JP6315467B2 - Network recovery system and program - Google Patents

Description

  The present invention is connected to a monitoring system that monitors a failure of a node occurring in at least one network including a plurality of nodes and a node control system that controls each node. When a failure occurs in any of the nodes, the node The present invention relates to a network recovery system and a program for giving an instruction to reconfigure a network for each service to a control system.

  Conventionally, research and technical development related to SDN (Software Defined Networking) / NFV (Network Function Virtualization) have been actively conducted. SDN / NFV is expected to shorten new service opening and reduce CAPEX (Capital expenditure) / OPEX (Operating expenditure). Here, in order to reduce OPEX, it is required to automate failure recovery on a virtual network that realizes SDN / NFV. For example, it is necessary to execute automatic recovery in service units, not automatic recovery from a microscopic viewpoint such as device failure units or virtual machine units.

  For example, in the physical resource control management system described in Patent Document 1, the hardware recovery method and the software recovery method cooperate with each other without impairing the recovery speed, and attempt to realize high-speed failover. That is, this physical resource control management system is composed of a hardware space 6100 and a software space 6500, and the failure management means 6552 controls the resource allocation means 6551 by linking hardware failure management and software failure management. doing.

JP 2008-107896 A

  However, considering that the service level agreement (SLA) differs for each service, there are cases where a single recovery means cannot cope with the problem. Further, the technique described in Patent Document 1 merely shows the operation inside the server, and is not sufficient for service restoration.

  FIG. 14 is a diagram illustrating a conventional network configuration example. The networks a to c include a plurality of nodes and are configured so that each node can connect to the Internet. Each node has a routing function, a firewall function, a virtual machine construction function, or the like. In FIG. 14, the monitoring system a monitors a failure in the network a, and the node control system a controls each node included in the network a. Similarly, the monitoring system b monitors a failure of the network b, and the node control system b controls each node included in the network b. Similarly, the monitoring system c monitors a failure in the network c, and the node control system c controls each node included in the network c.

  FIG. 15 is a diagram illustrating a logical configuration of the service a. The firewall FWa1 has a function of connecting the virtual machines VMa1 to a3 to the Internet. 16 to 18 are diagrams illustrating configuration examples of the service a. As shown in FIG. 16 to FIG. 18, the FWa 1 that provides the firewall function fails, the monitoring system b detects this, and the node control system b classified as the same network system is provided with a recovery means. May instruct the creation of virtual machines and provide firewall functions.

  However, if the node that accommodated the firewall FWa1 has failed, the service a cannot be continued as shown in FIG. 18 even if the function of the firewall FWa1 is implemented by another node. There is a possibility that the virtual machine VMa1 is generated. Also, as shown in FIG. 18, when straddling a network for recovery, there is a possibility that the function of the firewall FWa1 cannot be constructed.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a network recovery system and program capable of performing optimal recovery processing for each service.

  (1) In order to achieve the above object, the present invention takes the following measures. That is, the network recovery system of the present invention is connected to a monitoring system that monitors a failure of a node occurring in at least one network including a plurality of nodes and a node control system that controls each of the nodes. When a failure occurs, the network recovery system gives an instruction to the node control system to reconfigure the network on a service basis, and includes failure information indicating that a failure has occurred in any node from the monitoring system. Based on the failure information and the network configuration information, using an information integration unit that acquires and outputs the acquired failure information and a plurality of algorithms for creating a recovery plan for recovering a network failure , Using free network resources for each service An optimal design unit that creates an old plan; and an instruction creation unit that creates network setting information for restoration using the restoration plan and outputs the created network setting information to the node control system; It is characterized by providing.

  In this way, it is possible to acquire failure information indicating that a failure has occurred in any node from the monitoring system, output the acquired failure information, and create multiple recovery plans for creating a recovery plan that restores network failures. Create a recovery plan that uses available network resources for each service based on failure information and network configuration information, and use the recovery plan to create network setting information for recovery. Since the network setting information created for the node control system is output, it is possible to automatically perform service restoration when an alarm indicating the occurrence of a failure is triggered. In addition, since a recovery plan using free resources of the network is created for each service, it is possible to realize a recovery process in an optimal form for each service. Furthermore, since it specializes in instructing instead of controlling recovery, it is possible to realize a system that performs automatic recovery for each service without making a major change to the existing system configuration.

  (2) In the network recovery system of the present invention, the information integration unit records network configuration information, resource information, and failure information, and the optimum design unit is based on the network configuration information, resource information, and failure information. And identifying nodes constituting the recovery plan.

