JP2018067759A - Loop failure detection method, network device, and loop failure detection program - Google Patents

Abstract

An object of the present disclosure is to detect the occurrence of a loop fault and specify a loop source. In the present disclosure, when a network device 91 receives a communication frame from outside the ring network, the network device 91 encapsulates the communication frame and transmits the first frame after limiting the communication frame within the ring network. Transfer procedure and when the network device 91 receives a communication frame from the ring network, the transmission source address is rewritten to an XOR composite value obtained by XORing the transmission source address and the address of the own device, and then transmitted. The second transfer procedure is performed, and in the second transfer procedure, the source address of the received communication frame is compared with the XOR composite value described in the normal value determination table held in advance. Then, it is determined whether or not a loop fault has occurred in the ring network. [Selection] Figure 9

Description

  The present disclosure relates to detection of loop faults in ring networks.

  A ring network (hereinafter referred to as NW) is an NW that physically connects NW devices in a ring shape and logically has a bus-side topology. A loop fault is a fault that originally has a logical topology on the bus side, but logically becomes a ring-shaped NW due to a special fault.

  When a frame for multiple destinations or a frame whose destination is unknown flows, the NW device copies and transmits the frame to all ports. However, in the case of Condition 1 or Condition 2, a network NW state called NW storm (broadcast storm) or NW meltdown occurs.

  Condition 1: If a loop failure occurs and a frame copied by itself before the frame copy process is completed, the frame is further copied including the returned frame. As a result, the transmission frame is exponentially amplified, the traffic becomes congested, and communication cannot be performed at all.

  Condition 2: When NW devices belonging to m rings simultaneously cause a loop failure in n out of m rings (m ≧ n) at the same time, 2n copied frames are returned. X (2m-1) copies are to be copied. Among them, n rings (2n × (2n-1)) are returned, so that the transmission frame is exponentially amplified in order to copy (2m-1) times and transmit it. As a result, traffic becomes congested and communication cannot be performed at all.

  FIG. 1 illustrates a case where two (n = 2) rings fail simultaneously under condition 2. (1) One frame whose destination is unknown flows into the NW device. (2) Two frames are copied by the NW device and transmitted from the left and right ports belonging to the respective ring type NW # 1 and # 2. As a result, the number of frames increases from one to four, which is 2n times. (3) The four copied frames are returned to the NW device. (4) Each is copied by the NW device and transmitted to each ring again. At this time, since it is transmitted from all ports of the ring type NW except the reception port, it is transmitted from the port (2n-1). As a result, the number increases from 4 to 12 (2n-1) times. (5) The 12 copied frames are returned to the NW device. In this way, by repeating (4) and (5), the frames multiply exponentially, consumes all the NW band, and cannot communicate at all.

  If a loop failure occurs alone, nothing happens, but if the above condition 1 or condition 2 is met, the entire NW immediately melts down. For this reason, in order to prevent NW meltdown in advance, it is important to detect even a single loop failure at an early stage.

  In general, a check code is provided to detect garbled data frames. For example, there is FCS (CRC 32) added to the end of the Ethernet (registered trademark) frame. When bit corruption occurs due to noise received from the transmission path, the presence or absence of bit corruption can be detected with this check code. However, this check code may not work. This is a case where bit corruption of a data frame occurs in the NW device. For example, consider a case where a certain data frame is received at a certain port A of the NW device and then output from another port B.

(1) When a data frame (for example, an ether frame) is received at port A, the check code is referred to and verified to check for the occurrence of bit corruption.
(2) The data frame format is converted into an internal representation unique to the NW device. Remove the check code at this time.
(3) Rewrite the priority identifier and customer identifier as necessary.
(4) Transfer to port B.
(5) The transferred data is converted from the internal representation unique to the NW device into a predetermined frame format (eg, Ethernet frame).
At this time, a check code is again calculated and added to the frame and transmitted.

  In the process of (3) to (4), if bit corruption occurs due to a failure of the NW device, as shown in FIG. 2, a check code is calculated and assigned to the frame in which bit corruption exists, and then transmitted. Therefore, the check code is correct despite the presence of bit corruption, and the bit corruption cannot be detected by the check code. A mechanism for adding / verifying a check code may be provided each time inside the NW device, but the NW device becomes complicated (expensive). In particular, since a loop failure is often caused by an internal failure of the NW device, it is necessary to take measures to detect bit corruption that is difficult to detect with a check code.

  If the received frame is verified and does not match the check code, it is discarded as an illegal frame. On the other hand, the check code is removed and converted into an internal representation unique to the NW device. When bit corruption occurs due to a failure of the NW device, transfer is made to port B with bit corruption inherent. At this time, a check code is calculated and assigned in a state where bit corruption is inherent. For this reason, the check code cannot detect garbled bits. On the check code, since it is a correct value, it is transferred with a bit garbled. The received frame is verified, but since the check code is generated with data that is garbled, it is not treated as an illegal frame and is not discarded.

  The main detection methods for NW loop failures include a first detection method using a loop detection frame, a second detection method for monitoring traffic volume, a third detection method for monitoring routing table rewrite frequency, and TTL (HopLimit). ) Can be used as an example.

