JP7104327B2 - Information processing device, virtual machine management program and virtual machine management method - Google Patents

Description

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

In the field of information processing, virtualization technology that operates multiple virtual computers (sometimes called virtual machines or logical hosts) on physical computers (sometimes called physical machines or physical hosts) It's being used. Software such as an OS (Operating System) can be executed on each virtual machine. Physical machines that utilize virtualization technology run software to manage multiple virtual machines. For example, software called a hypervisor may allocate the processing power of a CPU (Central Processing Unit) and the storage area of a RAM (Random Access Memory) to a plurality of virtual machines as arithmetic resources.

By the way, a service is used in which the right to use a resource of a computer system is lent to a user via a network, and the user is charged according to the usage status such as the usage amount and usage time of the resource. Such a service usage form is sometimes called Infrastructure as a Service (IAAS). In a computer system, virtualization technology is used from the viewpoint of separation of processing between users and ease of billing management, and one or more virtual machines may be executed for each user. Further, in such a service, a method of suppressing an excessive allocation of resources to a user is considered.

For example, a management device has been proposed in which one or more virtual machines are arranged in each of the first and second virtualization systems, and a certain process is distributed to each virtual machine. The proposed management device is the first when the load of the virtual machine in the first virtual system satisfies the first load condition and the load of the virtual machine in the second virtual system satisfies the second load condition. Allows the addition of virtual machines to the 2 virtualization systems.

In addition, by changing the dummy cycle added to the instruction execution processing according to the usage status of the user's virtual computer and the usage conditions of the user's computer, the CPU performance of the virtual computer can be changed without any special operation. There are also suggestions for methods.

It also has first and second databases running on virtual machines with different performances, and replaces the virtual machines used by the system in response to external instructions to scale up and down the performance of the system. There are also suggestions for methods.

Furthermore, there is also a proposal of an autoscale method in which the virtual machines of the business system before the update are deleted according to the processing load of the virtual machines of the entire business system before and after the update.

JP-A-2015-153402 Japanese Unexamined Patent Publication No. 2006-65430 Japanese Unexamined Patent Publication No. 2012-215937 Japanese Unexamined Patent Publication No. 2017-11164

When the frequency of use of a virtual machine by a user increases, it is conceivable to start a virtual machine having a higher performance than the virtual machine and switch to operation using the high-performance virtual machine. At this time, if the image data of the virtual machine stored in the auxiliary storage device is loaded into the memory of the physical machine and the virtual machine is started, it takes time to start the virtual machine.

Therefore, in preparation for the increasing frequency of use of virtual machines, it is conceivable to operate relatively high-performance virtual machines in advance and provide them to users. However, when the image data of the virtual machine is loaded into the memory and operated, the user is charged a usage fee for resources such as a CPU and RAM. The higher the performance of the virtual machine used by the user (the larger the resource allocation amount), the larger the billing amount. For this reason, the user may be charged by the relatively high-performance virtual machine even when the virtual machine is used relatively infrequently, and the user may be charged excessively. be.

In one aspect, it is an object of the present invention to provide an information processing device, a virtual machine management program, and a virtual machine management method that enable rapid switching of virtual machines while reducing resource usage.

In one aspect, an information processing device is provided. This information processing device has a storage unit and a processing unit. The storage unit stores information indicating the correspondence between a plurality of virtual machines having different performances. The processing unit sends an instruction for arranging a plurality of image data corresponding to a plurality of virtual machines to a storage device to the physical machine, monitors the load of the first virtual machine running on the physical machine, and the load is the first. When the threshold value of is exceeded, the image data of the second virtual machine having higher performance than that of the first virtual machine is loaded from the storage device into the memory based on the information stored in the storage unit, and the second virtual machine is loaded. Sends an instruction to the physical machine to prepare, and when the load exceeds the second threshold, which is higher than the first threshold, it sends an instruction to switch the first virtual machine to the second virtual machine. ..

Also, in one aspect, a virtual machine management program is provided.
Also, in one aspect, a virtual machine management method is provided.

On the one hand, it reduces resource usage and allows for rapid switching of virtual machines.

It is a figure which shows the information processing system of 1st Embodiment. It is a figure which shows the example of the information processing system of the 2nd Embodiment. It is a block diagram which shows the hardware example of the management server. It is a figure which shows the functional example of the management server. It is a figure which shows the functional example of a VM host. It is a figure which shows the example of VM deployment of a VM host. It is a figure which shows the example of the condition table. It is a figure which shows the example of the management table. It is a figure which shows the preparation example of downsizing. It is a figure which shows the example of stopping downsizing and the example of execution. It is a figure which shows the preparation example of upsizing. It is a figure which shows the example of stopping the upsizing and the example of execution. It is a flowchart which shows the processing example of the management server. It is a flowchart which shows the example of VM switching control (downsize). It is a flowchart which shows the example of downsizing post-processing. It is a figure which shows the specific example of downsizing post-processing. It is a flowchart which shows the example of VM switching control (upsize). It is a flowchart which shows the example of the upsizing post-processing. It is a figure which shows the specific example of the upsizing post-processing. It is a flowchart which shows the example of unnecessary VM deletion.

Hereinafter, the present embodiment will be described with reference to the drawings.
[First Embodiment]
The first embodiment will be described.

FIG. 1 is a diagram showing an information processing system according to the first embodiment.
The information processing system 1 includes an information processing device 10, a physical machine 20, and a storage device 30. The information processing device 10, the physical machine 20, and the storage device 30 are connected to the network 40. The information processing device 10 manages the operation of the virtual machine in the physical machine 20.

The physical machine 20 executes a virtual machine using the hardware 21 of the physical machine 20. The physical machine 20 executes the hypervisor 22. The hypervisor 22 allocates the resources of the hardware 21 included in the physical machine 20 to the virtual machine. The hardware 21 includes a memory 21a and a processor 21b. For example, the hypervisor 22 allocates the storage resource of the memory 21a and the arithmetic resource of the processor 21b to the virtual machine 23, and operates the virtual machine 23 on the physical machine 20. For example, the virtual machine 23 communicates with a client terminal (not shown) operated by the user via the network 40, and supports a user's work by a predetermined software function (for example, a Web server function). I will provide a. For example, the user is charged a usage fee according to the usage status of the virtual machine 23 (resource amount, usage time, etc. of the virtual machine 23).

The storage device 30 is a storage that stores data used for processing of the information processing device 10 and the physical machine 20. The storage device 30 may be built in the physical machine 20.
The information processing device 10 has a storage unit 11 and a processing unit 12. The storage unit 11 may be a volatile storage device such as a RAM (Random Access Memory) or a non-volatile storage device such as an HDD or a flash memory. The processing unit 12 may include a CPU, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), and the like. The processing unit 12 may be a processor that executes a program. A "processor" may also include a set of multiple processors (multiprocessors).

The storage unit 11 stores information indicating the correspondence between a plurality of virtual machines having different performances. For example, the storage unit 11 stores information (for example, storage of each data) indicating to the virtual machine 23 the data of the virtual machine having higher performance than the virtual machine 23 and the data of the virtual machine having lower performance than the virtual machine 23. Information indicating the position) is stored.

The processing unit 12 transmits to the physical machine 20 an instruction for arranging a plurality of image data corresponding to a plurality of virtual machines having different performances in the storage device 30. Here, the image data is data in a state where the virtual machine is shut down and saved in the storage device 30, and in addition to the information of the software included in the virtual machine, the information related to the amount of resources allocated to the virtual machine is displayed. Can include. The image data may be called a saved image or saved data of a virtual machine.

Here, in the following, "virtual machine" may be abbreviated as VM (Virtual Machine). Further, "virtual machine image data" may be abbreviated as VM data. The physical machine 20 stores a plurality of VM data of a plurality of virtual machines having different performances in the storage device 30 in response to an instruction from the processing unit 12.

For example, the storage device 30 stores VM data 31, 32, 33. VM data 31, 32, 33 are created by specifying different performances from one virtual machine image including software such as a certain OS or application. The performance of the virtual machine is determined by the amount of resources allocated to the hardware 21 such as the memory 21a and the processor 21b. For example, the larger the allocation amount of the memory 21a and the allocation amount of the processor 21b, the higher the performance. Further, the smaller the allocated amount of the memory 21a and the allocated amount of the processor 21b, the lower the performance.

For example, the VM data 31 is the data of the highest performance virtual machine among the VM data 31, 32, 33. The VM data 32 is data of a virtual machine having intermediate performance among the VM data 31, 32, 33. The VM data 33 is the data of the virtual machine having the lowest performance among the VM data 31, 32, 33. Intermediate performance is referred to as "M (Medium)". Higher performance than intermediate performance is referred to as "L (Large)". The fact that the performance is lower than the intermediate performance is referred to as "S (Small)".

