JP4086839B2 - Network communication system and failure detection notification method - Google Patents

Description

  The present invention relates to a network communication system and a failure detection notification method for performing data communication using a transmission line connected between network devices, and in particular, transmission / reception transmission lines between network devices are connected independently. In communication between networks and devices with the auto-negotiation function disabled, two independent transmission paths for transmission and reception are connected, and when data is sent and received between each other, the failure of the transmission path is communicated to the partner device. Regarding failure notification.

  Conventionally, in communication using twisted pair cables or optical fiber cables defined in IEEE 802.3 (IEEE: The Institute of Electrical and Electronics Engineers), as shown in FIG. The device 500 and the network device 510 are connected to each other on a one-to-one basis. The transmission path connecting the network devices 500 and 510 uses a transmission path 521 and a transmission path 522 that are independent for data transmission / reception. For example, in 1000Base-SX using a multimode optical fiber, one is for transmission and the other is for reception.

  In this communication method, a method for detecting a communication abnormality (link failure) in a transmission path between two network devices is defined in IEEE 802.3 by an auto-negotiation function. In this method, when a link failure occurs in the transmission path 522 where data is transferred from the network device 510 to the network device 500, the network device 500 on the data receiving side links with the alarm detection function of the received data. Detect failure.

  After detecting the link failure, as the auto-negotiation function, the network device 500 adds the information on the occurrence of the link failure to the transmission data for auto-negotiation to the network device 510 and transmits it to the transmission path 521, thereby The device 510 is notified of the occurrence of a failure. As described above, in auto negotiation, state monitoring is performed between network devices.

  However, although the auto-negotiation function is defined by IEEE 802.3, depending on the communication method, it is permitted to be an optional function or to be used as invalid. Therefore, in communication between network devices, the auto-negotiation function is not necessarily enabled and used. For example, in connection with a network device of a different vendor, there is a problem that the normal operation or the normal operation does not occur due to the compatibility, and the auto negotiation function is sometimes used as invalid.

  In this way, in communication between two network devices with the auto-negotiation function disabled, mutual monitoring between network devices is not performed, and there is a state where a link is established independently for each direction of transmission / reception signal. It will be. For this reason, when a link failure in only one direction has occurred in the transmission path between the network devices, there is no means for reliably detecting the generated failure by the network devices at both ends.

  A method is known that solves this problem using only existing technology without using a new control program (see, for example, Patent Document 1). In this method, as shown in FIG. 6, for example, a signal indicating availability is periodically transmitted from the transmission control unit 531 of the network device 530 to the transmission path 551 during a period in which data is not transmitted. Then, when the reception control unit 532 cannot detect the availability signal from the transmission line 552 for a certain period or more, it is determined that a failure has occurred. Then, under the control of the line control unit 534, the transmission control unit 531 intentionally stops sending any signal to the transmission line 551 that transmits data to the network device 540.

  Therefore, since the network device 540 determines that the input signal is completely stopped in the received data abnormality detection function, it can detect that this link failure has occurred. As described above, in the technique described in Patent Document 1, even when a failure occurs in a transmission path in only one direction between two connected network devices, the failure is surely generated in both network devices. It shows how it can be detected.

JP 2001-103062 (first page-second page, FIG. 1)

  In the failure detection notification method described in Patent Document 1 described above, communication on both sides of the transmission path is stopped even when a failure has occurred on only one transmission path. The communication will be stopped. In this case, there is a problem in that it is not possible to satisfy the desire to operate in only one direction.

  In addition, in a transmission line that uses link aggregation (IEEE 802.3) and bundles a plurality of transmission lines into one to ensure high bandwidth and redundancy, even if one transmission line cannot be used. Since communication is continued using the remaining transmission path, it is not necessary to stop communication in both directions.

  Therefore, an object of the present invention is that a communication path between networks and devices is connected independently. In communication between networks and devices with the auto-negotiation function disabled, a failure occurs in a transmission path in only one direction. In this case, a failure detection notification method that can reliably detect the occurrence of a failure in both network devices without stopping communication in the other direction is provided.

