JP2010226545A - Network node, network, and tunnel switching method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a network node capable of reducing an increase in the load of a virtual node and a decrease in data transfer performance. <P>SOLUTION: The network node 1 includes: a tunnel switch 11 for discriminating whether traffic from the external is general traffic or overlay traffic; a flow table 14b for associating the header information of a packet with the processing contents of the packet and recording both as a flow entry; and an overlay flow monitoring function 113 for monitoring an input port in the tunnel switch and the processing contents of the packet of the overlay traffic in the tunnel switch and recording the monitored results in the flow table 14b, wherein the network node processes the packet of the overlay traffic in the tunnel switch based on reference results of the flow table with the header information extracted from the packet of the overlay traffic as a retrieving key. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a network node, a network, a tunnel switching method used therefor, and a program therefor, and more particularly to a tunnel switching method that distributes tunnel traffic of an overlay network in a network constituting the overlay network.

  The network device related to the present invention specializes in data transfer, processes a packet of arriving data using a protocol installed in the device, and transfers the packet to a destination.

  However, in the future Internet environment, in order to provide various network applications and services, a plurality of virtual networks (overlay networks) are constructed on one real network (underlay network) such as an IP (Internet Protocol) network. It is thought.

  In this Internet environment, a plurality of virtual nodes are created using a virtualization technology for one physical network device, and a routing protocol corresponding to the service is mounted on the virtual node, thereby supporting each service. Multiple overlay networks can be created.

  However, in this method, when creating a plurality of virtual nodes in one physical network device, since the traffic of the overlay network is processed on the virtual node, processing overhead, transfer delay, and high load on the virtual node The problem occurs.

  As for overlay network nodes and network virtualization, for example, there is a technique described in Non-Patent Document 1 described later. Non-Patent Document 1 discloses research projects for large-scale overlay networks around the world.

  In the technique described in Non-Patent Document 1, a virtualization technique can be used to create a plurality of slices (virtual nodes) on a server, which are provided to a plurality of researchers. Each server is located in various parts of the world and connected via the Internet.

  In each server, resources [CPU (central processing unit), memory, etc.] are shared and provided to a plurality of researchers as a slice (virtual node) unit. Researchers collect slices connected via the Internet, create a testbed for a global virtual network, and load and test their own network-related applications and protocols on the testbed. .

PlanetLab: An Overtest Tested for Broadcast-Coverage Services [http: // www. planet-lab. org / files / pdn / PDN-03-009 / pdn-03-009. pdf]

  However, in the above-described virtual network, a plurality of servers (nodes) are placed all over the world, and the servers are connected via the Internet as end nodes. However, the core part of the network inside the Internet is a general network device (underlay network), and even if the entire overlay network is created, it is affected by the packet processing of the general network device. There is.

  In the above virtual network, a virtual node is created on each server, but the traffic of each tunnel in the overlay network is forwarded to each virtual node, processed on the virtual node, and then forwarded to the destination. ing. However, virtual nodes are created at the application level, and when each virtual node is responsible for both routing protocol processing and data transfer processing, the load on the virtual node increases and data transfer performance decreases. is there.

  Accordingly, an object of the present invention is to provide a network node, a network, a tunnel switching method used therefor, and a program therefor that can solve the above-described problems and can reduce an increase in load on a virtual node and a decrease in data transfer performance. There is.

The first network node according to the present invention includes a tunnel switch for identifying whether external traffic is general traffic or overlay traffic;
A flow table that records packet header information and processing contents of the packet in association with each other, and a flow entry;
An overlay flow monitoring function for monitoring an input port in the tunnel switch and processing contents of the overlay traffic packet in the tunnel switch and recording the monitoring result in the flow table;
Based on the reference result of the flow table using the header information extracted from the packet of the overlay traffic as a search key, the packet of the overlay traffic is processed in the tunnel switch.

The second network node according to the present invention constitutes an underlay network, and a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes processes overlay traffic. A network node that
In each of the plurality of virtual nodes, an overlay routing function and an overlay forwarding function for processing the overlay traffic are provided.
Underlay routing function and underlay forwarding function for processing general traffic,
A tunnel switch for identifying whether the traffic from the outside is the general traffic or the overlay traffic;
A flow table for associating and recording packet header information and processing of the packet as a flow entry;
An overlay flow monitoring function for monitoring an input port in the tunnel switch and processing contents of the overlay traffic packet in the tunnel switch and recording the monitoring result in the flow table;
The flow table is referenced using a header information search key extracted from the overlay traffic packet after the first overlay traffic packet, and the overlay traffic packet is processed in the tunnel switch based on the reference result. Yes.

  A network according to the present invention includes the above-described network node.

According to the first tunnel switching method of the present invention, an underlay network is configured, and a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes receives overlay traffic. A tunnel switching method used for a network node to be processed,
The network node is provided with a tunnel switch for identifying whether external traffic is general traffic or the overlay traffic, and a flow table that records packet header information and processing contents of the packet in association with each other as a flow entry,
The network node performs an overlay flow monitoring process of monitoring an input port in the tunnel switch and a processing content for the packet of the overlay traffic in the tunnel switch and recording the monitoring result in the flow table;
The tunnel switch processes the overlay traffic packet in the tunnel switch based on the reference result of the flow table using the header information extracted from the overlay traffic packet as a search key.

