JP2013090218A - Communication system including autonomous detection function of silent failure - Google Patents

Abstract

PROBLEM TO BE SOLVED: To autonomously identify a failure point (abnormality occurrence point) internal of a relay device.SOLUTION: The relay device includes: an addition means for adding a path identifier, uniquely indicating a passage route in a device, to a packet to be transmitted; when the path identifier is extracted from the passing packet, a plurality of monitoring means for respectively notifying passage information including the extracted path identifier and information indicative of a monitoring point in the device; based on the passage information notified from the plurality of monitoring means, a decision means for recording the passage state of the packet into a path table, and for detecting a failure point in the device according to the passage state recorded in the path table.

Description

  The present invention relates to a relay device applied to a communication system including a silent failure autonomous detection function.

In a communication system that performs client-client communication and server-client communication through an IP (Internet Protocol) network such as the Internet or an intranet, a plurality of network relay devices that relay (transfer) data to be transmitted in packet form Routers are arranged on the IP network.

  Normally, in such a communication system, when an abnormality such as packet loss or transfer delay occurs, a failure notification is sent from the corresponding router to the maintenance operation system. A maintenance operator (monitoring person) of the maintenance operation system can know that a specific router cannot transfer a packet based on the failure notification.

  However, even though an abnormality has occurred, the router itself may not be detected due to deterioration or failure, and the maintenance operation system may not be notified of the failure. In this case, the maintenance operator can recognize the failure only when there is an abnormality report from the user after the communication is affected, for example.

  Such a fault that is latent in the communication system and cannot detect the occurrence of abnormality autonomously is called a silent fault. In recent years, there has been a demand for a method for early detection of this silent failure in an IP network.

JP 2007-28305 A

  FIG. 1 illustrates a communication system that employs one method for detecting the silent failure described above. In the communication system 1A, a plurality of probe devices P1, P2, P3, P4, and P5 are arranged at monitoring points between routers R1, R2, R3, and R4 on the IP network 2, respectively.

  The probe devices P1 to P5 detect abnormalities that the routers R1 to R4 cannot detect themselves by monitoring the arrival state of the packet that is the transmission target data. And abnormality notification is performed outside the maintenance operation system 4 etc. by cooperation of probe device P1-P5 and loss spot specific device 3.

  More specifically, as shown in FIG. 2, the probe devices P1 and P2 assign probe device identifiers ID (P1) and ID (P2) to each packet (fragmented packet) as a passage history. Each packet arrives at the loss point identifying device 3 via a plurality of probe devices P1, P2. Loss location identifying device 3 can identify the path of the packet and the location where the loss of the packet has occurred (sometimes simply referred to as the loss location) based on the passage history included in the arriving packet. It is.

  In the example shown in FIG. 2, the probe device P1 arranged between the routers R1 and R2 sequentially receives packets including sequence numbers seq # N, seq # N + 1, seq # N + 2, and seq # N + 3 in the IP header, It transmits after giving probe apparatus identifier ID (P1) to each packet as a passage history.

  The probe device P2 arranged between the routers R2 and R3 receives a packet including the sequence numbers seq # N, seq # N + 2 and seq # N + 3 in the IP header, but the sequence number seq # N + 1 following the sequence number seq # N Since the packet including the sequence number seq # N + 2 is not received, the packet including the sequence number seq # N + 2 is transmitted as the passage history without adding the probe device identifier ID (P2).

  When the loss location identifying device 3 receives the packet transmitted from the probe device P2 via the router R3, the loss location identifying device 3 uses the probe device identifiers ID (P1) and ID (P2) in the passage history given to the arrival (reception) packet. Based on this, the packet path is specified. Here, the loss location identifying device 3 identifies (recognizes) the location where the loss occurred between the probe devices P1 and P2, since the passage history probe device identifier ID (P2) is not assigned to the packet with the sequence number seq # N + 2. Is possible.

  In the detection of the failure by the probe devices P1 and P2 and the loss location specifying device 3 described above, each probe device P1 and P2 gives a passage history to each packet, and the loss location specifying device 3 makes a final pass determination of each packet. Do. For this reason, when each packet is lost in the middle, and the identifier (probe device identifier) indicating the packet loss occurrence location does not reach the loss location identification device 3, the loss occurrence location cannot be identified (see FIG. 3). .

  Further, since the passage monitoring is performed with the sequence numbers seq # N, seq # N + 1, seq # N + 2, and seq # N + 3, the following restrictions (1), (2), and (3) are unavoidable.

  (1) It cannot be applied to retransmission packets. In other words, since the sequence numbers seq # N and seq # N + 2 are not continuous for the retransmitted packet, no probe device identifier is assigned. As a result, the location where the retransmission packet loss occurs cannot be identified (see FIG. 4).

  (2) Not applicable to single packet. That is, for a single packet that is not a plurality of fragmented packets, the probe device identifier cannot be assigned based on the sequence number. Therefore, the location where the loss occurs cannot be specified (see FIG. 5).

(3) Only applicable to communication conforming to the RTP (Real-time Transfer Protocol) protocol. The RTP protocol is a communication protocol for data stream transfer that requires real-time performance such as video and audio distribution.

  In the failure detection method described above, there remains a fundamental problem that the failure location (abnormality occurrence location) inside each router cannot be specified.

