JP2011061598A - Overlay network management device, physical node device, network control device, method and program for managing overlay network - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To distribute a processing load generated when a failure occurs, across a plurality of resources in a physical node. <P>SOLUTION: When an overlay network management part 1 newly receives a request to construct an overlay network, a resource usage calculation part 14 calculates an expected value of an amount of resources used by each CPU when the failure occurs in a component on a physical network used by a newly constructed overlay network, and a resource usage comparison part 15 performs a comparison of the expected values of the resource usage of each CPU that is calculated by the resource usage calculation part 14 and selects the CPU whose expected value is minimum as the CPU of which a transfer process part of the newly constructed overlay network is composed. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an overlay network management device, a physical node device, a network control device, an overlay network management method, and an overlay network management program, and more particularly to an overlay network management device and an overlay network that can distribute a processing load generated at the time of failure to a plurality of resources in a physical node. The present invention relates to a management method and an overlay network management program.

  In a network (NW) such as the Internet, with the diversification of user requests, it is desired to construct a variety of networks having characteristics according to each user request. As a technique for realizing such various networks, there is an overlay network technique for constructing a plurality of logical networks using a physical network. In an overlay network, a virtual node (virtual node, Overlay Node) is built on an actual physical node (physical node, Underlay Node), and this virtual node is connected. Provide an overlay network. In this virtual node, a network having desired characteristics can be constructed by limiting the bandwidth of the link.

FIG. 28 is a diagram illustrating an example of an overlay network. In the figure, a physical network (underlay network) 100 includes physical nodes 101 to 105 and links connecting them. Overlay networks 110, 120, and 130 are constructed using this physical network 100. The overlay network 110 is configured to include virtual nodes 111 to 114 and links connecting them. The overlay network 120 includes virtual nodes 121 to 124 and links connecting them. The overlay network 130 includes virtual nodes 131 to 133 and links connecting them.
In the overlay network, a necessary node is selected from physical nodes constituting the physical network, and a virtual router is constructed using the selected node. Then, an overlay network is constructed by connecting these virtual routers. Which physical node is selected differs for each overlay network.

FIG. 29 is a diagram illustrating a configuration example of a physical node. The physical node 101c includes a switch 171 that controls a transfer destination of data transmitted through the network, and CPUs (Central Processing Units) 161-1 to 161-N (N is a CPU) for operating software that controls the switch. A positive integer). The CPUs 161-1 to 161-N execute virtualization software 181-1 to 181-N, respectively. By executing these virtual software, a virtual machine is configured on the CPUs 161-1 to 161-N. Furthermore, a virtual router operates in the virtual machine. The virtual router corresponds to the virtual node of the overlay network. Depending on which physical node each overlay network selects, which overlay network has a virtual router on the physical node differs. For example, the virtual router of the overlay network 110 and the virtual router of the overlay network 120 exist on the physical node 104, but the virtual router of the overlay network 130 does not exist.
As shown in FIG. 29, when a plurality of CPUs exist in a physical node and creation of a plurality of virtual machines is required, a mechanism for determining which virtual machine is operated by which CPU is required. For such virtual machine placement problems, the current load on the CPU is monitored and a new virtual machine is created on the CPU with the lowest load for the purpose of making the load among CPUs as uniform as possible. There exists a technique to perform (see Patent Document 1).

JP 2005-115653 A

However, the conventional technology does not consider load distribution when a failure occurs. Therefore, when a failure occurs, the route recalculation load of the overlay network is concentrated on one CPU, which may increase the time required for route recalculation.
This problem will be described with reference to FIG. The network configuration shown in the figure is the same as the network configuration described in FIG. In FIG. 30, a virtual link (overlay link) between the virtual node 111 and the virtual node 114 uses a physical link (underlay link) between the physical node 101 and the physical node 104. When realized, when a link failure occurs between the physical node 101 and the physical node 104 of the physical network 100, the influence of the failure also affects the overlay network 110, and the virtual link is used without using the physical link where the failure has occurred. It is necessary to recalculate the route to realize The same applies to the case where the virtual link between the virtual node 121 and the virtual node 124 of the overlay network 120 is realized using the physical link between the physical node 101 and the physical node 104.
On the other hand, in the overlay network 130, when each virtual link is realized without using a physical link between the physical node 101 and the physical node 104, route recalculation processing is unnecessary, and processing during normal operation is performed. Can be executed. Under such circumstances, in the physical node 101, the virtual router for the virtual node 111 of the overlay network 110 and the virtual router for the virtual node 121 of the overlay network 120 operate on the same CPU, and the virtual node 131 of the overlay network 130. When the virtual router for the virtual node is operating on another CPU, the virtual router for the virtual node 111 and the virtual router for the virtual node 121 are recalculated on the operating CPU, whereas the virtual router for the virtual node 131 In the CPU on which the virtual router operates, route recalculation does not occur. As described above, when the CPU in which the route recalculation occurs is biased, the route recalculation time may increase.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide an overlay network management apparatus capable of distributing the processing load generated when a failure occurs to a plurality of CPUs in a physical node.

  [1] The present invention has been made to solve the above-described problems, and an overlay network management apparatus according to an aspect of the present invention is any physical node in a physical network including a physical node having a plurality of CPUs or those An overlay network management apparatus that manages an overlay network constructed using components of a physical link that connects physical nodes of the physical node, and is newly constructed when a request to construct a new overlay network is input A resource usage calculation unit that calculates an expected value of a resource amount used by each of the plurality of CPUs when a failure occurs in a component of the physical network used by the overlay network, and the resource usage calculation unit calculates CPU with the smallest expected resource amount Characterized by comprising a resource use amount comparing unit for selecting as the CPU should constitute the transfer processing unit of the overlay networks that valley building, the.

  [2] Further, an overlay network management apparatus according to an aspect of the present invention is the above-described overlay network management apparatus, wherein a failure that causes a failure to occur on a physical network and a failure occurrence probability are stored in association with each other. Occurrence probability database, failure load amount calculation formula database for storing a formula for calculating the resource amount used by the CPU when a failure occurs for each overlay network, and which CPU each overlay network is built with The virtual machine state monitor unit to be identified, the underlay link search unit for determining the physical link path for realizing the virtual link of the newly constructed overlay network, and the components of the physical network used by the already constructed overlay network, Overlay network A risk factor database stored for each, and the resource usage calculation unit is configured to monitor the virtual machine state monitor for each component of a physical network used by a newly constructed overlay network and for each CPU. Among the overlay networks identified as being constructed using the CPU, the overlay network constructed using the constituent elements is detected based on the data of the risk factor database. When a failure occurs in the component, a resource amount used by the CPU for the overlay network is calculated based on an equation stored in the failure load amount calculation formula database, and the calculated resource amount and The failure probability database Based on the failure probability for each component of reading, calculating the expected value of the amount of resources each of which uses a plurality of CPU, and characterized in that.

  [3] An overlay network management apparatus according to an aspect of the present invention is the above-described overlay network management apparatus, in which an overlay network newly constructed is based on a physical link path determined by the underlay link search unit. A risk factor extracting unit that extracts components of a physical network to be used, and the risk factor database is an overlay that has already been constructed with the components of the physical network used by the overlay network extracted by the risk factor extracting unit; As a constituent element of a physical network used by the network, each overlay network is stored.

  [4] An overlay network management apparatus according to an aspect of the present invention is the overlay network management apparatus described above, and communication data of each virtual machine that realizes a virtual node of an already constructed overlay network on the CPU. Each overlay network based on the information indicating the communication destination or transmission source and the information on the physical link path determined by the underlay link search unit. Further comprising a communication monitoring unit that acquires information indicating physical nodes and physical links that realize the information, and the risk factor database includes information indicating physical nodes and physical links that realize each overlay network acquired by the communication monitoring unit. Remember that every overlay network And butterflies.

  [5] An overlay network management apparatus according to an aspect of the present invention is the overlay network management apparatus described above, and acquires information on a physical node that realizes each virtual node from each physical node or each virtual node. In addition, it further includes a switch control monitor unit that monitors transmission data passing through each physical node and acquires information on physical links that realize each virtual link, and the risk factor database is acquired by the switch control monitor unit. The information of the physical node that realizes each virtual node and the information of the physical link that realizes the virtual link are stored for each overlay network.

  [6] A physical node device according to an aspect of the present invention is a physical node device that has a plurality of CPUs and constructs a virtual node of an overlay network, and a request to construct a new overlay network is input. And each of the plurality of CPUs when a failure occurs in any physical node in the physical network used by the newly constructed overlay network or a component of the physical link connecting the physical nodes. A resource usage calculation unit that calculates an expected value of the resource amount and a transfer processing unit of an overlay network that newly constructs a CPU that has the minimum expected value of the resource amount calculated by the resource usage calculation unit should be configured And a resource usage comparison unit that is selected as a CPU.

  [7] A physical node device according to an aspect of the present invention is the above-described physical node device, wherein a failure occurrence probability is stored in association with a failure occurrence probability and a component that may cause a failure on the physical network. Identifies the database, the failure load amount calculation formula database that stores a formula for calculating the resource amount used by the CPU for each overlay network when a failure occurs, and which CPU is used to construct each overlay network A virtual machine state monitor unit, an underlay link search unit that determines a physical link path for realizing a virtual link of a newly constructed overlay network, and components of the physical network used by the already constructed overlay network Risk factor database to be stored for each , And the resource usage calculation unit is configured by the virtual machine state monitor unit using the CPU for each component of the physical network used by the newly constructed overlay network and for each CPU. Among the overlay networks identified as being configured, an overlay network constructed using the component is detected based on the data of the risk factor database, and a failure has occurred in the component for the detected overlay network In this case, a resource amount used by the CPU for the overlay network is calculated based on an equation stored in the failure load amount calculation equation database, and is read from the calculated resource amount and the failure occurrence probability database. Failure occurred in each component Based on the rate, it calculates the expected value of the amount of resources each of which uses a plurality of CPU, and characterized in that.

  [8] A physical node device according to an aspect of the present invention is the physical node device described above, which is a physical node used by an overlay network newly constructed based on a physical link route determined by the underlay link search unit. A risk factor extracting unit that extracts network components; and the risk factor database includes a physical network component used by the overlay network extracted by the risk factor extracting unit and an already constructed overlay network. As a constituent element of the physical network to be used, each overlay network is stored.

  [9] A physical node device according to an aspect of the present invention is the above-described physical node device, and monitors communication data of each virtual machine that realizes a virtual node of an already constructed overlay network on the CPU. The information indicating the communication destination or transmission source is obtained, and each overlay network is realized based on the information indicating the communication destination or transmission source and the physical link route information determined by the underlay link search unit. A communication monitor unit that acquires information indicating a physical node device and a physical link, and the risk factor database includes information indicating the physical node device and physical link that realizes each overlay network acquired by the communication monitor unit. It is memorized for every overlay network.

