KR101527373B1 - Divided hierarchical network system based on software defined networks - Google Patents

Abstract

The present invention relates to a divided hierarchical network system based on a software defined network (SDN) capable of hierarchically operating a network by dividing the network in order to improve expandability and flexibility in the network based on SDN. According to the present invention, the divided hierarchical network system based on SDN comprises an edge controller performing the steps of: creating forwarding information on an inquiry for a flow forwarding of a lower layer, and responding the created forwarding information; creating mapping information to allow a plurality of edge ports of a plurality of switches constituting the lower layer to correspond to a plurality of virtual ports of a virtual switch, respectively; and inquiring the forwarding information on a flow converted into the virtual port in accordance with the edge port based on the mapping information, to an upper layer, when the forwarding information on the inquiry for the flow forwarding having the edge port received from the lower layer can not be created.

Description

[0002] DIVIDED HIERARCHICAL NETWORK SYSTEM BASED ON SOFTWARE DEFINED NETWORKS [0003]

The present invention relates to an SDN-based partitioned hierarchical network system capable of hierarchically operating a network in order to improve scalability and flexibility in a SDN (Software Defined Network) based network.

The network environment evolves around hardware, and it can not actively and flexibly cope with network changes and is not easy to expand. Vendors or other models of the same vendor may not be compatible with each other. Software defined network (SDN) technology and openflow protocol, which separates data plane and control play, are emerging as one of the alternative technologies, and there are a lot of cases applied as usecase now.

As the network environment becomes more complex, more security, automation, and scalability are required, and SDN-based networks are architectures that can meet these requirements. However, there is a limit to the number of switches that the controller of the control plane can handle. These limitations can reduce the scalability and flexibility of the network environment.

1. OpenFlow Switch Specification version 1.4.0 (Wire Protocol 0x05), October 14, 2013 [https://www.opennetworking.org/images/stories/downloads/sdn-resources/onf-specifications/openflow/openflow-spec- v1.4.0.pdf] 2. Software-Defined Networking: The New Norm for Netwrks, ONF White Paper, April 13, 2012 [https://www.opennetworking.org/images/stories/downloads/sdn-resources/white-papers/wp-sdn- newnorm.pdf]

Accordingly, it is an object of the present invention to provide an SDN-based partitioned network system capable of dividing an entire network and hierarchically managing divided networks.

In addition, by specifying partitioning requirements, it reduces errors due to network partitioning. Almost all controllers manage only the topology of the directly connected subnetwork, thereby reducing the load on the controller.

The SDN-based partitioned hierarchical network system according to the present invention generates and responds to forwarding information for a lower layer flow forwarding query, and transmits each of a plurality of edge ports of a plurality of switches constituting the lower layer as one virtual The mapping information is generated so as to correspond to a plurality of virtual ports of the switch, and when forwarding information for a flow forwarding query having an edge port received from a lower layer can not be generated, And an edge controller for inquiring the upper layer about the forwarding information about the flow converted into the virtual port corresponding to the port.

An SDN-based network system according to another embodiment of the present invention includes an orchestrator, a root controller, and a plurality of switches, wherein the orchestrator includes one switch or a predetermined number of switches connected to each other And an edge controller generator for generating an edge controller for controlling the edge network, wherein the first edge controller generated by the edge controller generator comprises: A topology management unit for managing a topology of a first edge network, which is a lower layer of the controller, an entry manager for generating and responding to forwarding information for a lower layer flow forwarding inquiry, and a plurality of edge ports Mapping each virtual port to a plurality of virtual ports of one virtual switch A mapping unit for generating information; When the forwarding information for the flow forwarding query having the edge port received from the lower layer is not generated, forwarding information for the flow converted into the virtual port corresponding to the edge port based on the mapping information, And a conversion unit inquiring into the layer.

