JP6260324B2 - Network system and network management method - Google Patents

Description

  The present invention relates to a network system and a network management method applied to an OpenFlow network including a plurality of controllers.

  In recent years, OpenFlow has been developed that facilitates the construction of highly flexible networks using inexpensive switches. OpenFlow separates the packet transfer function and the path control function of the conventional network device into an OpenFlow switch (OFS) and an OpenFlow controller (OFC). The OFS has only a packet transfer function, and the OFC centrally controls packet transfer of a plurality of OFSs (hereinafter referred to as OFS groups).

  In OpenFlow, a series of communications determined by a combination of a MAC address, an IP address, a port number, and the like is defined as a “flow”, and path control in units of flows is realized. The OFC sets a flow entry in the OFS using a message defined by the open flow protocol. OFS processes the received packet according to the contents of the flow entry.

  One of the OFCs is a controller that provides a virtual network function. Such an OFC calculates a flow entry according to a request from a user so that the OFS group behaves as a virtual bridge or a virtual router. Then, the OFC sets the calculated flow entry in the OFS group. There is a limit to the number of flow entries that can be set in the OFS. Therefore, in order to save the number of flow entries, the OFC often sets only the flow entries related to the flow in which communication actually occurs in the OFS. Specifically, the OFC calculates a communication route based on the virtual network setting and the physical network topology information when the OFS receives a PacketIn message which is a kind of OpenFlow message. . Then, the OFC sets a dynamic flow entry in which an idle timeout value and a hard timeout value are set in the OFS.

  In recent years, with the increase in the number of OFC types and the diversification of demands for OpenFlow networks, integrated controllers for managing a plurality of heterogeneous OFCs have been developed. Even in an OFC that provides a similar virtual network, an API (Application Programming Interface) provided to a user is different for each type of OFC. The integrated controller conceals API differences between OFCs and provides a common API to users. By introducing such an integrated controller, the user can easily control the heterogeneous OFC. In addition, it is possible to easily shift to a different type of OFC. “Migration to a different OFC” means that an OFS managed by a certain OFC is transferred to the management of another type of OFC.

  In Patent Document 1, in an OpenFlow network including a plurality of OFCs, OFS broadcasts an inquiry requesting OFC information to the network, and based on the return information for the inquiry, the inquiry destination of the flow entry A technique for selecting an OFC is described.

JP 2013-143698 A

  Here, consider a case where the integrated controller performs a transition to a different OFC in an environment where a plurality of different OFCs are managed. Hereinafter, the OFC that is the migration source is referred to as the migration source OFC. The OFC that is the migration destination is referred to as the migration destination OFC.

  The OFC sets a flow entry in the OFS according to the request from the user in accordance with the state of the physical network. There are many types of flow entries that satisfy user requests. In addition, the flow entries set in the OFS for each type of OFC are greatly different. For this reason, it is difficult for an OFC of a type different from the OFC to interpret the intention of the content of the flow entry set by a certain OFC.

  Therefore, when an OFS controlled by a certain OFC is transferred under the control of another heterogeneous OFC, the transfer destination OFC cannot maintain and manage the flow entry set in the OFS by the transfer source OFC. Therefore, the transfer destination OFC deletes the flow entry set in the OFS immediately after the transfer. Thereafter, the migration destination OFC resets the flow entry to OFS.

  In this way, immediately after the migration, there are few flow entries set in the OFS. Therefore, all the packets transferred according to the flow entry set in the OFS before the transfer are transferred via OFC immediately after the transfer because the flow entry corresponding to the packet is not set in the OFS. . Therefore, immediately after the migration, the load of the migration destination OFC increases, and the network throughput of the OpenFlow network temporarily decreases.

  Therefore, the present invention provides a network system and a network management method capable of reducing a decrease in network throughput when a switch group controlled by a controller is transferred to another controller in an OpenFlow network. For the purpose.

