMC") which is a clone of the virtual machine 331 in operation, and instructs the hypervisor 320 to start the virtual machine 332. The hypervisor 320 starts the virtual machine 332 according to the instruction.

[Step S515] The monitoring unit 120 specifies the name "VMC" of the virtual machine 332 to the load measuring unit 230 of the load balancer 200, and instructs the load measuring unit 230 to measure the response time of the virtual machine 332. The load measurement unit 230 transmits a request to the virtual machine 332 based on the request information stored in the storage unit 210 in response to the measurement instruction, and measures the response time for the request. For example, the load measurement unit 230 transmits a plurality of requests to the virtual machine 332, and sets the average value of the response times for each request as the response time of the virtual machine 332.

[Step S516] The monitoring unit 120 acquires the measurement result of the response time of the virtual machine 332 from the load measuring unit 230. The monitoring unit 120 sets the acquired response time in the measured value column of the response time of the record of the purpose "response improvement" in the virtual machine management table 113.

[Step S517] The monitoring unit 120 determines whether or not the response time of the virtual machine 332 is less than the response time threshold value (reference value × safety factor) required for the virtual machine 331 in operation. When the response time is less than the threshold value, the monitoring unit 120 sets "OK" in the response time determination result column of the record of the purpose "response improvement" in the virtual machine management table 113, and proceeds to the process in step S518. When the response time is equal to or greater than the threshold value, the monitoring unit 120 sets "NG" in the response time determination result column of the record of the purpose "response improvement" in the virtual machine management table 113, and steps S504 (FIG. FIG. 17).

[Step S518] The monitoring unit 120 instructs the hypervisor 320 in the server 300 to increase the amount of resources allocated to the virtual machine 331 in operation. For example, the monitoring unit 120 notifies the hypervisor 320 of the resource configuration of the virtual machine for the purpose “response improvement” of the virtual machine management table 113, and sets the resource configuration of the virtual machine 331 in operation as the notified resource configuration. Instruct the hypervisor 320.

[Step S519] The monitoring unit 120 updates the resource configuration of the virtual machine 331 in operation in the virtual machine management table 113. For example, the monitoring unit 120 writes a copy of the resource configuration of the record of the purpose "response improvement" by overwriting the resource configuration of the record of the purpose "operation".

In this way, the amount of resources allocated to the virtual machine can be changed as the amount of requests to the virtual machine changes. For example, if the amount of requests to a virtual machine is on the rise, the amount of resources allocated to the virtual machine can be increased without waiting for the response time to deteriorate. As a result, it is possible to suppress the delay in changing the amount of resources of the virtual machine. That is, the amount of resources can be changed more quickly than when the amount of resources allocated to the virtual machine is changed after detecting that the response time of the virtual machine has deteriorated. As a result, it is possible to prevent the response time from deteriorating due to the overload of the virtual machine.

If the amount of requests to the virtual machine is decreasing, the amount of resources allocated to the virtual machine can be reduced. As a result, it is possible to prevent excessive allocation of resources to the virtual machine and effectively utilize the resources of the entire system. Moreover, if the user using the virtual machine pays for the amount of resources allocated to the virtual machine, the cost of using the virtual machine can be reduced by minimizing the amount of resources. Can be done.

Further, in the second embodiment, it is confirmed that the condition regarding the response time is satisfied by the improvement machine before changing the amount of resources allocated to the virtual machine in operation. This prevents the amount of resources allocated from being excessively reduced.

Further, in the second embodiment, the priority for changing the amount of resources is determined for each type of resource according to the usage rate of the resource. Then, the resource amount change is preferentially determined from the resource of the higher priority type, and the response time condition is verified using the improvement machine. As a result, the response time when the resource is changed can be verified in order from the type of resource that is likely to cause performance deterioration, and the verification process can be performed efficiently.

[Other embodiments]
In the second embodiment, the load balancer that relays the request and the response measures the request amount and the response time, but the request amount and the response time can also be measured by another method. For example, the virtual machine monitoring device 100 can capture packets communicated via the business network 41, analyze the captured packets, and measure the request amount and response time for a specific virtual machine.

Although the embodiment has been illustrated above, the configuration of each part shown in the embodiment can be replaced with another having the same function. Further, any other components or processes may be added. Further, any two or more configurations (features) of the above-described embodiments may be combined.