According to the present invention, when satisfying the segmentation start requirement and the segmenting requirements, the entire network can be divided and hierarchically managed. Since the controller of each layer can manage only the network of the immediately lower level, The load can be reduced and thus the security can be improved.

In addition, an equivalent topology corresponding to the virtualized topology can be created and managed to provide an optimal path for path computation.

1 is a block diagram of a network system based on SDN (Software Defined Network) according to an embodiment of the present invention;
FIG. 2 is a block diagram of an abstracted network system of the network system of FIG. 1;
FIG. 3 is a block diagram of an orchestrator of the network system of FIG. 1;
FIG. 4 is a block diagram of a root controller of the network system of FIG. 1;
FIG. 5 is a block diagram of the edge controller of FIG. 1;
FIG. 6 is a flowchart of an SDN network virtualization method of an orchestrator according to an embodiment of the present invention;
Figure 7 is a flow diagram of an SDN network virtualization method of an edge controller in accordance with an embodiment of the present invention;
FIG. 8 is a flowchart of a path request response method of an edge controller according to an embodiment of the present invention;
Figure 9 is a flow diagram of a method for virtualizing an SDN network of a root controller according to an embodiment of the present invention;
10 is a block diagram illustrating an example of a network topology, and
11 is a block diagram showing another example of the network topology.

Hereinafter, the present invention will be described in more detail with reference to the drawings.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between. Also, the fact that the first component and the second component on the network are connected or connected means that data can be exchanged between the first component and the second component by wire or wirelessly.

In addition, suffixes "module" and " part "for the components used in the following description are given merely for convenience of description, and do not give special significance or role in themselves. Accordingly, the terms "module" and "part" may be used interchangeably.

When such components are implemented in practical applications, two or more components may be combined into one component, or one component may be divided into two or more components as necessary.

FIG. 1 is a block diagram of a network system based on SDN (Software Defined Network) according to an embodiment of the present invention. FIG. 2 is a block diagram of an abstracted network system of the network system of FIG. Fig. 4 is a block diagram of the root controller of the network system of Fig. 1, and Fig. 5 is a block diagram of the edge controller of Fig.

1, the network system according to the present embodiment includes an orchestrator 100, a root contoller 200, an edge controller 300, and a switch 400 can do.

Referring to FIGS. 1 and 2, the edge controller 300 of the network system according to the present embodiment acts as a general controller for a switch of a lower layer, and switches to a higher layer controller (a root controller 200) I can act.

The switch 400 may be a physical switch or a virtual switch. The switch 400 processes the received packet and can relay the flow between network devices (not shown) connected to the switch. The switch 400 may have a flow table. The switch 400 may include a multiple flow table for pipeline processing of the open flow specification.

The flow table includes a flow entry that defines rules for how to handle the flow of a network device (not shown). A network device may include a terminal device to which data or information is exchanged, or a physical device or a virtual device that performs a specific function.

A flow may refer to a packet flow of a particular path according to a series of packets sharing a value of at least one header field from a single switch or a combination of multiple flow entries of multiple switches. The open-flow network can perform path control, fault recovery, load balancing and optimization on a flow-by-flow basis.

The switch 400 can be divided into a core switch between an ingress and egress edge switch and an edge switch between a flow according to the combination of multiple switches.

The flow entry of the flow table includes a match field describing a condition (matching rule) to match a packet, a counter to be updated if there is a matching packet, a timeout describing the time to be discarded in the switch, instructions such as an opaque type of cookie, a change of a packet described in a flow entry, an update of an action set, and a metadata update, And may include a tuple such as an action, a priority, or the like.

The switch 400 can extract the flow information from the received packet as user traffic. The flow information may include an ingress port, a packet inflow port of a packet's edge switch, packet header information (source and destination IP address, MAC address, port, and VLAN information, etc.), metadata, . The metadata may be optionally added in the previous table or may be data added in another switch. The switch 400 can search the table for a flow entry for a received packet by referring to the extracted flow information. When the flow entry is found, the switch 400 can process and manipulate the received packet according to the processing rule (action or instruction) of the retrieved flow entry. If the flow entry search fails, the switch 400 transmits the minimum data of the received packet or the received packet to the controller of the upper layer, inquires about the flow processing, and receives the flow entry from the controller of the upper layer.

