JP5780461B2 - Network relay system and control method of network relay system - Google Patents

Description

  The present invention relates to a network relay system and a method for controlling the network relay system.

  Conventionally, a system is known in which a first network relay device group including a plurality of network relay devices functions as a fabric node, and a second network relay device group including a plurality of network relay devices functions as a line node ( For example, see FIG. 8 of Patent Document 1.)

JP 2009-290271 A

  In the prior art, a failure may occur in any of the physical lines connecting the network relay device included in the first network relay device group and the network relay device included in the second network relay device group. is assumed.

  Therefore, the present invention provides a network relay system including a plurality of interface relays and a plurality of fabric relays connected to the plurality of interface relays, and is capable of switching a communication path when a failure occurs. A system and a control method for a network relay system are provided.

  In order to solve the above-described problem, according to a first aspect of the present invention, a plurality of interface repeaters and a plurality of fabric repeaters connected to the plurality of interface repeaters are provided, and the interface repeater includes: , LAG setting means for setting a link aggregation group at a port connected to the fabric repeater, and the fabric repeater transmits a first failure notification frame to the interface repeater. And the LAG setting means of the interface repeater excludes the port that has received the first failure notification frame from the link aggregation group when the first failure notification frame is received. Is provided.

  In order to solve the above problem, according to the second aspect of the present invention, a network relay system comprising a plurality of interface relays and a plurality of fabric relays connected to the plurality of interface relays The interface repeater sets a link aggregation group for a port connected to the fabric repeater, and the fabric repeater sends a first failure notification frame to the interface repeater. When the interface relay unit receives the first failure notification frame, the interface relay unit excludes the port that has received the first failure notification frame from the link aggregation group. Is done.

  According to the network relay system and the network relay system control method of the present invention, when a failure occurs in a network relay system including a plurality of interface relays and a plurality of fabric relays connected to the plurality of interface relays. In addition, the communication path can be switched.

1 is a diagram schematically illustrating a configuration example of a network relay system. It is a block diagram which shows roughly the functional structure of FS. It is a block diagram which shows schematically the functional structure of IFS. It is a figure for demonstrating the basic operation example of a network relay system. FIG. 10B is a diagram (1/2) for explaining operation example 1 when a failure is detected by the network relay system; FIG. 10B is a diagram (2/2) for explaining the operation example 1 when a failure is detected by the network relay system; It is a flowchart which shows the implementation | achievement method of the failure notification shown in FIG. It is a flowchart which shows the implementation | achievement method of the reset of the link aggregation group shown in FIG. It is a figure which shows the LAG setting table of each IFS before and after a communication failure generate | occur | produces in the operation example 1. FIG. FIG. 10B is a diagram (1/2) for explaining operation example 2 when a failure is detected by the network relay system; FIG. 10B is a diagram (2/2) for explaining the operation example 2 when a failure is detected by the network relay system; It is a flowchart which shows the implementation | achievement method of the failure notification shown in FIG. It is a flowchart which shows the implementation | achievement method of the failure notification shown in FIG. FIG. 10B is a diagram (1/2) for explaining operation example 3 when a failure is detected by the network relay system; FIG. 10B is a diagram (2/2) for explaining the operation example 3 when a failure is detected by the network relay system; It is a flowchart which shows the implementation | achievement method of the failure notification shown in FIG.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram schematically illustrating a configuration example of a network relay system 10 according to an embodiment. The network relay system 10 is suitable for a purpose of relaying a network frame transmitted / received between a plurality of terminal devices 12 in a facility such as a data center. The terminal device 12 is a server, a network repeater, etc., for example. The number of connected terminal devices 12 and the type thereof are not particularly limited.

The network relay system 10 includes, for example, about 24 IFSs (Interface Face Switches) 14a, 14b,. The network relay system 10 includes, for example, eight FSs (Fabric Switches) 16a to 16h as a plurality of fabric relays. The number of IFSs 14a to 14x and FSs 16a to 16h mentioned here is an example, and the present invention is not limited to this.
In the following, “IFS 14a to 14x” as an interface relay may be referred to as “IFS 14”, and “FS 16a to 16h” as a fabric relay may be referred to as “FS 16”.

  In this embodiment, for example, box-type switching hubs are used as the IFSs 14a to 14x and the FSs 16a to 16h. The switching hub used here is provided with layer 2 and layer 3 data transfer functions of, for example, an OSI (Open Systems Interconnection) reference model. Note that the basic structures and functions of the IFSs 14a to 14x and the FSs 16a to 16h may be common to each other.

  A plurality of terminal devices 12 are distributed and connected to any one of the IFSs 14a to 14x. Note that the number of terminal devices 12 connected to each of the IFSs 14a to 14x may not be uniform. In addition, there is no restriction on what kind of terminal device 12 is connected to which IFS 14a to 14x. Further, the terminal device 12 may be a personal computer or a workstation other than the server.

  In the network relay system 10, the IFSs 14a to 14x and the FSs 16a to 16h are connected to each other by physical lines. The physical line is, for example, an optical fiber cable or a twisted pair cable. In FIG. 1, reference numerals are omitted for individual physical lines. Although some of the IFSs 14a to 14x are not shown, they are also individually connected to the FSs 16a to 16h via physical lines. The physical line is connected to ports (not shown) of the IFSs 14a to 14x and the FSs 16a to 16h.

  In each of the IFSs 14a to 14x, link aggregation groups, that is, LAGs 18a, 18b,..., 18x are set to the ports to which the FSs 16a to 16h are connected. Link aggregation is a function that treats a plurality of physical lines logically as one line. In each of the IFSs 14a to 14x, the plurality of ports to which the LAGs 18a, 18b, ..., 18x are set operate logically as one port.

