JP2016144133A - Control device, control program and communication system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently utilize a computation resource in a network according to a guarantee request.SOLUTION: A processor executes the following processing: reception processing for receiving a guarantee request related to the operation guarantee of a load target program, which is a program to be loaded on a plurality of transfer devices; selection processing for selecting, based on management information, a transfer device group configuring a route, from among the plurality of transfer devices; determination processing for determining, according to the guarantee request received by the reception processing, one virtual machine out of a first virtual machine and a second virtual machine configured in the transfer device selected by the selection processing, to be a loading destination of the load target program; and transmission processing for transmitting information related to the load target program to the load destination virtual machine determined by the determination processing. When the transfer device, forming the load destination virtual machine among the plurality of transfer devices, receives information related to the load target program, the transfer device executes the load target program on the load destination virtual machine.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a control device, a control program, and a communication system that control a plurality of transfer devices that transfer data in a network.

  In recent years, NFV (Network Functions Virtualization) technology has attracted attention. NFV is a technology for realizing network functions such as packet transfer and firewall realized on a general-purpose server by software on nodes constituting the network. In NFV, network functions can be dynamically added, deleted, and changed, the network configuration can be easily changed, and physical resources in the network can be effectively used.

  In the NFV environment, it is necessary to deploy a function in each node so that the operation guarantee requirement for each function executed on the node is satisfied and the calculation resources of each node are efficiently used as the entire network. At present, the network maintainer manages the calculation resources of each node and determines the software load target node based on the remaining calculation resources.

  Patent Document 1 describes an example in which a controller that manages a network distributes a load to each node that configures the network. Japanese Patent Laid-Open No. 2004-228561 “provides a virtual network that can effectively use network resources by using a weighting factor for a difference between a physical node resource and a physical node position of a requested virtual node position when creating a virtual node”. be written.

JP 2012-199644 A

  However, the technique described in Patent Document 1 does not guarantee the calculation resource allocation in each node that satisfies the execution requirements such as the guaranteed quality and priority required for each node. Therefore, in the technology described in Patent Document 1, in each NFV environment, each node is configured to satisfy the operation guarantee requirement for each function executed on the node and efficiently use the calculation resources of each node as the entire network. There is a problem that the function cannot be deployed.

  An object of the present invention is to efficiently use computational resources in a network in response to a guarantee request.

  A communication system according to one aspect of the invention disclosed in the present application is a communication system including a plurality of transfer devices that transfer data in a network and a control device that controls the plurality of transfer devices. Each of the transfer devices includes a first virtual machine that executes the program so as to comply with the operation guarantee of the program, and a second virtual machine that does not need to comply with the operation guarantee of the program when the program is executed. And the control device includes a processor and a storage device that stores route management information indicating a route set in the network, and the processor loads the plurality of transfer devices. A reception process for receiving a guarantee request for guaranteeing the operation of the load target program that is a target program to be loaded, and the management information And selecting the transfer device group that constitutes the path from the plurality of transfer devices and the transfer device selected by the selection process according to the guarantee request received by the acceptance process. A determination process for determining one of the first virtual machine and the second virtual machine as a load destination of the load target program, and a load destination virtual machine determined by the determination process A transmission process that transmits information related to the load target program, and the transfer apparatus that constructs the load destination virtual machine among the plurality of transfer apparatuses receives the information related to the load target program, The load target program is executed on a virtual machine.

  According to a typical embodiment of the present invention, it is possible to efficiently use computational resources in a network in response to a guarantee request. Problems, configurations, and effects other than those described above will become apparent from the description of the following embodiments.

It is explanatory drawing which shows the structural example of the communication system concerning a present Example. It is a block diagram which shows the example of an internal structure of a service controller. It is explanatory drawing which shows the example of a memory content of the software function table shown in FIG. It is explanatory drawing which shows the example of a memory content of the calculation resource management table shown in FIG. It is explanatory drawing which shows the example of a memory content of a candidate node table. It is a block diagram which shows the internal structural example of a node. It is explanatory drawing which shows the example of the memory content of a calculation resource table. It is explanatory drawing which shows the example of a sequence of a communication system. FIG. 9 is a flowchart showing a detailed processing procedure example of a load target node / VM selection process (step S806) shown in FIG. 8. FIG.

  In this embodiment, a virtual machine is applied to each node in the network to be controlled to construct a plurality of virtual machines (Virtual Machines, “VMs”), thereby calculating computing resources (for example, CPUs) in each node. (Central Processing Unit)) is divided in VM units. Prior to loading the software to be applied in the network to each node, the service controller that manages each node performs a load-guaranteed request for the software and the calculation resource of each node in the network according to the load-destination request. Determine the nodes and VMs within the nodes.

  As a result, it is possible to secure software calculation resources (for example, CPU usage rate) in accordance with a request for guaranteeing the operation of the software and efficiently use the calculation resources in the network. In addition, since the software is executed on the nodes in units of VMs, calculation resources can be acquired for each VM. Therefore, it is possible to easily specify the calculation resource of the software. Furthermore, by automating these processes, the burden on the network maintainer can be reduced.

<Configuration example of communication system>
FIG. 1 is an explanatory diagram of a configuration example of a communication system according to the present embodiment. First, the connection relationship in the communication system 100 will be described. The communication system 100 includes a node group Ns on the packet transport network 101 to be controlled and a service controller 103. In the communication system 100, for example, the service controller 103 provides a function to add value to the packet transport network 101 such as a traffic monitor and a firewall to the node group Ns as software.

  Node group Ns includes edge nodes EN1 and EN2 and relay nodes RN1 to RN5. Each of the edge nodes EN1 and EN2 is a transfer device arranged at an end point of the packet transport network 101. In FIG. 1, the edge node EN1 is connected to the network 111, and the edge node EN2 is connected to the network 112. In the following description, when any of the edge nodes EN1 and EN2 is acceptable, it is expressed as “edge node EN”.

  Each of the relay nodes RN <b> 1 to RN <b> 5 is a transfer device arranged at a relay point of the packet transport network 101. In the following description, when any of the relay nodes RN1 to RN5 may be used, it is expressed as “relay node RN”. In the following description, when either the edge node EN or the relay node RN may be used, it is expressed as “node N”. The number of nodes N is not limited to the number shown in FIG.

  The node N constructs a virtual machine (VM). The VM is an environment for executing the program P loaded from the service controller 103. Specifically, for example, two types of VMs are constructed. One is a VM that executes a program so as to comply with the operation guarantee of the program. This VM corresponds to an operation guarantee program VM described later. The other is a VM that does not have to comply with the operation guarantee of the program when executing the program. This VM corresponds to a best effort program VM described later. The node N is, for example, a transfer device that transfers data, but may be an L2 switch or a router.

  The maintenance person / other management system 102 refers to a maintenance person who manages the packet transport network 101 or an upper management system connected to the service controller 103. The maintenance person / other management system 102 gives an instruction to the service controller 103.

