JP6696947B2 - Virtual machine emulation system, virtual machine emulation method and computer program - Google Patents

Description

  The present invention relates to a virtual machine emulation system, a virtual machine emulation method, a computer program, and a virtual machine used in, for example, a cloud server business.

  A virtual machine that is emulated by software on an OS (Operating System) of a physical machine is known. In this case, a cluster is used to improve the processing capacity or secure the availability for the OS level or cloud base. Cluster methods include, for example, an active standby method and a replication method.

  In Patent Document 1, each core of the first processor of the call processing server belonging to the active system is assigned to each VM (Virtual Machine) of the call processing server, and the second processor of the call processing server belonging to the standby system is provided. Describes a technique for allocating a part of a plurality of cores of the above to a plurality of VMs that are standby systems.

  Patent Document 2 describes a technique of setting an arbitrary number of virtual machines of an arbitrary capacity in a resource pool by a resource pool in which a user sets an upper limit value of the resource capacity.

JP, 2017-27166, A JP, 2005-046078, A

  When a virtual machine is handled in the short term like cloud computing, a physical machine including an unused core is searched and a virtual machine is constructed. However, a virtual machine could only exist on one physical machine. That is, the virtual machine could be constructed only with an unused core configuration in the physical machine.

  Thus, for example, a physical machine having less than the desired number of cores (eg, 3 cores) in a virtual machine (eg, 2 cores) is unused and thus inefficient. Further, the processing capacity of the virtual machine cannot be higher than that of the physical machine, and the load may be concentrated on a specific physical machine.

  Also, in order to increase the availability of virtual machines, multiple virtual machines may be multiplexed. However, as described above, since a plurality of virtual machines are packaged in the same physical machine (physical CPU (Central Processing Unit)), when the physical CPU itself is broken, all virtual machines go down. There is a risk.

  Therefore, it is assumed that virtual machines are multiplexed among different physical machines. For example, it is assumed to use a cluster configuration on the application level. However, in this case, it is necessary to transfer the data from the primary virtual machine to the secondary virtual machine in order to synchronize the data. Since this data transfer has a large processing load, if it is executed frequently, the processing performance of the application may deteriorate.

  The above problems cannot be solved by the techniques of Patent Document 1 and Patent Document 2.

  The present invention has been made to solve the above problems, and provides a virtual machine emulation system, a virtual machine emulation method, a computer program, and a virtual machine capable of improving the performance and availability of a virtual machine. The purpose is to do.

  The virtual machine emulation system of the present invention is provided in at least one physical machine among a plurality of physical machines, and collectively manages all of the plurality of cores included in each of the plurality of physical machines. When constructing a virtual machine, at least one of the cores is designated from each of the at least two physical machines, and a core management unit that assigns the designated plurality of cores as virtual CPUs of the virtual machine is provided.

  A virtual machine emulation method of the present invention constructs a virtual machine by collectively managing a plurality of cores included in each of a plurality of physical machines in at least one of the plurality of physical machines. At this time, at least one of the cores is designated from each of the at least two physical machines, and the designated plurality of cores are assigned as virtual CPUs of the virtual machine.

  A computer program of the present invention collectively manages a plurality of cores included in each of a plurality of physical machines in a computer of at least one of the plurality of physical machines to construct a virtual machine. At this time, at least one core is designated from each of the at least two plurality of physical machines, and a process of assigning the designated plurality of cores as a virtual CPU of the virtual machine is executed.

  In the virtual machine of the present invention, the number of provisional processing instructions to which the plurality of cores selected from at least one of the at least two physical machines among the plurality of physical machines each having a plurality of cores are allocated It has an idea CPU.

  According to the present invention, it is possible to improve the processing capacity and availability of virtual machines.

