JP2012114645A - Network fabric switch system and fault detection method for the network fabric switch system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a network fabric switch system for reliably detecting the occurrence of a fault when the fault occurs in any one of fabric switches even in a case with a configuration where the multiple fabric switches are mutually connected by multiple physical lines.SOLUTION: Multiple fabric switches FSW6-FSW20 are respectively connected to respective interface fabric switches (IFSWs) 1, 2, 4. Link aggregation groups (LAGs) 34, 36, 38 are set in the respective IFSWs 1, 2, 4 with respect to ports to which the FSW6-FSW20 are connected. The FSW6-FSW20 transmit detection frames to the IFSWs, 1, 2, 4. The IFSWs 1, 2, 4 also transmit detection frames to the FSW6-FSW20. When the detection frame is not received in any one of the IFSWs 1, 2, 4 or the FSW6-FSW20, the occurrence of a fault is determined.

Description

  The present invention relates to a network relay system that relays network frames transmitted and received between terminal devices such as servers, personal computers, and workstations, and a failure detection method for such a network relay system.

  Conventionally, as a system that virtually operates as a single device using a plurality of network relay devices, a first network relay device group including a plurality of network relay devices functions as a fabric node and includes a plurality of network relay devices. A system in which a second network relay device group functions as a line node is known (see, for example, 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 having a function of reliably detecting a failure in any physical line, even if a plurality of repeaters are connected to each other via a plurality of physical lines. A fault detection method for a relay system and a network relay system is provided.

  In order to solve the above problems, a network relay system of the present invention is a network relay system that relays network frames transmitted and received between a plurality of terminal devices, and a plurality of terminal devices connected to each other. Provided in each interface repeater, a plurality of fabric repeaters interconnected with each interface repeater through a plurality of physical lines individually assigned to each of the interface repeaters, and a plurality of interface repeaters Setting means for setting a link aggregation group that logically bundles a plurality of physical lines when relaying a network frame to a plurality of fabric repeaters for a plurality of ports, and all fabric repeaters Individual physics for interface repeaters All the interfaces connected to each other by sending detection frames to all fabric repeaters for every physical line from all ports of the interface repeater within the link aggregation group When the detection frame transmission / reception means that enables bidirectional transmission / reception of the detection frame between the repeater and the fabric repeater and the detection frame cannot be received by at least one of the interface repeaters or the fabric repeater, Failure determination means for determining that a failure has occurred in relaying the network frame.

  In order to solve the above problems, the failure detection method of the network relay system according to the present invention includes a plurality of interface repeaters to which any of a plurality of terminal devices that transmit and receive network frames are connected, and a plurality of interface repeaters. It consists of a plurality of fabric repeaters interconnected with each interface repeater through a plurality of physical lines individually assigned to each interface repeater, and a plurality of ports provided in each interface repeater On the other hand, a network relay system that detects a failure occurring in a network relay system in which a link aggregation group in which a plurality of physical lines are logically bundled is set when a network frame is relayed between a plurality of fabric relays. Failure detection method, each While a detection frame is transmitted from an optical repeater to all interface repeaters through individual physical lines, detection is made to all fabric repeaters for each physical line from all ports of the interface repeater within the link aggregation group. A detection frame transmission / reception step that enables bidirectional transmission / reception of detection frames between all interface relays and fabric relays connected to each other by transmitting a frame, and at least one of the interface relays Alternatively, a failure determination step of determining that a failure has occurred in the relay of the network frame when the detection frame cannot be received by the fabric relay device.

  According to the network relay system and the failure detection method of the network relay system of the present invention, even when a plurality of repeaters are connected to each other through a plurality of physical lines, when any of the repeaters fails This can be reliably detected.

1 is a diagram schematically illustrating a configuration example of a network relay system. It is a figure which shows roughly the example of a structure of the network relay system in case IFSW, FSW, and a server are accommodated in the rack. It is a block diagram which shows roughly the functional structure of IFSW and FSW. It is a figure which shows roughly the structural example of the network relay system at the time of monitoring the communication performed between IFSW and FSW. It is a figure (1/2) which shows roughly composition of LAG at the time of communication failure in a 1st embodiment, and its operation. FIG. 3 is a diagram (2/2) schematically showing the configuration and operation of the LAG when a communication failure occurs in the first embodiment. 6 is a flowchart schematically showing processing executed by a failure determination unit of the FSW when a communication failure occurs between the FSW and IFSW shown in FIG. 5. 6 is a flowchart schematically showing processing executed by a LAG setting unit of IFSW when a communication failure occurs between the FSW and IFSW shown in FIG. 5. 6 is a table showing a LAG setting table of each IFSW before and after a communication failure occurs between the FSW and IFSW shown in FIG. 5. It is a figure (1/2) which shows roughly the composition and operation of LAG at the time of the occurrence of communication failure in a 2nd embodiment. It is a figure (2/2) which shows schematically the structure and operation | movement of LAG when the communication failure generate | occur | produces in 2nd Embodiment. 11 is a flowchart schematically showing processing executed by an IFSW failure determination unit and LAG setting unit when a communication failure occurs between the FSW and IFSW shown in FIG. 10. 11 is a flowchart schematically showing processing executed by a failure determination unit of the FSW when a communication failure occurs between the FSW and IFSW shown in FIG. It is a figure (1/2) which shows roughly the composition and operation of LAG at the time of the occurrence of communication failure in a 3rd embodiment. It is a figure (2/2) which shows schematically the structure and operation | movement of LAG when the communication failure generate | occur | produces in 3rd Embodiment. 15 is a flowchart schematically showing processing executed by an IFSW failure determination unit and a LAG setting unit when a communication failure occurs between the FSW and IFSW shown in FIG. 14.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. This embodiment is (1) a form as a network relay system including a plurality of IFSWs (Interface Fabric Switches) and a plurality of FSWs (Faric Switches), and (2) executed by at least any IFSW or FSW in the network relay system. It includes a form as a failure detection method for detecting that a failure has occurred in communication.