  In the first detection method, a loop detection frame is always communicated, and when the frame returns, it is determined that a loop failure has occurred (see, for example, Patent Document 1). In the first detection method, when an NW storm (broadcast storm) accompanying a loop failure occurs at the same time as the failure and the NW is congested, the loop failure detection frame that should be returned is lost due to the congestion and returned. There is a problem that failure detection cannot be performed even though a loop failure has occurred. In the first detection method, when a loop detection frame is rewritten to an invalid frame by a failed NW device, even if a loop failure occurs, the frame should be returned to the original state. Since it returns as a frame, there is a problem that failure detection may not be possible even though a loop failure has occurred.

  In the second detection method, since a loop failure often occurs and congestion due to NW storm often occurs, the traffic flow of a port belonging to the ring is measured, and a specific traffic (such as a broadcast frame) is measured. If the amount increases rapidly, it is determined that a loop failure has occurred. The second detection method has a problem that a loop failure without NW meltdown cannot be detected. Further, since the second detection method determines a loop failure only by the magnitude of traffic, there is a problem that it cannot be determined from which NW device in the network the frame that caused the loop failure (trigger) flows.

  The third detection method determines that a loop failure has occurred when rewriting of the routing table of the NW device (for example, FDB in the L2 switch) is frequently performed at the port belonging to the ring. In the third detection method, when the amount of traffic looping in the ring is small due to a loop failure (when NW meltdown is not accompanied), the rewriting of the routing table is also infrequent and a loop failure occurs. Nevertheless, there is a problem that failure detection may not be possible. In addition, since the third detection method determines a loop failure only by the rewriting frequency of the routing table, it cannot be determined from which NW device in the network the frame that caused the loop failure (trigger) flows. There's a problem.

  The fourth detection method refers to the value of the number of devices that have passed through the NW device in the frame (for example, TTL (HopLimit) in the IP payload portion of L3), and determines that a loop has occurred when the value reaches a certain value. To do. In the fourth detection method, when TTL (HopLimit) of the IP payload is used, since the fold is 8 bits, a loop failure has not occurred in a large-scale NW having more than 256 NW devices to be routed. Nevertheless, there is a problem that erroneous detection may occur in which a loop failure has occurred. In addition, since the fourth detection method determines only by the number of devices that pass through, there is a problem that it cannot be determined from which NW device in the network the frame that caused the loop failure (trigger) has flowed. In addition, the fourth detection method is that when the TTL (HopLimit) value is rewritten to an incorrect value in a failed NW device in the NW, when a loop failure occurs, every time it passes through the failed device, TTL (HopLimit) ) There is a problem that the value is reset to an incorrect value, and the frame is continually turned around in the NW, and the occurrence of a loop failure may not be detected.

JP 2013-153293 A

  Even if the first detection method to the fourth detection method are combined, loop failure detection cannot be avoided and false detection cannot be avoided (Problem 1), or even if the loop can be detected correctly, the loop source cannot be identified ( There is a problem 2). Therefore, an object of the present disclosure is to detect the occurrence of a loop failure and specify a loop source.

The loop failure detection method according to the present disclosure is:
When a network device in a ring network receives a communication frame from outside the ring network, the source address and the destination address are encapsulated in the communication frame to obtain the address of the network device in the ring network. A first transfer procedure for transmission after limiting to
When a network device in a ring network receives a communication frame from another network device in the ring network, the transmission source address is an XOR composite value obtained by XORing the transmission source address and the address of the own device. Execute the second transfer procedure to transmit after rewriting,
In the second transfer procedure, a loop failure occurs in the ring network by comparing the transmission source address of the received communication frame with the XOR composite value described in the normal value determination table held in advance. The presence or absence of is determined.

The network device according to the present disclosure is:
A network device provided in a ring network,
When a communication frame encapsulated in the ring network unit is received, an XOR operation is performed using the transmission source address and the address of the own device included in the communication frame, and the transmission source address is overwritten.
By comparing the XOR composite value after the XOR operation with the XOR composite value held in the device itself, it is determined whether or not a loop failure or bit corruption has occurred, and in the case of a loop failure, the loop source is specified.

  The loop failure detection program according to the present disclosure is a program for causing a computer to execute each procedure provided in the loop failure detection method according to the present disclosure, and is a program for causing the computer to function as a network device according to the present disclosure. The program can be recorded on a recording medium or provided through a network.

  According to the present disclosure, it is possible to detect the occurrence of a loop failure and specify a loop source.