As the network environment becomes more complex, more security, automation, and scalability are required, and SDN-based networks are architectures that can meet these requirements. However, there is a limit to the number of switches that the controller of the control plane can handle. These limitations can reduce the scalability and flexibility of the network environment. Accordingly, the present invention can manage a divided network hierarchically by dividing an entire network, reduce errors due to network division by specifying division requirements, allow almost all controllers to manage only the topology of a directly connected sub-network, The load on the controller can be reduced.

Referring to FIG. 3, the orchestrator 100 may include an overall topology management unit 110, a topology abstraction unit 120, and an edge controller generation unit 130. In the present embodiment, the orchestrator 100 is shown separately from the root controller 200, but the root controller 200 may replace the function of the orchestrator 100.

The overall topology management unit 110 can manage the entire topology information by building the network topology information based on the connection relationship of the entire real switches 400. [ The overall topology management unit 110 can monitor the connection status of all the actual switches. A physical switch, as used herein, may include a virtual switch as well as a physical switch. However, an actual switch means a switch having only a switching function located in a data plane, and excludes a virtual switch as an edge controller in which the functions of a controller and a switch described later are mixed. However, in order to avoid confusion with what the term virtual switch refers to herein, it is assumed that the actual switch is a physical switch.

 The overall topology management unit 110 may collect the connection relationships and the connection states of the physical switches from the root controller 200 or the edge controller 300. The overall topology management unit 110 may communicate with the edge controller 300 to receive port map information for a virtual port corresponding to a real port to be described later.

The topology abstraction unit 120 may divide all or a part of the entire topology into edge division networks based on the entire topology information by the overall topology management unit 110. [ The edge partitioning network can be a hierarchical structure.

The requirement for starting the segmentation by the topology abstraction unit 120 is not particularly limited. For example, it can be divided into an edge segmentation network from the initial network construction, and divided into a plurality of edge segmentation networks at a certain scale or more.

The edge controller generating unit 130 may generate an edge controller 300 that controls the edge partitioning network divided by the topology abstraction unit 120. A detailed description of the edge controller 300 will be described later. For flexibility and ease of management, the edge controller 300 is preferably a virtual machine (VM) that is created virtually rather than a physical device.

The topology abstraction unit 120 and the edge controller generation unit 130 may limit the construction of the edge division network and the generation of the edge controller. As a limiting requirement, there may be an example where the devices in the switching function in the edge network are linked networks that are directly connected to each other. As another example, the multi-link may be excluded between both edge division networks. In this embodiment, the topology abstraction unit 120 constructs physical switches SW11 to SW15 as one edge division network, and physical switches SW21 to SW25 as another edge division network. The edge controller generation unit 130 generates the edge controllers vSW1 and vSW2 to manage the respective edge division networks.

The orchestrator 100 may further include an equivalent topology generator 140. It can be used to search the optimal path between both nodes, especially the shortest hop. In this embodiment, the root controller 200 can not know the internal topology of the edge controller 300. Figure 10 shows an example of a network topology. For example, referring to FIG. 10, FIG. 10 (a) shows the overall network topology of the physical switches sw6 to sw8, FIG. 10 (b) And an edge controller (vSW6 to vSW8). When viewed from the root controller, the shortest hop path from switch vSW7 to switch vSW8 in the lower layer is vSW7 -> vSW8. However, in the overall topology, vSW7 -> vSW6 -> vSW8 is the shortest hop path. Therefore, in order to correct this, the equivalent topology generation unit 140 needs to generate the equivalent topology as shown in FIG. 10 (c).