  The problem is to provide a technology that makes it possible to autonomously identify a failure location (abnormality occurrence location) inside the relay device.

In order to solve the above-mentioned problem, the relay device adds an additional means for adding a path identifier uniquely indicating a passing route in the device to the transmission target packet; and when the path identifier is extracted from the packet that passes, A plurality of monitoring means for notifying each of passing information including a path identifier and information indicating a monitoring location in the apparatus; and a path table indicating a passing status of the packet based on the passing information notified from the plurality of monitoring means. And determining means for detecting a fault location in the apparatus in accordance with the passing situation recorded in the path table.

  According to the disclosed relay apparatus, a silent failure in the apparatus can be detected autonomously.

  Other objects, features and advantages will become apparent upon reading the detailed description set forth below when taken in conjunction with the drawings and the appended claims.

The figure for demonstrating the related art which detects a silent failure. The figure for demonstrating the related art which detects a silent failure. The figure for demonstrating the problem of related technology. The figure for demonstrating the problem of related technology. The figure for demonstrating the problem of related technology. 1 is a block diagram illustrating a configuration of a communication system according to an embodiment. The block diagram which shows the structure of the router in the communication system of one embodiment. The figure explaining the transmission object packet and path identifier in the router shown in FIG. The flowchart for demonstrating the process in the router shown in FIG. The flowchart for demonstrating the process in the router shown in FIG. The flowchart for demonstrating the process in the router shown in FIG. The figure explaining the path table in the router shown in FIG. The block diagram which shows the structure of the communication system of a modification.

  Hereinafter, further detailed description will be given with reference to the accompanying drawings. The drawings show preferred embodiments. However, it can be implemented in many different forms and is not limited to the embodiments described herein.

[Communications system]
Referring to FIG. 6 showing a system configuration in one embodiment, a communication system 1 including a silent failure autonomous detection function is arranged on an IP network 2 such as the Internet or an intranet, and relays (transfers) transmission target data in a packet form. A plurality of routers R11, R12, R13, and R14 as network relay devices.

  These routers R11 to R14 function as an ingress edge node, a relay node, and an egress edge node, and for example, cause a plurality of communication terminal devices 5 and 6 connected to the IP network 2 to perform communication between clients and clients. . Further, when the server is connected to the IP network 2, the routers R <b> 11 to R <b> 14 cause the communication terminal device 5 to perform server-client communication, for example.

  In this communication system 1, when an abnormality such as a packet loss or a transfer delay occurs inside the router, a failure site notification is sent from the corresponding router (for example, router R12) to the maintenance operation system 4 through the network line. The maintenance operator (monitoring person) of the maintenance operation system 4 recognizes that the specific router cannot transfer the packet based on the failure part notification including the IP address of the router and the trouble part.

As will be described in detail later, in this communication system 1, each of the routers R11 to R14 adds a path identifier indicating a passing route inside the router to a transmission target packet that passes through the router R11 to R14. Monitor and store the packet passage status of each part in the path table. Each router R11-R14 determines the presence / absence of packet transfer in each part based on the information stored in this path table, and autonomously identifies the fault location inside the router when an abnormality occurs.

[Router configuration]
In the communication system 1 shown in FIG. 6, the routers R11 to R14 functioning as network relay devices adopt the detailed configuration shown in FIG.

  As shown in FIG. 7, each router has a plurality (n; n ≧ 2) of input (ingress) interface units IFI1 to IFIn and a plurality (m; m ≧ 2) of output (exit) interface units IFE1 to IFEm. , A switch unit SW, and a control unit CONT. In each router, the input interface units IFI1 to IFIn, the output interface units IFE1 to IFEm, and the switch unit SW are connected to the control unit CONT through the control line C1.

  Each input interface unit IFI1 to IFIn includes an input packet buffer PBI and an input network processor NPI. Each output interface unit IFE1 to IFEm includes an output packet buffer PBO and an output network processor NPO. In FIG. 7, only the detailed configuration of the input interface unit IFIn and the output interface unit IFE1 is shown for simplification.

Each input network processor NPI of the input interface units IFI1 to IFIn includes a destination address extracting unit 11, an identifier assigning unit 12, a packet monitoring unit 13, and an IP forwarding table (FIB: Forwarding Information Base) 14 which is a simple routing table. Each input packet buffer PBI of the input interface units IFI <b> 1 to IFIn includes a packet monitoring unit 15. Here, the FIB 14 is provided in a memory (not shown) of the input network processor NPI. Further, the destination address extracting unit 11, the identifier assigning unit 12, and the packet monitoring units 13 and 15 can be implemented as functional components that perform functions described in detail later by executing a program developed in the memory of the input network processor NPI. It is.

  Each output network processor NPO of the output interface units IFE1 to IFEm includes a packet monitoring unit 21 and an identifier removing unit 22. Each output packet buffer PBO of the output interface units IFE 1 to IFEm includes a packet monitoring unit 23. Here, the packet monitoring units 21 and 23 and the identifier removing unit 22 can be implemented as functional components that perform the functions described in detail later by executing a program developed in a memory (not shown) of the output network processor NPO.

