JP2016048850A - Packet transfer device, packet transfer system and packet transfer method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a packet transfer device of which the fault monitoring performance is improved, a packet transfer system and a packet transfer method.SOLUTION: The packet transfer device includes a route processing part and a monitor processing part. The route processing part exchanges route information of a packet by establishing a communication session based on a first protocol with another device, transfers the packet on the basis of the exchanged route information and, when a request message requesting the establishment of a communication session based on the first protocol is received from the other device, transmits the route information to the other device. The monitor processing part starts fault monitor of a communication with the other device by establishing a communication session based on a second protocol with the other device after the route processing part is made wait transmission processing of the route information in the case where fault monitor setting included in the request message is effective, and makes the route processing part transmit the route information after the fault monitor is started.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a packet transfer device, a packet transfer system, and a packet transfer method.

  As the network scale increases, the importance of packet routing between networks, that is, routing performance, has increased. The router transfers a packet according to a routing protocol such as BGP (Border Gateway Protocol) and OSPF (Open Shortest Path First) (for example, refer to Patent Documents 1 and 2). Note that BGP and OSPF are defined in RFCs (Request For Comments) 4271 and 3328, respectively.

JP 2012-175425 A JP 2002-359638 A

  When a router starts up or detects a link with an adjacent node, it exchanges packet routing information with the adjacent node router according to the routing protocol, and then sends and receives monitoring packets at regular intervals. To monitor communication failures. For example, in the case of BGP, keep alive packets are transmitted and received at intervals of 60 seconds (default value), and in the case of OSPF, hello packets are transmitted and received at intervals of 10 seconds (default value).

  As described above, in the routing protocol, since the time interval of failure monitoring is defined in seconds, there is a problem that it takes a long time from the occurrence of a communication failure to the detection.

  Therefore, in order to detect a communication failure early, for example, a BFD (Bidirectional Forwarding Detection) protocol that monitors a communication failure by transmitting and receiving hello packets at intervals of several hundred milliseconds may be used together with a routing protocol. In this case, the router first establishes a communication session based on the routing protocol, and establishes a communication session based on the BFD protocol after the route of the packet in the network is determined. BFD is defined in RFC5880.

  However, if a communication failure occurs before the establishment of a communication session based on the BFD protocol, the communication failure is monitored at a long time interval based on the routing protocol, so it still takes a long time from the occurrence of the communication failure to the detection. . At this time, since the route information has been exchanged between the routers according to the routing protocol, each router starts transmission of the user packet.

  For this reason, many user packets are discarded in a long time from when a communication failure is detected by the failure monitoring according to the routing protocol until the route of the user packet is switched to a route that bypasses the failure. May decrease.

  Accordingly, the present invention has been made in view of the above problems, and an object thereof is to provide a packet transfer device, a packet transfer system, and a packet transfer method with improved fault monitoring performance.

  The packet transfer device described in this specification exchanges packet route information by establishing a communication session based on the first protocol with another device, and transfers the packet based on the exchanged route information. In the packet transfer device, when a request message requesting establishment of a communication session based on the first protocol is received from the other device, a route processing unit that transmits the route information to the other device, and included in the request message When the fault monitoring setting is valid, after waiting for the route processing unit to transmit the route information, the communication device based on the second protocol is established with the other device to establish a communication session with the other device. And a monitoring processing unit that causes the path processing unit to transmit the path information after the start.

  The packet transfer system described in the present specification exchanges packet route information by establishing a communication session based on a first protocol, and transfers a packet based on the exchanged route information. In the packet transfer system having the second packet transfer device, the first packet transfer device and the second packet transfer device include a failure monitoring setting relating to presence / absence of failure monitoring, and request establishment of a communication session based on the first protocol. When the failure monitoring setting included in the received request message is valid, after waiting for transmission processing of the route information, establishing a communication session based on the second protocol , Communication between the first packet transfer device and the second packet transfer device. Start the fault monitoring, transmitting and receiving together the path information after the start.

  The packet transfer method described in this specification exchanges packet route information by establishing a communication session based on the first protocol with another device, and transfers the packet based on the exchanged route information. In the packet transfer method, when a request message requesting establishment of a communication session based on the first protocol is received from the other device, and the failure monitoring setting included in the request message is valid, the route to the other device After waiting for the information transmission process, a communication session based on the second protocol is established with the other device, thereby starting communication failure monitoring with the other device. A method for transmitting the route information.

  Improve fault monitoring performance.

It is a block diagram which shows an example of a packet transfer system. It is a sequence diagram which shows the communication between routers in a comparative example. It is a sequence diagram which shows the communication between routers in an Example. It is a figure which shows an example of the operation | movement at the time of reception of a BGP OPEN message. It is a block diagram which shows an example of a BGP OPEN message. It is a figure which shows an example of the operation | movement at the time of reception of a hello packet. It is a sequence diagram which shows communication between routers when a BFD setting is invalid. It is a block diagram which shows an example of a router. It is a flowchart which shows an example of a path | route process. It is a flowchart which shows an example of the monitoring process of a failure.

  FIG. 1 is a configuration diagram illustrating an example of a packet transfer system. The packet transfer system includes three routers 1a to 1c as an example, but the number of routers is not limited.

  The routers 1a to 1c are examples of packet transfer apparatuses that transfer packets. For example, IP addresses “3.3.3.3”, “3.3.3.1”, and “3.3.3.2” are assigned to the routers 1a to 1c, respectively.

  The routers 1a to 1c are connected to each other via a communication cable such as an optical fiber. The routers 1a to 1c have ports (# 1 to # 3) for transmitting and receiving packets for each connection destination. Examples of packets to be transferred by the routers 1a to 1c include IP (Internet Protocol) packets, Ethernet (registered trademark) frames, and ATM (Asynchronous Transfer Mode) cells.

  The router 1b is connected to the terminal device 2a, the router 1a, and the router 1c via ports # 1 to # 3, respectively. The router 1a is connected to the router 1b and the router 1c via ports # 1 and # 2, respectively. The router 1c is connected to the terminal device 2b, the router 1a, and the router 1b via ports # 1 to # 3, respectively. There is no limitation on the connection form of the routers 1a to 1c.