The equivalent topology generation unit 140 may monitor whether there is a loop structure of edge division networks from the entire topology information. When the edge segment networks form a loop, the equivalent topology generation unit 140 preferably constructs an equivalent topology among the physical switches of the edge segment network constituting the loop.

The equivalent topology generation unit 140 may receive port map information for a virtual port of the edge controller 300 and a corresponding physical port from the edge controller 300, which will be described later. Based on the port map information, it is preferable that the identification information of the edge port of the equivalent topology is converted into the identification information of the virtual port. The port map information may be received through the communication between the edge topology controller 140 and the edge topology controller 140 and the overall topology manager 110 monitoring the connection information of the switch rather than the communication between the edge topology generator 140 and the edge controller 300 .

The equivalent topology information in which the identification information of the edge port is converted can be transmitted to the root controller 200.

In this embodiment, the controller is presented for the root controller 200 and the edge controller 300. These can serve as general controllers in SDN network systems. The root controller 200 manages at least one edge controller 300 of the lower layer and the edge controller 300 can manage at least one physical switch or at least one lower edge controller of the lower layer. Depending on the layer location, the name of the edge controller and the root controller may be relative, but in the absence of special mention in this specification, it is assumed that the root controller manages at least one edge controller directly below the highest layer.

In a SDN network system, a controller is a kind of command program that controls the SDN system, and can perform various complex functions such as routing, policy declaration, and security check. The controller can define the flow of packets occurring in the plurality of switches in the lower layer. The controller can calculate a route to be flowed by referring to a network topology or the like for a flow allowed in the network policy and then set an entry of the flow on a switch on the path. The controller can communicate with the switch using a specific protocol, for example, an open flow protocol. The communication channel of the controller and the switch can be encrypted by SSL.

4, the root controller 200 includes a switch communication unit 210, a lower topology management unit 220, a path determination unit 230, an entry management unit 240, (DB) 250, which may be referred to as an " entry database "

The lower topology management unit 220 can construct and manage the direct network topology information based on the connection relationship of the lower layer switches collected through the switch communication unit 210. [ In this embodiment, it is preferable that the lower layer switch communicating with the switch communication unit 210 is the edge control unit 300. [ The lower topology management unit 220 can receive the equivalent topology information from the equivalent topology generation unit 140 of the orchestrator 100 and manage the equivalent topology information. In the case of FIGS. 1 and 2, the entire topology information is the topology by the switches SW11 to SW15 and SW21 to SW25. In the entire topology, there is no structure in which the edge partitioning network forms a loop, so an equivalent topology is not constructed. The lower topology that the root controller 200 looks at, that is, the lower topology managed by the lower topology management unit 220, is the topology by the devices vSW1 and vSW2 in Fig. 10 (a) to 10 (c) respectively show the topology of the entire network, the subnetwork, and the equivalent network, respectively.

The path calculation unit 230 can obtain the transmission path of the packet received through the switch communication unit 210 and the action column to be executed by the switch on the transmission path based on the network topology information established in the topology management unit 220.

The entry management unit 240 registers the result calculated by the path calculation unit 230 in the entry DB 250 as a flow entry and responds to a request to add or update a flow entry or entries from the lower layer switch.

5, the root controller 200 includes a switch communication unit 310, a lower topology management unit 320, a path determination unit 330, an entry management unit 340, An entry database (DB) 350, a mapping unit 360, a conversion unit 370, and an upper-level controller communication unit 380 that communicates with the upper-layer controller.

The lower topology management unit 320 can construct and manage the direct network topology information based on the connection relationship of the lower layer switches collected through the switch communication unit 310. [ 1, the lower topology information managed by the lower topology management unit 320 of the edge controller 300 (SW1) is the topology by the switches SW11 to SW15.

