WO2012019464A1 - Inter-plate interconnection device and method in router cluster - Google Patents

Inter-plate interconnection device and method in router cluster Download PDF

Info

Publication number
WO2012019464A1
WO2012019464A1 PCT/CN2011/072869 CN2011072869W WO2012019464A1 WO 2012019464 A1 WO2012019464 A1 WO 2012019464A1 CN 2011072869 W CN2011072869 W CN 2011072869W WO 2012019464 A1 WO2012019464 A1 WO 2012019464A1
Authority
WO
WIPO (PCT)
Prior art keywords
central
ethernet
switch board
ethernet switch
frame
Prior art date
Application number
PCT/CN2011/072869
Other languages
French (fr)
Chinese (zh)
Inventor
叶傲
陈健
周海山
杨骐
杜相文
马苗苗
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2012019464A1 publication Critical patent/WO2012019464A1/en

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/40Constructional details, e.g. power supply, mechanical construction or backplane
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/35Switches specially adapted for specific applications
    • H04L49/351Switches specially adapted for specific applications for local area network [LAN], e.g. Ethernet switches

Definitions

  • the present invention relates to a technique for controlling surface interconnection, and in particular, to an apparatus and method for inter-board interconnection in a router cluster. Background technique
  • the core equipment of the network faces increasingly severe challenges, especially the capacity and scalability of the core router of the backbone network.
  • high-end core routers have high requirements in terms of capacity, security and reliability, and multi-service bearer capabilities.
  • the core routers of backbone networks must have good scalability and upgrade capabilities to adapt to the Internet. Changeable and rapid development trend.
  • Router cluster technology is the most effective way to solve the bottleneck of router capacity. It is a great leap and change in the history of router development. It breaks the traditional router expansion mode, while retaining the clarity of the network structure, facilitating operation management and facilitating maintenance. Without increasing maintenance costs.
  • Router clusters also known as router matrices or multi-chassis interconnects, are: By using parallel switching technology, two or more common core routers are connected in some way to form a multi-level multi-plane switching matrix system. It can work together and behave only as a logical router, thus breaking through the limitations of single chassis in terms of switching capacity, power consumption, heat dissipation, etc., and smoothly expanding to a larger capacity routing switching system.
  • the high-performance core router represents the development direction of the Internet, and it also best reflects the technical level of equipment manufacturers. Review the development history of core routers, Analysis of the current mainstream routing products features, we can see that the current core router technology development trends have the following aspects:
  • the capacity is getting higher and higher, and the port type is increased:
  • the main performance is the increase of the slot number and the high density of the line card.
  • the port capacity is increased from 2.5G to 10G, and then from 10G to 40G.
  • the type of interface is gradually increasing.
  • the system architecture is gradually improved: from the first generation single bus single CPU to the second generation single bus master and slave CPU, after the third generation single bus symmetric multi-CPU and fourth-generation multi-bus multi-CPU, to the present
  • the fifth generation shared memory and sixth generation crossbar system In terms of switching fabrics, high-end routers mainly use shared memory structures, and T-bit large-capacity routers mostly use multi-level switching fabrics to achieve future expansion to large-capacity non-blocking switching.
  • QoS Quality of Service
  • the control plane link provides a channel for control messages in the system.
  • the interaction between the control plane messages in the router cluster is generally implemented through the central switch box, which is equivalent to a message relay station, and the control plane links of each line card chassis are connected to On the center frame.
  • a link or Ethernet switch unit fails, it may affect a box or multi-frame user service.
  • a device for implementing control plane interconnection in a router cluster the first implementation scheme in the prior art is shown in FIG.
  • the device is in active/standby mode.
  • the two Ethernet switches in the chassis are in active/standby mode. Only the mainboard is switched, and the interconnect ports of the standby board are in the closed state. At a certain time, the main control board only
  • the Ethernet switch board can send and receive packets through an Ethernet switch board.
  • the center Ethernet switch board can only communicate with the link of an Ethernet switch board in the chassis.
  • the disadvantages of this implementation are: There are multiple single points of failure, such as when the link between the two central switching frames fails, resulting in a relationship between the two central switching frames, between the line card frame and the central switching frame. The communication chain is broken, and even the entire router cluster crashes. When there is a problem with the active and standby signal lines of the two Ethernet switch boards in the frame, the active/standby switchover cannot be completed. Finally, the communication link between the boards of the entire frame and other frames is disconnected. In general
  • FIG. 2 A device for implementing control plane interconnection in a router cluster, and a second implementation scheme in the prior art is shown in FIG. 2 .
  • the device uses a dual-NIC mode, that is, the board in the frame, such as the main control board, uses the dual-NIC type.
  • Each NIC is connected to an Ethernet switchboard in the frame.
  • the two NICs can send and receive packets at the same time.
  • the packet received by one NIC will not be sent to the Ethernet switchboard from another NIC.
  • the two Ethernet switching units in the frame can participate in packet exchange at the same time. There is no active/standby signal line or network port connection between them.
  • the main object of the present invention is to provide an apparatus and method for inter-board interconnection in a router cluster, which can improve reliability and solve the problem of communication disconnection between the line card frame and the central switching frame.
  • An apparatus for interconnecting boards in a router cluster comprising: at least two central switching frames, a line card frame, and a central Ethernet switchboard connected between the at least two central switching frames, the device further comprising: The Ethernet switch boards in the line card frame are respectively connected to the central Ethernet switch boards in the at least two central switch frames.
  • the first Ethernet switch board in the line card frame is interconnected with the first center Ethernet switch board, and the first Ethernet switch board is interconnected with the second center Ethernet switch board.
  • the second Ethernet switch board in the line card frame is interconnected with the first center Ethernet switch board, and the second Ethernet switch board is interconnected with the second center Ethernet switch board.
  • the first central Ethernet switchboard is interconnected with each of the Ethernet switchboards located in the other central switch chassis, and the second central Ethernet switchboard is interconnected with each of the Ethernet switchboards located in the other central switch chassis.
  • the communication between the central switch frame and the line card frame is established through the link between the interconnected first Ethernet switch board and the second center Ethernet switch board; or through the interconnected second Ethernet Establishing a communication between the central switching frame and the line card frame by using a link between the switch board and the second central Ethernet switch board;
  • the communication between the central switch fabric and the line card frame is established through the link between the interconnected first Ethernet switchboard and the first central Ethernet switchboard; or, through the interconnected second Ethernet exchange A link between the board and the first central Ethernet switch board establishes communication between the central switch frame and the line card frame.
  • the second Ethernet switch board and the first part are interconnected.
  • the link between the central switch fabric and the line card frame is established.
  • the communication between the central switch frame and the line card frame is established through the link between the first Ethernet switch board and the first center Ethernet switch board; or the first Ethernet switch board is interconnected.
  • a communication between the central switching frame and the line card frame is established with the link between the second central Ethernet switchboard.
  • the communication between the central switch fabric and the line card frame is established through the link between the first Ethernet switchboard and the second central Ethernet switchboard. Or establish a communication between the central switch frame and the line card frame through the link between the interconnected second Ethernet switch board and the first central Ethernet switch board; or through the interconnected second Ethernet switch board and the second center Ethernet switch board Linking between the central switch fabric and the line card frame; when the link between the second central Ethernet switch board and the second Ethernet switch board fails, the first Ethernet switch board interconnected with the first center Ethernet switch board Linking, establishing communication between the central switching frame and the line card frame; or establishing communication between the central switching frame and the line card frame through the link between the interconnected first Ethernet switching board and the second central Ethernet switching board; or through interconnection
  • the link between the second Ethernet switchboard and the first central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
  • a method for interconnecting boards in a router cluster comprising: connecting at least two central switch boards between the central switch frames, the method further comprising: respectively: separating the Ethernet switch boards in the line card frame The central Ethernet switch boards are connected to at least two central switching frames.
  • the method further includes: establishing, by using a link between the first Ethernet switch board and the second center Ethernet switch board, establishing communication between the central switch frame and the line card frame; or a chain between the interconnected second Ethernet switchboard and the second central Ethernet switchboard Road, establishing communication between the central switching frame and the line card frame;
  • the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the first Ethernet switch board and the first center Ethernet switch board; or The link between the second Ethernet switchboard and the first central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
  • the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected second Ethernet switch board and the second central Ethernet switch board; or, Establishing a communication between the central switch fabric and the line card frame by connecting the second Ethernet switch board and the first center Ethernet switch board;
  • the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected first Ethernet switch board and the first center Ethernet switch board; or, by interconnecting A link between the first Ethernet switchboard and the second central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
  • the method further includes: establishing a central switch frame by using a link between the first Ethernet switch board and the second center Ethernet switch board. Communication of the line card frame; or establishing communication between the central switch frame and the line card frame through the link between the interconnected second Ethernet switch board and the first central Ethernet switch board; or through the interconnected second Ethernet switch board and the second Establishing a communication between the central switch frame and the line card frame by the link between the central Ethernet switch boards;
  • the method further includes: establishing a central switch frame and a line card by using a link between the first Ethernet switchboard and the first central Ethernet switchboard.
  • the communication between the chassis and the line card frame is established through the link between the first Ethernet switch board and the second center Ethernet switch board; or the second Ethernet switch board and the first center Ethernet are interconnected.
  • the link between the switchboards establishes the communication between the central switch frame and the line card frame.
  • the Ethernet switch boards in the line card frame are respectively connected to the central Ethernet switch boards in at least two center switch frames.
  • FIG. 1 is a schematic structural diagram of an inter-board interconnection device of a router cluster in the prior art
  • FIG. 2 is a schematic structural diagram of an inter-board interconnection device of another router cluster in the prior art
  • Ethernet switch boards in the line card frame are respectively connected to the central Ethernet switch boards in at least two center switch frames.
  • connection in this paper is different from the “interconnection” which only includes physical connections in the general sense.
  • the “interconnection” in this paper means the interconnection of control planes between boards, which has both physical connection and software control. The relationship of the connection is not described.
  • a device for interconnecting boards in a router cluster which mainly includes the following contents:
  • the device includes: at least two central switching frames, a line card frame, and a central Ethernet switchboard connected between the at least two central switching frames, the device further comprising: the Ethernet switchboards in the line card frame and the at least The central Ethernet switch boards are connected in the center switch boxes.
  • first center Ethernet switch board located in one center switch frame is interconnected with the second center Ethernet switch board located in the other center switch frame, and the first Ethernet switch board located in one line card frame is interconnected with the first center Ethernet switch board.
  • the first Ethernet switchboard is interconnected with the second central Ethernet switchboard.
  • the second Ethernet switch board located in the line card frame is interconnected with the first center Ethernet switch board
  • the second Ethernet switch board is interconnected with the second center Ethernet switch board.
  • first central Ethernet switch board is interconnected with each of the Ethernet switch boards located in the other central switch frame
  • second central Ethernet switch board is interconnected with each of the Ethernet switch boards located in the other central switch frame
  • the communication between the central switch frame and the line card frame is established through the link between the interconnected first Ethernet switch board and the second center Ethernet switch board; or through the interconnected second Ethernet Establishing a communication between the central switching frame and the line card frame by using a link between the switch board and the second central Ethernet switch board;
  • the communication between the central switch fabric and the line card frame is established through the link between the interconnected first Ethernet switchboard and the first central Ethernet switchboard; or, through the interconnected second Ethernet exchange A link between the board and the first central Ethernet switch board establishes communication between the central switch frame and the line card frame.