  Examples of the terminal devices 2a and 2b include a personal computer and a server. As an example, the terminal devices 2a and 2b are assigned IP addresses “1.1.1.1” and “2.2.2.2”, respectively. The terminal devices 2a and 2b transmit / receive user packets to / from each other via the packet transfer system.

  The routers 1a to 1c control packet paths R1 and R2 connecting the terminal devices 2a and 2b based on the routing tables RTa to RTc. The routers 1a to 1c respectively generate or update the routing tables RTa to RTc by exchanging route information based on the routing protocol.

  In the routing tables RTa to RTc, a “destination” indicating a destination IP address of a user packet and a port number (# 1 to # 3) for outputting (transmitting) the user packet are registered in association with each other. . For example, the router 1a transmits a user packet whose destination IP address is “1.1.1.1” to the router 1b via the port # 1, and transmits a user packet whose destination IP address is “2.2.2.2”. The data is transmitted to the router 1c via the port # 2. The port numbers are shown as an example of parameters in the routing tables RTa to RTc, and the routing tables RTa to RTc include other parameters such as a gateway address and a route cost.

  The routers 1a to 1c exchange packet route information by establishing a communication session based on BGP, which is an example of a first protocol. The route information includes, for example, the IP address of the adjacent node and the route cost (distance) of the node. In this embodiment, BGP is given as an example of a routing protocol, but the present invention is not limited to this, and other routing protocols such as OSPF may be used.

  For example, the router 1b notifies the routers 1a and 1c of the IP address “1.1.1.1” and the route cost of the adjacent terminal device 2a. Further, the router 1a adds the IP address “1.1.1.1” notified from the router 1b and the route cost between the routers 1a and 1b to the notified route cost and notifies the router 1c. Further, the router 1c notifies the routers 1a and 1b of the IP address “2.2.2.2” and the route cost of the adjacent terminal device 2b. Furthermore, the router 1a adds the IP address “2.2.2.2” notified from the router 1c and the route cost between the routers 1c and 1a to the notified route cost and notifies the router 1b.

  Thus, the routers 1a to 1c register the IP address in the route information and the port number that received the route information in the “destination” and “port number” of the routing tables RTa to RTc, respectively. The routing tables RTa to RTc exemplify route information regarding routes R1 and R2 connecting the terminal devices 2a and 2b.

  In this example, the router 1c receives the route information including the same IP address “1.1.1.1” from the routers 1b and 1a, respectively, but registers the route information with the smaller route cost in the routing table RTc, The route information is stored as a spare. Further, the router 1b receives the route information including the same IP address “2.2.2.2” from the routers 1a and 1c, respectively, but registers the route information having the smaller route cost in the routing table RTb and the other route information. Is stored as a spare. The route information stored as a spare is described in parentheses of the routing tables RTb and RTc.

  That is, the routers 1a to 1c register route information corresponding to the route R1 having a short distance among the two routes R1 and R2 connecting the terminal devices 2a and 2b in the routing tables RTa to RTc. Thereby, the time required for communication between the terminal devices 2a and 2b is shortened.

  However, when a failure occurs in the route R1, the routers 1a to 1c register the route information corresponding to the other route R2, that is, the route information stored as a spare, in the routing tables RTa to RTc. For example, as indicated by a cross, when a failure occurs in the communication cable between the routers 1b and 1c, the routers 1b and 1c detect the communication failure and change the route R1 between the terminal devices 2a and 2b to the failure. The route is switched to the route R2 having no.

  At this time, the routers 1b and 1c delete the route information related to the route R1 where the failure has occurred from the routing tables RTb and RTc, respectively, and store them as spare route information. More specifically, the routing information of the IP address “2.2.2.2” and the port number “# 3” is deleted from the routing table RTb, and the IP address “1.1.1.1” and the port number “# 3” are deleted from the routing table RTc. The route information is deleted.

  The routers 1b and 1c begin to transfer user packets via the other route R2 by registering the route information (referred to in parentheses) initially stored as spares in the routing tables RTb and RTc. Thereby, communication between the terminal devices 2a and 2b is continued.

  Even if the routers 1b and 1c detect a communication failure, the router 1b and 1c continue to transmit user packets until the route information regarding the route R1 where the failure has occurred is deleted from the routing tables RTb and RTc. The user packet is discarded at the location where the failure occurs. Therefore, in order to reduce the number of discarded user packets, it is important to detect a communication failure early.

  The routers 1b and 1c transmit and receive keep-alive packets at intervals of 60 seconds (default value) according to BGP in order to detect communication failures. Therefore, assuming that the number of keep-alive packets not received until a failure is detected is 3, the failure detection requires 180 seconds (= 60 seconds × 3 times). In the case of OSPF, hello packets are transmitted and received at 10-second intervals (default value). Therefore, assuming that the number of hello packets not received before detecting a failure is 3 times, 30 seconds (= 10 seconds) for detecting the failure. X 3 times).

  Therefore, the routers 1a to 1c use the BFD protocol, which is an example of the second protocol, together with BGP in order to detect a communication failure at an early stage. According to the BFD protocol, it is possible to monitor a communication failure by transmitting and receiving hello packets at intervals of several hundred milliseconds.

  FIG. 2 is a sequence diagram illustrating communication between the routers 1b and 1c in the comparative example. In the comparative example, the routers 1a to 1c start monitoring communication failure by establishing a communication session based on BGP, establishing a communication session based on BFD, after exchanging route information. In this comparative example, communication between the routers 1b and 1c is described as an example, but communication between the routers 1a and 1c and between the routers 1a and 1b is performed in the same manner.

  The routers 1b and 1c are connected to each other to detect a link up (see reference sign SQ1). Note that link up and link down are detected at the ports of the routers 1b and 1c.

  Next, the routers 1b and 1c transmit and receive a BGP OPEN message to each other (see SQ2). The BGP OPEN message is an example of a request message for requesting establishment of a communication session based on BGP. By transmitting and receiving the BGP OPEN message, a communication session based on BGP is established between the routers 1b and 1c (see SQ3).

