JP2011180889A - Network resource management system, device, method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a network resource management system for automatically managing a network resource in a virtual environment. <P>SOLUTION: The network resource management system includes: a communication amount measuring means 10 for measuring a communication amount among virtual machines in a virtual environment where virtual machines operate in a plurality of physical servers; an expected value calculating means 12 for calculating an expected value of a network use rate when a physical server for operating a prescribed virtual machine is changed to a different physical server on the basis of the communication amount measured by the communication amount measuring means 10; a validity determination means 13 for determining whether or not a change of physical servers for operating the virtual machine is valid on the basis of the expected value calculated by the expected value calculating means 12; and a change means 16 for changing the physical server for operating the virtual machine on the basis of a result of determination of the validity determination means 13. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a network resource management system, a network resource management apparatus, a network resource management method, and a network resource management program that manage physical network resources in a virtual environment.

  Since hardware specifications have improved year by year, it has become easy to consolidate and operate an environment that has been operated by a plurality of servers and personal computers until now. For this reason, in the office, the virtualization of thin clients and the aggregation of multiple servers into physical servers is progressing.

  Normally, when multiple virtual computer environments (VM: Virtual Machine) are constructed and operated on one server, resources such as CPU, memory, network, disk IO, etc. are not exhausted. The design is done. However, unexpected cases may occur when the aggregation rate increases due to an instantaneous increase in load or hardware failure.

  In such a case, for example, a load can be distributed by performing a migration operation of moving a VM operating on a certain physical server to another physical server. Further, for example, Patent Literature 1 describes a system that supports migration work.

JP 2008-299791 A

  Assume a scene where a VM is constructed and operated. Here, multiple VMs run on software (VMM: Virtual Machine Monitor) that manages and provides VMs, and VMMs running on different servers are connected by a physical network (hereinafter referred to as a physical network). It is assumed that

  For example, when sending data from VM-1 running on VMM-A on one server to VM-2 running on VMM-B on a different server, VM-1 → VMM-A → VMM-B → Sends data via the VM-2 route. At this time, in the case of performing communication between VM-1 and VMM-A and communication between VMM-B and VM-2, a virtual network realized by VMM (hereinafter referred to as a virtual network) Because it is the communication in the section connected by, the process is executed only by VMM-A or VMM-B, and the actual physical network is not used.

  On the other hand, in communication between VMM-A and VMM-B, communication is performed via a physical network. A physical network has a lower limit of performance (for example, communication speed) compared to a virtual network. Therefore, when multiple VMs communicate with multiple VMs operating on other VMMs at the same time, the communication volume may reach the limit of the physical network communication capacity, which may cause delays in communication. .

  In such a case, if VM migration is performed using the system described in Patent Document 1, a group of VMs communicating without going through a physical network is prevented from being distributed to different servers. be able to.

  However, in the method described in Patent Document 1, when a manager or the like manually performs a migration operation, only the dependency between VMs is presented, and network resources in a virtual environment are automatically managed. I can't.

  Accordingly, an object of the present invention is to provide a network resource management system, a network resource management device, a network resource management method, and a network resource management program that can automatically manage network resources in a virtual environment.

  The network resource management system according to the present invention is based on a communication amount measuring unit that measures a communication amount between virtual machines in a virtual environment in which a virtual machine operates on a plurality of physical servers, and a communication amount measured by the communication amount measuring unit. The expected value calculation means for calculating the expected value of the network usage rate when the physical server that operates the predetermined virtual machine is changed to a different physical server, and the virtual machine based on the expected value calculated by the expected value calculation means. Including a validity determination unit that determines whether or not the change of the physical server to be operated is appropriate, and a change unit that changes the physical server that operates the virtual machine based on the determination result of the validity determination unit. To do.