  A network system according to the present invention includes a plurality of control devices that control a switch that transfers a packet based on a flow entry, and an integrated device that manages the plurality of control devices. The integrated device includes a switch that is controlled by the control device. When migrating to a control device other than the control device, the physical network topology information, virtual network setting information and flow information acquired from the migration source control device are transferred to the migration destination control device. Each control device includes a transfer processing unit to be transferred, and when the own device becomes the transfer destination, sets each switch to a switch arranged under the management of the own device based on each information received from the integrated device It includes a transition preparation unit that prepares a transition for calculating a flow entry.

  A network management method according to the present invention is a network management method in a network system including a plurality of control devices that control a switch that forwards a packet based on a flow entry, and an integrated device that manages the plurality of control devices. When a device shifts a switch controlled by the control device to the management of a control device different from the control device, the physical network topology information, the virtual network setting information, and the flow acquired from the transfer source control device Information is transferred to the destination control unit, and the destination control unit calculates the flow entry to be set on the switch placed under the management of the own unit after migration based on the information received from the integrated unit It is characterized by preparing for the transition.

  ADVANTAGE OF THE INVENTION According to this invention, when the switch group which a controller controls is transferred to the management of another controller in an open flow network, the fall of the throughput of a network can be reduced.

It is a block diagram which shows the structure of 1st Embodiment of the network system by this invention. It is a block diagram which shows the structure of transfer origin OFC. It is a block diagram which shows the structure of 1st Embodiment of transfer destination OFC. It is a block diagram which shows the structure of 1st Embodiment of an integrated controller. It is explanatory drawing which shows an example of a dynamic packet generated from dynamic flow information and dynamic flow information. It is explanatory drawing which shows other examples of dynamic flow information and the pseudo packet produced | generated from dynamic flow information. It is a flowchart which shows operation | movement of 1st Embodiment of a network system. It is explanatory drawing which shows an example of virtual network setting information. It is a block diagram which shows the structure of 2nd Embodiment of transfer destination OFC. It is a flowchart which shows operation | movement of 2nd Embodiment of a network system. It is a block diagram which shows the minimum structure of the network system by this invention.

Embodiment 1. FIG.
A first embodiment of the present invention will be described below with reference to the drawings.

  In this embodiment, a network system in which an integrated controller manages an OFC that provides a virtual network function is taken as an example. Although the setting interface of the virtual network differs depending on the type of OFC, in this embodiment, the driver unit of the integrated controller absorbs the difference and provides a common interface to the user.

  In the present embodiment, it is not necessary to mount a special function (for example, a function realized by the migration state management unit 240 and the pseudo packet generation unit 250 described later) in the migration source OFC. The migration source OFC uses only functions generally provided by many OFCs.

  FIG. 1 is a block diagram showing a configuration of a first embodiment of a network system according to the present invention.

  As illustrated in FIG. 1, the network system according to the present embodiment includes an integrated controller 10 and an OFC 200 that is a migration destination OFC. In addition to the OFC 200, the integrated controller 10 is connected to an OFC 100 serving as a migration source OFC. Note that the network system may include any number of migration destination OFCs. Further, any number of migration source OFCs may be connected to the integrated controller 10.

  In this embodiment, the integrated controller 10 migrates the OFS group controlled by the migration source OFC (OFC 100) under the management of the migration destination OFC (OFC 200).

  The OFC 100 is connected to an OFS group 1000 including a plurality of OFSs via a control line.

  The OFC 200 is a different type of OFC from the OFC 100. The OFC 200 can accept the same virtual network settings as the OFC 100. However, the OFC 200 and the OFC 100 calculate different flow entries for the same virtual network setting and set them in the OFS group 1000.

  The integrated controller 10 is connected to each OFC via a control line.

  Each OFS constituting the OFS group 1000 exchanges messages according to the OpenFlow protocol with the OFC 100. Each OFS performs packet rewriting, transfer, and the like in accordance with the flow entry set from the OFC. After the migration processing, the OFC 200 controls the OFS group 1000 instead of the OFC 100 and sets a flow entry.