  Next, the routers 1b and 1c transmit and receive a BGP UPDATE message to each other (see symbol SQ4). The BGP UPDATE message includes route information, and the routers 1b and 1c exchange route information by transmitting and receiving the BGP UPDATE message.

  The routers 1b and 1c generate routing tables RTb and RTc based on the exchanged route information. The routers 1b and 1c start transferring user packets based on the routing tables RTb and RTc (reference SQ5).

  Next, the routers 1b and 1c start monitoring a failure according to BGP (see SQ6). The routers 1b and 1c transmit keep-alive packets at a constant time interval T1 (for example, 60 seconds) (see SQ7). The keep alive packet is an example of a monitoring packet for monitoring a communication failure between the routers 1b and 1c.

  The routers 1b and 1c receive a keep-alive packet from the other routers 1c and 1b, thereby monitoring communication failures with the other routers 1c and 1b. More specifically, the routers 1b and 1c detect a communication failure when the keep-alive packets are not continuously received from the other routers 1c and 1b a predetermined number of times. For this reason, a communication failure between the routers 1b and 1c is easily detected. Note that transmission / reception of keep-alive packets is continued.

  The routers 1b and 1c wait for a predetermined time Tw after transmitting and receiving the BGP UPDATE message, and then transmit and receive a request message for requesting establishment of a communication session based on the BFD protocol (see SQ8). The predetermined time Tw is, for example, a time required for completing learning of the routes R1 and R2 by exchanging route information between all the routers 1a to 1c. By transmitting and receiving the request message, a communication session based on BFD is established between the routers 1b and 1c (see SQ9).

  Next, the routers 1b and 1c start monitoring a failure according to the BFD (see reference SQ10). The routers 1b and 1c transmit hello packets at a constant time interval T2 (for example, several hundred milliseconds, T2 <T1) (see SQ11). The hello packet is an example of a monitoring packet for monitoring a communication failure between the routers 1b and 1c.

  The routers 1b and 1c receive a hello packet from the other router 1c and 1b as a response to the transmitted hello packet, thereby monitoring a failure in communication with the other router 1c and 1b. More specifically, the routers 1b and 1c detect a communication failure when the hello packets are not continuously received from the other routers 1c and 1b a predetermined number of times. For this reason, a communication failure between the routers 1b and 1c is easily detected. In addition, transmission / reception of a hello packet is performed continuously thereafter. In this way, communication between the routers 1b and 1c in the comparative example is performed.

  In this comparative example, since a communication session based on the BFD protocol is established (reference SQ9), failure monitoring is performed at a time interval T2 shorter than the failure monitoring time interval T1 according to BGP. For this reason, it is possible to detect a communication failure at an early stage as compared with the failure monitoring according to BGP.

  However, if a communication failure occurs before the establishment of a communication session based on the BFD protocol, the communication failure is detected at a failure monitoring time interval T1 according to BGP, and thus it takes a long time from the occurrence of the communication failure to the detection. . At this time, since the route information has been exchanged between the routers 1c and 1b according to BGP (see symbol SQ4), the routers 1c and 1b have started transmitting user packets (see symbol SQ5).

  For this reason, many user packets are discarded in a long time from when a communication failure is detected by failure monitoring according to BGP until the route R1 of the user packet is switched to the route R2 that bypasses the failure. The quality may be reduced.

  Therefore, when the routers 1a to 1c of the embodiment receive the BGP OPEN message from the other routers 1a to 1c, the routers 1a to 1c follow the failure monitoring setting in the message before sending the route information to the other routers 1a to 1c. The failure monitoring performance is improved by establishing a BFD protocol communication session with 1a to 1c.

  FIG. 3 is a sequence diagram illustrating communication between the routers 1b and 1c in the embodiment. In this embodiment, communication between the routers 1b and 1c will be described as an example, but communication between the routers 1a and 1c and between the routers 1a and 1b is performed in the same manner.

  The routers 1b and 1c are connected to each other to detect a link-up (see symbol SQ21). Next, the routers 1b and 1c transmit and receive a BGP OPEN message to each other (see SQ22). As a result, a communication session based on BGP is established between the routers 1b and 1c (see SQ23).

  In the present embodiment, the BGP OPEN message includes BFD setting information indicating whether or not the BFD protocol is applied as an example of the failure monitoring setting. The routers 1b and 1c take out the BFD setting information from the received BGP OPEN message, and when the BFD setting is valid, the router 1b and 1c waits for the BGP UPDATE message transmission process (reference SQ24). The operation when the BFD setting is invalid will be described later.

  In this case, the routers 1b and 1c record the transmission source IP address of the BGP OPEN message in information recording means such as a memory. FIG. 4 shows an example of the operation when receiving the BGP OPEN message.

  Upon receiving the BGP OPEN message including the source IP address “3.3.3.2”, the destination IP address “3.3.3.1”, and the BFD setting information from the other router 1 c, the router 1 b receives the source IP address “3.3.3.2”. 3.3.3.2 "is recorded in the peer address table ATa. On the other hand, when receiving the BGP OPEN message including the source IP address “3.3.3.1”, the destination IP address “3.3.3.2”, and the BFD setting information from the other router 1b, the router 1c receives the source IP address “3.3.3.1”. Record the address “3.3.3.1” in the peer address table ATc. The IP addresses recorded in the peer address tables ATb and ATc are used for releasing the transmission of the BGP UPDATE message. The IP address of the transmission source is an example of first transmission source information indicating the transmission source of the BGP OPEN message.

  FIG. 5 is a configuration diagram showing an example of the BGP OPEN message. The BGP OPEN message includes “Version”, “My Autonomous System”, “Hold time”, “BGP Identifier”, “Opt Parm Len”, and an optional parameter area. The optional parameter area includes one or more sets of “Parm. Type”, “Parm. Length”, and “Parameter Value”.

  “Version” indicates a version of BGP, and “My Autonomous System” indicates an AS (Autonomous System) number to which the routers 1a to 1c belong. The AS number is an individual identification number assigned by IANA (Internet Assigned Numbers Authority) for each organization that performs route control in the Internet.