  In order to achieve the above object, according to the present invention, a network device that has detected a transmission path failure stops signal transmission to the transmission path to the network device of the connection partner for a certain period of time (output stop detection time). It has a function to do. In addition, it has a function of storing a transmission signal stopped for a certain period of time in a buffer mounted on the network device.

  In the failure detection notification method of the present invention, transmission / reception transmission paths between network devices are connected independently, and when a transmission path failure occurs in communication between network devices with the auto-negotiation function disabled, transmission is performed. The network device that has detected the path failure stops signal transmission to the transmission path to the network device of the connection partner for the output stop detection time. This output stop detection time is a time during which the connected network device can detect the stop of input data, and the time can be freely set as a variable parameter. Therefore, the partner connection apparatus can detect a transmission path failure.

  Further, the network device is equipped with a buffer, and transmission data writing / reading processing is always performed in a first-in / first-out manner. This buffer has a capacity sufficient to store data for a time longer than the output stop detection time. By providing this buffer, transmission data is accumulated in the buffer when signal transmission to the transmission path to the network device of the connection partner is stopped, and the accumulated data is sequentially transmitted when signal transmission suspension is canceled. To do. Since the buffer has a capacity sufficient to store data for longer than the output stop detection time, it is possible to cause the partner connection apparatus to detect a transmission path failure without damaging the transmission data.

  The first effect of the present invention is that, in the communication between network devices in which the auto-negotiation function is disabled, when both the network devices are notified of a failure using only the existing technology, the conventional technology uses a bidirectional transmission path. Although it has been stopped, according to the present invention, even when an abnormality occurs in one transmission line, it is possible to continue normal communication on the other transmission line without damaging the transmission data. This means that it is possible to construct a network that efficiently uses.

  According to the present invention, in communication between networks and devices, a link aggregation function is used, a plurality of transmission lines are bundled into one, and a transmission line in a configuration that ensures high bandwidth and redundancy is as follows: Even when one transmission line cannot be used, communication can be continued using the remaining transmission line. Therefore, when the failure detection notification method of the present invention is applied to communication between networks and devices using the link aggregation function, even if a transmission line failure occurs, data communication is continued without stopping the normal transmission line. Therefore, it is possible to obtain a second effect that it is possible to construct a network that secures a high bandwidth and uses the transmission path efficiently.

  The third effect of the present invention is that, in the failure detection notification method of the present invention, if the failure detection notification method of the present invention is applied to one of the two network devices, the above-described invention effect 1, 2 can be provided. For example, for a network device to which the failure detection notification method of the present invention is applied, the above effect can be expected without problems even when the network device of the connection partner is a general router, LAN, switch or the like. It can be done.

  The fourth effect of the present invention is that, in the present invention, since it is possible to determine whether the failure occurrence location is transmission / reception of the transmission path, the occurrence occurrence location at the time of failure, failure search, etc. are quickly performed. It will be able to be.

  The fifth effect of the present invention is that it is possible to easily adapt to various communication systems defined in IEEE 802.3 because it can be set according to the communication system to be connected. .

  The sixth effect of the present invention is that the present invention can be obtained easily and economically without the need for new development such as a protocol because the present invention only supports existing technology. .

  The present invention relates to a network communication system in which transmission / reception transmission paths between network devices are independently connected and an auto-negotiation function is disabled. At least one of the network devices is provided with a buffer unit, a reception control unit, an output stop control unit, and a transmission control unit.

  The buffer unit has a capacity capable of accumulating transmission data for at least the output stop detection time which is a time during which the opposite network device can detect the stop of the received data as an alarm. The reception control unit receives data from the opposite network device, but outputs a failure detection signal when it detects that a failure has occurred in the transmission path due to an abnormality in the received signal, and outputs a failure detection signal when it has recovered. Stop.

  The transmission control unit sequentially reads and transmits the data stored in the buffer unit by a first-in first-out method, but stops data transmission when a read processing stop signal is input. The output stop control unit repeats the output stop detection time and the output of the read processing stop signal at fixed time intervals that do not impair the data stored in the buffer unit every time a failure detection signal is input.

  FIG. 1 is a block diagram showing a network communication system to which the present invention is applied. In FIG. 1, a network device 100 is a network device such as a router or a LAN switch, and network devices 200, 300, and 400 are network devices such as client devices connected thereto.