  In this way, the information integration unit records network configuration information, resource information, and failure information, and the optimum design unit identifies the nodes that make up the recovery plan based on the network configuration information, resource information, and failure information. Recovery processing can be realized in an optimum manner for each service.

  (3) In the network recovery system of the present invention, the information integration unit shares the acquired failure information and outputs common failure information, while individualizing the recovery plan created by the optimum design unit The recovery plan is output to the instruction creating unit, and the instruction creating unit creates network setting information for recovery based on the individualized recovery plan.

  Thus, since the information integration unit shares the acquired failure information and outputs the common failure information, an increase in the recovery algorithm can be suppressed. In addition, since the recovery plan created by the optimum design unit is individualized and the individualized recovery plan is output to the instruction creation unit, it is possible to provide information that can be handled by each node control system.

  (4) In the network recovery system of the present invention, the information integration unit includes an information conversion table, and executes the sharing of the failure information and the individualization of the recovery plan using the information conversion table. It is characterized by.

  As described above, the sharing of the failure information and the individualization of the recovery plan are executed using the information conversion table, so that the process can be simplified and speeded up.

  (5) Further, in the network recovery system of the present invention, the instruction creating unit determines a priority for each service, and when the restoration of a plurality of services is necessary at the same time, the priority is given to the node control system. And outputting the network setting information.

  In this way, priority is set for each service, and when multiple services need to be restored at the same time, the priority and network setting information is output to the node control system. Can be executed.

  (6) The program of the present invention is connected to a monitoring system that monitors a failure of a node that occurs in at least one network including a plurality of nodes, and a node control system that controls each of the nodes. A server device program that gives an instruction to the node control system to reconfigure the network in units of services when a failure occurs in the node control system, and indicates that a failure has occurred in any node from the monitoring system Based on the failure information and the network configuration information, using one of a plurality of algorithms for acquiring information, outputting the acquired failure information, and creating a recovery plan for recovering a network failure Create a recovery plan that uses available network resources for each service And a process for generating network setting information for recovery using the recovery plan and outputting the generated network setting information to the node control system. It is characterized by that.

  In this way, it is possible to acquire failure information indicating that a failure has occurred in any node from the monitoring system, output the acquired failure information, and create multiple recovery plans for creating a recovery plan that restores network failures. Create a recovery plan that uses available network resources for each service based on failure information and network configuration information, and use the recovery plan to create network setting information for recovery. Since the network setting information created for the node control system is output, it is possible to automatically perform service restoration when an alarm indicating the occurrence of a failure is triggered. In addition, since a recovery plan using free resources of the network is created for each service, it is possible to realize a recovery process in an optimal form for each service. Furthermore, since it specializes in instructing instead of controlling recovery, it is possible to realize a system that performs automatic recovery for each service without making a major change to the existing system configuration.

  According to the present invention, it is possible to automatically perform service recovery triggered by the occurrence of an alarm indicating the occurrence of a failure. In addition, the restoration process can be realized in an optimum form for each service. Furthermore, it is possible to realize a system that performs automatic recovery for each service without making a major change to the existing system configuration.

It is a block diagram which shows schematic structure of the network recovery system which concerns on this embodiment. It is a figure which shows the example of sharing and individualization. 3 is a flowchart showing the operation of the optimum design unit 101. It is a figure which shows the example of a structure information database. It is a figure which shows the resource information example. It is a figure which shows the resource information example. It is a figure which shows the operation example of the network recovery system which concerns on this embodiment. It is a figure which shows the operation example of the network recovery system which concerns on this embodiment. It is a figure which shows the operation example of the network recovery system which concerns on this embodiment. It is a figure which shows the network structure which should be recovered. It is a figure which shows the modification of this embodiment. It is a figure which shows the logic structure of the service b. It is a figure which shows the traffic amount with respect to time. It is a figure which shows the example of a conventional network structure. It is a figure which shows the logical structure of the service a. It is a figure which shows the structural example of the service a. It is a figure which shows the structural example of the service a. It is a figure which shows the structural example of the service a.

  In order to reduce OPEX, the present inventors are required to automate failure recovery on a virtualized network that realizes SDN / NFV. However, it is not automatic recovery from a micro perspective, but on a service basis. Focusing on the need to execute automatic recovery, a system with an optimal recovery algorithm is provided for each service, and a configuration that directs the recovery to the node control system from that system is adopted. Thus, the present inventors have found that recovery processing can be realized and have come to the present invention.