  “Hold time” is used to calculate the value of the hold timer, which is the maximum value of the reception interval of consecutive keep-alive packets. “BGP Identifier” is an identifier of BGP determined at the time of activation. “Opt Parm Len” indicates the number of octets (bytes) in the option parameter area.

  “Parm. Type”, “Parm. Length”, and “Parameter Value” in the option parameter area indicate the type, length, and value of the parameter, respectively. “Parameter Value” has a variable data amount corresponding to “Parm. Length”. “Parameter Value” includes “Capability code”, “Capability length”, and “Capability Value” indicating the function, length, and value, respectively.

  The routers 1a to 1c use, for example, “Parm. Type” as BFD setting information. In this case, as the value of “Parm. Type”, BFD setting information can be indicated using 3 to 255 which are unassigned values in RFC4271. Further, when the assigned value 2 is used as the “Parm. Type” value, the routers 1a to 1c use the unassigned values 6 to 63 as the “Capability code” value. BFD setting information may be indicated.

  Referring to FIG. 3 again, after waiting for the BGP UPDATE message transmission process (reference SQ24), the routers 1b and 1c transmit and receive request messages for requesting establishment of a communication session based on the BFD protocol (reference SQ25). . As a result, a communication session based on the BFD is established between the routers 1b and 1c (see SQ26).

  Next, the routers 1b and 1c start failure monitoring according to the BFD (see symbol SQ27). The routers 1b and 1c transmit hello packets at a constant time interval T2 (see SQ28). In addition, transmission / reception of a hello packet is performed continuously thereafter.

  Next, when the IP address of the transmission source of the hello packet received by the communication session based on the BFD protocol matches the IP address recorded in the peer address tables ATb and ATc, the routers 1b and 1c send the BGP UPDATE message to each other. Transmission / reception is performed (reference SQ29). FIG. 6 shows an example of the operation when a hello packet is received.

  Upon receiving the hello packet including the source IP address “3.3.3.2” and the destination IP address “3.3.3.1” from the other router 1c, the router 1b peers with the source IP address “3.3.3.2”. The IP address “3.3.3.2” in the address table ATa is checked. On the other hand, when the router 1c receives the hello packet including the source IP address “3.3.3.1” and the destination IP address “3.3.3.2” from the other router 1b, the source IP address “3.3.3.1” is received. Is compared with the IP address “3.3.3.1” in the peer address table ATc. The source IP address is an example of second source information indicating the source of the hello packet.

  As a result of the collation, the routers 1b and 1c release the waiting state for transmission of the BGP UPDATE message because the IP address of the transmission source of the hello packet matches the IP address recorded in the peer address tables ATb and ATc. Thereby, the BGP UPDATE message is transmitted and received (reference SQ29), and the route information is exchanged between the routers 1b and 1c. On the other hand, if the IP addresses do not match as a result of the collation, the routers 1b and 1c maintain the BGP UPDATE message transmission standby state, so that the BGP UPDATE message is not transmitted.

  In this way, since the routers 1b and 1c collate the IP addresses of the transmission sources with the peer address tables ATb and ATc, it is easy for the routers 1c and 1b that are the destinations of the BGP UPDATE message that has been waiting to start the failure monitoring. Can be detected. Therefore, the routers 1b and 1c can easily detect the release timing of the transmission wait state of the BGP UPDATE message.

  The routers 1b and 1c generate routing tables RTb and RTc based on the exchanged route information. The routers 1b and 1c start transferring user packets based on the routing tables RTb and RTc (reference SQ30).

  Next, the routers 1b and 1c start monitoring a failure according to BGP (see reference SQ31). Routers 1b and 1c transmit keep-alive packets at a constant time interval T1 (see SQ32). Note that transmission / reception of keep-alive packets is continued.

  As described above, the routers 1b and 1c transmit the BGP OPEN message that includes the BFD setting regarding the presence / absence of failure monitoring according to the BFD protocol and requests establishment of a communication session based on BGP to each other. The routers 1b and 1c receive the BGP OPEN message from the other routers 1c and 1b. When the BFD setting included in the received BGP OPEN message is valid, the routers 1b and 1c perform route information transmission processing to the other routers 1c and 1b. Wait.

  Thereafter, the routers 1b and 1c establish a communication session based on the BFD with the other routers 1c and 1b to start monitoring the mutual communication failure, and then route to the other routers 1c and 1b. Send information.

  Therefore, if the BFD settings of the other routers 1c and 1b are valid, the routers 1b and 1c start fault monitoring in accordance with the BFD protocol before starting the transfer of user packets based on the route information after the exchange. be able to. Therefore, the routers 1b and 1c can detect a failure in accordance with the BFD protocol from the time when the transfer of the user packet is started.

  The routers 1b and 1c transmit and receive hello packets at a time interval T2 shorter than the failure monitoring by the communication session based on BGP in the failure monitoring by the communication session based on the BFD protocol. Therefore, the routers 1b and 1c can detect a communication failure at an early stage.

  In addition, when the BFD setting is invalid, the routers 1b and 1c monitor a communication failure with the other routers 1c and 1b by establishing a communication session based on BGP. Therefore, the routers 1b and 1c can perform failure monitoring even when the BFD setting is invalid.

  FIG. 7 is a sequence diagram illustrating communication between routers when the BFD setting is invalid. In this embodiment, communication between the routers 1b and 1c will be described as an example, but communication between the routers 1a and 1c and between the routers 1a and 1b is performed in the same manner.

  The routers 1b and 1c are connected to each other to detect a link up (see symbol SQ41). Next, the routers 1b and 1c transmit and receive a BGP OPEN message to each other (see SQ42). As a result, a communication session based on BGP is established between the routers 1b and 1c (see SQ43).

  The routers 1b and 1c extract BFD setting information from the received BGP OPEN message. When the BFD setting is invalid, the routers 1b and 1c transmit and receive a BGP UPDATE message to each other (reference SQ44). The routers 1b and 1c generate routing tables RTb and RTc based on the exchanged route information. The routers 1b and 1c start transferring user packets based on the routing tables RTb and RTc (reference SQ45).