  The network device 100 is provided with interface units 106, 107, 108 in a one-to-one correspondence with the network devices 200, 300, 400, and the interface units 106, 107, 108 are controlled by the host control unit 103. .

  These network devices are connected by, for example, a transmission path 121 for transmitting data from the network device 100 to the network device 200 and a transmission path 122 for transmitting data from the network device 200 to the network device 100, and IEEE802. The communications specified in .3 are performed mutually.

  FIG. 2 shows an embodiment of the network communication system of the present invention. 2, the interface unit 106 includes a transmission control unit 101, a reception control unit 102, a buffer unit 104, and an output stop control unit 105. The network device 200 includes a transmission control unit 201, a reception control unit 202, and a higher-level control unit. 203.

  The reception control unit 102 is an interface unit that receives a signal transmitted from the network device 200. Data received by the reception control unit 102 is sequentially transferred to the host control device 103. In addition, an alarm detection function for received data is provided to detect that a transmission path failure has occurred due to an abnormality in the received signal. In the present invention, this alarm detection function may take any form, and it is sufficient if it has a general alarm detection function.

  For example, a loss of input signal (LOSS OF SIGNAL), loss of synchronization of input frame (LOSS OF FRAME), CRC check error of input data, and the like can be detected. Then, a failure detection signal is output to the output stop control unit 105 as the detected alarm information. In addition, when the alarm that has occurred is recovered, a failure detection signal output to the output stop control unit 105 is stopped.

  The buffer unit 104 is a memory that can be read and written by a first-in first-out method. Signals to be transmitted from the network device 100 to the network device 200 are sequentially written from the host control unit 103 to the buffer unit 104, and are sequentially read from the buffer unit 104 to the transmission control unit 101. The capacity of the memory of the buffer unit 104 is required to be a minimum that can store data for the output stop detection time (details will be described later) or longer.

  The transmission control unit 101 is an interface unit that outputs a signal to be transmitted to the network device 200. Since the signal to be transmitted to the network device 200 is stored in the buffer unit 104, the reading process is sequentially performed and the read signal is output to the network device 200. If a read process stop signal has been received from the output stop control unit 105, the read process to the buffer unit 104 is stopped and no signal is transmitted to the network device 200. Furthermore, a function of restarting the reading process to the buffer unit 104 when the reading process stop signal is not input is provided.

  The output stop control unit 105 receives a failure detection signal from the reception control unit 102. When a failure detection signal is received, a function of outputting a read processing stop signal to the transmission control unit 101 for a predetermined time is provided. This fixed time is hereinafter referred to as output stop detection time for convenience. The output stop detection time is a time during which the partner network device 200 can detect a stop of input data as an alarm.

  That is, it is a time longer than the standard time for determining the stop of input data as a transmission path failure, which is set in the alarm detection function as the system on the network device 200 side. Further, since the time varies depending on the communication system used and the network device to be connected, a function that can be freely set in the output stop control unit 105 as a variable parameter is provided.

  When the network device to be transmitted / received is 200, the upper control unit 103 sequentially writes the transmission data to the buffer unit 104. The reception data has a function of receiving from the reception control unit 102. The function of the host control unit 103 can be appropriately selected depending on what device the network device 200 is. For example, an internetwork device such as a router has a relay engine and receives it. A function of relaying data received from the control unit 102 to another interface unit is provided.

  On the other hand, the transmission control unit 201 of the network device 200 is an interface unit that outputs a signal to be transmitted to the network device 100, and the reception control unit 202 is an interface unit that receives a signal transmitted from the network device 100. Unit 203 controls transmission control unit 201 and reception control unit 202.

FIG. 2 shows an example in which the present invention is applied only to the network device 100 side. However, the present invention is applied to the network device 200 side, that is, the buffer unit 104 and the output stop control unit 105 are used in the network device 100. However, the same effect can be obtained even when the network device 200 is provided.
[Description of Operation 1]
FIG. 3 is a flowchart showing the operation of the network device 100 together with the operation of the network device 200 from the normal operation to the occurrence of a failure, recovery from the failure, and returning to the normal operation state in the network communication system shown in FIG.