  The network resource management device according to the present invention is based on the traffic volume measuring means for measuring the traffic volume between virtual machines in a virtual environment in which the virtual machine operates on a plurality of physical servers, and the traffic volume measured by the traffic volume measuring means. The expected value calculation means for calculating the expected value of the network usage rate when the physical server that operates the predetermined virtual machine is changed to a different physical server, and the virtual machine based on the expected value calculated by the expected value calculation means. A validity determination unit that determines whether or not the change of the physical server to be operated is valid, and a change unit that changes the physical server that operates the virtual machine based on the determination result of the validity determination unit And

  The network resource management method according to the present invention measures a communication amount between virtual machines in a virtual environment where virtual machines operate on a plurality of physical servers, and operates a physical server that operates a predetermined virtual machine based on the measured communication amount. Calculate the expected value of the network usage rate when changing to a different physical server, determine whether the change of the physical server that operates the virtual machine is appropriate based on the calculated expected value, and determine whether the virtual machine is based on the determination result The physical server that operates is changed.

  A network resource management program according to the present invention provides a computer with a communication amount measurement process for measuring a communication amount between virtual machines in a virtual environment in which virtual machines operate on a plurality of physical servers, and a predetermined amount based on the measured communication amount. Expected value calculation processing to calculate the expected value of network usage when the physical server that operates the virtual machine is changed to a different physical server, and whether the change of the physical server that operates the virtual machine based on the calculated expected value is appropriate A validity determination process for determining whether or not, and a change process for changing a physical server on which the virtual machine is operated based on the determination result are executed.

  According to the present invention, network resources can be automatically managed in a virtual environment.

It is a block diagram which shows an example of a structure of the network resource management system by this invention. FIG. 6 is an explanatory diagram illustrating a communication state between an agent on a VMM of each server and a manager 133 on a monitoring server 130. It is a functional block diagram which shows an example of a function structure of a network resource management system. It is explanatory drawing which shows an example of communication between VM. It is explanatory drawing which shows an example of the state in which communication is performed. It is explanatory drawing which shows an example of the state after hot migration. It is explanatory drawing which shows an example of the state after hot migration. It is a flowchart which shows the process example which a network resource management system performs. It is explanatory drawing which shows an example of communication information. It is a block diagram which shows the minimum structural example of a network resource management system.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an example of the configuration of a network resource management system according to the present invention. As shown in FIG. 1, the network resource management system includes a server 100, a server 110, a server 120, a monitoring server 130, a storage 140, an FC (Fibre Channel) switch 150, and a network switch 160.

  Specifically, the server 100 is realized by an information processing apparatus such as a personal computer that operates according to a program, and a NIC (Network Interface Card) 101 and an HBA (Host Bus Adapter) for connecting to a SAN (Storage Area Network). Equipped with 102. In the server 100, the VMM 103 is operating, and on the VMM 103, the VMs 104 to 108 are operating. Note that the VM is specifically realized by a CPU, a memory, or the like of the server 100 that operates according to a program.

  Since the server 110 and the server 120 have the same configuration as the server 100, description thereof is omitted. In this embodiment, an example in which three servers on which VMM is operating is installed will be described. However, the number is not limited to three, and two or more servers may be installed.

  Specifically, the monitoring server 130 is realized by an information processing apparatus such as a personal computer that operates according to a program, and includes the NIC 131. In the monitoring server 130, the OS 132 is operating. The monitoring server 130 has a function of collecting information for monitoring each VM by using a network connected to the network switch 160.

  Specifically, the storage 140 is realized by a storage device such as an optical disk device or a magnetic disk device.

  Next, the manager 133 operating on the monitoring server 130 and the agent operating on each server will be described with reference to FIG. FIG. 2 is an explanatory diagram showing the communication status between the agents (109, 119 and 129) on the VMM of each server and the manager 133 of the monitoring server 130.

  The manager 133 of the monitoring server 130 operates on the VMM of each server by communicating with the agents (109, 119, and 129) operating on the VMM of each server (100, 110, and 120). It has a function of managing the communication status of VMs (104, 105, 114, 115, 124 and 125). For example, the agent 109 measures the communication destinations and communication amounts of the VM 104 and VM 105. The manager 133 manages the communication status of each VM by receiving and storing information indicating the communication destination and the communication amount measured from the agent 109 every predetermined period. The manager 133 is specifically realized by the CPU of the monitoring server 130 that operates according to a program. Each agent is specifically realized by the CPU executing processing according to the VMM.

