KR101090651B1 - Virtual machine monitor and multi-processor system - Google Patents

Abstract

On the virtual machine monitor having the exclusive allocation function of the I / O device, it has an interface for acquiring the physical position information of the I / O device, and uses the obtained physical position information to allocate resources to the virtual server according to the specified policy. Optimize. The virtual machine monitor has an interface for allocating resources according to a given policy (a parameter for deciding what to give priority to resource allocation) with respect to the I / O device, the number of CPUs, and the amount of memory required for the guest OS. In addition, the virtual machine monitor is provided with an interface for appropriately converting and notifying the physical location information of the resource allocated by the virtual machine monitor to the guest OS.

Multiprocessor System, CPU Socket, CPU Core, Memory Controller, Memory Interface, Memory, Hub, Adapter, Service Processor, Console

Description

VIRTUAL MACHINE MONITOR AND MULTI-PROCESSOR SYSTEM}

The present invention relates to an improvement of a virtual calculator in which a virtual server is operated on a physical calculator by a virtual machine monitor to allocate an I / O device to the virtual server.

With the recent progress of semiconductor technology, a multi-core processor that integrates a plurality of cores on one die and a memory controller integrated processor that mounts a memory controller on a processor die have emerged. In order to effectively utilize the integrated calculator resources in this way, there are many movements that reduce costs by aggregating processes distributed in a plurality of servers into one server. The means which becomes effective at the time of aggregation of such a server is a method of operating a some operating system on one server by server division. Server partitioning includes a physical partitioning method that supports partitioning by hardware on a node-by-node or component basis such as a processor (core) or an I / O device, and a logical partitioning method realized by firmware called a hypervisor or a virtual machine monitor. There is this.

In the logical partitioning system, each operating system (guest OS) is executed on a logical processor provided by a virtual machine monitor, and a plurality of logical processors are mapped to physical processors by a virtual machine monitor, thereby partitioning the partition into finer units than nodes. can do. The processor (core) can also be executed by switching one physical processor (core) to time division between a plurality of logical partitions. As a result, it is possible to generate more logical partitions than the number of physical processors (cores) and execute them simultaneously. As a virtual machine monitor software for the purpose of logical division, the technique as described in patent document 1 is typical.

However, the number of I / O devices that are inherently difficult to integrate is not reduced because of the need to have a place of input / output (path) as compared to a processor or a memory, which is relatively easy to integrate, so that the balance between a conventional CPU and an I / O device is poor. There is a tendency to collapse. In order to increase the number of I / O devices, a countermeasure that the slot is expanded by using an I / O switch is considered. However, the distance between the processor, memory, and I / O devices caused by I / O switches has created a case where I / O performance cannot be sufficiently derived.

Therefore, in the past, an approach to ensure sufficient I / O performance by assigning an I / O device shared by a plurality of virtual servers to a specific virtual server has been adopted. As a function of supporting exclusive application of I / O devices by such a virtual server, as disclosed in Non-Patent Document 1, VT-d or the like established by Intel Corporation is known.

On the other hand, with the progress of multicoreization and the emergence of processors incorporating memory controllers, there is a tendency that the arrangement of resources such as processor memory I / O devices is unbalanced. In order to secure performance and reliability on such an unbalanced system, resource allocation using physical location information is required. In a conventional OS, there is a structure called Affinity control that associates a specific processor with a memory, and ACPI (Advanced Configuration and Power Interface) is defined as a standard interface for acquiring physical position information in order to support this (Non-Patent Document 2). ). In this affinity control, resources are allocated based on what CPU and memory the OS or application uses.

[Patent Document 1] US Patent No. 6496847

[Non-Patent Document 1] Intel Virtualization Technology for Directed I / O Architecture Specification, [online], Intel Corp., August 24, 2007 Search, Internet <ftp://download.intel.com/technology/computing/vptech / Intel (r) _VT_for_Direct_IO.pdf>

[Non-Patent Document 2] Advanced Configuration and Power Interface Specification Revision 3.0, [online], Hewlett-Packard, Intel, Microsoft, Phoenix, and Toshiba, retrieved August 24, 2007, Internet <http: //www.acpi. info />

However, in the conventional OS as described above, the position information of the processor and the memory can be controlled using the structure of the ACPI. However, since the I / O device is a resource commonly referred to from the entire application, there is no concept of affinity control using the physical location information of the I / O device. In fact, in the System Resource Affinity Table (SRAT) or System Locality Distance Information Table (SLIT) defined within ACPI, only the physical location information of the processor and memory is targeted, and the physical location information of the I / O device is managed. Is out.

On the other hand, when the virtual machine monitor is used to assign an I / O device to a specific virtual server, the physical location information of the I / O device is an important parameter in securing the performance and reliability of the virtual server. . However, in the conventional ACPI-based interface, there is no means for acquiring I / O physical position information.

Also, even if the virtual machine monitor allocates the appropriate resources to the virtual server, if the guest OS on the virtual server cannot use the physical location information correctly, the affinity control of the guest OS will not function correctly, resulting in physical There was a problem in that performance and reliability equivalent to that of the server could not be secured.

The present invention has an interface for acquiring physical location information of an I / O device on a virtual machine monitor having an exclusive allocation function of an I / O device, and uses the acquired physical location information to allocate resources to a virtual server. The goal is to optimize according to the specified policy. In addition, it has an interface for converting the physical location information acquired by the virtual machine monitor to the virtual server appropriately and informs the virtual server, so that the same affinity control can be executed for the guest OS on the virtual server as when the guest OS is executed on the physical server. The purpose.

The present invention provides a virtual machine monitor that connects one or more processors, one or more memories, one or more I / O devices to an internal network, and allocates the processors, memory, and I / O devices to a virtual server. A multiprocessor system, wherein the virtual machine monitor acquires physical hardware information for acquiring configuration information of hardware including physical location information of hardware including the processor and memory of the multiprocessor system, an I / O device, and a network And a reception unit which receives a generation request including a number of processors and a memory amount of the virtual server to be generated, an allocation policy of I / O devices and resources, and an I / O device based on the received generation request. After assigning to a server, the processor and memory may be added to satisfy the allocation policy. An allocation processing unit for allocating to the server is provided.

The virtual machine monitor further includes a notification unit for notifying the virtual server of the physical location information of the processor, memory and I / O device allocated to the virtual server.

Therefore, the present invention can obtain the physical location information of the I / O device and use the acquired physical location information to optimize the allocation of resources to the virtual server according to the allocation policy of the resources specified by the generation request. do.

In addition, by notifying the virtual server of the physical location information of the resource allocated by the virtual machine monitor, the same control as on the physical server can be realized on the virtual server.

EMBODIMENT OF THE INVENTION Hereinafter, one Embodiment of this invention is described based on an accompanying drawing.

FIG. 1 is a block diagram showing a first embodiment and showing a relationship between a multiprocessor system (calculator) to which the present invention is applied, and a virtual machine monitor and a guest OS operating thereon.

The multiprocessor system 100 connects one or more CPU sockets (processor package) 110, memory controller 130, and I / O hubs 160a, 160b to inter-module connection I / F (interface) 200. The connected configuration is adopted. In addition, the inter-module connection I / F 200 constitutes an internal network of the multiprocessor system 100. Here, the connection between modules between the CPU socket 110 and the memory controller 130, between the CPU socket 110 and the I / O hubs 160a and 160b, and between the memory controller 130 and the I / O hubs 160a and 160b. The I / F 200 need not necessarily be the same I / F. Even if they were different I / F, the following description does not produce a difference. In addition, the generic terms of the I / O hubs 160a and 160b are referred to as I / O hubs 160. In addition, a chip set in which the functions of the memory controller 130 and the I / O hub 160 are mounted may exist. In addition, as shown in FIG. 1, the structure as the memory controller 130 is mounted on the CPU socket 110 may be sufficient. CPU socket 110 includes one or more CPU cores (processor cores) 120. The memory controller 130 is connected to at least one dual inline memory module (DIMM) 150 through the memory I / F 140. I / O hub 160 mounts one or more I / O adapters 180a through 180f through I / O connected I / F 170, and in front of I / O adapters 180a through 180f. 190d is connected. In addition, an I / O bridge or an I / O switch may be provided in front of the I / O connection I / F 170 to configure multiple stages of I / O. In addition, the I / O adapters 180a to 180f are constituted of a NIC, a host bus adapter (HBA), or the like, and the generic name thereof is referred to as the I / O adapter 180.

There is at least one service processor 220 on the multiprocessor system 100 and collects physical location information or connection information of each module (physical component) through the module information acquisition I / F 210. The service processor 220 may be mounted in an external form as an option of the multiprocessor system 100 or may be mounted as one of the I / O devices. In addition, the function of the service processor 220 may be provided on a separate computer connected via a LAN or the like. Moreover, an I / O device is comprised with the apparatus which inputs and outputs data.

On the multiprocessor system 100, a virtual machine monitor 300 is executed to divide resources on the multiprocessor system 100 into one or more virtual servers and provide them to the guest OSs 360a-c. The virtual machine monitor 300 receives the physical configuration information and the connection information of the module collected by the service processor 220 through the physical hardware configuration information acquisition interface 320. In addition, the generic terms of the guest OSs 360a to 360c are referred to as guest OS 360.

The guest OS 360 notifies the virtual machine monitor 300 by the logical hardware request I / F 350 of the resource required as the virtual server. In addition, at the start of the guest OS 360 (that is, at the start of the virtual server), the virtual machine requests the I / F 310 set by the administrator of the multiprocessor system 100 to the console 230. The monitor 300 is notified. After the virtual machine monitor 300 considers the resource allocation policy included in the logical hardware request I / F 350 with reference to the physical-logical hardware allocation table 310 or the like on the virtual machine monitor 300, Information of necessary physical resources is obtained through the physical-logical hardware allocation interface 330. The virtual machine monitor 300 includes a logical hardware allocation processing unit 801 for allocating physical resources as logical hardware based on the acquired physical resource information and the logical hardware request I / F 310. The information of the physical resource allocated by the virtual machine monitor 300 is converted into logical hardware configuration information by the virtual machine monitor 300, and is transmitted to the guest OS 360 through the logical hardware configuration information notification I / F 340. You are notified.

