JP6743368B2 - Virtual infrastructure host, virtual infrastructure host control method, virtual infrastructure host program, and communication system - Google Patents

Description

The present invention relates to a virtual base host, a virtual base host control method, a virtual base host program, and a communication system.

A virtual environment in which a plurality of virtual machines are simultaneously operated on a physical device called one virtual infrastructure host has become widespread. FIG. 1 shows a configuration example of a system in a virtual environment.

The client device 10 in FIG. 1 remotely controls the virtual machine 21 on the virtual base host 20, for example, using the remote desktop connection function of the OS (Operating System) installed in itself. Then, the client device 10 receives the result based on the remote operation from the virtual machine 21.

For example, when the virtual machine 21 of the virtual infrastructure host 20 in FIG. 1 is remotely operated by the client device 10, the virtual machine 21 requests the hypervisor 22 to allocate the resource of the hardware 23. Then, the virtual machine 21 uses the resource of the hardware 23 allocated by the hypervisor 22 to perform the operation based on the remote operation. For example, the virtual machine 21 outputs an image, a sound, etc. to the client device 10.

When requested to use the hardware 23 from the virtual machine 21, the hypervisor 22 of the virtual infrastructure host 20 in FIG. 1 allocates the resources of the hardware 23 to the virtual machine 21. Note that the hypervisor 22 appropriately allocates the resources of the hardware 23 even when the virtual machines 21 are requested to use the hardware 23 at the same time, and thereby the applications in the virtual machines 21 are It enables simultaneous processing such as.

The hardware 23 of the virtual base host 20 is realized by, for example, a graphic board. In this case, the virtual machine 21 can cause the resources of the hardware 23 allocated to itself to perform the graphic processing. That is, even if the client machine 10 requests the virtual machine 21 to execute an application that requires high graphic processing performance, the virtual machine 21 can handle this. In relation to the above, in recent years, a GPU virtualization method in which a resource of a GPU (Graphics Processing Unit) is allocated to each of virtual machines in a virtual environment in a one-to-one manner has become widespread. As a GPU virtualization method, for example, vGPU by Grid GPU of NVIDIA is widely known.

Further, as a technique related to the GPU virtualization method, Patent Document 1 discloses a technique related to a virtualized graphic architecture that provides benefits for compatibility, security, digital rights management, and the like. ..

JP 2006-190281 A

The GPU resource allocation in the GPU virtualization method is performed based on, for example, the graphic performance required for each user in advance with a performance measurement tool called a profiler, and based on the result of the digitization. Further, for example, the allocation of the GPU resource is performed by allocating the GPU resource with the same size for each user and then adjusting the GPU resource based on the operation feeling for each user.

The above-mentioned GPU resource allocation is mainly performed manually. Therefore, a GPU resource having an appropriate size may not be allocated to the virtual machine. For example, this is a case where the size of the GPU resource allocated to the virtual machine does not meet the size required by the application executed in the virtual machine. Further, for example, it is a case where a GPU resource having a size larger than necessary is allocated to the virtual machine.

In the former case, the user of the client device that logs into the virtual machine and causes the virtual machine to execute an application feels stress due to display delay of images transferred from the virtual machine, poor followability of mouse operations, and the like. However, even if the allocated GPU resource does not satisfy the required size, the GPU resource once allocated is still maintained.

On the other hand, in the latter case, there arises a problem that the utilization efficiency of the GPU resource is deteriorated.

However, Patent Document 1 neither describes nor suggests a method for solving the above problems.

The present invention has been made in view of the above problems, and provides a virtual base host, a virtual base host control method, a virtual base host program, and a communication system that can appropriately allocate a GPU resource of an appropriate size. The purpose is to

In order to achieve the above object, the virtual infrastructure host according to an aspect of the present invention, when a virtual machine, a GPU resource assigned to the virtual machine, and a size change request for the GPU resource are notified from the virtual machine, And a hypervisor for changing the size of the GPU resource.