  The switch unit SW includes a switch 41 and a plurality of packet monitoring units 42, 43, 44, 45, and 46. The switch unit SW interconnects the input interface units IFI1 to IFIn and the output interface units IFE1 to IFEm via a switch fabric switch 41 having a cross-connect function. Here, the packet monitoring units 42 to 46 can be implemented as functional components that fulfill the functions described in detail later by executing a program developed in a memory of a processor (not shown) in the switch unit SW.

The control unit CONT includes a routing control unit 51, an identifier assigning unit 52, a packet monitoring unit 53, a routing table 54, a passage determination unit 55, and a path table 56. The control packet buffer PBC of the control unit CONT includes a packet monitoring unit 57. Here, the routing table 54 and the path table 56 are provided in a memory (not shown) of the control unit CONT. Further, the routing control unit 51, the identifier assigning unit 52, the packet monitoring units 53 and 57, and the passage determination unit 55 are functions to be described in detail later by executing a program developed in a memory of a processor (not shown) in the control unit CONT. It can be implemented as a functional component that fulfills.

[Router function]
When a transmission target packet (which may be simply referred to as a packet) arrives at the nth input interface unit IFIn of the router whose detailed configuration is illustrated in FIG. 7, the destination address extraction unit 11 of the input network processor NPI The destination address (IP address) “a” is extracted from the IP header of the packet to be transmitted that has arrived (received), and is notified to the identifier assigning unit 12.

  The identifier assigning unit 12 determines the output interface number “# E1” corresponding to the first output interface unit IFE1 from the stored information of the IP forwarding table (FIB) 14 based on the notified destination address “a”. The output interface identifier “E1” is added to the transmission target packet as a header in the router. In the FIB 14, a correspondence relationship between the destination address and the output interface number is set in advance by an instruction from the control unit CONT.

  Further, the identifier assigning unit 12 adds a path identifier “X” that uniquely indicates the route through which the packet passes within the router to the transmission target packet as an intra-router header. The identifier assigning unit 52 functions in the same manner as the identifier assigning unit 12. Details of this path identifier will be described later.

  The packet monitoring unit 13 monitors a path identifier of a packet that passes through, and notifies the passage determination unit 55 of the control unit CONT of packet passage information. Other packet monitoring units 15, 21, 23, 42 to 46, 53, and 57 existing in each part (monitoring part) in the router similarly monitor the path identifiers of the packets that pass through and determine whether to pass the packet passing information. Notify unit 55.

  The input packet buffer PBI temporarily stores a packet when the packet is transferred to the switch unit SW.

  The switch unit SW switches the packet to the output interface unit IFE1 by the switch 41 in accordance with the output interface identifier “E1” added to the packet. Note that the processor for this switching control is not shown.

  The output packet buffer PBO of the output interface unit IFE1 temporarily stores packets when transferring the packets to the output network processor NPO.

  The identifier removing unit 22 removes the in-router header including the output interface identifier “E1” and the path identifier “X” added to the packet.

  The passage determination unit 55 of the control unit CONT records the packet passage state in the path table 56 based on the notified packet passage information. The path table 56 holds the packet passage status of each monitored location according to an instruction from the passage determination unit 55. In addition, the passage determination unit 55 monitors (refers to) the path table 56, detects a failure inside the router according to the packet passage state, and notifies the maintenance operation system 4 of the failure.

In the router of the communication system 1 according to this embodiment, the routing control unit 51, the identifier assigning unit 52, the packet monitoring unit 53, the routing table 54, and the packet monitoring unit 57 of the control packet buffer PBC are connected with other routers. It functions when exchanging (route) information. In this case, the passage determination unit 55 records the packet passage state in the path table 56 based on the packet passage information notified from the packet monitoring units 53 and 57, etc., and monitors the path table 56 to determine the packet passage state. In response, a failure inside the router is detected and notified to the maintenance operation system 4.

[Path identifier]
As described above, the identifier assigning unit 12 adds a path identifier uniquely indicating a passing (forwarding) route in the router together with the output interface identifier to the transmission target packet as an intra-router header (intra-device header). Here, the details of the path identifier will be described with reference to FIG.

  As shown in FIG. 8, the transmission target packet transferred in the router has a configuration in which a path identifier PID is added to a normal IP packet including an IP header and data (packet data) as a part of the header in the router. The path identifier PID includes fields of version F1, header length F2, router input interface number (input physical port number) F3, router output interface number (output physical port number) F4, and packet number F5.

  In the path identifier PID, the version F1 represents the version of the path identifier with 4 bits. Here, the version F1 indicates whether the transmission target packet that the packet monitoring unit existing at each monitoring location in the router passes is monitored in units of routes (passing routes) or in units of packets for each passing route. The version F1 is notified in advance to the identifier assigning units 12 and 52 by the control unit CONT based on an instruction from the maintenance operation system 4, for example.

  In the path identifier PID, the header length F2 represents the length of the entire path identifier in 1 byte. The router input interface number F3 represents the interface numbers of the input interface units IFI1 to IFIn when the transmission target packet arrives at the router. The output interface number F4 of the router represents the interface number of the output interface units IFE1 to IFEm when the transmission target packet is transmitted from the router.

  Further, in the path identifier PID, the packet number F5 is a sequence number for uniquely indicating a fragmented packet passing through the same route in the router, and can be obtained from the IP header. The fields of the input interface number F3, the output interface number F4, and the packet number F5 are variable length.