  Now, a case will be described in which the network devices 100 and 200 perform mutual communication by Ethernet using the transmission paths 121 and 122. At this time, it is assumed that the network devices 100 and 200 communicate without using the auto-negotiation function.

  Data to be transmitted from the network device 100 to the network device 200 is written from the upper control unit 103 to the buffer unit 104. Next, the data is read from the buffer unit 104 to the transmission control unit 101 by the first-in first-out method. The read signal is sent from the transmission control unit 101 to the opposite network device 200. Further, data received from the network device 200 is received by the reception control unit 102 and then transferred to the host control unit 103. The host control unit 103 performs necessary data transfer as needed, such as transmitting the received data to another interface unit.

  Now, the reception control unit 102 always monitors the received signal for alarm by the alarm detection function (step S1 in FIG. 3). This alarm monitoring detects, for example, a loss of input signal (LOSS OF SIGNAL), loss of synchronization of input frame (LOSS OF FRAME), and CRC check error of input data.

  While in a normal data communication state, the reception control unit 102 passes the received data to the host control unit 103. When a transmission line failure occurs in the transmission line 122 in the link (Yes in step S2), in the reception control unit 102 of the network device 100, the alarm detection function detects a failure in the transmission line 122 (step S3 in FIG. 3). ). The reception control unit 102 that has detected the transmission path failure discards the received data. Therefore, although data is flowing on the transmission path 222, it is substantially the same as when transmission from the network device 200 is stopped.

  In addition, the reception control unit 102 that has detected the transmission path failure transmits a failure detection signal to the output stop control unit 105 (step S4). The output stop control unit 105 that has received the failure detection signal outputs a read process stop signal to the transmission control unit 101 during the output stop detection time in order to stop the reading process to the buffer unit 104 (step S5). The transmission control unit 101 stops the reading process and completely stops the output signal to the network device 200 in accordance with the reading process stop signal (step S6). During this time, transmission data written from the higher control unit 103 to the buffer unit 104 is stored in a memory in the buffer unit 104.

  On the other hand, the reception control unit 202 of the network device 200 also constantly monitors the received signal for alarm using the alarm detection function (step S21). This alarm monitoring is the same as that in step S1, but the detection of the disconnection of the input signal can include the one caused by the stop of the transmission data output (step S6).

  That is, when the state where all input signals from the transmission path 121 in the link are stopped is continued for the output stop detection time (Yes in step S22), this input data stop can be detected as a transmission path failure. By setting the time, the network device 200 can determine that a transmission path failure has occurred (step S23). With the above method, it is possible to detect the occurrence of a transmission path failure in both the network device 100 and the network device 200.

  When the output stop detection time has elapsed, the output stop control unit 105 stops outputting the read processing stop signal notified to the transmission control unit 101 (step S7), restarts the read processing to the buffer unit 104, and The signal output process to the device 200 is resumed (step S8).

  While the reading process from the transmission control unit 101 is stopped, the transmission data written from the higher control unit 103 to the buffer unit 104 is stored in the memory in the buffer unit 104. Therefore, the reading process to the buffer unit 104 is performed. As a result, the stored data is also output to the network device 200 in order. Since the memory capacity of the buffer unit 104 has a capacity for storing data for the output stop detection time, the communication on the transmission path 121 can be continued without damaging the transmission data. As a result, the communication on the transmission line 122 is stopped during the transmission line failure, but the communication on the transmission line 121 is continued.

  Even after transmission path failure detection (step S3), the reception control unit 102 continues to monitor the received signal in the same manner as in step S1 (step S9). However, even if a transmission line abnormality is detected during this period, it has already been transmitted to the network device (step S6), and is not transmitted again.

When the transmission line abnormality disappears (No in step S10), the transmission line abnormality is recovered, so that the process returns to step S1, and the reception control unit 102 passes the received data to the host control unit 205.
[Description 2 of operation]
In the failure detection notification method described above, the transmission line failure on the network device 200 side is caused by the transmission line 121 or is caused by the transmission line 122 (leading to the continuation of the stop of the input signal). I can't identify any of them. In order to solve this, a method for enabling identification of a faulty part of the transmission path is shown in the flowchart of FIG.