  When each IFS 14a to 14x transmits the network frame received from each terminal device 12 to the FSs 16a to 16h, the network frame from any port within the corresponding LAG 18a, 18b,. Decide to send. In this algorithm, when a network frame is received from the connected terminal device 12, an operation (for example, four arithmetic operations) is performed using a MAC address, an IP address, etc. indicating destination information and transmission source information stored in the frame. The port number for transmission is determined from the value obtained thereby. At this time, an identification number (INDEX) may be set for each port in the LAGs 18a, 18b,..., 18x in advance, and the value obtained by the above calculation may be associated with the port identification number. In any case, when each IFS 14a-14x uniquely determines a port to be used for transmitting the network frame based on the unique information in the network frame, the network frame received from the terminal device 12 is transmitted to the FSs 16a-16h. The network communication system 10 as a whole secures a necessary communication band.

Further, in the network relay system 10 of the present embodiment, when a network frame is transmitted and received between any two terminal devices 12 by the above algorithm, communication at the time of transmission (uplink) and communication at the time of reception (downlink) The paths are matched with each other.
For example, consider a case where two terminal devices 12 respectively connected to two different IFS 14a and IFS 14b transmit and receive network frames. In this case, the network frame transmitted from the terminal device 12 (source) connected to the IFS 14a is transferred to the IFS 14b via the specific FS 16d, for example, by the above algorithm, and the destination (destination) terminal from the IFS 14b. It is transferred to the device 12. At this time, in the IFS 14b, since the LAG 18b is set to the ports facing all the FSs 16a to 16h, the network frame is not transmitted back from the IFS 14b to the LAG 18b.

  Conversely, when the terminal device 12 connected to the IFS 14b is the transmission source, the network frame is transmitted from the IFS 14b to the FS 16d, and further transmitted from the FS 16d to the destination terminal device 12 through the IFS 14a. Here again, in the IFS 14a, since the LAG 18a is set to the ports facing all the FSs 16a to 16h, the network frame is not transmitted back in the LAG 18a.

  As described above, when a network frame is transmitted and received between two terminal devices 12 connected to two different IFSs 14a and IFS 14b, ports used in the IFS 14a and IFS 14b that relay the network frame are bidirectionally matched. Here, the combination of the two terminal devices 12 connected to the IFS 14a and the IFS 14b has been described as an example. Similarly, when transmitting / receiving between the terminal devices 12 of other combinations, bidirectional transmission and reception are also performed. The communication path matches the port used.

  In addition, when a VLAN (Virtual Local Area Network) is set for each terminal device 12 connected to the network relay system 10, the IFSs 14a to 14x and the FSs 16a to 16h each transmit and receive a network frame using a tag VLAN in principle. I do. At this time, all VLAN information assigned to each terminal device 12 is registered in each port of the FSs 16a to 16h. When transmitting network frames, the IFSs 14a to 14x and the FSs 16a to 16h transmit VLAN information in a tagged state according to each received network frame. Thereby, even if a VLAN is set for each terminal device 12, the communication path at the time of transmission and the communication path at the time of reception match between different terminal devices 12.

  The above is the basic configuration and network frame relay operation of the network relay system 10 of the present embodiment. In addition, in the present embodiment, a function (failure detection function) for quickly detecting a failure (for example, a physical line abnormality) in the communication state between the plurality of IFSs 14a to 14x and the FSs 16a to 16h occurs. This is realized in each IFS 14a to 14x and each FS 16a to 16h.

  Hereinafter, the configurations of the IFSs 14a to 14x and the FSs 16a to 16h will be described in detail. As described above, hereinafter, “IFS14a to 14x” and “FS16a to 16h” may be collectively referred to as “IFS14” and “FS16”, respectively.

[Fabric repeater]
FIG. 2 is a block diagram schematically showing a functional configuration of the FS 16. That is, the basic configurations and functions of the individual FSs 16a to 16h are common.

  The FS 16 has a plurality of ports 46a, 46b,. It is assumed that the number of ports of the FS 16 covers the number of installed IFSs 14a to 14x in the network relay system 10. In the following, “ports 46a, 46b,..., 46x” are also referred to as “ports 46”. The number of ports 46 is not limited to the number shown.

In addition to the port 46, the FS 16 has a frame transfer processing unit 48 and a memory unit 50.
Among these, the memory part 50 is comprised by RAM (random access memory) which can be rewritten, for example. The memory unit 50 stores an FDB (forwarding database) 50a. The memory unit 50 stores identification information 50b for identifying the FS 16 and the port 46.

The IFSs 14a to 14x are connected to the ports 46, respectively. When the port 46 receives the network frame transmitted from each of the IFSs 14 a to 14 x, the port 46 transfers the received network frame to the frame transfer processing unit 48.
Each port 46 transmits the network frame transferred from the frame transfer processing unit 48 to any of the IFSs 14a to 14x.

  The frame transfer processing unit 48 is realized by an integrated circuit such as an FPGA (Field Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit).

  The frame transfer processing unit 48 transfers the network frame to the transfer destination port 46 based on the destination information in the network frame. The frame transfer processing unit 48 creates and updates the FDB 50a based on the source address information in the network frame.

In the FDB 50a, the transmission source information in the network frame received at the port 46 and the number of the port 46 that received the network frame are registered in association with each other.
In other words, when the network frame is transferred from the port 46, the frame transfer processing unit 48 relays the network frame to the transfer destination port 46 based on the information registered in the FDB 50a.