The path calculation unit 330 can obtain the transmission path of the packet received through the switch communication unit 310 and the action column to be executed by the switch on the transmission path based on the network topology information established in the topology management unit 320. If there is no flow entry for the received packet, the path calculation unit 330 may transmit all or a part of the packet to the conversion unit 370. [

The entry management unit 340 registers the result calculated by the path calculation unit 330 in the entry DB 350 as a flow entry and responds to a request to add or update a flow entry or entries from the lower layer switch.

The mapping unit 360 converts the edge segmentation network into one virtual switch and converts the edge port of the edge segmentation network into the virtual port of the converted virtual switch when the edge controller is created by the orchestrator 100 . Referring to FIGS. 1 and 2, the edge controller SW1 generates virtual switch identification information vSW1 in order to make the subnetworks SW11-SW15 appear as one virtual switch through the mapping unit 360, The actual edge port p11.1 can be converted to vp1.1. The mapping unit 360 may generate and manage mapping information that associates actual port identification information and virtual port identification information. The mapping unit 360 may transmit the mapping information to the request of the orchestrator 100.

The converting unit 370 can convert the actual port information of the packet received from the path calculating unit 330 into virtual port information using the mapping information of the mapping unit 360. [ The converting unit 370 can inquire the forwarding information about the packet to the controller (the root controller 200 in this embodiment) of the upper layer through the host controller communication unit 380 with respect to the converted packet.

The converting unit 370 can convert the virtual port information of the forwarding message or the flow entry of the root controller 200 received through the host controller communication unit 380 into actual port information using the mapping information. The entry converted by the conversion unit 370 may be stored in the entry DB 350 through the entry management unit 340. [

6 is a flowchart of an SDN network virtualization method of an orchestrator according to an embodiment of the present invention.

Referring to FIG. 6, the orchestrator 100 may virtualize the entire network according to a preset condition (S410). The topology abstraction unit 120 virtualizes the switch 400 when the entire topology management unit 110 of the orchestrator 100 detects that the switch 400 is constructed when the preset condition is established in the network, May generate the edge controller 300 (S430). The topology abstraction unit 120 of the orchestrator 100 uses the entire topology information of the overall topology management unit 110 to calculate the abstraction It can decide whether to start. When the number of the switches 400 exceeds the minimum number, the topology abstraction unit 120 divides the switches 400, which are directly connected to each other by a predetermined number of ranges, into one edge division network . Corresponding to the number of edge division networks, the edge controller generation unit 130 can generate or delete the edge controller 300. [

The topology abstraction unit 120 of the orchestrator 100 may generate a wait flag and broadcast it to the entire network while starting virtualization (S420). This is because an error does not occur when calculating the path setting when the edge segment network construction is completed.

The topology abstraction unit 120 may virtualize the edge partitioning network if it satisfies predetermined conditions such as a virtualization start requirement and an edge partitioning network requirement (S430).

The edge controller generating unit 130 may generate and connect the edge controller 300 to the generated edge partitioning network or may delete the edge controller 300 of the deleted edge partitioning network in operation S440.

The equivalent topology generation unit 140 monitors whether there is a loop structure of edge division networks from the entire topology information (S460), and if there is a loop structure, the equivalent topology information may be generated and transmitted to the root controller 200 (S470 ).

If the orchestrator 100 determines that the construction of the equivalent topology information is completed or does not need to be established, the orchestrator 100 may broadcast a wait flag release message to the entire network through the topology abstraction unit 120 (S480).

7 is a flowchart of an SDN network virtualization method of an edge controller according to an embodiment of the present invention.

Referring to FIG. 7, when the edge controller 300 is generated by the edge controller generating unit 130, the necessary application can be driven (S520). A user request or a predetermined application can be driven from an application for performing a controller and a switch function. In this specification, the components called the switch communication unit, the topology management unit, the path calculation unit, and the like can be driven as a kind of application.

After the preparation for driving is completed, the edge controller 300 generates a virtual switch identifier, and then converts the physical port into a virtual port and maps the virtual port to the virtual switch (S520). When the mapping is completed, the edge controller 300 can send a completion message to the orchestrator 100 to help determine whether to release the waiting flag when the edge controller 300 is ready to operate.