  More specifically, the packet monitoring unit existing at each monitoring location in the router monitors the transmission target packet passing based on the version F1 for each passing route, and notifies the passage determining unit 55 of the packet passing information. When the monitoring form is adopted, the identifier assigning unit 12 adds a path identifier PID uniquely indicating a passing route in the router, that is, a set of the input interface number and the output interface number, to the transmission target packet as an intra-router header.

  Accordingly, when the first monitoring mode is adopted, the path identifier PID is given to the path identifier PID by the identifier adding unit 12 and the input interface number F3 (for example, F3 = # In = 5) and the output interface number F4 (for example, F4 = # E1 = 1). ) Is set, but the packet number F5 is not set. In the path identifier field of the path table 56, “51” that uniquely indicates the set of the input interface number F3 and the output interface number F4 is recorded.

  In addition, when each packet monitoring unit adopts the second monitoring mode in which the transmission target packet passing based on version F1 is monitored for each passing route and the packet determination information is notified to the passage determining unit 55, an identifier is added. The unit 12 is a packet that is a sequence number for uniquely indicating a fragmented packet that passes through the same route in the router, in addition to information that uniquely indicates a passing route (a pair of an input interface number and an output interface number) in the router. A path identifier PID including a number is added to the transmission target packet as an intra-router header.

  Therefore, when the second monitoring mode is adopted, the input interface number F3 (for example, F3 = # In = 5) and the output interface number F4 (for example, F4 = # E1 = 1) are applied to the path identifier PID by the identifier adding unit 12. ), A packet number F5 (for example, F5 = 045) is set. In the path identifier column of the path table 56, “51045” indicating the input interface number F3, the output interface number F4, and the packet number F5 is recorded.

[Router operation]
<First monitoring mode>
When the packet monitoring unit present at each monitoring location in the router monitors the path identifier PID of the transmission target packet that passes through, in the unit of the passing route, and adopts the first monitoring form in which the packet passing information is notified to the passage determining unit 55, An operation example in the router will be described with reference to related drawings.

  When a packet arrives at the nth input interface unit IFIn of the router whose detailed configuration is illustrated in FIG. 7, the destination address extraction unit 11 of the input network processor NPI receives the destination from the IP header of the packet to be transmitted that has arrived (received). The address (IP address) “a” is extracted and notified to the identifier assigning unit 12 (processes 901 and 902 in FIG. 9).

  The identifier assigning unit 12 determines (extracts) the output interface number “# E1” corresponding to the first output interface unit IFE1 from the storage information of the FIB 14 based on the notified destination address “a”, and in the switch 41 An output interface identifier “E1” for switching is generated and added to the transmission target packet as a part of the header in the router (processes 903, 904, and 905 in FIG. 9).

  Further, the identifier assigning unit 12 creates a path identifier (PID) “X” that uniquely indicates a passing route PR1 in the router, that is, a set of an input interface number and an output interface number, and adds an intra-router header to the transmission target packet. And then sent out (processes 906 and 907 in FIG. 9).

  When the first monitoring mode is adopted, the path identifier “X” is given an input interface number F3 (for example, F3 = # In = 5) and an output interface number F4 (for example, F4 = # E1 = 1) by the identifier adding unit 12. ) Is set, but the packet number F5 is not set.

  The packet monitoring unit 13 monitors the passing packet, extracts the path identifier “X”, and determines the packet passage information including the path identifier “X” and passage part information (here, the NPI number) as the passage determination of the control unit CONT. After notifying the unit 55 through the control line C1, the packet is transmitted (processing 1001, 1002, 1003, 1004 in FIG. 10).

  Other packet monitoring units 15, 21, 23, 42, 43, 45, and 46 existing at each monitoring location in the router similarly monitor packets passing through to extract a path identifier “X”, and packet passing information To the passage determination unit 55.

  The passage determination unit 55 sequentially records the packet passage state in the path table 56 based on the packet passage information notified from the packet monitoring unit 13 or the like, and updates the overwriting (process 1005 in FIG. 10).

The passage determination unit 55 determines that an abnormality has occurred when the transmission target packet does not output from the router within a predetermined time after it arrives at the router. In this example, the passage determination unit 55 creates a new entry in the path table 56, then monitors the path table 56, and if no packet passage is recorded in all the monitored parts even after waiting for a predetermined time, It is determined that the failure has occurred, and the identified failure part is notified to the maintenance operation system 4.

  Next, the processing in the passage determination unit 55 will be described in detail.

  When receiving the packet passage information including the path identifier “X” and the passage part information from the packet monitoring unit 13 or the like, the passage determination unit 55 determines that there is an entry in the path table 56 (processing 1101 and 1102 in FIG. 11). ).

  If the path identifier “X” as an entry has already been recorded in the path table 56, the packet passage status is recorded in the path table 56 and overwritten (step 1105 in FIG. 11).

  If the path identifier “X” as an entry does not exist in the path table 56, a new entry is created (recorded) in the path identifier column of the path table 56 (process 1103 in FIG. 11).

  When creating a new entry, the passage determination unit 55 sets (starts: ON) the failure determination timer TM1, and records the packet passage state in the path table 56 (processing 1104 and 1105 in FIG. 11).