  • the communication between the central switch frame and the line card frame is established through the link between the interconnected second Ethernet switch board and the second central Ethernet switch board; or, through the interconnected second Ethernet Establishing a communication between the central switching frame and the line card frame by using a link between the switch board and the first central Ethernet switch board;
  • the communication between the central switch frame and the line card frame is established through the link between the first Ethernet switch board and the first center Ethernet switch board; or the first Ethernet switch board is interconnected.
  • a communication between the central switching frame and the line card frame is established with the link between the second central Ethernet switchboard.
  • the communication between the central switch frame and the line card frame is established through the link between the interconnected first Ethernet switch board and the second center Ethernet switch board. Or establish a communication between the central switch frame and the line card frame through the link between the interconnected second Ethernet switch board and the first central Ethernet switch board; or through the interconnected second Ethernet switch board and the second center Ethernet switch board Link, establishing communication between the central switching frame and the line card frame;
  • the link between the second central Ethernet switchboard and the second Ethernet switchboard fails, the communication between the first Ethernet switchboard and the first central Ethernet switchboard is established, and communication between the central switch frame and the line card frame is established.
  • multiple links between the central switch frame and the line card frame can implement link aggregation using the Link Aggregation Control Protocol (LACP).
  • LACP Link Aggregation Control Protocol
  • the benefit of this is to avoid loops.
  • the present invention increases the interconnection between the second central Ethernet switch board and the first Ethernet switch board.
  • the link aggregation is not used, the first center Ethernet switch board and the second center Ethernet A loop is formed between the switch board and the first Ethernet switch board, and after the link aggregation is performed, the first Ethernet switch board is respectively interconnected with the first center Ethernet switch board and the second center Ethernet switch board.
  • a link is actually handled as a link aggregation group, which is equivalent to a link, thus avoiding loops.
  • the present invention further solves the problem of communication disconnection between the line card frame and the central switching frame, and further increases other possible interconnections in the device, covering all possible interconnections in the device, and may include the interior of the central switching frame.
  • Inter-board control plane interconnection inter-board control plane interconnection inside the line card frame
  • inter-board control plane interconnection between the central switching frames inter-board control plane interconnection between the central switching frames
  • inter-board control plane interconnection between the line card frame and the central switching frame inter-board control plane interconnection between the line card frame and the central switching frame.
  • the upgrade technology of the invention is safe, reliable, flexible and convenient, and can also ensure the smooth upgrade of the control plane.
  • the board interconnected by the control plane of the board includes: a central Ethernet switch board and an Ethernet switch board, and includes The central Ethernet switch board and peripheral boards other than the Ethernet switch board, such as the main control board.
  • the first level is: the peripheral board is exchanged on the main control board; the second level is the exchange on the Ethernet switch board; the third level is the center Ethernet exchange. Exchange on the board.
  • the central Ethernet switchboard is used to interconnect the interfaces in the router cluster, and the Ethernet switchboards of each frame are connected to the central Ethernet switchboard, thus implementing the hub for control communication between the interfaces.
  • the central Ethernet switchboard is implemented in both BOOT and major versions. When one central Ethernet switchboard is in the BOOT, the other can also implement the stacking function in the large version.
  • the BOOT refers to the degree of system startup. In order to adapt to an increasingly complex application environment in the electronic field, such as the communication industry and the automation industry, a corresponding main control central processor (CPU) is usually designed according to requirements. ), and BOOT is the basis for guiding the CPU system to start.
  • the large board is corresponding to BOOT.
  • BOOT corresponds to the bios setting and hardware detection phase when the computer starts, that is, the stage before starting the operating system; and the large version corresponds to the operating system startup phase.
  • the Ethernet switch board has two functions. One is to connect the central Ethernet switch board to communicate with other frames. The other is to establish the communication link between the peripheral boards in the frame. The communication between the peripheral boards of this frame must pass through the Ethernet switch board.
  • the Ethernet switchboards in this subrack are connected to each other. The connected ports are the stacking ports. The stacking function of the Ethernet switch board is implemented.
  • the peripheral board, such as the main control board is only a whole, just like an Ethernet switchboard. .
  • the Ethernet switchboard of the central switch chassis also needs to be connected to the central Ethernet switchboard of another central switch chassis. That is, the Ethernet switchboards of each chassis are connected to the central Ethernet switchboard. As a result, the redundancy of the link is achieved. Higher, the reliability will be improved accordingly.
  • the peripheral boards in each frame include the main control board, the line card, and the switching network board.
  • Each peripheral board has its own board.
  • the small switch chip, the peripheral board is connected to the Ethernet switch board in the frame through the small switch chip. Establish a communication link. In this way, the peripheral boards can communicate with each other through the Ethernet switch board.
  • a peripheral board is connected to two Ethernet switches. The two links are connected to each other through LACP.
  • Link aggregation allows one or more links to be aggregated together to form a link aggregation group.
  • the link aggregation group is treated as a separate link, that is, the ability to present the same MAC interface to the MAC application layer.
  • a packet When a packet is sent through an aggregation group formed by two ports on the peripheral board, it is determined according to the load sharing principle, which port in the aggregation group is sent. Different types of packets may be sent from different ports, so that communication bandwidth and reliability are both Can improve.
  • the advantages of the link aggregation and stacking of the present invention are as follows: Compared with the prior art, the second scheme of the prior art has a further defect: the switching time of the dual network card is relatively slow. When the communication between the main control board 21 and the main control board 23 is reached, the egress switch board 31, the central Ether switch board 41, and the Ether switch board 35 are selected. When a link in the middle fails, it needs to be very After a long time can be detected, and then switch to other links, the long interruption of control plane communication will affect the user's business operation. In short, the reliability is relatively low, and the performance of link switching is also relatively low.
  • the present invention implements link aggregation on all the boards through the small switch chip, and implements the stacking function on all the boards, and also implements the stacking function for the Ethernet switch board and the central Ethernet switch board.
  • the redundancy of the link is higher, the reliability and stability are greatly improved, and the bandwidth is also increased.
  • the boards can back up each other and share part of the load.
  • one or several links appear in the system. The fault is detected for up to 100 milliseconds, and the control plane communication is restored. When an Ethernet switchboard or the central Ethernet switchboard is reset, the communication between other boards is normal after 100 milliseconds. The impact of this type of fault on control plane communication is extremely low.
  • the central Ethernet switch board implements stacking in the BOOT.
  • the Ethernet switch board it is not aware of whether it is in the BOOT or in the large version.
  • the startup time of the Ethernet switchboard in the control center is made, so that they have a sequence and have Time difference to enter the large version, so as to ensure that they do not enter the large version at the same time, because the BOOT to the large version needs to load the kernel image, the board does not have the swap function during the process of loading the kernel image.
  • the communication interruption time between the boxes is relatively long. When they enter the large version when they are different, the system can be powered on and started to start, and the control plane communication between the frames will not be broken.
  • the invention is illustrated by way of example below.
  • the control plane interconnection device in the router cluster of the present invention includes: two central switching frames and one line card frame.
  • the interconnection of router clusters is generally one.
  • the present invention is applicable to the above several interconnection modes.
  • Each central switch chassis has a central Ethernet switch board.
  • the central switch frame has two central Ethernet switch boards.
  • Each frame has two Ethernet switch boards, and there are also some peripheral boards, such as line cards and main control boards.
  • peripheral boards such as line cards and main control boards.
  • each peripheral board is the same for the control plane interconnection. Now, only the main control board is taken as an example for explanation.
  • the boards in the chassis include: the central Ethernet switch board, the Ethernet switch board, and the peripheral boards except the center Ethernet switch board and the Ethernet switch board, such as the main control board).
  • the central Ethernet switch board it can be connected to the Ethernet switch board of the local chassis through the standby board or through the optical fiber.
  • the ports of the standby board and the optical fiber ports can be aggregated into one link aggregation group. This physical difference is not seen for the LACP protocol.
  • the link between the main control board and the Ethernet switch board is used.
  • the link ports on the main control board are negotiated with the ports on the Ethernet switch board to implement link aggregation.
  • the two Ethernet switch boards are stacked on the main control board. It is like a single board, so you can aggregate several ports into one aggregation group, just like a port for the upper application.
  • the main control board can have two or three ports, and even more ports are connected to one Ethernet switch board instead of one port, but at least one must be used to improve redundancy and reliability. . Number of interconnected ports It is mainly determined by the number of ports on the main control board d and the switch chip.
  • the ports on the peripheral boards are all set up, and they can also be aggregated. Then look at the operation and interconnection of the Ethernet switch board.
  • the stacking link between the two Ethernet switch boards can be two or more. This is just a schematic diagram. If there are multiple, link aggregation is required, and stack packets are sent and received through the stack link.
  • the Ethernet switch board of the center switch chassis needs to be connected to the central Ethernet switch board between the chassis. When a central Ethernet switch board fails, the control switch between the central switch frames and between the line card frame and the central switch frame is not controlled. Chain, and make the whole system paralyzed.
  • the two Ethernet switch boards are connected to the central Ethernet switch board between the two switches to prevent two faults.
  • the ports connected to the same card on the two Ethernet switch boards belong to a link aggregation group. They know each other's ports belong to which link aggregation group and whether the port is added to the link aggregation group. It is necessary to notify the other party through the timing synchronization and change synchronization of the stack information to maintain a stable stacking mode.
  • the two synchronization modes are used to maintain the ports of the two Ethernet switch boards. That is, the other port is added to the link aggregation group on the card or removed from the link aggregation group, and the port and link aggregation group are implemented. The information is consistent.
  • the port information is the information that the port is added to or removed from the link aggregation group and the link aggregation group number corresponding to the port.
  • After receiving the port information compare and update the port information that you have reserved.
  • the board will use this port. Add or delete from the link aggregation group, and immediately send a notification message to the other party.
  • the link aggregation group corresponding to the port is added or deleted. .
  • the two Ethernet switches need to maintain a global table of the ports that the other party has joined in the link aggregation group.
  • the contents are: link aggregation group number of the port, port number, and module number of the port.
  • the module number is the chip number of the device where the port resides. It is used to distinguish the Ethernet switch boards in the stack.
  • Timing synchronization is mainly to make up for the lack of synchronization failure. If the synchronization fails, the timing synchronization can ensure the consistency of the information at both ends.
  • the stack synchronization information between the two Ethernet switch boards is not sent as a type of message.
  • the format of the LACP packet is the packet header of the LACP packet.
  • the packet body stores the stack synchronization information.
  • IP Internet Protocol
  • MAC Media Access Control
  • the data contents of the stack synchronization are: link aggregation group number, port number, and synchronization type. Among them, the synchronization type specifies whether the synchronization or change synchronization.
  • Synchronous synchronization only synchronizes the port number of the link aggregation group and the link aggregation group number of the port.
  • the change synchronization only synchronizes the port number and port of the board that has just been deleted or added from the link aggregation group.
  • the stacking and link aggregation of the central Ethernet switchboard is performed, and the stacking and link aggregation are performed.