In addition, the multiprocessor system 100 has an input device and an output device, and a console 230 used by an administrator is connected to give a command to the virtual machine monitor 300 or the service processor 220 or to monitor the virtual machine. The processing result is received from the 300 or the service processor 220 and displayed on the output device.

<Configuration of Multiprocessor Systems>

In addition, the structure of the multiprocessor system 100 in this 1st Embodiment is demonstrated in detail. Four CPU sockets 110a to 110d are connected in a ring shape by the inter-module connection I / F 200. Each CPU socket 110 has two CPU cores 120, and there are eight CPU cores 120 throughout the multiprocessor system 100. In the following description, these eight CPU cores may be referred to as serial numbers # 0 to # 7 of the CPU core from the left side of the figure.

The memory controller 130 on each CPU socket 110 has four memory I / Fs 140, respectively, and four DIMMs (memory) 150 are connected to each other. Here, for simplicity of explanation, it is assumed that each DIMM 150 has 1 GB, 4 GB per CPU socket, and 16 GB of memory for the entire multiprocessor system. In the following description, the serial numbers of the 16 DIMMs 150 may be referred to as DIMM # 0 to # 15 from the left side of the drawing.

The I / O hub 160 is composed of two I / O hubs 160a and 160b, each having two inter-module connection I / Fs 200, and the I / O hub 160a is a CPU socket. 110a and 110b and the I / O hub 160b are connected to the CPU sockets 110c and 110d, respectively. I / O hub 160 has four I / O-connected I / Fs 170, each with an I / O slot 175 in front of I / O-connected I / Fs 170, and an I / O adapter 180 is connected.

There are eight I / O slots 175 throughout the multiprocessor system 100. If the serial number of the I / O slot 175 is numbered # 0-7 from the left side of the drawing, I / O slot # 0 Is I / O adapter 180a, I / O slot 180 is I / O adapter 180b, I / O slot # 3 is I / O adapter 180c, and I / O slot # 4 is I The / O adapter 180d is connected to the I / O slot # 5 with the I / O adapter 180e and the I / O slot # 7 with the I / O adapter 180f. In this example, nothing is connected to I / O slot # 1 and I / O slot # 6.

FIG. 64 illustrates a virtual machine monitor 300, guest OS1 360a, guest OS2 360b, and guest OS3 360c running on the multiprocessor system 100 corresponding to the configuration of FIG. It is a schematic diagram. The program of the virtual machine monitor 300 resides on either one of the I / O devices 190 or on the ROM 201 of the service processor 200 and is stored upon startup of the multiprocessor system 300. It is deployed from memory 150 (in this example, memory 150 # 0 in this example) and executed on any one of virtual machine CPU cores 120. CPU core 120 executing virtual machine monitor 300 May be fixed. For example, the CPU cores to be executed may be variable depending on the operating states of the CPU cores, such as empty CPU cores and CPU cores with less processing load.

The guest OS 360 has a program on any one of the I / O devices 190 assigned to the logical server 370 partitioned by the virtual machine monitor 300, and starts up the logical server 370. Is developed on the memory 150 assigned to the reference numeral 370 and executed by the CPU core 300 assigned to the reference numeral 370. In the example of FIG. 64, guest OS1 360a is deployed on memory 150 # 5, guest OS2 360b is on memory 150 # 10, and guest OS3 360c is deployed on memory 150 # 13. It is executed. In addition, in this example, although the guest OS 360 was arrange | positioned on one memory module for the sake of clarity, the case where it distribute | distributes on several memory modules depending on the size of a guest OS and setting of memory interleaving is also shown. There may be.

2 to 6 show the physical hardware configuration information acquisition I / F 320 of the multiprocessor system 100 obtained by the virtual machine monitor 300 via the service processor 220. 2 shows details of the physical hardware configuration information acquisition I / F 320. The physical hardware configuration information acquisition I / F 320 includes four types of tables (physical component scheme 400, I / O adapter scheme 450, inter-component distance correspondence table 500, and physical network scheme 550). Table). These tables are generated (or updated) by querying the information of the hardware resources collected by the service processor 220 through the module information acquisition I / F 210 by the virtual machine monitor 300 at a predetermined period or at a predetermined timing. ) The physical hardware configuration information acquisition I / F 320 set by the virtual machine monitor 300 is stored in the physical location information storage memory 165 or the memory 150 of the I / O hub 160 shown in FIG. 1. do.

3 shows the configuration of the physical component configuration table 400 of the physical hardware configuration information acquisition I / F 320. The physical component scheme 400 includes a resource # 405 indicating a serial number of a resource, a resource type 410 indicating a type of resource, a range 415 indicating a range of resources corresponding to the serial number, and a series of components. A component # 420 representing a number, a component type 425 representing a type of a component, a power consumption 430 of a resource specified by a serial number, and a power consumption 435 of a component for operating the resource.

In the description of the present embodiment, the component # 420 points to an object that is physically inserted or removed, such as the CPU socket 110 or the I / O hub 160, and a plurality of resources are provided. It may be connected on one component # 420. In the example of this embodiment, the CPU core 120 or the DIMM 150 is connected to a component called the CPU socket 110, and the I / O adapter 180 is connected to the component called the I / O hub 160. do. If the memory controller 130 is separate from the CPU socket 110, the chip set is present, or the like, they may also be independent components. The physical component composition table 400 is a table which shows the inclusion relationship of such a resource and a component. Indicates that the resource represented by scope 415 is included in component # 420. When the resource type 410 is a memory, a range of physical addresses is shown in addition to the number of the corresponding DIMM 150. In this example, since 1 GB per 1 DIMM is divided into 4 parts according to the CPU socket to which 16 GB from 0x0_0000_0000 to 0x0_3_FFFF_FFFF belong. The power consumption [W] in the case where each resource is in the active state is indicated by reference numeral 430, and the power consumption in the case where the component based on the resource is in the active state is indicated by reference numeral 435. For example, it is assumed that core # 0 is in operation and core # 1 is not in operation. In this case, the power consumption of the resource is 20 because it is 20 per core, but since the power consumption of the CPU socket of the component on which it is based is 80 according to the physical component scheme 400, the total 20 + 80 = 100 is only core # 0 running. Power consumption when there is one. In this case, in order to reduce the size of the physical component scheme 400, a plurality of resources belonging to the same component are shown as one entry in the range 415, but different entries may be used for each resource.

In addition, the power consumption 430 of the resource and the power consumption 435 of the component use data previously set in the ROM 221 of the service processor 220. The ROM also stores information related to component performance, such as power consumption of the I / O adapter 180, bandwidth and latency between components.

4 shows a configuration of the I / O adapter scheme 450. The I / O adapter scheme 450 includes an I / O adapter # () that represents an entry for every I / O adapter 180 on the multiprocessor system 100 and indicates the serial number of the I / O adapter 180. 455), an I / O adapter 460 indicating an identifier of the I / O adapter, an I / O slot # 465 equipped with this I / O adapter, an adapter type 470 indicating a type of the I / O adapter, Each item of the power consumption 475 of this I / O adapter is comprised. The adapter type 470 describes a type of adapter called a network interface card (NIC) and a host bus adapter (HBA).

5 shows a configuration of the distance correspondence table 500 between components. The component-to-component distance correspondence table 500 includes an entry corresponding to all components on the multiprocessor system 100, a component # 420 indicating an identifier of a component, a component type 425 indicating a type of a component, It consists of a distance 510 between pair components representing the distance between this component and another component. The distance between paired components 510 is also divided by all components and represents the distance from the component of component # 420 to the component of the distance between paired components 510. When there are N components throughout the system, the distance between pair components 510 constitutes a matrix of N × N. Typically, the distance to reach itself is considered zero, so zero is usually arranged in the diagonal portion of this matrix. In this embodiment, since the performances of the inter-module connection I / F 200 are all assumed to be equivalent, the number of times that the target component is reached after passing through the inter-module connection I / F 200 is set as the distance. This distance is treated as physical location information of the component. For example, the distance from the CPU socket 110a to the I / O hub 160b is, as shown in FIG. 1, when the distance from the CPU socket 110a to the CPU socket 110d represents the inter-module connection I / F 200. Since it passes once and passes through the module-to-module connection I / F 200 from CPU socket 110d to I / O hub 160b once, the distance between components totals two.

In addition, when the inter-module connection I / F is nonuniform, a method of representing the distance using the total value of the latency 570 in the physical network configuration table 550 described later is considered.

In addition, in this embodiment, since the memory 150 is directly connected to the CPU socket 110, since the distance between the CPU socket 110 and the memory 150 is 0, it is not included in the distance correspondence table 500 between components. When 150 is connected to the internal network (inter-module connection I / F 200), the memory 150 is added to component # 420, and the distance between each component is added to the distance correspondence table 500 between components. You can set it.

In addition, the distance correspondence table 500 between components can be generated by the virtual machine monitor 300 based on the information of the physical hardware configuration information acquisition I / F 320. In addition, the service processor 220 may generate the distance correspondence table 500 between components based on the information of the physical hardware configuration information acquisition I / F 320 and notify the virtual machine monitor 300.

6 shows a configuration of a physical network configuration table 550. Physical network scheme 550 is a network # 555 representing the serial number of the connection of the inter-module connection I / F 200 for all the inter-module connection I / F 200 on the multiprocessor system 100. And an item indicating which component 420b is connected from which component 420a, and a band 560 and a latency 570 between the component 420a and the component 420b. In the present embodiment, the band 560 of the network connecting between the CPU sockets 110 is 6, whereas the network band 560 connecting the CPU socket 110 and the I / O hub 160 is half. We assumed to be three. The unit of the band 560 is, for example, [Gbps], and the unit of the latency 570 is, for example, [nsec].