  When all the packet passage states are recorded in the path table 56, the passage determination unit 55 resets (stops: OFF) the failure determination timer TM1 and ends the processing (processing 1106 and 1107 in FIG. 11).

  The passage determination unit 55 determines whether or not the failure determination timer TM1 set in the process 1104 has expired while repeating the processes 1101 to 1106 (process 1108 in FIG. 11).

  When the failure determination timer TM1 expires, the passage determination unit 55 refers to the path table 56, determines that there is a failure in the monitoring portion where no packet passage state is recorded, and notifies the maintenance operation system 4 of the failure portion ( Processes 1109 and 1110 in FIG.

  Subsequently, the passage determination unit 55 clears the record of the corresponding entry in the path table 56, resets the failure determination timer TM1, and ends the process (processes 1111 and 1112 in FIG. 11).

  In the example shown in FIG. 12A, X = “51” is recorded as a new entry in the path identifier column of the path table 56. In addition, a packet passage state (ON: passage) is recorded in each monitoring part (monitoring part) column of the record corresponding to the path identifier “51” of the path table 56.

  The target monitoring parts in the router are the input interface units IFI1 to IFIn, the output interface units IFE1 to IFEm, and the switch unit SW, but in this example, they are further subdivided to increase the fault detection accuracy. Therefore, as shown in FIG. 12A, the monitoring part column of the path table 56 includes the input network processor NPI, the input packet buffer PBI, the input and output of the switch unit SW, the output packet buffer PBO, and the output network processor NPO. Is registered in advance.

For example, when a packet loss occurs due to a failure inside the switch 41 of the switch unit SW, the packet passes through the packet monitoring units 13, 15, and 43 until the input of the switch 41, but the packet monitoring unit after the output of the switch 41 Since the packets do not pass through 45, 23, and 21, no packet passes are recorded in the output of the switch unit SW of the path table 56, the output packet buffer PBO, and the monitoring part column of the output network processor NPO.

  When the router is operating normally, the packet to be transmitted from which the header in the router including the output interface identifier “E1” and the path identifier “X” has been removed by the identifier removing unit 22 is transmitted from the router.

<Second monitoring mode>
Next, the packet monitoring unit existing at each monitoring location in the router monitors the path identifier PID of the transmission target packet passing for each passing route, and notifies the passage determining unit 55 of the packet passing information. In the case of adopting the monitoring mode, an example of the operation in the router will be described with reference to related drawings.

  When a packet arrives at the nth input interface unit IFIn of the router whose detailed configuration is illustrated in FIG. 7, the destination address extraction unit 11 of the input network processor NPI uses the destination address “a” from the IP header of the arrived transmission target packet. Is extracted and notified to the identifier assigning unit 12 (processes 901 and 902 in FIG. 9).

  Based on the notified destination address “a”, the identifier assigning unit 12 extracts the output interface number “# E1” corresponding to the first output interface unit IFE1 from the storage information of the FIB 14, and performs switching in the switch 41. Output interface identifier “E1” is added as a part of the header in the router to the transmission target packet (processes 903, 904, and 905 in FIG. 9).

  Further, the identifier assigning unit 12 uniquely indicates a fragmented packet that passes through the same route in the router, in addition to information that uniquely indicates a passing route PR1 in the router, that is, a set of an input interface number and an output interface number. A path identifier (PID) “X” (“X1”, “X2”...) Including a packet number which is a sequence number is created and further added as a header in the router to the transmission target packet (FIG. 9). Processing 906, 907). In the process of creating the path identifiers “X1”, “X2”, etc., the identifier assigning unit 12 extracts the sequence number of the fragmented packet from the IP header of the transmission target packet and sets it as the packet number.

  In the case of adopting the second monitoring mode, the path identifier “X1” is given an input interface number F3 (for example, F3 = # In = 5) and an output interface number F4 (for example, F4 = # E1 = 1) by the identifier adding unit 12. ), A packet number F5 (for example, F5 = 001) is set. For the next path identifier “X2”, the identifier assigning unit 12 uses the input interface number F3 (for example, F3 = # In = 5) and the output interface number F4 (for example, F4 = # E1 = 1) together with the packet number F5. (For example, F5 = 002) is set.

  Subsequently, the packet monitoring units 13, 15, 43, 45, 23, and 21 existing at each monitoring location in the router monitor the passing packet and pass the path in the same manner as the processing 1001, 1002, 1003, and 1004 described above. The identifiers “X1”, “X2” and the like are extracted, and packet passage information is notified to the passage determination unit 55.

  Similar to the processing 1005 described above, the passage determination unit 55 sequentially records the packet passage state in the path table 56 based on the packet passage information notified from the packet monitoring unit 13 or the like, and updates the overwriting.

  Similarly to the above-described processes 1101 to 1112, the passage determination unit 55 creates a new entry in the path table 56 and then monitors the path table 56. If the passage is not recorded, it is determined that an abnormality has occurred, and the identified faulty part is notified to the maintenance operation system 4.

  In the example shown in FIG. 12B, X1 = “51001”, X2 = “51002”, and the like are recorded in the path identifier column of the path table 56 as new entries. Further, the packet passage status (ON: passage) is recorded in each monitoring part column of the record corresponding to the path identifiers “51001” and “51002” of the path table 56.