[First Embodiment]
FIG. 1 is a diagram schematically illustrating a configuration example of a network relay system. This network relay system mainly includes a plurality of IFSWs 1, 2, 4 (interface relays) and a plurality of FSWs 6, 8, 10, 12, 14, 16, 18, 20 (fabric relays), and these are mutually connected. It is composed of a plurality of physical lines to be connected (reference numerals are omitted). IFSW1, 2, 4 and FSW6, 8, 10, 12, 14, 16, 18, 20 shown as examples in FIG. 1 are data transfer functions such as layer 2 and layer 3 of the OSI (Open Systems Interconnection) reference model, for example. Is a switching hub. The basic structures and functions of the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20 may be common to each other. In the present embodiment, among the switching hubs, those connected to terminal devices such as the server 22 to the server 32 are referred to as “IFSW”, and those connected to the IFSW are referred to as “FSW”. Each of the servers 22, 24, 26, 28, 30, and 32 is a server used by the user in the backbone network, such as a file server or a mail server, or a network repeater for relaying data in the user's network. Terminal equipment. In addition to the IFSWs 1, 2, 4, some other IFSWs (not shown) are arranged in the network relay system, and these are individually FSWs 6, 8, 10, 12, 14, 16, 18 and 20 are connected by physical lines.

  Each of the IFSWs 1, 2, and 4 includes a plurality of ports (not shown), and some of these ports are ports arranged in the FSWs 6, 8, 10, 12, 14, 16, 18, and 20, respectively. It is connected to the. Further, in the IFSW 1, terminal devices such as a plurality of servers 22 and 24 and a plurality of servers (not shown) are connected to the other ports. Similarly to IFSW1, IFSW2 and 4 have servers 26 and 28, servers 30 and 32, and a plurality of other servers (not shown) connected to some ports. In the present embodiment, the server 22 to the server 32 are exemplified as terminal devices, but the terminal device may be a personal computer, a workstation, or the like.

  In the IFSW1, a LAG 34 using a link aggregation function is set in each port to which the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 are connected. Similarly to IFSW1, IFSW2 and 4 have LAGs 36 and 38 respectively set in the ports to which FSWs 6, 8, 10, 12, 14, 16, 18, and 20 are connected.

  When each IFSW 1, 2, 4 transmits a network frame received from each server 22, 24, 26, 28, 30, 32 to the FSW 6, 8, 10, 12, 14, 16, 18, 20, a predetermined algorithm From which port among the ports set in the respective LAGs 34, 36, 38 is determined to transmit the network frame. In this algorithm, for example, when an identification number (INDEX) is set in advance for each port constituting the LAG and a network frame transmitted from the servers 22 to 32 is received, destination information and transmission stored in the network frame are transmitted. By associating a value obtained by calculating a MAC address or an IP address indicating original information with the above identification number, a port to which received data is to be transferred is uniquely determined. Thereby, it is possible to distribute the load generated when each IFSW 1, 2, 4 transfers the network frame received from the servers 22-32 to the FSWs 6-20.

  In addition, since each IFSW 1, 2, and 4 can uniquely determine the port to which the network frame is to be transmitted by the above algorithm, the communication path on the transmission side of the network frame between the servers 22 to 32 and the reception side Communication paths can be matched. For example, when the server 22 and the server 32 transmit / receive network frames to / from each other, the frames transmitted from the server 22 are transferred to the FSW 20 via the IFSW 1 according to the above algorithm, and further transferred from the FSW 20 to the server 32 via the IFSW 4. Is done. Conversely, the frame transmitted from the server 32 is transferred to the FSW 20 via the IFSW 4 and further transferred from the FSW 20 to the server 22 via the IFSW 1. As described above, when a network frame is transmitted and received between a certain two servers 22 and 32, the ports used by IFSW1 and IFSW4 match in both directions.

  In addition, IFSW1, 2, 4 are transferred between them by reconfiguring the configuration of LAGs 34, 36, 38 set so far, for example, when a communication failure occurs between IFSW1 and FSW6. The network frame can be switched and transferred to another communication path instantly. The operations of IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, 20 before and after the occurrence of a communication failure will be described in more detail later with reference to another drawing.

  The IFSWs 1, 2, 4 are respectively connected to the FSWs 6, 8, 10, 12, 14, 16, 18, 20, and each IFSW 1, 2, 4 is configured by configuring each port connecting them as one LAG. In addition, network frames transferred between the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 can be efficiently distributed. Accordingly, the total amount of data to be transferred per unit time between these is the amount of data that can be transferred per unit time between IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, 20 And the desired data transfer efficiency can be maintained (so-called non-blocking can be realized).