  Next, the functional configuration of the network resource management system in the present embodiment will be described with reference to FIG. FIG. 3 is a functional block diagram illustrating an example of a functional configuration of the network resource management system. As shown in FIG. 3, the network resource management system includes a communication amount measurement unit 10, a movement cost calculation unit 11, an expected value calculation unit 12, a validity determination unit 13, a resource determination unit 14, and a movement unit 15.

  Specifically, the communication amount measuring means 10 is realized by an agent and a manager. The communication amount measuring means 10 has a function of measuring the communication amount between VMs.

  Specifically, the movement cost calculation means 11 is realized by a CPU of an information processing apparatus that operates according to a program. The movement cost calculation means 11 has a function of calculating resources (for example, CPU, memory, physical network, etc.) necessary for moving the VM as movement costs.

  Specifically, the expected value calculation means 12 is realized by a CPU of an information processing apparatus that operates according to a program. The expected value calculation means 12 has a function of calculating the expected value of the physical network usage when the VM is moved. In addition, the expected value calculation means 12 has a function of calculating an expected value of the physical network usage rate when the VM is moved. For example, the expected value calculation means 12 identifies at least one of the combinations of VMs having the largest communication amount among the communication between VMs operating in different VMMs, and operates the identified VMs in the same VMM. Calculate the expected value of physical network usage when one VM is moved.

  Specifically, the validity determination means 13 is realized by a CPU of an information processing apparatus that operates according to a program. The validity determination means 13 has a function of determining the validity of moving the VM based on the movement cost calculated by the movement cost calculation means 11 and the expected value calculated by the expected value calculation means 12. .

  Specifically, the resource determination unit 14 is realized by a CPU of an information processing apparatus that operates according to a program. The resource determination unit 14 has a function of determining whether or not resources (for example, CPU and memory) of the destination VMM are exhausted by moving the VM. Whether the resource determination unit 14 can avoid running out of resources of the migration destination and the migration source VMM by moving VMs operating on different VMMs (hereinafter referred to as VM exchanges). It has a function to judge. Note that resource depletion in the present embodiment refers to a state where the amount of resources necessary for the server to perform each process is insufficient, for example.

  Specifically, the moving means 15 is realized by a CPU of an information processing apparatus that operates according to a program. The moving unit 15 has a function of executing a process of moving to another VMM, for example, using a technique called hot migration without stopping the operating VM. Note that moving the VM means changing the VMM that operates the VM from the VMM that is currently operating to a different VMM.

  Next, the operation of the network resource management system will be described with reference to FIGS.

  FIG. 4 is an explanatory diagram illustrating an example of communication between VMs. Each server illustrated in FIG. 4 performs communication via the network switch 160 from each NIC. Communication between servers is performed via a physical network. In FIG. 4, communication via the physical network is represented by a solid line.

  As shown in FIG. 4, data transmitted from the VM always passes through the virtual switch. In addition, when data is transmitted from the VM to the outside (that is, when data is transmitted to a VM or the like operating on another VMM via a physical network), the transmitted data is mounted on the server. The network switch 160 is reached from the existing NIC.

  In communication between VMs operating on the same VMM, only a virtual switch is used, and therefore a physical network is not used. In FIG. 4, communication other than the physical network (that is, communication in the virtual network) is represented by a broken line.

  FIG. 5 is an explanatory diagram illustrating an example of a state in which communication is performed. In the example shown in FIG. 5, a large amount of data is communicated between VM-A 205 and VM-B 214, and the communication path is represented by a bold line.

  Communication using a virtual network is performed between VM-A 205 and NIC 201 and between VM-B 214 and NIC 211. On the other hand, communication via the actual physical network is performed between the virtual switch A207 and the virtual switch B217.

  Here, since the amount of communication between the VM-A 205 and the VM-B 214 is very large, it is expected that the communication speed of all other VMs that perform communication using the NIC 201 or the NIC 211 will decrease. This is because multiple VMs use a physical network that has a lower communication speed limit than a virtual network via a single or a small number of NICs. Because it interferes with the communication of other VMs.