The above is the details of the physical hardware configuration information acquisition I / F 320. The service processor 220 collects these pieces of information from each resource component through the module information acquisition I / F 210. The virtual machine monitor 300 obtains the physical hardware configuration information acquisition I / F 320 from the service processor 220. As for the physical network configuration table 550 and the distance correspondence table 500 between components, the information which cannot be obtained only by the service processor 220 inquiring the module alone is determined or initialized in the multiprocessor system 100. At the time of change, a method of automatically detecting a connection relationship using an inquiry protocol through the module information acquisition I / F 210 and the inter-module connection I / F 200, or a system management tool used by an administrator at the time of configuration change. It is conceivable that the service processor 220 or the like store information on the outside of the multiprocessor system 100 and configure it based on the information.

<Configuration of Virtual Server>

7-9, the structure of the logical hardware request I / F 350 of the virtual server 1_370a in 1st Embodiment, and the virtual server corresponding to it is shown.

7 illustrates a logical hardware request I / F requesting a resource required for the startup of the virtual server 1_370a (see FIGS. 9 and 16) to the virtual machine monitor 300 when the guest OS 360 is activated. The configuration of 350a) is shown. The logical hardware request I / F 350a is set by the administrator in the console 230.

The logical hardware request I / F 350a includes a # 351 indicating the serial number of the virtual server, a guest name 352 indicating the identifier of the guest OS 360 running on the virtual server, and an I / A assigned to the virtual server. O adapter 353, number of CPU cores 354 allocated to virtual server, amount of memory 355 allocated to virtual server, resource allocation policy 356 indicating policy of resource allocated to virtual server, priority of virtual server It consists of a diagram 357. The virtual server # 351 is an identifier for identifying the virtual server on the system and does not need to be set by the requesting guest. The guest name 352 is information for identifying each guest OS 360. For example, the type of the guest OS 360 (Windows (registered trademark) / Linux, etc.) is also included therein. I / O adapter 353 is a list of I / O adapters 180 required by the guest OS. I / O adapters 180a and 180c are required here. The number of CPU cores 354 is the number of CPU cores required by the guest OS 360a. This requires four cores. The memory amount 355 is a memory amount required by the guest OS 360a. 4GB is required here. The resource allocation policy 356 represents a policy when allocating resources, and is a key parameter in the present invention. As the resource allocation policy 356, the following policy is considered, for example.

Performance First: Choose an arrangement that will shorten the distance between components.

CPU-memory priority: Make the distance between the CPU and memory shorter (valid when the CPU has a lot of memory access)

CPU-I / O priority: Make the distance between the CPU and I / O devices shorter (valid when there are many I / O interrupts)

I / O device-memory priority: Make the distance between the I / O device and the memory shorter (valid when there are many DMA transfers from the I / O device).

CPU-I / O device-memory priority: Place each resource close together (when balancing the performance of the entire virtual server).

Reliability Priority: Reduces the number of common components and networks between virtual servers.

Band Priority: Increases the effective band of the network

Power saving priority: Reduces power consumption throughout the system

In the following description of this embodiment, an example of the policy to set is CPU-I / O priority among performance priority. However, since the virtual server 1 360a is the first virtual server on the multiprocessor system 100, it is highly likely that the requested resource can be obtained basically. Priority 357 is a parameter used to determine which virtual server to give priority when there is a conflict between resources requested between virtual servers. As an example of showing the priority, it is assumed that the integer is higher and the higher the value is, the higher the priority is. However, in the first embodiment, it is assumed that all virtual servers have the same priority.

In the present embodiment, only one policy is provided to the virtual server. However, among the configurations that satisfy the primary policy and provide a primary policy with a higher priority and a secondary policy with a lower priority, the secondary policy is also possible. You may select a configuration that satisfies you. Also contemplated are interfaces that provide numerical weighting for multiple policies. Here, for the sake of simplicity, only one policy is assumed.

FIG. 8 corresponds to a result of the logical hardware allocation processing unit 801 of the virtual machine monitor 300 assigning physical resources based on the logical hardware request I / F 350a of the virtual server 1_370a shown in FIG. The configuration and contents of a physical-logical hardware assignment table 310 are shown. In addition, the process which the virtual machine monitor 300 allocates a hardware resource is explained in full detail later in FIGS. 55-63.

The physical-logical hardware allocation table 310 includes a virtual server # 351 representing the serial number of the virtual server, an on / off 311 indicating the startup state of the virtual server, and a resource representing the resource used by the virtual server. The resource # includes an I / O adapter 312, a CPU core 313, a memory 314, a use component # 315, and a use network # 316. In this embodiment, the I / O adapters # 1, 3, CPU cores # 0, 1, 2, 3, DIMMs # 4, 5, 6, and 7 are replaced with virtual servers in accordance with the logical hardware request I / F 350a. 1_370a). In the memory 314 of the used resource #, corresponding physical addresses range from 0x1_0000_0000 to 4 GB (this 0x1_0000_0000 is sometimes referred to as the base address of the virtual server 1). In this example, 110a, 110b, and 160a are set to the use component # 315 as identifiers of the components loaded with these resources, and # 1, 5, and 6 are set to the use network # 316 as the network # to be used for connection. Indicates.

The specific procedure for generating the physical-logical H / W allocation table 310 of FIG. 8 from the logical hardware (HW) request I / F 350 of FIG. 7 will be described later.

9, based on the logical hardware request I / F 350a and the physical-logical hardware (HW) allocation table 310, the logical hardware allocation processing unit 801 of the virtual machine monitor 300 is a multiprocessor system ( The virtual server 1_370a allocated on 100 is shown. A portion enclosed by a dotted line in the drawing represents resources allocated to the virtual server 1_370a among the resources of FIG. 1.

<Operation in the First Embodiment>

10 to 17, the operation of the first embodiment will be described.

10 and 11 show the logical hardware for the next virtual server 2 after the virtual server 1_370a is allocated as shown in FIG. 9 by the logical hardware allocation processing unit 801 of the virtual machine monitor 300. The required I / Fs 350b and 350c are shown. The logical hardware request I / Fs 350b and 350c both require the same I / O adapter 353, the number of CPU cores 354, and the memory amount 355. The difference is that only the resource allocation policy 356 is "CPU-memory priority" in the logical hardware request I / F 350b, whereas "I / O-memory priority" is in the logical hardware request I / F 350c. It is. In the following example, different resources are allocated by the difference of these policies.

12 and 13 show a result of the logical hardware allocation processing unit 801 of the virtual machine monitor 300 assigning to the logical hardware request I / F 350 shown in FIGS. 10 and 11 for the virtual server 2. Two physical-logical hardware assignment tables 310 are shown. Both I / O adapters 312 are identical to # 2, but use CPU cores 313 and memory 314 are different. In the physical-logical hardware assignment table 310b of FIG. 12, the CPU cores # 4, 5 and the DIMMs # 8, 9, 10, and 11 are assigned. In the physical-logical hardware assignment table 310c of FIG. CPU cores # 6, 7 and DIMM # 0, 1, 2, 3 are allocated. In addition to these two methods, resource allocation methods are conceivable, but the allocation method that essentially does not change is excluded, and candidates of several allocation methods are compressed.

In the logical hardware request I / F 350b of FIG. 10, since the CPU-memory priority policy is requested, the CPU cores # 4 and # 5 of the CPU socket 110c and the memory connected to the CPU socket 110c. By assigning # 8 to 11 to the virtual server 2 and setting the resource distance to 0, the short distance between the CPU and the memory is selected. On the other hand, since the logic hardware request I / F 350c shown in Fig. 11 requires an I / O-memory priority policy, the I / O adapter # 2 is provided so that the distance between the I / O device and the memory is shortest. The nearest memory # 0 to # 3 is selected. In this way, the logical hardware allocation processing unit 801 of the virtual machine monitor 300 can optimize the hardware resources allocated to the virtual server according to the requested policy.

FIG. 14 shows an inter-resource distance calculation table 600b in which the virtual machine monitor 300 calculates the distance between each resource, corresponding to the allocation table 310b shown in FIG. 12. The resource-to-resource calculation table 600b includes a virtual server number # 351 representing an identifier (or serial number) of the virtual server, and for each virtual server number # 351, for calculating the distance between resources in the virtual server. Table. In the category 601, there are three types of categories corresponding to the resource allocation policy: CPU-memory, I / O-memory, and CPU-I / O. The virtual machine monitor 300 according to the component # and the distance correspondence table 500 between components belonging to the component 601 from the resource from 602 to the resource to 603 for each category 601. Calculate In this example, since CPU cores # 4, # 5 and DIMMs # 8, 9, 10, and 11 are all mounted on the same component 110c, the distance between components 604 becomes zero. On the other hand, the distance between the component 160a on which I / O adapter # 2 is mounted and the component 110c on which DIMMs # 8, 9, 10, and 11 are mounted is 2 according to the distance correspondence table 500 between components. The distance 604 is two. The distance between CPU cores # 4, 5 and I / O adapter # 2 is likewise equal to two. Finally, the sum 605 of the distances between components for each category 601 is calculated. In this example, the CPU-memory is 0, the I / O-memory is 2, and the CPU-I / O is 4.

FIG. 15 shows an inter-resource distance calculation table 600c corresponding to the physical-logical hardware assignment table 310c shown in FIG. Since the distance between the component 110d on which the CPU cores # 6 and # 7 are mounted and the component 110a on which the DIMMs # 0, 1, 2, and 3 are mounted is 1, the inter-component distance 604 becomes one. Meanwhile, since the distance between the component 160a on which the I / O adapter # 2 is mounted and the component 110a on which the DIMMs # 0, 1, 2, and 3 are mounted is 1, the distance between components is 604. Finally, since the distance between the component 110d on which the CPU cores # 6 and 7 are mounted and the component 160a on which the I / O adapter # 2 is mounted is two, the distance between components 604 becomes two. The total of the distances between the components 605 is 2 for CPU-memory, 1 for I / O-memory and 4 for CPU-I / O.