  The service controller 103 is a control device that connects to each node N and provides various services to the node N in accordance with instructions from the maintainer / other management system 102. For example, when the instruction from the maintenance person / other management system 102 is the route setting request r1, the service controller 103 generates route setting information RTI in accordance with the route setting request r1, and routes it in the packet transport network 101. Set. The route setting request r1 includes information for specifying the edge node EN as the start point, information for specifying the edge node EN as the end point, a port number to be used, a necessary communication band (bps), and the like. In FIG. 1, as an example, a selection route RT indicated by a bold line is set. A known method may be used as a method for generating the route setting information RTI.

  When the instruction from the maintenance person / other management system 102 is a service provision request r2, the service controller 103 determines a node N to load the program P according to the service provision request r2. The service provision request r2 is a service program provided by the node N, and information (request program identification information) for identifying a program P (hereinafter referred to as a request program) that is requested to be loaded into the node N, Information (guarantee request) that specifies the operation guarantee content is included.

  Next, information held by the service controller 103 and the node N will be described. The service controller 103 has a software function table T1, a calculation resource management table T2, a communication resource management table T3, and a candidate node table T5. The software function table T1 is information defining software functions. Here, the software is a program P that the service controller 103 loads into the node N. The software function table T1 is information prepared in advance. Details of the software function table T1 will be described with reference to FIG.

  The calculation resource management table T2 is a table for managing the calculation resources of the node N. The calculation resource of the node N is, for example, a CPU usage rate or a memory usage rate. In this example, the CPU usage rate will be described as an example. When the program loaded from the service controller 103 is executed on the node N, the CPU usage rate increases. The calculation resource is managed for each VM constructed on the node N, for example. Details of the calculation resource management table T2 will be described with reference to FIG.

  The communication resource management table T3 is a table for managing communication resources of the node N. The communication resource of the node N is, for example, the communication band (Gbps) of the node N, the usage status of the port, and the like. The service controller 103 is set in the packet transport network 101 by referring to the communication resource management table T3 according to the route setting request r1 from the maintenance person / other management system 102 (specifying the edge nodes EN at both ends of the route). The route setting information RTI indicating the route to be generated can be generated. Here, the route set in the packet transport network 101 may be the shortest route or a load-distributed route so that the communication resources of the relay nodes RN are uniform. A policy regarding what kind of route the service controller 103 sets may be stored in advance.

  The candidate node table T5 is a table for managing candidates of nodes to which the program P is loaded. The node to which the program P is loaded is a node on the path set in the packet transport network 101. The service controller 103 refers to the candidate node table T5 according to the service provision request r2 (specifying the service (program P) executed by the node) from the maintenance person / other management system 102, and determines the node to which the program P is loaded. decide. Details of the candidate node table T5 will be described with reference to FIG.

  The node N has a calculation resource table T4. The calculation resource table T4 is a table for recording the calculation resource of the node N (CPU usage rate in this example). The calculation resource is managed for each VM, for example. The calculation resource of the node N is sent from the node N to the service controller 103 and used for updating the calculation resource management table T2. Details of the calculation resource table T4 will be described with reference to FIG.

  Next, the operation of the communication system 100 will be described. The service controller 103 generates route setting information RTI by referring to the communication resource management table T3 according to the route setting request r1 from the maintenance person / other management system 102. The service controller 103 transmits the route setting information RTI to the nodes N (for example, edge nodes EN1, EN2, relay nodes RN1 to RN3) on the selected route RT defined by the route setting information RTI. The node N that has received the route setting information RTI performs routing according to the route setting information RTI. Thereby, the selection route RT is set. Further, the service controller 103 associates the information for identifying the generated route setting information RTI with the information for specifying the node N on the selected route RT defined in the route setting information RTI in a route management table (not shown). Store. That is, the route management table is a table that records a plurality of route setting information RTIs and the node N of each route setting information RTI in association with each other.

  Next, the service controller 103 refers to the candidate node table T5 in accordance with the service provision request r2 from the maintainer / other management system 102, and the VM calculation resource of the node N on the selected route RT when the program P is executed. Confirm. Then, the service controller 103 loads the program P into the VM of the node N on the selected route RT that has sufficient computing resources even if the program P is executed. The node loaded with the program P executes the program P on the designated VM. As a result, it is possible to secure the calculation resource of the program P for which operation guarantee is requested, and to efficiently use the calculation resource in the packet transport network 101. Also, by automating these processes, it is possible to reduce the burden on the maintainer of the packet transport network 101.

<Internal Configuration Example of Service Controller 103>
FIG. 2 is a block diagram illustrating an internal configuration example of the service controller 103. The service controller 103 includes a software function table T1, a calculation resource management table T2, a communication resource management table T3, a maintainer / other management system IF 501, a path generation unit 502, a node resource monitoring unit 503, and a software management unit. 504, a software load unit 505, and a node IF 506.

  Specifically, the software function table T1, the calculation resource management table T2, and the communication resource management table T3 are realized by a storage device (not shown) such as a memory or a hard disk, for example. Specifically, the maintenance / other management system IF 501 and the node IF 506 are realized by, for example, an interface circuit (not shown) including a port. Specifically, the path generation unit 502, the node resource monitoring unit 503, the software management unit 504, and the software load unit 505 are realized, for example, by causing a processor (not shown) to execute a program stored in a storage device. Is done.

  The maintenance person / other management system IF 501 receives a route setting request r1 and a service provision request r2 from the maintenance person / other management system 102. The maintenance person / other management system IF 501 identifies the type of the received request, sends a route setting request r1 to the route generation unit 502, and sends a service provision request r2 to the software management unit 504.

  The route generation unit 502 generates route setting information RTI by referring to the communication resource management table T3 according to the route setting request r1, and generates a route management table. The route generation unit 502 transmits the generated route setting information RTI to the node N included in the selected route RT specified by the route setting information RTI via the node IF 506.

  The node resource monitoring unit 503 monitors the calculation resource of the node N. Specifically, for example, the node resource monitoring unit 503 transmits a calculation resource acquisition request to the node N at a predetermined timing. The node N that has received the acquisition request refers to the calculation resource table T4, reads the CPU usage rate for each execution program 402 from the CPU usage rate 403 for each VM type 401, and sends it to the node resource monitoring unit 503. The node resource monitoring unit 503 receives the CPU usage rate, which is the calculation resource information RSI from the node N, via the node IF 506, and sends the received CPU usage rate to the software management unit 504.

  The software management unit 504 uses the CPU usage rate from the node resource monitoring unit 503 to update the CPU usage rate 304 of the calculation resource management table T2. In addition, when receiving the service provision request r2, the software management unit 504 creates a candidate node table T5 and is requested by the service provision request r2 to which VM of which node N among the nodes N included in the selected route RT. The candidate node table T5 is used to determine whether to execute the requested program. The software management unit 504 sends the determination result and the identification information of the requested program included in the service provision request r2 to the software loading unit 505.

  The software load unit 505 reads the request program stored in the service controller 103 from the storage device based on the request program identification information from the software management unit 504. Then, the software loading unit 505 specifies the VM of the corresponding node N according to the determination result of the software management unit 504, and loads the read program P as a request program.