  FIG. 2 is a block diagram showing the configuration of the migration source OFC (OFC 100). As illustrated in FIG. 2, the OFC 100 includes an OF (OpenFlow) message processing unit 110, a physical network management unit 120, and a virtual network management unit 130.

  The OF message processing unit 110 exchanges messages with the OFS according to the OpenFlow protocol.

  The physical network management unit 120 holds physical network topology information (hereinafter referred to as physical network information).

  The virtual network management unit 130 holds virtual network setting information (hereinafter referred to as virtual network setting information).

  The OFC 100 controls the OFS group 1000 according to the virtual network setting notified from the integrated controller 10. Specifically, the OFC 100 receives virtual network setting information from the integrated controller 10 and stores it in the virtual network management unit 130. The OFC 100 calculates the transfer destination physical switch port of the target packet according to the virtual network setting information when the OF message processing unit 110 receives the PacketIn message from the OFS. Then, the OFC 100 sets a flow entry for the OFS on the path between the transfer source and transfer destination of the packet according to the physical network information held by the physical network management unit 120.

  FIG. 3 is a block diagram showing the configuration of the first embodiment of the migration destination OFC. As shown in FIG. 3, the migration destination OFC (OFC 200) includes an OF message processing unit 210, a physical network management unit 220, a virtual network management unit 230, a migration state management unit 240, and a pseudo packet generation unit 250. Including.

  The OF message processing unit 210, the physical network management unit 220, and the virtual network management unit 230 are the same as the OF message processing unit 110, the physical network management unit 120, and the virtual network management unit 130 of the OFC 100 illustrated in FIG.

  The migration state management unit 240 manages information necessary for calculating a flow entry to be set in the OFS after migration.

  The pseudo packet generation unit 250 generates a pseudo packet including a PacketIn message that the migration source OFC will receive from the OFS group 1000.

  The OF message processing unit 210, the physical network management unit 220, the virtual network management unit 230, the migration state management unit 240, and the pseudo packet generation unit 250 are realized by a CPU of a computer that operates according to a program, for example. The program is stored in a storage device (not shown) of the OFC 200, for example. The CPU of the OFC 200 reads the program, and operates as an OF message processing unit 210, a physical network management unit 220, a virtual network management unit 230, a migration state management unit 240, and a pseudo packet generation unit 250 according to the program. In addition, the OF message processing unit 210, the physical network management unit 220, the virtual network management unit 230, the migration state management unit 240, and the pseudo packet generation unit 250 may be realized by separate hardware.

  FIG. 4 is a block diagram illustrating a configuration of the integrated controller 10 according to the first embodiment. As shown in FIG. 4, the integrated controller 10 includes a setting reception unit 11, a migration processing unit 12, and a driver unit 13 (driver units 13-1 and 13-2). The driver unit 13-1 is a driver installed for the OFC 100. The driver unit 13-2 is a driver installed for the OFC 200.

  The setting reception unit 11 receives a setting change request from the user. As an interface for accepting a setting change request, a REST (Representational State Transfer) API or the like can be considered. The request from the user includes the controller identifier. Therefore, the setting reception unit 11 can determine the type of the controller according to the identifier and pass the request to the appropriate driver unit 13.

  The setting reception unit 11 receives a migration process start request from the user.

  When the migration processing unit 12 receives a migration processing start request from the user via the setting reception unit 11, the migration processing unit 12 starts the migration processing. During the migration process, the migration processing unit 12 acquires information necessary for the migration process from the migration source OFC 100 and passes it to the migration destination OFC 200.

  The driver unit 13 converts the request received from the setting reception unit 11 into a format that can be processed by the OFC as a control target, and sends the converted request to the OFC. The driver unit 13 is installed in the integrated controller 10 for each type of OFC to be controlled.