In order to achieve the above object, a communication system according to an aspect of the present invention includes a virtual infrastructure host and an information processing device connected to the virtual infrastructure host, wherein the virtual infrastructure host includes a virtual machine and the virtual machine. The virtual machine includes a GPU resource assigned to a machine and a hypervisor that changes the size of the GPU resource when the virtual machine is notified of a request to change the size of the GPU resource. When the predetermined operation is detected, the size change request is notified to the hypervisor.

In order to achieve the above object, a control method of a virtual base host according to an aspect of the present invention is a control method of a virtual base host including a virtual machine and a GPU resource allocated to the virtual machine, If the machine requests a change in the size of the GPU resource, the size of the GPU resource is changed.

In order to achieve the above object, a program of a virtual base host according to an aspect of the present invention is a computer of a virtual base host including a virtual machine and a GPU resource allocated to the virtual machine, and the virtual machine is the GPU. When it is requested to change the size of the resource, the processing for changing the size of the GPU resource is executed.

According to the present invention, an appropriately sized GPU resource can be appropriately allocated.

1 shows an example of the configuration of a system in a virtual environment composed of a client device 10 and a virtual infrastructure host 20. It is an example of a block configuration diagram of a virtual base host 30 according to the first embodiment of the present invention. 6 is a flowchart showing an example of operation of the virtual base host 30 according to the first exemplary embodiment of the present invention. It is an example of a system configuration diagram of a communication system 100 according to a second embodiment of the present invention. It is an example of a block configuration diagram of a virtual base host 200 according to a second embodiment of the present invention. It is a flowchart which shows an example of operation|movement of the virtual machine 220 and the hypervisor 230 which concern on the 2nd Embodiment of this invention. It is a flowchart which shows an example of operation|movement of the hypervisor 230 which concerns on the 2nd Embodiment of this invention.

(First embodiment)
A first embodiment of the present invention will be described.

FIG. 2 shows a block configuration diagram of the virtual base host 30 according to the first embodiment of the present invention. As shown in FIG. 2, the virtual infrastructure host 30 includes a GPU resource 31, a virtual machine 32, and a hypervisor 33.

The GPU resource 31 is assigned to the virtual machine 32. The GPU resource performs graphic processing of an application executed on the virtual machine 32, for example.

The virtual machine 32 notifies the hypervisor 33 of the size change request of the GPU resource 31. For example, the virtual machine 32 according to the present embodiment determines that the GPU resource 31 allocated to the virtual machine 32 is insufficient when the user detects an invalid input operation to the application on the virtual machine 32 multiple times. To do. Then, the virtual machine 32 notifies the hypervisor 33 of the size change request of the GPU 31.

The hypervisor 33 changes the size of the GPU resource 31 when notified of the size change request of the GPU resource 31 from the virtual machine 32.

Next, the operation of the virtual infrastructure host 30 will be described with reference to FIG.

(S101: Operation of Virtual Machine 32)
The virtual machine 32 notifies the hypervisor 33 of the size change request of the GPU resource 31. Here, for example, when the virtual machine 32 according to the present embodiment detects an invalid input operation to the application on the virtual machine 32 from the user a plurality of times, the GPU resource 31 allocated to the virtual machine 32 becomes insufficient. Determine that Then, the virtual machine 32 notifies the hypervisor 33 of the size change request of the GPU 31.

(S102: Operation of hypervisor 33)
The hypervisor 33 changes the size of the GPU resource 31 when notified of the size change request of the GPU resource 31 from the virtual machine 32.

As described above, the hypervisor 33 in the virtual infrastructure host 30 according to the present embodiment changes the size of the GPU resource 31 when the virtual machine 32 is notified of the size change request of the GPU resource 31. That is, the virtual machine 32 can operate using a GPU resource having an appropriate size.

From the above, according to this embodiment, it is possible to appropriately allocate a GPU resource of an appropriate size.