  The node IF 506 transmits the route setting information RTI and the request program (program P) from the route generation unit 502 to the node N.

<Example of stored contents of table of service controller>
FIG. 3 is an explanatory diagram showing an example of stored contents of the software function table T1 shown in FIG. The software function table T1 includes a program name 201, a load type 202, an applicable node type 203, and an assumed CPU usage rate 204. The software function table T1 is information that the service controller 103 holds in advance.

  The program name 201 is identification information that uniquely identifies the program P to be loaded into the node N. The load type 202 is information that defines the type of load of the program P specified by the program name 201, and includes a constant load and a variable load. The constant load indicates that the load when the program P is executed is constant. The fluctuating load indicates that the load when the program P is executed fluctuates.

  The application node type 203 is information indicating the type of the node N to which the program P specified by the program name 201 is applied. That is, the applicable node type 203 is information indicating the type of the node N that executes the program P specified by the program name 201. When the application node type 203 is “edge”, it can be applied only to the edge node EN. When the application node type 203 is “edge / relay”, it is applicable to both the edge node EN and the relay node RN. When the application node type 203 is “relay”, it is applicable only to the relay node RN.

  The assumed CPU usage rate 204 is a predicted value of the CPU usage rate of the program P specified by the program name 201. The assumed CPU usage rate 204 is a predicted value of the CPU usage rate per CPU.

  FIG. 4 is an explanatory diagram showing an example of stored contents of the calculation resource management table T2 of the selected route RT shown in FIG. The calculation resource management table T2 includes a node name 301, a VM type 302, a loaded program 303, and a CPU usage rate 304. The calculation resource management table T2 is a table that is updated with information from the node N. When a plurality of route setting information RTIs are generated, the calculation resource management table T2 is created for each route setting information RTI.

  The node name 301 is identification information that uniquely identifies the node N. The node name 301 may be set in advance. In addition, when the route setting information RTI is generated, the software management unit 504 of the service controller 103 may store the node N in the selected route RT in the node name 301.

  The VM type 302 is information indicating the type of VM operating on the node N specified by the node name 301. The VM type 302 “operation guarantee program VM” indicates that the VM executes the operation guarantee program. The operation guarantee program is a program P that requires immediate and complete operation guarantee regardless of the load state of the computing resource in the node N.

  The VM type 302 “VM for best effort program” indicates that the VM executes the best effort program. The best effort program is a program P having a guarantee level that does not require a strict guarantee of operation and is allowed not to be executed until a resource is available depending on the load state of the calculation resource in the node N. .

  The VM type 302 is set in advance by the maintenance person / other management system 102.

  The loaded program 303 is a program P loaded on the node N specified by the node name 301. The loaded program 303 is classified for each VM type 302. The loaded program 303 stores the program name of the loaded program P after the service controller 103 loads the program P to each node N.

  The CPU usage rate 304 is a CPU usage rate in the VM defined by the VM type 302 for the loaded program 303. Regarding the CPU usage rate 304, when the software management unit 504 of the service controller 103 receives the CPU usage rate from the node N, the software management unit 504 stores the received CPU usage rate in the CPU usage rate 304 as it is. Alternatively, a predetermined statistical value may be calculated using the received CPU usage rate, and the calculation result may be stored in the CPU usage rate 304.

  When the VM type 302 is the VM for the operation guarantee program, for each CPU usage rate 304 of the loaded program 303, the maximum value of the CPU usage rate collected from the same node N is the same for each loaded program 303 in the CPU usage rate 304. Stored. By setting the maximum value, calculation resource allocation can be performed so that the operation guarantee program VM surely executes the loaded program 303.

  When the VM type 302 is a VM for the best effort program, for each CPU usage rate 304 of the loaded program 303, the average value of the CPU usage rates collected from the same node N is loaded into the CPU usage rate 304. Stored for each. By setting the average value, it is possible to perform calculation resource allocation assuming a case where the load varies.

  FIG. 5 is an explanatory diagram of an example of stored contents of the candidate node table T5. The candidate node table T5 includes a node 601 in the selected path, a candidate flag 602, a CPU usage rate 603, a necessary CPU usage rate 604, and an assumed CPU usage rate 605 after loading. The candidate node table T5 is created by the software management unit 504 when the software management unit 504 of the service controller 103 receives the service provision request r2.

  The selected route node 601 is identification information for uniquely identifying the node N constituting the selected route RT. In the example of FIG. 1, EN1, EN2, RN1, RN2, and RN3 are stored in the selected route node 601 as identification information of the node N that constitutes the selected route RT. The identification information “EN” of the node N indicates that the node N is an edge node. The identification information “RN” of the node N indicates that the node N is a relay node. In the selected route node 601, after the route generation unit 502 of the service controller 103 determines the selected route, the node name of the node N on the selected selected route is stored by the software management unit 504.

  The candidate flag 602 is identification information indicating whether the node 601 in the selected route corresponds to the applicable node type 203 corresponding to the request program in the software function table T1 of FIG. In this example, “1” is applicable and “0” is not applicable.

  For example, when the requested program is “continuity characteristics test”, the software management unit 504 refers to the software function table T1 and sets the application node type 203: “edge” corresponding to the program name 201: “continuity characteristics test”. Identify. When the node 601 in the selected path in FIG. 5 is “EN1”, the candidate flag corresponding to “EN1” is “1” because it corresponds to the applicable node type 203: “edge”. The same applies to the node 601 in the selection path “EN2”. On the other hand, when the node 601 in the selected route is “RN1”, “RN2”, or “RN3”, the candidate node 602 remains “0” because it does not correspond to the applicable node type “edge”.

  The CPU usage rate 603 corresponds to the CPU usage rate 304 of the loaded program 303 in the VM type 302 according to the operation guarantee requirement information (guarantee request) of the requested program included in the service provision request r2 in the calculation resource management table T2. It is the total value. For example, when the guarantee request is “guarantee”, the VM type 302 is “VM for operation guarantee program”. Accordingly, in the calculation resource management table T2, when the node name 301 is “EN1”, the CPU usage rates 304 of “bandwidth control”, “firewall”, and “connectivity monitoring” that are the loaded programs 303 are “18”. “40%”, which is the sum of “%”, “17%”, and “5%”, is stored in the CPU usage rate 603 of the node 601 in the selected path: “EN1” in the candidate node table T5.

  The same applies when the guarantee request is “guaranteed or best effort”. “Warranty or best effort” is a guarantee request that is treated as “guaranteed” as much as possible, and is handled as “best effort” when the load destination of the requested program is not determined in “guaranteed”. In the above example, when the guarantee request is “best effort”, the VM type 302 is “VM for best effort program”. Therefore, in the calculation resource management table T2, in the case of the node name 301: “EN1”, the CPU usage rates of “7 delay time measurement”, “continuity characteristics test” and “traffic monitor” which are the loaded programs 303 are “7”. “20%”, which is the sum of “%”, “5%”, and “8%”, is stored in the CPU usage rate 603 of the node 601 in the selected path “EN1” in the candidate node table T5.