  Next, the routers 1b and 1c start monitoring a failure according to BGP (see SQ46). Routers 1b and 1c transmit keep-alive packets at a constant time interval T1 (see SQ47). Note that transmission / reception of keep-alive packets is continued. Thus, in the embodiment, communication between the routers 1b and 1c is performed.

  Next, the configuration of the routers 1a to 1c will be described.

  FIG. 8 is a configuration diagram illustrating an example of the routers 1a to 1c. The routers 1a to 1c include a CPU (Central Processing Unit) 10, a ROM (Read Only Memory) 11, a RAM (Random Access Memory) 12, a parameter storage memory 13, a switching circuit 14, and a plurality of ports (# 1 to #N ( N: an integer of 2 or more)) 15.

  The CPU 10 is connected to a ROM 11, a RAM 12, a parameter storage memory 13, and a switching circuit 14 via a data bus 16 so that signals can be input and output with each other. The ROM 11 stores a program for driving the CPU 10. The RAM 12 functions as a working memory for the CPU 10.

  The plurality of ports 15 include, for example, a PHY (Physical layer) chip, and transmit and receive packets to and from the other routers 1a to 1c. The switching circuit 14 is connected to a plurality of ports 15, outputs a packet input from the CPU 10 to a designated port 15, and outputs a packet input from each port 15 to the CPU 10.

  The parameter storage memory 13 is an example of a storage unit that records information, and stores operation setting information 130, a routing table 131, and a peer address table 132. The operation setting information 130 is information relating to the operation of the routers 1a to 1c, and includes, for example, BFD information indicating whether the BFD protocol is supported, that is, whether the BFD setting is valid or invalid. The parameter storage memory 13 may be a non-volatile memory such as a flash memory.

  The routing table 131 is as described with reference to the routing tables RTa to RTc in FIG. The peer address table 132 is as described with reference to FIGS.

  When the CPU 10 reads a program from the ROM 11, a monitoring processing unit 100, a path processing unit 101, and a packet processing unit 102 are formed as functions. The packet processing unit 102 discriminates the content of the packet received via each port 15 and outputs it to the monitoring processing unit 100 and the route processing unit 101. The packet processing unit 102 identifies a user packet, a BGP OPEN message, a BGP UPDATE message, and the like, and notifies the monitoring processing unit 100 and the route processing unit 101 of them.

  The route processing unit 101 generates a BGP OPEN message. At this time, the path processing unit 101 refers to the operation setting information 130 and, when the BFD setting is valid, adds the BFD setting information to the BGP OPEN message. The generated BGP OPEN message is transmitted via the port 15 to the routers 1a to 1c of the adjacent nodes.

  Further, when receiving the BGP OPEN message from the other routers 1a to 1c, the route processing unit 101 generates a BGP UPDATE message including the route information. The path processing unit 101 transmits a BGP UPDATE message to the routers 1 a to 1 c that are the transmission sources of the BGP OPEN message via the port 15.

  When receiving the BGP UPDATE message from the other routers 1a to 1c, the route processing unit 101 updates the routing table 131 based on the route information included in the received BGP UPDATE message. The route processing unit 101 starts transferring user packets based on the updated routing table 131. That is, the route processing unit 101 establishes a BGP-based communication session with the other routers 1a to 1c, thereby exchanging the route information of the user packet, and transfers the user packet based on the exchanged route information. .

  For example, when the transfer of the user packet is started, the path processing unit 101 changes the completion flag from “0” to “1” to notify the monitoring processing unit 100 of the start of the transfer of the user packet.

  The monitoring processing unit 100 performs failure monitoring of communication with the other routers 1a to 1c in accordance with at least one of the BGP and BFD protocols. When the BFD setting included in the BGP OPEN message received from the other routers 1 a to 1 c is valid, the monitoring processing unit 100 records the IP address of the transmission source of the BGP OPEN message in the peer address table 132.

  In addition, when the BFD setting included in the BGP OPEN message received from the other routers 1a to 1c is valid, the monitoring processing unit 100 causes the route processing unit 101 to wait for a BGP UPDATE message, that is, route information transmission processing. For example, the monitoring processing unit 100 instructs the route processing unit 101 to wait for transmission processing by setting the standby flag to “1”.

  The monitoring processing unit 100 causes the route processing unit 101 to wait for the BGP UPDATE message transmission processing, and then establishes a communication session based on the BFD protocol with the other routers 1a to 1c, so that the other routers 1a to 1 The trouble monitoring of communication with 1c is started. More specifically, the monitoring processing unit 100 establishes a communication session by transmitting / receiving a BFD session establishment request message to / from other routers 1a to 1c. In addition, the monitoring processing unit 100 starts monitoring a communication failure with the other routers 1a to 1c by transmitting and receiving hello packets to and from the other routers 1a to 1c.

  After the failure monitoring is started, the monitoring processing unit 100 checks the IP address of the transmission source of the hello packet received from the other routers 1a to 1c by the communication session based on the BFD protocol with the IP address recorded in the peer address table 132. . If the IP addresses match as a result of the collation, the monitoring processing unit 100 causes the route processing unit 101 to transmit a BGP UPDATE message. At this time, for example, the monitoring processing unit 100 instructs the route processing unit 101 to cancel the standby state of the transmission processing by setting the standby flag to “0”. As a result, the route processing unit 101 starts transferring the user packet.

  As described above, the monitoring processing unit 100 starts monitoring a failure of communication with the other routers 1a to 1c before the route processing unit 101 exchanges route information with the other routers 1a to 1c. Therefore, the monitoring processing unit 100 can perform failure monitoring at a time interval T2 shorter than the time interval T1 for failure monitoring by BGP from the start of user packet transfer. Therefore, the routers 1a to 1c can detect a communication failure with the other routers 1a to 1c at an early stage and reduce the number of discarded packets.

  When the start of the transfer of the user packet is notified from the route processing unit 101, the monitoring processing unit 100 starts monitoring a failure of communication with the other routers 1a to 1c according to BGP. More specifically, the monitoring processing unit 100 starts monitoring the communication failure with the other routers 1a to 1c by transmitting and receiving keepalive packets to and from the other routers 1a to 1c.