  The IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20 are respectively independent box-type switching hubs. Therefore, it is not necessary to arrange these components as a single unit, and it is possible to realize a connection configuration that considers the arrangement of a plurality of servers more flexibly than, for example, a network relay system configured using a chassis type switching hub. . Note that the connection form considering these arrangements will be described in more detail later with reference to FIG.

  If a VLAN (Virtual Local Area Network) is set for each of the servers 22, 24, 26, 28, 30, 32, IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, Each of the network units 20 transmits / receives a network frame using a tag VLAN in principle. At this time, all the VLAN information assigned to the servers 22 to 32 is registered in the ports of the FSWs 6, 8, 10, 12, 14, 16, 18, and 20. The IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20 transmit the VLAN information in a tagged state according to each received network frame when transmitting the network frame. . As a result, even if a VLAN is set in each of the servers 22, 24, 26, 28, 30, and 32, the communication path on the transmission side of the network frame and the communication path on the reception side are matched between the servers. Can do. Therefore, an operator or the like who manages this network relay system can easily manage communications executed between the servers 22 to 32 individually.

  FIG. 2 shows the configuration of the network relay system when the IFSWs 1, 2, 4, FSWs 6, 8, 10, 12, 14, 16, 18, 20, and the servers 22, 24, 26, 28, 30, 32 are stored in a rack. It is a figure which shows an example schematically. In the network relay system shown in FIG. 1, for example, IFSW1,2,4, FSW6,8,10,12,14,16,18,20 and servers 22,24,26,28,30,32 are racked in the data center. A configuration example in the case of storing in a case will be described.

  In one rack 40, IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, 20 are accommodated. The other racks 42 and 44 house a plurality of servers 22, 24, 26, 28, 30 and 32, respectively. At this time, the FSWs 6, 8, 10, 12, 14, 16, 18, 20 and the IFSWs 1, 2, 4 are connected to the ports SW of the IFSWs 1, 2, 4 similarly to the configuration (physical topology configuration) shown in FIG. 8, 10, 12, 14, 16, 18, and 20 are connected. The servers 22, 24, 26, 28, 30, and 32 accommodated in the racks 42 and 44 are connected to IFSWs 1, 2, and 4, respectively.

  In this way, a plurality of IFSWs 1, 2, 4 and FSW6 to FSW20 can be housed and arranged in one rack.

  Further, the IFSWs 1, 2, 4 are respectively stored in racks 42, 44 and other racks (not shown), and the servers 22, 24, 26, 28, 30, connected to the IFSWs 1, 2, 4 are stored. 32 may be stored in each rack in which the above-described IFSWs 1, 2, and 4 are stored.

  In any case, in the present embodiment, the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20 may be arranged in one place or distributed. Therefore, it is possible to flexibly cope with future server expansion and rearrangement.

  FIG. 3 is a block diagram schematically showing the functional configuration of the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20. In this embodiment, IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, 20 have the same basic configuration and function as a switching hub. Here, for the sake of convenience, the functional configuration will be described using IFSW1.

  The IFSW 1 includes a port 46a to a port 46j, a frame transfer processing unit 48, a memory unit 50, a LAG setting unit 52, and a failure determination unit 54. The memory unit 50 includes an FDB (Forwarding Data Base) 50a and a LAG setting table 50b. The failure determination unit 54 includes a detection frame generation unit 54a and a failure notification frame generation unit 54b.

  Of the ports 46a to 46j, for example, the ports 46a to 46h are connected to the FSWs 6, 8, 10, 12, 14, 16, 18, and 20, respectively, and the ports 46i and 46j are the servers 22 and 24, respectively. It is assumed that it is connected to When the ports 46 a to 46 h receive the network frame transferred from each of the FSWs 6, 8, 10, 12, 14, 16, 18, 20 or the servers 22, 24, transfer them to the frame transfer processing unit 48. Also, the network frame transferred from the frame transfer processing unit 48 is transmitted.

  The frame transfer processing unit 48 relays the network frame transferred from the ports 46a to 46j to the transfer destination port. The LAG setting unit 52 sets the LAG 34 with the ports 46a to 46h as one group based on the link aggregation function, and stores this information in the LAG setting table 50b. In this embodiment, the LAG setting unit 52 arranged in the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 does not set LAG for each port.

  The frame transfer processing unit 48 specifies the transfer destination port based on the destination information and source information of the network frame transferred from the ports 46a to 46j, and the information stored in the FDB 50a and the LAG setting table 50b of the memory unit 50. Network frames. For example, when the server 22 sends data to the server 24, when the network frame transmitted from the server 22 is transferred to the frame transfer processing unit 48 via the port 46i, the frame transfer processing unit 48 is stored in the network frame. The FDB 50 a is referred to based on the destination information indicating the server 24 and the transmission source information indicating the server 22. When the information indicating the server 24 is registered in the FDB 50a in association with the port 46j to which the server 24 is to be transmitted, the network frame is transferred to the port 46j. In addition, when the server 22 transmits data to the servers 26 to 32 connected to the other IFSWs 2 and 4, the frame transmitted from the server 22 is transferred to the frame transfer unit 42 via the port 40 i. When the destination information of the network frame is registered in association with the ports 46a to 46h set in the LAG 34 with reference to the FDB 50a, the transfer processing unit 48 determines the network frame based on a predetermined algorithm. Transmission from another port (for example, port 46a).