FIG. 6 is an explanatory diagram when two (n = 2) rings simultaneously fail under a condition 2 loop failure. It is explanatory drawing about the problem of related technology. 2 illustrates an example of an encapsulation according to the present disclosure. A specific example in the case where the XOR composite value is not applied to the transmission source address is shown. A specific example in the case where the XOR composite value is applied to the transmission source address is shown. It is explanatory drawing which shows the specific example of the loop source which concerns on this indication. The specific example of the XOR operation which concerns on this indication is shown. 1 illustrates an example of a loop fault detection system according to the present disclosure. The structural example of the NW apparatus which concerns on this indication is shown. An example of an outline of a frame processing procedure when a frame is input from the ring port 14L is shown. An example of port status monitoring of the ring port 14L using the “normal value determination table” is shown. An example of a “transmission source NW device address rewriting table” is shown. An example of a “normal value determination table” is shown. An example of a “ring-type intra-NW routing table” is shown. An example of an outline of a frame processing procedure when a frame is input from the ring port 14R is shown. An example of port status monitoring of the ring port 14R using the “normal value determination table” is shown. An example of a “transmission source NW device address rewriting table” is shown. An example of a “normal value determination table” is shown. An example of a “ring-type intra-NW routing table” is shown. An example of steps S121 to S124 of the frame processing procedure when input from a subscriber port is shown. An example of steps S125 to S127 of the frame processing procedure when input from a subscriber port is shown. An example of steps S131 to S133 of the frame processing procedure when input from a subscriber port is shown. An example of port status monitoring of a subscriber port using a “capsuling table” is shown. An example of a “capsuling table” is shown. An example of the forwarding port determination table for floods within a ring is shown. An example of the table for floods in a ring for normal time is shown. An example of the table for floods in a ring for the time of a path | route failure is shown. An example of the frame discard table for flooding within a ring is shown. An example of procedure 1 to procedure 3 in the operation of the NW controller will be shown. An example of the procedure 4 to the procedure 7 in the operation of the NW controller will be shown. An example of a table | surface (T1) is shown. An example of a table | surface (T2) is shown. An example of a table | surface (T3) is shown. An example of the SQR code in each procedure in the operation of the NW controller is shown. An example of steps S151 to S154 in the processing of the NW controller when a notification of a change in the port link state is received is shown. An example of steps S155 to S159 in the processing of the NW controller when receiving a notification of a change in the port link state is shown. An example of steps S160 to S162 in the processing of the NW controller when receiving a notification of a change in the port link state is shown.

  Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In addition, this indication is not limited to embodiment shown below. These embodiments are merely examples, and the present disclosure can be implemented in various modifications and improvements based on the knowledge of those skilled in the art. In the present specification and drawings, the same reference numerals denote the same components.

The present disclosure includes the following configurations.
First configuration: A normal communication frame is re-encapsulated in ring NW units, and the source address and destination address of the frame are limited to NW devices in the ring.
Second configuration: In addition, each time the device passes through the NW device in the ring, the source address is overwritten with the NW device address of the own device by the XOR operation.
Third configuration: Furthermore, it is determined from the list of XOR composite values held in each NW device whether or not a loop fault has occurred or a bit is garbled.

  FIG. 3 shows an example of encapsulation when the IEEE 802.1ah technology is used. By encapsulating in units of rings, even if a loop failure occurs, the address is limited to the address of the NW device in the ring, making it easy to identify the NW device. Thereby, this indication solves subject 2 by providing the 1st composition.

  When the addresses are encapsulated by the NW devices that form the ring according to the first configuration, the combinations are finite, a realistic order is obtained, and a table can be formed.

  On the other hand, when the XOR composite value is applied to the transmission source address, it can be distinguished from the transmission source address and the reception port whether the frame is looped or normal (not looped). Therefore, using the list of XOR composite values, it is determined whether or not a loop fault has occurred or a bit garble has occurred. Thereby, this indication solves subject 1 by providing the 2nd and 3rd composition.

  4 and 5 show specific examples of the case where the XOR composite value is not applied to the transmission source address and the case where it is applied. When the XOR composite value is not applied to the transmission source address, as shown in FIG. 4, the transmission source addresses of the received frames are both “NW device (A3)”, and it cannot be determined whether or not it is looped. On the other hand, by applying the XOR composite value to the transmission source address, the transmission source address of the received frame can be distinguished from “A3vA4” if it is normal and “A1vA2” if it is a loop.

  FIG. 6 shows a specific example of the loop source. Combinations of normal source addresses (XOR composite values) received by the NW device (A1) fall within a finite and realistic order. A list of transmission source addresses (XOR composite values) of non-looped frames to be received at the left port of the NW device A1 is “A4” from A4, which is a number obtained by subtracting 1 from the number of devices constituting the ring, There are three types, “A3vA4” from A3 and “A2vA3vA4” from A2. Anything other than these may be discarded.

  Further, in the port on the left side of the NW device A1, when discarding, it is possible to determine a loop generation source and occurrence of bit corruption from the transmission source address. For example, if “A1vA2vA3vA4” or “0 (NULL value)”, the loop source is A1, if “A1vA2vA3”, the loop source is A4, and if “A1vA2”, the loop source is A3. If “A1”, the loop source is A2, and otherwise, it can be determined as a bit error.

  A list of transmission source addresses (XOR composite values) of non-looped frames to be received at the right port is obtained by subtracting 1 from the number of devices constituting the ring, “A2” from A2, “ A3vA4 ”and“ A4vA3vA2 ”from A4. Anything other than these may be discarded.

  Further, when discarding, it is possible to determine a loop generation source and occurrence of bit corruption from the transmission source address. If "A1vA2vA3vA4" or "0 (NULL value)", the loop source is A1, if "A1vA3vA4", the loop source is A1, if "A1vA4", the loop source is A3, and "A1" If so, the loop source is A4, otherwise a bit error.