For example, when the processing unit 12 receives the user's instruction to deploy the virtual machine 23 having the performance "M", the processing unit 12 inputs the VM data 31, 32, 33 of the performances "L", "M", and "S" to the physical machine. 20 stores the data in the storage device 30. Information on the amount of resources allocated in the performance "L" with respect to the performance "M" and the amount of resources allocated in the performance "S" with respect to the performance "M" is stored in advance in the storage unit 11. The processing unit 12 loads the VM data 32 into the memory 21a by the physical machine 20, and executes the virtual machine 23 having the performance “M” on the physical machine 20. The processing unit 12 receives from the physical machine 20 a notification of the result of the arrangement of the VM data 31, 32, 33 in the storage device 30 and the completion of starting the virtual machine 23, and information indicating the correspondence relationship between a plurality of virtual machines having different performances. Is generated and stored in the storage unit 11.

The processing unit 12 monitors the load acquired from the first virtual machine running on the physical machine 20. When the load of the first virtual machine exceeds the first threshold value, the processing unit 12 loads the image data of the second virtual machine, which has higher performance than that of the first virtual machine, from the storage device 30 into the memory 21a. Then, an instruction to prepare the second virtual machine is transmitted to the physical machine 20.

For example, the processing unit 12 acquires a load from the virtual machine 23 at predetermined time intervals. The load is, for example, the processor utilization by the virtual machine 23. The load may be expressed in other indicators, such as memory usage, in place of or in conjunction with processor usage.

When the load of the virtual machine 23 exceeds the first threshold value, the processing unit 12 loads the VM data 31 from the storage device 30 into the memory 21a, and prepares a virtual machine having higher performance than the virtual machine 23. The instruction is transmitted to the physical machine 20. The ready state is to put the corresponding virtual machine into a bootable state (paused state). Based on the instruction, the physical machine 20 loads the VM data 31 into the memory 21a so that the virtual machine (virtual machine 24) corresponding to the VM data 31 can be started.

The processing unit 12 continues to monitor the load of the first virtual machine. When the load of the first virtual machine exceeds the second threshold value higher than the first threshold value, the processing unit 12 transmits an instruction to switch the first virtual machine to the second virtual machine to the physical machine 20.

For example, when the load of the virtual machine 23 exceeds the second threshold value, the processing unit 12 transmits an instruction to switch the virtual machine 23 to the virtual machine 24 to the physical machine 20. Based on the instruction, the physical machine 20 starts the virtual machine 24 by the VM data 31 loaded in the memory 21a. The physical machine 20 switches the virtual machine 23 to the virtual machine 24. Specifically, the physical machine 20 starts the virtual machine 24 and switches the provider of a predetermined service such as the Web server function to the user from the virtual machine 23 to the virtual machine 24.

The physical machine 20 has, for example, a load balancer. By changing the request from the client terminal assigned to the virtual machine 23 so as to allocate it to the virtual machine 24, the load balancing unit can switch the virtual machines without stopping the service provision to the user.

According to the information processing system 1, a plurality of image data of a plurality of virtual machines having different performances are stored in the storage device 30. The load of the first virtual machine running in the information processing system 1 (physical machine 20) is monitored. When the load of the first virtual machine exceeds the first threshold value, the image data of the second virtual machine having higher performance than that of the first virtual machine is loaded from the storage device 30 into the memory 21a, and the second virtual machine is loaded. The machine is ready. When the load of the first virtual machine exceeds the second threshold value higher than the first threshold value, the first virtual machine is switched to the second virtual machine.

This reduces resource usage and enables rapid switching of virtual machines.
Specifically, the information processing device 10 operates the virtual machine 23 in a normal state (when the virtual machine 23 has a relatively low load), and when the load of the virtual machine 23 becomes higher than the first threshold value, the virtual machine 23 is virtual. The machine 24 is loaded into the memory 21a to prepare it. Therefore, in the normal state, it is not necessary to load the virtual machine 24 into the memory 21a, and the amount of the memory 21a used can be suppressed. As a result, it is possible to suppress the charge associated with the use of the memory 21a for the user.

Further, while the virtual machine 24 is in the ready state, the virtual machine 24 is not started and the processor 21b is not used. Therefore, in the prepared state of the virtual machine 24, the charge associated with the use of the processor 21b to the user can be suppressed.

Moreover, since the VM data 31 has already been loaded into the memory 21a, when the load of the virtual machine 23 exceeds the second threshold value, the physical machine 20 quickly starts the virtual machine 24, and the virtual machine It is possible to switch from 23 to the virtual machine 24.

In this way, resource usage can be reduced and virtual machines can be switched quickly. In addition, by reducing the amount of resources used, it is possible to reduce the amount charged to the user. In addition, it is possible to switch to a high-performance virtual machine at an appropriate timing when it is estimated that the load is increasing, and it is possible to suppress the delay in the processing performed by the virtual machine.

After switching, the physical machine 20 may stop the virtual machine 23 and, for example, save the VM data 32 from the memory 21a to the storage device 30 after a predetermined time has elapsed. As a result, the amount of resources used by the virtual machine 23 can be further reduced.

Further, the processing unit 12 transmits an instruction to prepare the virtual machine 24 to the physical machine 20 when the time when the load of the virtual machine 23 continuously exceeds the first threshold reaches the first time. You may. Then, when the time when the load of the virtual machine 23 continuously exceeds the second threshold value reaches the second time, an instruction to switch the virtual machine 23 to the virtual machine 24 may be transmitted to the physical machine 20. As a result, it is possible to appropriately detect that the load on the virtual machine 23 tends to increase and prepare to start the virtual machine 24. Further, it is possible to appropriately detect that the virtual machine 23 is in a high load state and switch from the virtual machine 23 to the virtual machine 24.

Further, after the processing unit 12 transmits a command to prepare the virtual machine 24, if a predetermined time elapses without switching from the virtual machine 23 to the virtual machine 24, the image data of the virtual machine 24 is stored in the memory 21a. An instruction to delete from may be sent to the physical machine 20. As a result, when it is determined that the virtual machine 23 does not become a relatively high load state and switching to the virtual machine 24 is unnecessary, the usage amount of the memory 21a can be suppressed.

Further, when the load of the virtual machine 23 drops below the third threshold value lower than the first threshold value, the processing unit 12 loads the VM data 33 into the memory 21a to generate a virtual machine having lower performance than the virtual machine 23. You may instruct the physical machine 20 to be ready. Then, when the load of the virtual machine 23 drops below the fourth threshold value, which is lower than the third threshold value, the processing unit 12 may instruct the physical machine 20 to switch the virtual machine 23 to a low-performance virtual machine. .. In response to the instruction, the physical machine 20 starts a low-performance virtual machine based on the VM data 33 loaded in the memory 21a, and switches from the virtual machine 23 to the low-performance virtual machine. This makes it possible to switch virtual machines quickly while further reducing resource usage. In addition, by reducing the amount of resources used, the amount charged to the user can be further reduced.

Further, when a predetermined period elapses without changing to the ready state after another virtual machine (for example, the virtual machine 23) having a performance different from that of the virtual machine 23 is saved in the storage device 30, the processing unit 12 causes the other virtual machine. An instruction to delete the image data of the virtual machine of the above from the storage device 30 may be transmitted to the physical machine 20. As a result, when it is estimated that the load state of the virtual machine 23 is relatively stable, the amount of the storage device 30 used can be suppressed. By reducing the amount of storage device 30 used, the amount charged to the user can be further reduced.

[Second Embodiment]
Next, a second embodiment will be described.
FIG. 2 is a diagram showing an example of the information processing system of the second embodiment.

The information processing system of the second embodiment has a management server 100 and VM hosts 200, 300, .... The information processing system of the second embodiment makes the virtual machine executed by the VM hosts 200, 300, ... Available to the client 50. The management server 100 and the VM hosts 200, 300, ... Are connected to the network 60. The network 60 is connected to the network 70. The client 50 is connected to the network 70. For example, the network 60 is a LAN (Local Area Network). For example, the network 70 is a WAN (Wide Area Network), the Internet, or the like.

The management server 100 is a server computer that manages the operation of virtual machines operating on VM hosts 200, 300, ....
The VM hosts 200, 300, ... Are server computers that execute virtual machines. Each of the VM hosts 200, 300, ... Executes a hypervisor, and the functions of the hypervisor allocate the CPU and RAM resources of each of the VM hosts 200, 300, ... To the virtual machine.

The client 50 is a client computer operated by a user. The client 50 functions as, for example, a Web browser, and accesses the Web service provided by the virtual machine on the VM hosts 200, 300, ... By the Web browser.

The information processing system of the second embodiment provides IaaS. That is, the information processing system of the second embodiment lends the right to use the resources of the computer system to the user via the network, and charges the user according to the usage status such as the usage amount and usage time of the resource (metered amount). Billing).

FIG. 3 is a block diagram showing a hardware example of the management server.
The management server 100 includes a CPU 101, a RAM 102, an HDD 103, an image signal processing unit 104, an input signal processing unit 105, a medium reader 106, and a NIC 107. Each of these hardware is connected to the bus of the management server 100. The CPU 101 corresponds to the processing unit 12 of the first embodiment. The RAM 102 or the HDD 103 corresponds to the storage unit 11 of the first embodiment.