  For example, when a packet loss occurs due to a failure inside the switch 41 of the switch unit SW, the packet passes through the packet monitoring units 13, 15, and 43 until the input of the switch 41, but the packet monitoring unit after the output of the switch 41 Since the packets do not pass through 45, 23, and 21, no packet passes are recorded in the output of the switch unit SW of the path table 56, the output packet buffer PBO, and the monitoring part column of the output network processor NPO.

  When the router is operating normally, the packet to be transmitted from which the header in the router including the output interface identifier “E1” and the path identifiers “X1” and “X2” is removed by the identifier removing unit 22 is transmitted from the router.

<Deformation example>
In the communication system 1 according to the embodiment shown in FIG. 6, routing information may be exchanged between the routers R <b> 11 to R <b> 14 on the IP network 2.

  In the router (specific router) whose detailed configuration is illustrated in FIG. 7, the routing control unit 51, the identifier assigning unit 52, the packet monitoring unit 53, the routing table 54, and the packet monitoring unit 57 of the control packet buffer PBC are the other routers. It works when exchanging routing information with.

  When exchanging the routing information, a control packet including the routing information is transmitted from the routing control unit 51 of the specific router to the other router by the passing route PR2 in FIG. 7, and the routing control unit 51 of the specific router from the other router. There is a form in which a control packet including routing information is received by the passing route PR3 in FIG. The first monitoring mode and the second monitoring mode described above can be selectively applied to both the control packet transmission mode and the control packet reception mode.

  For example, when the first monitoring mode is adopted in the mode in which the control packet is transmitted by the passing route PR2, the routing control unit 51 includes the destination address “a” corresponding to another router in the IP header of the transmission target packet (control packet). And the information which shows the control packet containing routing information is set, and the identifier provision part 52 is notified.

  The identifier assigning unit 52 extracts the output interface number “# E1” corresponding to the first output interface unit IFE1 from the stored information of the routing table 54 based on the notified destination address “a”, and performs switching in the switch 41. Output interface identifier “E1” is generated and added to the transmission target packet as a part of the header in the router.

Further, the identifier assigning unit 52 uses the information indicating the control packet including the notified routing information as the passing route in the router, the information indicating the transmission of the control packet instead of the input interface number, and the output interface number. A path identifier “M1” is generated and added to the transmission target packet as a header in the router, and then transmitted.

  When the control packet transmission form is adopted, the transmission target packet is input from the control packet buffer PBC to the switch 41 via the packet monitoring unit 44.

  The packet monitoring units 53, 57, 44, 45, 23, and 21 existing at each monitoring location in the router monitor the passing packet and monitor the path identifier “M 1” in the same manner as the processes 100 1, 100 2, 100 3, and 100 4 described above. ”Is extracted, and packet passage information is notified to the passage determination unit 55.

  Similar to the processing 1005 described above, the passage determination unit 55 sequentially records the packet passage state in the path table 56 based on the packet passage information notified from the packet monitoring unit 53 and the like, and updates the overwriting.

  Similarly to the above-described processes 1101 to 1112, the passage determination unit 55 creates a new entry in the path table 56 and then monitors the path table 56. If the passage is not recorded, it is determined that an abnormality has occurred, and the identified faulty part is notified to the maintenance operation system 4.

  In the example shown in FIG. 12C, M1 = “01” (where “0” is information indicating control packet transmission) is recorded as a new entry in the path identifier column of the path table 56. In addition, the packet passage status (ON: passage) is recorded in each monitoring part column of the record corresponding to the path identifier “01” of the path table 56.

  As shown in FIG. 12C, in the monitoring part column of the path table 56, the control unit CONT, the control packet buffer PBC, the input and output of the switch unit SW, the output packet buffer PBO, and the output network processor NPO are registered in advance. Has been.

  For example, when a packet loss occurs due to a failure inside the switch 41 of the switch unit SW, the packet passes through the packet monitoring units 53, 57, and 44 up to the input of the switch 41, but the packet monitoring unit after the output of the switch 41 Since the packets do not pass through 45, 23, and 21, no packet passes are recorded in the output of the switch unit SW of the path table 56, the output packet buffer PBO, and the monitoring part column of the output network processor NPO.

  When the router is operating normally, the packet to be transmitted from which the header in the router including the output interface identifier “E1” and the path identifier “M1” has been removed by the identifier removing unit 22 is transmitted from the router.

  Next, for example, when the first monitoring form is adopted in the form in which the control packet arrives through the passing route PR3, when the packet arrives at the nth input interface unit IFIn of the router, the destination address of the input network processor NPI The extraction unit 11 extracts the destination address “aa” addressed to itself from the IP header of the arrived transmission target packet (control packet), and notifies the identifier addition unit 12 of the destination address.

  The identifier assigning unit 12 refers to the information indicating the control packet including the routing information from the IP header, based on the notified destination address “aa” addressed to itself, and outputs the input interface number and the output as the passing route in the router. A path identifier “M2” indicating information indicating the arrival of the control packet is created instead of the interface number, and is added to the transmission target packet as a header in the router and then transmitted.

  When the control packet incoming form is adopted, the transmission target packet is input from the input packet buffer PBI to the switch 41 via the packet monitoring unit 43 and output to the packet monitoring unit 44. The transmission target packet is input to the routing control unit 51 via the control packet buffer PBC and the packet monitoring unit 53 and processed.