  Here, in the case of “transmission source address = own device NW device address”, it may be determined to be a loop. However, it is only one unit (transmission source device) that determines whether or not it is a loop frame. That is, there is a problem that a loop occurs if this device fails. In addition, there is a problem that a loop is missed if the source address is garbled. On the other hand, when the XOR composite value is applied to the transmission source address, it is possible to determine whether or not the frame is a loop frame by all the devices constituting the ring, and it is possible to cope with the garbled transmission source address.

Also, if the XOR composite value is used, it can be considered that it returns to the original value after several turns due to the nature of the XOR operation, and becomes the same value as a normal frame. For this problem, the loop frame is determined by all the devices constituting the ring.
FIG. 7 shows a specific example of the XOR operation. For example, the NW device A4 determines as follows.
・ 0 lap → 1st lap: “A1vA2vA3vA4” (detectable as an abnormal value)
-1st lap ⇒ 2nd lap: "0 (NULL value)" (detectable as an abnormal value)
・ 2nd lap ⇒ 3rd lap: "A1vA2vA3vA4"
・ 3rd lap ⇒ 4th lap: “0 (NULL value)”
For example, the NW device A2 determines as follows.
-0th lap: “A1” (correctly transferred as normal value)
-First lap: “A2vA3vA4” (A loop can be detected as an abnormal value)
-Second lap: "A1"
-Third lap: "A2vA3vA4"
In this way, detection and destruction can be performed in any of the NW devices constituting the ring until the second round.

  FIG. 8 illustrates an example of a loop failure detection system according to the present disclosure. In the loop failure detection system, network devices 91 are connected in a ring shape. Each NW device 91 includes a table processing unit 11 for each port.

  It is assumed that one PC is connected to each NW device 91 and each PC is communicating. Each NW device 91 holds a table in which a frame processing method is described, and performs processing such as frame rewriting / transfer processing / discard based on the table. When processing a frame not described in the table or when a link failure occurs, each NW device 91 makes an inquiry to the NW controller 92 and processes the frame based on a response from the NW controller 92. In response to an inquiry from each NW device 91, the NW controller 92 adds, deletes, and rewrites the table of the NW device 91.

  FIG. 9 illustrates a configuration example of the NW device according to the present disclosure. The NW device 91 according to the present embodiment includes a port and table processing units 11L, 11R, and 11A for each port, a routing unit 12, an NW controller inquiry unit 13, a capsuling unit 16, and a decapsuling unit 17. The table processing unit 11 is associated with each port. The table processing units 11L, 11R, and 11A hold a table in which a frame processing method is described, and perform processing such as frame rewriting / transfer processing / discard based on the table. The encapsulation unit 16 performs a encapsulation process for frames received from outside the ring network. The decapsulation unit 17 performs decapsulation processing on the encapsulated frame received from within the ring network.

  Regarding the processing of frames not described in the table, the table processing units 11L, 11R, and 11A make an inquiry to the NW controller 92 via the NW controller inquiry unit 13, and perform the frame processing based on the response from the NW controller 92. Do. Further, when a change occurs in the port state such as a link breakage, the NW controller inquiry unit 13 notifies the NW controller 92 that the port state has changed, to the table processing units 11L, 11R, and 11A.

  The table processing unit 11A functions as a first transfer unit and executes a first transfer procedure. The subscriber port 15 is a port that receives a communication frame from outside the ring network. The table processing units 11L and 11R function as a second transfer unit and execute a second transfer procedure. The ring ports 14L and 14R are ports that receive communication frames from other network devices in the ring network.

(Table processing units 11L and 11R)
FIG. 10 shows an example of an outline of a frame processing procedure when a frame is input from the ring port 14L. When a frame is input from the ring port 14L, the table processing units 11L and 11R refer to the “transmission source NW device address rewriting table” and rewrite the transmission source NW device address (S111). Then, referring to the “normal value determination table”, the source NW device address is confirmed (S112). If the transmission source NW device address is correct, not an illegal value, the transfer destination port is determined with reference to the “ring type intra-NW routing table” (S114). If the source NW device address is an illegal value, the frame is discarded and the maintenance person is notified (S113). Here, in the case of “A4vA3vA2vA1”, reception may be performed at “destination unknown address”, so only frame discard is performed.

  If the destination NW device address is another device, a frame is output from the transfer destination port (ring port 14R) (S115). When the transmission destination NW device address is its own device, the decapsuling unit 17 performs decapsulation processing (S116), and performs reference processing of the “capsuling table” of the subscriber port as shown in FIG. 24 (S117). When the transmission destination address is described in the “capsuling table”, the decapsulated frame is transmitted from the subscriber port 15. When the destination address is “address unknown”, the process proceeds to step S121 described later. If the destination NW device address is “destination unknown address”, the frame is copied and both processes are executed. As a result, “output frame from transfer destination port”, “decapsulation processing”, and “referencing capsule table” are executed.

  FIG. 11 shows an example of port state monitoring of the ring port 14L using the “normal value determination table”. It is determined whether or not the link state of the “ring port 14L” port has changed, and if so, the NW controller 92 is notified of the changed port and its port state.