  • the way is the same as the Ethernet switch board.
  • the difference is that it implements the stacking function in the BOOT.
  • the other can also be stacked in the large version.
  • the stack information format in the BOOT and the processing of receiving the stack information should be the same as those in the large version.
  • the calculation of the module number of the port on the central Ethernet switch board must also be consistent.
  • the stack port is disconnected, causing the link between the two central Ethernet switch boards to fail. In this case, the other should take this from the BOOT to the center in the large version.
  • the port of the Ethernet switch is removed from the link aggregation group. After the large version is entered, the port connected to the Ethernet switch performs LACP packet negotiation. However, the static aggregation mode cannot be used because the ports are aggregated together. It is known that the system identifier of the LACP packet sent by each card in the system is represented by the chassis number, slot number, and CPU number. The port of the central Ethernet switch board is received. After the LACP packet is received, the chassis number of the other party is extracted, and the extracted chassis number is associated with the link aggregation group number to be added by the port. The rule is not only one type, as long as there is one type - corresponding The relationship is OK. Now set the rule as the rack number plus one. The value is the link aggregation group number.
  • the rules for the two central Ethernet switch boards are Straight, so that between them there is no need to send additional information to determine the link aggregation group number.
  • the port is added to the link aggregation group. If the stack port is added to the link aggregation group, the stack interface sends the stack information to the peer.
  • the connection between the Ethernet switchboard and the Ethernet switchboard between the chassis is not as good as that of Figure 3.
  • an Ethernet switchboard can connect two or more ports to the central Ethernet switchboard, or one between them.
  • the connection to one is not cross-connected, except that the connection shown in Figure 3 uses the least amount of fiber to maximize reliability. In the process of upgrading, there will be no upgrade failure due to the connection mode slightly different from Figure 3.
  • connection and configuration between the board and the Ethernet switch board, the Ethernet switch board, and the center Ethernet switch board are completed.
  • the control plane interconnection of the entire cluster is completed.
  • the solution of the invention not only makes the control plane interconnection more secure and reliable, but also makes the interconnection between the line card frame and the central switching frame more convenient and flexible, and the connection between the links and the links does not affect the quality of the system communication. the amount.
  • a method for interconnecting boards in a router cluster comprising: connecting at least two central switch boards between the central switch frames, the method further comprising: respectively: separating the Ethernet switch boards in the line card frame The central Ethernet switch boards are connected to at least two central switching frames.
  • the method includes: interconnecting a first central Ethernet switch board located in one central switch frame with a second center Ethernet switch board located in another central switch frame, and the first Ethernet switch board located in a line card frame and the first The central Ethernet switchboard interconnects the first Ethernet switchboard with the second central Ethernet switchboard.
  • the method further includes: establishing, by the link between the interconnected first Ethernet switchboard and the second central Ethernet switchboard, establishing communication between the central switch frame and the line card frame; or Establishing a communication between the central switch fabric and the line card frame by connecting the second Ethernet switch board and the second center Ethernet switch board;
  • the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the first Ethernet switch board and the first center Ethernet switch board; or The link between the second Ethernet switchboard and the first central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
  • the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected second Ethernet switch board and the second center Ethernet switch board; or Establishing a communication between the central switch fabric and the line card frame by connecting the second Ethernet switch board and the first center Ethernet switch board;
  • the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected first Ethernet switch board and the first center Ethernet switch board; or, by interconnecting A link between the first Ethernet switchboard and the second central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
  • the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected first Ethernet switch board and the second central Ethernet switch board; or connecting the second Ethernet switch board and the first center Ethernet through the interconnected
  • the communication between the central switching frame and the line card frame is established by the link between the switchboards, or the communication between the central switch frame and the line card frame is established through the link between the second Ethernet switch board and the second central Ethernet switch board;
  • the method further includes: establishing a central switch frame and a line card by using a link between the first Ethernet switchboard and the first central Ethernet switchboard.
  • the communication between the chassis and the line card frame is established through the link between the first Ethernet switch board and the second center Ethernet switch board; or the second Ethernet switch board and the first center Ethernet are interconnected.
  • the link between the switchboards establishes the communication between the central switch frame and the line card frame.

Abstract

The present invention discloses an inter-plate interconnection device in a router cluster, wherein the device includes: at least two central switching frames and a line card frame, wherein the central Ethernet switching plates of the at least two central switching frames are connected with each other and the Ethernet switching plates of the line card frame are connected with the central Ethernet switching plates of the at least two central switching frames respectively. The present invention also discloses an inter-plate interconnection method in a router cluster, wherein the method includes: connecting the central Ethernet switching plates of the at least two central switching frames with each other; connecting the Ethernet switching plates of the line card frame with the central Ethernet switching plates of the at least two central switching frames respectively. With the device and method of the present invention, the reliability is increased and the problem of the communication link interrupt between the line card frame and the central switching frames is solved.

Description

路由器集群中板间互联的装置及方法 技术领域  Device and method for interconnecting boards in router cluster
本发明涉及控制面互联的技术, 尤其涉及一种路由器集群中板间互联 的装置及方法。 背景技术  The present invention relates to a technique for controlling surface interconnection, and in particular, to an apparatus and method for inter-board interconnection in a router cluster. Background technique
随着互联网技术的高速发展,新业务不断涌现,特别是以点对点(P2P ) 和网络电视(IPTV )为代表的视频业务的崛起, 极大地刺激了带宽的需求, 互联网从最初的聊天、 电子邮件、 Web浏览等需求, 发展到今天的 P2P下 载, 视频聊天、 播客等。 无论是从应用模式, 还是网络流量, 都发生了翻 天覆地的变化。 因此网络核心设备面临越来越严峻的挑战, 尤其是对骨干 网核心路由器的容量和扩展性提出了极高的要求。 例如在 IP网络的电信化 过程中对高端核心路由器在容量、 安全可靠性、 多业务承载能力等方面提 出了极高的要求, 骨干网核心路由器必须具有良好的扩展性和升级能力, 以适应互联网多变、 飞速发展的趋势。  With the rapid development of Internet technology, new services are emerging, especially the rise of video services represented by peer-to-peer (P2P) and Internet TV (IPTV), which greatly stimulates the demand for bandwidth, the Internet from the initial chat, email Demand for web browsing, etc., to today's P2P downloads, video chats, podcasts, etc. Both the application model and the network traffic have undergone earth-shaking changes. Therefore, the core equipment of the network faces increasingly severe challenges, especially the capacity and scalability of the core router of the backbone network. For example, in the process of telecommunications of IP networks, high-end core routers have high requirements in terms of capacity, security and reliability, and multi-service bearer capabilities. The core routers of backbone networks must have good scalability and upgrade capabilities to adapt to the Internet. Changeable and rapid development trend.
路由器集群技术是目前解决路由器容量瓶颈的最有效方式, 是路由器 发展史上一个极大的飞跃和改变, 它打破了传统的路由器扩展模式, 同时 保留了网络结构的清晰度, 便于运营管理, 方便维护而不增加维护成本。 路由器集群又称路由器矩阵或多机框互联, 即为: 通过釆用并行交换技术, 将两台或两台以上普通核心路由器通过某种方式连接, 共同组成一个多级 多平面的交换矩阵系统, 使其能够协同工作, 并且对外只表现为一台逻辑 路由器, 从而突破单机箱在交换容量、 功耗、 散热等方面的限制, 平滑扩 展为更大容量的路由交换系统。 高性能核心路由器代表着互联网的发展方 向, 也最能体现出设备厂商的技术水平。 回顾核心路由器的发展历程, 分 析目前主流路由产品的功能特征, 可以看出目前核心路由器的技术发展趋 势主要有以下几个方面: Router cluster technology is the most effective way to solve the bottleneck of router capacity. It is a great leap and change in the history of router development. It breaks the traditional router expansion mode, while retaining the clarity of the network structure, facilitating operation management and facilitating maintenance. Without increasing maintenance costs. Router clusters, also known as router matrices or multi-chassis interconnects, are: By using parallel switching technology, two or more common core routers are connected in some way to form a multi-level multi-plane switching matrix system. It can work together and behave only as a logical router, thus breaking through the limitations of single chassis in terms of switching capacity, power consumption, heat dissipation, etc., and smoothly expanding to a larger capacity routing switching system. The high-performance core router represents the development direction of the Internet, and it also best reflects the technical level of equipment manufacturers. Review the development history of core routers, Analysis of the current mainstream routing products features, we can see that the current core router technology development trends have the following aspects:
1、 容量越来越高, 端口类型增多: 主要表现为槽位数的提高和线卡高 密度化。 首先端口能力从 2.5G提升到 10G, 再从 10G提升到 40G。 另外, 接口类型也逐步增多。  1. The capacity is getting higher and higher, and the port type is increased: The main performance is the increase of the slot number and the high density of the line card. First, the port capacity is increased from 2.5G to 10G, and then from 10G to 40G. In addition, the type of interface is gradually increasing.
2、 提高安全可靠性: 在网络发展初期, 由于流量较小, 网络攻击也没 有现在这么频繁, 路由器设备的安全问题主要体现在其自身的安全可靠性 上, 如网络设备的电气特性等。 随着网络的不断发展和多业务需求的增加, 系统越来越庞大, 一个小模块故障就影响到整个系统, 网络安全保障对路 由器的安全可靠性提出了更高的要求, 包括对网络攻击的防范和感知。  2. Improve security and reliability: In the early stage of network development, because of the small traffic and the lack of network attacks, the security of router devices is mainly reflected in its own security and reliability, such as the electrical characteristics of network devices. With the continuous development of the network and the increasing demand for multiple services, the system is getting larger and larger. A small module failure affects the entire system. Network security guarantees higher requirements for the security and reliability of the router, including attacks on the network. Prevention and perception.
3、 系统架构逐步完善: 从第一代单总线单 CPU发展到第二代单总线 主从 CPU、之后经历了第三代单总线对称式多 CPU和第四代多总线多 CPU, 到目前的第五代共享内存式和第六代交叉开关体系。 而在交换结构方面, 高端路由器主要釆用共享内存结构, T比特大容量路由器多数釆用多级交换 结构, 以实现未来扩展到大容量无阻塞交换。  3, the system architecture is gradually improved: from the first generation single bus single CPU to the second generation single bus master and slave CPU, after the third generation single bus symmetric multi-CPU and fourth-generation multi-bus multi-CPU, to the present The fifth generation shared memory and sixth generation crossbar system. In terms of switching fabrics, high-end routers mainly use shared memory structures, and T-bit large-capacity routers mostly use multi-level switching fabrics to achieve future expansion to large-capacity non-blocking switching.
4、 提升服务质量(QoS ): 运营级的 IP网络必须能够承载多种业务, 同时保证端到端的 QoS。 对于无连接的 IP网络来说, 实现服务质量保障一 直是个棘手的问题。 近年来, IP 网络的服务质量保障问题已经得到运营商 的极大重视, 设备厂商也为此提出了不同的解决方案。 目前, 保障服务质 量的技术手段主要有如下几种: 对流量进行整形、 重排; 解决头阻塞问题; 实现交换结构的无阻塞交换等。  4. Improve Quality of Service (QoS): Carrier-class IP networks must be able to carry multiple services while ensuring end-to-end QoS. For a connectionless IP network, achieving quality of service is always a tricky issue. In recent years, the service quality assurance problem of IP networks has been greatly valued by operators, and equipment manufacturers have proposed different solutions for this. At present, the technical means for guaranteeing service quality are mainly as follows: shaping and rearranging traffic; solving the problem of head blocking; implementing non-blocking switching of the switching fabric.