  Steps T8 in FIG. 4 are the same as those in the flowchart in FIG. 3, but after the transmission data is restarted (step T8) due to the stop of the output of the read stop signal (step T7 in FIG. 4), the network device 100 If the transmission path failure is continued in the alarm detection function (step T9) in the reception control unit 102 (Yes in step T10), a read stop signal is output again from the output control unit 105 after the output stop interval has passed ( Step T5). As a result, transmission data to the network device 200 is again stopped for the output stop detection time (step T6). Thereafter, the output stop / output stop release cycle (steps T5 to T10) is repeated.

  In this failure detection notification method, the network device 200 side can confirm that transmission path failure detection due to the stop of the input signal occurs periodically. According to the configuration of FIG. 2, since it is clear that the periodic transmission line failure occurring in the output stop interval is a transmission line failure of the transmission line 122, it is detected as a failure part of the transmission side transmission line 122. It is possible to perform (step T26). On the other hand, if the detected transmission line failure is not periodic but is a single failure detection or continuous failure detection, it can be detected as a failure point in the transmission side transmission line 121 (step T25).

  If the output stop interval is too short, the writing of the next data starts before all the data stored for the output stop time is output to the buffer unit 104, so according to the network load and system configuration, A mechanism that can be set in the output stop control unit 105 as a variable parameter so that it can be freely adjusted is provided.

  Thereafter, the network device 200 when the failure is recovered determines that the data has been normally input by confirming that the data has been normally input for a time longer than the output stop interval. be able to. When the failure is recovered, the reception control unit 102 of the network device 100 can confirm data reception from the network device 200.

  Next, a specific setting example of the output stop interval will be shown. In this example, the distance delay time of the transmission line is not taken into consideration, and a “zero” delay is assumed. Further, when the communication between the network and the device is performed according to the communication standard 100Base-T, the bandwidth of the transmission path is 100 Mbps, and the use band of data transmission / reception is 80% as a network load. In addition, it is assumed that the standard time for determining the stop of input data as a transmission path failure in the network device 200 is 1 sec. Therefore, the output stop detection time set in the output stop control unit 105 is 1 sec (in this example, no margin is considered).

  Here, the memory capacity of the buffer unit 104 of the network device 100 is 100 Mbit because it is sufficient to have a capacity for 1 sec which is the output stop detection time. Next, signal transmission is stopped for 1 sec due to a transmission path failure, and the output stop interval is calculated from the state where signal output is resumed from the state where 100% transmission data is stored in the buffer unit 104.

  Since the bandwidth usage rate is 80%, the maximum bandwidth from which transmission data stored in the buffer unit 104 is read is 20%, and therefore 100 Mbps × 20% = 20 Mbps. Therefore, the time until all the transmission data stored in the buffer unit 104 is read is 100 Mbit ÷ 20 Mbps = 5 sec. As a result, it can be seen that the output stop interval may be set to 5 seconds or more.

Block diagram of a network communication system to which the present invention is applied The block diagram which shows one Example of the network communication system of this invention The flowchart which shows the 1st operation example of this invention. The flowchart which shows the 2nd operation example of this invention. Block diagram of the first conventional example Block diagram of second conventional example

Explanation of symbols

100, 200 Network device 101, 201 Transmission control unit 102, 202 Reception control unit 103, 203 Upper control unit 104 Buffer unit 105 Output stop control unit 300, 400 Network device 121-126 Transmission path

Claims (6)