Further, in view of the above, the virtual base host 30 according to the present embodiment is a virtual machine for three-dimensional CAD (Computer Aided Design) and three-dimensional CG (Computer Graphics) that require high graphic processing performance. An appropriate size resource can be allocated to 32. As a result, the virtual infrastructure host 30 can increase the number of virtual machines 32 that it can accommodate.

(Second embodiment)
Next, a second embodiment of the present invention will be described.

FIG. 4 shows a system configuration diagram of the communication system 100 according to the second embodiment of the present invention. As shown in FIG. 4, the communication system 100 includes a virtual infrastructure host 200, and a client device 300 that logs into a virtual machine 220 (described later) included in the virtual infrastructure host 200 and causes the virtual machine 220 to execute an application.

As shown in FIG. 5, the virtual base host 200 according to the second embodiment of the present invention includes a GPU 210 including virtual GPUs 211-1 to 211-n, virtual machines 220-1 to 220-n, and a hypervisor 230. And

Hereinafter, when it is not necessary to distinguish between the virtual GPUs 211-1 to 211-n for description, the virtual GPUs 211-1 to 211-n will be referred to as virtual GPUs 211 for description. Further, when it is not necessary to distinguish between the virtual machines 220-1 to 220-n for description, the virtual machines 220-1 to 220-n are referred to as virtual machines 220 for description.

Here, the virtual infrastructure host 200 is an example of the virtual infrastructure host 30 in the first embodiment of the present invention. The virtual GPU 211 is an example of the GPU resource 31 according to the first embodiment of this invention. The virtual machine 220 is an example of the virtual machine 32 according to the first embodiment of this invention. The hypervisor 230 is an example of the hypervisor 33 in the first embodiment of the present invention.

The GPU 210 is hardware that performs graphic processing of an application that the client device 300 that has logged in to the virtual machine 220 causes the virtual machine 220 to execute, and includes virtual GPUs 211-1 to 211-n. The virtual GPUs 211-1 to 211-n are resources of the GPU 210 and are assigned to the virtual machines 220-1 to 220-n in a one-to-one manner. Here, for example, the virtual GPU 211-1 performs graphic processing of an application executed by the virtual machine 220-1. Further, for example, the virtual GPU 211-2 performs graphic processing of an application executed by the virtual machine 220-2. The size of the virtual GPUs 211-1 to 211-n may be changed by the allocation changing unit 233 of the hypervisor 230, as described later.

The virtual machine 220 executes an application installed in itself, etc., based on an instruction from the client device 300 logged in to the virtual machine 220. As illustrated in FIG. 5, the virtual machine 220 includes an application control unit 221, a graphic driver control unit 222, a detection unit 223, and a first communication processing unit 224.

The application control unit 221 controls execution of applications installed in the virtual machine 220 based on an instruction from the client device 300. When controlling an application that requires graphic processing, the application control unit 221 notifies the graphic driver control unit 222 to that effect.

The graphic driver control unit 222 performs the following operation when notified from the application control unit 221 that the application requiring graphic processing is controlled. In this case, the graphic driver control unit 222 requests the virtual GPU 211 of the GPU 210 to perform the graphic processing via the graphic driver installed in the virtual machine 220. Here, the graphic driver installed in the virtual machine 220 manages address information indicating the location of the virtual GPU 211 assigned to the virtual machine 220 and size information indicating the size.

Further, the graphic driver control unit 222 notifies the first communication processing unit 224 of the fact that the size of the virtual GPU 211 allocated to the virtual machine 220 has been expanded and the address information and size information of the resource of the additionally allocated GPU 210. When received from the hypervisor 230, the following operation is performed. In this case, the graphic driver control unit 222 causes the graphic driver installed in the virtual machine 220 to additionally manage the received address information and size information. As a result, the virtual machine 220 can cause the virtual GPU 211 whose size has been increased to perform graphic processing.