  In each of the IFSWs 1, 2, and 4, it is preferable that the ports 46a to 46h that connect the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 are matched. That is, if the ports connecting the FSWs 6, 8, 10, 12, 14, 16, 18, 20 are matched in each IFSW 1, 2, 4, the port that transmits the network frame uniquely determined by the above algorithm Are common to the IFSWs 1, 2, and 4. As a result, the network frame transferred between the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 is an upstream communication that is transferred from one server to the other server. The same communication path is transferred for the path and the downstream communication path transferred from the other server to the one server. For this reason, the communication speed of the frames passing through the upstream and downstream communication paths between the servers 22 to 32 can be kept constant, and efficient communication can be executed without narrowing the bandwidth.

  The failure determination unit 54 monitors the communication state with the FSWs 6, 8, 10, 12, 14, 16, 18, 20 connected to the individual ports 46a to 46h constituting the LAG 34 in the detection frame generation unit 54a. A detection frame is generated and transmitted from the ports 46a to 46h at predetermined intervals. If the failure determination unit 54 cannot receive a detection frame at a port connected to any of the FSWs 6, 8, 10, 12, 14, 16, 18, 20, the failure notification frame generation unit 54b A frame is generated and transmitted from the port that cannot receive the detection frame to the FSW connected to the port. Further, in the case of the third embodiment described below, the failure determination unit 54 is connected to a port that can no longer receive a detection frame when it cannot receive a detection frame transmitted from an FSW connected to any of the ports. The transmission of the detection frame toward the FSW is stopped.

Further, the failure determination unit 54 provided in each of the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 generates a detection frame in the detection frame generation unit 54a, and all the units connected to the IFSWs 1, 2, and 4 are connected. This is transmitted from the port at predetermined intervals. Further, when the failure determination unit 54 provided in each of the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 cannot receive the detection frame transmitted from the IFSW connected to any of the ports, the failure determination unit 54 The notification frame generation unit 54b generates a failure notification frame and transmits it from all ports connected to the IFSWs 1, 2, and 4.
Note that the process executed by the failure determination unit 54 will be described in more detail later with reference to FIGS.

  FIG. 4 is a diagram schematically showing a configuration example of a network relay system when monitoring communication executed between IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, and 20. is there. In the network relay system shown in FIG. 1, each of the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 transmits and receives detection frames, so that the ports connected to each other (physical Monitor the communication status (including the line). Note that arrows indicated by solid lines and dotted lines in FIG. 4 indicate the transmission direction of the detection frame.

  The IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, 20 generate detection frames in the detection frame generation unit 54a. In this detection frame, identification information indicating the transmission source IFSW 1, 2, 4 and FSW 6, 8, 10, 12, 14, 16, 18, 20 and identification information indicating each port that transmits the detection frame are stored. ing. In the FSWs 6, 8, 10, 12, 14, 16, 18, and 20, the generated detection frames are sent from each port to the connection destination IFSWs 1, 2, and 4 at predetermined intervals (for example, every several tens of seconds to several minutes). Send to. In the IFSWs 1, 2, and 4, the generated detection frames are sent from the ports 46a to 46h belonging to the LAGs 34, 36, and 38 to the FSWs 6, 8, and 10 at predetermined intervals (also every several tens of seconds to several minutes). , 12, 14, 16, 18, and 20 are transmitted.

  In this way, IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18, and 20 identify each other's communication status by transmitting and receiving detection frames to each other. Can do. On the other hand, when the detection frame cannot be received, it is possible to detect that a failure has occurred in communication performed with the transmission source of the detection frame. Note that processing in IFSW and FSW after the occurrence of a failure will be described in more detail later with reference to FIGS.

  5 and 6 are diagrams schematically showing the configuration and operation of the LAGs 34, 36, 38 when a communication failure occurs in the first embodiment. Normally, the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20 monitor the communication state of each other by transmitting and receiving detection frames, but in the following, as an example, A case where a failure occurs in communication between IFSW1 and FSW6 will be described.

  In FIG. 5, (A): When a failure occurs in communication from IFSW1 to FSW6, the detection frame transmitted from IFSW1 does not reach FSW6 within a predetermined time. At this time, the FSW 6 determines that a failure has occurred in the communication executed between the above in the failure determination unit 54 (failure determination step).

  In FIG. 5, (B): The FSW 6 generates a failure notification frame storing information indicating that a communication failure has occurred in the failure notification frame generation unit 54b, identification information indicating the FSW 6, and identification information indicating each port. This is transmitted from all ports connected to IFSW1, 2, 4 (failure notification step).

  FIG. 6: Upon receiving the failure notification frame transmitted from the FSW 6, each IFSW 1, 2, 4 excludes the port 46 a (specific port) that has received the frame from each LAG 34, 36, 38. The LAG setting unit 52 of each IFSW 1, 2, 4 resets the link aggregation group with the port 46a excluded (regrouping step). Therefore, the network frame transmitted from the port 46a before the failure occurs is newly determined from the other ports 46b to 46h constituting the LAGs 34, 36, and 38 based on a predetermined algorithm (for example, Sent from port 46b).

  FIG. 7 is a flowchart schematically showing processing executed by the failure determination unit 54 of the FSW 6 when a communication failure occurs between the FSW 6 and the IFSW 1 shown in FIG. FIG. 8 is a flowchart schematically showing processing executed by the LAG setting unit 52 of the IFSWs 1, 2, and 4 when a communication failure occurs between the FSW 6 and the IFSW 1 shown in FIG. Hereinafter, each processing will be described.