  The FS 16 further includes a detection frame generation unit 52, a failure determination unit 54, and a failure notification frame generation unit 56 as a configuration for monitoring the communication state (mainly detecting a failure). Among these, the detection frame generation unit 52 is configured as a detection frame transmission unit (second detection frame transmission unit) included in each FS 16. The failure determination unit 54 is configured as a failure determination unit (second failure determination unit) included in each FS 16. Moreover, the failure notification frame generation unit 56 is configured as a failure notification frame transmission unit included in each FS 16.

  The detection frame generation unit 52 generates a detection frame (second detection frame) in order to monitor the communication state with the IFSs 14 a to 14 x connected to the port 46. At this time, the detection frame includes identification information (MAC address or the like) for identifying the source FS16 (individual FSs 16a to 16h) and identification information for identifying the ports 46a, 46b,. For example, a port number) is stored.

  The detection frame generated by the detection frame generation unit 52 is transmitted from each port 46 to the connected IFSs 14a to 14x via the frame transfer processing unit 48. At this time, the detection frame is transmitted from each port 46 every predetermined time (for example, every several tens of milliseconds to several seconds).

When the failure determination unit 54 detects a communication state failure, the failure determination unit 54 causes the failure notification frame generation unit 56 to generate a failure notification frame.
Here, in the present embodiment, the detection frames are transmitted from the IFSs 14a to 14x to the connection destinations FSs 16a to 16h every predetermined time (every tens of milliseconds to several seconds). In other words, the IFSs 14a to 14x and the FSs 16a to 16h monitor each other's communication state by transmitting and receiving detection frames to each other.
Accordingly, when the reception of the detection frame cannot be confirmed at any port 46 within the predetermined time, the failure determination unit 54 of the FS 16 determines that a failure has occurred.

The failure notification frame generation unit 56 generates a failure notification frame (first failure notification frame) to notify the IFSs 14a to 14x that a communication failure has occurred. At this time, the identification information 50b stored in the memory unit 50 is stored in the failure notification frame.
That is, the failure notification frame identifies identification information indicating that a communication failure has occurred, identification information identifying the FSs 16a to 16h of the transmission source, and ports 46a, 46b,..., 46x that transmit the failure notification frame. The identification information to be stored is stored.

  In FS 16, the failure notification frame generated by the failure notification frame generation unit 56 is transmitted from all the ports 46 a, 46 b,..., 46 x connected to the IFSs 14 a to 14 x via the frame transfer processing unit 48.

[Interface repeater]
FIG. 3 is a block diagram schematically showing a functional configuration of the IFS 14. That is, the basic configurations and functions of the individual IFSs 14a to 14x are common. The IFS 14 and FS 16 have the same basic configuration and function as a switching hub. In addition, FIG. 3 shows a characteristic configuration of the IFS 14.

  The IFS 14 has a plurality of ports 58a to 58j. The IFS 14 includes a frame transfer processing unit 60, a memory unit 62, and a LAG setting unit 64.

  The memory unit 62 is configured by a RAM, for example. The memory unit 62 stores an FDB 62a and a LAG setting table 62b. The memory unit 62 collectively stores identification information indicating that a failure has occurred in the communication state, identification information identifying the IFSs 14a to 14x, identification information identifying the ports 58a to 58j, and the like as identification information 62c. ing.

  In FIG. 3, ports 58a to 58h shown above the frame transfer processing unit 60 are connected to the FSs 16a to 16h, respectively. Ports 58 i and 58 j shown below the frame transfer processing unit 60 are connected to different terminal devices 12, respectively. Hereinafter, “ports 58a to 58j” are also referred to as “ports 58”.

  In the IFS 14, the FSs 16a to 16h are connected to the ports 58a to 58h in the order of arrangement. The IFSs 14a to 14x shown in FIG. 1 are mutually connected to the FSs 16a to 16h basically in the arrangement order shown in FIG. That is, in this embodiment, in each IFS 14a-14x, it is preferable that the same FS16a-16h is connected to the same port (port of the same port number).

  In FIG. 3, ten ports 58a to 58j are shown, but the number of ports is not limited to the number shown. The number of ports covers the number of installed FSs 16 a to 16 h in the network relay system 10.

  The ports 58a to 58h transfer the network frame received from the FSs 16a to 16h to the frame transfer processing unit 54. The ports 58a to 58h transmit the network frame transferred from the frame transfer processing unit 60 to the connection destinations FSs 16a to 16h.

  The other ports 58 i and 58 j each transfer the network frame received from the terminal device 12 to the frame transfer processing unit 60. The ports 58 i and 58 j transmit the network frame transferred from the frame transfer processing unit 60 to the terminal device 12 that is the connection destination.

The frame transfer processing unit 60 is realized by an integrated circuit such as an FPGA or an ASIC, for example, similarly to the frame transfer processing unit 48 of the FS 16 shown in FIG.
The frame transfer processing unit 60 transfers the network frame to the transfer destination port 58 based on the destination information in the network frame. The frame transfer processing unit 60 creates and updates the FDB 62a based on the source address information in the network frame.

  In the FDB 62a, the transmission source information in the network frame received at the port 58 and the number of the port 58 that received the network frame are registered in association with each other.

  The frame transfer processing unit 60 of the IFS 14 relays the network frame to the ports 58i and 58j connected to the destination terminal device 12 with reference to the FDB 62a. Further, the frame transfer processing unit 60 refers to the LAG setting table 62b together with the FDB 62a, and relays the network frame to the ports 58a to 58h connected to the destination FSs 16a to 16h.