  In this embodiment, the manager uses the agent to monitor the communication status, and specific VMs have been using a large amount of bandwidth for a long time, and after that, it can be expected that a large amount of communication will be performed regularly thereafter. At that time, the VMs are controlled to operate on the same VMM.

  At this time, it is assumed that the VMM resources (for example, CPU, memory, disk IO, etc.) at the aggregation destination are sufficiently free. In this embodiment, the manager can move without stopping the VM by hot migration.

  FIG. 6 is an explanatory diagram showing a state after hot migration. In the example shown in FIG. 6, VM-A206 and VM-B214 are exchanged from the state shown in FIG. In the state shown in FIG. 6, VM-A 205 and VM-B 214 communicate using a large amount of bandwidth, but this is communication within virtual switch A. Therefore, the communication between VM-A 205 and VM-B 214 does not affect the physical network (communication between NIC 201, NIC 211, NIC 221 and network switch 160).

  Therefore, by shifting from the state shown in FIG. 5 to the state shown in FIG. 6, it is possible to expect a reduction in the bandwidth usage rate of the entire physical network. Further, when only the virtual network is used, communication can be performed regardless of the status of communication using the physical network of another VM, so that the throughput of the VM-A 205 and the VM-B 214 can be improved. In this embodiment, VM-A206 and VM-B214 are replaced in order to operate VM-A205 and VM-B214 on the same VMM. However, the present invention is not limited to this, as shown in FIG. , Either VM-A205 or VM-B214 may be simply moved. Further, for example, the VM-A 205 and the VM-B 214 may be moved to the VMM-C and operated.

  FIG. 8 is a flowchart showing an example of processing executed by the network resource management system. Here, it is assumed that the VM is operating in the virtualization environment shown in FIG. Further, as shown in FIG. 2, it is assumed that the manager 133 of the monitoring server 130 manages the communication status of each VM.

  In the virtual environment, each agent (209, 219, and 229) continuously measures the traffic volume and communication destination of the VM operating on the VMM. Then, the manager 133 receives communication information indicating the communication destination and the communication amount of each VM for each predetermined period from each agent (step S1). Next, the manager 133 calculates the usage amount of the physical network based on the received communication information (step S2).

  FIG. 9 is an explanatory diagram illustrating an example of communication information. In the tables shown in FIGS. 9A to 9C, the portion surrounded by a thick line is communication between VMs operating on the same VMM. That is, a portion surrounded by a bold line is communication that is not used without going through a physical network.

  Upon reception of communication information (for example, information shown in FIG. 9A), the manager 133 calculates a communication amount between different VMMs as a physical network usage amount. For example, the manager 133 calculates the total amount of traffic between the VM running on the VMM-A203 and the VM running on the VMM-B213, thereby communicating between the VMM-A203 and the VMM-B213. Calculate the amount.

  In the example shown in FIG. 9A, the manager 133 calculates 830 as a value indicating the usage amount of the physical network between the VMM-A 203 and the VMM-B 213. Further, the manager 133 calculates a value indicating the amount of use of the physical network between the VMM-A 203 and the VMM-C 223 as 412. Further, the manager 133 calculates 1108 as a value indicating the usage amount of the physical network between the VMM-B213 and the VMM-C223. Then, the manager 133 calculates a value indicating the total usage amount of the physical network as 2350. It can be said that the processing of steps S1 and S2 is executed by the traffic measuring means 10 realized by the agent and the manager.

  Next, the movement cost calculation means 11 calculates a resource (for example, CPU, memory, etc.) necessary for moving a predetermined VM to another VMM as the movement cost. In addition, the expected value calculation unit 12 specifies a combination of VMs having the largest communication amount among the communication between VMs operating in different VMMs, and at least one of them is operated so that the specified VMs operate on the same server. The expected value of the physical network usage that is reduced when one VM is moved is calculated (step S3). In this embodiment, an example is described in which the expected value calculation unit 12 calculates the expected value of the physical network usage when the VM is moved. However, the expected value of the physical network usage rate may be calculated. Good. In this case, the expected value calculation means 12 can calculate the value obtained by dividing the calculated expected value of the usage amount of the physical network by the communication capacity of the physical network as the expected value of the physical network usage rate.