  [10] A network control apparatus according to an aspect of the present invention is a network control apparatus for controlling an overlay network constructed using a physical network component including a physical node having a plurality of CPUs. When a request for constructing an overlay network is input to the node, an expected amount of resources used by each of the plurality of CPUs when a failure occurs in a component of the physical network used by the newly constructed overlay network A resource usage calculation unit that calculates a value, and a CPU that has a minimum expected value of the resource amount calculated by the resource usage calculation unit is selected as a CPU that should constitute a transfer processing unit of a newly constructed overlay network A failure occurs on the resource usage comparison unit and the physical network. A failure occurrence probability database that associates and stores component elements and failure occurrence probabilities, and a failure load amount calculation equation database that stores equations for calculating the amount of resources used by the CPU when a failure occurs for each overlay network; A virtual machine state monitor unit that identifies which CPU is used to construct each overlay network, an underlay link search unit that determines a physical link path for realizing a virtual link of a newly constructed overlay network, and Obtain information on the physical nodes that realize each virtual node from the risk factor database that stores the components of the physical network used by the constructed overlay network for each overlay network and each physical node or each virtual node. Pass through each physical node A switch control monitor unit that monitors transmission data to be acquired and acquires information on physical links that realize each virtual link, and the resource usage calculation unit is a physical network used by a newly constructed overlay network Among the overlay networks that are identified as being constructed by the virtual machine state monitor unit using the CPU for each component and each CPU, an overlay network constructed using the component is identified as the risk. The amount of resources used by the CPU for the overlay network when a failure occurs in the component of the detected overlay network is detected based on the data in the factor database, and the failure load amount calculation formula database Calculate based on the formula read from And calculating an expected value of the amount of resources used by each of the plurality of CPUs based on the amount of the resource and the failure occurrence probability of each component read from the failure occurrence probability database, and the risk factor database includes the switch The information of the physical node that realizes each virtual node and the information of the physical link that realizes the virtual link acquired by the control monitor unit is stored for each overlay network.

  [11] An overlay network management method according to an aspect of the present invention uses any physical node in a physical network including a physical node having a plurality of CPUs, or a component of a physical link that connects the physical nodes. An overlay network management method of an overlay network management apparatus that manages an overlay network to be constructed, wherein when the overlay network management apparatus receives a request to construct a new overlay network, the overlay network to be newly constructed A resource usage calculation step of calculating an expected value of the resource amount used by each of the plurality of CPUs when a failure occurs in a component of the physical network used by the overlay network management device; A resource usage comparison step of selecting a CPU having the minimum expected value of the resource amount calculated in the resource usage calculation step as a CPU that should constitute a transfer processing unit of a newly constructed overlay network. It is characterized by that.

  [12] An overlay network management method according to an aspect of the present invention is the overlay network management method described above, wherein the overlay network management device includes a component that can cause a failure on a physical network, a failure occurrence probability, Occurrence probability database that stores information in association with each other, failure load amount calculation formula database that stores formulas for calculating the amount of resources used by the CPU when a failure occurs for each overlay network, and an already built overlay network A risk factor database that stores the components of the physical network used by each overlay network, and the overlay network management method includes: A virtual machine state monitoring step that identifies whether the overlay network management device is constructed, and an underlay link search step in which the overlay network management device determines a physical link path that realizes a virtual link of the overlay network to be newly constructed. The resource usage calculation step is constructed using the CPU in the virtual machine state monitoring step for each physical network component used by the newly constructed overlay network and for each CPU. The overlay network constructed using the component is detected based on the data of the risk factor database, and a failure has occurred in the component for the detected overlay network. In this case, the amount of resources used by the CPU for the overlay network is calculated based on the formula stored in the failure load amount calculation formula database, and the calculated resource amount and the fault occurrence probability database are read out. The expected value of the amount of resources used by each of the plurality of CPUs is calculated based on the failure occurrence probability in each component.

  [13] An overlay network management method according to an aspect of the present invention is the overlay network management method described above, based on the physical link path determined by the overlay network management apparatus in the underlay link search step. A risk factor extracting step of extracting a physical network component used by the newly constructed overlay network, wherein the risk factor database is extracted by the risk factor extracting step, and the physical network configuration used by the overlay network is extracted The element is stored for each overlay network as a component of a physical network used by an already constructed overlay network.

  [14] An overlay network management method according to an aspect of the present invention is the overlay network management method described above, wherein the overlay network management apparatus realizes a virtual node of an already constructed overlay network on the CPU. Monitoring communication data of each virtual machine to acquire information indicating a communication destination or transmission source, information indicating the communication destination or transmission source and physical link route information determined in the underlay link search step And a communication monitoring step of acquiring information indicating physical nodes and physical links that realize each overlay network, and the risk factor database realizes each overlay network acquired in the communication monitoring step. Physical nodes and objects Stores information indicating a link for each overlay network, characterized in that.

  [15] An overlay network management method according to an aspect of the present invention is the overlay network management method described above, in which the overlay network management device is configured to physically implement each virtual node from each physical node or each virtual node. The risk factor database further includes a switch control monitoring step of acquiring node information, monitoring transmission data passing through each physical node, and acquiring information on a physical link that realizes each virtual link. The physical node information realizing each virtual node and the physical link information realizing the virtual link acquired in the switch control monitoring step are stored for each overlay network.

  [16] A program according to an aspect of the present invention is constructed using any physical node in a physical network including a physical node having a plurality of CPUs or a component of a physical link that connects the physical nodes. An overlay network management program for managing an overlay network, wherein when a request for constructing a new overlay network is input to a computer, a failure occurs in a component of the physical network used by the newly constructed overlay network. A resource usage calculation step for calculating an expected value of a resource amount used by each of the plurality of CPUs when it occurs, and a CPU having a minimum expected value of the resource amount calculated in the resource usage calculation step, Newly constructed overlay net Characterized in that to execute the resource usage comparison step of selecting as the CPU should constitute transfer processing unit over click, the.

  [17] A program according to an aspect of the present invention is the above-described program, wherein the computer stores a failure occurrence probability in which a component that may cause a failure on a physical network is associated with the failure occurrence probability. A database, a failure load amount calculation formula database for storing a formula for calculating the resource amount used by the CPU when a failure occurs for each overlay network, and components of the physical network used by the already constructed overlay network, A risk factor database stored for each overlay network, a virtual machine state monitoring step for identifying which CPU is used to construct each overlay network in the computer, and a virtual of the newly constructed overlay network Realize the link An underlay link search step for determining a physical link route, and in the resource usage calculation step, for each component of the physical network used by the newly constructed overlay network and for each CPU, Of the overlay network identified as being constructed using the CPU in the virtual machine state monitoring step, the overlay network constructed using the component is detected based on the data of the risk factor database and detected. For the overlay network, when a failure occurs in the component, the amount of resources used by the CPU for the overlay network is calculated based on an equation stored in the failure load amount calculation formula database, Calculated And over scan weight, based on the failure occurrence probability of each component read from the failure occurrence probability database, characterized in that to calculate the expected value of the amount of resources each of which uses a plurality of CPU.

  [18] A program according to an aspect of the present invention is the above-described program, which is used by the overlay network newly constructed on the computer based on the physical link path determined in the underlay link search step. A risk factor extracting step for extracting physical network components is further executed, and the physical network components used by the overlay network extracted in the risk factor extracting step are already built in the risk factor database. As a constituent element of a physical network used by the network, each overlay network is stored.

  [19] A program according to an aspect of the present invention is the above-described program, wherein the computer monitors communication data of each virtual machine that realizes a virtual node of an overlay network that has already been constructed on the CPU. Information indicating the communication destination or transmission source, and based on the information indicating the communication destination or transmission source and the physical link path information determined in the underlay link search step, Further, a communication monitor step for acquiring information indicating physical nodes and physical links to be realized is further executed, and information indicating physical nodes and physical links for realizing each overlay network acquired in the communication monitor step is stored in the risk factor database. Remember each overlay network And features.

  [20] A program according to an aspect of the present invention is the above-described program, wherein the computer acquires information on a physical node that realizes each virtual node from each physical node or each virtual node, and The transmission data passing through each physical node is monitored to further execute a switch control monitor step for acquiring information on the physical link that realizes each virtual link, and the risk factor database is acquired in the switch control monitor step. Information on a physical node that realizes each virtual node and information on a physical link that realizes a virtual link are stored for each overlay network.

  According to the present invention, the processing load generated when a failure occurs can be distributed to a plurality of CPUs in a physical node.

It is a block diagram which shows schematic structure of the physical node in the 1st Embodiment of this invention. It is a block diagram which shows another schematic structure of the physical node in the embodiment. It is a block diagram which shows schematic structure of the overlay network management part 1 in the embodiment. 4 is a flowchart illustrating a processing procedure for selecting a CPU by an overlay network management unit 1 in the embodiment. It is a figure which shows the example of the overlay network construction request | requirement in the embodiment. In the same embodiment, it is a figure which shows the example of virtual machine name information which the input part 10 produces | generates. In the same embodiment, it is a figure which shows the example of the physical link path | route information which the underlay link search part 21 produces | generates. In the embodiment, it is a figure which shows the example of the data which the risk factor database 20 memorize | stores. In the same embodiment, it is a figure which shows the example of the risk factor use overlay network information which the resource competition search part 13 produces | generates. In the same embodiment, it is a figure showing the example of virtual machine realization CPU information which the virtual machine state monitor part 19 outputs. 4 is a diagram illustrating an example of data stored in a failure load amount calculation formula database 18 in the embodiment. FIG. 4 is a diagram illustrating an example of data stored in a failure occurrence probability database in the same embodiment. FIG. It is a block diagram which shows schematic structure of the overlay network management part 2 in the 2nd Embodiment of this invention. In the embodiment, it is a figure which shows the example of the data which the risk factor database 33 memorize | stores. In the embodiment, it is a figure which shows the example of the information which shows the identifier of the virtual node corresponding to the physical node contained in the overlay network construction | assembly request | requirement memorize | stored in a node corresponding relationship memory | storage part. In the same embodiment, it is a figure which shows the example of the information which matches the path | route of a physical link, and a virtual link which the link corresponding relationship memory | storage part memorize | stores. In the same embodiment, it is a figure which shows the example of the risk factor database 33 before and behind the communication monitor part 32 adding a line. In the same embodiment, it is a figure which shows the example of the risk factor database 33 before and after the communication monitor part 32 writes the name of a virtual node, and the name of a physical node. In the same embodiment, it is a figure which shows the example of the risk factor database 33 before and behind the communication monitor part 32 integrating a line. In the same embodiment, it is a figure which shows the example of the risk factor database 33 before and behind the communication monitor part 32 integrating a line. It is a figure which shows the structural example of the network in the 3rd Embodiment of this invention. 2 is a configuration diagram illustrating a schematic configuration of a physical node 201 in the embodiment. FIG. It is a block diagram which shows schematic structure of the network control part 210 in the embodiment. It is a block diagram which shows another schematic structure of the network control part 210 in the embodiment. It is a block diagram which shows schematic structure of the overlay network management part 3 in the embodiment. In the same embodiment, it is a figure which shows the example of the switch control information which the switch control monitor part 41 acquires from a switch. In the same embodiment, it is a figure which shows the example of the switch control information which the switch control monitor part 41 acquires from a transfer process part. It is a figure which shows the example of an overlay network. It is a figure which shows the structural example of a physical node. It is a figure which shows the example which a failure generate | occur | produced on the physical network.