  The setting reception unit 11, the migration processing unit 12, and the driver unit 13 are realized by a CPU of a computer that operates according to a program, for example. The program is stored in a storage device (not shown) of the integrated controller 10, for example. The CPU of the integrated controller 10 reads the program and operates as the setting reception unit 11, the migration processing unit 12, and the driver unit 13 according to the program. Moreover, the setting reception part 11, the transfer process part 12, and the driver part 13 may be implement | achieved by separate hardware.

  Next, the operation of this embodiment will be described.

  First, pseudo packet generation processing in the pseudo packet generation unit 250 of the OFC 200 will be described.

  FIG. 5 is an explanatory diagram showing an example of the dynamic flow information and a pseudo packet generated from the dynamic flow information. FIG. 6 is an explanatory diagram showing another example of the dynamic flow information and a pseudo packet generated from the dynamic flow information. The dynamic flow information is flow information that is dynamically generated by the OFC. The dynamic flow information includes information indicating a dynamic flow. A dynamic flow refers to communication between terminals using an open flow network. As shown in FIGS. 5 (A) and 6 (A), the flow information is composed of a flow entrance port, a match condition, and an action.

  The pseudo packet generation unit 250 of the migration destination OFC (OFC 200) generates a packet including the protocol header of the highest layer in each flow information as a pseudo packet. For example, for the flow information shown in FIG. 5 (A) that includes only the source MAC address in the match condition, a pseudo packet including an Ethernet (registered trademark) header is generated as shown in FIG. 5 (B). . Further, for example, for the flow information shown in FIG. 6A including the destination IP address in the match condition, as shown in FIG. 6B, a pseudo packet including the Ethernet header and the IP header is generated.

  For values that cannot be reproduced from the flow information, the pseudo packet generator 250 randomly generates values and sets them in the pseudo packet. The transmission source address “192.168.50.1” in the IP header of the pseudo packet illustrated in FIG. 6B is an example of a value randomly generated by the pseudo packet generation unit 250. Since the pseudo packet is not transmitted outside the OFC 200, there is no problem even if an arbitrary value is set in this way. Even if the OFC 200 generates a flow entry that is not originally used for communication from the pseudo packet, there is no major problem other than unnecessarily increasing the number of entries in the OFS flow table.

  Next, the processing of the network system at the time of shifting to the different OFC in this embodiment will be described. FIG. 7 is a flowchart showing the operation of the first embodiment of the network system.

When the migration processing unit 12 of the integrated controller 10 receives the migration processing start request via the setting reception unit 11, the migration processing unit 12 acquires the following information from the virtual network management unit 130 of the OFC 100 via the driver unit 13-1. (Step S101).
・ Virtual network setting information ・ Dynamic flow information

  This dynamic flow information is flow information dynamically created by the OFC 100 based on settings indicated by physical network information, virtual network setting information, and the like held by the OFC 100.

  FIG. 8 is an explanatory diagram of an example of virtual network setting information. In the example illustrated in FIG. 8, it is assumed that the OFC 100 and the OFC 200 provide a function capable of assigning each terminal connected to the OpenFlow network to the virtually created bridge using the MAC address as a key.

  The migration processing unit 12 of the integrated controller 10 acquires physical network information from the physical network management unit 120 of the OFC 100 via the driver unit 13-1 (Step S102).

  The migration processing unit 12 transmits the information acquired in the processes of steps S101 and S102 to the migration state management unit 240 of the OFC 200 via the driver unit 13-2 (step S103). Thereby, the OFC 200 can recognize that the own device has become the migration destination. Then, the OFC 200 starts migration preparation (the processing in steps S104 to S108).

  The migration state management unit 240 of the OFC 200 sets the virtual network setting information acquired from the integrated controller 10 in the virtual network management unit 230 (step S104).