In addition, the graphic driver control unit 222 notifies the first communication processing unit 224 of the fact that the size of the virtual GPU 211 assigned to the virtual machine 220 is reduced and the address information and size information of the virtual GPU 211 after the size reduction. When received from the visor 230, the following operation is performed. In this case, the graphic driver control unit 222 replaces the address information and the size information of the virtual GPU 211 currently managed by the graphic driver installed in the virtual machine 220, with the address information and the size information of the virtual GPU 211 after the size reduction. Manage size information. As a result, the virtual machine 220 can cause the virtual GPU 211, which has been reduced in size, to perform graphic processing.

The detection unit 223 performs the following operation when the application installed in the virtual machine 220 is controlled by the application control unit 221, that is, when the application is executed by the client device 300. In this case, the detection unit 223 detects a predetermined operation on the client device 300. Here, the predetermined operation is performed when the user of the client device 300 feels stress due to a processing delay of an application executed by the virtual machine 220 caused by a size shortage of the virtual GPU 211 assigned to the virtual machine 220. This is the operation to be performed. The predetermined operation is, for example, an operation of repeatedly inputting the same instruction to the client apparatus 300 within a predetermined time, an operation of repeatedly inputting an instruction in a screen area in which the input on the client apparatus 300 is invalidated, or the like. Is. Then, when the detection unit 213 detects the above-described predetermined operation, the detection unit 213 notifies the hypervisor 230 via the first communication processing unit 224 of a size increase request for the virtual GPU 211 assigned to the virtual machine 220.

The first communication processing unit 224 exchanges various information with the hypervisor 230.

As shown in FIG. 5, the hypervisor 230 includes a second communication processing unit 231, a monitoring unit 232, and an allocation changing unit 233.

The second communication processing unit 231 exchanges various information with the virtual machine 220.

The monitoring unit 232 monitors the usage rates of the virtual GPUs 211-1 to 211-n in the GPU 210.

The allocation changing unit 233 operates as follows when the second communication processing unit 231 receives a size increase request for the virtual GPU 211 allocated to the virtual machine 220 from the virtual machine 220. In this case, the allocation changing unit 233 confirms the usage rate of the virtual GPU 211 allocated to the virtual machine 220 other than the virtual machine 220 that has notified the size increase request via the monitoring unit 232, and responds to the size increase request. It is determined whether or not it is possible.

When it is determined that the size expansion request can be met, the allocation changing unit 233 expands the size of the virtual GPU 211 corresponding to the size expansion request by additionally allocating the resource of the GPU 210. Then, the allocation changing unit 233 expands the size of the virtual GPU 211 allocated to the virtual machine 220 to the virtual machine 220 that has notified the size expansion request, and the address information and size information of the additionally allocated GPU 210. And notify. At this time, when there is a virtual machine 220 whose size has been expanded in the allocated virtual GPU 211, in addition to the virtual machine 220 that has notified the size expansion request, the allocation changing unit 233 sends the virtual machine 220 to the virtual machine 220. Also, the same notification as above is performed. Further, at this time, when there is a virtual machine 220 whose size has been reduced in the allocated virtual GPU 211, the allocation changing unit 233 informs the virtual machine 220 that the size of the virtual GPU 211 has been reduced. The address information and the size information of the virtual GPU 211 after the size reduction is notified.

On the other hand, when the allocation changing unit 233 determines that the size expansion request cannot be met, the size change of the virtual GPU 211 allocated to the virtual machine 220 is not possible with respect to the virtual machine 220 that has notified the size expansion request. Notify that it is possible.