综上所述, 在路由器集群中, 可靠性在技术发展趋势中尤为重要, 其 中控制面的可靠性尤其备受关注, 控制平面链路为系统中控制消息提供通 道。 目前, 路由器集群中框间控制面消息的交互, 普遍通过中心交换机框 中来实现, 相当于一个消息中转站, 每个线卡机框的控制面链路都连接到 中心机框上。 当某条链路或以太网交换单元出现故障可能影响到一个框或 多框的用户业务。 In summary, in the router cluster, reliability is particularly important in the technology development trend, in which the reliability of the control plane is particularly concerned, and the control plane link provides a channel for control messages in the system. Currently, the interaction between the control plane messages in the router cluster is generally implemented through the central switch box, which is equivalent to a message relay station, and the control plane links of each line card chassis are connected to On the center frame. When a link or Ethernet switch unit fails, it may affect a box or multi-frame user service.
在路由器集群中实现控制面互联的装置, 现有技术中的第一种实现方 案如图 1 所示。 该装置釆用主备的方式, 即为: 每个框内的两个以太交换 板是主备的方式, 只有主板交换, 备板的互联端口都处于关闭状态, 在某 一个时刻主控板只能通过框内的一个以太交换板进行收发报文, 中心以太 交换板也只与框内的一个以太交换板的链路畅通。 这种实现方案的缺点为: 存在多处的单点故障, 如当两个中心交换框之间的链路出现故障, 会导致 两个中心交换框之间、 线卡框和中心交换框之间通信断链, 乃至整个路由 器集群崩溃。 当框内的两个以太交换板的主备信号线存在问题时, 而主备 倒换就不能完成, 最终使整个框的单板与其它框的通信链路断开。 总的来 说就是可靠性、 稳定性很低。  A device for implementing control plane interconnection in a router cluster, the first implementation scheme in the prior art is shown in FIG. The device is in active/standby mode. The two Ethernet switches in the chassis are in active/standby mode. Only the mainboard is switched, and the interconnect ports of the standby board are in the closed state. At a certain time, the main control board only The Ethernet switch board can send and receive packets through an Ethernet switch board. The center Ethernet switch board can only communicate with the link of an Ethernet switch board in the chassis. The disadvantages of this implementation are: There are multiple single points of failure, such as when the link between the two central switching frames fails, resulting in a relationship between the two central switching frames, between the line card frame and the central switching frame. The communication chain is broken, and even the entire router cluster crashes. When there is a problem with the active and standby signal lines of the two Ethernet switch boards in the frame, the active/standby switchover cannot be completed. Finally, the communication link between the boards of the entire frame and other frames is disconnected. In general, reliability and stability are low.
在路由器集群中实现控制面互联的装置, 现有技术中的第二种实现方 案如图 2所示。 该装置釆用的是双网卡方式, 即为: 框内的单板如主控板 釆用双网卡上式, 每个网卡分别连接框内的一个以太交换板, 两个网卡可 以同时收发报文, 一个网卡收到的报文不会从另外一个网卡发送给以太交 换板。 框内的两个以太交换单元可以同时参与报文交换, 相不影响, 它们 之间既没有主备信号线也没有网口相连。 这种方案的控制面链路虽然不存 在单点故障, 但是其缺点在于: 当有两处链路存在故障时, 就可能出现一 个框与其它的框通信链路都中断, 例如当以太交换板 35上与中心以太交换 板 41相连的端口故障, 且中心以太交换板 42上与以太交换板 36相连的端 口故障, 那么会导致线卡框与两个中心交换框间的通信断链。 可见: 釆用 现有技术, 无论哪种实现方案, 都无法避免线卡框和中心交换框间的通信 断链问题, 可靠性很低, 目前迫切需要一种可靠性高的路由器集群中实现 板间互联的实现方案。 发明内容 A device for implementing control plane interconnection in a router cluster, and a second implementation scheme in the prior art is shown in FIG. 2 . The device uses a dual-NIC mode, that is, the board in the frame, such as the main control board, uses the dual-NIC type. Each NIC is connected to an Ethernet switchboard in the frame. The two NICs can send and receive packets at the same time. The packet received by one NIC will not be sent to the Ethernet switchboard from another NIC. The two Ethernet switching units in the frame can participate in packet exchange at the same time. There is no active/standby signal line or network port connection between them. Although there is no single point of failure in the control plane link of this scheme, the disadvantage is that when there are two links faults, a frame and other frame communication links may be interrupted, for example, when the Ethernet switchboard The port connected to the central Ethernet switch board 41 is faulty, and the port connected to the Ethernet switch board 36 on the central Ethernet switch board 42 is faulty, which causes the communication between the line card frame and the two central switch frames to be broken. It can be seen that: 现有Using the existing technology, no matter which implementation scheme, the communication disconnection problem between the line card frame and the central switching frame cannot be avoided, and the reliability is very low. Currently, a highly reliable router cluster implementation board is urgently needed. The implementation of the interconnection. Summary of the invention
有鉴于此, 本发明的主要目的在于提供一种路由器集群中板间互联的 装置及方法, 能提高可靠性, 解决了线卡框和中心交换框间的通信断链问 题。  In view of the above, the main object of the present invention is to provide an apparatus and method for inter-board interconnection in a router cluster, which can improve reliability and solve the problem of communication disconnection between the line card frame and the central switching frame.
为达到上述目的, 本发明的技术方案是这样实现的:  In order to achieve the above object, the technical solution of the present invention is achieved as follows:
一种路由器集群中板间互联的装置, 该装置包括: 至少两个中心交换 框、 线卡框, 所述至少两个中心交换框之间的中心以太交换板相连, 该装 置还包括: 所述线卡框中的以太交换板分别与所述至少两个中心交换框中 的中心以太交换板相连。  An apparatus for interconnecting boards in a router cluster, the apparatus comprising: at least two central switching frames, a line card frame, and a central Ethernet switchboard connected between the at least two central switching frames, the device further comprising: The Ethernet switch boards in the line card frame are respectively connected to the central Ethernet switch boards in the at least two central switch frames.
其中, 位于所述线卡框的第一以太交换板与第一中心以太交换板互联, 所述第一以太交换板与第二中心以太交换板互联。  The first Ethernet switch board in the line card frame is interconnected with the first center Ethernet switch board, and the first Ethernet switch board is interconnected with the second center Ethernet switch board.
其中, 位于所述线卡框的第二以太交换板与第一中心以太交换板互联, 所述第二以太交换板与第二中心以太交换板互联。  The second Ethernet switch board in the line card frame is interconnected with the first center Ethernet switch board, and the second Ethernet switch board is interconnected with the second center Ethernet switch board.
其中, 第一中心以太交换板与位于另一个中心交换框的各个以太交换 板皆互联, 第二中心以太交换板与位于另一个中心交换框的各个以太交换 板皆互联。  The first central Ethernet switchboard is interconnected with each of the Ethernet switchboards located in the other central switch chassis, and the second central Ethernet switchboard is interconnected with each of the Ethernet switchboards located in the other central switch chassis.
其中, 第一中心以太交换板本身故障时, 通过互联的第一以太交换板 与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者 通过互联的第二以太交换板与第二中心以太交换板间的链路, 建立中心交 换框与线卡框的通信;  When the first center Ethernet switch board itself fails, the communication between the central switch frame and the line card frame is established through the link between the interconnected first Ethernet switch board and the second center Ethernet switch board; or through the interconnected second Ethernet Establishing a communication between the central switching frame and the line card frame by using a link between the switch board and the second central Ethernet switch board;
第二中心以太交换板本身故障时, 通过互联的第一以太交换板与第一 中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过 互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框 与线卡框的通信。  When the second central switchboard itself fails, the communication between the central switch fabric and the line card frame is established through the link between the interconnected first Ethernet switchboard and the first central Ethernet switchboard; or, through the interconnected second Ethernet exchange A link between the board and the first central Ethernet switch board establishes communication between the central switch frame and the line card frame.
其中, 第一以太交换板本身故障时, 通过互联的第二以太交换板与第 二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通 过互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换 框与线卡框的通信; Wherein, when the first Ethernet switch board itself fails, the second Ethernet switch board and the first part are interconnected. Establishing a central switching frame and a line card through the link between the second Ethernet switchboard and the first central Ethernet switchboard. The link between the central switch fabric and the line card frame is established. Box communication
第二以太交换板本身故障时, 通过互联的第一以太交换板与第一中心 以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过互联 的第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线 卡框的通信。  When the second Ethernet switch board itself is faulty, the communication between the central switch frame and the line card frame is established through the link between the first Ethernet switch board and the first center Ethernet switch board; or the first Ethernet switch board is interconnected. A communication between the central switching frame and the line card frame is established with the link between the second central Ethernet switchboard.
其中, 第一中心以太交换板与第一以太交换板间链路故障时, 通过互 联的第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与 线卡框的通信; 或者通过互联的第二以太交换板与第一中心以太交换板间 的链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以太交换 板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 第二中心以太交换板与第二以太交换板间链路故障时, 通过互联的第 一以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框 的通信; 或者通过互联的第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以太交换板与第一 中心以太交换板间的链路, 建立中心交换框与线卡框的通信。  When the link between the first central Ethernet switchboard and the first Ethernet switchboard is faulty, the communication between the central switch fabric and the line card frame is established through the link between the first Ethernet switchboard and the second central Ethernet switchboard. Or establish a communication between the central switch frame and the line card frame through the link between the interconnected second Ethernet switch board and the first central Ethernet switch board; or through the interconnected second Ethernet switch board and the second center Ethernet switch board Linking between the central switch fabric and the line card frame; when the link between the second central Ethernet switch board and the second Ethernet switch board fails, the first Ethernet switch board interconnected with the first center Ethernet switch board Linking, establishing communication between the central switching frame and the line card frame; or establishing communication between the central switching frame and the line card frame through the link between the interconnected first Ethernet switching board and the second central Ethernet switching board; or through interconnection The link between the second Ethernet switchboard and the first central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
一种路由器集群中板间互联的方法, 该方法包括: 将至少两个中心交 换框之间的中心以太交换板相连, 该方法还包括: 将所述线卡框中的以太 交换板分别与所述至少两个中心交换框中的中心以太交换板相连。  A method for interconnecting boards in a router cluster, the method comprising: connecting at least two central switch boards between the central switch frames, the method further comprising: respectively: separating the Ethernet switch boards in the line card frame The central Ethernet switch boards are connected to at least two central switching frames.
其中, 该方法还包括: 将位于所述线卡框的第一以太交换板与第一中 心以太交换板互联, 将所述第一以太交换板与第二中心以太交换板互联。  The method further includes: interconnecting the first Ethernet switch board located in the line card frame with the first center Ethernet switch board, and interconnecting the first Ethernet switch board with the second center Ethernet switch board.