  The migration state management unit 240 sets the physical network information acquired from the integrated controller 10 in the physical network management unit 220 (step S105). The physical network management unit 220 operates based on the physical network information set from the migration state management unit 240 until the process of step S112 is completed.

  The transition state management unit 240 inputs the dynamic flow information acquired from the integrated controller 10 to the pseudo packet generation unit 250. Then, the pseudo packet generation unit 250 generates a pseudo packet based on the input dynamic flow information (step S106).

  The migration state management unit 240 inputs the pseudo packet generated by the pseudo packet generation unit 250 to the virtual network management unit 230 (step S107).

  The virtual network management unit 230 calculates a flow entry to be set in the OFS after migration based on the pseudo packet, the virtual network setting information, and the physical network information. The virtual network management unit 230 calculates a flow entry by the same method as when OFS is connected, that is, by the same method as when a PacketIn message is received from the OFS (step S108).

  After the virtual network management unit 230 calculates the flow entries corresponding to all dynamic flows, the migration state management unit 240 notifies the integrated controller 10 of completion of migration preparation (step S109).

  The migration processing unit 12 of the integrated controller 10 requests the OFC 100 to disconnect from the OFS group 1000 (step S110). At this time, it is necessary to set the OFS group 1000 so that the OFS group 1000 that is disconnected from the OFC 100 is connected to the OFC 200 immediately after the disconnection. For example, the user or the like sets the IP address of the OFC 200 in advance in each OFS of the OFS group 1000 as the OFC having the next highest connection priority after the OFC 100.

  When the OFS group 1000 is connected to its own apparatus, the physical network management unit 220 of the OFC 200 deletes all flow entries set in each OFS of the OFS group 1000 (step S111).

  The physical network management unit 220 sets the flow entry calculated by the virtual network management unit 230 in step S108 in each OFS (step S112).

  Thereafter, the physical network management unit 220 updates the physical network information held by itself according to the actual physical network change.

  As described above, in this embodiment, the integrated controller acquires physical network information, virtual network setting information, and dynamic flow information from the migration source OFC. Then, the integrated controller passes the information to the migration destination OFC. Then, based on the information received from the integrated controller, the migration destination OFC generates a PacketIn message that the migration source OFC would have received from the OFS group, and calculates a flow entry to be set in the OFS. Then, the transfer destination OFC sets this flow entry in the OFS immediately after the transfer. With such a configuration, when migrating to a different OFC, the migration destination OFC can maintain and manage the flow entry set in the OFS by the migration source OFC. That is, even immediately after the transition to the different OFC, the flow entry can be set in the OFS. Therefore, it is possible to reduce a decrease in network throughput that occurs during the transition to a different type of OFC.

  In the present embodiment, the case has been described in which only the migration destination OFC (OFC 200) includes the migration state management unit and the pseudo packet generation unit. However, the migration source OFC includes the migration state management unit and the pseudo packet generation unit. May be.

Embodiment 2. FIG.
Hereinafter, a second embodiment of the present invention will be described with reference to the drawings.

  FIG. 9 is a block diagram illustrating the configuration of the second embodiment of the migration destination OFC.

  As shown in FIG. 9, in addition to the configuration of the first embodiment, the migration destination OFC (OFC 300) in the second embodiment includes a migration processing physical network management unit 310 that operates only during migration processing, and a migration And a processing virtual network management unit 320.

  With the configuration illustrated in FIG. 9, the migration destination OFC can accept an OFS controlled by another OFC even in an operating state in which the OFS is already controlled. Note that the migration destination OFC needs to acquire the virtual network and physical network information of the migration source OFC and integrate it with existing information held by itself. The migration destination OFC calculates a flow entry to be set in the OFS group after migration based on the integrated information. In order for the migration destination OFC to integrate physical networks, the migration destination OFC needs to acquire physical network boundary information held by the integrated controller 10.