The CPU 101 is a processor that executes a program instruction. The CPU 101 loads at least a part of the programs and data stored in the HDD 103 into the RAM 102 and executes the program. The CPU 101 may include a plurality of processor cores. Further, the management server 100 may have a plurality of processors. The processes described below may be performed in parallel using multiple processors or processor cores. Also, a set of multiple processors may be referred to as a "multiprocessor" or simply a "processor".

The RAM 102 is a volatile semiconductor memory that temporarily stores a program executed by the CPU 101 and data used by the CPU 101 for calculation. The management server 100 may include a type of memory other than RAM, or may include a plurality of memories.

The HDD 103 is a non-volatile storage device that stores software programs such as an OS, middleware, and application software, and data. The management server 100 may be provided with other types of storage devices such as a flash memory and an SSD (Solid State Drive), or may be provided with a plurality of non-volatile storage devices.

The image signal processing unit 104 outputs an image to the display 111 connected to the management server 100 in accordance with a command from the CPU 101. As the display 111, any kind of display such as a CRT (Cathode Ray Tube) display, a liquid crystal display (LCD: Liquid Crystal Display), a plasma display, and an organic EL (OEL: Organic Electro-Luminescence) display can be used.

The input signal processing unit 105 acquires an input signal from the input device 112 connected to the management server 100 and outputs the input signal to the CPU 101. As the input device 112, a pointing device such as a mouse, a touch panel, a touch pad, or a trackball, a keyboard, a remote controller, a button switch, or the like can be used. Further, a plurality of types of input devices may be connected to the management server 100.

The medium reader 106 is a reading device that reads programs and data recorded on the recording medium 113. As the recording medium 113, for example, a magnetic disk, an optical disk, a magneto-optical disk (MO), a semiconductor memory, or the like can be used. The magnetic disk includes a flexible disk (FD) and an HDD. Optical discs include CDs (Compact Discs) and DVDs (Digital Versatile Discs).

The medium reader 106 copies, for example, a program or data read from the recording medium 113 to another recording medium such as the RAM 102 or the HDD 103. The read program is executed by, for example, the CPU 101. The recording medium 113 may be a portable recording medium and may be used for distribution of programs and data. Further, the recording medium 113 and the HDD 103 may be referred to as a computer-readable recording medium.

The NIC 107 is an interface that is connected to the network 60 and communicates with another computer via the network 60. The NIC 107 is connected to a communication device such as a switch or a router by a cable.

The VM hosts 200, 300, ... And the client 50 are also realized by the same hardware as the management server 100.
FIG. 4 is a diagram showing a functional example of the management server.

The management server 100 includes a storage unit 120, a VM deployment unit 130, a load calculation unit 140, a VM switching processing unit 150, a VM state management unit 160, and a VM deletion unit 170. The storage unit 120 is realized by the storage area of the RAM 102 or the HDD 103. The VM deployment unit 130, the load calculation unit 140, the VM switching processing unit 150, the VM state management unit 160, and the VM deletion unit 170 are realized, for example, by the CPU 101 executing the program stored in the RAM 102.

The storage unit 120 stores a management table for managing the operating status of the virtual machine.
The VM deployment unit 130 instructs the VM hosts 200, 300, ... To deploy the virtual machine. When deploying a virtual machine with a performance specified by a user, the VM deployment unit 130 saves a virtual machine having a performance higher than the specified performance and a virtual machine having a performance lower than the specified performance in a saved state. The VM deployment command to be deployed is transmitted to the VM hosts 200, 300, ....

Here, the performance of the virtual machine is determined, for example, by the amount of CPU or RAM allocated to the virtual machine. The performance of a virtual machine is sometimes referred to as flavor. For example, the VM hosts 200, 300, ... Can deploy a plurality of virtual machines having different performances from one virtual machine image.

Information indicating the amount of resources allocated to a high-performance virtual machine by one stage and information indicating the amount of resources allocated to a low-performance virtual machine by one stage with respect to the amount of resources allocated to the virtual machine instructed by the user. Is stored in the storage unit 120 in advance. For example, the allocated resource amount is given in advance as information on the difference with respect to the resource amount specified by the user. For example, in the difference information, the difference in the CPU allocation amount is represented by the number of CPUs, the number of clocks, and the like, and the difference in the RAM allocation amount is represented by the RAM size and the like. Based on the information stored in the storage unit 120, the VM deployment unit 130 is a virtual machine having one step higher performance than the performance of the virtual machine and one step higher than the performance of the virtual machine for the virtual machine instructed to be deployed by the user. Only determine the amount of resources allocated to low-performance virtual machines.

The load calculation unit 140 periodically acquires load information indicating the load of the virtual machine from the virtual machines operating on the VM hosts 200, 300, ..., And calculates the load of the virtual machine based on the load information. The load of the virtual machine is, for example, the CPU usage rate of the virtual machine with respect to the CPU assigned to the virtual machine. However, the load on the virtual machine is in addition to the CPU usage rate, or instead of the CPU usage rate, the RAM usage rate by the virtual machine, the number of people logged in to the virtual machine, the number of disk accesses by the virtual machine, and the communication traffic by the virtual machine. It may be evaluated by the amount or the like. For example, when evaluating the load of a virtual machine using a plurality of indexes such as CPU usage rate and RAM usage rate, the load calculation unit 140 normalizes each index and obtains the average of each normalized index. It is conceivable to calculate the load on the virtual machine.

The VM switching processing unit 150 issues a VM switching instruction for switching from a virtual machine having the current performance to a virtual machine having a different performance to the VM hosts 200, 300, ... Based on the load of the virtual machine acquired by the load calculation unit 140. Send.

The VM state management unit 160 acquires the state of the virtual machine from the VM hosts 200, 300, ..., And records it in the management table stored in the storage unit 120.
The VM deletion unit 170 uses the VM switching processing unit 150 to specify a virtual machine that has not been switched for a predetermined period of time or longer, and issues a VM deletion instruction for deleting the switching destination candidate virtual machine for the specified virtual machine to the VM host. Send to 200, 300, ....

FIG. 5 is a diagram showing a functional example of the VM host.
The VM host 200 has a VM deployment execution unit 210, a load information notification unit 220, a VM switching execution unit 230, a VM status notification unit 240, and a VM deletion execution unit 250. In the VM deployment execution unit 210, the load information notification unit 220, the VM switching execution unit 230, the VM status notification unit 240, and the VM deletion execution unit 250, for example, the processor of the VM host 200 stores a program stored in the RAM of the VM host 200. It is realized by executing it.

The VM deployment execution unit 210 receives a VM deployment command from the VM deployment unit 130, and deploys the virtual machine on the VM host 200. Virtual machine deployment is performed by specifying the amount of resources to be allocated (that is, performance) for a virtual machine image. As described above, when deploying a virtual machine having a certain performance, the VM deployment execution unit 210 deploys a virtual machine having a higher performance than the virtual machine and a virtual machine having a lower performance. The high-performance virtual machine and the low-performance virtual machine are, for example, in a state of being saved in an auxiliary storage device such as an HDD or SSD provided in the VM host 200, or an external storage connected to the network 60 (save state). It is said that. The CPU and RAM of the VM host 200 are not used for the virtual machine in the saved state.

The load information notification unit 220 periodically acquires load information indicating the load of the virtual machine running on the VM host 200, and transmits it to the load calculation unit 140.
The VM switching execution unit 230 receives a VM switching command from the VM switching processing unit 150, and executes VM switching based on the VM switching command. The function of the VM switching execution unit 230 is realized by, for example, the load balancing function of the VM host 200.

The VM status notification unit 240 acquires the status of the virtual machine managed by the VM host 200 in response to the VM switching by the VM switching execution unit 230, and transmits the acquired virtual machine status to the VM status management unit 160.

The VM deletion execution unit 250 receives a VM deletion command from the VM deletion unit 170, and deletes the virtual machine managed by the VM host 200 based on the VM deletion command.
Other VM hosts including the VM host 300 also have the same functions as the VM host 200.

FIG. 6 is a diagram showing an example of VM deployment of a VM host.
The VM host 200 has hardware 200h. The hardware 200h includes a CPU 201, a RAM 202, and an HDD 203. The VM host 200 may have other types of storage devices such as SSDs in place of or in combination with HDD 203. The hardware 200h also includes other hardware such as a NIC used for operating a virtual machine (not shown).

The VM host 200 has a hypervisor 260. The hypervisor 260 allocates the resources of the CPU 201, RAM 202, and HDD 203 to the virtual machine, and operates the virtual machine on the VM host 200.

The VM host 200 has, for example, virtual machines 261,262,263, ..., 270 and a load balancer 280. The virtual machines 261,262,263 provide, for example, a Web server function. The virtual machine 270 provides, for example, an AP server function and executes processing according to a processing request received from any of the virtual machines 261, 262, and 263. The load balancing unit 280 has a function of distributing requests to a plurality of virtual machines. Here, the load balancing unit 280 is used for the VM switching process. The load balancing unit 280 may be realized by a virtual machine.