  FIG. 12 shows an example of a “transmission source NW device address rewriting table”. Since the source / destination address is limited to the NW device address in the ring, the combinations are also limited, and all combinations can be written out statically in the list. That is, no combination explosion occurs. As a remark, an example of normal value determination is shown. The number of entries (rules) in the table is “the number of NW devices constituting the ring” × 2 [pieces]. Normally, it is discarded in the next process (normal value determination), but it is preferable to rewrite in order to identify the loop source when a failure such as the table contents being rewritten illegally occurs. .

  FIG. 13 shows an example of a “normal value determination table”. The number of entries (rules) in the table is “the number of NW devices constituting the ring” × 2 + 1 [pieces]. Since the source / destination address is limited to the NW device address in the ring, the combinations are also limited, and all combinations can be statically written in the list. That is, no combination explosion occurs.

  FIG. 14 shows an example of a “ring type intra-NW routing table”. The number of entries (rules) in the table is 3 [pieces]. This table can also be written out statically.

  FIG. 15 shows an example of an outline of a frame processing procedure when a frame is input from the ring port 14R. Steps S111 to S117 are executed in the same manner as when a frame is input from the ring port 14L. However, the transfer destination port in step S115 is the ring port 14L. In the case of “A2vA3vA4vA1”, since reception may be performed at “destination unknown address”, only frame discard is performed.

  FIG. 16 shows an example of port status monitoring of the ring port 14R using the “normal value determination table”. FIG. 17 shows an example of a “transmission source NW device address rewriting table”. FIG. 18 shows an example of a “normal value determination table”. FIG. 19 shows an example of a “ring type intra-NW routing table”. In these examples, the frame is the same as the ring port 14L when a frame is input from the ring port 14L, but the transfer order of the NW device is different from the ring port 14L. Different.

(Table processing unit 11A)
20 to 22 show an example of an outline of the frame processing procedure when input is made from the subscriber port 15. When a frame is input from the subscriber port 15, the table processing unit 11A refers to the “capsuling table”, and the encapsulation unit 16 performs the encapsulation process (S121). The “capsuling table” describes, for each subscriber port 15, a transmission destination address before encapsulation, an address after encapsulation, and a transfer destination port.

  When the transmission destination address is not registered in the encapsulation table, or when the transfer destination port is “undefined route”, the NW controller inquiry unit 13 asks the NW controller 92 of the address and transfer destination port after the encapsulation. An inquiry is made (S122). The table processing unit 11A performs a capsuling process based on the inquiry result, and adds an entry to the “capsuling table”. Here, if the destination NW device address is “destination unknown address”, the table processing unit 11A does not add an entry to the table. If the input port and the transfer destination port are the same, the table processing unit 11A discards the frame.

  The table processing unit 11A determines whether the “transmission source NW device address” and the “transmission destination NW device address” are the same (S123). If they are the same, the table processing unit 11A performs decapsulation processing (S124), and outputs a frame from the transfer destination port (subscriber port 15).

  When the “transmission source NW device address” and the “transmission destination NW device address” are different in step S123, the table processing unit 11A determines whether the “transmission destination NW device address” is a “destination unknown address” (S125). ). In the case of “address unknown”, the table processing unit 11A performs decapsulation processing (S126), and outputs a frame from the transfer destination port (subscriber port 15).

  In step S125, if the address is not “destination unknown address”, the table processing unit 11A determines whether the “forwarding destination port” is “route undefined” (S127). If the route is not “undefined”, the table processing unit 11A outputs a frame from the transfer destination port. Here, when the input port and the transfer destination port are the same, the table processing unit 11A discards the frame.

  If “Route Undefined” in step S127, the process proceeds to the in-ring flood process. In the in-ring flood process, the table processing unit 11A refers to the “in-ring flood transfer destination port determination table” as shown in FIG. 25, and “ring port # 1” and “ring port # 2” as transfer destinations. Is determined (S131). The table processing unit 11A determines whether or not the ring type NW is in failure (path failure) (S132). If the ring type NW is not in failure (path failure) in step S132, the table processing unit 11A refers to the “in-ring flood source NW device address rewriting table [normal]” as shown in FIG. The “source NW device address” is rewritten (S133). On the other hand, when the ring type NW is in failure (path failure) in step S132, the table processing unit 11A creates a “ring in-ring flood source NW device address rewriting table [at the time of path failure]” as shown in FIG. Refer to and rewrite the “transmission source NW device address” (S134).

  After that, the table processing unit 11A refers to the “in-ring flood frame discard table” as shown in FIG. 28 and determines whether the input port and the forwarding port are the own ring (S135). When the input port and the transfer destination port are not their own rings, the table processing unit 11A outputs frames from both ring ports (ring port L and ring port R). On the other hand, when the input port and the transfer destination port are the same, the table processing unit 11A discards the frame that has flowed in from its own ring.

  FIG. 23 shows a port status monitoring example of the subscriber port 15 using the “capsuling table”. It is determined whether or not the link status of the “subscriber port” port has changed, and if so, the NW controller 92 is notified of the changed port and its port status.

  FIG. 24 shows an example of a “capsuling table”. In the case of a frame not in the table or a frame in which the “forwarding port” column is registered as “undefined route”, the NW controller inquiry unit 13 inquires the NW controller 92. In response to the inquiry, the NW controller 92 returns a transmission destination address (NW device address) to be encapsulated. The table processing unit 11A adds the response from the NW controller 92 to the encapsulation table. However, when the “destination NW device address” is the “destination unknown address”, the table processing unit 11A does not add to the encapsulation table. The operation of the NW controller will be described later.