[Transfer processing within the same IFS 14]
For example, it is assumed that the terminal device 12 connected to the port 58i sends data to the terminal device 12 connected to another port 58j. In this case, the terminal device 12 serving as a transmission source transmits a network frame including data to the IFS 14. The transmitted network frame is received by the port 58 i and transferred to the frame transfer processing unit 60.

  The frame transfer processing unit 60 refers to the FDB 62a based on the destination information and the transmission source information stored in the network frame. At this time, when the destination information indicating another terminal device 12 is registered in the FDB 62a in association with the port 58j, the frame transfer processing unit 60 relays the network frame to the port 58j. The port 58j transmits a network frame to the terminal device 12 that is the connection destination.

[Transfer processing between different IFSs 14]
When the terminal device 12 connected to the port 58i transmits data to the terminal device 12 connected to the other IFSs 14a to 14x, the network frame transmitted from the terminal device 12 as the transmission source is received by the port 58i. To the frame transfer processing unit 60.

  When the destination information stored in the network frame is registered in association with the ports 58a to 58h constituting the LAG 18, the frame transfer processing unit 60 refers to the LAG setting table 62b. Here, “LAG 18a, 18b,..., 18x” shown in FIG. 1 are collectively referred to as “LAG 18”.

  The frame transfer processing unit 60 selects a port 58 to transmit a network frame from among the plurality of ports 58 registered in the LAG setting table 62b based on a predetermined algorithm. The frame transfer processing unit 60 relays the network frame to the selected port 58, and the selected port 58 transmits the network frame to the connection destination FSs 16a to 16h.

  In addition, the LAG setting unit 64 as the LAG setting unit creates and updates the LAG setting table 62b. Specifically, the LAG setting unit 64 registers the ports 58a to 58h connected to the FSs 16a to 16h in the LAG setting table 62b in association with the group number (LAG ID). That is, each port 58 registered in the LAG setting table 62b belongs to a common link aggregation group (LAG18).

  The IFS 14 further includes a detection frame generation unit 66, a failure determination unit 68, and a failure notification frame generation unit 70 as a configuration for monitoring the communication state (in particular, detecting a failure). Among these, the detection frame generation unit 66 is configured as a detection frame transmission unit (first detection frame transmission unit) included in each IFS 14. The failure determination unit 68 is configured as a failure determination unit (first failure determination unit) included in each IFS 14. The failure notification frame generation unit 70 is configured as a failure notification unit included in each IFS 14.

  Similarly to the FS 16 illustrated in FIG. 2, the detection frame generation unit 66 generates a detection frame (first detection frame) in order to monitor the communication state with the FSs 16 a to 16 h connected to the ports 58 a to 58 h. At this time, the detection frame stores identification information for identifying the transmission source IFSs 14a to 14x and identification information for identifying the ports 58a to 58h that transmit the detection frame. That is, the detection frame generation unit 66 generates a detection frame in which the identification information 62b stored in the memory unit 62 is stored.

  The detection frame generated by the detection frame generation unit 66 is transmitted from the ports 58a to 58h to the connection destination FSs 16a to 16h via the frame transfer processing unit 60 every predetermined time (every tens of milliseconds to several seconds). Sent.

When the failure determination unit 68 detects a communication state failure, the failure determination unit 68 causes the failure notification frame generation unit 70 to generate a failure notification frame (second failure notification frame).
Specifically, when the failure determination unit 68 cannot confirm the reception of the detection frame transmitted from the FS 16a to 16h within the predetermined time at any of the ports 58a to 58h, a failure occurs in the communication state. It is determined that

  When the failure determination is made, the failure notification frame generation unit 70 generates a failure notification frame to notify the FSs 16a to 16h that a failure has occurred in the communication state. At this time, the failure notification frame stores identification information indicating that a communication failure has occurred, identification information identifying the transmission source IFSs 14a to 14x, and identification information identifying the ports 58a to 58h that transmit the failure notification frame. ing.

Here, in the FS 16 described above, the failure notification frame is transmitted from all the ports 48 to which the IFS 14 is connected.
However, in the IFS 14, the failure notification frame generated by the failure notification frame generation unit 70 is transmitted only from the port 58 in which the reception of the detection frame cannot be confirmed. Therefore, the failure notification frame is not transmitted from the port 58 receiving the detection frame.

The port 58 receiving the detection frame from the FS 16 transmits the detection frame even after a failure occurs.
As a method for notifying the FSs 16a to 16h that a failure has occurred in the communication state, the failure determination unit 68 does not cause the failure notification frame generation unit 70 to generate a failure notification frame, but instead of the detection frame. There is also a method of stopping the transmission of the detection frame from the port 58 for which reception cannot be confirmed. An example of the operation of the failure determination unit 68 will be described later.

  In the present embodiment, the LAG setting unit 64 updates the LAG setting table 62b in order to avoid discarding the network frame due to the communication failure that has occurred. Specifically, when a failure notification frame is received from the connection destination FS 16a to 16h at any of the ports 58a to 58h, the LAG setting unit 64 corresponds to the identification information indicating the transmission source FS 16a to 16h and the corresponding information. Ports 58a to 58h to be excluded (deleted) from the LAG setting table 62b. In other words, the LAG setting unit 64 excludes the port 58 that has received the failure notification frame from the members of the LAG 18.

Then, the LAG setting table 62 b updated by the LAG setting unit 64 is newly stored (overwritten) in the memory unit 62.
Further, the frame transfer processing unit 60 refers to the updated LAG setting table 62b and transmits the network frame to the connection destination FSs 16a to 16h.
Therefore, the network frame transmitted / received at the port 58 (the port 58 that received the failure notification frame) before the occurrence of the communication failure is transmitted / received via the other port 58 newly determined based on a predetermined algorithm. Frame discarding is avoided. This is an example of failure avoidance due to degeneration of LAG18.