It is a block diagram which shows the structural example of the virtual machine emulation system which concerns on the 1st Embodiment of this invention. 3 is a flowchart for explaining an operation example of the virtual machine emulation system shown in FIG. 1. It is a block diagram which shows the structural example of the virtual machine emulation system which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the structural example of the virtual machine emulation system which concerns on the 3rd Embodiment of this invention. It is an image figure of a virtual machine. It is a figure which shows the example of the 1st virtual machine constructed | assembled by 3rd Embodiment. It is a figure which shows the example of the 2nd virtual machine constructed | assembled by 3rd Embodiment. It is a block diagram which shows the structural example of the core management part shown in FIG. It is a block diagram which shows the structural example of the core control part shown in FIG. 9 is a flowchart for explaining an example of a virtual machine allocation operation in the virtual machine emulation system according to the third exemplary embodiment. 8 is a flowchart showing an example of an operation when a virtual CPU receives a processing command from an OS / application after the assigned virtual machine has been operated in the virtual machine emulation system of the present exemplary embodiment. It is a flow chart which shows an example of operation data update operation in a virtual machine emulation system of this embodiment. 6 is a flowchart showing an example of a failure monitoring operation after the assigned virtual machine has been activated in the virtual machine emulation system of the exemplary embodiment.

[First Embodiment]
(Description of configuration)
FIG. 1 is a block diagram showing a configuration example of a virtual machine emulation system 1 according to the first exemplary embodiment of the present invention.

  The virtual machine emulation system 1 emulates a virtual machine on a physical machine. The virtual machine emulation system 1 includes a core management unit (a core management unit 4 in FIG. 1) provided in at least one physical machine of a plurality of physical machines (the physical machine 2 and the physical machine 3 are illustrated in FIG. 1). And a core management unit 5). The core management unit is an example of core management means.

  The core management unit 4 and the core management unit 5 collectively manage all of the plurality of cores of the physical CPUs of the physical machines 2 and 3, respectively. Then, when constructing the virtual machine, at least one of the core management unit 4 and the core management unit 5 designates at least one core from each of at least two physical machines (for example, the physical machines 2 and 3) and is designated. The plurality of cores are assigned as virtual CPUs of the virtual machine.

The core management unit has a function of constructing a virtual machine including only cores in one physical machine, in addition to the above-described function of constructing a virtual machine including a core designated from each of at least two physical machines. Can be provided.
(Explanation of operation)
FIG. 2 is a flow chart for explaining an operation example (virtual machine emulation method) of the virtual machine emulation system 1 shown in FIG.

  In at least one physical machine of the physical machines 2 and 3, for example, the physical machine 2, the core management unit 4 collectively manages a plurality of cores included in the physical machines 2 and 3 (steps). S1). When constructing a virtual machine, the core management unit 4 designates at least one core from each of at least two physical machines (for example, the physical machines 2 and 3), and designates the designated cores as virtual CPUs of the virtual machine. Assign as.

The configuration and operation of the virtual machine emulation system (particularly, the core management unit) of the first embodiment will be described in detail in the third embodiment shown below.
(Explanation of effects)
In the first embodiment described above, at least one core management unit designates at least one core from each of at least two physical machines, and assigns the designated multiple cores as virtual CPUs of virtual machines. That is, in this embodiment, one virtual machine can be constructed across a plurality of physical machines. As a result, the processing capacity of the virtual machine can be increased to one physical machine or more. Further, it is possible to prevent the load (the number of used cores) from being biased only to a specific physical machine.

  Further, in this embodiment, the virtual CPU of the virtual machine is configured by the cores of different physical machines. Even if one physical CPU fails, the other physical CPU is unlikely to fail. Therefore, according to the present embodiment, it is possible to avoid a situation where the entire virtual machine goes down.

In summary, according to this embodiment, it is possible to improve the processing capacity and availability of the virtual machine.
[Second Embodiment]
FIG. 3 is an example in which the virtual machine emulation system 1 shown in FIG. 1 or the virtual machine emulation system 10 shown in FIG. 4 described later is realized by a computer, and a virtual machine according to a second embodiment of the present invention. 3 is a block diagram showing a configuration example of an emulation system 100. FIG. The virtual machine emulation system 100 includes a storage unit 102 and a calculation unit 104 including a processor and the like. The storage unit 102 is a computer-readable recording medium and stores a computer program 150. The computer program 150 is, for example, a program for causing the calculation unit 104 to execute the processing shown in FIG. 2 and the processing shown in FIG. 10. The computer program 150 is executed by, for example, the computing unit 104 (computer) of at least one physical machine of the plurality of physical machines.

  According to the second embodiment described above, it is possible to improve the processing capacity and availability of the virtual machine, as in the first and third embodiments described later.