其中, 第一中心以太交换板本身故障时, 该方法还包括: 通过互联的 第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡 框的通信; 或者通过互联的第二以太交换板与第二中心以太交换板间的链 路, 建立中心交换框与线卡框的通信; When the first center Ethernet switch board itself is faulty, the method further includes: establishing, by using a link between the first Ethernet switch board and the second center Ethernet switch board, establishing communication between the central switch frame and the line card frame; or a chain between the interconnected second Ethernet switchboard and the second central Ethernet switchboard Road, establishing communication between the central switching frame and the line card frame;
第二中心以太交换板本身故障时, 该方法还包括: 通过互联的第一以 太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通 信; 或者, 通过互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信。  When the second center Ethernet switch board itself is faulty, the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the first Ethernet switch board and the first center Ethernet switch board; or The link between the second Ethernet switchboard and the first central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
其中, 第一以太交换板本身故障时, 该方法还包括: 通过互联的第二 以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的 通信; 或者, 通过互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信;  When the first Ethernet switch board itself is faulty, the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected second Ethernet switch board and the second central Ethernet switch board; or, Establishing a communication between the central switch fabric and the line card frame by connecting the second Ethernet switch board and the first center Ethernet switch board;
第二以太交换板本身故障时, 该方法还包括: 通过互联的第一以太交 换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过互联的第一以太交换板与第二中心以太交换板间的链路, 建立 中心交换框与线卡框的通信。  When the second Ethernet switch board itself is faulty, the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected first Ethernet switch board and the first center Ethernet switch board; or, by interconnecting A link between the first Ethernet switchboard and the second central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
其中, 第一中心以太交换板与第一以太交换板间链路故障时, 该方法 还包括: 通过互联的第一以太交换板与第二中心以太交换板间的链路, 建 立中心交换框与线卡框的通信; 或者通过互联的第二以太交换板与第一中 心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互联 的第二以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线 卡框的通信;  When the link between the first central Ethernet switch board and the first Ethernet switch board is faulty, the method further includes: establishing a central switch frame by using a link between the first Ethernet switch board and the second center Ethernet switch board. Communication of the line card frame; or establishing communication between the central switch frame and the line card frame through the link between the interconnected second Ethernet switch board and the first central Ethernet switch board; or through the interconnected second Ethernet switch board and the second Establishing a communication between the central switch frame and the line card frame by the link between the central Ethernet switch boards;
第二中心以太交换板与第二以太交换板间链路故障时, 该方法还包括: 通过互联的第一以太交换板与第一中心以太交换板间的链路, 建立中心交 换框与线卡框的通信; 或者通过互联的第一以太交换板与第二中心以太交 换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以 太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通 信。 本发明的装置中线卡框中的以太交换板分别与至少两个中心交换框中 的中心以太交换板相连。 When the link between the second central switchboard and the second Ethernet switchboard is faulty, the method further includes: establishing a central switch frame and a line card by using a link between the first Ethernet switchboard and the first central Ethernet switchboard. The communication between the chassis and the line card frame is established through the link between the first Ethernet switch board and the second center Ethernet switch board; or the second Ethernet switch board and the first center Ethernet are interconnected. The link between the switchboards establishes the communication between the central switch frame and the line card frame. In the device of the present invention, the Ethernet switch boards in the line card frame are respectively connected to the central Ethernet switch boards in at least two center switch frames.
釆用本发明, 由于这种互联, 涵盖了装置中线卡框和中心交换框间所 有可能的互联, 因此使得可靠性得到大大的提高, 而且稳定性也得到大大 的提高, 从而解决了线卡框和中心交换框间的通信断链问题。 附图说明  By using the invention, since the interconnection covers all possible interconnections between the line card frame and the central switching frame in the device, the reliability is greatly improved, and the stability is also greatly improved, thereby solving the line card frame. Communication disconnection problem between the central exchange box. DRAWINGS
图 1为现有技术中一种路由器集群的板间互联装置的结构示意图; 图 2为现有技术中另一种路由器集群的板间互联装置的结构示意图; 图 3为本发明装置实施例的路由器集群的板间互联装置的结构示意图。 具体实施方式  1 is a schematic structural diagram of an inter-board interconnection device of a router cluster in the prior art; FIG. 2 is a schematic structural diagram of an inter-board interconnection device of another router cluster in the prior art; Schematic diagram of the inter-board interconnection device of the router cluster. detailed description
本发明的基本思想是: 线卡框中的以太交换板分别与至少两个中心交 换框中的中心以太交换板相连。  The basic idea of the invention is that the Ethernet switch boards in the line card frame are respectively connected to the central Ethernet switch boards in at least two center switch frames.
这里需要指出的是: 本文中的 "互联" 区别于通常意义上仅仅包括物 理连接的 "互连", 本文中的 "互联" 表示板间控制面互联, 既有物理连接 关系, 又有软件控制实现连接的关系, 不作赘述。  It should be pointed out here that the "interconnection" in this paper is different from the "interconnection" which only includes physical connections in the general sense. The "interconnection" in this paper means the interconnection of control planes between boards, which has both physical connection and software control. The relationship of the connection is not described.
下面结合附图对技术方案的实施作进一步的详细描述。  The implementation of the technical solution will be further described in detail below with reference to the accompanying drawings.
一种路由器集群中板间互联的装置, 主要包括以下内容:  A device for interconnecting boards in a router cluster, which mainly includes the following contents:
该装置包括: 至少两个中心交换框、 线卡框, 至少两个中心交换框之 间的中心以太交换板相连, 该装置还包括: 所述线卡框中的以太交换板分 别与所述至少两个中心交换框中的中心以太交换板相连。  The device includes: at least two central switching frames, a line card frame, and a central Ethernet switchboard connected between the at least two central switching frames, the device further comprising: the Ethernet switchboards in the line card frame and the at least The central Ethernet switch boards are connected in the center switch boxes.
这里, 位于一个中心交换框的第一中心以太交换板与位于另一个中心 交换框的第二中心以太交换板互联, 位于一个线卡框的第一以太交换板与 第一中心以太交换板互联, 第一以太交换板与第二中心以太交换板互联。  Here, the first center Ethernet switch board located in one center switch frame is interconnected with the second center Ethernet switch board located in the other center switch frame, and the first Ethernet switch board located in one line card frame is interconnected with the first center Ethernet switch board. The first Ethernet switchboard is interconnected with the second central Ethernet switchboard.
这里, 位于线卡框的第二以太交换板与第一中心以太交换板互联, 第 二以太交换板与第二中心以太交换板互联。 Here, the second Ethernet switch board located in the line card frame is interconnected with the first center Ethernet switch board, The second Ethernet switch board is interconnected with the second center Ethernet switch board.
这里, 第一中心以太交换板与位于另一个中心交换框的各个以太交换 板皆互联, 第二中心以太交换板与位于另一个中心交换框的各个以太交换 板皆互联。  Here, the first central Ethernet switch board is interconnected with each of the Ethernet switch boards located in the other central switch frame, and the second central Ethernet switch board is interconnected with each of the Ethernet switch boards located in the other central switch frame.
这里, 第一中心以太交换板本身故障时, 通过互联的第一以太交换板 与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者 通过互联的第二以太交换板与第二中心以太交换板间的链路, 建立中心交 换框与线卡框的通信;  Here, when the first center Ethernet switch board itself fails, the communication between the central switch frame and the line card frame is established through the link between the interconnected first Ethernet switch board and the second center Ethernet switch board; or through the interconnected second Ethernet Establishing a communication between the central switching frame and the line card frame by using a link between the switch board and the second central Ethernet switch board;
第二中心以太交换板本身故障时, 通过互联的第一以太交换板与第一 中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过 互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框 与线卡框的通信。  When the second central switchboard itself fails, the communication between the central switch fabric and the line card frame is established through the link between the interconnected first Ethernet switchboard and the first central Ethernet switchboard; or, through the interconnected second Ethernet exchange A link between the board and the first central Ethernet switch board establishes communication between the central switch frame and the line card frame.
这里, 第一以太交换板本身故障时, 通过互联的第二以太交换板与第 二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通 过互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换 框与线卡框的通信;  Here, when the first Ethernet switch board itself fails, the communication between the central switch frame and the line card frame is established through the link between the interconnected second Ethernet switch board and the second central Ethernet switch board; or, through the interconnected second Ethernet Establishing a communication between the central switching frame and the line card frame by using a link between the switch board and the first central Ethernet switch board;
第二以太交换板本身故障时, 通过互联的第一以太交换板与第一中心 以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过互联 的第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线 卡框的通信。  When the second Ethernet switch board itself is faulty, the communication between the central switch frame and the line card frame is established through the link between the first Ethernet switch board and the first center Ethernet switch board; or the first Ethernet switch board is interconnected. A communication between the central switching frame and the line card frame is established with the link between the second central Ethernet switchboard.
这里, 第一中心以太交换板与第一以太交换板间链路故障时, 通过互 联的第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与 线卡框的通信; 或者通过互联的第二以太交换板与第一中心以太交换板间 的链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以太交换 板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 第二中心以太交换板与所述第二以太交换板间链路故障时, 通过互联 的第一以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线 卡框的通信; 或者通过互联的第一以太交换板与第二中心以太交换板间的 链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以太交换板 与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信。 Here, when the link between the first central Ethernet switch board and the first Ethernet switch board is faulty, the communication between the central switch frame and the line card frame is established through the link between the interconnected first Ethernet switch board and the second center Ethernet switch board. Or establish a communication between the central switch frame and the line card frame through the link between the interconnected second Ethernet switch board and the first central Ethernet switch board; or through the interconnected second Ethernet switch board and the second center Ethernet switch board Link, establishing communication between the central switching frame and the line card frame; When the link between the second central Ethernet switchboard and the second Ethernet switchboard fails, the communication between the first Ethernet switchboard and the first central Ethernet switchboard is established, and communication between the central switch frame and the line card frame is established. Or establish a communication between the central switch frame and the line card frame through the link between the interconnected first Ethernet switch board and the second central Ethernet switch board; or through the interconnected second Ethernet switch board and the first center Ethernet switch board The link establishes communication between the central switching frame and the line card frame.
这里需要指出的是, 中心交换框与线卡框之间的多条链路可以釆用链 路聚合控制协议(LACP )实现链路聚合。这样做的好处是避免环路的产生。 以第一中心以太交换板、 第二中心以太交换板、 与第一以太交换板之间的 互联而言, 为了解决线卡框和中心交换框间的通信断链问题, 除了将第一 中心以太交换板与第一以太交换板互联之外, 本发明增加了第二中心以太 交换板与第一以太交换板的互联, 如果不釆用链路聚合, 第一中心以太交 换板、 第二中心以太交换板、 与第一以太交换板之间的互联就产生了环路, 而釆用链路聚合后, 第一以太交换板分别与第一中心以太交换板和第二中 心以太交换板互联的两条链路, 实际上作为一个链路聚合组处理, 相当于 一条链路, 从而避免了环路。  It should be noted here that multiple links between the central switch frame and the line card frame can implement link aggregation using the Link Aggregation Control Protocol (LACP). The benefit of this is to avoid loops. In order to solve the problem of communication disconnection between the line card frame and the central switch frame, in addition to the first center Ethernet switch, the first center Ethernet switch board, the second center Ethernet switch board, and the first Ethernet switch board are interconnected. The present invention increases the interconnection between the second central Ethernet switch board and the first Ethernet switch board. If the link aggregation is not used, the first center Ethernet switch board and the second center Ethernet A loop is formed between the switch board and the first Ethernet switch board, and after the link aggregation is performed, the first Ethernet switch board is respectively interconnected with the first center Ethernet switch board and the second center Ethernet switch board. A link is actually handled as a link aggregation group, which is equivalent to a link, thus avoiding loops.