  That is, the network recovery system of the present invention is connected to a monitoring system that monitors a failure of a node occurring in at least one network including a plurality of nodes and a node control system that controls each of the nodes. When a failure occurs, the network recovery system gives an instruction to the node control system to reconfigure the network on a service basis, and includes failure information indicating that a failure has occurred in any node from the monitoring system. Based on the failure information and the network configuration information, using an information integration unit that acquires and outputs the acquired failure information and a plurality of algorithms for creating a recovery plan for recovering a network failure , Using free network resources for each service An optimal design unit that creates an old plan; and an instruction creation unit that creates network setting information for restoration using the restoration plan and outputs the created network setting information to the node control system; It is characterized by providing.

  As a result, the present inventors have made it possible to automatically perform service recovery triggered by the occurrence of an alarm indicating the occurrence of a failure. In addition, it is possible to realize recovery processing in an optimal form for each service, and furthermore, it is possible to realize a system that performs automatic recovery for each service without making major changes to the existing system configuration. . Embodiments of the present invention will be specifically described below with reference to the drawings.

  FIG. 1 is a block diagram showing a schematic configuration of a network recovery system according to the present embodiment. This network recovery system includes an optimal network recovery system 100 and a monitoring integrated system 200 connected to the monitoring systems a to c and the node control systems a to c. The optimum network restoration system 100 includes an optimum design unit 101, an instruction creation unit 103, and a communication unit 104. The monitoring integration system 200 includes an information integration unit 201 and a communication unit 209, and includes configuration information 203, resource information 205, and failure information 207 as a database. Note that the optimum network recovery system 100 and the monitoring integrated system 200 may be constructed on the same server.

  The networks a to c include a plurality of nodes (not shown) and are configured such that each node can connect to the Internet. In FIG. 1, a monitoring system a monitors a failure of the network a, and the node control system a controls each node included in the network a. Similarly, the monitoring system b monitors a failure of the network b, and the node control system b controls each node included in the network b. Similarly, the monitoring system c monitors a failure in the network c, and the node control system c controls each node included in the network c.

  The information integration unit 201 of the monitoring integrated system 200 regenerates failure information, that is, commonizes and individualizes failure information. For example, in a network switch or the like, in general, the name of an interface or the like may differ depending on the manufacturer or the like. That is, if an attempt is made to create a recovery process by referring to failure information from a network device as it is, a recovery algorithm is required for each manufacturer, leading to an increase in the number of algorithms. However, for example, when "GigabitEthernet (registered trademark) 1/1" as a network device fails and when "GigabitEthernet (registered trademark) 1/0/1" fails, the process of restoring the failure is different. It is not considered. Therefore, the information integration unit 201 shares fault information from network devices. For example, it has a function to unify differences in host naming rules and device interface names as described below.

  For example, if a network device is "Shinjuku Switch 01 GigabitEthernet (registered trademark) 1/1", this is called "Shinjuku Switch 01 Interface1", and another network device is "Shinjuku_Switch 02 GigabitEthernet (registered trademark)". ) 1/0/1 ”, this is“ Shinjuku Switch 02 Interface1 ”. Note that these conversion processes can also be performed using a conversion table.

  The optimum design unit 101 in the optimum network restoration system 100 has a plurality of restoration algorithms. For example, in FIG. The optimum design unit 101 performs a design for recovery based on the failure information shared by the information integration unit 201. Based on the design, instructions are given to the node control systems a to c. Here, it is necessary to individualize the instructions so that each node control system ac can understand. Therefore, the optimal network restoration system 100 temporarily returns the setting information related to the instruction to the information integration unit 201, and the information integration unit 201 personalizes it. The optimum design unit 101 transmits the individualized setting information to the instruction creation unit 103, and the instruction creation unit 103 creates instructions for the node control systems a to c. FIG. 2 shows an example of sharing and individualization.

  The optimum design unit 101 refers to the failure information 207 and the configuration information 203 and creates a recovery plan for free resources for each service. As shown in FIG. 1, the optimum design unit 101 is provided with a plurality of algorithms for creating a recovery plan, and is basically prepared for each service SLA. In other words, service A may wish to recover by reducing the delay amount as much as possible between the functions realizing the service, and the physical distance between the functions is as much as possible considering fault tolerance. You may want a recovery that will grow. The optimum design unit 101 selects an algorithm for each service according to the flowchart shown in FIG.