  Step S100: The failure determination unit 54 of the FSW 6 determines whether or not a detection frame has been received at each port. If a detection frame is received at each port (Yes), this process is repeated. When the detection frame is not received at any or all of the ports (No), the process proceeds to the next step S102.

  Steps S102 and S104: Next, when the failure determination unit 54 of the FSW 6 has not received the detection frame even after a predetermined time has elapsed (Yes), it generates a failure notification frame and connects all the IFSWs 1, 2, and 4 A failure notification frame is transmitted (flooded) from the port, and this processing ends (END). When the detection frame is received within the predetermined time (step S102: No), steps S100 to S102 are repeated.

[See Figure 8]
Steps S200, S202: Each IFSW 1, 2, 4 confirms whether or not a failure notification frame has been received at the port connected to the FSW 6, and if a failure notification frame is received (Yes), then the LAG setting unit 52 The port that received this failure notification frame is excluded from the LAGs 34, 36, and 38, and this process is terminated. Although the failure notification frame is constantly monitored in step S200 here, the process of step S202 is executed in sequence using the reception of the failure notification frame as a trigger without placing step S200 in particular. It is good.

  Next, FIG. 9 is a table showing the LAG setting table 50b of each IFSW 1, 2, and 4 before and after a communication failure occurs between the FSW 6 and the IFSW 1 shown in FIG.

[Before failure]
In FIG. 9, (A): Before the failure occurs, the LAG setting table 50b of the IFSWs 1, 2, and 4 shows the ports 46a to 46h belonging to the LAG in the LAG belonging port column on the left side. The LAG ID column on the right side shows the LAG number. At this time, the ports 46a to 46h shown in the LAG belonging port column belong to the common LAG 34 (LAG 36, 38) shown in the LAG ID column. In the present embodiment, one LAG is set for each IFSW 1, 2, 4, but a plurality of LAGs may be set.

[After failure]
In FIG. 9, (B): When the LAG setting unit 52 of the IFSW 1, 2, 4 receives the failure notification frame transmitted from the FSW 6 at the port 46a, the port 46a is excluded from the LAG 34 (LAG 36, 38) and re-executed. Perform grouping. At this time, the LAG affiliation port column shows ports 46b to 46h excluding the port 46a.

  As described above, the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 always monitor each other's communication state for each port, and if a communication failure occurs, the communication passes through this interval. The network frame to be transmitted is transmitted from the ports belonging to the remaining LAGs. That is, when a communication failure occurs, the communication path of the network frame passing through the section where the failure has occurred is instantly switched to another communication path, and the time for discarding the network frame due to the communication failure can be greatly shortened. Thereby, the reliability of the communication performed between each server 22-32 can be improved.

  In addition, although the configuration of each LAG 34, 36, 38 is temporarily degenerated when a failure occurs, by applying the above algorithm in the reconfigured LAG 34, 36, 38, transmission / reception continues between the servers 22-32. Network frame relay can continue. After that, when physical line or port restoration work is performed, LAG 34, 36, 38 settings are restored to the initial state, thereby efficiently restarting network frame relay using all bandwidths. can do.

[Second Embodiment]
10 and 11 are diagrams schematically illustrating the configuration and operation of the LAGs 34, 36, and 38 when a communication failure occurs in the second embodiment. Similarly to the first embodiment, in the second embodiment, the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20 each communicate with each other by transmitting and receiving detection frames. Is monitoring. In the second embodiment, it is assumed that a failure occurs in communication from the FSW 6 to the IFSW 1.

  In FIG. 10, (A): When a failure occurs in communication from the FSW 6 to the IFSW 1, the detection frame transmitted from the FSW 6 does not reach the IFSW 1 within a predetermined time. At this time, the IFSW 1 determines that a failure has occurred in the communication executed between the failure determination unit 54 (failure determination step).

  In FIG. 10, (B): IFSW1 generates a failure notification frame in the failure notification frame generation unit 54b and transmits it from the port 46a connected to the FSW 6 (initial failure notification step).

[Refer to FIG. 11]
11A, when the FSW 6 receives the failure notification frame transmitted from the IFSW 1, the FSW 6 generates a failure notification frame in the failure notification frame generation unit 54b of the FSW 6, and all of the failure notification frames are connected to the IFSWs 1, 2, and 4. From the port (total failure notification step).

  In FIG. 11, (B): When receiving the failure notification frame transmitted from the FSW 6, the IFSWs 2 and 4 exclude the port 46a that has received the frame from the LAGs 36 and 38, respectively. In addition, the IFSW 1 excludes the port 46a that has not received the detection frame within a predetermined time from the LAG 34. The LAG setting unit 52 of each IFSW 1, 2, 4 resets the link aggregation group with the port 46a excluded (regrouping step). Therefore, the network frame transmitted / received at this port 46a before the failure occurs is sent through a port newly determined based on a predetermined algorithm from among the other ports 46b to 46h constituting the LAGs 34, 36, 38. Sent and received.

  FIG. 12 is a flowchart schematically showing processing executed by the failure determination unit 54 and the LAG setting unit 52 of the IFSW 1 when a communication failure occurs between the FSW 6 and the IFSW 1 shown in FIG. FIG. 13 is a flowchart schematically showing processing executed by the failure determination unit 54 of the FSW 6 when a communication failure occurs between the FSW 6 and the IFSW 1.