这里需要指出的是, 本发明除了解决线卡框和中心交换框间的通信断 链问题, 进一步增加了装置中其他可能的互联, 涵盖了装置中所有可能的 互联, 可以包括中心交换框内部的板间控制面互联、 线卡框内部的板间控 制面互联、 中心交换框之间的板间控制面互联、 线卡框和中心交换框间的 板间控制面互联。 这样做的好处是使得整个装置的可靠性得到提高。 由于 这种互联是多级交互的互联方式, 且涵盖了装置中所有可能的互联, 因此 使得可靠性得到大大的提高, 而且稳定性也得到大大的提高。 本发明的方 案还涉及到路由器集群扩容时控制面升级的技术, 釆用本发明的升级技术, 安全可靠、 灵活方便, 同时也能保证控制平面的平滑升级。 本发明板间控 制面互联的单板除了包括: 中心以太交换板、 以太交换板之外, 还包括除 中心以太交换板和以太交换板之外的外围单板比如主控板。 It should be noted here that the present invention further solves the problem of communication disconnection between the line card frame and the central switching frame, and further increases other possible interconnections in the device, covering all possible interconnections in the device, and may include the interior of the central switching frame. Inter-board control plane interconnection, inter-board control plane interconnection inside the line card frame, inter-board control plane interconnection between the central switching frames, and inter-board control plane interconnection between the line card frame and the central switching frame. The benefit of this is that the reliability of the entire device is improved. Since this interconnection is a multi-level interconnection method and covers all possible interconnections in the device, the reliability is greatly improved, and the stability is greatly improved. The solution of the invention also relates to the technology of upgrading the control plane when the router cluster is expanded. The upgrade technology of the invention is safe, reliable, flexible and convenient, and can also ensure the smooth upgrade of the control plane. The board interconnected by the control plane of the board includes: a central Ethernet switch board and an Ethernet switch board, and includes The central Ethernet switch board and peripheral boards other than the Ethernet switch board, such as the main control board.
上述多级交互的互联方式釆用三级交换的方式时, 第一级为: 外围单 板如主控板上的交换; 第二级为以太交换板上的交换; 第三级为中心以太 交换板上的交换。  When the above-mentioned multi-level interconnection method uses the three-level switching mode, the first level is: the peripheral board is exchanged on the main control board; the second level is the exchange on the Ethernet switch board; the third level is the center Ethernet exchange. Exchange on the board.
针对中心以太交换板而言, 中心以太交换板用来实现路由器集群中各 个框间的互联, 且各个框的以太交换板都与中心以太交换板相连, 也就实 现了框间控制面通信的枢纽功能。 中心以太交换板在 BOOT和大版本中都 实现堆叠, 当一个中心以太交换板在 BOOT中, 另一个在大版本中也能实 现堆叠功能。 其中, 针对所述 BOOT而言, 所述 BOOT指系统启动程度, 为了适应电子领域如通信行业和自动化行业日益复杂的应用环境, 通常情 况下, 都会针对需求设计相应的主控中央处理器(CPU ), 而 BOOT则是引 导 CPU系统启动的基础。 针对所述大版本而言, 大板本是与 BOOT相对应 的。 BOOT是对应于计算机启动时 bios设置、 硬件检测阶段, 也就是启动 操作系统之前的阶段; 而大版本是对应于操作系统启动运行阶段。  For the central Ethernet switchboard, the central Ethernet switchboard is used to interconnect the interfaces in the router cluster, and the Ethernet switchboards of each frame are connected to the central Ethernet switchboard, thus implementing the hub for control communication between the interfaces. Features. The central Ethernet switchboard is implemented in both BOOT and major versions. When one central Ethernet switchboard is in the BOOT, the other can also implement the stacking function in the large version. For the BOOT, the BOOT refers to the degree of system startup. In order to adapt to an increasingly complex application environment in the electronic field, such as the communication industry and the automation industry, a corresponding main control central processor (CPU) is usually designed according to requirements. ), and BOOT is the basis for guiding the CPU system to start. For the large version, the large board is corresponding to BOOT. BOOT corresponds to the bios setting and hardware detection phase when the computer starts, that is, the stage before starting the operating system; and the large version corresponds to the operating system startup phase.
针对以太交换板而言, 以太交换板有两个功能, 一个是与中心以太交 换板连接实现本框与其它框的通信; 另一个是建立本框内外围单板之间的 通信链路, 即本框的外围单板之间通信都必须经过以太交换板。 本框内的 以太交换板有链路相连, 相连的端口成为堆叠口, 用来实现以太交换板的 堆叠功能, 使其外围单板如主控板感知的只是一个整体, 就像一个以太交 换板。 中心交换框的以太交换板也需要与另一个中心交换框的中心以太交 换板相连, 也就是说, 各个框的以太交换板都与中心以太交换板相连, 这 样一来, 链路的冗余度更高, 可靠性也就相应提高了。  For the Ethernet switch board, the Ethernet switch board has two functions. One is to connect the central Ethernet switch board to communicate with other frames. The other is to establish the communication link between the peripheral boards in the frame. The communication between the peripheral boards of this frame must pass through the Ethernet switch board. The Ethernet switchboards in this subrack are connected to each other. The connected ports are the stacking ports. The stacking function of the Ethernet switch board is implemented. The peripheral board, such as the main control board, is only a whole, just like an Ethernet switchboard. . The Ethernet switchboard of the central switch chassis also needs to be connected to the central Ethernet switchboard of another central switch chassis. That is, the Ethernet switchboards of each chassis are connected to the central Ethernet switchboard. As a result, the redundancy of the link is achieved. Higher, the reliability will be improved accordingly.
针对除中心以太交换板和以太交换板之外的外围单板而言, 针对每个 框内的外围单板包括主控板、 线卡、 交换网单板, 每个外围单板上都有自 己的小交换芯片, 外围单板通过此小交换芯片与本框内的以太交换板相连, 建立起通信链路。 这样外围单板通过以太交换板就能相互通信了。 外围单 板与两个以太交换板相连, 相连的两条链路通过 LACP 实现链路聚合, 链 路聚合允许一条或多条链路聚合到一起, 形成链路聚合群, 从而使得 MAC 应用层可以把链路聚合群当作一条单独的链路来对待,也就是为 MAC应用 层呈现同一 MAC接口的能力。报文经过外围单板两个端口所形成的聚合组 发送时, 根据负荷分担原则, 决定从此聚合组中的哪个端口发送, 不同类 型的报文可能从不同端口发送, 使通信带宽和可靠性都能提高。 For the peripheral boards except the central Ethernet switch board and the Ethernet switch board, the peripheral boards in each frame include the main control board, the line card, and the switching network board. Each peripheral board has its own board. The small switch chip, the peripheral board is connected to the Ethernet switch board in the frame through the small switch chip. Establish a communication link. In this way, the peripheral boards can communicate with each other through the Ethernet switch board. A peripheral board is connected to two Ethernet switches. The two links are connected to each other through LACP. Link aggregation allows one or more links to be aggregated together to form a link aggregation group. The link aggregation group is treated as a separate link, that is, the ability to present the same MAC interface to the MAC application layer. When a packet is sent through an aggregation group formed by two ports on the peripheral board, it is determined according to the load sharing principle, which port in the aggregation group is sent. Different types of packets may be sent from different ports, so that communication bandwidth and reliability are both Can improve.
这里需要指出的是, 本发明釆用链路聚合和堆叠的好处在于: 对比本 发明与现有技术可知, 现有技术的第二种方案还有一个缺陷为: 双网卡的 切换时间比较慢, 达到秒级, 如主控板 21和主控板 23之间通信时选择的 通路是以太交换板 31、 中心以太交换板 41至以太交换板 35 , 当中间某条 链路出现故障后, 需要很长的时间才能检测出来, 再切换到其它链路, 控 制面通信中断的时间长就会影响到用户的业务运行。 总之就是可靠性比较 低, 链路切换的性能也比较低。 而本发明在所有单板之间都通过小交换芯 片进行互联, 在所有单板上都实现了链路聚合, 另外对于以太交换板和中 心以太交换板还实现了堆叠功能, 一方面使系统中链路的冗余度更高, 可 靠性、 稳定性大幅度提高, 带宽也增加了, 单板之间既能相互备份又可以 分担部分负荷; 另一方面当系统中一条或几条链路出现故障, 最长 100 毫 秒就会检测出来, 控制面通信就恢复正常, 当某个以太交换板或中心以太 交换板出现故障复位时, 也是最长 100 毫秒后其它单板之间的通信就会正 常, 这一类故障对控制面通信的影响降到极低。 中心以太交换板在 BOOT 中实现堆叠,对以太交换板来说是感知不到其在 BOOT中还是在大版本中, 再在 BOOT里控制中心以太交换板的启动时间, 使它们有先后顺序并且有 时间差地进入大版本, 以此保证它们不同时进入大版本, 因为 BOOT到大 版本需要加载内核映像, 加载内核映像的过程中单板的没有交换功能而使 各个框之间的通信中断时间比较长, 当它们不同时进入大版本时, 可以实 现系统加电启动到启动完成, 框间的控制面通信不会断。 It should be noted that the advantages of the link aggregation and stacking of the present invention are as follows: Compared with the prior art, the second scheme of the prior art has a further defect: the switching time of the dual network card is relatively slow. When the communication between the main control board 21 and the main control board 23 is reached, the egress switch board 31, the central Ether switch board 41, and the Ether switch board 35 are selected. When a link in the middle fails, it needs to be very After a long time can be detected, and then switch to other links, the long interruption of control plane communication will affect the user's business operation. In short, the reliability is relatively low, and the performance of link switching is also relatively low. The present invention implements link aggregation on all the boards through the small switch chip, and implements the stacking function on all the boards, and also implements the stacking function for the Ethernet switch board and the central Ethernet switch board. The redundancy of the link is higher, the reliability and stability are greatly improved, and the bandwidth is also increased. The boards can back up each other and share part of the load. On the other hand, one or several links appear in the system. The fault is detected for up to 100 milliseconds, and the control plane communication is restored. When an Ethernet switchboard or the central Ethernet switchboard is reset, the communication between other boards is normal after 100 milliseconds. The impact of this type of fault on control plane communication is extremely low. The central Ethernet switch board implements stacking in the BOOT. For the Ethernet switch board, it is not aware of whether it is in the BOOT or in the large version. In the BOOT, the startup time of the Ethernet switchboard in the control center is made, so that they have a sequence and have Time difference to enter the large version, so as to ensure that they do not enter the large version at the same time, because the BOOT to the large version needs to load the kernel image, the board does not have the swap function during the process of loading the kernel image. The communication interruption time between the boxes is relatively long. When they enter the large version when they are different, the system can be powered on and started to start, and the control plane communication between the frames will not be broken.