  However, as described above, the monitoring processing unit 100 transmits and receives hello packets to and from the other routers 1a to 1c at a time interval T2 shorter than the failure monitoring by the communication session based on BGP in the failure monitoring by the communication session based on the BFD. Therefore, the failure monitoring according to the BFD protocol can detect the failure earlier than the failure monitoring according to BGP. Note that, when reducing the amount of traffic between the routers 1a to 1c, the monitoring processing unit 100 may perform only failure monitoring according to the BFD protocol among the BGP and BFD protocols.

  Further, when the BFD setting included in the BGP OPEN message received from the other routers 1a to 1c is invalid, the monitoring processing unit 100 does not establish a BFD protocol communication session with the other routers 1a to 1c. For this reason, the monitoring processing unit 100 monitors the communication failure with the other routers 1a to 1c in accordance with BGP without recording the IP address in the peer address table 132 and instructing standby for the transmission process of the BGP UPDATE message. Start.

  Thus, even when the other routers 1a to 1c do not support the BFD protocol, the monitoring processing unit 100 can detect a failure by performing the failure monitoring according to BGP.

  FIG. 9 is a flowchart illustrating an example of route processing. The path processing unit 101 determines whether or not a link up with the other routers 1a to 1c has been detected (step St1). The link-up is detected by each port 15 and notified to the route processing unit 101 via the bus 16.

  The path processing unit 101 ends the process when no link up is detected (No in Step St1). When the path processing unit 101 detects a link up (Yes in step St1), the path processing unit 101 sets a completion flag to “0” (step St2). As described above, the completion flag indicates to the monitoring processing unit 100 that processing is not completed when it is “0”, and indicates completion of processing when it is “1”.

  Next, the path processing unit 101 generates a BGP OPEN message (step St3). The BGP OPEN message is formed in the format shown in FIG. 5, for example, but is not limited thereto.

  Next, the path processing unit 101 refers to the operation setting information 130 and determines whether or not the BFD setting is valid (step St4). When the BFD setting is valid (Yes in step St4), the route processing unit 101 adds BFD setting information to the generated BGP OPEN message (step St5). Further, when the BFD setting is invalid (No in Step St4), the path processing unit 101 performs the process in Step St6 described below without adding the BFD setting information.

  Next, the path processing unit 101 transmits a BGP OPEN message to the other routers 1a to 1c via the port 15 (step St6). Next, the path processing unit 101 determines whether or not a BGP OPEN message has been received from the destination routers 1a to 1c (step St7). If the BGP OPEN message has not been received (No in step St7), the path processing unit 101 executes the process in step St7 again.

  When the route processing unit 101 receives the BGP OPEN message (Yes in step St7), the route processing unit 101 generates a BGP UPDATE message including the route information (step St8). The BGP UPDATE message is generated, for example, in a format according to the contents specified in RFC4271.

  Next, the path processing unit 101 determines whether or not the standby flag is “1” (step St9). As described above, the standby flag is used by the monitoring processing unit 100 to instruct the standby of the transmission process of the BGP UPDATE message. When the standby flag is “1”, it means an instruction to put the BGP UPDATE message transmission process in a standby state, and when it is “0”, it means an instruction to cancel the standby state of the BGP UPDATE message transmission process.

  When the standby flag is “1” (Yes in step St9), the route processing unit 101 executes the process in step St9 again. That is, the path processing unit 101 sets the BGP UPDATE message transmission process to the standby state while the standby flag is “1”.

  Further, when the standby flag is “0” (No in Step St9), the path processing unit 101 transmits the generated BGP UPDATE message (Step St10). That is, when the standby flag becomes “0”, the path processing unit 101 cancels the standby state of the BGP UPDATE message transmission process.

  Next, the path processing unit 101 determines whether or not a BGP UPDATE message has been received from the other routers 1a to 1c of the transmission destination in step St6 (step St11). If the BGP UPDATE message has not been received (No in step St11), the path processing unit 101 executes the process in step St11 again.

  Further, when the route processing unit 101 receives the BGP UPDATE message (Yes in Step St11), the route processing unit 101 updates the routing table 131 based on the route information included in the BGP UPDATE message (Step St12). Next, the route processing unit 101 starts transferring the user packet in accordance with the updated routing table 131 (step St13). As described above, when the route processing unit 101 receives the BGP OPEN message from the other routers 1a to 1c, the route processing unit 101 transmits the BGP UPDATE message to the routers 1a to 1c. Can exchange packets.

  Next, the path processing unit 101 sets the completion flag to “1” (step St14) and ends the process. In this way, route processing is performed.

  FIG. 10 is a flowchart illustrating an example of a failure monitoring process. The monitoring processor 100 determines whether a BGP OPEN message has been received from the other routers 1a to 1c (step St21). If the BGP OPEN message is not received (No in step St21), the monitoring processing unit 100 ends the process.

  Further, when the BGP OPEN message is received (Yes in step St21), the monitoring processing unit 100 determines whether or not the BFD setting in the BGP OPEN message is valid (step St22). When the BFD setting is invalid (No in Step St22), the monitoring processing unit 100 executes a process in Step St33 described later.

  If the BFD setting is valid (Yes in step St22), the monitoring processing unit 100 sets the standby flag to “1” (step St23). As a result, the monitoring processing unit 100 causes the route processing unit 101 to wait for the BGP UPDATE message transmission processing, and establishes a communication session based on the BFD protocol as described below.

  Next, as described with reference to FIG. 4, the monitoring processing unit 100 records the IP address of the transmission source of the received BGP OPEN message in the peer address table 132 (step St24). Next, the monitoring processing unit 100 generates a BFD session establishment request message (step St25).

  Next, the monitoring processing unit 100 transmits a BFD session establishment request message to the other routers 1a to 1c that have transmitted the BGP OPEN message (step St26). Next, the monitoring processor 100 determines whether or not a BFD session establishment request message has been received from the other routers 1a to 1c (step St27). When the BFD session establishment request message has not been received (No at Step St27), the monitoring processing unit 100 executes the process at Step St27 again.