  For example, the movement cost calculation unit 11 specifies a combination having the largest communication amount among communication between VMs via a physical network. In the example shown in FIG. 9A, the movement cost calculation means 11 identifies VM-B216 and VM-C225 as the combination having the largest communication volume via the physical network.

  Next, the movement cost calculation means 11 calculates the movement cost when exchanging VM-C225 with VM-B214 or VM-B215 and the movement cost when exchanging VM-B216 with VM-C224 or VM-C226. To do. Although an example of exchanging VMs will be described here, the movement cost calculation means 11 includes a movement cost for simply moving VM-C225 to VMM-B and a movement cost for moving VM-B216 to VMM-C. May be calculated.

  Next, the expected value calculation means 12 uses the physical network usage reduced when VM-C225 is replaced with VM-B214 or VM-B215, or when VM-B216 is replaced with VM-C224 or VM-C226. The expected value of each is calculated.

  For example, when the expected value calculation means 12 replaces VM-B216 and VM-C226, as shown in FIG. 9B, the use of the physical network between VMM-A203 and VMM-B213 The expected value of quantity is calculated as 788. In addition, the expected value calculation means 12 calculates the expected value of the physical network usage amount between the VMM-A 203 and the VMM-C 223 as 454. In addition, the expected value calculation means 12 calculates the expected value of the physical network usage amount between VMM-B213 and VMM-C223 as 405. Then, the expected value calculation means 12 calculates the expected value of the total usage of the physical network as 1647.

  Next, the validity determination unit 13 determines whether or not to move the VM based on the movement cost calculated by the movement cost calculation unit 11 and the expected value calculated by the expected value calculation unit 12 (step S4).

  For example, the validity determination unit 13 determines that the VM should not be moved when the movement cost exceeds a predetermined threshold. Further, for example, the validity determination unit 13 determines that the movement cost should not be moved or exchanged for the VM whose movement cost does not exceed a predetermined threshold and the expected value calculated by the expected value calculation unit 12 is the lowest. . That is, the validity determination means 13 determines that the movement is appropriate when the movement cost does not exceed a predetermined threshold and the amount of use of the physical network is the smallest.

  If it is determined in step S4 that the VM should be moved, the resource determination unit 14 determines whether the movement destination VMM resource is depleted by moving the VM (step S5). For example, the resource determination unit 14 receives resource information indicating the resource from each agent, and determines whether or not the resources of the migration destination VMM are depleted based on the received resource information.

  If it is determined in step S5 that the resources of the migration destination VMM will not be exhausted, the migration means 15 performs hot migration to migrate a predetermined VM to another VMM (step S7).

  Further, when it is determined in step S5 that the resources of the migration destination VMM are exhausted, the resource determination unit 14 determines whether or not resource exhaustion can be avoided by replacing the VM (step S6). . For example, the resource determination unit 14 determines whether or not resource depletion can be avoided based on the resource information of the migration destination VMM and the resource information of the migration source VMM.

  If it is determined in step S6 that resource depletion cannot be avoided, the network resource management system shifts the processing to step S3 and repeats the processing from step S3 onward.

  On the other hand, if it is determined in step S6 that resource depletion can be avoided, the moving means 15 replaces the predetermined VM with a VM operating in another VMM (step S8). Here, the moving means 15 exchanges VM-B216 and VM-C226. That is, the moving means 15 moves VM-B216 from VMM-B213 to VMM-C223, and moves VM-C226 from VMM-C223 to VMM-B213.

  Thereafter, the network resource management system repeats the processing from step S3 to step S7 or S8 until it is determined in step S4 that the VM should not be moved. For example, the moving means 15 exchanges VM-A 205 and VM-B 214 from the state shown in FIG. 9B (FIG. 9C).

  In this way, depending on the availability of VMM resources, the VM can be moved to one VMM or moved (exchanged) with each other as shown in FIG. 5 and FIG. VMs can be moved without depleting resources other than network such as memory. If there is a possibility that resources other than the VMM network may be exhausted by moving the VM, the VM is not moved. These determinations are made by the monitoring server 130 (for example, the validity determination unit 13 and the resource determination unit 14).