The candidate which satisfies the resource allocation policy most is selected from the resource allocation candidates which calculated the distance-to-resources distance calculation table 600 as mentioned above. In the case of "CPU-memory priority" at 350b, a resource allocation is selected in which the value of the CPU-memory of the total 605 of the component-to-component distances becomes small. In contrast, in the case of "I / O-memory priority", resource allocation is selected in which the value of the I / O-memory of the total 605 of the distances between components becomes small. When the physical-logical hardware allocation tables 310b and 310c are compared, resource allocation of reference numeral 310b for "CPU-memory priority" and reference numeral 310c for "I / O-memory priority" is selected, respectively. .

FIG. 16 is a configuration diagram of the virtual server 2_370b corresponding to the physical-logical hardware assignment table 310b shown in FIG. The virtual server 2_370b is composed of a CPU socket 110c, a memory connected thereto, and an I / O adapter 180b, as indicated by a dashed line in the figure.

FIG. 17 is a configuration diagram of a virtual server corresponding to the physical-logical hardware assignment table 310c shown in FIG. As shown by a dashed line in the figure, the virtual server 2_370b includes CPU cores # 6, 7 on the CPU socket 110d, DIMMs # 0, 1, 2, 3, and I / O on the CPU socket 110a. It consists of an adapter 180b.

According to the above example, even if the same I / O adapter, the number of CPU cores, and the memory amount are requested due to the difference in the resource allocation policy of the logical hardware request I / F 350, different resources can be allocated. The administrator can configure the desired virtual server automatically.

<Modification 1>

Subsequently, a first modification of the first embodiment will be described. 18-20, the logical hardware request I / F 350d of the virtual server 1_370a in the modification 1 is shown in FIG. 18, and from this logical hardware request I / F 350d, a virtual machine monitor ( The physical-logical hardware assignment table 310d as a result of the assignment by the logical hardware assignment processing unit 801 of 300 is shown in FIG. 19, and the configuration of the virtual server 1_370a is shown in FIG. Requiring reference numerals 180a and 180c as the I / O adapter is the same as in the first embodiment described above. However, unlike the first embodiment described above, the virtual server 1 requires two CPU cores and 8 GB of memory. In response to this request, the logical hardware allocation processing unit 801 of the virtual machine monitor 300 allocates resources as shown in Figs. 19 and 20 from the CPU-I / O priority policy, which is the request policy.

Subsequently, after the allocation of the virtual server 1 shown in Figs. 18 to 20 in Figs. 21 to 28, the resource request of the virtual server 2 is made by the logical hardware request I / F 350e shown in Fig. 21 and shown in Fig. 22. The case where the resource request of the virtual server 2 by the logical hardware request I / F 350f is performed is shown. Virtual server 2 requests the I / O adapter 180b, the number of CPU cores 2, and 4 GB of memory in the same manner as in the first embodiment. In contrast, in the logical hardware request I / F 350e of FIG. 21, "CPU-memory priority" as the resource allocation policy 356, and "CPU-I / O priority" as the resource allocation policy 356 at 350f. Is specified.

23 and 24 are examples of different resource allocations for the virtual server 2 of the physical-logical hardware allocation tables 310e and 310f of FIGS. 21 and 22, respectively. The logical hardware allocation processing unit of the virtual machine monitor 300 is shown. 801 indicates the result of the assignment. In the physical-logical hardware assignment table 310e, the CPU core and the memory are allocated on the same CPU socket 110c. In the physical-logical hardware assignment table 310f, the CPU core is located on the CPU socket 110a. The DIMM is allocated on the CPU socket 110d.

25 and 26 show the results of calculating the distance calculation tables 600e and 600f between resources according to the allocation of the physical-logical hardware allocation tables 310e and 310f. In the inter-resource distance calculation table 600e shown in FIG. 25, the sum of the inter-component distances 605 is 0 for CPU-memory, 2 for I / O-memory, and 4 for CPU-I / O. On the other hand, in the inter-resource distance calculation table 600f of FIG. 26, the sum of the distances between components is 605 for CPU-memory, 2 for I / O-memory, and 2 for CPU-I / O. As a result, in the case of "CPU-memory priority", the assignment of the reference numeral 310e having a small sum 605 of the component-to-component distances of the CPU-memory is CPU-I / O in the case of "CPU-I / O priority". The allocation of reference numeral 310f with a small sum 605 of the inter-component distances is selected.

FIG. 27 is a configuration diagram of a virtual server corresponding to the physical-logical hardware allocation table 310e of FIG. 23. The virtual server 2_370b is composed of CPU cores # 4, 5, DIMMs # 8, 9, 10, 11, and an I / O adapter 180b on the CPU socket 110c.

FIG. 28 is a configuration diagram of a virtual server corresponding to the physical-logical hardware allocation table 310f of FIG. 24. The virtual server 2_370b is composed of CPU cores # 0, 1 on the CPU socket 110a, DIMMs # 12, 13, 14, 15 on the CPU socket 110d, and the I / O adapter 180b.

The above is the first modification of the first embodiment.

<Modification 2>

Subsequently, a second modification of the first embodiment will be described. 29 to 31 show the logical hardware request I / F 350g of the virtual server 1_370a in the second modified example in FIG. 29, and the physical-logical hardware allocation table 310g in FIG. , And the configuration of the virtual server 1_370a are shown in FIG. The requirements of the reference numerals 180a and 180c as the I / O adapter are the same as those of the first embodiment described above, but only one CPU core and only 2 GB of memory are required. Further, as the policy, the CPU-memory-I / O priority is used to select an arrangement in which the distance between the components is as close as possible. In this case, since it is the first virtual server, the CPU socket 110a close to the I / O hub 160a to which the necessary I / O adapters 180a and 180c are connected is first selected, and the CPU core and DIMM on the reference numeral 110a are selected. By assigning to this policy can be satisfied.

32 to 34 show three logical hardware request I / F 350s that differ only in the resource allocation policy 356 for virtual server2. In common, the request I / O adapter 353 has the same reference numeral 180b, the number of required CPU cores 354 is 1, the request memory amount 355 is 2GB, and the priority is 5. The logical hardware request I / F 350h of FIG. 32 designates "reliability priority" as the resource allocation policy 356. FIG. The logical hardware request I / F 350i in FIG. 33 designates "band priority" as the resource allocation policy 356. FIG. The logical hardware request I / F 350j in FIG. 34 designates "power saving priority" as the resource allocation policy 356. FIG. In the following, it will be shown which configuration is selected according to each policy.

35 to 36 show a physical-logical hardware allocation table 310 corresponding to two resource allocations. In the physical-logical hardware assignment table 310h of FIG. 35, CPU cores # 2, DIMMs # 4, and 5 on CPU sockets 110b different from those used by virtual server 1 are allocated as CPU cores and memories. The first priority of reliability is satisfied. On the other hand, in the physical-logical hardware allocation table 310i of FIG. 36, CPU cores # 1, DIMMs # 2, and 3 on the same CPU socket 110a as those used by the virtual server 1 are allocated, and the requested bandwidth is given priority. We are satisfied with policy.

If the allocation policy is not the same distance priority as the above-described embodiment, a separate index is required to replace the inter-resource distance calculation table 600 as a reference for resource allocation.

37 to 38 show a component network assignment table 650 which replaces the distance calculation table 600 between the resources. The component network allocation table 650 is a table for the logical hardware allocation processing unit 801 of the virtual machine monitor 300 to check a component, a network, and an effective network band shared among a plurality of virtual servers. A virtual server # 351 representing an identifier or serial number of a public server; a public component # 651 representing an identifier of a component shared between virtual servers; a public identifier representing a network used to use a shared component. Corresponding to both the items of the network # 652 and the networks used in each virtual server, each item is referred to as the network # 653, the common number 654, and the effective band 655.

The component network assignment table 650h of FIG. 37 is a table corresponding to the resource allocation of the physical-logical hardware assignment table 310h of FIG. 35. The public component # 651 corresponds to the I / O hub 160a, and no public network # 652 exists. Since the network # 653 is not shared with each network # 653, the common number 654 is one. The effective band 655 is a value obtained by dividing the value of the band 560 corresponding to each network # 555 in the physical network configuration table 550 by the common number 654 (that is, the number of virtual servers). In this case, since the common water is 1, the value of the band 560 of the network becomes the effective band 655 as it is.

The component network assignment table 650i of FIG. 38 is a table corresponding to the resource allocation of the physical-logical hardware assignment table 310i of FIG. 36. Common component # 651 corresponds to I / O hub 160a and CPU socket 110a. In addition, as public network # 652, network # 5 which connects I / O hub 160a and CPU socket 110a is corresponded. In this case, since the common number 654 is two, the effective band 655 is half the value of the band 560 of the original network.

If the resource allocation policy 356 is &quot; reliability priority &quot;, a configuration is selected such that the number of shared components and the number of public networks become as small as possible. In this case, since the I / O adapters required by the virtual server 1 and the virtual server 2 are on the same I / O hub 160a, it is impossible to set the number of shared components to zero. However, when the component network assignment tables 650h and 650i corresponding to the physical-logical hardware assignment tables 310h and 310i are compared, the one corresponding to the physical-logical hardware assignment table 310h shown in FIG. You can see that the number of public components is small. Therefore, for the request of the logical hardware request I / F 350h of Fig. 32 in which the requested policy is &quot; reliability priority, &quot; the allocation of the physical-logical hardware assignment table 310h shown in Fig. 35 is selected.

When the resource allocation policy 356 is "band priority", a configuration is selected such that the resource allocation policy 356 becomes large so as to become the effective band of the network. In this case, when the component network allocation table 650h and 650i of FIG. 37 are compared, the effective band is larger in the structure of the component network allocation table 650h. Therefore, for the request of the logical hardware request I / F 350i whose policy is "band priority", allocation of the physical-logical hardware assignment table 310h of FIG. 37 in which the effective band is maximized is selected.