Virtual machines 261,262,263 are instances of virtual machines launched from a common virtual machine image. However, the virtual machines 261,262,263 have different performances (flavors) from each other. The virtual machine 261 has the highest performance among the virtual machines 261,262,263. The virtual machine 262 has an intermediate performance among the virtual machines 261,262,263. The virtual machine 263 has the lowest performance among the virtual machines 261,262,263.

The distinguished name of the virtual machine 261 is "VM1-L". The distinguished name of the virtual machine 262 is "VM1-M". The distinguished name of the virtual machine 263 is "VM1-S". The distinguished name of the virtual machine 270 is "VM2".

During normal operation, a virtual machine 262 with intermediate performance runs and is used by users. The operating state is described as "Active". During normal operation, the virtual machines 261,263 are in a saved state saved in HDD 203 or external storage. The evacuation state is described as "Shelved". In the saved state, the corresponding virtual machine (for example, virtual machines 261,263) is shut down, the VM data of the virtual machine is stored in the HDD 203 or the external storage, and the VM data is deleted from the RAM of the VM host 200. To. The saved state may be, for example, a state in which an instance of the virtual machine is saved in the auxiliary storage device by the save command in Openstack (registered trademark). When the virtual machine in the Shelved state is restarted, the VM data is expanded in the RAM and then started, so the time required for starting is relatively long.

As the state of the virtual machine, there is also a paused state in which the image data of the virtual machine is loaded in the RAM of the VM host 200, but the virtual machine is not started. The paused state is marked as "Stopped". In the paused state, the corresponding virtual machine is not started and the CPU of the VM host 200 is not used. The paused state may be, for example, a state in which the state of the virtual machine is stored in the main storage device of the VM host and the virtual machine is suspended by the pause command in Openstack. In the Stopped state, the VM data of the virtual machine is expanded on the RAM, but the operation is stopped. When restarting a virtual machine in the Stopped state, the virtual machine can be started by allocating a CPU, so the time required for starting is relatively short (it can be restarted faster than restarting from the Shelved state).

FIG. 7 is a diagram showing an example of a condition table.
The condition table 151 shows the conditions for the VM switching process. The VM switching processing unit 150 determines whether to switch the virtual machine 262 to the virtual machine 261 or the virtual machine 263 based on the load of the virtual machine 262. For example, the VM switching processing unit 150 makes the determination based on the conditions shown in the condition table 151. The condition table 151 includes items of classification, condition name, condition content, and processing category.

In the classification item, information indicating whether the corresponding condition is a condition for downsizing the virtual machine or a condition for upsizing the virtual machine is registered. “Downsize” indicates that it is a condition for downsizing (that is, reducing the performance) of the virtual machine. "Upsize" indicates that the condition for upsizing (that is, improving the performance) of the virtual machine. The name of the condition is registered in the item of the condition name. The specific content of the condition is registered in the item of the condition content. In the item of processing category, the processing category corresponding to the corresponding condition is registered. The processing categories include "switching preparation" and "switching execution". "Preparation for switching" indicates that the VM data in the saved state is loaded into the RAM of the VM host and put into the paused state (Stopped). "Switching execution" indicates that a CPU is assigned to a virtual machine in a suspended state and started, and the service provider is switched from the existing virtual machine to the started virtual machine.

According to the condition table 151, there are conditions D1, D2, U1 and U2.
"Condition D1" indicates that when the "CPU usage rate of less than 20% continues for 10 minutes or more" of the virtual machine, the command of "preparation for switching" of "downsize" is transmitted to the VM host.

"Condition D2" indicates that when the "CPU usage rate of less than 10% continues for 30 minutes or more" of the virtual machine, the instruction of "switching execution" of "downsize" is transmitted to the VM host. For example, the condition content of the "condition D2" is determined based on the time immediately after the "condition D1" is satisfied and the "switching preparation" is performed.

"Condition U1" indicates that when the "CPU usage rate exceeding 50% continues for 10 minutes or more" of the virtual machine, the command of "upsize" "preparation for switching" is transmitted to the VM host. However, as will be described later, the threshold value of the CPU usage rate of "CPU usage rate 50%" under the condition U1 is an initial value and is variable according to the load status of the virtual machine.

"Condition U2" indicates that when the virtual machine "exceeds 90% of the CPU usage rate continues for 30 minutes or more", the instruction of "switching execution" of "upsize" is transmitted to the VM host. For example, the condition content of the "condition U2" is determined based on the time immediately after the "condition U1" is satisfied and the "switching preparation" is performed.

FIG. 8 is a diagram showing an example of a management table.
The management table 121 is stored in the storage unit 120. The management table 121 includes items of VM name, state, type, condition D1 satisfaction, condition D2 satisfaction, condition U1-CPU usage threshold, condition U1 satisfaction, condition U2 satisfaction, performance, and state update date and time.

The identification name of the virtual machine is registered in the item of VM name. The status of the corresponding virtual machine is registered in the status item. The type of virtual machine is registered in the type item. There are two types of virtual machines. The first type is "running VM", which indicates that it is a running virtual machine. The second type is a "switching destination VM" indicating that the virtual machine is a candidate virtual machine to be switched to for a running virtual machine. For example, a running (switching source) virtual machine and a switching destination candidate virtual machine are associated with each other by a VM name. In one example, a virtual machine that shares the left part of the hyphen symbol "-" (for example, the part of "VM1" such as "VM1-M") included in the VM name is in the relationship between the switching source and switching destination candidates. be. In the following description, an operating (switching source) virtual machine may be referred to as a "switching source VM". In addition, a virtual machine that is a candidate for a switching destination may be referred to as a "switching destination VM".

In the item of condition D1 satisfaction, a flag indicating whether or not condition D1 is satisfied is registered. In the item of condition D2 satisfaction, a flag indicating whether or not condition D2 is satisfied is registered. In the item of the condition U1-CPU usage threshold, the current setting value of the CPU usage threshold in the condition U1 for the corresponding virtual machine is registered. In the item of condition U1 satisfaction, a flag indicating whether or not condition U1 is satisfied is registered. In the item of condition U2 satisfaction, a flag indicating whether or not condition U2 is satisfied is registered. Information indicating the performance of the corresponding virtual machine is registered in the performance item. In the example of FIG. 8, a predetermined performance is represented by “M”, a performance lower than the performance indicated by the performance “M” is represented by “S”, and a performance higher than the performance indicated by the performance “M” is represented by “L”. ". In the status update date and time item, the last date and time when the status was updated is registered for the corresponding virtual machine.

For virtual machines other than the state "Active" (type "Operating VM"), the items of condition D1 satisfaction, condition D2 satisfaction, condition U1-CPU usage threshold, condition U1 satisfaction, and condition U2 satisfaction are "Null". ". The initial value of the condition D1 satisfaction, condition D2 satisfaction, condition U1 satisfaction, and condition U2 satisfaction of the virtual machine in the state "Active" is "False". The initial value of the item of the condition U1-CPU usage threshold of the virtual machine in the state "Active" is "50%".

For example, in the management table 121, the VM name is "VM1-M", the state is "Active", the type is "Operating VM", the condition D1 is satisfied "False", the condition D2 is satisfied "False", and the condition U1- Records that the CPU usage threshold is "50%", the condition U1 is satisfied is "True", the condition U2 is satisfied is "False", the performance is "M", and the status update date and time is "2018-06-02 14:30" are registered. Has been done. This record indicates that the virtual machine 262 is in the active state, the performance is "M", and the last state update date and time is 14:30 on June 2, 2018. Further, for the virtual machine 262, the conditions D1 and D2 are not satisfied, the CPU usage threshold of the condition U1 is 50%, the condition U1 is satisfied, and the condition U2 is not satisfied. show.

Further, in the management table 121, the VM name is "VM1-S", the status is "Shelved", the type is "switching destination VM", the performance is "S", and the status update date and time is "2018-05-31 15:00". The record is registered. This record shows that the virtual machine 263 is in the saved state (Shelved), is a candidate for switching destination of the virtual machine 262, the performance is "S", and the final state update date is 15:00 on May 31, 2018. Indicates that it is a minute.

Further, in the management table 121, the VM name is "VM1-L", the status is "Shelved", the type is "switching destination VM", the performance is "L", and the status update date and time is "2018-06-01 15:30". The record is registered. In this record, the virtual machine 261 is in the saved state (Shelved), it is a candidate for the switching destination of the virtual machine 262, the performance is "L", and the final state update date is 15:30 on June 1, 2018. Indicates that it is a minute.

In the management table 121, an example in which the virtual machine 262 and the virtual machines 261,263, which are candidates for switching the virtual machine 262, are registered is shown, but the same as in the management table 121 when other virtual machines exist. Record is registered.

Further, the management server 100 holds information indicating the VM host on which the virtual machine of the VM name is deployed with respect to the VM name, and the virtual machine of the VM name is deployed based on the VM name. You can identify the VM host you are using. Further, the management server 100 holds a path indicating the storage location of the VM data corresponding to the VM name with respect to the VM name, and sets the VM in the ready state (paused state) based on the VM name. Data may be specified for the VM host.