8 is a flowchart of a path request response method of an edge controller according to an embodiment of the present invention.

Referring to FIG. 8, the edge controller 300 may receive a packet-in message, which is a message inquiring a packet path, from a switch in a lower layer through the switch receiver 310 (S610).

The edge controller 300 may determine whether there is an entry matching the packet of the packet-in message (S620).

If the entry exists, it can be determined whether or not there is a wait flag in the orchestrator 100 (S630).

If there is a wait flag in the orchestrator 100, the edge controller 300 waits for the wait flag to be released and then determines whether there is an entry for the received packet again (S620).

If there is no waiting flag in the orchestrator 100, the edge controller 300 can determine whether the database of the upper layer controller, that is, the root controller 200, should be updated (S645).

For example, when the connection of the switch to the lower network of the edge controller 300 is down, or after the down connection, or when a new switch is added, the lower topology of the edge controller 300 changes, This is because the port of the virtual switch to view changes. 11 is a block diagram illustrating an example of a network topology. The switch sw9 group and the switch sw10 group in Fig. 11A can be virtualized into respective edge division networks. In this case, the controller of the upper layer (the root controller 200) determines the lower network topology as shown in FIG. 11 (b). When a problem occurs as shown in FIG. 11 (c) in the link, since the direct connection between the switches of the existing edge partitioning network vSW9 is cut off, the edge partitioning network must be rebuilt. When the edge controller 300 inquires of the root controller 200 about the packet path setting before the rebuilding, the root controller 200 refers to the topology in FIG. 11 (b) instead of FIG. 11 (d) I will respond. The edge controller 300 will inquire back to the root controller 200 because the packet transfer to the flow entry by the root controller 200 is not performed. To avoid such loop iterations, the edge controller 300 needs to determine if there is a change in the topology of the subnetwork and needs to update the database of the root controller 200.

If it is not necessary to update the result of the update determination (S645), the edge controller 300 can send a packet-out message to the switch requesting the packet path through the entry DB 350 to transmit the packet to the proper port (S690). The edge controller 300 may transmit a flow change message to the switch on the path in advance. Such a flow change message may be set to always be sent with a packet-out message.

If it is necessary to update the result of the update determination (S645), for example, if it is necessary to notify the root controller 200 that the switch structure is as shown in Fig. 11 (d), the edge controller 300 updates its sub- And then notifies the root controller 200 of this. 11 (d), the edge controller 300 may cause the root controller 200 to appear as two switches, or a new edge controller may be created by the orchestrator 100. When the update is completed, there is no flow entry for the received packet of the edge controller 300. That is, it is the same as the case where there is no result entry in S620 of the existence determination of the received packet.

(S620) If there is no result entry, the edge controller 300 converts the edge port of the received packet to a virtual port (S650), and sends a packet-in message to the root controller 200 (S660). Upon receiving the packet-out message from the root controller 200 in operation S670, the edge controller 300 converts the virtual port of the packet-out message to an actual port in operation S680, and then, based on the internal network topology information, And then transmits the packet-out message to the inquiring switch (S690). The related entry may be updated in the entry DB 350.

9 is a flowchart of a SDN network virtualization method of a root controller according to an embodiment of the present invention.

Referring to FIG. 9, the root controller 200 may receive a packet-in message from a lower layer switch, i.e., the edge controller 300 (S710). The root controller 200 may calculate the appropriate path based on the information of the packet-in message. If there is equivalent topology information in the calculated appropriate path (S720), the root controller 200 may reflect the equivalent topology first (S730). In step S740, the root controller 200 determines whether there is an entry corresponding to the packet of the packet-in message, and if there is no entry, transmits the packet discard message to the lower layer in step S760. On the basis of this, the root controller 200 may transmit the packet-out message to the switch of the inquiring lower layer (S750).