  FIG. 3 is a flowchart showing the operation of the optimum design unit 101. First, when a failure notification is received (step S1), the affected service is identified with reference to the configuration information database (step S2). Next, the service with the highest priority among the services that have not been recovered is identified (step S3). Next, a location to be restored is identified (step S4), resource information is referred to, and a node that can be used from a CPU load or the like is selected (step S5). Next, a usable node configuration is patterned (step S6). In step S6, for example, the position of the node is determined as follows. That is, each of the three nodes “Fukuoka Node (VM)”, “Sapporo Node (VM)”, “Tokyo Node (VM)”, “Tokyo Node (VM)”, “Osaka Node (VM)” Associate each of the two nodes.

  Next, referring to the resource information, a link that can be used from the expected free bandwidth is selected (step S7), the configuration information database is referred to, and the algorithm to be executed is specified (step S8). If an algorithm is specified, the algorithm is executed and a combination candidate is narrowed down (step S9). Next, the narrowed candidates are arbitrarily extracted (step S10). Next, it is determined whether there are still affected services (step S11). If there are still affected services, the process proceeds to step S3. On the other hand, if there is no affected service in step S11, a recovery control instruction is created (step S12), and the process ends.

  In FIG. 1, the instruction creation unit 103 creates settings for recovery from the recovery plan. The instruction creation unit 103 has a driver for each of the node control systems a to c. Each node control system ac may exist for each network or each control device type. The instruction creation unit 103 has a function of knowing the role of each of the node control systems a to c and the control target node, and determining which node control system should be instructed to execute a given recovery plan. Have. In addition, an instruction setting is created through individualization of recovery plan information from the information integration unit 201 and each node control system driver.

  This instruction setting may be a setting for calling an API (Application Programming Interface) of the node control systems a to c, or a script describing a CLI (Command Line Interface).

FIG. 4 is a diagram illustrating an example of the configuration information database. As shown in FIG. 4, priority is defined for the service. When a plurality of service restorations are requested at the same time, restoration processing is executed in order of priority.
5 and 6 are diagrams showing examples of resource information. These pieces of resource information are used for resource search for performing recovery.

  Next, the operation of the network recovery system configured as described above will be described. 7 to 10 are diagrams illustrating an operation example of the network recovery system according to the present embodiment. In FIG. 7, this network recovery system includes an optimal network recovery system 100 and a monitoring integrated system 200 connected to the monitoring systems a to c and the node control systems a to c. The networks a to c include a plurality of nodes and are configured so that each node can connect to the Internet. In FIG. 7, the monitoring system a monitors a failure of the network a, and the node control system a controls each node included in the network a. Similarly, the monitoring system b monitors a failure of the network b, and the node control system b controls each node included in the network b. Similarly, the monitoring system c monitors a failure in the network c, and the node control system c controls each node included in the network c.

  The network recovery system according to the present embodiment includes an optimal recovery unit corresponding to a service and has the following characteristics. That is, the optimal network recovery system 100 has a plurality of recovery algorithms. An increase in algorithms can be prevented by a re-forming function for sharing or individualizing failure information. In addition, it is possible to give priority to each service and restore a plurality of services continuously.

  The network recovery system can allocate resources at the time of recovery. That is, it is possible to identify the recovery position by grasping the network topology information and the network resource information. Moreover, it is possible to realize optimal resource allocation based on the past resource usage status, and to realize stable operation after recovery.

  In addition, this network recovery system can instruct recovery control to an existing control system. That is, by grasping the role of the control system, it is possible to distribute control to an appropriate control system.

  As shown in FIG. 8, when the physical node accommodated by the firewall FWa1 fails, the monitoring system b detects the failure and transmits the detected information to the monitoring integrated system 200. As shown in FIG. 8, in the monitoring integrated system 200, the information integration unit reforms the failure information in a format that can be understood by the optimum network recovery system 100 (commonization). Then, the reformed (shared) fault information is transmitted to the optimum network restoration system 100. Here, the service a adopts the logical configuration shown in FIG.

  Next, as shown in FIG. 9, the optimum network recovery system 100 refers to the configuration information in the monitoring integrated system 200 and reconstructs the network configuration to be recovered. The network configuration to be restored adopts a configuration as shown in FIG. 10, for example. Next, the optimal network recovery system 100 selects the recovery algorithm a, for example, “the delay to the Internet is within 100 ms” in order to satisfy the SLA of the service a, and tries to recover the network. At this time, the CPU usage amount, traffic usage amount, and the like used by the virtual machine VMa1 are read from the monitoring integrated system 200, and a resource that has no problem even if the VM construction or the like is executed is searched. After determining the resources to be used, the node control systems a to c are instructed to recover. Each of the node control systems a to c that has received an instruction from the optimal network recovery system 100 actually controls the node and performs network recovery. After the node control, the node control systems a to c notify the optimum network restoration system 100 to that effect, and the optimum network restoration system 100 transmits the restored configuration to the monitoring integrated system 200 and updates the information.