  The required CPU usage rate 604 is a CPU usage rate required for executing the requested program. The software management unit 504 refers to the software function table T1 in FIG. 3, and when the load type 202 of the request program specified by the program name 201 is “constant load”, the software management unit 504 sets the assumed CPU usage rate 204 to the required CPU usage rate. As a necessary CPU usage rate 604. On the other hand, when the load type 202 of the requested program specified by the program name 201 is “variable load”, the software management unit 504 uses the maximum value of the assumed CPU usage rate 204 as the required CPU usage rate, and the required CPU usage rate 604. To store.

  For example, when the requested program is “bandwidth control”, the load type 202 is “constant load” and the assumed CPU usage rate 204 is “20%”, so “20%” is stored in the required CPU usage rate 604. Is done. On the other hand, when the request program is “continuity characteristic test”, the load type 202 is “fluctuating load” and the assumed CPU usage rate 204 is “0 to 10%”, so the required CPU usage rate 604 is “10%”. Is stored. FIG. 5 shows an example in which the request program of the service provision request r2 is “continuity characteristic test” and the guarantee request is “guarantee”.

  The post-load assumed CPU usage rate 605 is a predicted value of the CPU usage rate in the VM that executes the requested program at the load destination node N when the requested program is loaded into the load destination node N. Specifically, for example, the post-load estimated CPU usage rate is a total value of the CPU usage rate 603 and the required CPU usage rate 604. The post-load assumed CPU usage rate 605 stores a value calculated by the software management unit 504 based on the CPU usage rate 603 and the required CPU usage rate 604.

<Internal configuration example of node N>
FIG. 6 is a block diagram illustrating an internal configuration example of the node N. The node N includes a network interface (NW-IF) 701, a VM management unit 702, an operation guarantee program VM 703, a best effort program VM 704, a route table 705, an input / output processing unit 706, a port 707, Have

  The NW-IF 701 is an interface through which the node N communicates with the service controller 103. Data received by the NW-IF 701 from the service controller 103 includes path setting information RTI and a program P. The data sent from the NW-IF 701 to the service controller 103 includes calculation resource information RSI.

  The VM management unit 702 constructs a VM inside the node N (operation guarantee program VM 703, best effort program VM 704), and manages the constructed VM. The VM management unit 702 activates a guest OS (Operating System) and allocates a processor for each VM type. For example, when the node N has x processors, the VM management unit 702 assigns y processors to the operation guarantee program VM 703 and z processors to the best effort program VM 704 (y + z ≦ x).

  It is preferable that the VM management unit 702 does not share the processor between the operation guarantee program VM 703 and the best effort program VM 704. Then, the VM management unit 702 activates the guest OS for each VM, and constructs an operation guarantee program VM 703 and a best effort program VM 704. Each VM is loaded with the program P loaded from the service controller 103.

  In addition, the VM management unit 702 includes a calculation resource monitoring unit 721, a program storage unit 722, and the calculation resource table T4 described above. The calculation resource monitoring unit 721 monitors the usage status of calculation resources such as a CPU usage rate for each VM. For example, the calculation resource monitoring unit 721 repeatedly (for example, periodically) transmits a calculation resource acquisition request to each VM, and uses a calculation resource usage status (for example, a program executed on the VM from the VM). , CPU usage rate) is acquired and stored in the calculation resource table T4. Specifically, the computing resource monitoring unit 721 is realized, for example, by causing a processor to execute a host OS of a node N or an application running on the host OS.

  The program storage unit 722 is a storage area for storing the program P loaded as a request program from the service controller 103. The program P stored in the program storage unit 722 is sent to the corresponding VM. Specifically, the program storage unit 722 is realized by a storage device in the node N, for example.

  The operation guarantee program VM 703 is a VM that executes an operation guarantee program by an assigned processor. The operation guarantee program is a program P that requires immediate and complete operation guarantee regardless of the load state of the computing resources in the node N such as the CPU usage rate. The operation guarantee program VM 703 includes a program instance 731 and a software function execution unit 732. The program instance 731 is an operation guarantee program that runs on the operation guarantee program VM 703. The software function execution unit 732 is a guest OS that executes the program instance 731.

  The best effort program VM 704 is a VM that executes a best effort program by an assigned processor. The best effort program does not require a strict guarantee of operation and guarantees that it will not be executed until the computing resource is free, depending on the load situation of the computing resource in the node N such as the CPU usage rate. Level program P. The best effort program VM 704 includes a program instance 741 and a software function execution unit 742. The program instance 741 is a best effort program that operates on the best effort program VM 704. The software function execution unit 742 is a guest OS that executes the program instance 741. Upon receiving a calculation resource acquisition request from the calculation resource monitoring unit 721, both VMs 703 and 704 return CPU usage rates for the program instances 731 and 741 that are operating.

  As described above, by dividing the VM for each guarantee request in the node N, the calculation resources in the node N can be clearly divided and the program operation guarantee request can be satisfied. In the present embodiment, an example is shown in which one operation guarantee program VM 703 and one best effort program VM 704 are constructed. A plurality of operation guarantee program VMs 703 may be constructed. In this case, the program instances 731 that operate in the respective operation guarantee program VMs 703 are arranged so as not to overlap. Similarly, a plurality of best effort program VMs 704 may be constructed. In this case, the program instances 741 that operate in each best effort program VM 704 are arranged so as not to overlap.

  The route table 705 is a table that stores route setting information RTI developed from the service controller 103 via the NW-IF 701.

  The input / output processing unit 706 refers to the route table 705, determines the output destination port 707 of the packet input from a certain port 707, and passes the packet to the determined output destination port 707. The input / output processing unit 706 analyzes a packet input from a certain port 707 and transfers the packet to the operation guarantee program VM 703 or the best effort program VM 704 according to the analysis result. The port 707 transmits and receives packets.

  The calculation resource table T4 is specifically realized by a storage device (not shown) such as a memory or a hard disk. Specifically, the VM management unit 702, the operation guarantee program VM 703, and the best effort program VM 704 are realized, for example, by causing a processor to execute a program stored in a storage device. Specifically, the NW-IF 701 and the port 707 are realized by, for example, an interface circuit (not shown). Specifically, the path table 705 and the input / output processing unit 706 are realized by, for example, an FPGA (Field-Programmable Gate Array).

<Example of stored contents of table in node>
FIG. 7 is an explanatory diagram of an example of the contents stored in the calculation resource table T4. The calculation resource table T4 includes a VM type 401, an execution program 402, and a CPU usage rate 403. The calculation resource table T4 is a table that is updated according to information from the service controller 103 and the usage status of the CPU in the VM. The example of FIG. 7 shows an example of the calculation resource table T4 that the edge node EN1 has.

  As for the VM type 401, when the VM type is designated from the service controller 103, the node N stores the VM type in the VM type 401.

  The execution program 402 is a program P that is loaded from the service controller 103 and executed on the VM of the VM type 401. For the execution program 402, when the program P is loaded from the service controller 103, the node N stores the program name in the execution program 402.