  Hereinafter, an operation example by the network relay system 10 will be described by dividing it into a plurality of patterns.

[Basic operation example]
FIG. 4 is a diagram for explaining a basic operation example of the network relay system 10. In FIG. 4 and subsequent figures, in order to prevent complication, illustration of some of the IFSs 14a to 14x and the FSs 16a to 16h is omitted as appropriate. On the other hand, in order to facilitate understanding of the operation example, the ports 46a to 46x and the ports 58a to 58h related to the description are added.

  When the communication state of the network relay system 10 is good, the IFSs 14a to 14x and the FSs 16a to 16h transmit and receive the detection frames 72 and 74 to each other at predetermined intervals. Specifically, the detection frames 72 are transmitted from the IFSs 14a to 14x toward the FSs 16a to 16h, and the detection frames 74 are also transmitted from the FSs 16a to 16h toward the IFSs 14a to 14x. . Note that arrows shown by solid lines and dotted lines in FIG. 4 indicate the transmission directions of the detection frames 72 and 74, respectively. Here, in order to prevent complication, the detection frames 72 and 74 are shown only for some transmission / reception paths (the same applies to the following drawings).

As described above, the network relay system 10 allows the IFSs 14a to 14x and the FSs 16a to 16h to transmit and receive the detection frames 72 and 74 to each other, so that the ports 58a to 58h and the ports 46a to 46x are connected to each other. (Including communication).
In particular, in the present embodiment, LAGs 18a to 18x are set in the ports 58a to 58h of the IFSs 14a to 14x, but the detection frames 74 are transmitted from all of the ports 58a to 58h through the physical lines.
Accordingly, the IFSs 14a to 14x and the FSs 16a to 16h receive the detection frames 74 and 72 (signs are in the order of correspondence, and the same applies hereinafter) within a predetermined time, thereby confirming that the communication state of each other is normal. Can do.

On the other hand, when the reception of the detection frames 74 and 72 cannot be confirmed within a predetermined time by any of the IFSs 14a to 14x or the FSs 16a to 16h, the detection that the reception has not been confirmed by the IFSs 14a to 14x or the FSs 16a to 16h It can be determined that a failure has occurred in the communication state with the transmission source of the frames 74 and 72. Such determination is performed by the failure determination unit 54 shown in FIG. 2 or the failure determination unit 68 shown in FIG.
In the following description, “determining that a failure has occurred” is also referred to as “failure detection”.

[Operation example 1 when a failure is detected]
5 and 6 are diagrams for explaining an operation example 1 when a failure is detected by the network relay system 10. Here, as an example, a case where a failure occurs in the communication state between the IFS 14a and the FS 16a will be described.

  As shown in FIG. 5A, when a failure occurs in the communication state from the IFS 14a to the FS 16a (such as a broken physical line), the detection frame 74 transmitted from the IFS 14a is sent to the FS 16a within a predetermined time. It will not reach. At this time, the FS 16a determines that a failure has occurred in the communication state with the IFS 14a by the failure determination unit 54 shown in FIG.

As shown in FIG. 5B, the FS 16a that has detected a failure transmits a failure notification frame 76 from each port 46a to 46x to each IFS 14a to 14x. A broken arrow shown in FIG. 5B indicates the transmission direction of the failure notification frame 76.
The failure notification frame 76 is generated in the failure notification frame generation unit 56 shown in FIG. At this time, the failure notification frame 76 includes information indicating that a communication failure has occurred, identification information for identifying the FS 16a, and identification information for each port 46a to 46x at the time of transmission.

[When receiving a failure notification frame]
As shown in FIG. 6, when each IFS 14a-14x receives the failure notification frame 76 from the FS 16a, the port 46a that has received this is excluded from the members of the respective LAGs 18a-18x.

  Specifically, the LAG setting unit 64 of each IFS 14a to 14x resets the link aggregation group in a state where the port 46a is excluded from the members. In other words, the LAG setting unit 64 of each IFS 14a to 14x regroups each LAG 18a to 18x. In FIG. 6, not all physical lines are shown in the LAGs 18a to 18x, but the ports 58b to 58h that are not excluded from the members belong to the LAGs 18a to 18x. In FIG. 6, the arrows along the physical line indicate not only the detection frames 72 and 74 but also the transmission direction of the network frame.

  Accordingly, the network frame transmitted from the port 46a before the occurrence of the failure is transmitted from a port (for example, the port 46b) newly determined based on a predetermined algorithm from the other ports 46b to 46h constituting the LAGs 18a to 18x. Will be sent. As a result, the communication band of each of the LAGs 18a to 18x decreases temporarily, but communication can be continued and the discard of the frame can be avoided.

[Method for realizing operation example 1]
FIG. 7 is a flowchart showing a method for realizing the failure notification shown in FIG. FIG. 8 is a flowchart showing a method for realizing the resetting of the link aggregation group shown in FIG.

[Failure notification processing]
First, a method for realizing a failure notification by the FS 16 of the operation example 1 will be described.
In the failure notification process shown in FIG. 7, the failure determination unit 54 of the FS 16 checks whether or not the detection frame 74 has been received at each of the ports 46a to 46x (step S100). When the reception of the detection frame 74 is confirmed (Yes), this process is repeated.