  As described above, according to the present embodiment, it is possible to automatically perform service restoration triggered by the occurrence of a warning indicating the occurrence of a failure. In addition, the restoration process can be realized in an optimum form for each service. Furthermore, it is possible to realize a system that performs automatic recovery for each service without making a major change to the existing system configuration.

(Modification)
FIG. 11 is a diagram illustrating a modification of the present embodiment. In this modification, it is assumed that there are a plurality of services, for example, service a and service b are provided simultaneously. The service a has the logical configuration shown in FIG. 15, and the service b has the logical configuration shown in FIG. As shown in FIG. 11, when a node accommodating the load balancer LBb1 fails, not only the service b but also the service a is affected. Therefore, the network recovery system according to the present embodiment refers to the configuration information and performs recovery processing from a service with a high priority.

  In this way, when priority is set for each service and multiple services need to be restored at the same time, the priority and network setting information are output to the node control system to recover from the higher priority service. Processing can be executed.

  FIG. 13 is a diagram illustrating traffic volume with respect to time. The network recovery system according to the present embodiment periodically acquires traffic information and holds it for an arbitrary period. Then, a traffic trend in an arbitrary period is predicted using linear approximation as shown in FIG. 13, for example. By using the predicted traffic volume when creating a recovery plan, stable operation after recovery can be achieved.

100 Optimal Network Recovery System 101 Optimal Design Unit 103 Instruction Creation Unit 104 Communication Unit 200 Monitoring Integrated System 201 Information Integration Unit 203 Configuration Information 205 Resource Information 207 Fault Information 209 Communication Unit

Claims (6)

When a failure occurs in any of the nodes connected to a monitoring system that monitors a node failure that occurs in a plurality of interconnected networks that include a plurality of nodes and a node control system that controls each node , A network recovery system that gives an instruction to the node control system to reconfigure a network in units of services whose communication paths cross each network,
An information integration unit that acquires failure information indicating that a failure has occurred in any node from the monitoring system, and outputs the acquired failure information;
A recovery plan for reconstructing the logical configuration of the network before the occurrence of the failure of the network to be recovered based on the failure information and recovering the network failure based on the reconfigured information of the logical configuration of the network before the occurrence of the network An optimal design unit that checks the free resources of the network for each service and creates a recovery plan using any of a plurality of algorithms for creating
A network recovery system comprising: an instruction creation unit that creates network setting information for restoration using the restoration plan and outputs the created network setting information to the node control system. The information integration unit records network configuration information, resource information, and failure information,
The network recovery system according to claim 1, wherein the optimum design unit identifies a node constituting a recovery plan based on the network configuration information, resource information, and failure information. The information integration unit shares the acquired failure information different for each network and outputs common failure information, while the recovery plan created by the optimal design unit is individualized to suit each network, and the individualized recovery plan Is output to the instruction creation unit,
The network restoration system according to claim 1 or 2, wherein the instruction creating unit creates network setting information for restoration based on the individualized restoration plan.   4. The network recovery system according to claim 3, wherein the information integration unit includes an information conversion table, and executes the sharing of the failure information and the individualization of the recovery plan using the information conversion table.   The instruction creating unit determines a priority for each service, and outputs the priority and the network setting information to the node control system when a plurality of services need to be restored simultaneously. The network restoration system according to any one of claims 1 to 4. When a failure occurs in any of the nodes connected to a monitoring system that monitors a node failure that occurs in a plurality of interconnected networks that include a plurality of nodes and a node control system that controls each node , A program of a server device that gives an instruction to reconfigure a network in units of services whose communication path crosses each network, to the node control system,
Processing for acquiring failure information indicating that a failure has occurred in any node from the monitoring system, and outputting the acquired failure information;
A recovery plan for reconstructing the logical configuration of the network before the occurrence of the failure of the network to be recovered based on the failure information and recovering the network failure based on the reconfigured information of the logical configuration of the network before the occurrence of the network Using one of multiple algorithms to create a network, checking the free resources of the network for each service, creating a recovery plan,
Generating a network setting information for recovery using the recovery plan and outputting the generated network setting information to the node control system; and causing the computer to execute a series of processes. Program.

Images