  Step S300: The failure determination unit 54 of IFSW1 determines whether or not a detection frame has been received at each port. If a detection frame is received at each port (Yes), this process is repeated. If a detection frame has not been received at any or all of the ports (No), the process proceeds to the next step S302.

  Steps S302 and S304: Next, the failure determination unit 54 of IFSW1 repeats steps S300 to S302 when a detection frame is received within a predetermined time (step S302: No). On the other hand, the failure determination unit 54 of IFSW1 generates a failure notification frame and transmits a failure notification frame from the port 46a connected to the FSW 6 when a detection frame has not been received even after a predetermined time has elapsed (step S302: Yes). To do.

  Step S304: The LAG setting unit 52 of IFSW1 excludes the port 46a that has not received the detection frame within a predetermined time from the LAG 34, and ends (END) this process.

[See Fig. 13]
Steps S400 and S402: The FSW 6 confirms whether or not a failure notification frame has been received at the port 46a connected to the IFSW 1, and if a failure notification frame is received (Yes), then the failure determination unit 54 of the FSW 6 transmits the failure notification frame. Generate. Then, this is transmitted from the ports connected to the IFSWs 1, 2, and 4, respectively, and this processing is ended (END).

  When the failure notification frame transmitted by the FSW 6 is received by the IFSWs 2 and 4, each IFSW 2 and 4 executes the LAG resetting process shown in FIG. 8 to exclude the port 46a from the respective LAGs 36 and 38. Reconfigure (regroup) in that state.

  As described above, in the second embodiment, as in the first embodiment, when a communication failure occurs, each of the IFSWs 1, 2, and 4 instantaneously sets the communication path of the network frame transferred via the IFSW1 and the FSW6. It is possible to switch to another communication path. In addition, although the configuration of each LAG is temporarily degenerated when a failure occurs, by applying the above algorithm within the reconfigured LAG 34, 36, 38, the network frames that are continuously transmitted and received between the servers 22-32 Relay can continue. Then, when physical line and port restoration work is performed thereafter, the network frames are efficiently relayed using all the bands by restoring the settings of LAG 34, 36, and 38 to the initial state. can do.

[Third Embodiment]
FIG. 14 is a diagram schematically illustrating the configuration and operation of the LAGs 34, 36, and 38 when a communication failure occurs in the third embodiment. In the third embodiment as well as in the first embodiment and the second embodiment, the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, 20 are respectively detected frames shown in FIG. By transmitting and receiving, each other's communication status is monitored. In the third embodiment, a case where a failure occurs in communication from the FSW 6 to the IFSW 1 will be described as a modification of the second embodiment.

  In FIG. 14, (A): When a failure occurs in communication from the FSW 6 to the IFSW 1, the detection frame transmitted from the FSW 6 does not reach the IFSW 1 within a predetermined time. At this time, the IFSW 1 primarily determines that a failure has occurred in communication executed between the failure determination unit 54 (primary failure determination procedure).

  In FIG. 14, (B): IFSW1 determines that a communication failure has occurred, stops transmission of the detection frame from port 46a to FSW6 (detection frame transmission stop procedure). The IFSW 1 excludes the port 46a connected to the FSW 6 from the LAG 34. For this reason, the FSW 6 cannot receive the detection frame within a predetermined time at the port 46a connected to the IFSW 1 and secondarily determines that a communication failure has occurred between the IFSW 1 and the FSW 6 (secondary failure). Judgment procedure).

[See Fig. 15]
In FIG. 15, (A): The FSW 6 generates a failure notification frame in the failure notification frame generation unit 54b as in the first embodiment and the second embodiment, and all of them are connected to the IFSWs 1, 2, and 4. Is transmitted from the other port (failure notification step).

  In FIG. 15, (B): When receiving the failure notification frame transmitted from the FSW 6, the IFSWs 2 and 4 exclude the port 46a that has received the frame from the LAGs 36 and 38, respectively. Each IFSW 1, 2, 4 determines a port to be newly transferred with respect to the network frame transferred from the port 46a before the failure occurs. The LAG setting unit 52 of each IFSW 2 and 4 resets the link aggregation group in a state where the port 46a is excluded (regrouping step). 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 34, 36, and 38 in each of the IFSWs 1, 2, and 4.

  FIG. 16 is a flowchart schematically showing processing executed by the failure determination unit 54 and the LAG setting unit 52 of the IFSW 1 when a communication failure occurs between the FSW 6 and the IFSW 1 shown in FIG. The procedure will be described below.

  Step S500: The failure determination unit 54 of IFSW1 determines whether or not a detection frame has been received at each of the ports 46a to 46h belonging to the LAG 34. When the detection frame is received at each of the ports 46a to 46h (Yes), this process is repeated. If no detection frame is received at any or all of the ports (No), the process proceeds to the next step S502.

  Step S502: Next, the failure determination unit 54 of IFSW1 confirms whether or not a predetermined time has elapsed since the previous detection frame was received at each port. In each port, when the predetermined time has not elapsed since the previous detection frame was received (No), steps S500 to S502 are repeated. On the other hand, in any or all ports, when a predetermined time has elapsed since the previous detection frame was received (Yes), the failure determination unit 54 determines that a communication failure has occurred. At this time, the LAG setting unit 52 excludes the port 46a that has not received the detection frame within the predetermined time from the LAG 34 (step S504).