Next, an example of switching the virtual machine by the VM host 200 based on the instruction of the management server 100 will be described. Although the description will be focused on the VM host 200, the same control is applied to other VM hosts. First, a switching example for downsizing a virtual machine will be described.

FIG. 9 is a diagram showing an example of preparation for downsizing.
Virtual machines 261,262,263 are deployed in the VM host 200. The virtual machine 262 is in an active state. The virtual machines 261,263 are in the evacuated state (Shelved).

The VM switching processing unit 150 periodically monitors the load of the virtual machine 262 calculated by the load calculation unit 140.
The VM switching processing unit 150 detects that the condition D1 is satisfied for the virtual machine 262. Then, the VM switching processing unit 150 transmits a VM state change command for changing the virtual machine 263 from the saved state (Shelved) to the paused state (Stopped) to the VM host 200.

When the VM switching execution unit 230 receives the VM state change instruction, the VM switching execution unit 230 loads the VM data of the virtual machine 263 into the RAM of the VM host 200 based on the VM state change instruction, and puts the virtual machine 263 in the suspended state (Stopped). Change to. The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 263 has been changed to the paused state. Based on the notification, the VM state management unit 160 registers in the management table 121 that the virtual machine 263 has been suspended. In addition, the status update date and time is also updated along with the status update (the same applies hereinafter). In this way, the preparation for downsizing (preparation for switching) is completed.

The VM switching processing unit 150 registers in the management table 121 that the condition D1 is satisfied for the virtual machine 262, subsequently monitors the load of the virtual machine 262, and makes a determination based on the condition D2.

FIG. 10 is a diagram showing an example of stopping downsizing and an example of executing downsizing.
FIG. 10A shows an example of discontinuing downsizing. In the VM switching processing unit 150, when the condition D1 is satisfied and the condition D2 is not satisfied for the virtual machine 262 (the CPU usage rate does not become less than 10% continuously for 30 minutes after the condition D1 is satisfied). If), decide to stop downsizing. In this case, the VM switching processing unit 150 transmits a VM state change command for changing the virtual machine 263 from the suspended state (Stopped) to the saved state (Shelved) to the VM host 200.

When the VM switching execution unit 230 receives the VM state change instruction, the VM data of the virtual machine 263 is saved in the HDD 203 or the external storage based on the VM state change instruction, and is deleted from the RAM of the VM host 200. As a result, the virtual machine 263 is changed from the suspended state (Stopped) to the saved state (Shelved). The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 263 has been changed to the saved state. Based on the notification, the VM state management unit 160 registers in the management table 121 that the virtual machine 263 has been in the saved state. In this way, downsizing is stopped.

FIG. 10B shows an execution example of downsizing. The VM switching processing unit 150 determines that the virtual machine 262 is downsized when it detects that the condition D2 is satisfied after the condition D1 is satisfied. In this case, the VM switching processing unit 150 changes the virtual machine 263 from the suspended state (Stopped) to the operating state (Active), and transmits a VM switching command for switching to the virtual machine 263 to the VM host 200.

When the VM switching execution unit 230 receives the VM switching instruction, the VM switching execution unit 230 changes the virtual machine 263 from the suspended state (Stopped) to the operating state (Active) based on the VM switching instruction. That is, the VM switching execution unit 230 allocates a CPU to the virtual machine 263 and starts the virtual machine 263 from the VM data loaded in the RAM of the VM host 200. Then, the VM switching execution unit 230 controls the load balancing unit 280 to distribute the requests distributed to the virtual machine 262 to the virtual machine 263.

Further, the VM switching execution unit 230 changes the virtual machine 262 from the operating state (Active) to the paused state (Stopped). The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 263 has been changed to the operating state and that the virtual machine 262 has been changed to the suspended state. Based on the notification, the VM state management unit 160 registers in the management table 121 that the virtual machine 263 has been in the operating state and that the virtual machine 262 has been in the suspended state.

Further, the VM deployment unit 130 transmits a VM deployment command for deploying the virtual machine 264 having a performance lower than that of the virtual machine 263 in the saved state and a VM deletion command for deleting the virtual machine 261 to the VM host 200.

When the VM deployment execution unit 210 receives the VM deployment instruction and the VM deletion instruction, it deletes the virtual machine 261 based on these instructions (deletes the VM data of the virtual machine 261 from the HDD 203 or the external storage). In addition, the VM deployment execution unit 210 deploys the virtual machine 264 in a saved state (Shelved). For example, the VM name of the virtual machine 264 is "VM1-SS".

The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 261 has been deleted and that the virtual machine 264 has been deployed in the saved state. Based on the notification, the VM state management unit 160 deletes the record of the virtual machine 261 from the management table 121 and adds the record related to the virtual machine 264 to the management table 121.

Next, a switching example for upsizing the virtual machine will be described.
FIG. 11 is a diagram showing an example of preparation for upsizing.
Virtual machines 261,262,263 are deployed in the VM host 200. The virtual machine 262 is in an active state. The virtual machines 261,263 are in the evacuated state (Shelved).

The VM switching processing unit 150 periodically monitors the load of the virtual machine 262 calculated by the load calculation unit 140.
The VM switching processing unit 150 detects that the condition U1 is satisfied for the virtual machine 262. Then, the VM switching processing unit 150 transmits a VM state change command for changing the virtual machine 261 from the saved state (Shelved) to the paused state (Stopped) to the VM host 200.

When the VM switching execution unit 230 receives the VM state change instruction, the VM switching execution unit 230 loads the VM data of the virtual machine 261 into the RAM of the VM host 200 based on the VM state change instruction, and puts the virtual machine 261 in a suspended state (Stopped). Change to. The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 261 has been changed to the paused state. Based on the notification, the VM state management unit 160 registers in the management table 121 that the virtual machine 261 has been suspended. In this way, the preparation for upsizing (preparation for switching) is completed.

The VM switching processing unit 150 registers in the management table 121 that the condition U1 is satisfied for the virtual machine 262, subsequently monitors the load of the virtual machine 262, and makes a determination based on the condition U2.

FIG. 12 is a diagram showing an example of stopping the upsizing and an example of executing the upsizing.
FIG. 12A shows an example of discontinuing the upsizing. In the VM switching processing unit 150, when the condition U1 is satisfied and the condition U2 is not satisfied for the virtual machine 262 (the CPU usage rate does not exceed 90% continuously for 30 minutes after the condition U1 is satisfied). If), decide to stop upsizing. In this case, the VM switching processing unit 150 transmits a VM state change command for changing the virtual machine 261 from the suspended state (Stopped) to the saved state (Shelved) to the VM host 200.

When the VM deployment execution unit 210 receives the VM state change command, the VM deployment execution unit 210 saves the VM data of the virtual machine 261 in the HDD 203 or the external storage based on the VM state change command, and deletes the VM data from the RAM of the VM host 200. As a result, the virtual machine 261 is changed from the suspended state (Stopped) to the saved state (Shelved). The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 261 has been changed to the saved state. Based on the notification, the VM state management unit 160 registers in the management table 121 that the virtual machine 261 has been in the saved state. In this way, the upsizing is stopped.

In this case, the VM switching processing unit 150 increases the CPU usage threshold of the condition U1 by a predetermined amount (for example, about 10%) with respect to the virtual machine 262, and manages the increased CPU usage threshold. Register in table 121.

FIG. 12B shows an execution example of upsizing. The VM switching processing unit 150 determines that the virtual machine 262 is upsized when it detects that the condition U2 is satisfied after the condition U1 is satisfied. In this case, the VM switching processing unit 150 changes the virtual machine 261 from the suspended state (Stopped) to the operating state (Active), and transmits a VM switching command for switching to the virtual machine 261 to the VM host 200.

When the VM switching execution unit 230 receives the VM switching instruction, the VM switching execution unit 230 changes the virtual machine 261 from the suspended state (Stopped) to the operating state (Active) based on the VM switching instruction. That is, the VM switching execution unit 230 allocates a CPU to the virtual machine 261 and starts the virtual machine 261 from the VM data loaded in the RAM of the VM host 200. Then, the VM switching execution unit 230 controls the load balancing unit 280 to distribute the requests distributed to the virtual machine 262 to the virtual machine 261.

Further, the VM switching execution unit 230 changes the virtual machine 262 from the operating state (Active) to the paused state (Stopped). The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 261 has been changed to the operating state and that the virtual machine 262 has been changed to the suspended state. Based on the notification, the VM state management unit 160 registers in the management table 121 that the virtual machine 261 has been in the operating state and that the virtual machine 262 has been in the suspended state.

Further, the VM deployment unit 130 transmits a VM deployment command for deploying the virtual machine 265 having higher performance than the virtual machine 261 in the saved state and a VM deletion command for deleting the virtual machine 263 to the VM host 200.

When the VM deployment execution unit 210 receives the VM deployment instruction and the VM deletion instruction, it deletes the virtual machine 263 based on these instructions (deletes the VM data of the virtual machine 263 from the HDD 203 or the external storage). Further, the VM deployment execution unit 210 deploys the virtual machine 265 in a saved state (Shelved). For example, the VM name of the virtual machine 265 is "VM1-LL".