In the second tunnel switching method according to the present invention, an underlay network is configured, and a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes receives overlay traffic. A tunnel switching method used for a network node to be processed,
The network node is provided with a tunnel switch that identifies whether traffic from the outside is the general traffic or the overlay traffic, and a flow table that records packet header information and processing of the packet as a flow entry in association with each other,
Each of the plurality of virtual nodes is provided with an overlay routing function and an overlay forwarding function for processing the overlay traffic,
The network node monitors an underlay routing process and an underlay forwarding process for processing general traffic, an input port in the tunnel switch, and a processing content for a packet of the overlay traffic in the tunnel switch and An overlay flow monitoring process for recording a monitoring result in the flow table;
The tunnel switch refers to the flow table using a search key of header information extracted from the overlay traffic packet after the first overlay traffic packet, and based on the reference result, the tunnel switch refers to the overlay traffic packet. Is processing within.

The first program according to the present invention constitutes an underlay network, and a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes processes overlay traffic. A program to be executed by a central processing unit in a network node,
The network node is provided with a tunnel switch for identifying whether external traffic is general traffic or the overlay traffic, and a flow table that records packet header information and processing contents of the packet in association with each other as a flow entry,
An overlay flow monitoring process for monitoring an input port in the tunnel switch and a processing content for the packet of the overlay traffic in the tunnel switch and recording the monitoring result in the flow table;
The tunnel switch is caused to process the overlay traffic packet in the tunnel switch based on a reference result of the flow table using header information extracted from the overlay traffic packet as a search key.

The second program according to the present invention constitutes an underlay network, and a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes processes overlay traffic. A program to be executed by a central processing unit in a network node,
The network node is provided with a tunnel switch that identifies whether traffic from the outside is the general traffic or the overlay traffic, and a flow table that records packet header information and processing of the packet as a flow entry in association with each other,
Each of the plurality of virtual nodes is provided with an overlay routing function and an overlay forwarding function for processing the overlay traffic,
Underlay routing processing and underlay forwarding processing for processing general traffic, input port in the tunnel switch and processing contents of the overlay traffic packet in the tunnel switch are monitored and the monitoring result is the flow Including overlay flow monitoring to record in the table,
The tunnel switch is made to refer to the flow table using a search key of header information extracted from the overlay traffic packet after the first overlay traffic packet, and the overlay traffic packet is referred to the tunnel switch based on the reference result It is characterized by being processed within.

  By adopting the configuration and operation as described above, the present invention can reduce the increase in the load on the virtual node and the decrease in the data transfer performance.

It is a block diagram which shows the structural example of the network node by the 1st Embodiment of this invention. It is a block diagram which shows the internal structure of the network node by the 1st Embodiment of this invention. It is a block diagram which shows the basic structural example of the underlay network by the 1st Embodiment of this invention. It is a block diagram which shows the basic structural example of the overlay network (A) by the 1st Embodiment of this invention. It is a block diagram which shows the basic structural example of the overlay network (B) by the 1st Embodiment of this invention. It is a figure which shows the action by which the packet is processed when the packet of the first overlay traffic arrives in the 1st Embodiment of this invention. It is a figure which shows the structural example of the flow table by the 1st Embodiment of this invention. It is a figure which shows the method of producing a new overlay flow table by the overlay flow monitoring function in the 1st Embodiment of this invention. It is a figure which shows the structural example of the packet of the overlay traffic to which the 1st Embodiment of this invention is applied. It is a figure which shows the internal structure of the header field of the flow entry in the flow table of FIG. It is a flowchart which shows the process with respect to the first packet by the network node of FIG. FIG. 2 is a flowchart showing a process for a packet after the first overlay traffic packet by the network node of FIG. 1. FIG. FIG. 2 is a flowchart showing a process of recording the processing contents of the first overlay traffic packet in the overlay flow table with high priority by the network node of FIG.

  Next, embodiments of the present invention will be described with reference to the drawings. First, an outline of a network according to the present invention will be described. In the network according to the present invention, a plurality of virtual nodes are created on a network node in order to create a plurality of overlay networks.

  In the network node constituting the network according to the present invention, in order to process the tunnel (virtual link) traffic of the overlay network, the first packet of the overlay traffic is processed by the overlay routing function and the overlay forwarding function in the virtual node. Further, the network node transfers the overlay traffic at high speed by referring to the entry of the flow table in the tunnel switch for the overlay traffic that has arrived thereafter.

  In other words, the network node includes a tunnel switch that identifies whether traffic from the outside of the network node is general traffic or overlay traffic.

  In this tunnel switch, general packet data and overlay packet data are distinguished from each other by information of an outer IP (Internet Protocol) header, a tunnel header, and an inner IP header in the header of the packet. Further, the tunnel switch uses the information extracted from the header portion of the packet as a search key, refers to the entry recorded in the flow table, and determines the processing of the packet.

  In the packet processing (action) in the tunnel switch, the general traffic packet is distributed to the underlay forwarding function, and the processing of the underlay traffic is entrusted. The overlay traffic packet is decapped (removed: decapsulated) by the action defined in the packet processing of the flow entry, distributed to the output port of the tunnel switch, and transferred to the virtual node.

  Furthermore, for packets returning from the virtual node inside the network node to the tunnel switch, the IP header of the new outer is encap (encapsulated) by the action defined in the packet processing of the flow entry, and the underlay forwarding function Send to. The underlay forwarding function transfers an overlay traffic packet to the outside using information in the outer IP header.