  The packet monitoring units 13, 15, 43, 44, 57, and 53 existing at each monitoring location in the router monitor the packets that pass and monitor the path identifier “M2” in the same manner as the processing 1001, 1002, 1003, and 1004 described above. ”Is extracted, and packet passage information is notified to the passage determination unit 55.

  Similar to the processing 1005 described above, the passage determination unit 55 sequentially records the packet passage state in the path table 56 based on the packet passage information notified from the packet monitoring unit 13 or the like, and updates the overwriting.

  Similarly to the above-described processes 1101 to 1112, the passage determination unit 55 creates a new entry in the path table 56 and then monitors the path table 56. If the passage is not recorded, it is determined that an abnormality has occurred, and the identified faulty part is notified to the maintenance operation system 4.

  In the example shown in FIG. 12D, M2 = “50” (where “0” is information indicating the arrival of a control packet) is recorded as a new entry in the path identifier column of the path table 56. In addition, the packet passage state (ON: passage) is recorded in each monitoring part column of the record corresponding to the path identifier “50” of the path table 56.

  As shown in FIG. 12D, in the monitoring part column of the path table 56, the input network processor NPI, the input packet buffer PBI, the input and output of the switch unit SW, the control packet buffer PBC, and the control unit CONT are registered in advance. Has been.

  For example, when a packet loss occurs due to a failure inside the switch 41 of the switch unit SW, the packet passes through the packet monitoring units 13, 15, and 43 until the input of the switch 41, but the packet monitoring unit after the output of the switch 41 Since the packet does not pass through 44, 57, 53, the passage of the packet is not recorded in the output of the switch unit SW of the path table 56, the control packet buffer PBC, and the monitoring part column of the control unit CONT.

  In the above description, a packet including routing information has been described as an example, but other packets originating from or destined to the own apparatus can be similarly monitored.

[Modification of communication system]
In the communication system 1 including the silent failure autonomous detection function of the above-described embodiment, the passage determination unit 55 of the control unit CONT included in each router R11, R12, R13, R14 that forwards the transmission target packet is In cooperation with the path table 56 and the packet monitoring unit 13, the occurrence of an abnormality in the router is detected, and a failure location notification including the router IP address and failure location is sent to the maintenance operation system 4.

  As a result, the maintenance operation system 4 can recognize that the specific router cannot transfer the packet based on the failure site notification from each router.

  In the communication system 1B of the modified example illustrated in FIG. 13, the passage determination unit and the path table are provided in the control system 7 as a centralized control unit.

  In this control system 7, each router R11A to R14A adds a path identifier indicating a route inside the router to a transmission target packet passing through itself, monitors the path identifier in each part in the router, and The control system 7 is notified of the packet passing status together with the IP address of the router.

  In the control system 7 notified of the packet passage information including the path identifier and passage portion information and the IP address of the router, the passage determination unit (not shown) stores the packet passage information in the router-compatible path table (not shown). To do. The passage determination unit determines the presence / absence of packet transfer in each part of each router based on the storage information of this path table, and identifies the fault part of the faulty router. Then, the control system 7 notifies the maintenance operation system 4 of the faulty part including the router IP address and the faulty part.

  As a result, in the maintenance operation system 4, it is possible to recognize that the specific router cannot transfer the packet based on the notification of the faulty part from the control system 7.

  When the control system 7 including the passage determination unit and the path table is at the network level as the centrally arranged control unit as in the communication system 1A, the following advantages can be obtained compared to the communication system 1.

  (1) It is possible to apply the silent failure autonomous detection function to the communication system while suppressing changes to the existing router.

  (2) When the size of the path table managed by the router increases due to an increase in the traffic volume of the transmission target packet relayed by the router, it is necessary to increase the memory inside the router. In this case, a change to each router existing in the IP network can be suppressed, and this can be dealt with by newly establishing an external control system.

[Other variations]
In the embodiment and the modification described above, the identifier adding unit adds the path identifier to the transmission target packet, and the passing information including the path identifier extracted by each packet monitoring unit and the information indicating the monitoring location in the apparatus is obtained. By notifying each, the passage determination unit detects the failure point in the device in accordance with the passage state of the packet recorded in the path table. Is possible. In this case, each packet monitoring unit notifies only the event of packet passing to be recorded in the path table. Then, the passage determination unit does not perform the passage determination for each transmission route of the transmission target packet, but performs the passage determination on all the passage routes together. As a result, it is not possible to specify the fault location in units of passage routes, but it is possible to specify the fault location within the apparatus. In addition, by using a configuration in which a packet monitoring unit for notifying the passage determination unit of an event of packet passing is physically or logically arranged upstream of the destination address extraction unit, in combination with the above-described embodiment or this other modification Therefore, it is possible to identify a failure location in a wider range, such as detecting a failure in the destination address extraction unit.

  The processing in the embodiment and the modification described above is provided as a computer-executable program, and can be provided via a recording medium such as a CD-ROM or a flexible disk, and further via a communication line.

  In addition, each process in the above-described embodiment and modification can be performed by selecting and combining any or all of the processes.

[Others]
The following additional remarks are disclosed with respect to the above-described embodiment and modifications.