  If reception of the detection frame 74 cannot be confirmed at any of the ports 46a to 46x (step S100: No), the failure determination unit 54 of the FS 16 checks whether or not a predetermined time has elapsed (step S102). When reception of the detection frame 74 is confirmed within a predetermined time (step S102: No), steps S100 to S102 are repeated.

  When the predetermined time has passed in a state where the reception of the detection frame 74 cannot be confirmed (step S100: No), the failure notification frame generation unit 56 of the FS 16 generates a failure notification frame 76. When the FS 16 transmits (floods) the failure notification frame 76 from all the ports 46a to 46x connected to the IFSs 14a to 14x (step S104), the processing ends (END).

[LAG resetting process]
Next, a method for realizing link aggregation group resetting by the IFS 14a of the operation example 1 will be described.

  In the LAG resetting process shown in FIG. 8, the failure determination unit 68 of the IFS 14a checks whether or not the failure notification frame 76 has been received at the ports 58a to 58h connected to the FSs 16a to 16h (step S200). In particular, if the failure notification frame 76 has not been received (No), the processing is temporarily ended here.

  When it is confirmed that the failure notification frame 76 has been received (step S200: Yes), the LAG setting unit 64 excludes the port (here, port 58a) that has received the failure notification frame 76 from the members of the LAG 18a, and regroups it. (Step S202), and the process ends. Here, the failure notification frame 76 is constantly monitored in step S200. However, the processing from step S202 onward is triggered in sequence using the reception of the failure notification frame 76 as a trigger without placing step S200 in particular. It may be executed.

  FIG. 9 is a diagram illustrating the LAG setting table 62b of each of the IFSs 14a to 14x before and after a communication failure occurs in the operation example 1.

[Before failure]
As shown in the “LAG belonging port column” in FIG. 9A, all ports 58a to 58h are set as LAG belonging ports (members) in the LAG setting table 62b of the IFS 14a before the failure occurs. . Further, as shown in the “LAG ID column”, all the ports 58a to 58h belong to a common LAG 18a. In the present embodiment, one LAG is set for each IFS 14a to 14x, but a plurality of LAGs may be set.

[After failure]
When the IFS 14a receives the failure notification frame 76 from the FS 16a at the port 46a, as shown in FIG. 9B, the port 46a is excluded from the members belonging to the LAG 18a and regrouped (updated). As a result, the ports 46b to 46h excluding the port 46a are reset in the “LAG affiliation port column” as affiliation members after update.

[Summary of Operation Example 1]
As described above, according to the operation example 1, the FS 16 that has detected the occurrence of the failure transmits a failure notification frame to the IFS 14, and the IFS 14 performs an operation of excluding the port that has received the failure notification frame from the LAG. The FS 16 connected to the communication path where the failure has occurred is removed from the communication path used. That is, in operation example 1, when a communication failure occurs, the communication path of the network frame that passes through the section where the failure has occurred is switched to another communication route. Thereby, the reliability of communication between the several terminal devices 12 connected to different IFS14a-14x can be improved.

In addition, although the configuration of each LAG 18a to 18x temporarily degenerates when a failure occurs, by applying the above algorithm within the LAGs 18a to 18x after regrouping, the network frames continuously transmitted and received between the terminal devices 12 Relay can continue.
When a physical line or port recovery operation is performed thereafter, the LAGs 18a to 18x can be restored to the initial state, thereby efficiently restarting network frame relay using all the bands. .

  The above operation example 1 relates to the occurrence of a failure in the direction from the IFS 14 to the FS 16. On the other hand, when a failure occurs in the direction from the FS 16 to the IFS 14, the following operation examples 2 and 3 are performed.

[Operation example 2 when a failure is detected]
10 and 11 are diagrams for explaining an operation example 2 when a failure is detected by the network relay system 10. Here, as an example, a case where a failure occurs in the communication state between the IFS 14a and the FS 16a will be described, but the direction of communication in which the failure has occurred is different as described above.

  As shown in FIG. 10A, when a failure occurs in the communication state from the FS 16a to the IFS 14a, the detection frame 72 transmitted from the FS 16a does not reach the IFS 14a within a predetermined time. At this time, the IFS 14a determines that a failure has occurred in the communication state with the FS 16a by the failure determination unit 64 shown in FIG.

  Then, as shown in FIG. 10B, the IFS 14a that has detected the failure generates a failure notification frame 78 in the failure notification frame generation unit 70 shown in FIG. 3, and sends this to the FS 16a from the port 46a. To send.

  As shown in FIG. 11A, the FS 16a that has received the failure notification frame 78 transmits the failure notification frame 76 from all the ports 58a to 58h to the IFSs 14a to 14x. The failure notification frame 76 is generated by the failure notification frame generation unit 56 of the FS 16a.

  Note that the failure notification frame 76 from the FS 16a does not reach the IFS 14a in which a failure has occurred with the FS 16a. Therefore, the LAG setting unit 64 of the IFS 14a resets the LAG 18a excluding the port 58a that is connected to the FS 16a. FIG. 11B shows the LAG 18a reset.

  Then, as shown in FIG. 11B, when each IFS 14b-14x other than the IFS 14a receives the failure notification frame 76 from the FS 16a, the port 46a that has received this is excluded from the members of the respective LAGs 18a-18x. To do. That is, the LAG setting unit 64 of each IFS 14b to 14x regroups each LAG 18a to 18x in a state where the port 46a is excluded from the members.

[Method for realizing operation example 2]
FIG. 12 is a flowchart showing a method for realizing the failure notification shown in FIG. FIG. 13 is a flowchart showing a method for realizing the failure notification shown in FIG.