  Here, the boundary information of the physical network is information indicating the boundary of the physical network, and for example, two ports (for example, “port 30 of the switch 1” and “switch 2” at the position where the two physical networks are in contact with each other). For example, “port 40” ”. In this embodiment, physical network boundary information is stored in advance in the migration processing physical network management unit 310 by a user or the like. Note that the migration processing unit 12 of the integrated controller 10 may generate physical network boundary information based on the physical network information acquired from the physical network management units of the OFCs 100 and 200.

  The migration processing physical network management unit 310 and the migration processing virtual network management unit 320 are realized by a CPU of a computer that operates according to a program, for example. The program is stored in a storage device (not shown) of the OFC 300, for example. The CPU of the OFC 300 reads the program and operates as the migration processing physical network management unit 310 and the migration processing virtual network management unit 320 according to the program. Further, the migration processing physical network management unit 310 and the migration processing virtual network management unit 320 may be realized by separate hardware.

  The processing of the network system at the time of transition to the heterogeneous OFC in this embodiment will be described. FIG. 10 is a flowchart showing the operation of the second embodiment of the network system.

The migration processing unit 12 of the integrated controller 10 acquires the following information from the virtual network management unit 130 of the OFC 100 via the driver unit 13-1 (Step S201).
・ Virtual network setting information ・ Dynamic flow information

  The migration processing unit 12 acquires physical network information from the physical network management unit 120 of the OFC 100 via the driver unit 13-1 (Step S202).

  The migration processing unit 12 transmits the information acquired in the processes of steps S201 and S202 to the migration state management unit 240 of the OFC 300 via the driver unit 13-2 (step S203).

  The migration state management unit 240 of the OFC 300 integrates the virtual network setting information acquired from the integrated controller 10 and the virtual network setting information held by the virtual network management unit 230, and sets them in the migration processing virtual network management unit 320 ( Step S204).

  The migration state management unit 240 integrates the physical network information acquired from the integrated controller 10 and the physical network information held by the physical network management unit 220 based on the physical network boundary information, and performs physical network management for migration processing. Set in the unit 310 (step S205). Thereafter, the physical network management unit 220 notifies the migration processing physical network management unit 310 of the network-related event (PacketIn message or the like) received by the OF message processing unit 210. As a result, the generation or change of the flow entry based on the event that occurred during the migration process is reflected in the migration processing physical network management unit 310.

  The transition state management unit 240 inputs the dynamic flow information acquired from the integrated controller 10 to the pseudo packet generation unit 250. Then, the pseudo packet generation unit 250 generates a pseudo packet based on the input dynamic flow information (step S206).

  The migration state management unit 240 inputs the pseudo packet generated in step S206 to the migration processing physical network management unit 310 (step S207).

  The migration processing physical network management unit 310 calculates a flow entry to be set in the OFS after migration based on the pseudo packet, the virtual network setting information acquired from the migration processing virtual network management unit 320, and the physical network information. (Step S208). At this time, when receiving the notification from the physical network management unit 220, the migration processing physical network management unit 310 updates the flow entry based on the event corresponding to the notification.

  After the migration processing physical network management unit 310 calculates flow entries corresponding to all dynamic flows, the migration state management unit 240 notifies the integrated controller 10 of completion of migration preparation (step S209).

  The migration processing unit 12 of the integrated controller 10 requests the OFC 100 to disconnect from the OFS group 1000 (step S210).

  When the OFS group 1000 is connected to its own apparatus, the physical network management unit 220 of the OFC 300 deletes all the flow entries set in each OFS of the OFS group 1000 (step S211).

  The migration state management unit 240 copies the information held by the migration processing virtual network management unit 320 to the virtual network management unit 230 (step S212).

  The physical network management unit 220 sets the difference between the flow entry calculated in the operating state before the migration process and the flow entry held by the migration processing physical network management unit 310 in the OFS group 1000 (step S213).

  The migration state management unit 240 copies the information held by the migration processing physical network management unit 310 to the physical network management unit 220 (step S214).