<First Embodiment>
The first embodiment of the present invention will be described below with reference to the drawings.
FIG. 1 is a configuration diagram showing a schematic configuration of a physical node (physical node device) in the first embodiment of the present invention. In the figure, the physical node 101a includes a switch 171 and CPUs 161-1 to 161- (N + 1) (N is a positive integer).
The physical node 101a is one physical node on the physical network.
The switch 171 controls the transfer destination of transmission data through the network and transmits data.
Each of the CPUs 161-1 to 161-N executes virtualization software and constructs a virtual machine. A plurality of virtual machines can be constructed on one CPU. The CPU 161-(N + 1) executes the overlay network management unit 1.

Each virtual machine includes one transfer processing unit. The forwarding processing unit executes route calculation and forwarding table (Forwarding table) creation regarding the overlay network. The transfer processing unit controls data transfer between virtual machines by controlling the switch 171. One transfer processing unit corresponds to one virtual node. In the following, each overlay network is assumed to be composed of virtual nodes constructed on different physical nodes and links between these virtual nodes. Therefore, two or more virtual nodes constructed on the same physical node are not included in the same overlay network. That is, a plurality of virtual machines constructed on the same physical node are associated with any one of different overlay networks.
The CPU 161- (N + 1) implements the overlay network management unit 1 by executing software. As will be described later, the overlay network management unit 1 selects a CPU to be used for constructing a new overlay network from among a plurality of CPUs included in the physical node. In the present embodiment, each of the physical nodes having a plurality of CPUs includes the overlay network management unit 1, and the overlay network management unit 1 selects one of the CPUs included in the physical node including the overlay network. Select CPUs. Hereinafter, the physical node including the overlay network management unit is also referred to as “managed physical node (of the overlay network management unit)”. For example, the management target physical node of the overlay network management unit 1 included in the physical node 101a is the physical node 101a.

The overlay network management unit 1 may be realized on a virtual machine instead of an independent CPU.
FIG. 2 is a configuration diagram showing another schematic configuration of the physical node in the embodiment. In the figure, the physical node 101b includes a switch 171 and CPUs 162-1 to 162-N.
The switch 171 is the same as the switch 171 in FIG.
As with the CPUs 161-1 to 161-N in FIG. 1, the CPUs 162-1 to 162-N execute virtualization software and construct a virtual machine. Further, the CPU 162-N implements the overlay network management unit 1 by executing software on one virtual machine.

Note that the overlay network management unit 1 may be realized as an overlay network management device independent of a physical node. In this case, the overlay network management apparatus is connected to the management target physical node so that the overlay network management apparatus and the management target physical node can communicate with each other. The overlay network management apparatus may be realized by executing a program on a management server independent of the physical node.
In the following, it is assumed that a protocol in which route recalculation is executed in all nodes in the network when the network fails is operating on the overlay network. An example of such a protocol is OSPF (Open Shortest Path First).

FIG. 3 is a configuration diagram illustrating a schematic configuration of the overlay network management unit 1. In the figure, an overlay network management unit 1 includes an input unit 10, a transmission unit 24, an underlay link search unit 21, a risk factor extraction unit 11, a risk factor database (Database) 20, a resource selection unit 12, and an output unit 16. A failure occurrence probability database (DB) 17, a failure load amount calculation formula database (DB) 18, and a virtual machine state monitor unit 19. The resource selection unit 12 includes a resource competition search unit 13, a resource usage calculation unit 14, and a resource usage comparison unit 15.
The input unit 10 receives a request to construct a new overlay network on a physical node (hereinafter also referred to as “overlay network construction request”). Upon receiving the overlay network construction request, the input unit 10 generates a virtual machine name (virtual machine identifier) for identifying a virtual node of the newly constructed overlay network, and constructs the virtual machine name and the virtual machine. Generate virtual machine name information indicating the correspondence with the physical node. Here, a virtual node and a virtual machine that realizes the virtual node have a one-to-one correspondence. Therefore, not only a virtual machine but also a virtual node can be identified using the virtual machine name. The input unit 10 outputs the received overlay network construction request and the generated virtual machine name information to the underlay link search unit 21 and the transmission unit 24.
The transmission unit 24 communicates with the overlay network management unit 1 realized on another physical node.

The underlay link search unit 21 determines the path of the physical link for the virtual link included in the newly constructed overlay network. The underlay link search unit 21 outputs the determined physical link to the risk factor extraction unit 11.
The risk factor extraction unit 11 extracts the physical node included in the overlay network construction request and the physical link determined by the underlay link search unit 21 as a risk factor (failure factor). The risk factor extraction unit 11 writes the extracted risk factor in the risk factor database 20.
The risk factor database 20 stores risk factors extracted by the risk factor extraction unit 11.

The resource selection unit 12 determines which CPU is to operate the newly constructed overlay network transfer processing unit.
In the resource selection unit 12, the resource conflict search unit 13 selects an overlay network that uses the risk factor output from the risk factor extraction unit 11 from among the overlay networks that are already constructed using the management target physical node. Search by risk factor. The overlay network that has already been constructed using the management target physical node is an overlay network in which the CPU of the management target physical node performs route recalculation when a failure occurs. For example, when a failure occurs in any component of the physical network, the target of the path recalculation by the CPU of the physical node 101a includes a virtual node realized using the CPU of the physical node 101a. An overlay network that is only an overlay network and is constructed without using the CPU of the physical node 101a is not subject to route recalculation. Therefore, the resource conflict search unit 13 searches only the overlay network constructed using the management target physical node.
The resource usage calculation unit 14 calculates an expected value of how much resource each CPU uses when a failure occurs.
The resource usage comparison unit 15 compares the expected value E _k of the resource usage of each CPU calculated by the resource usage calculation unit 14 and newly transfers the CPU having the smallest value to the overlay network transfer processing unit. Is selected as the CPU to be realized.

The output unit 16 outputs information representing the CPU selected by the resource selection unit 12.
The failure occurrence probability database (DB) 17 stores risk factors and failure occurrence probabilities in association with each other. The failure load amount calculation formula database (DB) 18 stores a formula for calculating the CPU load amount required when a failure occurs for each overlay network. The failure occurrence probability database 17 and the failure load amount calculation formula database 18 are referred to when the resource use amount calculation unit 14 calculates the resource amount used by each CPU when a failure occurs.
The virtual machine state monitor unit 19 monitors which CPU of which virtual node of which overlay network is operating. The monitoring result of the virtual machine state monitoring unit 19 is referred to when the resource usage calculation unit 14 calculates the resource amount used by each CPU when a failure occurs.

Next, the operation of the overlay network management unit 1 will be described.
FIG. 4 is a flowchart showing a processing procedure for selecting a CPU that realizes a virtual node of an overlay network newly constructed by the overlay network management unit 1.
When the input unit 10 accepts an overlay network construction request input by a user from a terminal device outside the physical node or when the transmission unit 24 accepts an overlay network construction request from another physical node, the overlay network management unit 1 A process of selecting a CPU that realizes a virtual node of a newly constructed overlay network is started.
Here, there are usually a plurality of physical nodes, and the input unit 10 of any physical node accepts an overlay network construction request input by the user. The physical node that has received the overlay network construction request input by the user transmits the overlay network construction request to another physical node, and the transmission unit 24 of the other physical node accepts the transmitted overlay network construction request.

The overlay network construction request stores a physical node used in construction of the overlay network and information indicating which node to create a connection link as an overlay network.
FIG. 5 is a diagram illustrating an example of an overlay network construction request. As shown in the figure, an overlay network name and configuration information are stored in the overlay network construction request.
The overlay network name is a name of an overlay network to be newly constructed, and is information for identifying the overlay network. The configuration information is information indicating information on a physical node that constructs a virtual node and information on a link between the virtual nodes. Information on the link between the virtual nodes is expressed by associating the names of two physical nodes that construct the virtual node. In the example of FIG. 5, the node names of the physical nodes 1, 2, 5,... Are shown in the configuration information. These physical nodes are physical nodes that construct a virtual node of the overlay network 1 indicated by the overlay network name. In the configuration information, physical node 1 and physical node 2, physical node 1 and physical node 5,. As a result, the overlay network 1 includes a virtual node realized using the physical node 1 and a virtual node realized using the physical node 2, and a virtual node realized using the physical node 1 and a physical node. It is shown that there is a virtual link with a virtual node realized using the node 5.

  In step S101, the overlay network management unit 1 determines whether or not the input unit 10 has received the overlay network construction request. If the input unit 10 accepts (step S101: YES), the process proceeds to step S102, otherwise (step S101: NO), that is, if the transmission unit 24 accepts, the transmission unit 24 sends an overlay network construction request. Virtual machine name information, which will be described later, is output to the underlay link search unit 21 and then the process proceeds to step S103.

In step S102, the input unit 102 generates a virtual machine name to be assigned to a virtual machine for realizing a newly constructed overlay network, and generates virtual machine name information.
FIG. 6 is a diagram illustrating an example of virtual machine name information generated by the input unit 10. The figure shows an example in which the physical node 1, the physical node 2, the physical node 4, and the physical node 5 are included in the overlay network construction request. The input unit 10 includes physical nodes 1, 2, 4, 5 are associated with the virtual machine names “virtual machine 141”, “virtual machine name 142”, “virtual machine 144”, and “virtual machine name 145”.
This virtual machine name is also used as an identifier of the virtual node. When the transfer processing unit performs data transmission, the virtual machine name is included in the data and transmitted as the identifier of the source virtual node and the identifier of the destination virtual node.

The input unit 10 outputs the received overlay network construction request and virtual machine name information to the underlay link search unit 21 and the transmission unit 24.
The transmission unit 24 transmits the overlay network construction request and virtual machine name information received from the input unit 10 to the transmission unit 24 of the overlay network management unit 1 operating on another physical node.
An input device that receives an overlay network construction request may be provided separately from the physical node. When this input device accepts an overlay network construction request, it generates virtual machine name information and outputs the overlay network construction request and virtual machine name information to the transmission unit 24 of the overlay network management unit 1 operating on each physical node. To do. In this case, each overlay network management unit 1 does not need to include the input unit 10.