[Failure notification process (1)]
In the failure notification process (1) shown in FIG. 12, the failure determination unit 68 of the IFS 14 checks whether or not the detection frame 72 has been received at each of the ports 58a to 58h (step S300). When the reception of the detection frame 72 is confirmed (Yes), the failure determination unit 68 repeats this process.

  When the reception of the detection frame 72 cannot be confirmed at any or all of the ports 58a to 58h (step S300: No), the failure determination unit 68 of the IFS 14 checks whether a predetermined time has elapsed (step S302). ). If reception of the detection frame 72 is confirmed within a predetermined time (step S302: No), steps S300 to S302 are repeated.

  When the predetermined time has passed in a state where the reception of the detection frame 72 cannot be confirmed (step S300: No), the failure notification frame generation unit 70 of the IFS 14 generates a failure notification frame 78. The IFS 14 transmits a failure notification frame 78 from the port 58a to the FS 16a that has not been able to confirm the reception of the detection frame 72 (step S304).

  Then, the LAG setting unit 64 of the IFS 14a excludes the port (in this case, the port 58a) for which the reception of the detection frame 72 could not be confirmed within a predetermined time from the member, and regroups (updates) the LAG 18a (step) S306), the process is terminated.

[Failure notification process (2)]
Next, in the failure notification process (2) shown in FIG. 13, the failure determination unit 54 of the FS 16 checks whether or not the failure notification frame 78 has been received at each of the ports 46a to 46x (step S400).

  When the failure notification frame 78 is received (step S400: Yes), the failure notification frame generation unit 56 of the FS 16 generates a failure notification frame 76. And this is transmitted from all the ports 46a-46x connected to IFS14a-14x (step S402), and this process is complete | finished (END).

  Note that the failure notification frame 76 from the FS 16a does not reach the IFS 14a that has failed with the FS 16a, but the IFS 14a excludes the port 58a and resets the LAG 18a as described above. Communication with the other FSs 16b to 16h is possible.

  Further, when the failure notification frame 76 transmitted from the FS 16a is received by the other IFSs 14b to 14x, each IFS 14b to 14x executes the LAG resetting process shown in FIG. 8 and excludes the port 46a from the respective LAGs 18b to 18x. To reconfigure (regroup).

[Summary of Operation Example 2]
As described above, in the operation example 2, when a failure is detected in each of the IFSs 14a to 14x, the failure notification frame 78 is transmitted to the FSs 16a to 16h that are the transmission sources of the detection frames 72 that could not be confirmed. The failure notification frame 76 can be transmitted from the FSs 16a to 16h.
Similarly to the operation example 1, the network frame can be continuously relayed even after the LAGs 18a to 18x are degenerated.

[Operation example 3 when a failure is detected]
14 and 15 are diagrams for explaining an operation example 3 when a failure is detected by the network relay system 10. Here, similarly to the operation example 2, a case where a failure occurs in the communication state from the FS 16a to the IFS 14a will be described.

  As shown in FIG. 14A, when a failure occurs in communication from the FS 16a to the IFS 14a, the detection frame 72 transmitted from the FS 16a does not reach the IFS 14a within a predetermined time. At this time, the IFS 14a determines in the failure determination unit 68 that a failure has occurred in the communication state with the FS 16a.

  In this case, as shown in FIG. 14B, the IFS 14a stops transmitting the detection frame 74 to the FS 16a. At this time, the IFS 14a excludes the port 46a connected to the FS 16a from the members of the LAG 18a and performs reconfiguration (regrouping).

  Further, as shown in FIG. 14B, since the transmission of the detection frame 74 is stopped, the FS 16a can receive the detection frame 74 within a predetermined time at the port 46a connected to the IFS 14a. Disappear. Therefore, the failure determination unit 54 of the FS 16a determines that a failure has occurred in the communication state with the IFS 14a.

  Then, as shown in FIG. 15A, the FS 16a generates a failure notification frame 76 in the failure notification frame generation unit 56 in the same manner as in the operation example 1 and operation example 2 so far. Transmit from all ports 46a-46x connected to .about.14x.

  Similarly, for the IFS 14a in which a failure has occurred with the FS 16a, the failure notification frame 76 from the FS 16a does not arrive. However, as described above, the IFS 14a excludes the port 58a and resets the LAG 18a. Therefore, communication with the other FSs 16b to 16h is possible.

  Further, as shown in FIG. 15B, when the other IFSs 14b to 14x receive the failure notification frame transmitted from the FS 16a, the respective LAGs 18b to 18x are excluded from the port 46a that has received the failure notification frame. To reset.

  As described above, when the LAG 18a is reset in the IFS 14a that has detected the failure, and when the LAGs 18b to 18x are reset in the other IFSs 14b to 14x, each IFS 14a to 14x is set to the port 46a before the failure occurs. The port to be newly transferred is determined for the network frame transferred from the network. Accordingly, the port to which the network frame is to be newly transferred is determined based on a predetermined algorithm from among the other ports 46b to 46h belonging to the LAGs 18a to 18x in each of the IFSs 14a to 14x.

[Method for realizing operation example 3]
FIG. 16 is a flowchart showing a method for realizing the failure notification shown in FIG.

[Failure notification processing]
The failure determination unit 68 of the IFS 14 checks whether or not the detection frame 72 has been received at each of the ports 58a to 58h (step S500). When the reception of the detection frame 72 is confirmed (Yes), the failure determination unit 68 repeats this process.

  When the reception of the detection frame 72 cannot be confirmed at any of the ports 58a to 58h (step S500: No), the failure determination unit 68 of the IFS 14 checks whether a predetermined time has passed (step S502). If the reception of the detection frame 72 is confirmed within a predetermined time (step S502: No), steps S500 to S502 are repeated.