Here, when the second communication processing unit 231 receives a size increase request for the virtual GPU 211-1 from the virtual machine 220-1, the allocation changing unit 233 determines the usage rates of the virtual GPUs 211-2 to 211-n. Confirm through the monitoring unit 232. The allocation changing unit 233, for example, identifies the virtual GPU 211-2 that is allocated to the virtual machine 220-2 as the virtual GPU 211 whose usage rate is equal to or more than a predetermined time and is equal to or less than a predetermined threshold value, among the virtual GPUs 211-2 to 211-n. In that case, the size of the virtual GPU 211-2 is reduced. Then, the allocation changing unit 233 expands the size of the virtual GPU 211-1 by additionally allocating the resources of the GPU 210 that have been reduced. After that, the allocation changing unit 233 notifies the virtual machine 220-1 that the size of the virtual GPU 211-1 has been expanded, and the address information and size information of the additionally allocated resource of the GPU 210. Also, the allocation changing unit 233 notifies the virtual machine 220-2 that the size of the virtual GPU 211-2 has been reduced, and the address information and size information of the virtual GPU 211-2 after the size reduction. The size reduction of the virtual GPU 211-2 is performed within a range that does not hinder the graphic processing of the virtual GPU 211-2.

The client device 300 is, for example, a communication device such as a PC (Personal Computer) or a smartphone. The client device 300 logs in to the virtual machine 220 of the virtual infrastructure host 200. Then, the client device 300 instructs the virtual machine 220 to execute the application installed in the virtual machine 220.

Next, an operation of expanding the virtual GPU 211 assigned to the virtual machine 220 will be described with reference to FIG.

(S201: Operation in Virtual Machine 220)
When the application installed in the virtual machine 220 is controlled by the application control unit 221, the detection unit 223 performs the following operation when a predetermined operation of the user on the client device 300 is detected. In this case, the detection unit 223 notifies the hypervisor 230 of the size increase request of the virtual GPU 211 assigned to the virtual machine 220 via the first communication processing unit 224. Here, the predetermined operation of the user is, for example, an operation of repeatedly inputting the same instruction to the client device 300 within a predetermined time, or an instruction on the screen area where the input on the client device 300 is invalidated. Is an operation of repeatedly inputting.

(S202: Operation in Hypervisor 230)
The allocation changing unit 233 operates as follows when the second communication processing unit 231 receives a size increase request for the virtual GPU 211 allocated to the virtual machine 220 from the virtual machine 220. In this case, the allocation changing unit 233 confirms the usage rate of the virtual GPU 211 allocated to the virtual machine 220 other than the virtual machine 220 that has notified the size increase request via the monitoring unit 232, and responds to the size increase request. It is determined whether or not it is possible.

(S203: Operation in hypervisor 230)
When determining that the size expansion request in S202 can be met (YES in S202), the allocation changing unit 233 expands the size of the virtual GPU 211 corresponding to the size expansion request in S202 by additionally allocating the resource of the GPU 210. Then, the allocation changing unit 233 expands the size of the virtual GPU 211 allocated to the virtual machine 220 to which the virtual machine 220 has been notified of the size expansion request in S202, and indicates the resource address of the additionally allocated GPU 210. Notify information and size information.

On the other hand, when the allocation changing unit 233 determines that the size increase request in S202 cannot be met (NO in S202), the allocation changing unit 233 responds to the size increase request of the virtual GPU 211 to the virtual machine 220 that has notified the size increase request in S202. Notify that you cannot be done. In this case, the subsequent S204 is not performed.

(S204: Operation in Virtual Machine 220)
The graphic driver control unit 222 informs that the first communication processing unit 224 has expanded the size of the virtual GPU 211 allocated to the virtual machine 220 from the hypervisor 230, the address information and the size information of the additionally allocated GPU 210, Is notified, the following operation is performed. The graphic driver control unit 222 updates the address information and size information of the virtual GPU 211 managed by the graphic driver installed in the virtual machine 220 based on the notified address information and size information. Specifically, the graphic driver control unit 222 causes the graphic driver to additionally manage the notified address information and size information. As a result, the virtual machine 220 can cause the expanded virtual GPU 211 to perform graphic processing.

As described above, in the present embodiment, when the virtual machine 220 detects an operation caused by the user's stress on the client device 300, the virtual machine 220 notifies the hypervisor 230 of a size increase request for the virtual GPU 211 assigned to itself. To do. Then, the hypervisor 230 expands the size of the virtual GPU 211 corresponding to the size increase request based on the usage rate of the virtual GPU 211 assigned to the virtual machines 220 other than the virtual machine 220 that has notified the size increase request.