The present invention can also be embodied as computer-readable codes on a computer-readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like, and may be implemented in the form of a carrier wave (for example, transmission via the Internet) . The computer readable recording medium may also be distributed over a networked computer system so that computer readable code can be stored and executed in a distributed manner. And functional programs, codes, and code segments for implementing the present invention can be easily inferred by programmers skilled in the art to which the present invention pertains.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes and modifications may be made by those skilled in the art without departing from the spirit and scope of the present invention.

100: Orchestrator 200: Root controller
300: Edge controller 400: Switch

Claims (8)

Generates forwarding information for the flow forwarding query of the lower layer and responds to the generated forwarding information,
Generating mapping information so that each of the plurality of edge ports of the plurality of switches constituting the lower layer corresponds to a plurality of virtual ports of one virtual switch,
When the forwarding information for the flow forwarding query having the edge port received from the lower layer is not generated, forwarding information for the flow converted into the virtual port corresponding to the edge port based on the mapping information, And an edge controller for inquiring into the layer,
And the edge controller is regarded as a switch for making a flow forwarding inquiry in an upper layer. delete The method according to claim 1,
Wherein the edge controller generates forwarding information for the flow having the edge port based on the forwarding information for the transformed flow received from the upper layer, the mapping information, and the network topology of the lower layer, Based hierarchical network system. The method according to claim 1,
Wherein the plurality of switches constituting the lower layer constitute an edge division network connected to each other. 5. The method of claim 4,
Wherein the edge controller updates the mapping information when a switch having an edge port is added or deleted to the edge segment network. 6. The method of claim 5,
The edge controller transmits to the upper and lower layers a flag indicating that the topology of the lower layer is updated before updating the mapping information,
Wherein the forwarding of the flow associated with the topology update flag is suspended. Generates forwarding information for the flow forwarding query of the lower layer and responds to the generated forwarding information,
Generating mapping information so that each of the plurality of edge ports of the plurality of switches constituting the lower layer corresponds to a plurality of virtual ports of one virtual switch,
When the forwarding information for the flow forwarding query having the edge port received from the lower layer is not generated, forwarding information for the flow converted into the virtual port corresponding to the edge port based on the mapping information, And an edge controller for inquiring into the layer,
Wherein the plurality of switches constituting the lower layer constitute an edge division network connected to each other,
The lower layer of the root controller that responds to the forwarding information of the upper layer flow is composed of a plurality of edge controllers,
At least first to third edge controllers of the plurality of edge controllers form a loop topology,
Wherein the root controller has a plurality of equivalent topology information corresponding to the number of hops between all the nodes based on a topology of all the nodes constituting the lower layer of the first to third edge controllers,
Edge ports of the plurality of equivalent topology information are converted into virtual edge ports mapped by the first to third edge controllers,
And the root controller generates the plurality of equivalent topology information by considering the loop topology by the first to third controllers when generating forwarding information for the inquiry of the flow forwarding information between the virtual edge ports. SDN - based partitioned hierarchical network system. An SDN (Software Defined Network) based network system including an orchestrator, a root controller, and a plurality of switches,
The orchestrator includes:
A topology abstraction unit for setting one switch or a predetermined number or less of the connected switches to one edge division network; And
And an edge controller generator for generating an edge controller for controlling the edge division network,
Wherein the first edge controller generated by the edge controller generator comprises:
A topology management unit managing a topology of a first edge division network that is a lower layer of the first edge controller;
An entry manager for generating and responding to forwarding information for a lower layer forwarding query;
A mapping unit configured to generate mapping information such that each of a plurality of edge ports of a plurality of switches constituting the lower layer corresponds to a plurality of virtual ports of one virtual switch;
When the forwarding information for the flow forwarding query having the edge port received from the lower layer is not generated, forwarding information for the flow converted into the virtual port corresponding to the edge port based on the mapping information, The SDN-based network system comprising:

Images