  Next, in step S103, upon receiving an overlay network construction request from the input unit 10 or the transmission unit 24, the underlay link search unit 21 reads a virtual link from the overlay network construction request, and a physical link that realizes each virtual link. Determine the route. As shown in FIG. 5, in the overlay network construction request, the virtual link is indicated by a physical node that realizes virtual nodes at both ends of the link. The underlay link search unit 21 determines a physical link path for setting a virtual link between these physical nodes, and generates physical link path information indicating the determined physical link path.

FIG. 7 is a diagram illustrating an example of physical link path information generated by the underlay link search unit 21. In the example of the figure, it is shown that the virtual link between the virtual node realized using the physical node 1 and the virtual node realized using the physical node 2 is realized using the physical link 1. ing. In addition, it is shown that the virtual link between the virtual node realized using the physical node 1 and the virtual node realized using the physical node 5 is realized using the physical link 4 and the physical link 5. Has been.
The underlay link search unit 21 determines a route having the shortest distance between physical nodes as a physical link route. The method by which the underlay link search unit 21 determines the path of the physical link is not limited to the above method. For example, the underlay link search unit 21 may determine a path with a low physical link bandwidth usage as a physical link path. However, a plurality of overlay network management units 1 need to share the same physical link path information. Therefore, each overlay network management unit 1 uses the same physical link path determination method. Note that one overlay network management unit 1, for example, the overlay network management unit 1 that has received an overlay network construction request, may generate physical link path information and transmit it to each overlay network management unit 1. Thereby, each overlay network management unit 1 can share the same physical link path information.
The underlay link search unit 21 outputs the virtual machine name information, physical link path information, and overlay network construction request to the risk factor extraction unit 11.

Next, in step S104, when the risk factor extraction unit 11 receives an overlay network construction request or the like from the underlay link search unit 21, the component that becomes a risk factor from the output overlay network construction request and physical link path information. Is read.
The risk factor extraction unit 11 reads out all physical nodes and all physical links included in the newly constructed overlay network as risk factors. For example, the risk factor extraction unit 11 reads out the physical nodes 1, 2, 5,... Included in the network construction request in FIG. Further, the risk factor extraction unit 11 reads the physical link 1 from FIG. 7 as a physical link that realizes the virtual link indicated by the physical node 1 and the physical node 2 in the network construction request of FIG. 5 and adds it to the risk factor. Similarly, for the other virtual links included in the network construction request, the risk factor extraction unit 11 adds a physical link that realizes the virtual link as a risk factor.
Note that, as a smaller granularity, a switch port or the like to be allocated to the virtual link may be included in the risk factor.
The risk factor extraction unit 11 writes the read risk factor in the risk factor database 20. Further, the risk factor extraction unit 11 reads the read risk factor (risk factor of the newly constructed overlay network), virtual machine name information, and the name of the newly constructed overlay network (the overlay network name included in the overlay network construction request). Is output to the resource conflict search unit 13.

FIG. 8 is a diagram illustrating an example of data stored in the risk factor database 20. As shown in the figure, the risk factor database 20 stores an overlay network name and risk factor information in association with each other.
The overlay network name is the name of an overlay network that has already been constructed or newly constructed, and is information that identifies the overlay network. The risk factor information is information indicating a risk factor for each overlay network. In the example of FIG. 8, the overlay network 1 indicates that the physical node 1, the physical node 2, and the physical link 5 are included as risk factors. By using the risk factor database 20, it is possible to search for an overlay network in which a route recalculation occurs when a failure occurs in a certain risk factor (for example, the physical node 1).

Next, in step S105, when the resource contention search unit 13 receives a risk factor or the like from the risk factor extraction unit 11, the resource contention search unit 13 searches for an overlay network that has already been constructed using the management target physical node, and The overlay network using the risk factor received from the risk factor extraction unit 11 is searched for each risk factor, and the risk factor use overlay network information is generated.
Specifically, first, the resource conflict search unit 13 includes, as risk factors, a managed physical node, that is, a physical node in which the resource conflict search unit 13 is included, from the overlay network stored in the risk factor database 20. The overlay network is searched as an overlay network that has already been constructed using the management target physical node. For example, the resource conflict search unit 13 of the overlay network management unit 1 included in the physical node 1 uses an overlay network that includes the physical node 1 as the risk factor as an overlay network that has already been constructed using the managed node. Search from the overlay network stored in the database 20. Next, for each risk factor received from the risk factor extraction unit 11, the resource conflict search unit 13 creates an overlay network that includes the risk factor as a risk factor in an overlay network that has already been constructed using the managed physical node. Search further from inside. For example, when the physical factor 2 is included in the risk factor output by the risk factor extracting unit 11, the overlay network using the physical node 2 is stored in the overlay network stored in the risk factor database 20, and the above-mentioned The search further searches from the overlay network that has already been constructed using the managed physical node.
The searched overlay network is an overlay network in which the CPU of the managed physical node (physical node 1 in the above example) performs route recalculation when a failure occurs in each risk factor of the newly constructed overlay network. .

FIG. 9 is a diagram illustrating an example of the risk factor use overlay network information generated by the resource conflict search unit 13. In the figure, for example, overlay networks 1, 2,... Are realized using the physical node 1. The resource competition search unit 13 reads the correspondence relationship between the overlay network and the risk factor from the risk factor database 20 for each risk factor, and generates risk factor use overlay network information.
The resource conflict search unit 13 outputs the risk factor use overlay network information and the virtual machine name information to the resource usage calculation unit 14.

When the resource conflict search unit 13 searches the risk factor database 20, the risk factor database 20 stores risk factors of the already constructed overlay network (configuration of physical network used by the already constructed overlay network). In addition to the element), the risk factor of the newly constructed overlay network is also registered by the risk factor extraction unit 11. Therefore, when the resource conflict search unit 13 searches the risk factor database 20, a line including the name of the overlay network to be newly constructed (the overlay network name included in the overlay network construction request) in the overlay network column is searched. Exclude from Thereby, the resource competition unit 13 can acquire the information of the already constructed overlay network from the risk factor database 20.
Alternatively, after the overlay network is constructed based on the overlay network construction request, the risk factor extracting unit 11 may write the risk factor of the overlay network in the risk factor database 20. For example, after the overlay network is constructed, the overlay network management unit 1 receives a signal indicating that the overlay network is constructed. After the overlay network management unit 1 receives this signal, the resource competition unit 13 writes the risk factor of the overlay network in the risk factor database 20.

Next, in step S106, when the resource usage calculation unit 14 receives risk factor use overlay network information or the like from the resource conflict search unit 13, the expected value of how much resource each CPU uses when a failure occurs. Is calculated.
The resource usage calculation unit 14 calculates the amount of load generated in each CPU as the resource usage of each CPU when a failure occurs. In addition, you may make it use quantities other than the load amount of CPU, such as using the memory amount which each CPU uses for path | route recalculation as a resource amount. The resource usage calculation unit 14 refers to the failure occurrence probability database 17 and the failure load calculation formula database 18 to calculate the load generated in each CPU when a failure occurs.
First, the resource usage calculation unit 14 reads out the risk factor of the newly constructed overlay network from the risk factor column of the risk factor use overlay network information received from the resource conflict search unit 13, and fails as one of the read risk factors. The amount of load U _{k, i} generated in each CPU when the error occurs is calculated for each risk factor based on equation (1).

Here, k is an identifier for identifying the CPU. I represents a risk factor that assumes the occurrence of a failure. Further, j represents an overlay network that shares a target risk factor among overlay networks that have been constructed using the target physical node. That is, j represents an overlay network in which a route recalculation must be executed when a failure occurs due to a target risk factor. ρ _j is a CPU load necessary for the overlay network j when a failure occurs due to a target risk factor. The resource use amount calculation unit 14 outputs the CPU load amount from the virtual machine state monitor unit 19 and outputs virtual machine realization CPU information indicating which CPU is operating on which CPU, and a load amount calculation formula at the time of failure Calculation is performed using the formula stored in the database 18.

FIG. 10 is a diagram illustrating an example of virtual machine realization CPU information output from the virtual machine state monitor unit 19.
In the example of the figure, the virtual machines 111 and 141 are operating on the CPU 1 and the virtual machine 121 is operating on the CPU 2.
The virtual machine state monitoring unit 19 monitors which CPU of which overlay network virtual node is operating, and generates virtual machine realization CPU information. The virtualization software on each CUP manages information on the operating virtual machine and provides an interface for acquiring this information. The virtual machine state monitor unit 19 acquires information indicating which virtual machine is operating on which CPU from the virtualization software through this interface. The virtual machine information monitor unit 19 outputs the generated virtual machine realization CPU information to the resource usage calculation unit 14.

FIG. 11 is a diagram illustrating an example of a failure occurrence load amount calculation formula table stored in the harm load calculation formula database (DB) 18.
The failure load amount calculation formula database 18 stores a formula for calculating the load amount (CPU load amount) generated by the CPU when a failure occurs for each overlay network. These data are registered in advance in the failure load amount calculation formula database 18 by the administrator.
The amount of load required when a failure occurs depends on the number of nodes and links included in the target overlay network. For this reason, it is desirable that the formula for calculating the load generated by the CPU when a failure occurs is a function of the number of nodes and the number of links included in the target overlay network, as shown in FIG. As these formulas, known calculation formulas can be used. For example, in a network that executes OSPF, it is known that the CPU load required when a failure occurs is on the order of (NlogN). Here, N represents the number of nodes included in the network.

For each risk factor in the risk factor use overlay network information, the resource usage calculation unit 14 sequentially reads the name of the overlay network that uses the risk factor. For example, the resource usage calculation unit 14 first reads from the risk factor use overlay network information in FIG. 9 that the overlay network 1 is using the physical node 1 and performs the process described later. Next, the resource usage calculation unit 14 reads that the overlay network 2 is using the physical node 1 and performs a process described later. This is repeated for each overlay network that uses each risk factor.
Using the overlay network name read from the risk factor use overlay network information, the resource usage calculation unit 14 reads the identifier of the CPU that implements the virtual machine included in this overlay network from the virtual machine implementation CPU information. For example, for the overlay network 1 read from the risk factor use overlay network information of FIG. 9, the fact that the virtual machine 111 included in the overlay network 1 is operating on the CPU 1 is read from the virtual machine realization CPU information of FIG. .
Here, it can be determined from the virtual machine name, for example, which overlay network the virtual machine indicated in the virtual machine realization CPU information is included in. In the example of FIG. 10, the second number among the three-digit numbers included in the virtual machine name indicates the overlay network including the virtual machine. For example, the virtual machine 111 is included in the overlay network 1, the virtual machine 121 is included in the overlay network 2, and the virtual machine 141 is included in the overlay network 4.