  The above-described tunnel switch has an overlay flow monitoring function for monitoring an input port in the tunnel switch and recording the overlay network key of overlay traffic and the IP header information of the inner as a reference key. The overlay flow monitoring function records each action in which a packet of overlay traffic is processed in the network node in a temporary storage table that temporarily stores the information for each reference key.

  The overlay flow monitoring function, when information of Decap, Encap, and output port, which are processed actions of the flow of each overlay traffic, is temporarily stored in the temporary storage table, the information on the temporary storage table is used as a high priority entry. Record in the higher priority overlay flow table. This entry records three actions: Decap, Encap (with new IP header information), and output port information.

  Therefore, the tunnel switch refers to the overlay flow table having a high priority with the search key of the header information extracted from the overlay traffic packet after the first packet, and does not process the overlay traffic packet in the virtual node. In the tunnel switch, a new outer IP header is added by Decap, Encap, and output port which are three actions of packet processing of the entry.

  In the tunnel switch, the flow table is referenced based on the search key extracted from the packet of overlay traffic. If no entry is found in the flow table, the entry with the lowest priority is set with only the input port information of the search key. The packet is hit, and the packet is transferred to the underlay forwarding function.

  Thus, in the network node according to the present invention, the three actions of the packet processing are determined from the result of referring to the entry of the flow table with the search key extracted from the header of the overlay traffic packet. Can handle overlay traffic packets.

  Therefore, in the network node according to the present invention, even if the entire overlay network is created, the increase in the load on the virtual node and the decrease in the data transfer performance can be reduced without being affected by the packet processing of general network equipment. it can.

  FIG. 1 is a block diagram showing a configuration example of a network node according to the first embodiment of the present invention. In FIG. 1, the network node 1 according to the first embodiment of the present invention includes a tunnel switch 11, an underlay forwarding function 12 for processing general traffic, and a virtual node 13.

  The tunnel switch 11 includes a Decap action 111, an Encap action 112, an overlay flow monitoring function 113, and a physical network interface 114, and an action (not shown) for transferring to the physical network interface 114 by an entry in the flow table 14b. Have.

  The virtual node 13 includes an overlay forwarding function 131 that processes overlay traffic and a virtual network interface 132. The overlay flow table 14a having a high priority records entries from the overlay flow monitoring function 113.

  First, when the first external packet arrives at the physical network interface 114 of the network node 1, the tunnel switch 11 to which the physical network interface 114 is connected extracts the packet header, and the general traffic packet And overlay traffic packets.

  The tunnel switch 11 refers to the flow entry recorded in the flow table 14b using the extracted header information as a search key. The tunnel switch 11 transfers the packet determined to be a general traffic packet from the result of referring to the flow entry to the underlay forwarding function 12. The underlay forwarding function 12 processes the packet by an underlay packet transfer processing method and transfers the packet to the physical network interface 114.

  Further, the tunnel switch 11 removes the outer IP header from the Decap action 111 for the packet determined to be an overlay traffic packet as a result of referring to the flow entry, and performs the packet processing action described in the flow entry. Transfer to the destination virtual node 13 (see 210 in FIG. 1).

  In the virtual node 13, for the transferred packet, the overlay forwarding function 131 determines the transfer destination of the packet and outputs it to the virtual network interface 132.

  For the packet 211 returned from the virtual node 13 to the tunnel switch 11, the tunnel switch 11 refers again to the entry in the flow table 14b from the information of the connected virtual network interface 132, and the virtual link (tunnel) ) Is determined. The tunnel switch 11 adds the source address and destination address of the virtual link as a new outer IP header by the Encap action 112 which is a processing action of the packet processing, and transfers it to the underlay forwarding function 12.

  The underlay forwarding function 12 determines the packet transfer destination and output interface using the outer IP header, and transfers the packet to the physical network interface 114.

  Here, the overlay flow monitoring function 113, when the first overlay flow passes, the physical network interface 114, the virtual network interface 132 to which the virtual node 13 and the tunnel switch 11 are connected, the tunnel switch 11 and the underlay forwarding. Three locations of the port to which the function 12 is connected are monitored (see 213 in FIG. 1).

  The overlay flow monitoring function 113 records the action of processing the packet of the passed overlay traffic and the physical interface that is output last, and stores the information of Decap, Encap, and output interface that are the recorded packet processing actions. It is recorded in the temporary storage table 14c (see 216 in FIG. 1). In this case, “Key | Inner IP Header”, “Decap”, “Encap (New Outer IP Header)”, “Forward (Output I / F)”, “Header Fields” are recorded in the temporary storage table 14c. (See FIG. 1).

  After the determination of Decap, Encap, and output interface, which are the above three packet processing actions, is recorded in the temporary storage table 14c, the overlay flow monitoring function 113 uses the overlay network key, which is the reference key, and the inner IP. The address and the IP address of the outer before decapping are used as a search key, and three actions are registered as a new action group in the high-priority overlay flow table 14a, which is higher in the flow table 14b (see 215 in FIG. 1). ). “Action: Decap Encap Forward (Output I / F)” is registered in the overlay flow table 14a having a high priority.