  The CPU usage rate 403 is the CPU usage rate of the execution program 402 on the VM in the VM type 401. When there are a plurality of processors occupied by the VM, the CPU usage rate of all the processors is obtained. As for the CPU usage rate 403, the node N stores the CPU usage rate acquired when the VM of the VM type 401 executes the CPU usage rate acquisition command at a predetermined timing in the CPU usage rate 403. That is, the latest CPU usage rate is stored in the CPU usage rate 403. The acquisition command is issued at regular time intervals, for example.

<Sequence of Communication System 100>
FIG. 8 is an explanatory diagram showing a sequence example of the communication system 100. The maintenance person / other management system 102 gives a route setting request r1 to the service controller 103 (step S801). Upon receiving the route setting request r1, the service controller 103 determines a route and generates route setting information RTI (step S802). In FIG. 1, EN1 as the start edge node EN and EN2 as the end edge node EN are designated by the route setting request r1, and the selected route RT (edge node EN1, relay nodes RN1, RN2, RN3, edge node EN2) is determined. .

  Then, the service controller 103 transmits the route setting information RTI related to the determined selected route RT to the node N on the selected route RT (step S803). The node N that has received the route setting information RTI stores the received route setting information RTI in the route table 705, and sets a route according to the route setting information RTI (step S804).

  Thereafter, the maintenance person / other management system 102 gives a service provision request r2 to the service controller 103 (step S805). The service provision request r2 includes a program name that is identification information of the requested program and a guarantee request. There are three types of guarantee requests, for example, “guarantee”, “best effort”, and “guarantee or best effort”.

  “Warranty” is a guarantee request designated when the requested program is to be executed by the operation guarantee program VM 703. Specifically, for example, “guarantee” is a guarantee request indicating the guarantee content to comply with the operation guarantee of the load target program.

  “Best effort” is a guarantee request designated when the requested program is to be executed by the VM 704 for the best effort program. Specifically, for example, “best effort” is a guarantee request indicating a guarantee content that the operation guarantee of the load target program does not have to be observed.

  “Guaranteed or Best Effort” is a guarantee request designated when the requested program is to be executed by the operation guarantee program VM 703 as much as possible. Specifically, for example, “guarantee or best effort” is a guarantee request indicating the guarantee content that the guarantee content of “best effort” is applied unless the guarantee content of “guarantee” is satisfied. If the operation guarantee program VM 703 does not guarantee the operation, the request program is assigned to the best effort program VM 704.

  Upon receiving the service provision request r2, the service controller 103 executes a load target node / VM selection process (step S806). The load target node / VM selection process (step S806) is a process of selecting the load destination of the requested program, that is, the node N and the VM on the node N. Details of the load target node / VM selection process (step S806) will be described with reference to FIG.

  When the service controller 103 selects the load destination in the load target node / VM selection process (step S806), the service controller 103 transmits the request program P and the VM type of the VM that executes the program P to the load destination node N (step S806). S807). The node N stores the transmitted program P in the program storage unit 722 (step S808). Then, the node N allocates the program P as a program instance to the VM selected as the load destination indicated by the transmitted VM type, and executes the requested program on the VM (step S809).

<Load target node / VM selection process (step S806)>
FIG. 9 is a flowchart showing a detailed processing procedure example of the load target node / VM selection process (step S806) shown in FIG. The software management unit 504 of the service controller 103 sets the node name of the node N on the selected route RT determined in step S802 to the selected route node 601 of the candidate node table T5 (step S901). In the example of FIG. 1, EN1, EN2, RN1, RN2, and RN3 are stored in the selected route node 601 as identification information of the node N that constitutes the selected route RT.

  Next, the software management unit 504 of the service controller 103 searches the program name 201 based on the program name of the request program included in the received service provision request r2 with reference to the software function table T1 shown in FIG. The application node type 203 of the corresponding program name 201 is specified (step S902). For example, when the program name 201 of the request program is “bandwidth control”, the applicable node type 203 is “edge”. When the program name 201 of the request program is “traffic monitor”, the applicable node type 203 is “edge / relay”.

  Next, the software management unit 504 of the service controller 103 selects a candidate node from the nodes 601 in the selected route based on the applicable node type 203 specified in step S902 (step S903). For example, when the applicable node type 203 is “edge”, the software management unit 504 of the service controller 103 sets the edge node as a candidate node. That is, the software management unit 504 changes the candidate flag 602 in which the node 601 in the selected route in the candidate node table T5 is an “edge node” to “1”.

  When the applicable node type 203 is “edge / relay”, the software management unit 504 of the service controller 103 sets the edge node and the relay node as candidate nodes. That is, the software management unit 504 changes the candidate flag 602 in which the node 601 in the selected route in the candidate node table T5 is “edge node” and “relay node” to “1”. When the application node type 203 is “relay”, the software management unit 504 of the service controller 103 sets the relay node as a candidate node. That is, the software management unit 504 changes the candidate flag 602 in which the node 601 in the selected route in the candidate node table T5 is “relay node” to “1”. The service controller 103 sets a candidate node by setting the candidate flag 602 of the candidate node table T5 to “1”. In the example of FIG. 5, since the applicable node type 203 is “edge”, the candidate flag 602 is set to “1” for the nodes 601 in the selected route: the nodes EN1 and EN2.

  Thereafter, the software management unit 504 of the service controller 103 determines a guarantee request (step S904). Steps S904 to S908 are executed for each candidate node. In step S904, when the guarantee request is “best effort” (step S904: best effort), the software management unit 504 of the service controller 103 selects the best effort which is the VM type corresponding to the guarantee request among the VM types of the candidate nodes. The CPU CPU usage rate of the program VM is acquired and stored in the CPU usage rate 603 of the candidate node table T5 corresponding to the node 601 in the selection path of the candidate node (step S905), and the process proceeds to step S907.

  Specifically, for example, the software management unit 504 of the service controller 103 refers to the calculation resource management table T2 in FIG. 4 based on the program name of the received request program, and identifies the entry of the candidate node from the node name 301. . Since the guarantee request is “best effort”, the loaded program and CPU usage rate corresponding to the VM type 302: “VM for best effort program” are obtained from the loaded program 303 and CPU usage rate 304 in the entry of the candidate node. Identify. The software management unit 504 of the service controller 103 calculates the total value of the identified CPU usage rate 304 and stores the calculated total value in the CPU usage rate 603.

  For example, in FIG. 4, when the node name 301 of the candidate node is EN1 and the guarantee request is “best effort”, the loaded program 303 and the CPU usage rate 30 are “7%” of “delay time measurement”. “5%” in the “continuity characteristics test” and “8%” in the “traffic monitor”. Therefore, the software management unit 504 of the service controller 103 calculates “20%” as the total value of the CPU usage rate. Then, the software management unit 504 of the service controller 103 stores “20%”, which is the total value, in the CPU usage rate 603 of the candidate node “EN1” in the candidate node table T5.

  In step S904, when the guarantee request is “guarantee” or “guarantee or best effort” (step S904: “guarantee” or “guarantee or best effort”), the software management unit 504 of the service controller 103 selects the candidate node. The CPU usage rate of the operation guarantee program VM that is the VM type corresponding to the guarantee request among the VM types is acquired, and the CPU usage rate of the candidate node table T5 corresponding to the node 601 in the selection path of the candidate node is acquired. In step S906, the process proceeds to step S907.