Further, the resource usage calculation unit 14 uses the overlay network name read from the risk factor use overlay network information to calculate an equation for calculating the CPU load required at the time of the failure, as a failure load calculation equation database. 18 is read out. For example, the overlay network 1 read from the risk factor use overlay network information of FIG. 9 needs a CPU load of 2.3 × n × log (n) when a failure occurs, and the load amount at the time of failure occurrence of FIG. Read from the calculation formula table. The resource usage calculation unit 14 reads the number of nodes included in the overlay network from the risk factor database 33 and substitutes it into the formula read from the fault load calculation formula database 18 to calculate the CPU load required when a fault occurs. calculate.
For each risk factor of the risk factor use overlay network information, the resource usage amount calculation unit 14 calculates the CPU load amount required when a failure occurs for each overlay network that uses the risk factor, and calculates the calculated CPU load amount. The load amount U _{k, 1} of the equation (1) is calculated for each CPU.

Here, there are usually multiple risk factors for the newly constructed overlay network. Resource use amount calculating unit 14, a failure occurrence probability P _i of each risk factor, with the load of each CPU per risk factor calculated by equation (1), the load of the CPU required in the event of a failure The expected value E _{k of the} quantity is calculated based on the formula (2).

  Here, i represents a risk factor included in the newly constructed overlay network. Further, j represents an overlay network that shares a target risk factor among overlay networks that have been constructed using the target physical node as described above.

The failure occurrence probability database (DB) 17 stores risk factors and failure occurrence probabilities in association with each other. These data are registered in advance in the failure occurrence probability database 17 by the administrator.
FIG. 12 is a diagram illustrating an example of a failure occurrence probability table stored in the failure occurrence probability database. In the figure, the data in the resource column indicates a component of the physical network as a risk factor, and the data in the failure occurrence probability column indicates the failure occurrence probability in this component. For example, the failure occurrence probability at node 1 is 0.001. As the failure occurrence probability, for example, mean time between failures (MTBF) is used. Further, the failure occurrence probability may be a relative value between resources.
The resource usage calculation unit 14 outputs the calculated expected value of the resource usage of each CPU and the virtual machine name information to the resource usage comparison unit 15.

Returning to FIG. 4, in step S 107, when the resource usage comparison unit 15 receives the expected value of the resource usage of each CPU from the resource usage calculation unit 14, the expected resource usage of each CPU output. The CPU compares the values and selects the CPU having the smallest value as the CPU that should constitute the transfer processing unit of the newly constructed overlay network.
The resource usage comparison unit 15 compares the expected value E _k of the resource usage of each CPU calculated by the resource usage calculation unit 14 and newly transfers the CPU having the smallest value to the overlay network transfer processing unit. Is selected as the CPU to be configured.
The resource usage comparison unit 15 outputs information representing the selected CPU and virtual machine name information to the output unit 16.
When the output unit 16 receives information indicating the selected CPU or the like from the resource usage comparison unit 15, the output unit 16 includes a virtual node associated with the physical node including the overlay network management unit 1 including the output unit 16. The machine name is read from the virtual machine name information, and the read virtual machine name and information indicating the CPU selected by the resource selection unit 12 are output. The output unit 16 outputs, for example, information representing the CPU selected by the resource selection unit 12 to a management screen (not shown). Thereby, the administrator can confirm the selected CPU. Alternatively, the output unit 16 outputs information indicating the selection and the virtual machine name to the virtualization software operating on the selected CPU, and the received virtualization software assigns the new virtual machine to the virtual software. You may make it build with a machine name.

  The operation of the overlay network management apparatus 1 when receiving an overlay network construction request has been described above. On the other hand, when there is a request for deleting the overlay network, the risk factor extracting unit 11 deletes the data of the overlay network to be deleted from the risk factor database 20. As a result, when the overlay network construction request is received, the overlay network management unit 1 uses the CPU load amount for recalculating the route of the already deleted overlay network as the CPU that has realized this overlay network node. It is possible to calculate the CPU load amount more accurately.

  Next, the effect of this embodiment will be described. In the present embodiment, the resource conflict search unit 13 uses a function (node, port, etc.) corresponding to the risk factor extracted by the risk factor extraction unit 11 in the overlay network that has already been constructed using the target physical node. The resource usage calculation unit 14 calculates an expected value of how much resource each CPU uses when a failure occurs, and the resource usage comparison unit 15 calculates the resource usage. The expected value of the resource usage of each CPU calculated by the calculation unit 14 is compared, and the CPU having the smallest value is selected as the CPU that should constitute the transfer processing unit of the request overlay network. Thereby, the processing load generated when a failure occurs can be distributed to a plurality of CPUs in the physical node.

<Second Embodiment>
FIG. 13 is a configuration diagram showing a schematic configuration of the overlay network management unit 2 according to the second embodiment of the present invention. In the figure, the overlay network management unit 2 includes an input unit 30, a transmission unit 34, an underlay link search unit 31, a risk factor database 33, a resource selection unit 12, an output unit 16, a failure occurrence probability database 17, and a load amount at the time of failure. The calculation formula database 18, the virtual machine state monitor unit 19, and the communication monitor unit 32 are included. The resource selection unit 12 includes a resource competition search unit 13, a resource usage calculation unit 14, and a resource usage comparison unit 15.
In the figure, the same reference numerals (12, 13, 14, 15, 16, 17, 18, 19) are assigned to portions corresponding to the respective portions in FIG.
The overlay network management unit 2 is different from the overlay network management unit 1 of the first embodiment in that the communication monitor unit 32 writes risk factors in the risk factor database 33.
Note that the resource contention search unit 13 replaces the risk factor extracted by the risk factor extraction unit 11 of the first embodiment with the physical node information included in the network construction request and the physical information output by the underlay link search unit 31. Similar to the first embodiment, an overlay network using these physical nodes and physical links is searched using the link information.

Similar to the input unit 10 in FIG. 3, the input unit 30 generates virtual machine name information when receiving an overlay network construction request. The input unit 30 outputs the received overlay network construction request and the generated virtual machine name information to the underlay link search unit 31 and the transmission unit 34. Further, the input unit 30 outputs the virtual machine name information to the communication monitor unit 32.
Similar to the underlay link search unit 21 in FIG. 3, the underlay link search unit 31 determines the physical link route and generates physical link route information for the virtual link included in the overlay network construction request. The underlay link search unit 31 outputs the generated physical link path information to the resource conflict search unit 13 and the communication monitor unit 32.
The risk factor database 33 stores risk factors written by the overlay network.

The communication monitor unit 32 stores the physical node included in the overlay network construction request received from the input unit 30 and the corresponding virtual node identifier in an internal node correspondence storage unit (not shown). In addition, the communication monitor unit 32 associates the physical link path and the information specifying the virtual link output from the underlay link search unit 31 with each other and stores them in an internal link correspondence storage unit (not shown).
Further, the communication monitor unit 32 monitors communication data of each virtual machine executed by the virtualization software. A transfer processing unit belonging to a certain overlay network communicates only with a node in the overlay network to which the transfer processing unit belongs. Therefore, the destination virtual node and the source virtual node of data transmitted and received by the transfer processing unit belong to the same overlay network. The communication monitor unit 32 reads out the destination virtual node and the source virtual node from the communication data. The communication monitoring unit 32 specifies the physical node in which the read destination virtual node is constructed and the physical node in which the transmission source virtual node is constructed using the node correspondence storage unit. Further, the communication monitor unit 32 writes the identified physical node in the risk factor database 33.
In addition to the overlay network construction request and virtual machine name information received from the input unit 30, the transmission unit 34 also communicates information regarding communication between virtual nodes received from the communication monitor unit between the overlay network management units 2.

Next, the operation of the overlay network management unit 2 will be described.
The operation performed when the overlay network management unit 2 receives an overlay network construction request is the same as the flowchart of FIG. However, the risk factor extraction in step S104 is performed by the communication monitor unit 32 writing the risk factor in the risk factor database 42.
Similar to the input unit 30 in FIG. 3, the input unit 30 generates virtual machine name information upon receiving an overlay network construction request. The input unit 30 outputs the received overlay network construction request and the generated virtual machine name information to the underlay link search unit 31 and the transmission unit 34. Further, the input unit 30 outputs the virtual machine name information to the communication monitor unit 32. Similar to the transmission unit 24 in FIG. 3, the transmission unit 34 transmits and receives an overlay network construction request and virtual machine name information between the overlay network management units 2. The transmission unit 34 that has received the overlay network construction request and the virtual machine name information outputs the received overlay network construction request and virtual machine name information to the underlay link search unit 31. In addition, the transmission unit 34 outputs the received virtual machine name information to the communication monitor unit 32.
As in the case of the first embodiment, an input device may be provided separately from the overlay network management unit 2, and this input device may receive an overlay network construction request.
Similar to the underlay link search unit 21 in FIG. 3, the underlay link search unit 31 determines a physical link route for a virtual link included in a newly constructed overlay network, and generates physical link route information. The underlay link search unit 31 outputs the virtual machine name information and the physical link path information to the resource conflict search unit 13. The underlay link search unit 31 also outputs physical link path information to the communication monitor unit 32. As in the case of the first embodiment, each overlay network management unit 2 uses the same physical link route determination method. Alternatively, as in the case of the first embodiment, one overlay network management unit 2 may generate physical link path information and transmit it to each overlay network management unit 2.

The risk factor database 33 stores risk factors written by the communication monitor unit 32.
FIG. 14 is a diagram illustrating an example of data stored in the risk factor database 33. As shown in FIG. 8, the risk factor database 20 of the first embodiment stores the overlay network name and the risk factor information, whereas the risk factor database 33 replaces the overlay network name with the virtual node identifier. (Specifically, virtual machine name) is stored.

The communication monitoring unit 32 outputs information indicating the name of the physical node included in the overlay network construction request and the identifier of the corresponding virtual node, which is output from the input unit 30 or the transmission unit 34 as virtual machine name information. It is written as a node correspondence table in a relationship storage unit (not shown).
FIG. 15 is a diagram illustrating an example of information indicating a name of a physical node included in an overlay network construction request and an identifier of a corresponding virtual node stored in the node correspondence storage unit. As shown in the figure, the node correspondence storage unit accumulates and stores virtual machine name information output from the input unit 30 or the transmission unit 34.