  As described above, in this embodiment, the network resource management system includes a plurality of VMMs, a plurality of VMs operating on the VMM, and a monitoring server. Then, the monitoring server collects communication information of each VM from an agent operating on the VMM. The VMM also allocates resources such as CPU and memory to the VM, determines the priority and load status of other VMs, and reallocates resources in a timely manner.

  In addition, the manager on the monitoring server communicates with an agent that operates on the VMM, identifies a VM that is communicating in large quantities, and executes processing so that the VM operates on the same VMM. The network resource management system executes this process on the assumption that the resources of the aggregation destination VMM will not be exhausted, and does not execute the process if it is determined that the resources will be insufficient due to aggregation. Therefore, it can be said that the network resource management system has the following effects.

  The first effect is that, when specific VMs use a large amount of network bandwidth, it is possible to automatically reduce the usage rate of physical network resources. The reason is that VM-a and b communicate using a virtual network only by operating VMs (hereinafter referred to as VM-a and b) on the same VMM. This is because the physical network is not used.

  The second effect is that the communication speed of VM-a, b and the communication speed of VMs other than VM-a, b can be improved. The reason is that the VM-a, b communicates using only the virtual network due to the first effect, so that the processing can be performed without being influenced by the communication of other VMs. In addition, the physical network bandwidth hindered by VM-a, b is released, and as a result, the influence of VM-a, b on the communication of other VMs is reduced. This is because the communication speed of VMs other than a and b is improved.

  From the above, it can be said that the present embodiment includes the following solution to the problem to be solved.

  In the case of communication between VMs operating on different VMMs, a physical network is used, but in the case of communication between VMs operating on the same VMM, the physical network is not used. Processing is only on VMM. For this reason, in the present embodiment, for communication between VMs as much as possible, a virtual network is used and control is performed so as to reduce the amount of use of VMM physical network resources.

  For example, a VM that plays the role of a server on which an application is running (a computer that provides a specific function to a client computer) and a VM that takes the role of a database server are on different VMMs. Assuming that it is operating in

  Assuming that the application and the database regularly perform a large amount of communication, it is assumed that each VM operates between different VMMs, and thus uses a large amount of physical network bandwidth between VMMs.

  In such a case, VMs that perform a large amount of communication are operated with the same VMM by hot migration (a technique for moving from VMM to another VMM without stopping the VM). As a result, the two VMs operate on the same VMM and use a virtual network, so that communication can be performed without using a physical network.

  Monitoring of the communication status is performed by placing an agent on the VMM and sending the CPU, memory, disk IO, and network bandwidth usage per unit time to the monitoring server. Therefore, it is not necessary for the administrator or the like to manually perform the migration work, and network resources can be managed automatically.

  Next, the minimum configuration of the network resource management system according to the present invention will be described. FIG. 10 is a block diagram illustrating a minimum configuration example of the network resource management system. As shown in FIG. 10, the network resource management system includes a communication amount measuring unit 10, an expected value calculating unit 12, a validity determining unit 13, and a changing unit 16 as minimum components.

  In the network resource management system with the minimum configuration shown in FIG. 10, in a virtual environment in which virtual machines operate on a plurality of physical servers, the traffic volume measuring means 10 continuously measures the traffic volume between virtual machines. Then, the expected value calculation unit 12 calculates the expected value of the network usage rate when the physical server that operates the predetermined virtual machine is changed to a different physical server based on the communication amount measured by the communication amount measurement unit 10 . Next, the validity determination unit 13 determines whether or not the change of the physical server that operates the virtual machine is appropriate based on the expected value calculated by the expected value calculation unit 12. Next, the changing unit 16 changes the physical server that operates the virtual machine based on the determination result of the validity determining unit 13.

  Therefore, according to the network resource management system having the minimum configuration, network resources can be automatically managed in a virtual environment.

  In this embodiment, the characteristic configuration of the network resource management system as shown in the following (1) to (5) is shown.