以下对本发明进行举例阐述。  The invention is illustrated by way of example below.
装置实施例:  Device embodiment:
如图 3 所示, 本发明路由器集群中的控制面互联装置包括: 两个中心 交换框、 一个线卡框, 这个只是为了方便说明方法, 而进行的简化, 实际 上路由器集群的互联一般为一个中心交换框和四个线卡框、 两个中心交换 框和八个线卡框、 四个中心交换框和十六个线卡框。 本发明在上述几种互 联方式下都适用。 每个中心交换框都有一个中心以太交换板, 在一个中心 交换框和四个线卡框的方式下, 中心交换框有两个中心以太交换板。 每个 框都有两个以太交换板, 并且还有一些外围单板, 如线卡、 主控板, 为了 更能简洁的阐述本发明, 且各个外围单板对控制面互联来说都是一样的, 现只以主控板为例来进行说明。  As shown in FIG. 3, the control plane interconnection device in the router cluster of the present invention includes: two central switching frames and one line card frame. This is simplified for convenience of description. In fact, the interconnection of router clusters is generally one. Center switch box and four line card boxes, two center switch boxes and eight line card boxes, four center switch boxes, and sixteen line card boxes. The present invention is applicable to the above several interconnection modes. Each central switch chassis has a central Ethernet switch board. In a central switch frame and four line card frames, the central switch frame has two central Ethernet switch boards. Each frame has two Ethernet switch boards, and there are also some peripheral boards, such as line cards and main control boards. In order to explain the present invention more succinctly, each peripheral board is the same for the control plane interconnection. Now, only the main control board is taken as an example for explanation.
以下对装置中各部件的连接、 链路聚合和堆叠进行具体阐述。  The following is a detailed description of the connections, link aggregation, and stacking of the various components in the device.
所有框间的单板(这里的单板包括: 中心以太交换板、 以太交换板、 和除中心以太交换板和以太交换板之外的外围单板比如主控板)连线都是 以光纤连的, 对于中心以太交换板来说其与本框的以太交换板相连时可以 通过备板, 也可以通过光纤, 其走备板的端口和走光纤的端口可以聚合到 一个链路聚合组中, 对 LACP协议来说是看不到这种物理上的差别。  The boards in the chassis (the boards here include: the central Ethernet switch board, the Ethernet switch board, and the peripheral boards except the center Ethernet switch board and the Ethernet switch board, such as the main control board). For the central Ethernet switch board, it can be connected to the Ethernet switch board of the local chassis through the standby board or through the optical fiber. The ports of the standby board and the optical fiber ports can be aggregated into one link aggregation group. This physical difference is not seen for the LACP protocol.
先看主控板与以太交换板的连接, 主控板上互联端口通过与以太交换 板上的端口进行 LACP报文协商而实现链路聚合, 且两个以太交换板实现 堆叠对主控板来说就象单独的一个单板, 因此可以把若干个端口聚合到一 个聚合组中, 对上层应用来说就像一个端口。 主控板可有两个或三个端口, 甚至更多的端口与一个以太交换板相连, 而不是限制为一个端口, 但是至 少要有一个, 这样才能取到提高冗余度和可靠性的效果。 互联的端口个数 主要由主控板上 d、交换芯片的端口数量来决定。 The link between the main control board and the Ethernet switch board is used. The link ports on the main control board are negotiated with the ports on the Ethernet switch board to implement link aggregation. The two Ethernet switch boards are stacked on the main control board. It is like a single board, so you can aggregate several ports into one aggregation group, just like a port for the upper application. The main control board can have two or three ports, and even more ports are connected to one Ethernet switch board instead of one port, but at least one must be used to improve redundancy and reliability. . Number of interconnected ports It is mainly determined by the number of ports on the main control board d and the switch chip.
外围单板上的端口都设置好了, 也可以进行聚合了, 再看以太交换板 的操作以及互联情况。 两个以太交换板之间的堆叠链路可以是两个或多个, 这里只是一个示意图, 如果为多个的话, 需要进行链路聚合, 堆叠报文通 过堆叠链路收发。 本中心交换框的以太交换板还需要与框间的中心以太交 换板相连, 这样当一个中心以太交换板故障时, 不至于中心交换框之间以 及线卡框与中心交换框之间控制面断链, 而使整个系统瘫痪。 让两个以太 交换板与框间的中心以太交换板相连, 是为了防止两点故障, 即如果本中 心交换框中的中心以太交换板和一个以太交换板都故障时, 都不影响本框 与其它框的通信, 整个系统也能正常运行。 因为框内的单板都是插在固定 的槽位, 槽位数也是固定的, 这样就能保证事先知道以太交换板的哪些端 口连接同一个单板, 而连接一个单板的端口都应该加入一个链路聚合组, 最后就知道哪些端口属于一个链路聚合组, 因为两个中心以太交换板实现 堆叠, 所以与它们两个中心以太交换板相连的端口也属于一个聚合组, 有 了这种固定的关系就可以用 LACP协议进行静态聚合了。 以太交换板与其 它单板的连接情况都设计好了, 下一步就要进行堆叠了。 堆叠时两个以太 交换板上与同一个单板相连的端口都属于一个链路聚合组, 它们要互相知 道对方的哪些端口属于哪个链路聚合组以及端口是否加入到链路聚合组 中, 因此需要通过堆叠信息的定时同步和变化同步来通知对方, 达到维持 一种稳定的堆叠模式。 这两种同步方式来维护两个以太交换板的端口, 也 就是把对方的端口加入到本单板上的链路聚合组中或从链路聚合组中删 除, 并实现端口和链路聚合组的信息一致。 利用 LACP协议在静态聚合方 式下, 端口信息就是指端口加入到链路聚合组或从链路聚合组删除的信息 , 以及端口对应的链路聚合组编号。 一方面需要定时同步端口信息到对方, 就是将本以太交换板中的端口信息定时同步到另一个以太交换板, 对方接 收到端口信息后, 要和自己保留的端口信息进行比较并更新; 另一方面是 变化同步端口信息到另一个以太交换板, 当本板端口的状态发送了改变, 本板就要将这个端口从链路聚合组中添加或者删除, 同时要立即发送通知 消息到对方, 对方接收到端口的状态信息发生改变后, 就在这个端口对应 的链路聚合组中, 完成对这个端口添加或者删除动作。 两个以太交换板要 保持一张对方已加入链路聚合组中的端口的全局表, 内容有: 端口的链路 聚合组号, 端口号, 端口的模块号。 其中, 模块号是端口所在设备的芯片 号, 用来区分堆叠时的各个以太交换板。 当收到对方的变化同步或定时同 步报文后, 把报文中的信息和全局表中的信息进行比较, 更新全局表中的 信息, 以此来达到本板与对方的信息同步。 定时同步保证全局表中内容的 一致性, 变化同步保证全局表中内容的实时性, 其只是有端口状态发生改 变后同步一次, 否则不会进行变化同步。 定时同步主要是弥补变化同步失 败的不足, 如果变化同步失败, 过一段时间定时同步还是能保证两端信息 的一致。 The ports on the peripheral boards are all set up, and they can also be aggregated. Then look at the operation and interconnection of the Ethernet switch board. The stacking link between the two Ethernet switch boards can be two or more. This is just a schematic diagram. If there are multiple, link aggregation is required, and stack packets are sent and received through the stack link. The Ethernet switch board of the center switch chassis needs to be connected to the central Ethernet switch board between the chassis. When a central Ethernet switch board fails, the control switch between the central switch frames and between the line card frame and the central switch frame is not controlled. Chain, and make the whole system paralyzed. The two Ethernet switch boards are connected to the central Ethernet switch board between the two switches to prevent two faults. If the central Ethernet switch board and an Ethernet switch board in the center switch chassis are faulty, this frame does not affect the frame. Other frame communication, the entire system can also run normally. Because the boards in the frame are inserted in the fixed slots, the number of slots is fixed. This ensures that the ports of the Ethernet switch board are connected to the same board in advance, and the ports connected to one board should be added. A link aggregation group, and finally knows which ports belong to a link aggregation group. Because the two central Ethernet switch boards implement the stack, the ports connected to the two central Ethernet switch boards also belong to an aggregation group. Fixed relationships can be statically aggregated using the LACP protocol. The connection between the Ethernet switch board and other boards is designed. The next step is to stack. The ports connected to the same card on the two Ethernet switch boards belong to a link aggregation group. They know each other's ports belong to which link aggregation group and whether the port is added to the link aggregation group. It is necessary to notify the other party through the timing synchronization and change synchronization of the stack information to maintain a stable stacking mode. The two synchronization modes are used to maintain the ports of the two Ethernet switch boards. That is, the other port is added to the link aggregation group on the card or removed from the link aggregation group, and the port and link aggregation group are implemented. The information is consistent. In the static aggregation mode, the port information is the information that the port is added to or removed from the link aggregation group and the link aggregation group number corresponding to the port. On the one hand, you need to synchronize the port information to the other party. You can synchronize the port information in the Ethernet switch board to another Ethernet switch board. After receiving the port information, compare and update the port information that you have reserved. On the other hand, change the synchronization port information to another Ethernet switch board. When the status of the board port is changed, the board will use this port. Add or delete from the link aggregation group, and immediately send a notification message to the other party. After the status information of the received port is changed, the link aggregation group corresponding to the port is added or deleted. . The two Ethernet switches need to maintain a global table of the ports that the other party has joined in the link aggregation group. The contents are: link aggregation group number of the port, port number, and module number of the port. The module number is the chip number of the device where the port resides. It is used to distinguish the Ethernet switch boards in the stack. After receiving the change synchronization or timing synchronization message of the other party, the information in the message is compared with the information in the global table, and the information in the global table is updated, thereby achieving synchronization between the information of the board and the other party. Timing synchronization ensures the consistency of the content in the global table. The synchronization synchronization ensures the real-time nature of the content in the global table. It only synchronizes once the port status changes, otherwise the change synchronization will not occur. Timing synchronization is mainly to make up for the lack of synchronization failure. If the synchronization fails, the timing synchronization can ensure the consistency of the information at both ends.
两个以太交换板之间的堆叠同步信息不另外当作一种报文发送, 而用 The stack synchronization information between the two Ethernet switch boards is not sent as a type of message.
LACP报文的格式, 即用 LACP报文的报文头, 报文体存放堆叠同步信息。 这样处理的好处是当两个以太交换板只能收发 LACP报文, 而不能收发其 它报文的情况时, 避免堆叠出现故障, 而使控制面通信断链, 还有一个好 处是不需要互相知道对方网口的网际协议( IP )地址或媒体接入控制( MAC ) 地址。 堆叠同步的数据内容有: 链路聚合组号, 端口号, 同步类型。 其中, 同步类型指定时同步还是变化同步。 定时同步只同步本板中已加入链路聚 合组的端口号和端口所在的链路聚合组号, 变化同步只同步本板中刚从链 路聚合组中删除或添加的端口号和端口所在的链路聚合组号, 还有添加或 删除的信息。 The format of the LACP packet is the packet header of the LACP packet. The packet body stores the stack synchronization information. The advantage of this processing is that when two Ethernet switch boards can only send and receive LACP packets, but cannot send and receive other packets, the stack is prevented from being faulty, and the control plane communication is broken. Another advantage is that there is no need to know each other. The Internet Protocol (IP) address or Media Access Control (MAC) address of the other party's network port. The data contents of the stack synchronization are: link aggregation group number, port number, and synchronization type. Among them, the synchronization type specifies whether the synchronization or change synchronization. Synchronous synchronization only synchronizes the port number of the link aggregation group and the link aggregation group number of the port. The change synchronization only synchronizes the port number and port of the board that has just been deleted or added from the link aggregation group. Link aggregation group number, plus information added or deleted.