Further, the communication monitor unit 32 rewrites the physical link name in the virtual link column with the virtual machine name in the information that the physical link route information and the virtual link that the underlay link search unit 31 outputs as physical link route information. Then, it is written as a link correspondence table in an internal link correspondence storage unit (not shown).
FIG. 16 is a diagram illustrating an example of information associated with a physical link route and a virtual link stored in the link correspondence storage unit. The communication monitor unit 32 rewrites the physical node name in the virtual link column in the physical link route information received from the underlay link search unit 31 using the virtual machine name information. As shown in the figure, the link correspondence storage unit accumulates and stores physical link path information in which the communication monitor unit 32 rewrites the virtual link field.

  Further, the communication monitor unit 32 monitors communication data of each virtual machine executed by the virtualization software. A transfer processing unit belonging to a certain overlay network communicates only with a node in the overlay network to which the transfer processing unit belongs. Therefore, the destination virtual node and the source virtual node of data transmitted and received by the transfer processing unit belong to the same overlay network. The communication monitor unit 32 reads out the destination virtual node and the transmission source virtual node from the communication data. The communication monitoring unit 32 identifies the physical node in which the read destination virtual node is constructed and the physical node in which the transmission source virtual node is constructed using information stored in the node correspondence storage unit. . Further, the communication monitor unit 32 writes the identified physical node in the risk factor database 33.

At this time, the communication monitoring unit 32 determines whether or not the source virtual node and the destination virtual node are already stored in the virtual node identifier column of the risk factor database 33, and according to the determination result. The following processing is performed.
1. When it is determined that none of the virtual nodes is stored: a new row is added to the risk factor database 33, and the identifier of the source virtual node and the destination virtual node are added to the virtual node identifier column of the added row. The identifier is written, and the node name of the physical node in which the virtual node is constructed and the node name of the physical node in which the transmission source virtual node is constructed are written in the risk factor information column.
FIG. 17 is a diagram illustrating an example of the risk factor database 33 before and after the communication monitor unit 32 adds a row.
For example, it is assumed that the source virtual node is a virtual node represented by “virtual machine 111” and the destination virtual node is a virtual node represented by “virtual machine 112”. FIG. 6A shows the risk factor database 33 before the communication monitor unit 32 adds a row, and neither the source virtual node nor the destination virtual node is stored in the risk factor database.
In this case, the communication monitor unit 32 adds a row to the risk factor database 33 as shown in the row L171 of FIG. 7B, and the identifier “virtual machine 111” of the transmission source virtual node is displayed in the virtual node identifier column. And the identifier “virtual machine 112” of the destination virtual node is written. Further, the communication monitor unit 32 stores “physical node 1” corresponding to “virtual machine 111” and “virtual machine 112” corresponding to “virtual machine 112” from the node correspondence relationship table of FIG. "Physical node 2" is read out and written in the risk factor information column as shown in row L171.

2. When only one of the virtual nodes is stored: The virtual node that is not stored in the virtual node identifier column of the row in which this virtual node is stored (for example, the virtual node of the transmission source is virtual) If it is stored in the node identifier column, the identifier of the destination virtual node is added, and the physical node in which the non-stored virtual node is constructed in the risk factor information column (example above) Then, the name of the physical node on which the destination virtual node is constructed) is added to the risk factor information column.
FIG. 18 is a diagram illustrating an example of the risk factor database 33 before and after the communication monitor unit 32 writes the name of the virtual node and the name of the physical node.
For example, it is assumed that the source virtual node is a virtual node represented by “virtual machine 111” and the destination virtual node is a virtual node represented by “virtual machine 113”. FIG. 9A shows the risk factor database 33 before the communication monitor unit 32 writes the virtual node name and the physical node name, and the transmission source “virtual machine 111” is stored in the row L181. However, the destination “virtual machine 113” is not stored in the risk factor database.
In this case, the communication monitor unit 32 sets the destination in the virtual node identifier column of the row in which “virtual machine 111” is stored (row L181 in FIG. 11A), as in row L182 in FIG. The virtual machine identifier “virtual machine 113” is written. Further, the communication monitor unit 32 reads “physical node 3” corresponding to “virtual machine 113” from the node correspondence table in FIG. 15 stored in the node correspondence storage unit, and performs risk factors as shown in row L182. Write in the information column.

3. If both virtual nodes are stored, but stored as different network risk factors: consolidate their respective rows. For example, all the data in the row in which the transmission source virtual node is stored is added to the row in which the destination virtual node is stored, and then the row in which the transmission source virtual node is stored is deleted.
FIG. 19 is a diagram illustrating an example of the risk factor database 33 before and after the communication monitor unit 32 integrates rows.
For example, it is assumed that the source virtual node is a virtual node represented by “virtual machine 111” and the destination virtual node is a virtual node represented by “virtual machine 113”. FIG. 9A shows the risk factor database 33 before the communication monitor unit 32 merges the rows. The transmission source “virtual machine 111” is stored in the row L191, and the destination “virtual machine 113” is Stored in row L192.
In this case, the communication monitor unit 32, like the line L193 in FIG. 9B, among the virtual machine names in the line where the “virtual machine 113” is stored (the line L192 in FIG. What is not stored in the row in which the “virtual machine 111” is stored (row L191 in FIG. 11A) is changed to the row in which the “virtual machine 111” is stored (row L191 in FIG. 15A). Write in the virtual node identifier field. Similarly, the communication monitor unit 32 stores the row of “virtual machine 111” among the names of risk factors in the row (row L192 of FIG. 11A) in which “virtual machine 113” is stored ( What is not stored in the row L191) of FIG. 11A is written in the risk factor information column of the row where the “virtual machine 111” is stored (row L191 of FIG. 11A).
Thereafter, the communication monitor unit 32 deletes the row in which “virtual machine 113” is stored (row L192 in FIG. 11A). As a result, the risk factor database 33 shown in FIG.
4). When both virtual nodes are stored as risk factors of the same network: Since they are already stored as virtual nodes belonging to the same network, nothing is done.

Further, the communication monitoring unit 32 reads out the path of the physical link that realizes the virtual link between the transmission source virtual node and the destination virtual node from the link correspondence storage unit. And the above 1. ~ 3. In this case, the read physical link is added to the risk factor information column in the row where the data is added in the risk factor database 33.
FIG. 20 is a diagram illustrating an example of the risk factor database 33 before and after the communication monitor unit 32 integrates rows.
For example, the transmission source virtual node is a virtual node represented by “virtual machine 111”, the destination virtual node is a virtual node represented by “virtual machine 112”, Suppose that a row L201 in FIG.
In this case, the communication monitor unit 32 uses the physical link 1 as a physical link path that realizes a virtual link between the virtual node represented by “virtual machine 111” and the virtual node represented by “virtual machine 112”. The configuration is read from the link correspondence table in FIG. As shown in the row L202 in FIG. The read physical link name “physical link 1” is added to the risk factor information column of the row added in (the row L201 in FIG. 11A).
On the other hand, the above 4. If you do nothing.

Here, the communication monitor unit 32 monitors only data transmitted and received by the virtual machine executed on the managed physical node, and the virtual machine executed on the other physical node further executes on the other physical node. Data that is sent to and received from the virtual machine is not monitored.
Accordingly, the communication monitor unit 32, for example, a virtual link between a virtual node realized on a physical node other than the managed physical node and a virtual node realized on another physical node other than the managed physical node. The physical link path for realizing the above cannot be acquired from the communication data monitored by the communication monitor unit 32.
On the other hand, when a failure occurs in the network as described above, route recalculation is performed in all nodes. Therefore, the resource usage calculation unit 14 is configured so that all of the overlay networks constructed using the managed physical nodes CPU load needs to be calculated for risk factors.
Therefore, the communication monitoring unit 32 reads the destination virtual node identifier and the source virtual node identifier read from the virtual machine communication data obtained by monitoring (hereinafter also referred to as “transmission / reception node identifier”). Is output to the transmission unit 34. The transmission unit 34 transmits / receives the identifier of the transmission / reception node output from the communication monitoring unit 32 between the overlay network management units 2. When the transmission unit 34 receives the identifier of the transmission / reception node from the other overlay network management unit 2, the transmission unit 34 outputs the received identifier of the transmission / reception node to the communication monitoring unit 32. The communication monitor unit 32 updates the risk factor database 33 using the identifier of the transmission / reception node output from the transmission unit 34, similarly to the identifier of the transmission / reception node read from the communication data of the virtual machine.
As a result, the communication monitoring unit 32 creates a virtual link between a virtual node realized on a physical node other than the managed physical node and a virtual node realized on a physical node other than the managed physical node. It is possible to extract risk factors of the entire virtual network including the physical link path to be realized.

The effect in 2nd Embodiment is demonstrated. In the second embodiment, similar to the risk factor extraction unit 11 in the first embodiment, the communication monitor unit 32 extracts the risk factor. Therefore, as in the first embodiment, the processing that occurs at the time of failure The load can be distributed to a plurality of CPUs in the physical node.
As in the first embodiment, the overlay network management unit 2 of the second embodiment may be realized by dedicated one CPU in the physical node or may be realized on a virtual machine. Good. Or you may implement | achieve as an overlay network management apparatus independent of a physical node.

<Third Embodiment>
FIG. 21 is a diagram illustrating a configuration example of a network in the third embodiment of the present invention. In the figure, a physical network 200 includes physical nodes 201 to 205, links between these physical nodes, and a network control unit (network control device) 210. The network control unit 210 centrally manages a plurality of physical nodes 201 to 205. An overlay network is constructed on the physical network 200.

  FIG. 22 is a configuration diagram illustrating a schematic configuration of the physical node 201. In the figure, the physical node 201 includes a switch. Transfer processing functions such as route calculation and transfer table creation are executed by the network control unit 210. The same applies to other physical nodes.

FIG. 23 is a configuration diagram illustrating a schematic configuration of the network control unit 210. In the figure, the network control unit 210 includes CPUs 261-1 to 261- (N + 1).
As with the CPUs 161-1 to 161-N of the first embodiment, each of the CPUs 261-1 to 261-N executes virtualization software and constructs a virtual machine. Each virtual machine includes one transfer processing unit. One transfer processing unit is configured for one overlay network. The transfer processing unit executes route calculation and transfer table creation for one entire overlay network. The CPUs 261-1 to 261-N may cause a plurality of transfer processing units to operate on one OS. However, in order to ensure security and resources for each transfer processing unit, as shown in FIG. It is desirable to construct a virtual machine and execute the transfer processing unit in each virtual machine.
Since the network control unit 210 manages a plurality of physical nodes, these transfer processing units issue a switch control request to the plurality of physical nodes.
The CPU 261-(N + 1) implements the overlay network management unit 3 by executing software.