  Therefore, the first overlay traffic packet (Long Path Forwarding) 213 is received by the inside of the tunnel switch 11, the virtual network overlay forwarding function 131, and the underlay forwarding function 12 after the packet arrives at the network node 1 from the outside. , Decap, Encap, and output interface are determined in this order.

  The overlay flow monitoring function 113 monitors each process and records the monitoring result in the overlay flow table 14a having a high priority. As a result, the tunnel switch 11 performs processing for a packet (Short Path Forwarding) 214 that arrives after the packet 213 of the first overlay traffic with reference to the entry in the overlay flow table 14a having a high priority.

  That is, the tunnel switch 11 determines three actions for the packet processing from the result of referring to the search key extracted from the header of the packet of the overlay traffic and the flow entry in the overlay flow table 14a having a high priority. Therefore, the packet of overlay traffic can be processed in the tunnel switch 11.

  Therefore, in the network node 1 according to the first embodiment of the present invention, even if the entire overlay network is created, the increase in the load on the virtual node 13 and data are not affected by the packet processing of general network devices. Reduction in transfer performance can be reduced.

  FIG. 2 is a block diagram showing an internal configuration of the network node 1 according to the first embodiment of the present invention. In FIG. 2, a plurality of virtual nodes (A, B) 13a and 13b can be created in the network node 1.

  In the virtual nodes (A, B) 13a and 13b, overlay routing functions 133a and 133b for processing overlay traffic and overlay forwarding functions 131a and 131b are mounted.

  Further, in the network node 1, general traffic such as underlay traffic is processed by the underlay routing function 15 and the underlay forwarding function 12.

  Inside the network node 1, a tunnel switch 11 is mounted at a location connected to the physical network interface 114, and general packets and overlay packets of the arrived traffic are identified and recorded in the flow table [DB (database) 14]. Referring to the current flow entry, the action to be processed by the packet and the output interface of the transfer destination are determined.

  The tunnel switch 11 includes therein the overlay flow monitoring function 113 and an action 110 including a Decap action 111 and an Encap action 112.

  FIG. 3 is a block diagram showing a basic configuration example of an underlay network (real network) according to the first embodiment of the present invention, and FIG. 4 is an overlay network (virtual network) according to the first embodiment of the present invention. FIG. 5 is a block diagram showing a basic configuration example of (A), and FIG. 5 is a block diagram showing a basic configuration example of an overlay network (virtual network) (B) according to the first embodiment of the present invention.

  In FIG. 3, an underlay network 100 is composed of overlay nodes (network nodes) 1-0 to 1-4 that construct an overlay network. The overlay nodes 1-0 to 1-4 are virtual nodes (A) 13a-0 to 13a-2, 13a-4, virtual nodes (B) 13b-0, 13b-1, 13b-3, and 13b-4. And are connected by underlay links (physical links) 230-233.

  In FIG. 4, an overlay network (A) 140 includes virtual nodes (A) 13a-0 to 13a-2 and 13a-4 on overlay nodes 1-0 to 1-2 and 1-4 as overlay links (virtual links). (Tunnel) 141 to 144 are connected and constructed.

  In FIG. 5, the overlay network (B) 150 includes virtual nodes (B) 13b-0, 13b-1, 13b-3, 13b- on the overlay nodes 1-0, 1-1, 1-3, 1-4. 4 are connected by overlay links (virtual links) (tunnels) 151-154.

  FIG. 6 is a diagram showing an action of processing a packet when the first overlay traffic packet arrives in the first embodiment of the present invention. In FIG. 6, when a packet of the first overlay traffic from the outside arrives at the physical network interface 114 of the network node 1, the tunnel switch 11 to which the physical network interface 114 is connected extracts the header of the packet, Differentiate between traffic traffic packets and overlay traffic packets.

  The tunnel switch 11 refers to the flow entry recorded in the flow table 14b using the extracted header information as a search key. The tunnel switch 11 transfers the packet determined to be a general traffic packet from the result of referring to the flow entry to the underlay forwarding function 12. The underlay forwarding function 12 processes the packet by an underlay packet transfer processing method and transfers the packet to the physical network interface 114.

  Also, the tunnel switch 11 removes the outer IP header of the packet from the packet determined to be an overlay traffic packet by the Decap action 111 as a result of referring to the flow entry, and processes the packet described in the flow entry. Is transferred to the destination virtual node 13 (see 310 in FIG. 6).

  In the virtual node 13, for the transferred packet, the overlay forwarding function 131 determines the transfer destination of the packet and outputs it to the virtual network interface 132.

  The tunnel switch 11 refers to the entry in the flow table 14b again for the packet (311) returned from the virtual node 13 using the information of the connected virtual network interface 132 as a search key, and the virtual link ( Determine the source address and destination address of the tunnel.

  Thereafter, the tunnel switch 11 adds the source address and destination address of the virtual link as a new outer IP header by the Encap action 112, and forwards it to the underlay forwarding function 12.

  The underlay forwarding function 12 determines the packet transfer destination and output interface using the outer IP header, and sends the packet to the physical network interface 114.

  FIG. 7 is a diagram showing a configuration example of a flow table according to the first embodiment of the present invention. In FIG. 7, the flow table 14b records a flow entry 140b. The flow entry 140b includes a header field 141b ("header field # 1" to "header field #n") to be referred to and a reference to the header field 141b. The packet processing from the result and the action 142b ("action # 1" to "action #n") for judging the transmission destination are described.