  Specifically, for example, the software management unit 504 of the service controller 103 refers to the calculation resource management table T2 in FIG. 4 based on the program name of the received request program, and identifies the entry of the candidate node from the node name 301. . Since the guarantee request is “guaranteed” or “guaranteed or best effort”, the loaded program 303 corresponding to the VM type 302: “VM for operation guarantee program” and the CPU usage rate are set to the loaded program 303 in the candidate node entry. And the CPU usage rate 304. The software management unit 504 of the service controller 103 calculates the total value of the identified CPU usage rate 304 and stores the calculated total value in the CPU usage rate 603.

  For example, in FIG. 4, when the node name 301 of the candidate node is EN1 and the guarantee request is “guaranteed” or “guaranteed or best effort”, the loaded program 303 and the CPU usage rate 304 are “bandwidth control”. “18%”, “firewall” “17%”, and “connectivity monitoring” “5%”. Therefore, the software management unit 504 of the service controller 103 calculates “40%” as the total value of the CPU usage rate 304. Then, the software management unit 504 of the service controller 103 stores “40%”, which is the total value, in the CPU usage rate 603 of the candidate node “EN1” in the candidate node table T5.

  That is, the value stored in the CPU usage rate 603 is the VM type corresponding to the received guarantee request (if the guarantee request is “guaranteed” or “guaranteed or best effort”, “operation guarantee program VM”, guarantee request Is “best effort”, “VM for best effort program”)).

  In step S907, the software management unit 504 of the service controller 103 refers to the software function table T1 and acquires the load type and the assumed CPU usage rate of the request program as the load type 202 and the assumed CPU usage rate 204 (step S907). . Specifically, for example, the software management unit 504 of the service controller 103 searches the program name 201 based on the program name of the request program included in the received service provision request r2, and searches for the request program of the corresponding program name 201. The load type and the assumed CPU usage rate are specified from the load type 202 and the assumed CPU usage rate 204 in the corresponding program name 201 entry. For example, when the program name 201 of the request program is “bandwidth control”, the load type 202 is “constant load”, and the assumed CPU usage rate 204 is “20%”.

  Since the assumed CPU usage rate 204 is a predicted value of the CPU usage rate per processor, when the number of processors assigned to the VM of the candidate node corresponding to the guarantee request is 1, the obtained assumption is obtained. The CPU usage rate 204 is used as it is.

  On the other hand, when the number of processors assigned to the VM of the candidate node corresponding to the guarantee request is plural, the software management unit 504 of the service controller 103 uses the acquired assumed CPU usage rate 204 as a candidate corresponding to the guarantee request. Divide by the number of processors assigned to the VM of the node. For example, when the number of processors allocated to the VM of the candidate node corresponding to the guarantee request is four and the assumed CPU usage rate 204 acquired from the software function table T1 is “20%”, the assumed CPU usage after division is The rate is “5%”. In the following description, the case where the number of processors allocated to the VM of the candidate node corresponding to the guarantee request is one will be described as an example, but the assumed CPU usage rate after division is also assumed as the assumed CPU usage rate 204. Be treated.

  Then, the software management unit 504 of the service controller 103 stores the acquired assumed CPU usage rate 204 in the necessary CPU usage rate 604 of the candidate node candidate node table T5. Specifically, for example, when the request program is “traffic monitor”, the assumed CPU usage rate 204 is “10%”, so that the required CPU usage rate of the candidate node (for example, EN1) in the candidate node table T5 “10%” is stored in 604.

  Further, for example, when the load type 202 for the request program is “variable load”, the software management unit 504 of the service controller 103 acquires the assumed CPU usage rate 204 and sets the maximum value of the assumed CPU usage rate 204 as a candidate. Stored in the necessary CPU usage rate 604 of the node table T5. Specifically, for example, when the request program is “load balancer”, the assumed CPU usage rate 204 is “0 to 30%”, and therefore the CPU required for the candidate node (for example, EN1) in the candidate node table T5 In the usage rate 604, the maximum value “30%” of the assumed CPU usage rate 204: “0 to 30%” is stored.

  Thereafter, the software management unit 504 of the service controller 103 calculates a post-load assumed CPU usage rate 605 for each candidate node (step S908). Specifically, for example, the software management unit 504 of the service controller 103 adds the CPU usage rate 603 and the necessary CPU usage rate 604 for each entry whose candidate flag is “1” in the candidate node table T5. The post-load assumed CPU usage rate is calculated and stored in the post-load assumed CPU usage rate 605. In the entry of node 601: “EN1” in the selection path of the candidate node table T5 in FIG. 5, the CPU usage rate 603 is “40%” and the necessary CPU usage rate 604 is “10%”. “50%” is stored as the rate 605.

  Thereafter, the software management unit 504 of the service controller 103 determines whether there is a candidate node having a post-load assumed CPU usage rate 605 of 100% or less in the candidate node group (step S909). That is, the software management unit 504 of the service controller 103 determines, for each candidate node, whether or not the post-load assumed calculation resource is a calculation resource that can comply with the operation guarantee of the request program. When there is a candidate node whose estimated CPU usage rate 605 after loading is 100% or less (step S909: Yes), the software management unit 504 of the service controller 103 causes the CPU in the candidate node whose estimated CPU usage rate 605 after loading is 100% or less. The usage rate acquisition source VM (the best effort program VM for which the CPU usage rate has been acquired in step S905 or the operation guarantee program VM for which the CPU usage rate has been acquired in step S906) is determined as the load destination (step S910).

  When there are a plurality of candidate nodes whose post-load assumed CPU usage rate 605 is 100% or less, the CPU usage rate acquisition source VM of the candidate node having the lowest post-load expected CPU usage rate 605 may be determined as the load destination. . In addition, when there are a plurality of candidate nodes whose post-load assumed CPU usage rate 605 is 100% or less, the CPU usage rate acquisition source VM in any candidate node whose post-load assumed CPU usage rate 605 is equal to or lower than the predetermined CPU usage rate. May be determined as the load destination. Thereafter, the software management unit 504 of the service controller 103 transmits the requested program and the VM type corresponding to the guarantee request to the candidate node selected in step S910 (step S807).

  On the other hand, when there is no candidate node whose post-load assumed CPU usage rate 605 is 100% or less (step S909: No), the software management unit 504 of the service controller 103 determines a guarantee request (step S911). When the guarantee request is “guaranteed or best effort” (step S911: guarantee or best effort), the software management unit 504 of the service controller 103 updates the guarantee request to “best effort” (step S912), and proceeds to step S905. Return. In other words, in step 909, there is no candidate node with an assumed post-loading CPU usage rate 605 of 100% or less (step S909: No), so operation cannot be guaranteed at the load candidate node. Accordingly, by updating the guarantee request to “best effort” and returning to step S905, when the guarantee request is determined in step S911, step S914 described later is executed. That is, the operation of the requested program is to be performed with “guaranteed” but cannot be assured.