  Thereafter, the physical network management unit 220 updates the physical network information held by itself according to the actual physical network change. In addition, the physical network management unit 220 stops the event notification to the migration processing physical network management unit 310.

  As described above, in the present embodiment, the migration destination OFC (OFC 300) integrates information (physical network information and virtual network setting information) held by the own apparatus and information acquired from the integrated controller 10. Then, the migration destination OFC calculates the flow entry based on the integrated information. As a result, even if the migration destination OFC is in a state where the own device already holds physical network information or virtual network setting information, for example, in an operating state in which OFS has already been controlled, Can accept the OFS.

  The network system according to the present invention is applied to, for example, a data center.

  Next, the outline of the present invention will be described. FIG. 11 is a block diagram showing the minimum configuration of the network system according to the present invention. As shown in FIG. 11, the network system according to the present invention includes a plurality of control devices 30-1 to 30-30 that control switches (corresponding to each OFS in the OFS group 1000 shown in FIG. 1) that transfer packets based on flow entries. -N and an integrated device 20 (corresponding to the integrated controller 10 shown in FIG. 1) that manages the plurality of control devices 30-1 to 30-n. The integrated device 20 includes a control device (for example, shown in FIG. 1). When the switch controlled by the OFC 100 is controlled under the control of a control device different from the control device (e.g., equivalent to the OFC 200 shown in FIG. 1), the physical network The migration processing unit 21 (in the integrated controller 10 shown in FIG. 4) transfers the topology information, the virtual network setting information, and the flow information to the migration destination control device. Each control device is arranged under the management of its own device after the migration based on the information received from the integrated device 20 when the own device becomes the migration destination. The migration preparation unit 31 (in the OFC 200 shown in FIG. 3) that performs the migration preparation (corresponding to the processing of steps S104 to S108 shown in FIG. 7 or the processing of steps S204 to S208 shown in FIG. 10) for calculating the flow entry to be set in the switch. 9 corresponds to the migration state management unit 240 and the pseudo packet generation unit 250, or the migration state management unit 240, the pseudo packet generation unit 250, the migration processing physical network management unit 310, and the migration processing virtual network management unit 320 in the OFC 300 illustrated in FIG. )including.

  According to such a configuration, when the switch controlled by the control device is transferred under the control of a control device different from the control device, the type of the transfer source control device and the transfer destination control device are different. Even in such a case, the transfer destination control device can maintain and manage the flow entry set in the switch by the transfer source control device. Therefore, even immediately after the transition, the flow entry can be set in the switch. Therefore, it is possible to reduce a decrease in the throughput of the network that occurs when the switch controlled by the control device is transferred to the control of a control device different from the control device.

  When the migration processing unit 21 receives a notification of completion of migration preparation from the migration destination control device, the migration processing unit 21 disconnects the switch to be migrated from the migration source control device, and the migration preparation unit 31 of the migration destination control device When a switch to be migrated is connected to the own device, the flow entry registered in the switch may be deleted, and then the flow entry calculated in the migration preparation may be registered in the switch. According to such a configuration, it is possible to set a flow entry corresponding to the transfer destination control device in the switch to be transferred. Therefore, even if the type of the transfer source control device and the transfer destination control device are different, the transfer destination control device can reliably control the switch transferred under the management of the own device. Become.

  Also, the migration preparation unit 31 of the migration destination control device generates a pseudo packet including a message that the migration source control device would receive from the switch from the flow information received from the integration device 20, and sets the virtual network The flow entry may be calculated based on the information, the physical network topology information, and the pseudo packet. According to such a configuration, for example, by generating a pseudo packet including a PacketIn message that the migration source control device would have received, in the same manner as when the migration source control device was connected to the switch, Flow entries can be calculated.