  Here, when the response from the NW controller 92 is slow and frame processing is desired to be performed at high speed, the encapsulation unit 16 sets “destination NW device address” to “destination unknown address” and “transfer destination port” to “undefined route”. The processing may be performed after being encapsulated as “,” and the NW controller 92 may be inquired in parallel.

  The table processing unit 11A encapsulates the frame itself input from the subscriber port 15. FIG. 3 shows an example of a communication data frame. Thereby, the communication data frame including the transmission destination NW device address, the transmission source NW device address, the transmission destination address, the transmission source address, and the transmission source data communicates in the ring type NW. By encapsulating the ring type NW unit, even if a loop failure occurs, the address is limited, so that the NW device can be easily identified.

  FIG. 26 shows an example of a ring flood table for normal use. The intra-ring flood table is statically prepared in advance and includes a source NW device address rewriting table [normal]. The transmission source NW device address rewriting table stores an action when a condition is matched for a combination of a transmission source NW device address and a transmission destination NW device address. “Ring port L transmission source NW device address” is a transmission source NW device address of a frame transmitted from ring port 14L, and “transmission source NW address” of a frame transmitted from “ring port 14L” is “ The same XOR composite value as that of the frame transmitted from the “NW device (A3)” is rewritten, so that the frame is not transferred first from the NW device (A4) so that a loop does not occur. The “ring port R transmission source NW device address” is the transmission source NW device address of the frame transmitted from the ring port 14R, and the “transmission source NW address” of the frame transmitted from the “ring port 14R” is “ The frame is rewritten to the same XOR composite value as that of the frame transmitted from the “NW device (A4)” so that the frame is not transferred first from the NW device (A3) so that a loop does not occur.

  FIG. 27 shows an example of an in-ring flood table for path failure. At the time of failure, the “source NW device address” is not rewritten for any of the transmissions of the ring ports 14L and 14R. As a result, the frame reaches the entire ring type NW regardless of the failure location in the ring type NW.

(NW controller 92)
The loop failure detection system according to the present embodiment includes the NW controller 92 to solve the transmission destination NW device address at the time of encapsulation. The encapsulation unit 16 of each NW device 91 performs the encapsulation when receiving the frame from the subscriber port 15, and inquires the NW controller 92 to determine the transmission destination address for the encapsulation. 29 and 30 show an example of a rough flow.

  The destination NW device address to be encapsulated is searched from “table (T1)” and “table (T2)” (S141-1). If the search fails, that is, if there is no search result, it is searched whether there is a registration as the own NW device in the “table (T2)”, and it is determined that the communication is within the own device (S141-2). If the search fails, that is, if there is no search result, the “destination NW address” is set as “destination unknown address”, and the “forwarding destination port” is set as “undefined route” (S142).

  When the search is successful in step S141-1, the transfer destination port is searched from “table (T3)” (S143). If the search is unsuccessful, that is, if there is no search result or the route is indefinite, the transfer destination port is searched from “Table (T2)” (S144). Then, the search result is added to the “table (T3)” and updated (S145).

  The NW controller 92 returns the search result to the inquiring NW device 91 (S147), and adds / updates the record of the inquiry to the “table (T2)” (S148). ) "Is deleted (S149).

  The encapsulation unit 16 of each NW device 91 performs the encapsulation when receiving the frame from the subscriber port 15, and inquires the NW controller 92 to determine the transmission destination address for the encapsulation. At this time, the DB table possessed on the NW controller 92 side will be described. 29, 30 and 33 show examples of the table (T1), the table (T2) and the table (T3). The table (T1) is a correspondence table indicating which NW device belongs to which ring type NW, and records are prepared statically in advance. The columns of “link state of ring port # 1” and “link state of ring port # 2” dynamically change. The table (T2) is a correspondence table indicating what address (registered address) is connected to each port (subscriber port, ring port) of the NW device 91, and in response to an inquiry from the NW device 91. Change dynamically. The table (T3) is a correspondence table indicating to which port of the ring type NW the encapsulated frame is transferred, and dynamically changes in response to an inquiry from the NW device 91.

Details of the processing flow on the NW controller 92 side will be described.
Procedure 1 is performed before step S141. In the procedure 1, when each NW device 91 encapsulates the received frame, if it is not registered in the “capsuling table”, the NW device address is processed as the destination unknown address, and the received frame is Inquires about the “destination NW device address” and “transfer destination port” to be transferred to the controller and encapsulated. The parameters used for the inquiry are as follows.
・ "Own NW device address"
・ "Source address (before encapsulating)"
・ "Destination address (before encapsulating)"
・ "Input port"