FIG. 39 shows an allocation resource power consumption calculation table 700h corresponding to the physical-logical hardware allocation table 310h of FIG. 35 in which the requested policy is used in the case of power saving priority. This table is configured by the virtual machine monitor 300 for resources and components that are used throughout the system, not in virtual server units. The category 701 indicating whether the target hardware is a resource or a component, and the resource type / component type The reference numeral 702 indicates a used resource / component # 703 indicating an identifier of a used resource or component, a power consumption 704 indicating power consumption for each resource, and a total 705 of power consumption. Examples of resources include CPU cores, memory, I / O slots, and I / O adapters. The resources used are enumerated. In addition, the power consumption 704 sets the respective values with reference to the resource power consumption 430 of the physical component scheme 400 and the power consumption 475 of the I / O adapter scheme 450. . When the category 701 is an entry of a component, the components required when using each resource are enumerated, and the power consumption 704 of the component is set according to the component power consumption 435 of the physical component configuration table 400. The total power consumption 705 is calculated by the virtual machine monitor 300 as the total power consumption 704.

40 shows an allocation resource power consumption calculation table 700i corresponding to the physical-logical hardware allocation table 310i of FIG. Since the number of resources in use does not change in the physical-logical hardware allocation tables 310h and 310i, the item of resources is mostly the same, although the number # of the resources used is different. The difference is that the components used are three components in total in the physical-logical hardware assignment table 310h, but the components used in the physical-logical hardware assignment table 310i are two in total. For this reason, when the sum total 705 of power consumptions is compared, it is set to 265 [W] in the reference | standard 700i about the value of 345 [W] in the allocation resource power consumption calculation table 700h of FIG.

When the resource allocation policy 356 of the logical hardware request I / F 350 is "power saving priority", the virtual machine monitor 300 selects an allocation method such that the total power consumption 705 becomes smaller as follows. . Therefore, for the request of the logical hardware request I / F 350j of FIG. 34 whose policy is "power saving priority", the logical hardware allocation processing unit 801 of the virtual machine monitor 300 performs the physical-logic of FIG. The allocation of the hardware allocation table 310i is selected.

FIG. 41 is a configuration diagram of a virtual server corresponding to the physical-logical hardware assignment table 310h of FIG. 36.

FIG. 42 is a configuration diagram of a virtual server corresponding to the physical-logical hardware configuration table 310i of FIG. 37.

The above is the second modification of the first embodiment.

In addition, in the case of the policy that prioritizes power consumption in the above, the heat generation amount of each resource and component may be used instead of power consumption.

<Specific Procedure of Resource Allocation in First Embodiment>

Here, the procedure for generating the physical-logical hardware configuration table 310 from the logical hardware configuration request I / F 350 in the first embodiment will be described in detail with reference to FIGS. 55 to 62.

FIG. 55 shows a flowchart of the entire logical hardware allocation process performed by the logical hardware allocation processing unit 801 of the virtual machine monitor 300. First, the I / O adapter requested to the guest OS 360 is allocated, further selected and allocated among unallocated CPU memories, the evaluation is performed according to the requested policy, and the CPU / memory closest to the requested policy is assigned. The procedure of selecting a combination is performed.

First, the process proceeds to step 800. The logical hardware allocation processing unit 801 of the virtual machine monitor 300 secures an entry of the virtual server 351 to the physical-logical hardware allocation table 310. Proceed to step 805.

Step 805 determines whether the allocation request of the I / O adapter 353 requested by the logical hardware request I / F 350a is exclusive or shared. If the result of the determination is exclusive, the process proceeds to step 810 and, if shared, to step 830.

Step 810 determines whether the requested I / O adapter 353 is assignable. If it is already assigned to another virtual server and cannot be allocated, the process proceeds to step 820 and, if assignable, to step 830.

Step 820 is a step for responding to the error to the logic hardware request I / F 350. After the error response, the flow proceeds to step 890.

Step 830 is an I / O adapter assignment process. The flow of the subroutine of I / O adapter assignment processing 830 is shown in FIG. After the completion of this subroutine, the flow proceeds to step 840.

Step 840 is a step of setting the allocation candidate CPU 841 and the allocation candidate memory 842 to the empty set φ. Here, the allocation candidate CPU 841 and the allocation candidate memory 842 represent candidates for the combination of the CPU and the memory closest to the allocation policy among the combinations up to that point. At this point, since the CPU and memory have not yet been selected, the empty set φ is assumed. Proceed to step 850.

Step 850 is a process of selecting a CPU memory that satisfies the request from the unassigned CPU memory. The flow of the subroutine of the CPU / memory selection process 850 is shown in FIG. At the completion of this subroutine, if there is an error response, the flow proceeds to step 820. Otherwise, the flow proceeds to step 860.

Step 860 is a step for evaluating the CPU memory selected in step 850 according to the allocation policy. The flow of the subroutine of policy evaluation 860 is shown in FIG. 58. After completion of this subroutine, the flow advances to step 870.

Step 870 is for determining if there is still an unallocated CPU / memory combination. If it still remains, the process returns to step 850. Otherwise, the process proceeds to step 880.

Step 880 is a process of allocating the allocation candidate CPU 841 and the allocation candidate memory 842 to the virtual server 351 of the physical-logical hardware assignment table 310. The flow of the subroutine of the CPU / memory allocation process 880 is shown in FIG. When this subroutine is completed, the logical hardware allocation process is complete.

Step 890 terminates the entry of the virtual server 351 of the physical-logical hardware allocation table 310 when an error occurs in the middle of the allocation in step 850 (or when the resources allocated in step 810 are insufficient). to be. After the end of this entry, the logical hardware allocation process is complete.

56 shows the flow of the subroutine of the I / O adapter allocation process 830. FIG. First, the process proceeds to step 900.

Step 900 selects an I / O adapter # 455 corresponding to the I / O adapter 353 of the logical hardware request I / F 350a from the I / O adapter scheme 450, and selects the physical-logical hardware scheme. A step of adding the entry of the virtual server 351 of the 310 to the use I / O adapter 312. Proceed to step 901.

In step 901, the I / O slot # 465 corresponding to the I / O adapter 353 is selected from the I / O adapter scheme 450, and the resource type 410 of the physical component scheme 400 is set to I / O adapter # 53. Find the entry corresponding to I / O slot # 465 from the entry in the O slot, and place the corresponding component 420 into the usage component 315 of the entry of the virtual server 351 of the physical-logical hardware scheme 310. It is step to add. Proceed to step 902.

Step 902 is a step of determining whether all the I / O adapters 353 have been allocated. If no adapters have yet been allocated, the process returns to step 900. Otherwise, I / O adapter assignment processing 830 is complete.

Fig. 57 shows the flow of the subroutine of the CPU / memory selection process 850. Figs. First, the process proceeds to step 905.

Step 905 is a step of determining whether the number of CPU cores 354 and the memory amount 355 requested in the logical hardware request I / F 350a can be allocated. If unable to assign, the flow proceeds to step 907. If so, the flow proceeds to step 906.

In step 906, among the unallocated combinations, a CPU core and a memory satisfying the number of CPU cores 354 and the memory amount 355 requested in the logical hardware request I / F 350a are selected, and the assigned CPU ( 851) and the provisional allocation memory 852. As a result, the CPU / memory selection process 850 is completed.

Step 907 is an error response step. As a result, the CPU / memory selection process 850 is completed.

58 shows the flow of the subroutine of policy evaluation 860. First, the process proceeds to step 910.

Step 910 is a step of performing conditional branching in accordance with the allocation policy 356 of the logic hardware request I / F 350a. If the allocation policy 356 is one of CPU-memory priority, CPU-I / O priority, I / O-memory priority, and CPU-I / O-memory priority, the flow proceeds to step 920. If the policy 356 is a reliability priority, step 930 is reached. If policy 356 is band-first, flow proceeds to step 940. If the policy 356 is power saving priority, the flow proceeds to step 950.

Step 920 is the process of calculating the distance between components. Fig. 59 shows the flow of the subroutine. After completion of this subroutine, the process proceeds to step 911.

Step 930 is component sharing number calculation processing. Fig. 60 shows the flow of the subroutine. After completion, proceed to step 911.

Step 940 is an effective band calculation process. Fig. 61 shows the flow of the subroutine. After completion, proceed to step 911.

Step 950 is power consumption calculation processing. Fig. 62 shows the flow of the subroutine. After completion, proceed to step 911.

Step 911 is a step of determining whether the allocation candidate CPU 841 and the allocation candidate memory 842 are empty sets?. In the case of the empty set?, The flow advances to step 913. Otherwise, the flow advances to step 912.

Step 912 is a step of determining whether the provisional allocation policy value 853 is smaller than that of the assignment candidate policy value 843. Here, the policy value is an index for quantitatively evaluating the configuration of the allocated resource component according to each allocation policy 356, and a smaller value is defined to be closer to the requested policy. If the provisional allocation policy value 853 is smaller than the assignment candidate policy value 843, the flow proceeds to step 913. Otherwise, policy evaluation 860 is complete.

In step 913, the provisional allocation CPU 851, the provisional allocation memory 852, and the provisional allocation policy value 853 are assigned to the allocation candidate CPU 841, the allocation candidate memory 842, and the allocation candidate policy value 843, respectively. Assignment step. By this process, the combination of CPU / memory of the combination with the smallest policy value among allotted so far is maintained in the allocation candidate CPU 841, the allocation candidate memory 842, and the allocation candidate policy value 843. By this process, policy evaluation 860 is completed.