  The tunnel switch 11 extracts the outer IP header, the overlay network key, and the inner IP header from the information of the arrived overlay header, and uses the flow entry 140b in the flow table 14b as a search key. By referencing, the action and distribution of each packet processing is performed.

  FIG. 8 is a diagram showing a method for creating a new overlay flow table by the overlay flow monitoring function 113 in the first embodiment of the present invention.

  The overlay flow monitoring function 113 in the tunnel switch 11 includes an input network interface (physical network interface 114) and a virtual network interface 132 to which the virtual node 13 and the tunnel switch 11 are connected when the first overlay flow passes. Three locations of ports to which the tunnel switch 11 and the underlay forwarding function 12 are connected are monitored. The overlay flow monitoring function 113 records the action of processing the packet of the passed overlay traffic and the physical interface output last (see 216 in FIG. 8), and the recorded packet processing action Decap, The information of Encap and output interface is recorded in the temporary storage table 14c.

  That is, in the temporary storage table 14c, “key”, “Inner IP Src Addr”, “Inner IP Dst Addr” 142c, “Decap”: “Old Outer IP Header” 143c, “Encap”: “New Outer IP” A flow entry 141c including “Header” 144c and “Forward”: “Output Interface” 145c is recorded.

  After recording Decap, Encap, and output interface decisions, which are the above three packet processing actions, in the temporary storage table 14c, the overlay flow monitoring function 113 uses the overlay network key, which is the reference key, and the inner IP address. And the outer IP address before the Decap as a search key, and three actions as new action groups are recorded in the high priority overlay flow table 14a, which is higher in the flow table 14b.

  Therefore, after the packet of the first overlay traffic arrives at the network node 1 from the outside, the inside of the tunnel switch 11, the overlay forwarding function 131 of the virtual network, and the underlay forwarding function 12, the Decap, Encap, and output Interface decisions are processed in order.

  The overlay flow monitoring function 113 monitors each process, and records the monitoring result in the overlay flow table 14a having a high priority, so that a packet that arrives after the packet of the first overlay traffic is switched to the tunnel switch. 11 is processed with reference to an entry in the overlay flow table 14a having a high priority.

  Therefore, the tunnel switch 11 determines the three actions of the packet processing from the result of referring to the flow entry with the search key extracted from the header of the overlay traffic packet. Overlay traffic packets that arrive after the packet can be processed.

  FIG. 9 is a diagram showing a configuration example of an overlay traffic packet to which the first embodiment of the present invention is applied. In FIG. 9, the overlay traffic packet includes an outer IP address (outer Src IP Add 405, outer Dst IP Add 406) of the outer header 401 of the overlay traffic, an overlay network key 407 of the tunnel header 402, an inner The inner IP address (inner Src IP Add 408, inner Dst IP Add 409) of the header 403 is extracted, and using them as a search key, the flow entry of the flow table 14b is referred to, and the packet processing action is determined. In FIG. 9, reference numeral 404 denotes a payload.

  FIG. 10 shows the internal structure of the header field 141b of the flow entry 140b in the flow table 14b of FIG. In FIG. 10, the header field 141b of the flow entry 140b is composed of an outer IP header, an overlay network key, and a packet header portion extracted from the inner IP header.

  That is, the header field 141b includes “Input Port”, “Ether (registered trademark) Source”, “Ether (registered trademark) Dest”, “Ether (registered trademark) Type”, “VLAN (Virtual Local Area Network) id”, “Outer IP Src Addr”, “Outer IP Dst Addr”, “Tunnel Proto Type”, “Inner Proto Type”, “Over-key Key”, “Inner IP Src Addr”, “Inner IP”.

  In addition, since the header field 141b has a corresponding action 142b ("action: · Decap · Encap · Forward (Output I / F)"), an action for executing packet processing from an entry hit by a reference from the search key. Judgment can be made.

  As shown in FIG. 8, the overlay flow monitoring function 113 records the Decap, Encap, and output port (interface) of each process of the first overlay traffic in the overlay network key and the inner IP header unit, and temporarily stores them. Record in table 14c. After all three Decap, Encap, and output ports are recorded, the overlay flow monitoring function 113 records a new header field and three action groups in the higher priority overlay flow table 14a.

  11 is a flowchart showing processing for the first packet by the network node 1 in FIG. 1, and FIG. 12 is a flowchart showing processing for packets after the first overlay traffic packet by the network node 1 in FIG. FIG. 7 is a flowchart showing a process of recording the processing contents of the first overlay traffic packet in the overlay flow table 14a having a high priority by the network node 1 of FIG.

  Processing performed by the network node 1 will be described with reference to FIGS. The processes shown in FIGS. 11 to 13 are realized by executing a program stored in a memory or the like by a control unit [for example, a CPU (Central Processing Unit) or the like] in the network node 1.

  In the network node 1, the general traffic and the overlay traffic are identified with respect to the first packet by the tunnel switch 11, and when it is determined as the bucket of the overlay traffic, the flow is determined using the search key extracted from the header field. The action of each packet processing is determined with reference to the table 14b.

  In the network node 1, the search key information is extracted from the header part of the packet by the tunnel switch 11 (step S1 in FIG. 11), and it is determined whether the search key has an overlay network key (step S2 in FIG. 11).