The processing on the NW controller 92 side will be described below. FIG. 34 shows an example of the SQR code in each procedure.
Step S141-1: The destination NW device address for the encapsulation is searched from “table (T2)” and “table (T1)”. In particular,
Procedure 2-1: First, a list of “ring type NW-IDs” to which “inquired NW device 91 (own NW device address)” belongs is extracted from “table (T1)”. At this time, when communicating across a plurality of ring-type NWs, in order to prevent returning to the original ring-type NW, the “input port” is “ring port # 1” or “ring port # 2”. “Ring type NW-ID” is excluded.
Procedure 2-2: A list of “NW device addresses” belonging to the “ring type NW-ID” extracted in “procedure 2-1” is extracted from “table (T1)”. On the other hand, “self NW device address” is included in order to cope with the self-device return communication.
Procedure 2-3: Extract a list in which “Destination Address” is registered in “NW Device Address” extracted in “Procedure 2-2” from “Table (T2)” Procedure 2-4: “Table From (T2) ", from the" NW device address "list extracted in" Procedure 2-3 ", only one item whose aging time has not elapsed and whose" update time "is the latest is extracted. The search result is a “destination NW device address” to be encapsulated, and the process proceeds to step 3. In the case of “no search result”, the process proceeds to step 2-A (S141-2).

Procedure 2-A: It is determined whether the communication is within the NW device (own NW device) inquired using “table (T2)” (S141-2). In particular,
・ "NW device address" = "Own NW device address"
・ "Registered address" = "Destination address (before encapsulation)"
・ Extract and search “registered port” of one entry that satisfies all the conditions that the aging time has not passed and the “update time” is the latest one.
If you have search results,
・ "Inquiry NW address (own NW address)" for "destination NW address",
・ Proceed to step 4 (S147) using “forwarding port” as a search result.
In the case of “no search result”, “destination NW address” is set as “destination unknown address” and “transfer destination port” is set as “undefined route”, and the process proceeds to step 4 (S147).

Procedure 3: Search for “forwarding port” of the encapsulated frame (S143 to S145).
Step 3-1: Search from “table (T3)” (S143).
In particular,
"Sender NW device address" = "Own NW device address"
“Destination NW device address” = “Search result of procedure 2 (NW device address)”
The result of extracting and searching for “transfer destination port” of one entry satisfying all of the conditions is set as “transfer destination port”, and the process proceeds to step 4. On the other hand, if “no search result”, the process proceeds to step 3-2.

Procedure 3-2-1: Search from “table (T2)”.
In particular,
・ "NW device address" = "Own NW device address"
・ "Registered address" = "Destination address (before encapsulation)"
-The result of extracting and searching for "Registered port" of one entry that satisfies all the conditions that the aging time has not passed and the "update time" is the latest is the "forwarding port", and the procedure 3-2-2 Proceed to 2. In the case of “no search result”, “forwarding port” is set to “undefined route” and the process proceeds to step 4 (S146).

Procedure 3-2-2: The search result of procedure 3-2-1 is added to “table (T3)” (S145).
In particular,
"Sender NW device address" = "Own NW device address"
“Destination NW device address” = “Search result of procedure 2 (transmission destination NW device address to be encapsulated)”
・ "Destination port" = "Search result of procedure 3-2-1 (registered port)"
Is added to the “table (T3)” as a new record. If it is registered, “transfer destination port” = “search result of procedure 3-2-1 (registered port)”
Update with.

Procedure 4: Returns the result of “transmission destination NW device address” searched in “procedure 2” and “transfer destination port” searched in “procedure 3” to the inquired NW device 91.
Step 6: If the inquired “source address” and “own NW device address” are not registered in the “table (T2)”,
・ "Registered address" = "Sender address"
・ "NW device address" = "Own NW device address"
・ "Registered port" = "Input port"
・ "Update time" = "Current time"
Add a record.
If registered, “Registered port” = “Input port” of the record
・ "Update time" = "Current time"
Update with.
Procedure 7: Delete the record whose aging time (current time−update time)> aging time has passed from “Table (T2)”.

(Operation on the NW controller 92 side)
A process of the NW controller 92 when a notification of a change in the port link state is received will be described. Each NW device 91 notifies the NW controller 92 when the link state of each port changes. Using this, the NW controller 92 detects a failure / recovery of the ring-type NW and switches the path. 35 to 37, a rough processing flow is described.

  When the notification of the change of the link state of the port is received from the NW device, it is determined whether or not the notification port is a ring port, that is, whether or not it is a port registered in the “table (T1)” (S151). If it is a subscriber port, that is, not registered in the “table (T1)”, it is determined whether or not the notification content is “link down” (S152). In the case of “link down”, the record of the port is deleted from the table “(T2)” (S153), and the notified NW device is requested to delete the entry of the port of the “subscriber port encapsulation table”. (S154).

In the “table (T1)”, the “port state” is updated for the record in which “ring port # 1” or “ring port # 2” is “notification port” (S155).
In the “table (T2)”, the “registration port” is updated to “undefined route” for the record whose “registration port” is “notification port” (S156).
In “Table (T3)”, the “transfer destination port” of the record having the NW device address of the ring type NW-ID to which the notification port belongs in “transmission destination NW address” is updated to “undefined route” (S157). . Specifically, it is as follows.
(1) One “ring type NW-ID” is searched from the notification port from “table (T1)”.
(2) The “NW device address” list having the searched “ring type NW-ID” is acquired from “table (T1)”.
(3) In the “table (T3)”, the “transfer destination port” of the record having the “NW device address” list acquired in (2) as the “destination NW device address” is updated to “undefined route”.