In addition, in step 913 of FIG. 58, the assignment candidate policy value 843 is set in the form of inheriting the value of the provisional assignment policy value 853. In the first step 912, the initial value of the assignment candidate policy value 843 is set. Becomes negative, but the first time in step 840 of FIG. 55, since the allocation candidate CPU 841 and the allocation candidate memory 842 are set to?, The determination in step 911 always goes to step 913, and step 912 does not pass. Do not. The assignment candidate policy value 843 may be initialized to a predetermined maximum value.

59 shows a flow of the subroutine of the inter-component distance calculation 920. First, the process proceeds to step 921.

In step 921, the I / O adapter 353 of the virtual server 351 of the logical hardware request I / F 350a, the provisional allocation CPU 851, and the provisional allocation memory 852 are used. It is a step of creating the distance calculation table 600. Specifically,

(1) The component to which each resource belongs is obtained from the physical component scheme 400

(2) The distance between each component is set to the distance between components 604 of the distance calculation table 600 according to the distance correspondence table 500 between components.

(3) The sum of distances 604 between components for each category 601 is calculated and substituted into Σ 605.

It consists of a process called. Next, the flow advances to step 922.

Step 922 is a step of performing conditional branching in accordance with the allocation policy 356 of the logical hardware request I / F 350a. If the policy 356 is CPU-memory priority, the flow proceeds to step 923. If the policy 356 is of CPU-I / O priority, proceed to step 924. If policy 356 is an I / O-memory priority, step 925 is reached. If the policy 356 is of CPU-I / O-memory priority, then step 926 is reached.

Step 923 is a step of setting the Σ 605 corresponding to the CPU-memory of the category 601 of the inter-resource distance calculation table 600 as the provisional allocation policy value 853. As a result, the distance calculation 920 between components is completed.

Step 924 is a step of setting Σ 605 corresponding to the CPU-I / O of the category 601 of the inter-resource distance calculation table 600 as the provisional allocation policy value 853. As a result, the distance calculation 920 between components is completed.

Step 925 is a step of setting Σ 605 corresponding to the I / O-memory of the category 601 of the inter-resource distance calculation table 600 as the provisional allocation policy value 853. As a result, the distance calculation 920 between components is completed.

Step 926 is a step in which the sum of Σ 605 of all categories 601 of the inter-resource distance calculation table 600 is added to the allocation policy value 853. As a result, the distance calculation 920 between components is completed.

60 shows the flow of the subroutine of the component common water calculation 930. First, the process proceeds to step 931.

Step 931 includes all allocations, including the I / O adapter 353 of the logical hardware request I / F 350a, the provisional allocation CPU 851 and the virtual server 351 using the provisional allocation memory 852. It is a step which creates the component network assignment table 650 about the completed virtual server. Specifically,

(1) The component to which each resource belongs is obtained from the physical component scheme 400

(2) Set a common component between different virtual servers in the common component # 651.

(3) Set the number of shared components # to the common number 654.

It consists of a process called. Proceed to step 932.

Step 932 is a step in which the common number 654 corresponding to the virtual server 351 of the component network assignment table 650 is added to the allocation policy value 853. This completes the component common water calculation 930.

61 shows the flow of the subroutine of the effective band calculation 940. First, the process proceeds to step 941.

Step 941 allocates all of the allocations, including the I / O adapter 353 of the logical hardware request I / F 350a, the provisional allocation CPU 851 and the virtual server 351 using the provisional allocation memory 852. It is a step which creates the component network assignment table 650 about the completed virtual server. Specifically,

(1) The component to which each resource belongs is obtained from the physical component scheme 400

(2) A network used between components is obtained from the physical network configuration table 550 and set in the network # 653.

(3) Set the network # 653 to be shared between different virtual servers in the public network # 652.

(4) The band of each network is obtained from the band 560 of the physical network configuration table 550, and the value obtained by dividing this by the number of shares 654 is set in the effective band 655.

Processing is included. Proceed to step 942.

Step 942 is a step of setting the negative allocation policy value 853 to the negative of the effective band 655 corresponding to the virtual server 351 of the component network allocation table 650. The smaller the policy value 853 is defined here, the closer to the required policy. On the other hand, the larger the absolute value of the effective band is, the better. Therefore, by assigning a negative value to an effective allocation policy value 853, the configuration having the largest effective band is finally selected. As a result, the execution band calculation 940 is completed.

62 shows a flow of the subroutine of power consumption calculation 950. First, the process proceeds to step 951.

Step 951 includes all allocations, including the I / O adapter 353 of the logical hardware request I / F 350a, the provisional allocation CPU 851 and the virtual server 351 using the provisional allocation memory 852. It is a step of calculating the allocation resource consumption power calculation table 700 about the completed virtual server. Specifically,

(1) The component to which each resource belongs is obtained from the physical component scheme 400

(2) From the physical component scheme 400, the power consumption is calculated for the resource consumption power 430 and the components of all allocated resources, and set in the power consumption 704 of the calculation table 700.

(3) Calculate the sum of all power consumptions 704 and set it to Σ 705

It includes processing called. Proceed to step 952.

Step 952 is a step of setting the total power consumption 705 of the allocated resource power consumption calculation table 700 as the provisional allocation policy value 853. As a result, power consumption calculation 950 is completed.

63 shows the flow of the subroutine of the CPU / memory allocation processing 880. First, the process proceeds to step 960.

Step 960 adds the allocation candidate CPU 841 to the use CPU core 313 of the entry of the virtual server 351 of the physical-logical hardware scheme 310 and the allocation candidate memory 842 to the use memory 314. It is a step to do it. Proceed to step 961.

Step 961 finds an entry corresponding to the CPU core 313 from an entry in which the resource type 410 of the physical component scheme 400 is the CPU core, and replaces the corresponding component 420 in the physical-logical hardware scheme 310. The step of adding to the usage component 315 of the entry of the virtual server 351, and similarly to the usage component 315 of the entry corresponding to the memory 314 from the entry whose resource type 410 is a memory. Proceed to step 962.

Step 962 is a step of determining whether all allocation candidate CPU 841 and allocation candidate memory 842 have been allocated. If unallocated allocation candidate CPU 841 or allocation candidate memory 842 still remains, step 960 is reached. Otherwise CPU / memory allocation processing 880 is complete.

By the above-described series of operations, the virtual machine monitor 300 can create a physical-logical hardware assignment table 310 from the logical hardware request I / F 350 instructed by the console 230.

As described above, according to the first embodiment, the virtual machine monitor 300 obtains the distance between components indicating the physical position information of the components in the multiprocessor system 100, and advances the distance correspondence table 500 between components in advance. In addition, the physical component configuration table 400, the I / O adapter configuration table 450, and the physical network configuration table are set in advance by acquiring the configuration of the physical network and power consumption of each resource. The virtual machine monitor 300 receives the logical hardware request I / F 350, which is a request for creation (or setup) of a virtual server, from the console 230, and first, requests the requested I / O adapter 180. The exclusive allocation of the I / O adapter 180 is made to the requested virtual server. Next, the virtual machine monitor 300 refers to the virtual server from the distance correspondence table 500 or the physical component configuration table 400 with reference to the physical location information and power consumption of the resource, and performs exclusive allocation. By selecting a CPU and a memory satisfying the required policy from the device and assigning it to the virtual server, it is possible to provide an optimal virtual server for the required policy.

In particular, when running a plurality of virtual servers (guest OS) in one physical calculator, even if the policies are different for each virtual server, optimal configuration for each virtual server from the physical location information of the I / O device, CPU socket, and memory It is possible to automatically assign the, thereby improving the availability of the virtual calculator.

Second Embodiment Logical Hardware Configuration Notification I / F

43 to 48, a second embodiment of the present invention will be described. In addition, the structure of the multiprocessor system 100 of this 2nd Embodiment is the same as that of the said 1st Embodiment.

43 illustrates logical hardware request I / F 350k of virtual server 1_370a. In the logical hardware request I / F 350k, reference numerals 180d and 180f are requested as four CPU cores, 8GB of memory, and an I / O adapter.

FIG. 44 shows the physical-logical hardware assignment table 310k corresponding to the logical hardware request I / F 350k of FIG. 45 shows an arrangement of the virtual server 1_370a allocated on the multiprocessor system 100 of FIG. The part enclosed with a dotted line in the figure is the virtual server 1_370a.

Here, the operation of the guest OS 1_360a on the virtual server 1_370a will be considered. It is assumed that the guest OS 1_360a is an OS on the multiprocessor system 100 and has a function of Affinity control for CPU and memory based on ACPI as described above. By the function of the affinity control, CPU cores # 4, 5 and DIMMs # 8, 9, 10, 11 on the CPU socket 110c are allocated to the application on the guest OS 360, or the CPU on the CPU socket 110d is allocated. By allocating cores # 6, 7 and DIMMs # 12, 13, 14, 15, the latency of access from the CPU core to the memory is shortened to improve performance. However, in order to use such affinity control function, the guest OS 1_360a needs to know the hardware arrangement of the virtual server 1. Therefore, configuration information notification I / F (logical hardware configuration information notification I / F 340) from the virtual machine monitor 300 to the guest OS 360 is required.

46 shows physical hardware configuration information 750 on the virtual machine monitor 300. This information is made secondary from the physical-logical hardware allocation table 310, and indicates the time point of the physical address of the host CPU core # 751 and the allocated memory 150 indicating the identifier or serial number of the CPU core 120. Host physical address base 752, and host physical address range 753, indicating the amount of memory 150 allocated. However, even if this information is notified to the guest OS 1_360a as it is, the guest OS 1_360a cannot perform Affinity control correctly. This is because the host physical address base on the multiprocessor system 100 allocated by the virtual machine monitor 300 to the guest OS 360a and the guest physical address base on the guest OS 1_360a are different from each other.

48 shows the relationship between the host physical address space and the guest physical address space. When the guest OS 1_360a is assigned to 0x2_0000_0000 of the host physical address base, the address shift occurs so that 0x0_0000_0000 of the guest physical address base corresponds to 0x2_0000_0000 of the host physical address base. Therefore, when notifying the guest OS 1_360a of the logical hardware configuration information, it is necessary to notify in consideration of this misalignment.