  When the predetermined time has passed in the state where the reception of the detection frame 72 cannot be confirmed (step S500: No) (step S502: Yes), the LAG setting unit 64 of the IFS 14 has confirmed the reception of the detection frame 72 ( Here, the port 46) is excluded from the members, and the LAGs 18a to 18x are reset (step S504).

  Then, the detection frame generation unit 66 of the IFS 14 stops transmission of the detection frame 74 from the port for which reception of the detection frame 72 has not been confirmed (step S506). Specifically, the generation of the detection frame 72 transmitted from the port 58a excluded from the LAG 18a is stopped, the transmission of the detection frame from the port 58a is stopped, and the process ends (END).

  Thereafter, the FS 16a that has stopped transmitting the detection frame 72 transmits the failure notification frame 76 from all the ports 46a to 46x based on the failure notification processing shown in FIG. When the IFS 14b-14x other than the IFS 14a receives the failure notification frame 76 transmitted from the FS 16a, the IFS 14b-14x excludes the port that received the failure notification frame 76 from the members based on the LAG resetting process shown in FIG. Perform reconfiguration (regrouping).

[Summary of Operation Example 3]
As described above, in the operation example 3, when a failure is detected in each of the IFSs 14a to 14x, the transmission of the detection frame 74 to the FSs 16a to 16h, which is the transmission source of the detection frame 72 that could not be confirmed, is stopped. The previous FSs 16a to 16h can detect a failure, thereby transmitting the failure notification frame 76 from the FSs 16a to 16h.
Similarly to the operation example 1, the network frame can be continuously relayed even after the LAGs 18a to 18x are degenerated.

  As described above, according to the network relay system 10 of the present embodiment, the communication state between the IFSs 14a to 14x and the FSs 16a to 16h is constantly monitored, and when a failure occurs on the communication path, Detect faults. When a failure is detected, the network frame passing through the communication path in which the failure has occurred can be switched and transferred to another communication path, which can contribute to improving the reliability of communication.

  As a modified example, the FS 16 monitors the link state of the port 46 connected to the IFS 14, and transmits a failure notification frame to the IFS 14 when the link state of the port 46 is linked down. Also good.

DESCRIPTION OF SYMBOLS 10 Network relay system 12 Terminal device 14 (14a-14x) IFS (interface repeater)
16 (16a-16h) FS (Fabric Repeater)
46 (46a to 46x) Port 48 Frame transfer processing unit 52, 66 Detection frame generation unit 54, 68 Failure determination unit 56, 70 Failure notification frame generation unit 64 LAG setting unit 58 (58a to 58h) Port

Claims (4)

Multiple interface repeaters;
A plurality of fabric repeaters connected to the plurality of interface repeaters;
With
The interface repeater is
LAG setting means for setting a link aggregation group in a port connected to the fabric relay, and first detection frame transmission means for transmitting a first detection frame to the fabric relay ,
The fabric relay is
A failure notification frame transmitting means for transmitting a first failure notification frame to the interface repeater; and a second detection frame transmitting means for transmitting a second detection frame to the interface repeater ;
The LAG setting means of the interface repeater, when receiving the first failure notification frame, excludes the port that has received the first failure notification frame from the link aggregation group ,
The interface repeater is
When the reception of the second detection frame cannot be confirmed, a first failure determination unit that determines that a failure has occurred, and a case where the first failure determination unit determines that a failure has occurred, Failure notification means for notifying that a failure has occurred with respect to the fabric relay connected to the port for which the reception of the detection frame could not be confirmed,
The fabric relay is
A second failure determination unit that determines that a failure has occurred when the reception of the first detection frame cannot be confirmed, or when a failure has been notified from the interface repeater;
The failure notification frame transmission means of the fabric repeater transmits the first failure notification frame when the second failure determination means determines that a failure has occurred,
The failure notification means of the interface repeater stops the transmission of the first detection frame from the port where the reception of the second detection frame could not be confirmed, thereby causing a failure to the fabric repeater. Notify you of the occurrence,
Network relay system. The network relay system according to claim 1 ,
The LAG setting unit of the interface repeater excludes, from the link aggregation group, a port that could not confirm the reception of the second detection frame when the first failure determination unit determines that a failure has occurred. ,
Network relay system. A network relay system control method comprising a plurality of interface repeaters and a plurality of fabric repeaters connected to the plurality of interface repeaters,
The interface repeater sets a link aggregation group on the port connected to the fabric repeater,
The fabric repeater transmits a first failure notification frame to the interface repeater,
When the interface repeater receives the first failure notification frame, the interface repeater excludes the port that has received the first failure notification frame from the link aggregation group ,
The interface repeater transmits a first detection frame to the fabric repeater;
The fabric relay transmits a second detection frame to the interface relay;
The interface repeater determines that a failure has occurred when the reception of the second detection frame cannot be confirmed, and the fabric connected to the port that has failed to confirm the reception of the second detection frame Notify the repeater that a failure has occurred,
The fabric repeater determines that a failure has occurred when the reception of the first detection frame cannot be confirmed, or when a failure has been notified from the interface repeater, Send a failure notification frame
The interface repeater notifies the fabric repeater that a failure has occurred by stopping transmission of the first detection frame from a port that has not been confirmed to receive the second detection frame. To
Network relay system control method. In the control method of the network relay system according to claim 3 ,
When determining that a failure has occurred, the interface repeater excludes a port from which the reception of the second detection frame could not be confirmed from the link aggregation group.
Network relay system control method.

Images