From the above, according to this embodiment, it is possible to allocate a resource of an appropriate size.

In the present embodiment, when the client unit 300 detects an operation to reduce the size of the virtual GPU 211, the detection unit 223 issues a size reduction request for the virtual GPU 211 assigned to the virtual machine 220 to the first communication processing unit. The hypervisor 230 may be notified via 224. Here, the operation for requesting the size reduction of the virtual GPU 211 is not particularly limited, such as the operation of inputting a predetermined password.

In addition, in the present embodiment, when there is a resource that is not allocated to any virtual GPU 211 in the GPU 210, the allocation changing unit 233 notifies the virtual machine 220 that has issued the size expansion request of the virtual GPU 211 to the resource May be assigned.

Further, in the present embodiment, the virtual GPU 211 performs the graphic processing of the application executed by the virtual machine 220, but the present invention is not limited to this. The virtual GPU 211 may, for example, perform general-purpose calculation of various programs executed by the virtual machine 220. In this case, the hypervisor 230 may enlarge or reduce the virtual GPU 211 assigned to the virtual machine 220 according to the processing status of the general-purpose calculation by the virtual GPU 211.

In addition, in the present embodiment, the allocation changing unit 233 can perform the following operations in parallel when the allocated virtual GPU 211 has the virtual machine 220 whose size has been reduced in S203. In this case, the allocation changing unit 233 notifies the virtual machine 220 that the size of the allocated virtual GPU 211 has been reduced, and the address information and size information of the reduced size virtual GPU 211. Accordingly, the graphic driver control unit 222 of the virtual machine 220 is notified from the allocation changing unit 233 instead of the address information and size information currently managed by the graphic driver installed in the virtual machine 220. Address information and size information are managed. As a result, the virtual machine 220 can allow the virtual GPU 211, which has been reduced in size, to perform graphic processing.

In addition, in the present embodiment, the allocation changing unit 233 can also perform the following operations in parallel. The allocation changing unit 233 constantly monitors the usage rates of the virtual GPUs 211-1 to 211-n via the monitoring unit 232. The allocation changing unit 233 performs the following operation when the existence of the virtual GPU 211 whose usage rate is equal to or more than a predetermined time and equal to or less than a predetermined threshold is specified among the virtual GPUs 211-1 to 211-n. In this case, the allocation changing unit 233 determines that the resources of the GPU 210 more than necessary are allocated to the specified virtual GPU 211, and reduces the size of the virtual GPU 211. Then, the allocation changing unit 233 notifies the virtual machine 220 to which the virtual GPU 211 is allocated that the size of the virtual GPU 211 has been reduced, and the address information and the size information of the virtual GPU 211 after the size reduction. To do.

The operation of the allocation changing unit 233 will be described with reference to FIG.

(S301)
The allocation changing unit 233 constantly monitors the usage rates of the virtual GPUs 211-1 to 211-n via the monitoring unit 232. When the allocation changing unit 233 specifies the existence of the virtual GPU 211 whose usage rate is equal to or more than a predetermined time and is equal to or less than a predetermined threshold value among the virtual GPUs 211-1 to 211-n, the allocation changing unit 233 reduces the size of the specified virtual GPU 211.

(S302)
The allocation changing unit 233 reduces the size of the allocated virtual GPU 211 with respect to the virtual machine 220 to which the virtual GPU 211 whose size has been reduced in S301 is allocated, the address information of the virtual GPU 211 after the size reduction, and Notify with size information.

By the operation of the allocation changing unit 233, the virtual base host 200 can improve the use efficiency of the resources of the GPU 210.

Although the respective embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and further modification/replacement/replacement may be performed without departing from the basic technical idea of the present invention. Adjustments can be made. Moreover, you may implement combining each embodiment suitably.