The arithmetic unit 104 includes, for example, a CPU and a memory that stores an instruction executed by the CPU. The computer-readable recording medium is, for example, a non-transitory storage device. Examples of the non-transitory storage device include, for example, a portable medium such as an optical disk, a magnetic disk, a ROM (Read Only Memory), and a nonvolatile semiconductor memory, and a hard disk built in a computer system. Further, the computer-readable recording medium may be a temporary storage device. Examples of the temporary storage device include a signal line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, or a volatile memory inside a computer system.

[Third Embodiment]
(Description of configuration)
FIG. 4 is a block diagram showing a configuration example of a virtual machine emulation system 10 according to the third exemplary embodiment of the present invention. The virtual machine emulation system 10 emulates a virtual machine on a physical machine.

  The virtual machine emulation system 10 includes a core management unit (a core management unit 22 in FIG. 4) provided in at least one physical machine among a plurality of physical machines (the physical machine 12 and the physical machine 14 are illustrated in FIG. 4). And a core management unit 32). The core management unit is an example of core management means. Further, the virtual machine emulation system 10 includes a core control unit (the core control unit 24 and the core control unit 34 are illustrated in FIG. 4) provided for each of the physical machines 12 and 14. The core control unit is an example of core control means.

  When constructing a desired virtual machine, a user specifies a core management unit (for example, the core management unit 22 of the physical machine 12 in FIG. 4) provided in at least one physical machine. In the present embodiment, this core management unit is called a "designated core management unit" (an example of designated core management means). The designated core management unit, for example, the core management unit 22 functions as a first user interface machine for a user who wants to construct a desired virtual machine. The core management unit that manages cores when the cores managed by the designated core management unit are multiplexed is referred to as a “multiplex core management unit” (an example of a multiplex core management unit) in the present embodiment. The designated core management unit and the multiplexed core management unit have, for example, an Active / Active configuration (mutual standby configuration). During normal operation, each core management unit executes different processing, and if the designated core management unit fails (or if an abnormality is found), the processing executed by the core managed by the designated core management unit is The core is managed by the core management unit. In other words, when an abnormality occurs in the designated core management unit, the multiplexed core management unit becomes the designated core management unit.

  For the purpose of improving the performance (for example, processing speed) of a virtual machine operated on the virtual machine emulation system 10, the core management unit (core management unit 22) typically uses one physical machine (for example, It is provided in the physical machine 12). Of course, it is also possible to provide a plurality of core management units even for the purpose of improving performance.

  On the other hand, when the purpose is to multiplex the core configuration of one virtual machine in order to improve the availability of the virtual machines, the core management unit (core management unit 22, core management unit 32) includes at least two physical machines. 12 and 14 are provided respectively.

  The physical CPU 20 and the physical CPU 30 each include a plurality of cores. In the present embodiment, the case where each has six cores (the case of hexa cores) is taken as an example. For example, the physical CPU 20 includes cores A to F, and the physical CPU 30 includes cores G to L. Of course, the number of cores in one physical CPU is not limited to six.

  The core management units 22 and 32 collectively manage all the cores A to L of all the physical machines 12 and 14 in the virtual machine emulation system 10. Specifically, the core management units 22 and 32 acquire operation information on all the cores A to L in the system via the core control units 24 and 34.

  Then, when constructing the virtual machine, at least one of the core management units 22 and 32 creates at least one core from each of the at least two physical machines (for example, the physical machines 12 and 14) based on the operation information. The designated cores are assigned and assigned as virtual CPUs of the virtual machine. The operation information includes information about the designated status (usage status) of each core and the usage frequency of each core (the number of processing instructions received per unit time).

  The core management unit has a function of constructing a virtual machine including only cores in one physical machine, in addition to the above-described function of constructing a virtual machine including a core designated from each of at least two physical machines. Can be provided.

  In this case, at least one of the core management units 22 and 32 presents the operating information to the user, and the core assigned to the virtual machine by the user's designation based on the presentation (hereinafter, “designated core”). Sometimes referred to as “.”).

  In addition, the above-mentioned instruction specifically specifies, for example, which physical machine 12 or 14 to use the core management unit as a virtual machine (a plurality of core management units in the case of multiplexing) and which physical machine 12. , And specify how many 14 cores to use.