The VM status notification unit 240 notifies the VM status management unit 160 that the virtual machine 263 has been deleted and that the virtual machine 265 has been deployed in the saved state. Based on the notification, the VM state management unit 160 deletes the record of the virtual machine 263 from the management table 121 and adds the record related to the virtual machine 265 to the management table 121.

Next, the processing procedure by the management server 100 and the VM hosts 200, 300, ... Will be described. Hereinafter, the description will be focused on the VM host 200, but the procedure is the same for other VM hosts.

FIG. 13 is a flowchart showing a processing example of the management server.
(S1) The VM deployment unit 130 receives a user's instruction to deploy the virtual machine 262. Then, the VM deployment unit 130 deploys the virtual machine 262 in the active state, and deploys the virtual machine 261 having the performance one step higher than the virtual machine and the virtual machine 262 having the performance one step lower than the virtual machine in the shelved state. A VM deployment command is sent to the VM host 200. The VM host 200 deploys the virtual machine 262 in the Active state and deploys the virtual machines 261,263 in the Shelved state based on the VM deployment instruction. The VM state management unit 160 receives a notification from the VM host 200 that the deployment has been performed, and adds records of virtual machines 261,262,263 to the management table 121.

(S2) The load calculation unit 140 periodically receives the load information of the virtual machine 262 from the VM host 200, and calculates the load of the virtual machine 262 based on the load information. The VM switching processing unit 150 monitors the load status of the virtual machine 262 calculated by the load calculation unit 140.

(S3) The VM switching processing unit 150 performs VM switching control based on the load status of the virtual machine 262. Details of the VM switching control will be described later.
(S4) The VM deletion unit 170 deletes unnecessary VMs that delete the switching destination VM when the VMs are not switched for a certain period of time. Details of deleting unnecessary VMs will be described later. Then, the process ends.

Here, the management server 100 can execute the processes of steps S2 and S3 and the processes of steps S4 in parallel.
Next, the procedure of VM switching control of the virtual machine by the management server 100 and the VM host 200 will be described. First, the procedure of downsizing VM switching control will be described.

FIG. 14 is a flowchart showing an example of VM switching control (downsize).
The VM switching control (downsize) corresponds to step S3 (however, when downsizing is performed).

It is assumed that the virtual machines 261,262,263 are deployed on the VM host 200 at the time before step S10, the virtual machines 262 are in the operating state, and the virtual machines 261,263 are in the evacuated state.

(S10) The load calculation unit 140 acquires the load state of the virtual machine 262 from the VM host 200 every minute. However, 1 minute is an example of the cycle for acquiring the load state, and other cycles may be used. The load calculation unit 140 calculates the load of the virtual machine 262 and provides it to the VM switching processing unit 150.

(S11) The VM switching processing unit 150 determines whether or not the load of the virtual machine 262 satisfies the condition D1. When the load of the virtual machine 262 satisfies the condition D1, the process proceeds to step S14. If the load of the virtual machine 262 does not satisfy the condition D1, the process proceeds to step S12.

The VM switching processing unit 150 manages the virtual machine 262 to satisfy the condition D1 with respect to the load of the virtual machine 262 by using the flag of the condition D1 satisfying item in the management table 121. When the VM switching processing unit 150 detects that the load of the virtual machine 262 has changed from the state where the condition D1 is not satisfied to the state where the condition D1 is satisfied, the VM switching processing unit 150 tells the virtual machine 262 the condition D1 of the management table 121. Change the sufficiency from False to True. After the condition D1 is satisfied, the VM switching processing unit 150 satisfies the condition D1 of the management table 121 when the load status of the virtual machine 262 does not satisfy the condition D1 (the CPU usage rate falls below the threshold value of the condition D1). , Change from True to False.

(S12) The VM switching processing unit 150 determines whether or not the state of the switching destination VM (here, the virtual machine 263) is Shelved with reference to the management table 121. When the state of the switching destination VM is Shelved, the process proceeds to step S10. If the state of the switching destination VM is not Shelved, the process proceeds to step S13. Here, in step S12, the case where the state of the switching destination VM is not Shelved is the case where the state of the switching destination VM is Stopped.

(S13) The VM switching processing unit 150 transmits a VM state change command for changing the switching destination VM from Stopped to Shelved to the VM host 200. In the VM host 200, the virtual machine 263 is changed from Stopped to Shelved based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 263 in the management table 121 to Shelved, and updates the state update date. Then, the process proceeds to step S10.

(S14) The VM switching processing unit 150 refers to the management table 121 and determines whether or not the state of the switching destination VM is Stopped. When the state of the switching destination VM is Stopped, the process proceeds to step S16. If the state of the switching destination VM is not Stopped, the process proceeds to step S15. Here, in step S14, the case where the state of the switching destination VM is not Stopped is the case where the state of the switching destination VM is Shelved.

(S15) The VM switching processing unit 150 transmits a VM state change command for changing the switching destination VM from Shelved to Stopped to the VM host 200. In the VM host 200, the virtual machine 263 is changed from Shelved to Stopped based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 263 in the management table 121 to Stopped, and updates the state update date. Then, the process proceeds to step S10.

(S16) The VM switching processing unit 150 determines whether or not the load of the virtual machine 262 satisfies the condition D2. When the load of the virtual machine 262 satisfies the condition D2, the process proceeds to step S19. If the load of the virtual machine 262 does not satisfy the condition D2, the process proceeds to step S17.

The VM switching processing unit 150 manages the virtual machine 262 to satisfy the condition D2 with respect to the load of the virtual machine 262 by using the flag of the condition D2 satisfying item in the management table 121. When the VM switching processing unit 150 detects that the load of the virtual machine 262 has changed from the state where the condition D2 is not satisfied to the state where the condition D2 is satisfied, the VM switching processing unit 150 tells the virtual machine 262 the condition D2 of the management table 121. Change the sufficiency from False to True.

(S17) The VM switching processing unit 150 determines whether or not VM switching has not occurred for a predetermined time from the time when the switching destination VM is changed to Stopped in step S15. If the VM switching has not occurred for a predetermined time from the time when the switching destination VM is changed to Stopped, the process proceeds to step S18. If a predetermined time has not elapsed since the switching destination VM was changed to Stopped, the process proceeds to step S10. Here, the predetermined time is, for example, 30 minutes (a time equal to or longer than the monitoring time of the condition D2).

(S18) The VM switching processing unit 150 transmits a VM state change command for changing the switching destination VM from Stopped to Shelved to the VM host 200. In the VM host 200, the virtual machine 263 is changed from Stopped to Shelved based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 263 in the management table 121 to Shelved, and updates the state update date. Then, the process proceeds to step S10.

(S19) The VM switching processing unit 150 performs VM switching (downsize). Specifically, the VM switching processing unit 150 changes the switching destination VM from Stopped to Active, and sends a VM switching instruction to the VM host 200 to switch from the current VM (virtual machine 262) to the switching destination VM (virtual machine 263). Send. In the VM host 200, the virtual machine 263 is changed from Stopped to Active based on the VM switching instruction. Then, the load balancing unit 280 switches the distribution destination of the request from the client 50 from the virtual machine 262 to the virtual machine 263. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 263 in the management table 121 to Active, and updates the state update date. After that, the virtual machine whose load is monitored becomes the virtual machine after switching.

(S20) The VM switching processing unit 150 executes downsize post-processing. Details of downsize post-processing will be described later. Then, the process proceeds to step S10.
FIG. 15 is a flowchart showing an example of downsize post-processing.

The downsizing post-processing corresponds to step S20.
(S30) The VM switching processing unit 150 transmits a VM state change command for changing the switching source VM (here, the virtual machine 262) from Active to Stopped to the VM host 200. In the VM host 200, the virtual machine 262 is changed from Active to Stopped based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 262 in the management table 121 to Stopped, and updates the state update date.

(S31) The VM deployment unit 130 issues a VM deployment instruction to the VM host 200 to deploy a virtual machine (virtual machine 264) having a performance one step lower than that of the switching destination VM (here, virtual machine 263) in a saved state (shelved). Send. In the VM host 200, the virtual machine 264 is deployed in the saved state based on the VM deployment instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the deployment has been performed, and adds a record of the virtual machine 264 to the management table 121.

(S32) The VM deployment unit 130 transmits a VM deletion command to the VM host 200 to delete a virtual machine (here, virtual machine 261) having a performance one step higher than that of the switching source VM. In the VM host 200, the virtual machine 261 is deleted based on the VM delete instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the deletion has been performed, and deletes the record of the virtual machine 261 from the management table 121.

Next, a specific example of downsize post-processing in the case of downsizing from the virtual machine 262 to the virtual machine 261 will be described.
FIG. 16 is a diagram showing a specific example of downsize post-processing.

First, the VM switching control of the management server 100 switches the distribution destination of the request from the client 50 from the virtual machine 262 to the virtual machine 263 (step T10). At this timing, the virtual machine 261 is Shelved. The virtual machine 262 is Active. The virtual machine 263 is Active.