It should be noted that the disclosures of the above patent documents are incorporated herein by reference. Modifications and adjustments of the exemplary embodiment are possible within the scope of the overall disclosure (including the claims) of the present invention and based on the basic technical concept of the invention. Further, various combinations or selections of various disclosed elements are possible within the scope of the claims of the present invention. That is, it goes without saying that the present invention includes various variations and modifications that can be made by those skilled in the art according to the entire disclosure including the claims and the technical idea.

10 Client Device 20 Virtual Infrastructure Host 21 Virtual Machine 22 Hypervisor 23 Hardware 30 Virtual Infrastructure Host 31 GPU Resource 32 Virtual Machine 33 Hypervisor 100 Communication System 200 Virtual Infrastructure Host 210 GPU
211 Virtual GPU
220 virtual machine 221 application control unit 222 graphic driver control unit 223 detection unit 224 first communication processing unit 230 hypervisor 231 second communication processing unit 232 monitoring unit 233 allocation changing unit

Claims (8)

A virtual machine,
GPU resources allocated to the virtual machine,
A hypervisor for changing the size of the GPU resource in response to the request for changing the size of the GPU resource from the virtual machine,
When the virtual machine detects a predetermined operation in the information processing device , the virtual machine notifies the hypervisor of the size change request for requesting expansion of the size,
The predetermined operation is at least one of an operation of repeatedly inputting the same instruction within a predetermined time and an operation of repeatedly inputting an instruction in a screen area in which the input in the information processing device is invalidated,
Virtual infrastructure host characterized by the following. There are multiple virtual machines,
The GPU resource is allocated to each of the virtual machines,
When the hypervisor is notified by the first virtual machine of the virtual machines of a size change request for the first GPU resource allocated to the first virtual machine of the GPU resources, A size of the first GPU resource based on a usage rate of one or more second GPU resources of the GPU resources, which is assigned to one or more second virtual machines of the virtual machine. The virtual infrastructure host according to claim 1, wherein the virtual infrastructure host is changed. If the size change request to the first GPU resource is a request to increase the size of the first GPU resource,
When the hypervisor specifies the second GPU resource whose usage rate is equal to or more than a predetermined time and is equal to or less than a predetermined threshold, the hypervisor reduces the size of the identified second GPU resource, and the reduced second GPU is used. The virtual infrastructure host according to claim 2, wherein a resource is allocated to the first GPU resource. When the size change request to the first GPU resource is a request to reduce the size of the first GPU resource,
The virtual base host according to claim 2 or 3, wherein the hypervisor reduces the size of the first GPU resource. The hypervisor reduces the size of the GPU resource when the utilization rate of the GPU resource is equal to or more than a predetermined time and is equal to or less than a predetermined threshold value. The listed virtual infrastructure host. A virtual infrastructure host according to any one of claims 1 to 5,
An information processing device connected to the virtual infrastructure host,
Communications system. A method for controlling a virtual infrastructure host, comprising a virtual machine and a GPU resource allocated to the virtual machine,
When the virtual machine requests to change the size of the GPU resource, the size of the GPU resource is changed according to the request,
When the virtual machine detects a predetermined operation in the information processing device, the virtual machine notifies the hypervisor that changes the size of the GPU resource of the request to request the size increase,
The predetermined operation is at least one of an operation of repeatedly inputting the same instruction within a predetermined time and an operation of repeatedly inputting an instruction in a screen area in which the input in the information processing device is invalidated,
A method for controlling a virtual infrastructure host, which is characterized by the following. A computer of a virtual base host including a virtual machine and a GPU resource assigned to the virtual machine,
When the virtual machine requests a change in the size of the GPU resource, a process of changing the size of the GPU resource in response to the request,
Let it run
The virtual machine, when detecting a predetermined operation in the information processing device, notifies the hypervisor of a size change request that requests the size change requesting the size increase,
The predetermined operation is at least one of an operation of repeatedly inputting the same instruction within a predetermined time and an operation of repeatedly inputting an instruction in a screen area in which the input in the information processing device is invalidated,
A program for a virtual infrastructure host characterized by the following.

Images