Note that the overlay network management unit 3 may be realized not on an independent CPU but on a virtual machine.
FIG. 24 is a configuration diagram showing another schematic configuration of the network control unit 210 in the embodiment. In the figure, the network control unit 210 is configured to include CPUs 262-1 to 262-N.
As with the CPUs 261-2 to 161-N in FIG. 23, the CPUs 262-1 to 262-N execute virtualization software and construct a virtual machine. Further, the CPU 262-N implements the overlay network management unit 3 by executing software on one virtual machine.
The overlay network management unit 3 may be realized as an overlay network management device independent of the network control unit 210. In this case, the overlay network management apparatus is connected to each physical node directly or via the network control unit 210, and monitors the switch of each physical node. The overlay network management apparatus is connected to the network control unit 210 and monitors the virtualization software of each CPU.

FIG. 25 is a configuration diagram illustrating a schematic configuration of the overlay network management unit 3. In the figure, the overlay network management unit 3 includes an input unit 10, an underlay link search unit 21, a risk factor database 42, a resource selection unit 12, an output unit 16, a failure occurrence probability database 17, and a failure load amount calculation formula database 18. And a virtual machine state monitor unit 19 and a switch control monitor unit 41. The resource selection unit 12 includes a resource competition search unit 13, a resource usage calculation unit 14, and a resource usage comparison unit 15.
In the figure, the same reference numerals (10, 21, 12, 13, 14, 15, 16, 17, 18, 19) are assigned to portions corresponding to the respective portions in FIG. These are different from the respective units in FIG. 3 in that processing for selecting CPUs in a plurality of physical nodes, but are otherwise the same as the respective units in FIG.
As in the second embodiment, the resource conflict search unit 13 searches for physical nodes and underlay links included in the network construction request instead of the risk factor extracted by the risk factor extraction unit 11 of the first embodiment. Similar to the first embodiment, the physical link output by the unit 31 is searched for an overlay network using these physical nodes and physical links.
The overlay network management unit 3 is different from the overlay network management unit 1 of the first embodiment in that the switch control monitor unit 41 writes a risk factor in the risk factor database 33.

  The switch control monitor unit 41 monitors information related to switch control (hereinafter also referred to as “switch control information”) from the switch or the transfer processing unit. The switch control monitor unit 41 detects a physical link path corresponding to the virtual link by monitoring data transmitted on the physical network.

Next, the operation of the overlay network management unit 3 will be described.
The operation performed when the overlay network management unit 3 receives an overlay network construction request is the same as the flowchart of FIG. However, the risk factor extraction in step S104 is performed by the switch control monitor unit 41 writing the risk factor in the risk factor database 42.
The switch control monitor unit 41 monitors switch control information from the switch or the transfer processing unit. A transfer processing unit belonging to a certain overlay network exchanges switch control information only with a switch (physical node) used by the overlay network to which the transfer processing unit belongs. Therefore, the physical node that exchanges switch control information with the transfer processing unit is a physical node that is used by the overlay network corresponding to the transfer processing unit.

FIG. 26 is a diagram illustrating an example of switch control information that the switch control monitor unit 41 acquires from a switch (physical node).
In the figure, the physical node name indicates the name of the physical node from which information is acquired. The information exchange destination transfer processing unit name indicates the name of the transfer processing unit with which the physical node indicated by the physical node name exchanges switch control information. In the example shown in the figure, the physical node 201 exchanges switch control information with the transfer processing unit 283-1-1, the transfer processing unit 283-1-3, and the like.
The switch control monitor unit 41 acquires these pieces of information from each physical node. Then, the switch control monitor unit 41 groups the physical node names that exchange information with the same transfer processing unit, so that the transfer processing unit determines the physical node used by the overlay network that performs path calculation and transfer table creation. To detect.
For example, each of the physical nodes 201, 202, 203, and 205 exchanges switch control information with the transfer processing unit 283-1-1, and other physical nodes exchange switch control information with the transfer processing unit 283-1-1. If not, the switch control monitor unit 41 detects physical nodes 201, 202, 203, and 205 as physical nodes used by the overlay network on which the transfer processing unit 283-1-1 executes path calculation and transfer table creation.
The switch control monitor unit 41 performs grouping for each transfer processing unit, detects a physical node used by each overlay network, and writes it in the risk factor database 42.

FIG. 27 is a diagram illustrating an example of switch control information that the switch control monitor unit 41 acquires from the transfer processing unit.
In the figure, the transfer processing unit name indicates the name of the transfer processing unit from which the switch control monitor unit 41 obtains information. The information exchange destination physical node name indicates the name of the physical node with which the transfer processing unit indicated by the transfer processing unit name exchanges information related to physical node switch control. These physical nodes are physical nodes used by the overlay network in which the transfer processing unit executes path calculation and transfer table creation. In the example of the figure, the transfer processing unit 283-1-1 exchanges switch control information with the physical node 201, the physical node 202, and the like. The physical node 201, the physical node 202, and the like are physical nodes used by the overlay network in which the transfer processing unit 283-1-1 executes path calculation and transfer table creation.
The switch control monitor unit 41 obtains information on physical nodes for exchanging information from each transfer processing unit, thereby detecting the physical nodes used by each overlay network as in the case of FIG. Write to.

The switch control monitor unit 41 detects a physical link path that realizes a virtual link by monitoring data transmitted on the physical network.
The switch control monitor unit 41 reads the source virtual node and the destination virtual node from the data passing on each physical node. As a result, the switch control monitor unit 41 detects which physical node passes the data transmitted between the virtual nodes in which order, and acquires the path of the physical link that realizes the virtual link.
For example, when data whose destination is the virtual node 235 and the transmission source is the virtual node 231 is sequentially detected in the physical nodes 201, 204, and 205, the switch control monitor unit 41 moves from the virtual node 231 to the virtual node 235. The physical link between the physical nodes 201 and 204 and the physical link between the physical nodes 204 and 205 are acquired as physical link paths for realizing the virtual link.
The switch control monitor unit 41 writes the acquired physical link in the risk factor database 42 as a risk factor of the overlay network to which the virtual node belongs.
As described above, the switch control monitor unit 41 extracts risk factors of each overlay network and writes them in the risk factor database 42. Hereinafter, the resource selection unit 12 performs CPU selection processing similar to that of the resource selection unit 12 of the first embodiment for each physical node included in the network construction request, and the output unit 16 outputs the output of the first embodiment. Similarly to the unit 16, information indicating the CPU selected by the resource selection unit 12 and the virtual machine name are output.

The effect in the third embodiment will be described. In the second embodiment, a network environment is assumed in which the transfer control function for the overlay network operates in a distributed manner on each physical node. However, in the third embodiment, the transfer control function of a plurality of physical nodes is provided. It is possible to cope with a network environment that executes centrally.
As in the first embodiment, the overlay network management unit 3 includes a risk factor extraction unit instead of the switch control monitor unit 41, and the risk factor extraction unit extracts the risk factor and writes it in the risk factor database 42. You may do it. Also in this case, as with the case where the switch control monitor unit 41 is provided, the CPU can be selected for all the physical nodes with one overlay network management unit. Similarly, as in the second embodiment, the overlay network management unit 3 includes a communication monitor unit instead of the switch control monitor unit 41, and the communication monitor unit detects a risk factor and writes it in the risk factor database 42. You may do it.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design changes and the like within a scope not departing from the gist of the present invention.

  The present invention is suitable for use in an overlay network management device, a physical node device, a network control device, an overlay network management method, and an overlay network management program.

1, 2, 3 Overlay network management unit 10, 30 Input unit 11 Risk factor extraction unit 12 Resource selection unit 13 Resource competition search unit 14 Resource usage calculation unit 15 Resource usage comparison unit 16 Output unit 17 Occurrence probability database 18 Failure Time load calculation formula database 19 Virtual machine state monitoring unit 20, 33, 42 Risk factor database 21, 31 Underlay link search unit 24, 34 Transmission unit 32 Communication monitoring unit 41 Switch control monitoring unit 100, 200 Physical networks 101-105 201-205 Physical nodes 110, 120, 130 Overlay networks 111-114, 121-124, 131-133 Virtual nodes 161-1 to 161- (N + 1), 162-1 to 162-N, 261-1 to 261- (N + 1), 2 2-1~262-N CPU
171, 271 switches 181-1 to 181-N, 281-1 to 281-N virtualization software 182-1-1 to 182-N- (M _N +1), 282-1-1 to 282-N- (M _N + 1) virtual machines 183-1-1 to 183 -N-M _N , 283-1-1 to 283 -N-M _N transfer processing unit 210 network control unit

Claims (20)