  Step S506: Then, the failure determination unit 54 of IFSW1 stops generating the detection frame to be transmitted from the port 46a excluded from the LAG 34, stops transmission of the detection frame from the port 46a, and performs this process. End (END).

  At this time, the FSW 6 transmits a failure notification frame from all ports connected to the IFSWs 1, 2, and 4 based on the failure notification process shown in FIG. Further, when receiving the failure notification frame transmitted from the FSW 6, the IFSWs 2 and 4 exclude the received ports from the LAGs 36 and 38 based on the LAG resetting process shown in FIG. (Regroup).

  In this way, the IFSWs 1, 2, 4 and the FSWs 6, 8, 10, 12, 14, 16, 18, and 20 send and receive detection frames to each other, so that the normality of communication executed between the ports is always maintained. Can be confirmed. Also, by sending a failure notification frame from the FSW 6 that has detected the failure to all the IFSWs 1, 2, and 4 connected thereto, the IFSWs 1, 2, and 4 reconfigure the LAGs 34, 36, and 38, respectively. The network frame is prevented from being transferred to the section where the communication failure has occurred. Thereby, the servers 22, 24, 26, 28, 30, and 32 connected to the IFSWs 1, 2, and 4 can perform stable communication with each other. Similarly, although the configurations of the LAGs 34, 36, and 38 are temporarily degenerated when a failure occurs, the servers 22 to 32 are continuously applied by applying the above algorithm in the reconfigured LAGs 34, 36, and 38. It is possible to continue relaying network frames transmitted and received between them. After that, when physical line and port restoration work is performed, the LAG setting can be restored to the initial state, so that the relay of network frames can be efficiently resumed using all bands. .

  As described above, according to the network relay system of the present embodiment, each of the IFSWs 1, 2, 4 and FSWs 6, 8, 10, 12, 14, 16, 18 for communication executed between the plurality of servers 22 to 32. , 20 can transfer the network frame without reducing the transfer efficiency. Further, the communication state between 1, 2, 4 and FSW 6, 8, 10, 12, 14, 16, 18, 20 is constantly monitored, and if a failure occurs on the communication path, the network frame passing through this Can be instantaneously switched to another communication path and transferred, which contributes to improvement in communication reliability.

  Further, after the physical line and port restoration work, the LAG setting unit 52 automatically restores the group configuration of the LAGs 34, 36, and 38 to the initial state so that the logical configuration of the network relay system can be quickly established. Can be restored.

1,2,4 IFSW (interface repeater)
6, 8, 10, 12, 14, 16, 18, 20 FSW (fabric repeater)
22, 24, 26, 28, 30, 32 Server (terminal equipment)
34, 36, 38 LAG
46a-46j ports

Claims (12)