  When there is no overlay network key, the tunnel switch 11 refers to the flow table 14b and distributes the packet to the underlay forwarding function 12 (step S3 in FIG. 11).

  If there is an overlay network key, the tunnel switch 11 refers to the flow table 14b, executes the packet processing Decap action 111 on the packet, and transfers the packet to the virtual node 13 (step S4 in FIG. 11). In the virtual node 13, the packet is transferred to the transfer destination by the overlay forwarding function 131 (step S5 in FIG. 11).

  The tunnel switch 11 refers to the flow table 14b, executes an Encap action 112 for packet processing, and transfers the packet to the underlay forwarding function 12 (step S6 in FIG. 11). The underlay forwarding function 12 determines the transfer destination of the packet and sends it to the transfer destination (step S7 in FIG. 11).

  In the network node 1, after processing the packet of the first overlay traffic, a new high-priority overlay flow table 14a is created, so that the packet arriving later after that packet is the higher-priority overlay flow. By referring to the table 14a, Decap, Encap, and output port are determined, each action is processed in the tunnel switch 11, and the packet is sent to the transfer destination.

  In the network node 1, the search key information is extracted from the header part of the packet by the tunnel switch 11 (step S11 in FIG. 12), and the entry of the higher priority overlay flow table 14a is referred to (step S12 in FIG. 12). .

  If the search key does not hit the header field (step S13 in FIG. 12), the tunnel switch 11 refers to the entry in the lower flow table 14b (step S14 in FIG. 12).

  The tunnel switch 11 hits the header field with the search key (step S13 in FIG. 12), executes three packet processing actions: Decap, Encap, and output interface, and sends the packet (step S15 in FIG. 12).

  In the network node 1, when the packet of the first overlay traffic is processed, each processing action is recorded by the overlay flow monitoring function 113 of the tunnel switch 11, and the three action groups and the new header field are recorded in the high priority overlay. Record in the flow table 14a.

  In the network node 1, the physical port is monitored by the overlay flow monitoring function 113 of the tunnel switch 11, and the packet Decap action 111 is recorded (step S21 in FIG. 13).

  The overlay flow monitoring function 113 monitors the virtual port (virtual network interface 132) with the virtual node 13, and records the packet Encap action 112 (step S22 in FIG. 13).

  The overlay flow monitoring function 113 monitors the port with the underlay forwarding function 12 and records the action of the output destination of the packet (step S23 in FIG. 13).

  The overlay flow monitoring function 113 waits (step S25 in FIG. 13) unless three actions: Decap, Encap, and output destination are recorded (step S24 in FIG. 13).

  If the overlay flow monitoring function 113 records three actions: Decap, Encap, and the output destination (step S24 in FIG. 13), the flow entry is registered in the overlay flow table 14a having a higher priority (FIG. 13). Step S26).

  As described above, in this embodiment, a plurality of virtual nodes (A, B) 13a and 13b are created on the network nodes 1-0 to 1-4 in order to construct a plurality of overlay networks 140 and 150. The tunnel traffic packets of the overlay networks 140 and 150 are processed by the overlay routing functions 133a and 133b and the overlay forwarding functions 131a and 131b, the first packets of which are in the virtual nodes (A, B) 13a and 13b.

  Further, in the present embodiment, the entries in the overlay flow table 14a registered by the overlay flow monitoring function 113 are referred to the packets of the same traffic flow after the first packet, and according to the contents of the entries. Processing.

  Therefore, in the present embodiment, since the packets other than the first packet are processed by referring to the overlay flow table 14a and the flow table 14b in the tunnel switch 11 with respect to the overlay traffic packet, the overlay traffic processing speed is high. And the load caused by packet processing of the virtual nodes (A, B) 13a and 13b can be reduced.

  That is, in this embodiment, even if the entire overlay network is created, the load on the virtual nodes (A, B) 13a and 13b is increased and the data transfer performance is not affected by packet processing of general network devices. Can be reduced.

  In the above embodiment, the virtual node 13 and the virtual nodes (A, B) 13a and 13b have been described. However, the other overlay nodes 1-0 to 1-4 are configured and operated in the same manner as the network node 1. By so doing, the present invention is applied to the virtual nodes (A) 13a-0 to 13a-2, 13a-4 and the virtual nodes (B) 13b-0, 13b-1, 13b-3, 13b-4. be able to.

  As described above, the present invention processes the packet of the overlay traffic that arrives first on the virtual node, but since the subsequent packet can be processed in the tunnel switch 11, the processing performance of packet transfer is improved. Since the packet transfer processing load in the virtual node 13, virtual node (A) 13a, and virtual node (B) 13b can be reduced, more virtual nodes can be mounted.