  (1) A network resource management system includes a traffic volume measuring unit (for example, a traffic volume measuring unit) that measures a traffic volume between virtual machines in a virtual environment in which a virtual machine (for example, VM 104) operates on a plurality of physical servers (for example, server 100) Based on the traffic (for example, communication information) measured by the traffic measuring means 10 and the physical server that operates the predetermined virtual machine is changed to a different physical server. Expected value calculation means for calculating the expected value of the network usage rate (for example, realized by the expected value calculation means 12), and change of the physical server that operates the virtual machine based on the expected value calculated by the expected value calculation means Based on the judgment result of the validity judgment means (for example, realized by the validity judgment means 13) and the judgment result of the validity judgment means Changing means for changing the physical server operating the machine (e.g., as implemented by the moving means 15), characterized in that it comprises a.

  (2) In a network resource management system, resources required to change a physical server that operates a given virtual machine from a currently operating physical server to a different physical server (for example, CPU, memory, disk IO, etc.) When the travel cost calculated by the travel cost calculation means exceeds a predetermined threshold, the travel cost calculation means (for example, realized by the travel cost calculation means 11) is calculated. May be configured to determine that the change of the physical server that operates the virtual machine is not appropriate.

  (3) In the network resource management system, the expected value calculation means includes a combination of virtual machines having the largest traffic volume among communication between virtual machines operating on different physical servers (for example, as shown in FIG. 9A). The combination of VM-B216 and VM-C225) is specified, and the expected value of the network usage rate is calculated when the physical server that operates the virtual machine is changed so that the specified virtual machine operates on the same physical server The validity determination unit may be configured to determine that the change of the physical server that operates the virtual machine is appropriate when the expected value of the network usage rate calculated by the expected value calculation unit is the lowest. Good.

  (4) In a network resource management system, when a physical server that operates a predetermined virtual machine is changed from a currently operating physical server to a different physical server, it is determined whether or not the resources of the different physical servers are depleted ( For example, step S5 shown in FIG. 8) includes a resource determination unit (for example, realized by the resource determination unit 14), and the change unit changes the physical server that operates the virtual machine based on the determination result of the resource determination unit. It may be configured to change to a physical server (for example, as shown in FIG. 7, VM-B 214 is moved to VMM-A 203).

  (5) In the network resource management system, the resource determination means changes the resource of the second physical server when the physical server that operates the predetermined virtual machine is changed from the first physical server that is currently operating to the second physical server. When the resource is exhausted, it is determined whether or not the resource exhaustion can be avoided by exchanging the predetermined virtual machine with another virtual machine operating on the second physical server (for example, step S6 shown in FIG. 8). The changing unit changes the physical server that operates the predetermined virtual machine to the second physical server based on the determination result of the resource determining unit, and changes the physical server that operates the other virtual machine to the first physical server. It may be configured to change to a server (for example, as shown in FIG. 6, VM-A 204 and VM-B 214 are exchanged).

  A part or all of the above embodiments can be described as in the following supplementary notes, but is not limited thereto.

(Additional remark 1) Let a computer perform the movement cost calculation process which calculates a resource required when changing the physical server which operates a predetermined virtual machine from the physical server currently operated to a different physical server as a movement cost The network resource according to claim 10, wherein, in the validity determination process, when the calculated movement cost exceeds a predetermined threshold value, a process for determining that the change of the physical server that operates the virtual machine is not valid is executed. Management program.

  The present invention can be applied to an application for reducing the physical network bandwidth in a situation where a specific VM uses a large amount of network bandwidth and the physical network is depleted and communication is delayed.