It is determined whether the notification content is “link up” (S158). In the case of “link down”, it is notified that a failure has occurred in the ring type NW-ID NW device to which the notification port belongs (S159). Specifically, (1) One “ring type NW-ID” is retrieved from the notification port from “table (T1)”.
(2) The “NW device address” list having the searched “ring type NW-ID” is acquired from “table (T1)”.
(3) From the “NW device address” list acquired in (2), the occurrence of a failure is notified to the corresponding NW device, and the “destination NW device address” in the “subscriber port encapsulation table” of the NW device is A request is made to set the “forwarding port” to “undefined route” for the entry in the list acquired in (2).

It is confirmed that the ring type NW has completely recovered (S160). For example, it is confirmed that no path failure has occurred in the ring type NW. As a condition, determination is made based on whether or not all ring ports of the ring type NW-ID to which the notification port belongs are “link up”. Specifically, it is as follows.
(1) One “ring type NW-ID” is searched from the notification port from “table (T1)”.
(2) From the “table (T1)”, the number of the “port state” having the searched “ring type NW-ID” is “link down” is totaled.
(3) If the total number is “0”, all ring ports are “link up”.

When the ring type NW has completely recovered from the path failure, that is, all are “link up”, the ring type NW-ID NW device to which the notification port belongs is notified that the failure has been recovered (S161). Specifically, it is as follows.
(1) One “ring type NW-ID” is searched from the notification port from “table (T1)”.
(2) The “NW device address” list having the searched “ring type NW-ID” is acquired from “table (T1)”.
(3) From the “NW device address” list acquired in (2), the corresponding NW device is notified of the failure recovery, and the “destination NW device address” in the “subscriber port encapsulation table” of the NW device is A request is made to set the “forwarding port” to “undefined route” for the entry in the list acquired in (2).

If the ring type NW has not completely recovered from the path failure (No in step S160), the ring type NW-ID NW device to which the notification port belongs is notified that the failure has partially recovered (S162). . Specifically, it is as follows.
(1) One “ring type NW-ID” is searched from the notification port from “table (T1)”.
(2) The “NW device address” list having the searched “ring type NW-ID” is acquired from “table (T1)”.
(3) From the “NW device address” list acquired in (2), the corresponding NW device is notified of partial failure recovery as “failure occurrence”, and the “subscriber port encapsulation table” of the NW device For the entry whose “destination NW device address” is the list acquired in (2), a request is made to set “forwarding port” to “undefined route”.
This is processing when a part of the path faults are recovered in the case where multiple path faults occur in the ring type NW. By performing this process, the communication recovery of the NW device in which communication isolation has occurred is attempted.

(Point of invention)
(1) A normal communication frame is re-encapsulated in a ring-type NW unit (for example, using IEEE 802.1ah technology). The frame source address and destination address are limited to NW devices in the ring.
(2) Then, when performing frame transfer in the ring, the source address is overwritten with the NW device address of its own device by XOR operation every time it passes through the NW device in the ring.
(3) From the list of XOR composite values held in each NW device, it is determined whether or not a loop fault has occurred or a bit is garbled.
By implementing these,
• Realize loop detection and loop source identification with normal data frames.
By using the XOR composite value for loop detection, determination can be made regardless of the presence or absence of NW storm.
-By comparing the list and the calculated XOR composite value, it is possible to detect whether or not a bit is garbled, and it is possible to avoid detection of loop faults due to garbled bits.

  The network device of the present disclosure can be realized by a computer and a program, and can be recorded on a recording medium or provided through a network.

  The present disclosure can be applied to the information communication industry.

11L, 11R, 11A: Table processing unit 12: Routing unit 13: NW controller inquiry unit 14L, 14R: Ring port 15: Subscriber port 16: Encapsulation unit 17: Decapsulation unit 81: Ring type NW
82: Network 91: Network device 92: Network controller 93: Database

Claims (3)

When a network device in a ring network receives a communication frame from outside the ring network, the source address and the destination address are encapsulated in the communication frame to obtain the address of the network device in the ring network. A first transfer procedure for transmission after limiting to
When a network device in a ring network receives a communication frame from another network device in the ring network, the transmission source address is an XOR composite value obtained by XORing the transmission source address and the address of the own device. Execute the second transfer procedure to transmit after rewriting,
In the second transfer procedure, a loop failure occurs in the ring network by comparing the transmission source address of the received communication frame with the XOR composite value described in the normal value determination table held in advance. Loop failure detection method for determining the presence or absence of A network device provided in a ring network,
A first transfer for receiving a communication frame from outside the ring network and encapsulating the communication frame and transmitting a transmission source address and a transmission destination address after being limited to addresses of network devices in the ring network And
A second transfer unit that receives a communication frame from another network device in the ring network and rewrites the transmission source address to an XOR composite value obtained by XORing the transmission source address and the address of the own device. And comprising
The second transfer unit compares the transmission source address of the received communication frame with the XOR composite value described in the normal value determination table held in advance, thereby generating a loop failure in the ring network. Network device that determines the presence or absence of   A loop fault detection program for causing a computer to execute each procedure according to claim 1.

Images