  The designated core management unit and the multiplex core management unit communicate with each other, and obtain information (the operating information of cores of other physical machines) that they do not have from other core management units. The contents of the core management table 60 can be synchronized. That is, in the case of multiplexing, the system core management table 60, which will be described later, of each core management unit has the same content, and the content is operation information for all cores in the virtual machine emulation system 10.

  The core management units 22 and 32 may themselves determine the designated core by analyzing the operation information.

  When constructing the virtual machine, the core management units 22 and 32 instruct the respective core control units 24 and 34 to secure the designated core.

  Further, when a processing command is received from the OS or application after the virtual machine is constructed, the core management units 22 and 32 select a predetermined core (for example, the optimum core) from all the cores allocated as the virtual machine (virtual CPU). A core), and sends a processing command to the core control units 24 and 34 that control the specified core.

  The core control units 24 and 34 acquire operation information about all the cores of the physical machines 12 and 14 to which they belong, and report the acquired operation information to the core management units 22 and 32, respectively.

  When constructing the virtual machine, the core control units 24 and 34 secure the designated core for the virtual machine based on the designation core securing instruction from the core management units 22 and 32.

  When the virtual machine is started and receives the processing instruction from the core management units 22 and 32, the core control units 24 and 34 transfer the processing instruction to the designated core. The core control units 24 and 34 transmit the processing result regarding the processing instruction received from the core to the core management unit.

  FIG. 5 shows an image diagram of a virtual machine. FIG. 5A is an image diagram of a general virtual machine for comparison, that is, a virtual machine constructed in one physical machine. FIG. 5B is an image diagram of a virtual machine constructed in the third embodiment, that is, a virtual machine constructed across a plurality of physical machines. Although FIG. 5B shows a virtual machine constructed in one physical machine, this virtual machine is not always necessary.

  FIG. 6 is a diagram showing an example of the first virtual machine 41A constructed in this embodiment. The first virtual machine 41A is a virtual machine constructed especially for the purpose of improving performance. The first virtual machine 41A includes an OS, an application, and a virtual CPU 40A. The virtual CPU 40A has three core management units, which are designated core management units 22, a core control unit 24, all cores A to F of the physical machine 12, a core control unit 34 of the physical machine 14, and a physical machine 14. And cores G to I. In this case, the core management unit 32 of the physical machine 14 is unnecessary. The core control unit 34 of the physical machine 14 is controlled by the core management unit 22 of the physical machine 12. Since the first virtual machine 41A includes the virtual CPU 40A including nine cores, the first virtual machine 41A has more cores and higher performance than one physical machine. For example, the processing time can be shortened when the same processing is executed. The core configuration of the first virtual machine 41A shown in FIG. 6 is merely an example. It is sufficient if the total number of cores used in the first virtual machine 41A exceeds the number of cores in the physical machine. For example, the first virtual machine 41A can include a virtual CPU 40A including four cores of the physical machine 12 and four cores of the physical machine 14. In this case, similarly to the above, the number of cores of the first virtual machine 41A is eight, which exceeds six cores of one physical machine. Further, the designated core management unit does not necessarily have to be the core management unit 22, and may be the core management unit of another physical machine.

FIG. 7 is a diagram showing an example of the second virtual machine 41B constructed in this embodiment. The second virtual machine 41B is a virtual machine constructed especially for the purpose of improving availability. The second virtual machine 41B includes an OS, an application, and a virtual CPU 40B. The virtual CPU 40B includes a core management unit 22 that is a designated core management unit, a core control unit 24, cores A to F (designated cores), a core management unit 32 that is a multiplexed core management unit, and a core control unit 34. , Cores G to L (designated cores). In this case, information is shared between the core management unit 22 and the core management unit 32. In the second virtual machine 41B, cores A to F controlled by the core control unit 24 and cores G to L controlled by the core control unit 34 are multiplexed (also referred to as duplication or redundancy).

  FIG. 8 is a block diagram showing a configuration example of the core management unit 22 shown in FIGS. 4, 6, and 7. The core management unit 32 has the same configuration as the core management unit 22. Further, although the designated core management unit and the multiplexing core management unit do not operate in the same manner, they have different configurations to be used, and thus have potentially the same configurations themselves.

  The core management unit 22 includes a user interface unit 50, a virtual CPU processing unit 52, a multiplexing processing unit 54, a core control communication unit 56, a system core operation management unit 58, a system core management table 60, and a system monitor. And a section 62.