(Appendix 1)
An adding means for adding a path identifier uniquely indicating a passing route in the apparatus to the transmission target packet;
A plurality of monitoring means for notifying each of passing information including the extracted path identifier and information indicating a monitoring location in the apparatus when the path identifier is extracted from the packet passing;
Based on the passage information notified from the plurality of monitoring means, the passage state of the packet is recorded in a path table, and the failure location in the apparatus is detected according to the passage state recorded in the path table With means;
A relay device comprising:

(Appendix 2)
The adding means adds the path identifier to the transmission target packet as an in-device header,
The path identifier includes information indicating whether the monitoring unit present at each monitoring location in the apparatus monitors the transmission target packet in units of passing routes or in units of packets for each passing route. Relay device.

(Appendix 3)
3. The relay apparatus according to appendix 1 or 2, wherein the adding unit adds the path identifier indicating a set of an input interface number and an output interface number as an internal header to the transmission target packet as a passing route in the apparatus.

(Appendix 4)
The adding means includes a packet number which is a sequence number for indicating fragmented packets passing through the same passing route, in addition to information indicating a set of an input interface number and an output interface number as a passing route in the device. 3. The relay device according to appendix 1 or 2, wherein a path identifier is added to the transmission target packet as an in-device header.

(Appendix 5)
The relay device according to supplementary note 1, wherein the determination unit determines that an abnormality has occurred when the passage status of all the monitoring points in the device is not recorded in the path table within a predetermined time.

(Appendix 6)
The monitoring points in the apparatus by the plurality of monitoring means include a plurality of input interface units, a plurality of output interface units, and a switch unit that interconnects the plurality of input interface units and the plurality of output interface units. The relay device according to 1.

(Appendix 7)
When the addition means transmits a control packet including routing information, the path identifier indicating information indicating control packet transmission instead of the input interface number and the output interface number as a passing route in the device 5. The relay apparatus according to appendix 3 or 4, wherein the relay device is added to the transmission target packet.

(Appendix 8)
When the control packet including routing information arrives, the adding means, as the passing route in the device, shows the input interface number and the path indicating information indicating control packet arrival instead of the output interface number. The relay device according to appendix 1, wherein an identifier is added to the transmission target packet.

(Appendix 9)
The relay device according to supplementary note 1, wherein the determination unit and the path table are provided in an individually arranged control unit or a network level concentrated arrangement control unit.

(Appendix 10)
The relay device according to appendix 1, wherein the determination unit notifies the maintenance operation system of a detected failure location in the device.

(Appendix 11)
Adding a path identifier uniquely indicating the route through the device to the transmission target packet;
When the path identifier is extracted from the packet passing through, the passing information including the extracted path identifier and information indicating the monitoring location in the apparatus is notified from each of a plurality of monitoring locations;
Based on the passage information notified from the plurality of monitoring locations, the passage status of the packet is recorded in a path table, and a fault location in the apparatus is detected according to the passage status recorded in the path table;
A silent fault autonomous detection method in which a relay device processes this.

DESCRIPTION OF SYMBOLS 1 Communication system 1B Communication system 2 IP network 4 Maintenance operation system 5 Communication terminal device 6 Communication terminal device 7 Control system R11 Router R12 Router R13 Router R14 Router R11A Router R12A Router R13A Router R14A Router

Claims (7)

An adding means for adding a path identifier uniquely indicating a passing route in the apparatus to the transmission target packet;
A plurality of monitoring means for notifying each of passing information including the extracted path identifier and information indicating a monitoring location in the apparatus when the path identifier is extracted from the packet passing;
Based on the passage information notified from the plurality of monitoring means, the passage state of the packet is recorded in a path table, and the failure location in the apparatus is detected according to the passage state recorded in the path table With means;
A relay device comprising: The adding means adds the path identifier to the transmission target packet as an in-device header,
2. The path identifier includes information indicating whether the monitoring unit present at each monitoring location in the apparatus monitors the transmission target packet in units of passing routes or in units of packets for each passing route. Relay device. 3. The relay device according to claim 1, wherein the adding unit adds the path identifier indicating a set of an input interface number and an output interface number as an intra-device header to the transmission target packet as a passing route in the device. The adding means includes a packet number which is a sequence number for indicating fragmented packets passing through the same passing route, in addition to information indicating a set of an input interface number and an output interface number as a passing route in the device. The relay device according to claim 1 or 2, wherein a path identifier is added to the transmission target packet as an in-device header. The relay device according to claim 1, wherein the determination unit determines that an abnormality has occurred when the passage status of all the monitoring points in the device is not recorded in the path table within a predetermined time. The relay device according to claim 1, wherein the monitoring unit notifies the determination unit of the passing information based on the path identifier included in the transmission target packet originating from or addressed to the own device. Adding a path identifier uniquely indicating the route through the device to the transmission target packet;
When the path identifier is extracted from the packet passing through, the passing information including the extracted path identifier and information indicating the monitoring location in the apparatus is notified from each of a plurality of monitoring locations;
Based on the passage information notified from the plurality of monitoring locations, the passage status of the packet is recorded in a path table, and a fault location in the apparatus is detected according to the passage status recorded in the path table;
A silent fault autonomous detection method in which a relay device processes this.

Images