47 shows the logical hardware configuration information notification I / F 340 from the virtual machine monitor 300 to the guest OS 1_360a. The logical hardware configuration information notification I / F 340 is a guest CPU core # 341 indicating the identifier or serial number of the CPU core 120 assigned to the guest OS 360, and the base assigned to the guest OS 360. A guest physical address base 342 representing an address and a guest physical address range 343 representing an address range allocated to the guest OS 360. In FIG. 47, the guest CPU core # 341 is renumbered in order from zero in the guest OS 360. The guest physical address base 342 is a value obtained by subtracting the base address 0x2_0000_0000 of the virtual server 1_370a from the host physical address base 752. The guest physical address range 343 is set to the same value as the host physical address range 753. This makes use logic hardware configuration information available to the guest OS 1_360a for affinity control based on ACPI. The virtual machine monitor 300 generates this logical hardware configuration information notification I / F 340 and notifies the guest OS 360. Note that although only the CPU # 341 and the pair of address ranges of the memory are notified here, it is also possible to notify more advanced information (including information of a distance such as the distance calculation table 600 between resources). Do. Even in that case, the principle of converting a host physical address into a guest physical address does not change.

As described above, according to the second embodiment of the present invention, the virtual machine monitor 300 allocates an appropriate resource to the virtual server, and the guest OS 360 on the virtual server has a logical hardware configuration generated by the virtual machine monitor 300. By acquiring the information notification I / F 340, it is possible to correctly use the physical position information of the component to be used, and to operate the affinity control of the guest OS 360 correctly to ensure the same performance and reliability as that of the physical server. It becomes possible.

Third Embodiment (Reassignment of Resources)

A third embodiment of the present invention will be described with reference to FIGS. 43 to 45 and 49 to 54. Similarly to the second embodiment, according to the logical hardware request I / F 350k shown in FIG. 43, the virtual machine monitor 300 is virtual as shown in the physical-logical hardware assignment table 310k shown in FIG. 45 shows the result of assigning the server 1_370a.

With the virtual server 1_370a assigned as shown in FIG. 45, the logical hardware request I / F 350m for the virtual server 2 shown in FIG. 49 is input from the console 230 to the virtual machine monitor 300. Assume that In the logical hardware request I / F 350m of FIG. 49, the priority 357 is 10, and the value is larger than the priority 5 of the logical hardware request I / F 350k of the virtual server 1. FIG. In this case, the virtual machine monitor 300 is required to rearrange the resources already allocated to the virtual server 1 so as to satisfy the resource allocation policy of the higher priority virtual server 2.

FIG. 50 shows a physical-logical hardware allocation table 310m when the virtual machine monitor 300 allocates resources to the virtual server 2_370b without changing the resources allocated to the virtual server 1_370a. In addition, in Fig. 51, the physical-logical hardware allocation table in the case of accepting the allocated resource from the virtual server 1_370a once, allocating the resource to the virtual server 2_370b, and then newly assigning the resource to the virtual server 1_370a ( 310n). In addition, Fig. 52 shows an inter-resource distance calculation table 600m calculated according to the physical-logical hardware assignment table 310m of Fig. 50, and Fig. 53 is calculated according to the physical-logical hardware assignment table 310n of Fig. 51. The distance calculation table 600n between one resource is shown.

The resource allocation policy 356 of the logical hardware request I / F 350m in FIG. 49 is &quot; I / O-memory priority &quot;, and the inter-component component of the &quot; I / O-memory &quot; Comparing the sum of the distances 605, it is 1 in the calculation table 600n in FIG. 53 for 2 in the calculation table 600m in FIG. 52, and the resource relocation of the virtual server 2 is more relocated. It turned out that arrangement which satisfy | fills a policy is attained. FIG. 54 shows how the virtual machine monitor 300 allocates the virtual server 1_370a and the virtual server 2_370b on the multiprocessor system 100 according to the physical-logical hardware allocation table 310n.

As described above, when there are a plurality of logical hardware requests having different priorities, the resources are allocated in order from the high priority logical hardware request I / F 350, so that the resources are preferentially allocated to the higher priority requests. It becomes possible to assign. However, in the case of rearrangement of resources, there are cases in which the CPU and the memory need to be moved. Regarding the movement of the CPU, a copy of the contents of the register, a cache or flashing of the TLB, etc. are required, and a copy of the memory is required for the movement of the memory. Known or well-known methods may be applied to the specific method of moving the CPU or memory.

As described above, according to the third embodiment of the present invention, when the virtual server is newly assigned to the multiprocessor system 100, the resources that have already been allocated are released after the resources have been released in order of high priority. By reallocating, it becomes possible to construct a more optimal virtual server.

As described above, the present invention is applicable to a calculator system composed of a plurality of processors or I / O devices and divided into a plurality of virtual servers, and a virtual machine monitor thereof.

BRIEF DESCRIPTION OF THE DRAWINGS The block diagram which shows 1st Embodiment and shows the structure of the multiprocessor system and virtual machine monitor which apply this invention.

2 is an explanatory diagram showing a first embodiment and showing a configuration of physical hardware configuration information acquisition I / F.

3 is an explanatory diagram showing a first embodiment and illustrating a configuration of a physical component configuration table.

4 is an explanatory diagram showing a first embodiment and showing a configuration of an I / O adapter configuration table.

5 is an explanatory diagram showing a first embodiment and showing a configuration of a distance correspondence table between components;

6 is an explanatory diagram showing a first embodiment and showing a configuration of a physical network configuration table.

FIG. 7 is an explanatory diagram showing the first embodiment and showing the configuration of the logical hardware request I / F of the virtual server 1. FIG.

8 is an explanatory diagram showing a first embodiment and showing the configuration of a physical-logical hardware assignment table of the virtual server 1. FIG.

Fig. 9 is a block diagram showing the first embodiment and showing an example of the virtual server 1 on the multiprocessor system.

Fig. 10 is an explanatory diagram showing the first embodiment and showing the configuration of the logical hardware request I / F of the virtual server 2 having the policy of CPU-memory priority.

Fig. 11 is an explanatory diagram showing a first embodiment and showing the configuration of a logical hardware request I / F of virtual server 2 having a policy of I / O-memory priority.

Fig. 12 is an explanatory diagram showing a first embodiment and showing a configuration of a physical-logical hardware assignment table of virtual server 2 having a policy of CPU-memory priority.

Fig. 13 is an explanatory diagram showing a first embodiment and showing the configuration of a logical hardware request I / F of virtual server 2 having a policy of I / O-memory priority;

Fig. 14 is an explanatory diagram showing a first embodiment and showing a configuration of a distance calculation table between resources of virtual server 2 having a policy of CPU-memory priority;

Fig. 15 is an explanatory diagram showing a first embodiment and showing a configuration of a distance calculation table between resources of virtual server 2 having a policy of I / O-memory priority;

Fig. 16 is a block diagram showing the first embodiment and showing the configuration of virtual servers 1 and 2 on a multiprocessor system with a policy of CPU-memory priority.

Fig. 17 is a block diagram showing the first embodiment and showing the configuration of virtual servers 1 and 2 on a multiprocessor system with a policy of I / O-memory priority.

18 is an explanatory diagram showing a first modification and showing a configuration of a logic hardware request I / F.

19 is an explanatory diagram showing a first modification and showing the structure of a physical-logical hardware assignment table.

20 is a block diagram showing a first modification and showing the configuration of virtual server 1 on a multiprocessor system.

Fig. 21 is an explanatory diagram showing a first modification and showing the configuration of logical hardware request I / F of virtual server 2 with a policy of CPU-memory priority.

Fig. 22 is an explanatory diagram showing a first modification and showing the configuration of logical hardware request I / F of virtual server 2 with policy as CPU-I / O priority.

Fig. 23 is an explanatory diagram showing a first modification and showing the configuration of a physical-logical hardware assignment table of virtual server 2 with a policy of CPU-memory priority.

Fig. 24 is an explanatory diagram showing a first modification and showing the configuration of a physical-logical hardware assignment table of virtual server 2 having a policy of CPU-I / O priority.

25 is an explanatory diagram showing a first modification and showing a configuration of a distance calculation table between resources of virtual server 2 having a policy of CPU-memory priority;

FIG. 26 is an explanatory diagram showing a first modification and showing a configuration of a distance calculation table between resources of the virtual server 2 having the policy of CPU-I / O priority; FIG.

Fig. 27 is a block diagram of the virtual servers 1 and 2 on the multiprocessor system when the first modification is shown and the policy is CPU-memory priority.

Fig. 28 is a block diagram of the virtual servers 1 and 2 on the multiprocessor system in the first modification, when the policy is set to CPU-I / O priority.

29 is an explanatory diagram showing a second modification example and showing the configuration of the logical hardware request I / F of the virtual server 1. FIG.

30 is an explanatory diagram showing a second modification example and showing the structure of a physical-logical hardware assignment table of the virtual server 1. FIG.

Fig. 31 is a block diagram showing the second modification and showing the configuration of virtual server 1 on the multiprocessor system.

Fig. 32 is an explanatory diagram showing a second modification and showing the configuration of logical hardware request I / F of virtual server 2 with policy as the priority of reliability.

Fig. 33 is an explanatory diagram showing a second modification and showing the configuration of logical hardware request I / F of virtual server 2 with policy as band priority.

Fig. 34 is an explanatory diagram showing a second modification and showing the configuration of logical hardware request I / F of virtual server 2 with policy as the power saving priority;

Fig. 35 is an explanatory diagram showing a second modification and showing the configuration of the physical-logical hardware assignment table of the virtual servers 1 and 2 with policy as the reliability priority;

Fig. 36 is an explanatory diagram showing a second modification and showing the configuration of the physical-logical hardware assignment table of the virtual servers 1 and 2 with policy as the band first.