An overlay network management apparatus for managing an overlay network constructed by using any physical node in a physical network including a physical node having a plurality of CPUs or a component of a physical link connecting the physical nodes,
When a request to construct a new overlay network is input, the amount of resources used by each of the plurality of CPUs when a failure occurs in a component of the physical network used by the newly constructed overlay network A resource usage calculator that calculates an expected value;
A resource usage comparison unit that selects a CPU having a minimum expected value of the resource amount calculated by the resource usage calculation unit as a CPU that should constitute a transfer processing unit of a newly constructed overlay network; and
An overlay network management apparatus comprising: A failure occurrence probability database that associates and stores a failure occurrence probability and a component that may cause a failure on the physical network;
A failure load amount calculation formula database for storing a formula for calculating the resource amount used by the CPU for each overlay network when a failure occurs;
A virtual machine state monitor unit that identifies which CPU is used to construct each overlay network;
An underlay link search unit for determining a physical link path for realizing a virtual link of a newly constructed overlay network;
A risk factor database for storing, for each overlay network, the components of the physical network used by the already constructed overlay network;
Further comprising
The resource usage calculation unit
Among the overlay networks that are identified that the virtual machine state monitor unit is constructed using the CPU for each CPU and each physical network component used by the newly constructed overlay network A resource used by the CPU for the overlay network when a failure occurs in the component of the detected overlay network based on the data of the risk factor database. The amount is calculated based on an expression stored in the failure load amount calculation expression database, and the plurality of the plurality of the number is calculated based on the calculated resource amount and the failure occurrence probability in each component read from the failure occurrence probability database. Used by each CPU Calculating the expected value of the resource amount,
The overlay network management apparatus according to claim 1. Based on the physical link path determined by the underlay link search unit, further comprises a risk factor extraction unit that extracts components of the physical network used by the newly constructed overlay network,
The risk factor database stores, for each overlay network, the physical network components used by the overlay network extracted by the risk factor extraction unit as physical network components used by the already constructed overlay network.
The overlay network management device according to claim 2. Information indicating a communication destination or transmission source is obtained by monitoring communication data of each virtual machine that realizes a virtual node of an already constructed overlay network on the CPU, and obtaining information indicating the communication destination or transmission source. And a communication monitoring unit that acquires information indicating physical nodes and physical links that realize each overlay network based on the physical link path information determined by the underlay link search unit,
The risk factor database stores, for each overlay network, information indicating physical nodes and physical links that realize each overlay network acquired by the communication monitor unit.
The overlay network management device according to claim 2. From each physical node or each virtual node, obtain information on the physical node that realizes each virtual node, and monitor transmission data that passes through each physical node to obtain information on the physical link that realizes each virtual link Further comprising a switch control monitor unit,
The risk factor database stores, for each overlay network, information on physical nodes that realize each virtual node acquired by the switch control monitor unit and information on physical links that realize virtual links.
The overlay network management device according to claim 2. A physical node device having a plurality of CPUs and constructing a virtual node of an overlay network,
When a request to construct a new overlay network is input, a failure occurs in any physical node in the physical network used by the newly constructed overlay network or a component of a physical link that connects these physical nodes. A resource usage calculation unit that calculates an expected value of the resource amount used by each of the plurality of CPUs when
A resource usage comparison unit that selects a CPU having a minimum expected value of the resource amount calculated by the resource usage calculation unit as a CPU that should constitute a transfer processing unit of a newly constructed overlay network; and
A physical node device comprising: A failure occurrence probability database that associates and stores a failure occurrence probability and a component that may cause a failure on the physical network;
A failure load amount calculation formula database for storing a formula for calculating the resource amount used by the CPU for each overlay network when a failure occurs;
A virtual machine state monitor unit that identifies which CPU is used to construct each overlay network;
An underlay link search unit for determining a physical link path for realizing a virtual link of a newly constructed overlay network;
A risk factor database for storing, for each overlay network, the components of the physical network used by the already constructed overlay network;
Further comprising
The resource usage calculation unit
Among the overlay networks that are identified that the virtual machine state monitor unit is constructed using the CPU for each CPU and each physical network component used by the newly constructed overlay network A resource used by the CPU for the overlay network when a failure occurs in the component of the detected overlay network based on the data of the risk factor database. The amount is calculated based on an expression stored in the failure load amount calculation expression database, and the plurality of the plurality of the number is calculated based on the calculated resource amount and the failure occurrence probability in each component read from the failure occurrence probability database. Used by each CPU Calculating the expected value of the resource amount,
The physical node device according to claim 6, wherein: Based on the physical link path determined by the underlay link search unit, further comprises a risk factor extraction unit that extracts components of the physical network used by the newly constructed overlay network,
The risk factor database stores, for each overlay network, the physical network components used by the overlay network extracted by the risk factor extraction unit as physical network components used by the already constructed overlay network.
The physical node device according to claim 7, wherein: Information indicating a communication destination or transmission source is obtained by monitoring communication data of each virtual machine that realizes a virtual node of an already constructed overlay network on the CPU, and obtaining information indicating the communication destination or transmission source. And a communication monitoring unit that acquires information indicating a physical node device and a physical link that realize each overlay network based on the physical link path information determined by the underlay link search unit,
The risk factor database stores, for each overlay network, information indicating physical node devices and physical links that realize each overlay network acquired by the communication monitor unit.
The physical node device according to claim 7, wherein: A network controller for controlling an overlay network constructed using a physical network component including a physical node having a plurality of CPUs,
When a request to construct a new overlay network is input, the amount of resources used by each of the plurality of CPUs when a failure occurs in a component of the physical network used by the newly constructed overlay network A resource usage calculator that calculates an expected value;
A resource usage comparison unit that selects a CPU having a minimum expected value of the resource amount calculated by the resource usage calculation unit as a CPU that should constitute a transfer processing unit of a newly constructed overlay network; and
A failure occurrence probability database that associates and stores a failure occurrence probability and a component that may cause a failure on the physical network;
A failure load amount calculation formula database for storing a formula for calculating the resource amount used by the CPU for each overlay network when a failure occurs;
A virtual machine state monitor unit that identifies which CPU is used to construct each overlay network;
An underlay link search unit for determining a physical link path for realizing a virtual link of a newly constructed overlay network;
A risk factor database for storing, for each overlay network, the components of the physical network used by the already constructed overlay network;
From each physical node or each virtual node, obtain information on the physical node that realizes each virtual node, and monitor transmission data that passes through each physical node to obtain information on the physical link that realizes each virtual link A switch control monitor unit to
Comprising
The resource usage calculation unit
Among the overlay networks that are identified by the CPU to be constructed by the virtual machine state monitor unit for each CPU and each physical network component used by the newly constructed overlay network A resource used by the CPU for the overlay network when a failure occurs in the component of the detected overlay network based on the data of the risk factor database. The amount is calculated based on an equation read from the failure load amount calculation equation database, and based on the calculated resource amount and the failure occurrence probability in each component read from the failure occurrence probability database, the plurality of CPUs The re-use that each uses To calculate the expected value of over scan amount,
The risk factor database stores, for each overlay network, information on physical nodes that realize each virtual node acquired by the switch control monitor unit and information on physical links that realize virtual links.
A network control device. Overlay network management method for overlay network management apparatus for managing an overlay network constructed by using any physical node in a physical network including a physical node having a plurality of CPUs or components of a physical link connecting the physical nodes Because
When a request for constructing a new overlay network is input to the overlay network management apparatus, if a failure occurs in a component of the physical network used by the newly constructed overlay network, the plurality of CPUs A resource usage calculation step for calculating an expected value of the resource amount used by each;
Resource usage amount that the overlay network management device selects a CPU having the minimum expected value of the resource amount calculated in the resource usage amount calculating step as a CPU that should constitute a transfer processing unit of a newly constructed overlay network A comparison step;
An overlay network management method comprising: The overlay network management device includes:
A failure occurrence probability database that associates and stores a failure occurrence probability and a component that may cause a failure on the physical network;
A failure load amount calculation formula database for storing a formula for calculating the resource amount used by the CPU for each overlay network when a failure occurs;
A risk factor database for storing, for each overlay network, the components of the physical network used by the already constructed overlay network;
Comprising
The overlay network management method includes:
A virtual machine state monitoring step in which the overlay network management device identifies which CPU is used to construct each overlay network;
An underlay link search step in which the overlay network management device determines a physical link path for realizing a virtual link of a newly constructed overlay network;
Further comprising
The resource usage calculation step includes:
Among the overlay networks identified as being constructed using the CPU in the virtual machine state monitoring step for each component of the physical network used by the newly constructed overlay network and for each CPU, the configuration An overlay network constructed using elements is detected based on the data of the risk factor database, and the detected network is used by the CPU for the overlay network when a failure occurs in the component. The resource amount is calculated based on an equation stored in the failure load amount calculation formula database, and based on the calculated resource amount and the failure occurrence probability in each component read from the failure occurrence probability database, Each of multiple CPUs Calculating the expected value of the amount of resources used,
The overlay network management method according to claim 11, wherein: The overlay network management device further comprises a risk factor extraction step of extracting a constituent element of the physical network used by the newly constructed overlay network based on the physical link path determined in the underlay link search step,
The risk factor database stores, for each overlay network, the physical network components used by the overlay network extracted in the risk factor extraction step as physical network components used by the already constructed overlay network.
The overlay network management method according to claim 12, wherein: The overlay network management apparatus acquires information indicating a communication destination or transmission source by monitoring communication data of each virtual machine that realizes a virtual node of an already constructed overlay network on the CPU. A communication monitoring step of acquiring information indicating a physical node and a physical link realizing each overlay network based on information indicating a destination or a transmission source and information on a physical link path determined in the underlay link search step; Equipped,
The risk factor database stores, for each overlay network, information indicating physical nodes and physical links that realize each overlay network acquired in the communication monitoring step.
The overlay network management method according to claim 12, wherein: The overlay network management device acquires information on physical nodes that realize each virtual node from each physical node or each virtual node, and monitors transmission data passing through each physical node to realize each virtual link. A switch control monitoring step for obtaining information on the physical link to be performed,
The risk factor database stores, for each overlay network, information on a physical node that realizes each virtual node and information on a physical link that realizes a virtual link acquired in the switch control monitoring step.
The overlay network management method according to claim 12, wherein: An overlay network management program for managing an overlay network constructed by using any physical node in a physical network including physical nodes having a plurality of CPUs or components of a physical link connecting the physical nodes,
On the computer,
When a request to construct a new overlay network is input, the amount of resources used by each of the plurality of CPUs when a failure occurs in a component of the physical network used by the newly constructed overlay network A resource usage calculation step for calculating an expected value;
A resource usage comparison step of selecting a CPU having the smallest expected value of the resource amount calculated in the resource usage calculation step as a CPU to constitute a transfer processing unit of a newly constructed overlay network;
A program for running The computer
A failure occurrence probability database that associates and stores a failure occurrence probability and a component that may cause a failure on the physical network;
A failure load amount calculation formula database for storing a formula for calculating the resource amount used by the CPU for each overlay network when a failure occurs;
A risk factor database for storing, for each overlay network, the components of the physical network used by the already constructed overlay network;
Comprising
On that computer,
A virtual machine state monitoring step for identifying which CPU is used to construct each overlay network;
An underlay link search step for determining a physical link path for realizing a virtual link of a newly constructed overlay network;
Is executed further,
In the resource usage calculation step,
Among the overlay networks identified as being constructed using the CPU in the virtual machine state monitoring step for each component of the physical network used by the newly constructed overlay network and for each CPU, the configuration An overlay network constructed using elements is detected based on the data of the risk factor database, and the detected network is used by the CPU for the overlay network when a failure occurs in the component. The resource amount is calculated based on an equation stored in the failure load amount calculation formula database, and based on the calculated resource amount and the failure occurrence probability in each component read from the failure occurrence probability database, Each of multiple CPUs For calculated the expected value of the amount of resources used,
The program according to claim 16. In the computer,
Based on the physical link path determined in the underlay link search step, a risk factor extraction step of extracting a physical network component used by a newly constructed overlay network is further executed.
To store the physical network components used by the overlay network extracted in the risk factor extraction step in the risk factor database for each overlay network as physical network components used by the already constructed overlay network of,
The program according to claim 17. In the computer,
Information indicating a communication destination or transmission source is obtained by monitoring communication data of each virtual machine that realizes a virtual node of an already constructed overlay network on the CPU, and obtaining information indicating the communication destination or transmission source. And a communication monitoring step of acquiring information indicating physical nodes and physical links that realize each overlay network based on the physical link route information determined in the underlay link search step,
For storing information indicating physical nodes and physical links realizing each overlay network acquired in the communication monitoring step in the risk factor database for each overlay network,
The program according to claim 17. In the computer,
From each physical node or each virtual node, obtain information on the physical node that realizes each virtual node, and monitor transmission data that passes through each physical node to obtain information on the physical link that realizes each virtual link Further executing a switch control monitoring step,
In order to store, for each overlay network, information on a physical node that realizes each virtual node and information on a physical link that realizes a virtual link acquired in the switch control monitor step in the risk factor database.
The program according to claim 17.

Images