  The user interface unit 50 is a mechanism for exchanging information between the user and the virtual machine emulation system 10. The user interface unit 50 includes, for example, hardware such as a keyboard, a mouse, and a display, and software such as menus, icons, and windows on the screen.

  The virtual CPU processing unit 52 receives a processing command from the OS / application.

  As shown in FIG. 6, when only one core management unit (that is, the designated core management unit) is set in the virtual CPU, the processing instruction from the OS / application is executed by the designated core management unit. It is received by the virtual CPU processing unit 52.

  On the other hand, as shown in FIG. 7, when a plurality of core management units (that is, a designated core management unit and a multiplexed core management unit) are set in the virtual CPU, the processing instruction from the OS / application is, for example, , May always be once received by the virtual CPU processing unit 52 of the designated core management unit. Alternatively, the virtual CPU processing units 52 of the plurality of core management units each receive a processing instruction, and the virtual CPU processing unit 52 receives the processing instruction by determining whether the core to which the processing instruction is executed is a core that is not under its control. It is also possible to select processing instructions. For example, the core management unit can perform exclusive control when accessing a memory or I / O (Input / Output) to control the operation of the core. The mechanism of minimizing the exclusive control can be handled by the function of the OS.

  The multiplexing processing unit 54 communicates with a multiplexing core management unit (for example, the core management unit 32 in FIG. 4) when performing multiplexing.

  The core control communication unit 56 communicates with the core control unit 24. For example, the core control communication unit 56 transfers the processing instruction for the core designated in the virtual machine to the core control unit 24. Note that, for example, when the core management units are not multiplexed as shown in FIG. 6 and only the core management unit 22 is used as shown in FIG. 6, the core management unit 22 belongs. No physical machine 14 has a core management unit (not used). However, even in such a case, the cores to be controlled (cores G, H, and I) also exist in the physical machine 14. Therefore, in this case, the core control communication unit 56 of the core management unit 22 of FIG. 6, which is the designated core management unit, also receives the core control unit 34 of the physical machine 14 in which the core management unit 22 does not exist. Transfer processing instructions.

  The system core operation management unit 58 receives operation information about all cores controlled by the core control unit 24 from the core control unit 24, and stores the operation information in the system core management table 60. When the virtual machines are not multiplexed, the operation information on all the cores controlled by the core control unit is also received from the core control unit of the physical machine to which the designated core management unit does not belong.

  The system monitoring unit 62 performs fault monitoring.

  FIG. 9 is a block diagram showing a configuration example of the core control unit 24 shown in FIG. The configuration of the core control unit 34 is the same as the configuration of the core control unit 24.

  The core control unit 24 includes a core management communication unit 80, a core communication unit 82, a core operation management unit 84, and a core management table 86.

  The core management communication unit 80 communicates with the core management unit 22. For example, the core management communication unit 80 receives a processing instruction from the core management unit 22.

  The core communication unit 82 transfers the processing instruction received by the core management communication unit 80 to the core.

  The core operation management unit 84 is activated at regular intervals. The activated core operation management unit 84 collects core operation information by referring to the core management table 86, and transmits the collected operation information to the core management unit via the core management communication unit 80. The content of the operation information is as described above.

The core management table 86 stores operation information about cores in a physical machine to which the core management table 86 belongs.
(Explanation of operation)
FIG. 10 is a flowchart for explaining an example of a virtual machine allocation operation in the virtual machine emulation system 10. In the following description, it is assumed that the core management unit 22 of the physical machine 12 is designated as the designated core management unit. Of course, the designated core management unit is not limited to the core management unit 22.

  When allocating a virtual machine, the user interface unit 50 of the core management unit 22, which is the designated core management unit, receives an operation information request from the user (step S10). The user interface unit 50 reads the operation information (information on the operation of all the cores A to L in the virtual machine emulation system 10) from the system core management table 60 (step S11). The user interface unit 50 presents the operation information to the user as a response to the operation information request (step S12). Specifically, the user interface unit 50 displays the read operation information on a display or the like.

  The user interface unit 50 accepts a virtual machine allocation request including virtual machine allocation information, which is information such as physical machine specification, the number of cores to be used, and multiplexing specification, from the user who has recognized the operation information (step S13). .. The core control communication unit 56 is information based on the virtual machine allocation request accepted by the user interface unit 50, and requests a core reservation request, which is a request for securing a core, to a designated physical machine (for example, a physical machine). 12) to the core control unit (for example, the core control unit 24) (step S14).