Fig. 37 is a diagram illustrating a configuration of a component network assignment table of virtual servers 1 and 2 in which a policy is given priority in reliability, showing a second modification.

FIG. 38 shows a second modification and shows an example of a configuration of a component network assignment table of virtual servers 1 and 2 in which the policy is band-first; FIG.

39 is an explanatory diagram showing a second modification and showing a configuration of an allocation resource consumption power calculation table in the case where the policy is made with priority on reliability;

40 is an explanatory diagram showing a second modification and showing the configuration of a component network allocation table in the case where the policy is set to consume power first;

Fig. 41 is a block diagram of a multiprocessor system of virtual servers 1 and 2 in which a second priority is given and policy is first of all reliability.

Fig. 42 is a block diagram of a multiprocessor system of virtual servers 1 and 2, in which the policy is band-first, showing a second modification.

43 is an explanatory diagram showing a second embodiment and showing the configuration of a logical hardware request I / F of the virtual server 1;

Fig. 44 is an explanatory diagram showing the second embodiment and showing the structure of a physical-logical hardware assignment table of the virtual server 1;

Fig. 45 is a block diagram showing the second embodiment and showing the configuration of virtual server 1 on the multiprocessor system.

FIG. 46 is an explanatory diagram showing a second embodiment and showing the configuration of physical hardware configuration information; FIG.

Fig. 47 is an explanatory diagram showing the second embodiment and showing the configuration of the logical hardware configuration information notification I / F;

Fig. 48 is a map showing the second embodiment showing the relationship between the host physical address and the guest physical address.

FIG. 49 is an explanatory diagram showing a third embodiment and showing the configuration of the logical hardware request I / F of the virtual server 2. FIG.

FIG. 50 is an explanatory diagram showing a second embodiment and showing the configuration of the physical-logical hardware assignment table of the virtual servers 1 and 2; FIG.

Fig. 51 is an explanatory diagram showing a second embodiment and showing another configuration of the physical-logical hardware assignment table of the virtual servers 1 and 2;

FIG. 52 is an explanatory diagram showing a second embodiment and showing the configuration of a resource consumption power calculation table of virtual server 2; FIG.

FIG. 53 is an explanatory diagram showing a second embodiment and showing another configuration of the resource consumption power calculation table of the virtual server 2. FIG.

Fig. 54 is a block diagram showing the second embodiment and showing the virtual servers 1 and 2 on the multiprocessor system.

55 is a flowchart illustrating the first embodiment and explaining an example of logical hardware allocation processing performed in a virtual machine monitor.

Fig. 56 is a flowchart showing the first embodiment and similarly explaining the subroutine of the I / O adapter allocation processing of the logical hardware allocation processing.

Fig. 57 is a flowchart showing the first embodiment and similarly explaining the subroutine of the CPU and memory selection processing of the logical hardware allocation processing;

Fig. 58 is a flowchart showing the first embodiment and similarly explaining the subroutine of the policy evaluation processing of the logical hardware allocation processing;

Fig. 59 is a flowchart showing the first embodiment and similarly explaining the subroutine of the distance calculation processing between components of the logic hardware allocation processing;

Fig. 60 is a flowchart showing the first embodiment and similarly explaining the subroutine of the component common water calculation processing of the logical hardware allocation processing.

Fig. 61 is a flowchart showing the first embodiment and similarly explaining the subroutine of the effective band calculating process of the logical hardware allocation process;

Fig. 62 is a flowchart showing the first embodiment and similarly explaining a subroutine of power consumption calculation processing of logical hardware allocation processing;

Fig. 63 is a flowchart showing the first embodiment and similarly explaining the subroutines of the CPU and memory allocation processing of the logical hardware allocation processing;

64 shows a first embodiment and a configuration diagram of a multiprocessor system;

<Explanation of symbols for the main parts of the drawings>

100: multiprocessor system

110: CPU socket

120: CPU core

130: memory controller

140: memory interface

150: memory

160: I / O hub

170: I / O interface

180: I / O Adapter

190: I / O device

200: interface between modules

210: module information acquisition interface

220: service processor

300: virtual machine monitor

310: physical-logical hardware assignment table

320: physical hardware configuration information acquisition interface

330 physical-logical hardware assignment interface

340: logical hardware configuration information notification interface

350: logical hardware request interface

400: physical component scheme

500: Distance correspondence table between components

550 physical network scheme

600: distance calculation table between resources

650: component network allocation table

700: allocation lease power consumption meter calculation table

750: physical HW configuration information

801: Logical Hardware Allocation Processing Unit

Claims (17)

And a plurality of processors, a plurality of memories, and a computer resource including a plurality of I / O devices, the computer resources being grouped and connected to components such as a processor socket, an I / O hub, and the like. A virtual machine monitor that runs a virtual server on a multiprocessor system with each other connected to an internal network. The virtual machine monitor, Physical hardware for acquiring configuration information of hardware including information indicating an inclusion relationship between each of a plurality of processors of the multiprocessor system, a computer resource including a plurality of memories and a plurality of I / O devices, and a component to which they belong; An information acquisition unit, A reception unit which receives a generation request including the number and memory of the virtual server to be created and the allocation policy of the I / O device and the resource; An allocation processing unit for allocating the I / O device specified in the generated generation request to the virtual server and then allocating the virtual server with the required number of processors and the required amount of memory to satisfy the allocation policy. Virtual machine monitor comprising a. delete The method of claim 1, The reception unit receives a request for creating a second virtual server, The allocation processing unit, The allocation of the processor and memory already allocated to the allocated virtual server is released once, and the I / O device specified in the creation request of the second virtual server is allocated to the second virtual server, and the second virtual server is allocated. A processor group including an unallocated processor and a memory group including an unallocated memory to satisfy the number of processors and the amount of memory and resource allocation policy included in the generation request of the second virtual server; Selecting a processor and memory to be allocated and relocating the processor and memory of the virtual server based on an allocation policy of the allocated virtual server's resources. The method of claim 1, The physical hardware information acquisition unit may be configured to perform a process between the processor and the memory and the processor and the I / O from information indicating the inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories, and the plurality of I / O devices, and the components to which they belong. Find the distance between O devices, The allocation processing unit, And allocate a processor and memory to the virtual server to which the I / O device is allocated such that the sum of the distances of the processor, memory, and I / O device is minimized based on the allocation policy of the resource. monitor. The method of claim 1, The allocation processing unit, And assigning one I / O device of the plurality of I / O devices to one virtual server, avoiding sharing among the plurality of virtual servers. delete The method of claim 1, The physical hardware information acquisition unit obtains a distance between a processor and a memory from information indicating an inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories and the plurality of I / O devices, and a component to which the physical resources belong. The allocation processing unit, If the resource allocation policy prioritizes access to the memory of the processor, assigning the processor and memory to the virtual server to which the I / O device is allocated such that the sum of the distances between the processor and the memory is minimal. Virtual machine monitor, characterized in that. The method of claim 1, The physical hardware information acquisition unit may determine a distance between the processor and the I / O device from information indicating an inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories, and the plurality of I / O devices, and the components to which they belong. Finding, The allocation processing unit, If the allocation policy of the resource prioritizes access of a processor and an I / O device, the processor is assigned to the virtual server to which the I / O device is allocated such that the sum of the distances between the processor and the I / O device is minimized. And memory allocation for the virtual machine. The method of claim 1, The physical hardware information acquisition unit may determine a distance between a memory and an I / O device from information indicating an inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories, and the plurality of I / O devices, and the components to which they belong. Finding, The allocation processing unit, If the resource allocation policy prioritizes access of memory and I / O devices, the virtual server to which the I / O device is allocated has a processor such that the sum of the distances between the memory and the I / O devices is minimized. And memory allocation for the virtual machine. The method of claim 1, The physical hardware information acquiring unit includes a processor, a memory, and an I / O device based on information indicating an inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories, and the plurality of I / O devices, and the components to which they belong. Find the distance of each other, The allocation processing unit, When the resource allocation policy secures the performance of the entire virtual server, the processor is allocated to the virtual server to which the I / O device is allocated such that the sum of the distances between the processor, the memory and the I / O device is minimized. And memory allocation for the virtual machine. The method of claim 1, The physical hardware information acquiring unit is configured to select a processor, a memory, and an I / O device from information indicating an inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories, and the plurality of I / O devices, and the components to which they belong. Find the path of the internal network to which you are connecting, The allocation processing unit, If the resource allocation policy gives priority to the reliability of the virtual server, the number of internal networks that are duplicated in the internal network used by the already allocated virtual server and the internal network used by the new virtual server that has created the request. And allocate a processor and memory to the new virtual server to which the I / O device is allocated so that the sum of the sums is minimal. The method of claim 1, The physical hardware information acquiring unit is configured to select a processor, a memory, and an I / O device from information indicating an inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories, and the plurality of I / O devices, and the components to which they belong. Find the path of the internal network to which you are connecting, The allocation processing unit, When the resource allocation policy has priority over the internal network bandwidth, the internal network bandwidth used by the already allocated virtual server and the internal network used by the new virtual server that has created the request overlap with the internal network bandwidth. And a number divided by the number of virtual servers as an effective band, and a processor and memory are allocated to the new virtual server to which the I / O device is allocated such that the effective band is maximized. The method of claim 1, The physical hardware information acquisition unit includes the processor, the memory, and the I / O from the information representing the inclusion relationship between each of the computer resources including the plurality of processors, the plurality of memories, and the plurality of I / O devices, and the components to which the components belong. Calculate the power consumption of the device and the components to which each belongs, The allocation processing unit, If the resource allocation policy prioritizes power consumption, the power consumption of the processor and memory and I / O devices and components of the virtual server already allocated, and the processor and memory used by the new virtual server that generated the request. And allocate a processor and memory to the new virtual server to which the I / O device is assigned so that the sum of power consumption of the I / O device and the component is minimized. delete delete delete delete

Images