A network relay system that relays network frames transmitted and received between a plurality of terminal devices,
A plurality of interface repeaters to which any of the terminal devices is connected;
A plurality of fabric repeaters interconnected with each interface repeater through a plurality of physical lines individually assigned to each of the plurality of interface repeaters;
A link aggregation group in which a plurality of physical lines are logically bundled is set for a plurality of ports provided in each interface repeater when the network frame is relayed between the plurality of fabric repeaters. Setting means;
While transmitting detection frames from each fabric relay to all the interface relays through the individual physical lines, all the ports from all the ports of the interface relays in the link aggregation group are all the physical lines. A detection frame that enables transmission and reception of detection frames in both directions between all the interface relays and the fabric relays that are connected to each other by transmitting a detection frame to the fabric relay Transmitting and receiving means;
A network relay system comprising failure determination means for determining that a failure has occurred in the relay of the network frame when the detection frame cannot be received by at least one of the interface relay or the fabric relay. The network relay system according to claim 1,
When the failure determination unit determines that a failure has occurred when the specific fabric relay device out of a plurality cannot receive the detection frame from the specific interface relay device, the detection device receives the detection frame. A failure notification means for transmitting a failure notification frame for notifying that a failure has occurred to each interface relay from the specific fabric relay that could not be performed;
When the failure notification frame is received at a specific port among a plurality of the interface repeaters, the specific port is excluded from the link aggregation group set by the setting unit at each interface repeater A network relay system further comprising regrouping means for performing regrouping in the network. The network relay system according to claim 1,
When it is determined that a failure has occurred by the failure determination means triggered by the fact that the specific interface repeater cannot receive the detection frame from the specific fabric repeater, the specification that failed to receive the detection frame An initial failure notification means for transmitting a failure notification frame for notifying at an initial stage that a failure has occurred from the interface relay to the specific fabric relay;
When the failure notification frame is received by the specific fabric repeater, the failure notification frame for notifying all of the interface relays that a failure has occurred is transmitted from the specific fabric repeater. A total failure notification means;
When the failure notification frame is received at a specific port among a plurality of the interface repeaters, the specific port is excluded from the link aggregation group set by the setting unit at each interface repeater A network relay system further comprising regrouping means for performing regrouping in the network. The network relay system according to claim 1,
When the failure determination unit first determines that a failure has occurred when the specific interface relay device among a plurality of the interface relay devices cannot receive the detection frame from the specific fabric relay device, the detection Detection frame transmission stopping means for stopping transmission of the detection frame from the specific interface relay that has not received a frame to the specific fabric relay;
When transmission of the detection frame is stopped by the detection frame transmission stop unit, the detection frame cannot be received by the specific fabric relay from the specific interface relay whose transmission has been stopped. When the failure determination means secondarily determines that a failure has occurred, the specific fabric relay that has failed to receive the detection frame notifies each interface relay that a failure has occurred. Failure notification means for transmitting a failure notification frame for
When the failure notification frame is received at a specific port among a plurality of the interface repeaters, the specific port is excluded from the link aggregation group set by the setting unit at each interface repeater A network relay system further comprising regrouping means for performing regrouping in the network. In the network relay system according to claim 3 or 4,
The specific interface relay that has not received the detection frame is in a state in which the specific port connected to the specific fabric relay through the physical line is excluded from the link aggregation group set by the setting unit. A network relay system characterized by regrouping. In the network relay system according to any one of claims 1 to 5,
The detection frame transmission / reception means includes:
The detection frame is transmitted at predetermined time intervals from each of the fabric repeater and the interface repeater,
The failure determination means includes
The network relay system according to claim 1, wherein a failure occurs when the detection frame is not received within the predetermined time by either the fabric relay or the interface relay. A plurality of interface repeaters to which any of a plurality of terminal devices that transmit / receive network frames to / from each other are connected, and each interface repeater through a plurality of physical lines individually assigned to each of the plurality of interface repeaters And a plurality of fabric repeaters connected to each other, and a plurality of ports provided in each interface repeater are connected to the plurality of fabric repeaters when relaying the network frame. A failure detection method for a network relay system that detects a failure that has occurred in a network relay system in which a link aggregation group in which the physical lines are logically bundled is set,
While transmitting detection frames from each fabric relay to all the interface relays through the individual physical lines, all the ports from all the ports of the interface relays in the link aggregation group are all the physical lines. A detection frame that enables transmission and reception of detection frames in both directions between all the interface relays and the fabric relays that are connected to each other by transmitting a detection frame to the fabric relay A transmission and reception process;
Failure detection of a network relay system comprising a failure determination step of determining that a failure has occurred in the relay of the network frame when the detection frame cannot be received by at least one of the interface relay or the fabric relay Method. The failure detection method for a network relay system according to claim 7,
When it is determined in the failure determination step that a failure has occurred due to a failure of the specific fabric relay device among the plurality of specific fabric relay devices to be received from the specific interface relay device, the detection frame is received. A failure notification step of transmitting a failure notification frame for notifying that a failure has occurred to each interface relay from the specific fabric relay that could not be performed;
When the failure notification frame is received at a specific port among a plurality of the interface repeaters, the group is regrouped in a state in which the specific port is excluded from the link aggregation group set at each interface repeater. A failure detection method for a network relay system, further comprising a regrouping step for performing networking. The failure detection method for a network relay system according to claim 7,
When it is determined in the failure determination step that a specific interface relay unit has failed to receive the detection frame from the specific fabric relay unit, a failure has not occurred. An initial failure notification step of transmitting a failure notification frame for notifying the occurrence of a failure from the interface relay to a specific fabric relay in an initial stage;
When the failure notification frame is received by the specific fabric repeater, the failure notification frame for notifying all of the interface relays that a failure has occurred is transmitted from the specific fabric repeater. A total failure notification process;
When the failure notification frame is received at a specific port among a plurality of the interface repeaters, the group is regrouped in a state in which the specific port is excluded from the link aggregation group set at each interface repeater. A failure detection method for a network relay system, further comprising a regrouping step for performing networking. The failure detection method for a network relay system according to claim 7,
The failure determination step includes
A primary failure determination procedure for temporarily determining that a failure has occurred when a specific interface relay device among a plurality of the interface relay devices cannot receive the detection frame from the specific fabric relay device;
Detection that stops transmission of the detection frame from the specific interface repeater that has failed to receive the detection frame to the specific fabric repeater when it is primarily determined that a failure has occurred in the primary failure determination procedure Frame transmission stop procedure,
When transmission of the detection frame is stopped by the detection frame transmission stop procedure, the detection frame cannot be received by the specific fabric relay from the specific interface relay whose transmission is stopped Including a secondary failure determination procedure for secondary determination that a failure has occurred,
When it is determined that a failure has occurred in the secondary failure determination procedure, the specific fabric relay that has failed to receive the detection frame notifies the interface relay that a failure has occurred. A failure notification step of transmitting a failure notification frame;
When the failure notification frame is received at a specific port among a plurality of the interface repeaters, the group is regrouped in a state in which the specific port is excluded from the link aggregation group set at each interface repeater. A failure detection method for a network relay system, further comprising a regrouping step for performing networking. In the failure detection method of the network relay system according to claim 9 or 10,
When the specific interface repeater fails to receive the detection frame, the specific interface relay connects the specific port connected to the specific fabric repeater through the physical line in the link aggregation group. A failure detection method for a network relay system, further comprising a specific regrouping step of performing regrouping in a state excluded from the network. The failure detection method for a network relay system according to any one of claims 7 to 11,
In the detection frame transmission / reception step,
The detection frame is transmitted at predetermined time intervals from each of the fabric repeater and the interface repeater,
In the failure determination step,
A failure detection method for a network relay system, wherein a failure occurs when either the fabric repeater or the interface repeater fails to receive the detection frame within the predetermined time.

Images