1,1-0 to 1-4 Network node (Overlay node)
11 Tunnel switch
12 Underlay forwarding function
13 Virtual node 13a,
13a-0 to 13a-2,
13a-4 Virtual node (A)
13b,
13b-0, 13b-1,
13b-3, 13a-4 Virtual node (B)
14 Database
14a High priority overlay flow table
14b Flow table
14c Temporary storage table
15 Underlay routing function
100 Underlay network
110 Action
111 Decap action
112 Encap action
113 Overlay flow monitoring function
114 Physical network interface 131, 131a, 131b Overlay forwarding function
132 Virtual Network Interface
133a, 133b Overlay routing function
140,150 Overlay network 141-144
151-154 Overlay link (virtual link)
230-233 Underlay link (physical link)

Claims (31)

A tunnel switch that identifies whether external traffic is general traffic or overlay traffic,
A flow table that records packet header information and processing contents of the packet in association with each other, and a flow entry;
An overlay flow monitoring function for monitoring an input port in the tunnel switch and a processing content of the overlay traffic packet in the tunnel switch and recording the monitoring result in the flow table;
A network node, wherein a packet of the overlay traffic is processed in the tunnel switch based on a reference result of the flow table using header information extracted from the packet of the overlay traffic as a search key. A plurality of overlay networks are constructed by using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes is a network node for processing the overlay traffic.
In each of the plurality of virtual nodes, an overlay routing function and an overlay forwarding function for processing the overlay traffic are provided.
Underlay routing function and underlay forwarding function for processing general traffic,
A tunnel switch for identifying whether the traffic from the outside is the general traffic or the overlay traffic;
A flow table for associating and recording packet header information and processing of the packet as a flow entry;
An overlay flow monitoring function for monitoring an input port in the tunnel switch and a processing content of the overlay traffic packet in the tunnel switch and recording the monitoring result in the flow table;
Referencing the flow table using a search key of header information extracted from the overlay traffic packet after the first overlay traffic packet, and processing the overlay traffic packet in the tunnel switch based on the reference result A network node characterized by   The overlay flow monitoring function monitors an input port in the tunnel switch and records an overlay network key of the overlay traffic and an inner IP (Internet Protocol) header information as a reference key. 2. The network node according to 2.   The tunnel switch refers to the flow entry registered in the flow table using a search key extracted from the header information of the overlay traffic packet, and based on the reference result, the header information of the overlay traffic packet header information. 4. The network node according to claim 2, wherein an outer header is removed, and the packet from which the outer header is removed is transferred to a destination virtual node.   The tunnel switch refers to the flow entry registered in the flow table using a search key extracted from the header information of the packet of the overlay traffic, and determines the local address of the tunnel of the virtual link from the information of the referred input port. 5. The network node according to claim 2, wherein a remote address is acquired, added to the packet of the overlay traffic as a new outer header, and transferred to the underlay forwarding function.   6. The network node according to claim 5, wherein the underlay forwarding function sends out an overlay traffic packet to which the new outer header is added based on information of the outer header.   7. A header field is recorded in the flow table using header information extracted from the packet, and processing contents of the packet are recorded in association with each header field. A network node according to any of the above.   Referring to the flow entry of the flow table using the outer header extracted from the packet of the overlay traffic arriving from the tunnel switch, the key of the overlay network, and the IP header information of the inner as a search key 8. The network node according to claim 2, wherein a packet of the overlay traffic is processed according to a processing content corresponding to.   If the flow table is referenced based on the search key extracted from the packet of the overlay traffic and no entry is found in the flow table, the flow table has the lowest priority only with the input port information. 9. The network node according to claim 2, wherein an entry is hit and the packet is forwarded to the underlay forwarding function.   3. The overlay flow monitoring function is characterized in that the information of the outer head, which is the processing content when the packet of the overlay traffic that arrives first, is sequentially processed, the encapsulation, and the determination of the output port are recorded. The network node according to claim 9.   In the overlay flow monitoring function, after the information of the outer head is removed, the encapsulation, and the determination of the output port are recorded, the overlay network key and the inner IP header information are used as reference keys, and in units of the reference keys. The network node according to claim 10, wherein the order of removal of information of the outer head, encapsulation, and determination of an output port is recorded.   The overlay flow monitoring function removes the outer head information in units of the reference key, records the order of encapsulation and output port determination, and then adds a new header to the higher priority overlay flow table above the flow table. The network node according to claim 11, wherein a feed and processing contents corresponding to the feed are registered.   The tunnel switch refers to the overlay flow table based on a search key extracted from the overlay traffic packet, and processes the overlay traffic packet in the order of the processing contents recorded in the hit entry. The network node according to claim 12. A packet of initial overlay traffic is transferred by the overlay routing function and the forwarding function on the virtual node;
With respect to the overlay traffic packet after the first overlay traffic packet, the processing contents of the overlay traffic packet are determined by referring to the entry in the overlay flow table, and the tunnel switch is determined based on the processing content. 14. The network node according to claim 13, wherein the network node is processed in the network.   A network comprising the network node according to any one of claims 1 to 14. In this tunnel switching method, a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes processes overlay traffic. And
The network node is provided with a tunnel switch for identifying whether external traffic is general traffic or the overlay traffic, and a flow table that records packet header information and processing contents of the packet in association with each other as a flow entry,
The network node performs an overlay flow monitoring process of monitoring an input port in the tunnel switch and a processing content for the packet of the overlay traffic in the tunnel switch and recording the monitoring result in the flow table;
A tunnel switching method in which the tunnel switch processes a packet of the overlay traffic in the tunnel switch based on a reference result of the flow table using header information extracted from the packet of the overlay traffic as a search key . In this tunnel switching method, a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes processes overlay traffic. And