  The core management communication unit 80 of the core control unit 24 receives the core reservation request (step S15). The core management communication unit 80 updates the operation information on the core management table 86 (step S16). Specifically, the core management communication unit 80 changes the state of the core designated by the core reservation request to "in use". The core management communication unit 80 transmits to the core management unit 22 a core reservation response indicating that the core reservation has been completed, which is a response to the core reservation request (step S17).

  The core control communication unit 56 that has received the core allocation response determines whether or not the user's instruction (contents of the virtual machine allocation information) includes multiplexing designation (step S18). When the multiplexing designation is included (step S18 / Yes), the multiplexing processing unit 54 causes the multiplexing core management unit (for example, the core management unit 32 of the physical machine 14) to perform multiplexing including the above virtual machine allocation information. The request is transmitted (step S19).

  The multiplexing processing unit 54 of the core management unit 32, which is the multiplexing core management unit, receives the multiplexing request (step S20). The multiplexing processing unit 54 stores the virtual machine allocation information included in the received multiplexing request (step S21). The multiplexing processing unit 54 transmits a multiplexing response as a response to the multiplexing request to the core management unit 22 which is the designated core management unit (step S22).

  When the user's instruction does not include the multiplexing designation (step S18 / No) and when the multiplexing response is received from the core management unit 32 which is the multiplexing core management unit, the designated core management unit determines the virtual machine allocation information. Is stored (step S23). The user interface unit 50 outputs a virtual machine allocation response as a response to the virtual machine allocation request to the user (step S24). A virtual machine is assigned by a series of these processes.

  FIG. 11 is a flowchart showing an example of the operation when the virtual CPU receives a processing command from the OS / application after the assigned virtual machine has started operating.

  After running the virtual machine, the virtual CPU receives a processing command from the OS or application (step S30). Specifically, the processing instruction is received by the virtual CPU processing unit 52 of the core management unit that constitutes the virtual CPU. The virtual CPU processing unit 52 reads operating information from the system core management table 60 (step S31). The virtual CPU processing unit 52 selects an optimum core from among the plurality of cores allocated in the virtual machine based on at least the operation information (step S32). This optimal core selection process may be called scheduling. Further, the operation information used to select the optimum core is, for example, information on the usage frequency of each core (the number of processing instructions received per unit time). The virtual CPU processing unit 52 sends a processing command to the core control unit that controls the selected core (step S33). The core management communication unit 80 of the core control unit receives the processing instruction (step S34). The core management communication unit 80 increments the number of processing instructions of the core in the operation information of the core management table 86 for the core that executes the processing instruction (for example, the optimum core) (step S35). The core communication unit 82 transmits the processing instruction to the core (step S36).

  FIG. 12 is a flowchart showing an example of the operation information update operation.

  First, as described with reference to FIG. 11, the core control unit increments the number of processing instructions in the core management table 86 each time a processing instruction is received.

  Then, the core operation management unit 84 of the core control unit is activated at regular intervals. Specifically, for example, the core operation management unit 84 observes the measurement result of the timing means (not shown in FIG. 9) or confirms whether the time-up signal is received from the timing means. , It is determined whether or not the activation timing has arrived (step S39). When the activation timing has not come (step S39 / No), the process of step S39 is executed again. On the other hand, when the activation timing arrives (step S39 / Yes), the core operation management unit 84 activates and reads the number of processing instructions of each core from the core management table 86 (step S40). The core operation management unit 84 transmits the number of read processing instructions of each core to the core management unit (step S41). After the transmission, the core operation management unit 84 clears the number of processing instructions of each core in the core management table to zero (step S42).

  The system core operation management unit 58 of the core management unit receives the number of processing instructions of each core (step S43). The system core operation management unit 58 reflects the number of received processing instructions in the system core management table 60 (step S44).

  Through the above processing, all core management units can grasp the number of processing instructions per unit time (that is, the frequency of use) for all cores.

  When the virtual machine is not multiplexed and only the designated core management unit exists in the system (for example, in the case of FIG. 6), the core control existing in the physical machine in which the designated core management unit does not exist. The unit transmits the number of read processing instructions of each core to the designated core management unit.