1a, 1b Terminal device 2 Network 3 Server 3a-1, 3a-2, ... CPU
3b-1, 3b-2, ... Memory 3c Virtual machine 10 Information processing device 11 Storage unit 12 Processing unit

Claims (6)

Monitor the amount of requests sent per unit period to the first virtual machine running on the server ,
The amount of the request in the most recent first unit period is smaller than the amount of the request in the second unit period immediately before the first unit period, and the degree of decrease is the first threshold value. If it exceeds, the decision is made to reduce the amount of hardware resources available to run the first virtual machine.
When it is decided to reduce the amount of the hardware resource available for executing the first virtual machine, the amount after the reduction of the hardware resource is determined, which is the same as the amount after the reduction to the server. The acquisition of the execution instruction of the second virtual machine in which the amount of the hardware resource can be used and the performance information of the second virtual machine from the server are acquired while reducing the amount after the reduction. Repeatedly while the performance information meets the specified conditions,
Reducing the amount of the hardware resources available to execute the first virtual machine to the reduced amount when the performance of the second virtual machine finally meets a predetermined condition . Processing unit instructing the server,
Information processing device with. The processing unit
If the amount of the request in the previous SL first unit period, the has increased than the amount of the request in the second unit period, the degree of increase exceeds a second threshold value, the second Decided to increase the amount of said hardware resources available to run one virtual machine,
The instructions to the server instruct the server to increase the amount of hardware resources available to execute the first virtual machine.
The information processing device according to claim 1. The processing unit
If it decides to increase the amount of the hardware resources available to run the first virtual machine, for each type of said hardware resource, the amount that allows uses to execute the first VM Based on resource management information that indicates the priority of changes, the minimum amount available to run the first virtual machine, and the maximum amount available to run the first virtual machine. Therefore, it was decided to increase the hardware resource of the highest priority type within a range not exceeding the maximum value.
If it is determined to reduce the amount of the hardware resource available to execute the first virtual machine, the highest priority type of the hardware resource is set to the minimum value based on the resource management information. Decide to reduce within the range not below,
The information processing device according to claim 1 or 2. The processing unit
The usage rate of the first virtual machine for each type of the hardware resource is acquired from the server, and the higher priority of the hardware resource of the type with the higher usage rate is set in the resource management information.
The information processing device according to claim 3. On the computer
Monitor the amount of requests sent per unit period to the first virtual machine running on the server,
The amount of the request in the most recent first unit period is smaller than the amount of the request in the second unit period immediately before the first unit period, and the degree of decrease is the first threshold value. If it exceeds, the decision is made to reduce the amount of hardware resources available to run the first virtual machine.
When it is decided to reduce the amount of the hardware resource available for executing the first virtual machine, the amount after the reduction of the hardware resource is determined, which is the same as the amount after the reduction to the server. The acquisition of the execution instruction of the second virtual machine in which the amount of the hardware resource can be used and the performance information of the second virtual machine from the server are acquired while reducing the amount after the reduction. Repeatedly while the performance information meets the specified conditions,
Reducing the amount of the hardware resources available to execute the first virtual machine to the reduced amount when the performance of the second virtual machine finally meets a predetermined condition . Instruct the server,
A virtual machine monitoring program that executes processing. The server running the first virtual machine and
A relay device forwards the request transmitted from the terminal apparatus to the first virtual machine, counting the amount of requests per unit time that is transferred to the first virtual machine,
The amount of requests per unit period transferred to the first virtual machine is acquired from the relay device, and the amount of the requests in the most recent first unit period is the second immediately before the first unit period. If the amount of the request is less than the amount of the request in the unit period of, and the degree of the decrease exceeds the first threshold value, the amount of hardware resources available for executing the first virtual machine is reduced. If it is determined that the amount of the hardware resource available for execution of the first virtual machine is to be reduced, then the amount of the hardware resource after reduction is determined, the reduction to the server. The execution instruction of the second virtual machine that can use the same amount of the hardware resource as the later amount, and the acquisition of the performance information of the second virtual machine from the server are reduced by reducing the reduced amount. However, while the acquired performance information satisfies a predetermined condition, the amount of the hardware resource that can be used for executing the first virtual machine is repeatedly determined, and the performance of the second virtual machine is a predetermined condition. The information processing device instructing the server to reduce the amount to the amount after the reduction when is finally satisfied, and
Information processing system with.

Images