Next, the virtual machine 262 is changed to Stopped (step T11). A virtual machine 264 having a performance one step lower than that of the virtual machine 263 is deployed (step T12). The virtual machine 261 is deleted (step T13). As a result, the virtual machine 262 with the performance one step higher is deployed in Stopped with respect to the virtual machine 263 in operation, and the virtual machine 264 with the performance one step lower is deployed in Shelved (step T14). ..

After step T14, the management server 100 monitors the load status of the virtual machine 263, and performs VM switching control by the same procedure as the procedure for the virtual machine 262. The VM switching processing unit 150 changes the virtual machine 262, which is the switching source VM, to Shelved, for example, after a predetermined time (for example, 30 minutes) has elapsed after the virtual machine 263 is set to the Active state. This reduces the charge to the user.

Next, the procedure for upsizing VM switching control will be described.
FIG. 17 is a flowchart showing an example of VM switching control (upsize).
The VM switching control (upsize) corresponds to step S3 (however, when upsizing is performed).

It is assumed that the virtual machines 261,262,263 are deployed on the VM host 200 at the time before step S40, the virtual machines 262 are in the operating state, and the virtual machines 261,263 are in the evacuated state.

(S40) The load calculation unit 140 acquires the load state of the virtual machine 262 from the VM host 200 every minute. However, 1 minute is an example of the cycle for acquiring the load state, and other cycles may be used. The load calculation unit 140 calculates the load of the virtual machine 262 and provides it to the VM switching processing unit 150.

(S41) The VM switching processing unit 150 determines whether or not the load of the virtual machine 262 satisfies the condition U1. When the load of the virtual machine 262 satisfies the condition U1, the process proceeds to step S44. If the load of the virtual machine 262 does not satisfy the condition U1, the process proceeds to step S42.

The VM switching processing unit 150 manages the virtual machine 262 to satisfy the condition U1 with respect to the load of the virtual machine 262 by using the flag of the condition U1 satisfying item in the management table 121. When the VM switching processing unit 150 detects that the load of the virtual machine 262 has changed from the state in which the condition U1 is not satisfied to the state in which the condition U1 is satisfied, the VM switching processing unit 150 tells the virtual machine 262 that the condition U1 in the management table 121 is satisfied. Change the sufficiency from False to True. After the condition U1 is satisfied, the VM switching processing unit 150 satisfies the condition U1 of the management table 121 when the load status of the virtual machine 262 does not satisfy the condition U1 (the CPU usage rate falls below the threshold value of the condition U1). , Change from True to False.

(S42) The VM switching processing unit 150 determines whether or not the state of the switching destination VM (here, the virtual machine 261) is Shelved with reference to the management table 121. When the state of the switching destination VM is Shelved, the process proceeds to step S40. If the state of the switching destination VM is not Shelved, the process proceeds to step S43. Here, in step S42, the case where the state of the switching destination VM is not Shelved is the case where the state of the switching destination VM is Stopped.

(S43) The VM switching processing unit 150 transmits a VM state change command for changing the switching destination VM from Stopped to Shelved to the VM host 200. In the VM host 200, the virtual machine 261 is changed from Stopped to Shelved based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 261 in the management table 121 to Shelved, and updates the state update date. Then, the process proceeds to step S40.

(S44) The VM switching processing unit 150 refers to the management table 121 and determines whether or not the state of the switching destination VM is Stopped. When the state of the switching destination VM is Stopped, the process proceeds to step S46. If the state of the switching destination VM is not Stopped, the process proceeds to step S45. Here, in step S44, the case where the state of the switching destination VM is not Stopped is the case where the state of the switching destination VM is Shelved.

(S45) The VM switching processing unit 150 transmits a VM state change command for changing the switching destination VM from Shelved to Stopped to the VM host 200. In the VM host 200, the virtual machine 261 is changed from Shelved to Stopped based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 261 in the management table 121 to Stopped, and updates the state update date. Then, the process proceeds to step S40.

(S46) The VM switching processing unit 150 determines whether or not the load of the virtual machine 262 satisfies the condition U2. When the load of the virtual machine 262 satisfies the condition U2, the process proceeds to step S50. If the load of the virtual machine 262 does not satisfy the condition U2, the process proceeds to step S47.

The VM switching processing unit 150 manages the virtual machine 262 to satisfy the condition U2 with respect to the load of the virtual machine 262 by using the flag of the condition U2 satisfying item in the management table 121. When the VM switching processing unit 150 detects that the load of the virtual machine 262 has changed from the state in which the condition U2 is not satisfied to the state in which the condition U2 is satisfied, the VM switching processing unit 150 tells the virtual machine 262 that the condition U2 in the management table 121 is satisfied. Change the sufficiency from False to True.

(S47) The VM switching processing unit 150 determines whether or not VM switching has not occurred for a predetermined time from the time when the switching destination VM is changed to Stopped in step S45. If the VM switching has not occurred for a predetermined time from the time when the switching destination VM is changed to Stopped, the process proceeds to step S48. If a predetermined time has not elapsed since the switching destination VM was changed to Stopped, the process proceeds to step S40. Here, the predetermined time is, for example, 30 minutes (a time equal to or longer than the monitoring time of the condition U2).

(S48) The VM switching processing unit 150 transmits a VM state change command for changing the switching destination VM from Stopped to Shelved to the VM host 200. In the VM host 200, the virtual machine 261 is changed from Stopped to Shelved based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 261 in the management table 121 to Shelved, and updates the state update date.

(S49) The VM state management unit 160 increases the condition U1-CPU usage threshold in the management table 121 by 10%. For example, the VM state management unit 160 may add 10% to the condition U1-CPU usage threshold or multiply the condition U1-CPU usage threshold by 1.1. Regarding the increase amount of the condition U1-CPU usage rate threshold value, 10% is an example and may be another increase amount. Then, the process proceeds to step S40.

(S50) The VM switching processing unit 150 performs VM switching (upsize). Specifically, the VM switching processing unit 150 changes the switching destination VM from Stopped to Active, and sends a VM switching instruction to the VM host 200 to switch from the current VM (virtual machine 262) to the switching destination VM (virtual machine 261). Send. In the VM host 200, the virtual machine 261 is changed from Stopped to Active based on the VM switching instruction. Then, the load balancing unit 280 switches the distribution destination of the request from the client 50 from the virtual machine 262 to the virtual machine 261. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 261 in the management table 121 to Active, and updates the state update date. After that, the virtual machine whose load is monitored becomes the virtual machine after switching.

(S51) The VM switching processing unit 150 executes upsize post-processing. Details of upsize post-processing will be described later. Then, the process proceeds to step S40.
Here, as in step S49, by increasing the condition U1-CPU usage rate threshold value, it is possible to prevent the switching destination VM from being immediately shifted to Stopped. As a result, the amount of resources (RAM 202) used can be suppressed. In addition, by reducing the amount of resources used, the amount charged to the user can be reduced.

FIG. 18 is a flowchart showing an example of upsize post-processing.
The upsize post-processing corresponds to step S51.
(S60) The VM switching processing unit 150 transmits a VM state change command for changing the switching source VM (here, the virtual machine 262) from Active to Stopped to the VM host 200. In the VM host 200, the virtual machine 262 is changed from Active to Stopped based on the VM state change instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the change has been made, changes the state of the virtual machine 262 in the management table 121 to Stopped, and updates the state update date.

(S61) The VM deployment unit 130 issues a VM deployment instruction to the VM host 200 to deploy a virtual machine (virtual machine 265) having a performance one step higher than that of the switching destination VM (here, virtual machine 261) in a saved state (shelved). Send. In the VM host 200, the virtual machine 265 is deployed in the saved state based on the VM deployment instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the deployment has been performed, and adds a record of the virtual machine 265 to the management table 121.

(S62) The VM deployment unit 130 transmits a VM deletion command to the VM host 200 to delete a virtual machine (here, virtual machine 263) having a performance one step lower than that of the switching source VM. On the VM host 200, the virtual machine 263 is deleted based on the VM delete instruction. Further, the VM state management unit 160 receives a notification from the VM host 200 that the deletion has been performed, and deletes the record of the virtual machine 263 from the management table 121.

Next, a specific example of the upsizing post-processing in the case of upsizing from the virtual machine 262 to the virtual machine 263 will be described.
FIG. 19 is a diagram showing a specific example of upsize post-processing.

First, the VM switching control of the management server 100 switches the distribution destination of the request from the client 50 from the virtual machine 262 to the virtual machine 261 (step T20). At this timing, the virtual machine 261 is Active. The virtual machine 262 is Active. Virtual machine 263 is Shelved.

Next, the virtual machine 262 is changed to Stopped (step T21). A virtual machine 265 having a performance one step higher than that of the virtual machine 261 is deployed (step T22). The virtual machine 263 is deleted (step T23). As a result, the virtual machine 262 with the performance one step lower is deployed in Stopped with respect to the virtual machine 261 in operation, and the virtual machine 265 with the performance one step higher is deployed in Shelved (step T24). ..