  When the monitoring processing unit 100 receives a BFD session establishment request message (Yes in step St27), the monitoring processing unit 100 starts transmitting a hello packet to the other routers 1a to 1c based on the BFD protocol (step St28). Next, the monitoring processor 100 determines whether or not a hello packet has been received from the other routers 1a to 1c (step St29). When the hello packet is not received (No at Step St29), the monitoring processing unit 100 executes the process at Step St29 again.

  When receiving the hello packet (Yes in step St29), the monitoring processing unit 100 checks the IP address of the transmission source of the hello packet with the peer address table 132 as described with reference to FIG. 6 (step St30). . If the IP addresses do not match as a result of the collation (No in step St31), the monitoring processing unit 100 executes the process in step St29 again.

  If the IP addresses match as a result of the collation (Yes in step St31), the monitoring processing unit 100 deletes the IP address from the peer address table 132 (step St32). Next, the monitoring processor 100 sets the standby flag to “0” (step St33).

  As a result, since the standby state of the BGP UPDATE message transmission processing is released, the route processing unit 101 can exchange route information with the other routers 1a to 1c and start transferring user packets. At this time, since the monitoring processing unit 100 has already started the failure monitoring according to the BFD protocol, even if a communication failure occurs with the other routers 1a to 1c, the monitoring processing unit 100 can detect the failure early. .

  Next, the monitoring processing unit 100 determines whether or not the completion flag is “1” (step St34). When the completion flag is “0” (No in Step St34), the monitoring processing unit 100 executes the process in Step St34 again.

  Further, when the completion flag is “1” (Yes in Step St34), the monitoring processing unit 100 starts transmission of keepalive packets to the other routers 1a to 1c based on BGP (Step St35). Thereby, failure monitoring according to BGP is performed. In this way, the monitoring process is performed.

  As described above, the packet transfer apparatuses 1a to 1c exchange packet path information by establishing a communication session based on the first protocol with other apparatuses, and packets based on the exchanged path information. Forward. Each of the packet transfer apparatuses 1a to 1c includes a route processing unit 101 and a monitoring processing unit 100.

  When the route processing unit 101 receives a request message for requesting establishment of a communication session based on the first protocol from another device, the route processing unit 101 transmits route information to the other device. When the failure monitoring setting included in the request message is valid, the monitoring processing unit 100 causes the route processing unit 101 to wait for transmission processing of route information, and then establishes a communication session based on the second protocol with another device. By establishing, the failure monitoring of communication with other devices is started. The monitoring processing unit 100 causes the route processing unit 101 to transmit route information after the start.

  According to the above configuration, when the route processing unit 101 receives a request message for requesting establishment of a communication session based on the first protocol from another device, the route processing unit 101 transmits route information to the other device. Packets can be transferred by exchanging route information.

  In addition, when the failure monitoring setting included in the request message is valid, the monitoring processing unit 100 causes the route processing unit to wait for the route information transmission processing and then communicates with another device based on the second protocol. Is established, failure monitoring of communication with other devices is started. The monitoring processing unit 100 causes the route processing unit 101 to transmit route information after starting failure monitoring.

  Therefore, the packet transfer apparatuses 1a to 1c can start monitoring a failure in communication with another device before starting the packet transfer according to the failure monitoring setting. For this reason, the packet transfer apparatuses 1a to 1c can detect a failure according to the second protocol from the time when the packet transfer is started.

  Therefore, according to the packet transfer apparatuses 1a to 1c according to the embodiment, the failure monitoring performance is improved.

  In addition, the packet transfer system according to the embodiment exchanges packet route information by establishing a communication session based on the first protocol, and transfers the packets based on the exchanged route information. 1c and second packet transfer apparatuses 1a to 1c.

  The first packet transfer apparatuses 1a to 1c and the second packet transfer apparatuses 1a to 1c transmit and receive request messages for requesting establishment of a communication session based on the first protocol, including failure monitoring settings relating to presence / absence of failure monitoring. When the failure monitoring setting included in the received request message is valid, the first packet transfer device 1a to 1c and the second packet transfer device 1a to 1c wait for the route information transmission processing and then switch to the second protocol. Establish a communication session based on.

  As a result, the first packet transfer apparatuses 1a to 1c and the second packet transfer apparatuses 1a to 1c start monitoring the communication failure between the first packet transfer apparatuses 1a to 1c and the second packet transfer apparatuses 1a to 1c. The first packet transfer apparatuses 1a to 1c and the second packet transfer apparatuses 1a to 1c transmit / receive path information to / from each other after the failure monitoring is started.

  Since the packet transfer system according to the embodiment includes the same configuration as the above-described packet transfer apparatuses 1a to 1c, the same effects as the above-described contents are obtained.

The packet transfer method according to the embodiment exchanges packet route information by establishing a communication session based on the first protocol with another device, and transfers the packet based on the exchanged route information. The method includes the following steps (1) to (3).
Step (1): A request message for requesting establishment of a communication session based on the first protocol is received from another device.
Step (2): When the failure monitoring setting included in the request message is valid, after waiting for transmission processing of the route information to another device, a communication session based on the second protocol is established with the other device. By doing so, failure monitoring of communication with other devices is started.
Step (3): The route information is transmitted to another device after the start.

  Since the packet transfer method according to the embodiment includes the same configuration as that of the packet transfer apparatuses 1a to 1c, the same effects as those described above can be obtained.

  The above-described embodiment is an example of a preferred embodiment of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the scope of the present invention.