  On the other hand, when the virtual machine is multiplexed (for example, in the case of FIG. 7), the core control unit transmits the number of processing instructions only to the core management unit in the physical machine to which the core control unit belongs. In this case, data synchronization between the designated core management unit and the multiplexing core management unit (synchronization of the system core management table 60) is performed separately.

  FIG. 13 is a flowchart showing an example of the fault monitoring operation after the assigned virtual machine is activated.

  This fault monitoring is executed when the system monitoring unit 62 of the core management unit is activated at regular intervals. Specifically, for example, the system monitoring unit 62 observes a measurement result of a clocking means (not shown in FIG. 8) or confirms whether or not a time-up signal is received from the clocking means. It is determined whether or not the activation timing has arrived (step S49). When the activation timing has not come (step S49 / No), the process of step S49 is executed again. On the other hand, when the activation timing comes (step S49 / Yes), the system monitoring unit 62 activates and determines whether or not the core management unit to which the system monitoring unit 62 belongs operates as a "multiplexing core management unit" (step S50). ). When the core management unit to which it belongs operates as the "multiplexing core management unit" (step S50 / Yes), the system monitoring unit 62 transmits a "health check signal" to the designated core management unit (step S51). .. The system monitoring unit 62 determines whether the response from the designated core management unit to the health check signal has been received within the predetermined time (step S52). When the response cannot be received within the predetermined time (step S52 / No), the system monitoring unit 62 sets the core management unit to which the system monitoring unit 62 belongs to the “designated core management unit” (step S53).

  When the core management unit to which the self belongs operates as the “designated core management unit” (step S50 / No), when a response to the health check signal is received within a predetermined time (step S52 / Yes), and in step S53. When the process ends, the process of the following step S54 is executed.

The system monitoring unit 62 monitors the state of the core control unit. Specifically, the system monitoring unit 62 transmits a health check signal to the core control unit under the control of the core management unit to which the system monitoring unit 62 belongs (step S54). In this case, the core control unit under management is not necessarily a core control unit belonging to the same physical machine, but may include a core control unit belonging to another physical machine. The system monitoring unit 62 determines whether the response from the core control unit to the health check signal has been received within the predetermined time (step S55). When the response from the core control unit has not been received (step S55 / No), the system monitoring unit 62 sets the state of the core managed by the core control unit in the system core management table 60 to "unusable". Change (step S56). When the process of step S56 is completed and when the response from the core control unit can be received (step S55 / Yes), the process of step S49 is executed again.
(Explanation of effects)
In the third embodiment described above, at least one core management unit designates at least one core from each of at least two physical machines, and assigns the designated multiple cores as virtual CPUs of virtual machines. . That is, in this embodiment, one virtual machine can be built across a plurality of physical machines. As a result, the processing capacity of the virtual machine can be increased to one physical machine or more. Further, it is possible to prevent the load (the number of used cores) from being biased only to a specific physical machine.

  Further, the virtual CPU of the virtual machine is composed of cores of different physical machines. Even if one physical CPU fails, the other physical CPU is unlikely to fail. Therefore, it is possible to prevent the entire virtual machine from going down.

  In summary, according to this embodiment, it is possible to improve the processing capacity and availability of the virtual machine.

  Further, in the case of the present embodiment, the virtual machine is configured by the cores of a plurality of physical machines, but is configured by one system. Therefore, it is not necessary to consider on the virtual OS, and it is possible to multiplex and speed up, or to simplify the scale of circuits and devices.

  Further, in the case of the present embodiment, although the physical machines are straddled, since there is only one virtual OS, normal OS maintenance is sufficient unlike the cluster configuration. That is, according to this embodiment, maintainability can be improved.

  Furthermore, according to the present embodiment, a virtual machine can be constructed in core units without being restricted by a physical machine, and thus it is possible to use a physical machine in which only one core is vacant or an old physical machine. That is, according to this embodiment, resources can be effectively used.

  Moreover, in the present embodiment, cores assigned to virtual machines are managed by the entire system, not by physical machine. Therefore, when a plurality of virtual machines are constructed on a physical machine, they can be constructed seamlessly.