In a network communication system in which transmission / reception transmission paths between network devices are independently connected and the auto-negotiation function is disabled,
At least one of the network devices,
A buffer unit having a capacity capable of storing transmission data for at least the output stop detection time which is a time during which the opposite network device can detect the stop of the received data as an alarm; and
A reception control unit that receives data from the opposite network device but outputs a failure detection signal when detecting that a failure has occurred in the transmission path due to an abnormality in the received signal, and stops the failure detection signal when the failure is recovered; ,
An output stop control unit that outputs a read processing stop signal only once when the failure detection signal is input;
A network comprising: a transmission control unit that sequentially reads and transmits data stored in the buffer unit by a first-in first-out method, but stops data transmission when the read processing stop signal is input Communications system. In a network communication system in which transmission / reception transmission paths between network devices are independently connected and the auto-negotiation function is disabled,
At least one of the network devices,
A buffer unit having a capacity capable of storing transmission data for at least the output stop detection time, which is a time during which the opposite network device can detect a stop of received data as an alarm; and
A reception control unit that receives data from the opposite network device but outputs a failure detection signal when detecting that a failure has occurred in the transmission path due to an abnormality in the received signal, and stops the failure detection signal when the failure is recovered; ,
An output stop control unit that repeats the output stop detection time and the output of the read processing stop signal at a constant time interval that does not impair the data stored in the buffer unit each time the failure detection signal is input;
A network comprising: a transmission control unit that sequentially reads and transmits data stored in the buffer unit by a first-in first-out method, but stops data transmission when the read processing stop signal is input Communications system. In at least one of the network devices, a failure detection notification method in a network communication system in which transmission / reception transmission paths between network devices are independently connected and the auto-negotiation function is disabled,
Storing transmission data in a buffer having a capacity for storing transmission data for a time equal to or longer than an output stop detection time, which is a time during which an opposing network device can detect a stop of reception data as an alarm; and
Receiving data from the opposite network device and outputting a failure detection signal upon detecting that a failure has occurred in the transmission path due to an abnormality in the received signal;
When the failure detection signal is input, the output stop detection time, outputting the read processing stop signal only once,
The data stored in the buffer unit is sequentially read and transmitted in a first-in first-out method, but the data transmission is stopped when the read processing stop signal is input,
And a step of stopping the failure detection signal when the failure is recovered. A failure detection notification method in a network communication system in which transmission / reception transmission paths between network devices are independently connected and the auto-negotiation function is disabled, in at least one of the network devices,
Storing transmission data in a buffer having a capacity for storing transmission data for a time equal to or longer than an output stop detection time, which is a time during which an opposing network device can detect a stop of reception data as an alarm; and
Receiving data from the opposite network device and outputting a failure detection signal upon detecting that a failure has occurred in the transmission path due to an abnormality in the received signal;
Repeating the output stop detection time and the output of the read processing stop signal at fixed time intervals that do not impair the accumulated data in the buffer unit each time the failure detection signal is input;
The data stored in the buffer unit is sequentially read and transmitted in a first-in first-out method, but when the read processing stop signal is input, the data transmission is stopped;
And a step of stopping the failure detection signal when the failure is recovered. A program executed in at least one of the network devices in a network communication system in which transmission / reception transmission paths between network devices are independently connected and the auto-negotiation function is disabled,
A procedure for storing transmission data in a buffer having a capacity for storing transmission data for a time equal to or longer than an output stop detection time, which is a time during which an opposing network device can detect a stop of received data as an alarm; and
A procedure for receiving a data from the opposite network device and outputting a failure detection signal upon detecting that a failure has occurred in the transmission path due to an abnormality in the received signal;
When the failure detection signal is input, the output stop detection time, a procedure for outputting a read processing stop signal only once,
A procedure for sequentially reading and transmitting data stored in the buffer unit in a first-in first-out method, but stopping data transmission when the reading process stop signal is input,
And a procedure for stopping the failure detection signal when recovered. A program executed in at least one of the network devices in a network communication system in which transmission / reception transmission paths between network devices are independently connected and the auto-negotiation function is disabled,
A procedure for storing transmission data in a buffer having a capacity for storing transmission data for a time equal to or longer than an output stop detection time, which is a time during which an opposing network device can detect a stop of received data as an alarm; and
A procedure for receiving a data from the opposite network device and outputting a failure detection signal upon detecting that a failure has occurred in the transmission path due to an abnormality in the received signal;
A procedure for repeating the output stop detection time and the output of the read processing stop signal at a constant time interval that does not impair the accumulated data in the buffer unit each time the failure detection signal is input;
A procedure for sequentially reading and transmitting data stored in the buffer unit in a first-in first-out method, but stopping data transmission when the reading process stop signal is input,
And a procedure for stopping the failure detection signal when recovered.

Images