  In addition, the migration preparation unit 31 of the migration destination control device integrates the virtual network setting information held by the own device and the virtual network setting information received from the integration device 20, and further indicates the boundary of the physical network. Based on the information, the topology information of the physical network held by the own device and the topology information of the physical network received from the integration device 20 are integrated, and the configuration information of the virtual network integrated with the topology information of the integrated physical network, The flow entry may be calculated based on the pseudo packet. According to such a configuration, even if the transfer destination control device is in an operating state in which the switch is already controlled, the switch controlled by the transfer source control device can be accepted.

  Further, when the migration preparation unit 31 of the migration destination control device receives a message from the switch during the migration preparation, the migration preparation unit 31 may update the flow entry set in the switch based on the message. According to such a configuration, when an event such as a PacketIn message is received from the switch during preparation for migration when the migration destination control device is in an operating state, generation or modification of a flow entry based on the event is performed. It can be reflected in the flow entry set in the switch.

DESCRIPTION OF SYMBOLS 10 Integrated controller 11 Setting reception part 12, 21 Transition processing part 13-1, 13-2 Driver part 20 Integrated apparatus 30-1-30-n Control apparatus 31 Transition preparation part 100, 200, 300 OFC
110, 210 OF message processing unit 120, 220 Physical network management unit 130, 230 Virtual network management unit 240 Migration state management unit 250 Pseudo packet generation unit 310 Migration processing physical network management unit 320 Migration processing virtual network management unit

Claims (7)

A plurality of control devices that control switches that forward packets based on flow entries;
An integrated device for managing the plurality of control devices,
The integrated device is
When migrating a switch controlled by a control device to the control of a control device different from the control device, the physical network topology information, virtual network setting information, and flow information acquired from the migration source control device , Including a transfer processing unit to be transferred to the transfer destination control device,
Each control device
Migration preparation for performing migration preparation for calculating a flow entry to be set in a switch arranged under the management of the own device after migration based on each information received from the integrated device when the own device becomes a migration destination A network system characterized by including a part. When the migration processing unit receives a notification of completion of migration preparation from the migration destination control device, it disconnects the switch to be migrated from the migration source control device,
When the switch is connected to its own device, the migration preparation unit of the migration destination control device deletes the flow entry registered in the switch and then registers the flow entry calculated in the migration preparation in the switch. The network system according to claim 1. The migration preparation unit of the migration destination control device generates a pseudo packet including a message that the migration source control device would receive from the switch from the flow information received from the integration device, and sets the virtual network setting information and the physical network The network system according to claim 1, wherein a flow entry is calculated based on the topology information of and the pseudo packet. The migration preparation unit of the migration destination control device integrates the virtual network setting information held by itself and the virtual network setting information received from the integration device, and further, based on information indicating the boundary of the physical network. In addition, the topology information of the physical network held by the own device and the topology information of the physical network received from the integrated device are integrated, and the configuration information of the virtual network and the pseudo packet integrated with the integrated topology information of the physical network, The network system according to claim 3, wherein the flow entry is calculated based on: The migration preparation unit of the transfer destination control device updates the flow entry calculated in the migration preparation based on the message when receiving a message from the switch managed by the own device during the migration preparation. The network system described. A network management method in a network system comprising: a plurality of control devices that control a switch that forwards a packet based on a flow entry; and an integrated device that manages the plurality of control devices,
When the integrated device shifts the switch controlled by the control device to the control of a control device different from the control device, the topology information of the physical network and the virtual network setting information acquired from the control device of the transfer source And flow information to the destination control device,
The migration destination control device performs migration preparation for calculating a flow entry to be set in a switch arranged under the management of the own device after migration based on each information received from the integrated device. Network management method. When the integrated device receives a notification of completion of migration preparation from the migration destination control device, the switch to be migrated is disconnected from the migration source control device,
7. When the switch is connected to its own device, the control device of the migration destination deletes the flow entry registered in the switch and then registers the flow entry calculated in the migration preparation in the switch. The network management method described.

Images