After step T24, the management server 100 monitors the load status of the virtual machine 261 and performs VM switching control by the same procedure as the procedure for the virtual machine 262. The VM switching processing unit 150 changes the virtual machine 262, which is the switching source VM, to Shelved, for example, after a predetermined time (for example, 30 minutes) has elapsed after the virtual machine 261 is set to the Active state. This reduces the charge to the user.

Next, the procedure for deleting unnecessary VMs by the management server 100 and the VM host 200 will be described.
FIG. 20 is a flowchart showing an example of deleting unnecessary VMs.

Deleting unnecessary VMs corresponds to step S4.
(S70) The VM deletion unit 170 monitors the status update date of the switching destination VM at 1-minute intervals with reference to the management table 121. Here, regarding the monitoring cycle, 1 minute is an example, and other cycles may be used.

(S71) The VM deletion unit 170 determines whether or not one week has passed from the state update date recorded for the switching destination VM. If one week has passed from the state update date recorded for the switching destination VM, the process proceeds to step S72. If one week has not passed from the state update date recorded for the switching destination VM, the process proceeds to step S70. Here, in the above example, there are two switching destination VMs for the operating VM. The VM deletion unit 170 determines step S71 for each switching destination VM, for example.

(S72) The VM deletion unit 170 deletes the switching destination VM for which one week has passed from the state update date. Specifically, the VM deletion unit 170 transmits a VM deletion command for deleting the switching destination VM to the VM host 200. The VM deletion execution unit 250 deletes the instance of the corresponding switching destination VM based on the VM deletion instruction. As a result, the VM data of the corresponding switching destination VM that has been in the Shelved state is deleted from the HDD 203 or the external storage. The VM state management unit 160 receives a notification from the VM host 200 that the deletion has been performed, and deletes the deleted switching destination VM record from the management table 121. Then, the process proceeds to step S70.

In this way, the management server 100 deletes the VM data of the switching destination VM whose state has not been updated for a relatively long period of time from the HDD 203 or the external storage among the switching destination VMs. This is because it is highly probable that switching to the switching destination VM will not be performed in the future. By deleting the VM data of the switching destination VM from the HDD 203 or the external storage, the resource usage can be further suppressed. Further, by suppressing the resource usage amount, when the user is charged for the use of the capacity of the auxiliary storage device such as the HDD 203 or the external storage, the charge amount for the user can be suppressed.

As described above, the management server 100 reduces the amount of resources used and enables rapid switching of virtual machines. Specifically, the management server 100 normally operates the virtual machine 262. When the load of the virtual machine 262 becomes higher than the first threshold value, the management server 100 loads the VM data of the virtual machine 261 having higher performance than the virtual machine 262 into the RAM of the VM host 200 to prepare the virtual machine 262. Therefore, in the normal state, it is not necessary to load the virtual machine 261 into the RAM, and the amount of the RAM used can be suppressed. As a result, it is possible to suppress the charge associated with the use of the RAM for the user.

The user may be charged for using the capacity of the auxiliary storage device such as the HDD 203 or the external storage. However, the charge for using the capacity of the auxiliary storage device is cheaper than the charge for using the capacity of the RAM, and the charge for the user when the virtual machine instance is saved in the auxiliary storage device. Can be kept low.

Further, while the virtual machine 261 is in the ready state (paused state), the virtual machine 261 is not started and the CPU of the VM host 200 is not used. Therefore, in the prepared state of the virtual machine 261, it is possible to suppress the charge for the user due to the use of the CPU.

Moreover, the VM data of the virtual machine 261 has already been loaded into the RAM of the VM host 200. Therefore, when the load of the virtual machine 262 exceeds the second threshold value higher than the first threshold value, the VM host 200 quickly starts the virtual machine 261 and moves from the virtual machine 262 to the virtual machine 261. Can be switched.

In this way, resource usage can be reduced and virtual machines can be switched quickly. In addition, by reducing the amount of resources used, it is possible to reduce the amount charged to the user.

After switching, the VM host 200 may stop the virtual machine 262 and, for example, save the VM data of the virtual machine 262 from the RAM of the VM host 200 to an auxiliary storage device or the like after a lapse of a predetermined time. As a result, the amount of resources used by the virtual machine 262 can be further reduced.

Further, when the load of the virtual machine 262 is lower than the third threshold value lower than the first threshold value, the management server 100 loads the VM data of the virtual machine 263 into the RAM of the VM host 200 to prepare it. Then, when the load of the virtual machine 262 is lower than the fourth threshold value, which is lower than the third threshold value, the management server 100 causes the VM host 200 to switch the virtual machine 262 to the low-performance virtual machine 263. As a result, the amount of resources used can be further reduced. By reducing the amount of resources used, the amount charged to the user can be reduced. Also, the VM data of the virtual machine 263 has already been loaded into the RAM of the VM host 200. Therefore, when the load of the virtual machine 262 is lower than the fourth threshold value, the VM host 200 can quickly start the virtual machine 263 and switch from the virtual machine 262 to the virtual machine 263. can.

The information processing of the first embodiment can be realized by causing the processing unit 12 to execute the program. Further, the information processing of the second embodiment can be realized by causing the CPU 101 to execute the program. The program can be recorded on a computer-readable recording medium 113.

For example, the program can be distributed by distributing the recording medium 113 on which the program is recorded. Alternatively, the program may be stored in another computer and distributed via the network. For example, the computer may store (install) a program recorded on the recording medium 113 or a program received from another computer in a storage device such as a RAM 102 or an HDD 103, read the program from the storage device, and execute the program. good.

1 Information processing system 10 Information processing device 11 Storage unit 12 Processing unit 20 Physical machine 21 Hardware 21a Memory 21b Processor 22 Hypervisor 23,24 Virtual machine 30 Storage device 31,32,33 VM data 40 Network

Claims (8)

A storage unit that stores information indicating the correspondence between multiple virtual machines with different performances,
An instruction for arranging a plurality of image data corresponding to the plurality of virtual machines in the storage device is transmitted to the physical machine, the load of the first virtual machine running on the physical machine is monitored, and the load is the first. When the threshold value is exceeded, the image data of the second virtual machine having higher performance than that of the first virtual machine is loaded into the memory from the storage device based on the information stored in the storage unit, and the second virtual machine is loaded. When an instruction to prepare the virtual machine is sent to the physical machine and the load exceeds a second threshold value higher than the first threshold value, the first virtual machine is sent to the second virtual machine. A processing unit that sends a switching command to the physical machine,
Information processing device with. The processing unit transmits an instruction to prepare the second virtual machine to the physical machine when the time when the load continuously exceeds the first threshold reaches the first time. When the time when the load continuously exceeds the second threshold reaches the second time, an instruction to switch the first virtual machine to the second virtual machine is transmitted to the physical machine.
The information processing device according to claim 1. After the processing unit transmits the command to prepare the second virtual machine, when a predetermined time elapses without switching from the first virtual machine to the second virtual machine, the first virtual machine Sends an instruction to delete the image data of the virtual machine 2 from the memory to the physical machine.
The information processing device according to claim 1 or 2. The processing unit transmits an instruction to delete the image data of the second virtual machine from the memory to the physical machine, and increases the first threshold value.
The information processing device according to claim 3. When the load of the first virtual machine falls below the third threshold value, which is lower than the first threshold value, the processing unit receives the information stored in the storage unit from the first virtual machine. An instruction to load the low-performance image data of the third virtual machine from the storage device into the memory and prepare the third virtual machine is transmitted to the physical machine, and the load is the third. When the value falls below the fourth threshold value, which is lower than the threshold value of, an instruction to switch the first virtual machine to the third virtual machine is transmitted to the physical machine.
The information processing device according to claim 1. When a predetermined period elapses without changing to the ready state after another virtual machine having a performance different from that of the first virtual machine is saved in the storage device, the processing unit performs image data of the other virtual machine. Is sent to the physical machine to delete the
The information processing device according to claim 1. On the computer
Sends an instruction to the physical machine to place multiple image data corresponding to multiple virtual machines with different performance in the storage device,
Monitor the load of the first virtual machine running on the physical machine and
When the load exceeds the first threshold value, the image data of the second virtual machine having higher performance than the first virtual machine is loaded from the storage device into the memory to prepare the second virtual machine. Is sent to the physical machine to
When the load exceeds a second threshold value higher than the first threshold value, an instruction to switch the first virtual machine to the second virtual machine is transmitted to the physical machine.
A virtual machine management program that executes processing. A virtual machine management method executed by an information processing system.
Multiple image data corresponding to multiple virtual machines with different performances are placed in the storage device,
Monitor the load of the first virtual machine running in the information processing system,
When the load exceeds the first threshold value, the image data of the second virtual machine having higher performance than the first virtual machine is loaded from the storage device into the memory to prepare the second virtual machine. ,
When the load exceeds a second threshold value higher than the first threshold value, the first virtual machine is switched to the second virtual machine.
Virtual machine management method.

Images