The network node is provided with a tunnel switch that identifies whether traffic from the outside is the general traffic or the overlay traffic, and a flow table that records packet header information and processing of the packet as a flow entry in association with each other,
Each of the plurality of virtual nodes is provided with an overlay routing function and an overlay forwarding function for processing the overlay traffic,
The network node monitors an underlay routing process and an underlay forwarding process for processing general traffic, an input port in the tunnel switch, and a processing content for a packet of the overlay traffic in the tunnel switch and An overlay flow monitoring process for recording a monitoring result in the flow table;
The tunnel switch refers to the flow table using a search key of header information extracted from the overlay traffic packet after the first overlay traffic packet, and based on the reference result, the tunnel switch refers to the overlay traffic packet. A tunnel switching method characterized by processing in a network.   The overlay flow monitoring process monitors an input port in the tunnel switch and records an overlay network key of the overlay traffic and an inner IP (Internet Protocol) header information as a reference key. 18. The tunnel switching method according to 17.   The tunnel switch refers to the flow entry registered in the flow table using a search key extracted from the header information of the overlay traffic packet, and based on the reference result, the header information of the overlay traffic packet header information 19. The tunnel switching method according to claim 17, wherein the outer header is removed, and the packet from which the outer header is removed is transferred to a destination virtual node.   The tunnel switch refers to the flow entry registered in the flow table using a search key extracted from the header information of the packet of the overlay traffic, and determines the local address of the tunnel of the virtual link from the information of the referred input port. The tunnel switching method according to any one of claims 17 to 19, wherein a remote address is acquired, added as a new outer header to the packet of the overlay traffic, and transferred to the underlay forwarding process.   21. The tunnel switching method according to claim 20, wherein, in the underlay forwarding process, an overlay traffic packet to which the new outer header is added is transmitted to the outside based on information on the outer header.   The header field is recorded using the header information extracted from the packet in the flow table, and the processing content of the packet is recorded in association with each header field. The tunnel switching method according to any one of the above.   Referring to the flow entry of the flow table using the outer header extracted from the packet of the overlay traffic arriving from the tunnel switch, the key of the overlay network, and the IP header information of the inner as a search key The tunnel switching method according to any one of claims 17 to 22, wherein a packet of the overlay traffic is processed according to a processing content corresponding to.   If the flow table is referenced based on the search key extracted from the packet of the overlay traffic and no entry is found in the flow table, the flow table has the lowest priority only with the input port information. The tunnel switching method according to any one of claims 17 to 23, wherein an entry is hit and the packet is transferred to the underlay forwarding function.   18. In the overlay flow monitoring process, the information of the outer head, which is the processing content when the overlay traffic packet that arrives first, is sequentially processed, the encapsulation, and the determination of the output port are recorded. The tunnel switching method according to claim 24.   In the overlay flow monitoring process, after the information of the outer head is removed, the encapsulation, and the determination of the output port are recorded, the key of the overlay network and the IP header information of the inner are used as reference keys. 26. The tunnel switching method according to claim 25, wherein the order of removal of information of the outer head, encapsulation, and determination of an output port is recorded.   In the overlay flow monitoring process, after the information of the outer head is removed, encapsulated, and the order of determining the output port is recorded in units of the reference key, a new header is added to the higher priority overlay flow table in the flow table. 27. The tunnel switching method according to claim 26, wherein a feed and processing contents corresponding to the feed are registered.   The tunnel switch refers to the overlay flow table based on a search key extracted from the overlay traffic packet, and processes the overlay traffic packet in the order of processing contents recorded in the hit entry. The tunnel switching method according to claim 27. A packet of initial overlay traffic is transferred by the overlay routing process and the forwarding process on the virtual node;
With respect to the overlay traffic packet after the first overlay traffic packet, the processing contents of the overlay traffic packet are determined by referring to the entry in the overlay flow table, and the tunnel switch is determined based on the processing content. 29. The tunnel switching method according to claim 28, wherein the tunnel switching method is performed in a network. While constructing an underlay network, a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes is executed by a central processing unit in the network node that processes the overlay traffic. A program to
The network node is provided with a tunnel switch for identifying whether external traffic is general traffic or the overlay traffic, and a flow table that records packet header information and processing contents of the packet in association with each other as a flow entry,
An overlay flow monitoring process for monitoring an input port in the tunnel switch and a processing content for the packet of the overlay traffic in the tunnel switch and recording the monitoring result in the flow table;
A program causing the tunnel switch to process a packet of the overlay traffic in the tunnel switch based on a reference result of the flow table using header information extracted from the packet of the overlay traffic as a search key. While constructing an underlay network, a plurality of overlay networks are constructed using a plurality of virtual nodes created in the own node, and each of the plurality of virtual nodes is executed by a central processing unit in the network node that processes the overlay traffic. A program to
The network node is provided with a tunnel switch that identifies whether traffic from the outside is the general traffic or the overlay traffic, and a flow table that records packet header information and processing of the packet as a flow entry in association with each other,
Each of the plurality of virtual nodes is provided with an overlay routing function and an overlay forwarding function for processing the overlay traffic,
Underlay routing processing and underlay forwarding processing for processing general traffic, input port in the tunnel switch and processing contents of the overlay traffic packet in the tunnel switch are monitored and the monitoring result is the flow Including overlay flow monitoring to record in the table,
The tunnel switch is made to refer to the flow table using a search key of header information extracted from the overlay traffic packet after the first overlay traffic packet, and the overlay traffic packet is referred to the tunnel switch based on the reference result A program characterized by being processed in a program.

Images