  Further, when the guarantee request is “guaranteed” (step S911: guarantee), there is no candidate node whose post-load assumed CPU usage rate 605 is 100% or less (step S909: No), so operation guarantee is performed at the load destination candidate node. become unable. In this case, the software management unit 504 of the service controller 103 outputs a notification that the request program load destination cannot be determined to the maintainer / other management system 102 (step S913). That is, although the requested program operation is to be performed with “guaranteed”, the operation cannot be guaranteed and cannot be changed to “best effort”, so the maintenance person / other management system 102 is prompted to reset the selected route RT. .

  When the guarantee request is “best effort” (step S911: best effort), the software management unit 504 of the service controller 103 determines the ratio of the CPU usage rate of the variable load program to the total CPU usage rate of the VM for the best effort program. The CPU usage rate acquisition source VM in the candidate node with the highest is determined as the load destination (step S914). Further, the software management unit 504 of the service controller 103 loads the CPU usage rate acquisition source VM in the candidate node in which the ratio of the CPU usage rate of the variable load program to the total CPU usage rate of the best effort program VM is equal to or higher than a predetermined rate. It may be determined first.

  The total CPU usage rate of the best effort program VM is a total value of the CPU usage rates 304 of the loaded program 303 of the best effort program VM 302 corresponding to the node name 301 of the candidate node in the calculation resource management table T2. For example, in FIG. 4, in the entry of the node name 301: “EN1”, the loaded program 303 of the VM 302 for best effort program is “delay time measurement”, “continuity characteristic test”, and “traffic monitor”. The respective CPU usage rates 304 are “7%”, “5%”, and “8%”. Therefore, the total CPU usage rate of the best effort program VM for the candidate node EN1 is “20%”.

  Of the loaded program 303 of the best effort program VM 302: “delay time measurement”, “continuity characteristic test”, and “traffic monitor”, the variable load program is “delay time measurement” and “continuity characteristic test”. . Therefore, the CPU usage rate of the variable load program is the total value “12%” of the CPU usage rates “7%” and “5%”.

  The ratio of the CPU usage rate of the variable load program to the total CPU usage rate of the best effort program VM is obtained, for example, by dividing the CPU usage rate of the variable load program by the total CPU usage rate of the best effort program VM. . In the case of the above example, 12% / 20% = 3/5 (= 60%).

  The software management unit 504 of the service controller 103 selects the candidate node with the highest ratio as the load destination. That is, since the CPU usage rate 304 that is the average value of the CPU usage rate of the variable load program is acquired in step S905, the higher the rate of the variable load program, the higher the probability that the load will decrease. This is because the CPU usage rate can be suppressed. Thereafter, the software management unit 504 of the service controller 103 transmits the requested program and the VM type corresponding to the guarantee request to the candidate node selected in step S914 (step S807).

  As described above, according to the present embodiment, it is possible to secure the calculation resource of the software for which the operation guarantee is requested and to efficiently use the calculation resource in the network. In addition, since the software is executed on the nodes in units of VMs, calculation resources can be acquired for each VM. Therefore, it is possible to easily specify the calculation resource of the software. Furthermore, by automating these processes, the burden on the network maintainer can be reduced.

  Further, the load target node / VM selection process (step S806) is executed with the service provision request r2 as a trigger, but the VM calculation resource of the node to which the request program is allocated has a predetermined threshold value (for example, CPU When the service controller 103 detects that the usage rate is 100% or more, the service controller 103 may re-execute the load target node / VM selection process (step S806).

  The predetermined threshold value here may be a value different for each VM type. For example, when the VM is an operation guarantee program VM and the service controller 103 detects that the calculation resource has reached 100% or more, the service controller 103 restarts the load target node / VM selection process (step S806). Run. In addition, the VM for the best effort program allows the CPU usage rate to exceed 100%, but if the service controller 103 detects that the state in which the CPU usage rate has exceeded 100% has continued for a predetermined time or longer, The controller 103 re-executes the load target node / VM selection process (step S806).

  In the case of re-execution, in the load target node / VM selection process (step S806), execution starts from step S904. Thereby, it is possible to dynamically suppress an increase in calculation resources even after the load destination is determined. When the request program is assigned to another node N, the program instance of the request program executed in the VM of the original node N is deleted.

  In the above-described embodiment, the program P is loaded. However, the program P may be stored in the node N in advance. In this case, instead of loading the program P, the service controller 103 loads the identification information (for example, the program name) of the program P, and the node N that has received the identification information of the program P receives the identification information of the program P by the VM management unit. What is necessary is just to build a program instance in the designated VM. In this way, by loading the identification information of the program P instead of the program P, it is possible to reduce the amount of communication from the service controller 103 to the node N.

  In the embodiment described above, the program P to be loaded to the node N is selected based on the route setting request r1 received from the maintenance / other management system 102 and the service provision request r2 received following the route setting request r1. However, the program P to be loaded into the node N may be selected based on the route setting request r1 and the service provision request r2 that are independently received by the service controller 103. That is, the service controller 103 receives a plurality of route setting requests r1, sets a plurality of selected routes RT in the packet transport network 101, and refers to the route information table based on the service provision request r2 received thereafter. Then, the selection route RT for loading the request program may be selected from the plurality of selection routes RT, and the program P to be loaded to the node N on the selected selection route RT may be selected.

  In this case, even if the service provision request r2 includes information for specifying the selection route RT for loading the request program, the service controller 103 sets the selection route RT for loading the request program together with the service provision request r2. The configuration may be such that information to be specified is received.

  With the configuration in which the service controller 103 receives the route setting request r1 and the service provision request r2 independently, the route setting and selection of the program P to be loaded are executed independently, for example, the route setting request r1 is not received. When the service provision request r2 is received, the program P can be loaded into one of a plurality of preset routes without changing the already set route, and the already loaded program Since P can be changed, the load destination of the program P can be determined flexibly. In addition, it is possible to efficiently use computational resources in the network.

  The present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the appended claims. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and the present invention is not necessarily limited to those having all the configurations described. A part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Moreover, you may add the structure of another Example to the structure of a certain Example. Moreover, you may add, delete, or replace another structure about a part of structure of each Example.

  In addition, each of the above-described configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, with an integrated circuit, and the processor realizes each function. It may be realized by software by interpreting and executing the program to be executed.

  Information such as programs, tables, and files that realize each function can be stored in a storage device such as a memory, a hard disk, and an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, and a DVD.

  Further, the control lines and the information lines are those that are considered necessary for the explanation, and not all the control lines and the information lines that are necessary for the mounting are shown. In practice, it can be considered that almost all the components are connected to each other.