DESCRIPTION OF SYMBOLS 10 Communication volume measurement means 11 Movement cost calculation means 12 Expected value calculation means 13 Validity determination means 14 Resource determination means 15 Movement means 16 Change means

Claims (10)

A traffic measuring means for measuring the traffic between the virtual machines in a virtual environment in which the virtual machines operate on a plurality of physical servers;
An expected value calculating unit that calculates an expected value of the network usage rate when the physical server that operates the predetermined virtual machine is changed to a different physical server, based on the communication amount measured by the communication amount measuring unit;
Based on the expected value calculated by the expected value calculating means, validity determining means for determining whether or not the change of the physical server that operates the virtual machine is appropriate;
A network resource management system, comprising: a change unit that changes a physical server that operates the virtual machine based on a determination result of the validity determination unit. Including a movement cost calculation means for calculating a resource required for changing a physical server that operates a predetermined virtual machine from a physical server that is currently operating to a different physical server as a movement cost,
The network according to claim 1, wherein the validity determination unit determines that the change of the physical server that operates the virtual machine is not valid when the movement cost calculated by the movement cost calculation unit exceeds a predetermined threshold. Resource management system. The expected value calculation means specifies a combination of virtual machines having the largest communication volume among the communication between virtual machines operating on different physical servers, and operates the specified virtual machines on the same physical server. To calculate the expected value of the network usage rate when the physical server that operates the virtual machine is changed,
The validity determination unit determines that the change of the physical server that operates the virtual machine is valid when the expected value of the network usage rate calculated by the expected value calculation unit is the lowest. The described network resource management system. When a physical server that operates a predetermined virtual machine is changed from a physical server that is currently operating to a different physical server, resource determination means that determines whether or not the resources of the different physical server are depleted,
The network resource according to any one of claims 1 to 3, wherein the changing unit changes a physical server that operates the virtual machine to the different physical server based on a determination result of the resource determining unit. Management system. The resource determining means is configured to change the predetermined physical server when the resource of the second physical server is depleted when the physical server that operates the predetermined virtual machine is changed from the first physical server currently operating to the second physical server. Determining whether or not the resource depletion can be avoided by exchanging the virtual machine with another virtual machine operating on the second physical server,
The changing means changes the physical server that operates the predetermined virtual machine to the second physical server based on the determination result of the resource determining means, and changes the physical server that operates the other virtual machine to the first physical server. The network resource management system according to claim 4, wherein the network resource management system is changed to one physical server. A traffic measuring means for measuring the traffic between the virtual machines in a virtual environment in which the virtual machines operate on a plurality of physical servers;
An expected value calculating unit that calculates an expected value of the network usage rate when the physical server that operates the predetermined virtual machine is changed to a different physical server, based on the communication amount measured by the communication amount measuring unit;
Based on the expected value calculated by the expected value calculating means, validity determining means for determining whether or not the change of the physical server that operates the virtual machine is appropriate;
A network resource management apparatus comprising: a changing unit that changes a physical server that operates the virtual machine based on a determination result of the validity determining unit. A movement cost calculation means for calculating a resource required for changing a physical server that operates a predetermined virtual machine from a currently operating physical server to a different physical server as a movement cost;
The network according to claim 6, wherein the validity determination means determines that the change of the physical server that operates the virtual machine is not valid when the movement cost calculated by the movement cost calculation means exceeds a predetermined threshold. Resource management device. Measure the amount of communication between the virtual machines in a virtual environment where the virtual machines operate on multiple physical servers,
Based on the measured traffic, calculate the expected value of the network usage rate when the physical server that operates a given virtual machine is changed to a different physical server,
Based on the calculated expected value, determine whether the change of the physical server that operates the virtual machine is appropriate,
A network resource management method, comprising: changing a physical server that operates the virtual machine based on a determination result. Calculate the resources required to change the physical server that runs a given virtual machine from the currently running physical server to a different physical server as the migration cost,
The network resource management method according to claim 8, wherein when the calculated movement cost exceeds a predetermined threshold, it is determined that the change of the physical server that operates the virtual machine is not appropriate. On the computer,
A traffic measurement process for measuring the traffic between the virtual machines in a virtual environment where the virtual machines operate on a plurality of physical servers;
An expected value calculation process for calculating an expected value of the network usage rate when the physical server that operates the predetermined virtual machine is changed to a different physical server based on the measured traffic;
A validity determination process for determining whether or not the change of the physical server that operates the virtual machine is appropriate based on the calculated expected value;
A network resource management program for executing change processing for changing a physical server that operates the virtual machine based on a determination result.

Images