In addition, the following additional notes are disclosed regarding the above description.
(Supplementary note 1) In a packet transfer device that exchanges packet route information by establishing a communication session based on the first protocol with another device, and forwards the packet based on the exchanged route information.
A route processing unit for transmitting the route information to the other device when receiving a request message requesting establishment of a communication session based on the first protocol from the other device;
When the failure monitoring setting included in the request message is valid, the route processing unit is made to wait for the route information transmission processing and then establishes a communication session based on the second protocol with the other device. And a monitoring processing unit that starts monitoring a failure of communication with the other device and causes the route processing unit to transmit the path information after the start.
(Additional remark 2) It further has a recording part which records information,
The monitoring processing unit
When the failure monitoring setting included in the request message is valid, the first transmission source information indicating the transmission source of the request message is recorded in the recording unit,
After starting failure monitoring of communication with the other device, second source information indicating the source of the monitoring packet received from the other device through a communication session based on the second protocol is collated with the first source information. Then, as a result of the collation, when the second transmission source information matches the first transmission source information, the packet processing apparatus according to appendix 1, wherein the path processing unit is configured to transmit the path information.
(Supplementary Note 3) The monitoring processing unit
When the failure monitoring setting is invalid, a communication session based on the first protocol is established, thereby performing communication failure monitoring with the other device,
The failure monitoring by the communication session based on the second protocol transmits / receives a monitoring packet to / from the other device at a shorter time interval than the failure monitoring by the communication session based on the first protocol. Packet transfer device.
(Additional remark 4) It has the 1st packet transfer apparatus and the 2nd packet transfer apparatus which exchange the path | route information of a packet by establishing the communication session based on a 1st protocol, and transfer a packet based on the said path | route information exchanged In the packet transfer system,
The first packet transfer device and the second packet transfer device are:
Including fault monitoring settings relating to the presence or absence of fault monitoring, sending and receiving request messages requesting establishment of a communication session based on the first protocol,
When the failure monitoring setting included in the received request message is valid, after waiting for transmission processing of the route information, establishing a communication session based on the second protocol, the first packet transfer device and Initiating communication failure monitoring between the second packet transfer devices,
A packet transfer system which transmits and receives the path information to each other after the start.
(Supplementary Note 5) The first packet transfer device and the second packet transfer device are:
If the fault monitoring setting included in the request message is valid, record first transmission source information indicating the transmission source of the request message;
After the failure monitoring is started, the second transmission source information indicating the transmission source of the monitoring packet received from the other of the first packet transfer device and the second packet transfer device by the communication session based on the second protocol 1 Check with the sender information,
The packet transfer system according to appendix 4, wherein when the second transmission source information matches the first transmission source information as a result of the collation, the route information is transmitted and received with each other.
(Appendix 6) The first packet transfer device and the second packet transfer device are:
When the fault monitoring setting is invalid, by establishing a communication session based on the first protocol, monitoring communication faults with each other,
The packet transfer system according to appendix 4 or 5, wherein in the failure monitoring by the communication session based on the second protocol, the monitoring packets are transmitted and received with each other at a shorter time interval than the failure monitoring by the communication session based on the first protocol. .
(Supplementary note 7) In a packet transfer method for exchanging packet route information by establishing a communication session based on the first protocol with another device and transferring a packet based on the exchanged route information,
Receiving a request message requesting establishment of a communication session based on the first protocol from the other device;
When the failure monitoring setting included in the request message is valid, after waiting for transmission processing of the route information to the other device, establishing a communication session based on the second protocol with the other device To start a fault monitoring of communication with the other device,
A packet transfer method comprising transmitting the route information to the other apparatus after the start.
(Supplementary note 8) When the fault monitoring setting included in the request message is valid, the first transmission source information indicating the transmission source of the request message is recorded,
After starting failure monitoring of communication with the other device, second source information indicating the source of the monitoring packet received from the other device through a communication session based on the second protocol is collated with the first source information. And
The packet transfer method according to appendix 7, wherein when the second source information matches the first source information as a result of the collation, the route information is transmitted to the other device.
(Supplementary note 9) When the failure monitoring setting is invalid, by establishing a communication session based on the first protocol, the failure monitoring of the communication with the other device is performed,
The failure monitoring by the communication session based on the second protocol transmits / receives a monitoring packet to / from the other device at a shorter time interval than the failure monitoring by the communication session based on the first protocol. Packet transfer method.

1a to 1c router 10 CPU
13 Parameter storage memory 15 Port 100 Monitoring processing unit 101 Route processing unit 130 Operation setting information 131 Routing table 132 Peer address table

Claims (5)

In a packet transfer device that exchanges packet route information by establishing a communication session based on the first protocol with another device, and forwards the packet based on the exchanged route information.
A route processing unit for transmitting the route information to the other device when receiving a request message requesting establishment of a communication session based on the first protocol from the other device;
When the failure monitoring setting included in the request message is valid, the route processing unit is made to wait for the route information transmission processing and then establishes a communication session based on the second protocol with the other device. And a monitoring processing unit that starts monitoring a failure of communication with the other device and causes the route processing unit to transmit the path information after the start. A recording unit for recording information;
The monitoring processing unit
When the failure monitoring setting included in the request message is valid, the first transmission source information indicating the transmission source of the request message is recorded in the recording unit,
After starting failure monitoring of communication with the other device, second source information indicating the source of the monitoring packet received from the other device through a communication session based on the second protocol is collated with the first source information. The packet transfer apparatus according to claim 1, wherein when the second transmission source information matches the first transmission source information as a result of the collation, the path processing unit is caused to transmit the path information. . The monitoring processing unit
When the failure monitoring setting is invalid, a communication session based on the first protocol is established, thereby performing communication failure monitoring with the other device,
The failure monitoring by the communication session based on the second protocol transmits and receives the monitoring packet with the other device at a shorter time interval than the failure monitoring by the communication session based on the first protocol. Packet transfer equipment. In a packet transfer system having a first packet transfer device and a second packet transfer device for exchanging packet route information by establishing a communication session based on a first protocol and transferring packets based on the exchanged route information ,
The first packet transfer device and the second packet transfer device are:
Including fault monitoring settings relating to the presence or absence of fault monitoring, sending and receiving request messages requesting establishment of a communication session based on the first protocol,
When the failure monitoring setting included in the received request message is valid, after waiting for transmission processing of the route information, establishing a communication session based on the second protocol, the first packet transfer device and Initiating communication failure monitoring between the second packet transfer devices,
A packet transfer system which transmits and receives the path information to each other after the start. In the packet transfer method of exchanging packet route information by establishing a communication session based on the first protocol with another device, and transferring the packet based on the exchanged route information,
Receiving a request message requesting establishment of a communication session based on the first protocol from the other device;
When the failure monitoring setting included in the request message is valid, after waiting for transmission processing of the route information to the other device, establishing a communication session based on the second protocol with the other device To start a fault monitoring of communication with the other device,
A packet transfer method comprising transmitting the route information to the other apparatus after the start.

Images