  Although the present invention has been described above using each embodiment, the technical scope of the present invention is not limited to the description of each embodiment. It is obvious to those skilled in the art that various changes or improvements can be added to each of the above embodiments. Therefore, it is needless to say that a form added with such changes or improvements is also included in the technical scope of the present invention. Further, the numerical values, the names of the components, and the like used in each of the embodiments described above are merely examples, and can be changed as appropriate.

1 Virtual Machine Emulation System 2, 3 Physical Machine 4, 5 Core Management Unit 10 Virtual Machine Emulation System 12, 14 Physical Machine 20, 30 Physical CPU
22, 32 Core management unit 24, 34 Core control unit 40A, 40B Virtual CPU
41A First virtual machine 41B Second virtual machine 50 User interface section 52 Virtual CPU processing section 54 Multiplexing processing section 56 Core control communication section 58 System core operation management section 60 System core management table 62 System monitoring section 80 Core management communication section 82 Core communication unit 84 Core operation management unit 86 Core management table 100 Virtual machine emulation system 102 Storage unit 104 Computing unit 150 Computer program

Claims (7)

A plurality of physical machines that collectively manages all of the plurality of cores having respectively, provided with plurality of Turkey A management means to build a virtual machine,
The plurality of core management units respectively construct the different virtual machines, and when constructing the virtual machines, designate at least one core from each of the at least two physical machines, and designate a plurality of designated cores. Has a function of allocating the core as a virtual CPU of the virtual machine,
At least one of the plurality of core managing means is designated as a designated core managing means for executing a process at a normal time, and at least one of the other core managing means is configured to: It is a multiplexed core management means that takes over the processing of the designated core management means and executes it.
The core management means determines whether it is the designated core management means or the multiplexed core management means, and when the self is the multiplexed core management means, When a health check signal is transmitted and a response to the health check signal cannot be received within a predetermined time, the multiplex core management means further has a function of making itself the designated core management means. ..   The core management unit collectively manages all the cores by using a system core management table that stores operation information about all the cores in the virtual machine emulation system. Virtual machine emulation system as described.   The virtual machine emulation system according to claim 2, wherein the operation information includes information on a usage status of the plurality of cores.   The virtual machine emulation system according to claim 3, wherein the operation information includes information about a usage frequency of the plurality of cores. Further comprising core control means provided in each of the plurality of physical machines, for controlling the plurality of cores of the respective physical machines,
The core management unit is optimal among the plurality of cores assigned to the virtual CPU based on the operation information when a processing command is received from the OS or application after the virtual machine is constructed. The virtual machine emulation according to any one of claims 2 to 4, characterized in that the core is selected, and the processing command is transmitted to the core control means that controls the selected core. system. Multiple physical machines to manage collectively a plurality of cores having respectively, a plurality of cores managing means for constructing a virtual machine,
Each of the different virtual machines is constructed, and when the virtual machine is constructed , at least one core is designated from each of the at least two physical machines, and a plurality of designated cores are virtualized by the virtual machine. Assigned as CPU ,
Further, at least one of the plurality of core managing means functions as a designated core managing means for executing a process at a normal time,
When the designated core management unit fails, at least one of the other core management units functions as a multiplexing core management unit that takes over and executes the processing of the designated core management unit,
Further, the core management means itself determines whether it is the designated core management means or the multiplexed core management means, and when it is the multiplexed core management means, the designation by another core management means is performed. A virtual machine emulation method in which a health check signal is transmitted to a core management unit, and when a response to the health check signal cannot be received within a predetermined time, the self is designated as the designated core management unit . A computer of at least two of the plurality of physical machines,
When managing a plurality of cores respectively possessed by the plurality of physical machines and constructing a virtual machine, at least one core is designated from each of the at least two physical machines, and designated. Executing a process of allocating a plurality of the cores as virtual CPUs of the virtual machine ,
In addition, at least one of the computers of the physical machine is caused to function as a designated core management unit that executes processing at normal time, and at least one of the computers of the other physical machines fails the designated core management unit. , Functioning as a multiplexing core management unit that takes over and executes the processing of the designated core management unit,
Further, each of the computers of the physical machines is made to judge whether it is the designated core managing means or the multiplexed core managing means, and when the self is the multiplexed core managing means, the designated core A computer program that transmits a health check signal to the management means and causes itself to function as the designated core management means when a response to the health check signal cannot be received within a predetermined time .

Images