100 communication system 101 packet transport network 103 service controller 502 route generation unit 503 node resource monitoring unit 504 software management unit 505 software load unit 702 VM management unit 706 input / output processing unit 721 calculation resource monitoring unit 722 program storage unit N node T1 software Function table T2 Calculation resource management table T4 Calculation resource table T5 Candidate node table

Claims (14)

A control device for controlling a plurality of transfer devices for transferring data in a network,
Each of the plurality of transfer devices includes a first virtual machine that executes the program so as to comply with the operation guarantee of the program, and a second that does not need to comply with the operation guarantee of the program when the program is executed. With a virtual machine,
The control apparatus includes a processor and a storage device that stores path management information indicating a path set in the network,
The processor is
A reception process for receiving a guarantee request for guaranteeing operation of a load target program that is a program to be loaded into the plurality of transfer devices;
A selection process for selecting, from the plurality of transfer devices, a transfer device group constituting the route based on the route management information;
Either one of the first virtual machine and the second virtual machine constructed in the transfer device selected by the selection process in response to the guarantee request received by the reception process. A determination process for determining a load destination of the load target program;
The control apparatus characterized by performing. The storage device further stores assumed calculation resources for each program,
The warranty request is:
Including the guarantee contents of the first guarantee contents that comply with the operation guarantee of the load target program and the second guarantee contents that do not need to comply with the operation guarantee of the load target program,
The processor is
For the transfer device selected by the selection process, an acquisition process for acquiring a current computing resource in any one of the virtual machines;
Based on the assumed calculation resource of the load target program stored in the storage device and the current calculation resource acquired by the acquisition process, the load target program is constructed in the selected transfer device. A calculation process for calculating an assumed calculation resource after loading when the virtual machine is loaded on any one of the virtual machines;
Run further,
The control device according to claim 1, wherein in the determination process, the processor determines the load destination virtual machine based on the post-load assumed calculation resource calculated by the calculation process.   In the acquisition process, the processor acquires the current calculation resource of the first virtual machine in the selected transfer device when any one of the guarantee contents is the first guarantee contents, 3. The control device according to claim 2, wherein when the guarantee content is the second guarantee content, a current calculation resource of the second virtual machine in the selected transfer device is acquired.   In the determination process, the processor determines whether the post-load assumed calculation resource is a calculation resource capable of complying with the operation guarantee of the load target program in any one of the virtual machines of the selected transfer device. If the post-load assumed calculation resource is a calculation resource that can comply with the operation guarantee of the load target program, the virtual machine of the selected transfer device is determined as the load destination of the load target program. The control device according to claim 2.   In the determination process, the processor includes a plurality of the one virtual machines of the selected transfer apparatuses in which the post-load assumption calculation resource can comply with the operation guarantee of the load target program in the transfer apparatus group. In this case, among the plurality of existing virtual machines, the one or more virtual machines of the selected transfer device whose estimated calculation resource after loading is the minimum or a predetermined threshold value or less are loaded with the load target program. The control device according to claim 4, wherein the control device is determined first.   In the determination process, the processor, when the post-load assumption calculation resource does not exist in the transfer device group, the one virtual machine of the selected transfer device that can comply with the operation guarantee of the load target program. 5. The control apparatus according to claim 4, wherein a load destination of the load target program is determined according to any one of the guarantee contents.   7. The determination process according to claim 6, wherein, in the determination process, when any one of the guarantee contents is the first guarantee contents, the processor outputs a notification that the load destination of the load target program cannot be determined. Control device. The storage device stores, for each program, load type information that defines a load type indicating whether the load is constant or variable,
In the determination process, when any one of the guarantee contents is the second guarantee contents, the processor, for each of the selected transfer devices, out of the group of programs being executed in the second virtual machine. Calculate the ratio of the program whose load type is fluctuating, and calculate the one virtual machine of the selected transfer apparatus whose calculated ratio for each selected transfer apparatus is the maximum or a predetermined ratio or more. The control apparatus according to claim 5, wherein the control apparatus determines a load destination of the load target program. The warranty request is:
If the first warranty content is not satisfied, the second warranty content further includes a third warranty content,
In the determination process, when any one of the guarantee contents is the third guarantee contents, the processor changes any one of the guarantee contents from the third guarantee contents to the second guarantee contents,
The control device according to claim 6, wherein the processor re-executes the acquisition process, the calculation process, and the determination process according to the changed second guarantee content.   The control device according to claim 1, wherein the processor executes a transmission process of transmitting information related to the load target program to a load destination virtual machine determined by the determination process.   3. The processor according to claim 2, wherein the processor re-executes the acquisition process, the calculation process, and the determination process when the calculation resource of the load destination becomes a predetermined threshold value or more. Control device. A control program executed by a processor of a control device that controls a plurality of transfer devices that transfer data in a network,
Each of the plurality of transfer devices includes a first virtual machine that executes the program so as to comply with the operation guarantee of the program, and a second that does not need to comply with the operation guarantee of the program when the program is executed. With a virtual machine,
The control apparatus includes the processor and a storage device that stores path management information indicating a path set in the network,
The control program is
In the processor,
A reception process for receiving a guarantee request for guaranteeing operation of a load target program that is a program to be loaded into the plurality of transfer devices;
A selection process for selecting, from the plurality of transfer devices, a transfer device group constituting the route based on the route management information;
Either one of the first virtual machine and the second virtual machine constructed in the transfer device selected by the selection process in response to the guarantee request received by the reception process. A determination process for determining a load destination of the load target program;
A control program characterized by causing The storage device further stores assumed calculation resources for each program,
The warranty request is:
Including the guarantee contents of the first guarantee contents that comply with the operation guarantee of the load target program and the second guarantee contents that do not need to comply with the operation guarantee of the load target program,
The control program is
In the processor,
For the transfer device selected by the selection process, an acquisition process for acquiring a current computing resource in any one of the virtual machines;
Based on the assumed calculation resource of the load target program stored in the storage device and the current calculation resource acquired by the acquisition process, the load target program is constructed in the selected transfer device. A calculation process for calculating an assumed calculation resource after loading when the virtual machine is loaded on any one of the virtual machines;
A transmission process for transmitting information related to the load target program to the load destination virtual machine determined by the determination process;
Is executed further,
13. The determination process according to claim 12, wherein the control program causes the processor to determine the load destination virtual machine based on the post-load assumed calculation resource calculated by the calculation process. Control program. A communication system comprising a plurality of transfer devices that transfer data in a network, and a control device that controls the plurality of transfer devices,
Each of the plurality of transfer devices includes a first virtual machine that executes the program so as to comply with the operation guarantee of the program, and a second that does not need to comply with the operation guarantee of the program when the program is executed. With a virtual machine,
The control apparatus includes a processor and a storage device that stores path management information indicating a path set in the network,
The processor is
A reception process for receiving a guarantee request for guaranteeing operation of a load target program that is a program to be loaded into the plurality of transfer devices;
A selection process for selecting, from the plurality of transfer devices, a transfer device group constituting the route based on the management information;
Either one of the first virtual machine and the second virtual machine constructed in the transfer device selected by the selection process in response to the guarantee request received by the reception process. A determination process for determining a load destination of the load target program;
A transmission process for transmitting information related to the load target program to the load destination virtual machine determined by the determination process;
Run
Among the plurality of transfer devices, the transfer device that constructs the load destination virtual machine is:
A communication system, wherein when the information related to the load target program is received, the load target program is executed on the load destination virtual machine.

Images