最后就进行中心以太交换板的堆叠和链路聚合了, 其堆叠和链路聚合 的方式跟以太交换板一样, 不同的是其在 BOOT中实现堆叠功能, 当一个 中心以太交换板在 BOOT中另一个在大版本中也能堆叠。 BOOT中的堆叠 信息格式、 对收到堆叠信息的处理和大版本中应该一致, 中心以太交换板 上端口的模块号的计算方式也要保持一致。当一个中心以太交换板从 BOOT 中进入大版本时, 堆叠端口就会断开, 使两个中心以太交换板之间的链路 不通, 这时另一个应该把这个从 BOOT到大版本中的中心以太交换板的端 口从链路聚合组中删除, 当进入大版本后, 其与以太交换板相连的端口进 行 LACP报文协商, 但不能通过静态聚合的方式, 因为哪些端口聚合到一 起事先是不知道的, 所以只能通过其它方式了, 此方式就是把系统中各个 单板所发送 LACP报文中的系统标识符用机框号、 槽位号、 CPU号表示, 中心以太交换板的端口收到 LACP报文后从中提取对方的机框号, 然后把 提取的机框号和端口所要加入的链路聚合组号通过一种规则对应起来, 规 则不是只有一种, 只要存在一种——对应的关系就可以了, 现在就把规则 定为机架号加一所得值就是链路聚合组号, 两个中心以太交换板所用的规 则要一直, 这样它们之间就不需要发送额外的信息来确定链路聚合组号。 协商成功后端口加入链路聚合组, 如果堆叠端口也加入链路聚合组, 那么 就立即通过堆叠口发送堆叠信息给对方。 中心以太交换板与框间的以太交 换板的连接并不要象图 3那样连接才能正常工作, 如一个以太交换板可以 出两个或更多的端口与中心以太交换板相连, 或它们之间一对一的相连不 交叉连接, 只是图 3 所示的连接是用最少的光纤而使可靠性达到最高。 在 升级的过程中不会因为连接方式稍与图 3不配, 就出现升级失败的情况。 Finally, the stacking and link aggregation of the central Ethernet switchboard is performed, and the stacking and link aggregation are performed. The way is the same as the Ethernet switch board. The difference is that it implements the stacking function in the BOOT. When one central Ethernet switchboard is in the BOOT, the other can also be stacked in the large version. The stack information format in the BOOT and the processing of receiving the stack information should be the same as those in the large version. The calculation of the module number of the port on the central Ethernet switch board must also be consistent. When a central Ethernet switch board enters a large version from the BOOT, the stack port is disconnected, causing the link between the two central Ethernet switch boards to fail. In this case, the other should take this from the BOOT to the center in the large version. The port of the Ethernet switch is removed from the link aggregation group. After the large version is entered, the port connected to the Ethernet switch performs LACP packet negotiation. However, the static aggregation mode cannot be used because the ports are aggregated together. It is known that the system identifier of the LACP packet sent by each card in the system is represented by the chassis number, slot number, and CPU number. The port of the central Ethernet switch board is received. After the LACP packet is received, the chassis number of the other party is extracted, and the extracted chassis number is associated with the link aggregation group number to be added by the port. The rule is not only one type, as long as there is one type - corresponding The relationship is OK. Now set the rule as the rack number plus one. The value is the link aggregation group number. The rules for the two central Ethernet switch boards are Straight, so that between them there is no need to send additional information to determine the link aggregation group number. After the negotiation is successful, the port is added to the link aggregation group. If the stack port is added to the link aggregation group, the stack interface sends the stack information to the peer. The connection between the Ethernet switchboard and the Ethernet switchboard between the chassis is not as good as that of Figure 3. For example, an Ethernet switchboard can connect two or more ports to the central Ethernet switchboard, or one between them. The connection to one is not cross-connected, except that the connection shown in Figure 3 uses the least amount of fiber to maximize reliability. In the process of upgrading, there will be no upgrade failure due to the connection mode slightly different from Figure 3.
上面所述的单板比如主控板与以太交换板间、 以太交换板间、 中心以 太交换板间的连接和设置都完成了, 那么整个集群的控制面互联就完成了。 本发明的方案不仅使控制面互联更安全可靠, 还让线卡框和中心交换框的 互联更方便灵活, 它们之间增减链路连接的时候都不会影响系统通信的质 量。 The connection and configuration between the board and the Ethernet switch board, the Ethernet switch board, and the center Ethernet switch board are completed. The control plane interconnection of the entire cluster is completed. The solution of the invention not only makes the control plane interconnection more secure and reliable, but also makes the interconnection between the line card frame and the central switching frame more convenient and flexible, and the connection between the links and the links does not affect the quality of the system communication. the amount.
一种路由器集群中板间互联的方法, 该方法包括: 将至少两个中心交 换框之间的中心以太交换板相连, 该方法还包括: 将所述线卡框中的以太 交换板分别与所述至少两个中心交换框中的中心以太交换板相连。  A method for interconnecting boards in a router cluster, the method comprising: connecting at least two central switch boards between the central switch frames, the method further comprising: respectively: separating the Ethernet switch boards in the line card frame The central Ethernet switch boards are connected to at least two central switching frames.
这里, 该方法包括: 将位于一个中心交换框的第一中心以太交换板与 位于另一个中心交换框的第二中心以太交换板互联, 将位于一个线卡框的 第一以太交换板与第一中心以太交换板互联, 将第一以太交换板与第二中 心以太交换板互联。  Here, the method includes: interconnecting a first central Ethernet switch board located in one central switch frame with a second center Ethernet switch board located in another central switch frame, and the first Ethernet switch board located in a line card frame and the first The central Ethernet switchboard interconnects the first Ethernet switchboard with the second central Ethernet switchboard.
这里, 第一中心以太交换板本身故障时, 该方法还包括: 通过互联的 第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡 框的通信; 或者通过互联的第二以太交换板与第二中心以太交换板间的链 路, 建立中心交换框与线卡框的通信;  When the first central switchboard itself is faulty, the method further includes: establishing, by the link between the interconnected first Ethernet switchboard and the second central Ethernet switchboard, establishing communication between the central switch frame and the line card frame; or Establishing a communication between the central switch fabric and the line card frame by connecting the second Ethernet switch board and the second center Ethernet switch board;
第二中心以太交换板本身故障时, 该方法还包括: 通过互联的第一以 太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通 信; 或者, 通过互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信。  When the second center Ethernet switch board itself is faulty, the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the first Ethernet switch board and the first center Ethernet switch board; or The link between the second Ethernet switchboard and the first central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
这里, 第一以太交换板本身故障时, 该方法还包括: 通过互联的第二 以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的 通信; 或者, 通过互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信;  Here, when the first Ethernet switch board itself is faulty, the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected second Ethernet switch board and the second center Ethernet switch board; or Establishing a communication between the central switch fabric and the line card frame by connecting the second Ethernet switch board and the first center Ethernet switch board;
第二以太交换板本身故障时, 该方法还包括: 通过互联的第一以太交 换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过互联的第一以太交换板与第二中心以太交换板间的链路, 建立 中心交换框与线卡框的通信。  When the second Ethernet switch board itself is faulty, the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected first Ethernet switch board and the first center Ethernet switch board; or, by interconnecting A link between the first Ethernet switchboard and the second central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
这里, 第一中心以太交换板与第一以太交换板间链路故障时, 该方法 还包括: 通过互联的第一以太交换板与所述第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以太交换板与第一 中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互 联的第二以太交换板与第二中心以太交换板间的链路, 建立中心交换框与 线卡框的通信; Here, when the link between the first central Ethernet switch board and the first Ethernet switch board is faulty, the method The method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected first Ethernet switch board and the second central Ethernet switch board; or connecting the second Ethernet switch board and the first center Ethernet through the interconnected The communication between the central switching frame and the line card frame is established by the link between the switchboards, or the communication between the central switch frame and the line card frame is established through the link between the second Ethernet switch board and the second central Ethernet switch board;
第二中心以太交换板与第二以太交换板间链路故障时, 该方法还包括: 通过互联的第一以太交换板与第一中心以太交换板间的链路, 建立中心交 换框与线卡框的通信; 或者通过互联的第一以太交换板与第二中心以太交 换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以 太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通 信。  When the link between the second central switchboard and the second Ethernet switchboard is faulty, the method further includes: establishing a central switch frame and a line card by using a link between the first Ethernet switchboard and the first central Ethernet switchboard. The communication between the chassis and the line card frame is established through the link between the first Ethernet switch board and the second center Ethernet switch board; or the second Ethernet switch board and the first center Ethernet are interconnected. The link between the switchboards establishes the communication between the central switch frame and the line card frame.
以上所述, 仅为本发明的较佳实施例而已, 并非用于限定本发明的保 护范围。  The above is only the preferred embodiment of the present invention and is not intended to limit the scope of the present invention.

Claims

权利要求书 Claim
1、 一种路由器集群中板间互联的装置, 所述装置包括: 至少两个中心 交换框、 线卡框, 所述至少两个中心交换框之间的中心以太交换板相连, 其特征在于, 所述装置还包括: 所述线卡框中的以太交换板分别与所述至 少两个中心交换框中的中心以太交换板相连。  A device for interconnecting boards in a router cluster, the device comprising: at least two central switching frames and a line card frame, wherein the central Ethernet switch boards are connected between the at least two central switching frames, wherein The device further includes: the Ethernet switch boards in the line card frame are respectively connected to the central Ethernet switch boards in the at least two center switch frames.
2、 根据权利要求 1所述的装置, 其特征在于, 位于所述线卡框的第一 以太交换板与第一中心以太交换板互联, 所述第一以太交换板与第二中心 以太交换板互联。  2. The device according to claim 1, wherein the first Ethernet switch board located in the line card frame is interconnected with the first center Ethernet switch board, the first Ethernet switch board and the second center Ethernet switch board. interconnected.
3、 根据权利要求 1所述的装置, 其特征在于, 位于所述线卡框的第二 以太交换板与第一中心以太交换板互联, 所述第二以太交换板与第二中心 以太交换板互联。  The device according to claim 1, wherein the second Ethernet switch board located in the line card frame is interconnected with the first center Ethernet switch board, and the second Ethernet switch board and the second center Ethernet switch board interconnected.
4、 根据权利要求 1所述的装置, 其特征在于, 第一中心以太交换板与 位于另一个中心交换框的各个以太交换板皆互联, 第二中心以太交换板与 位于另一个中心交换框的各个以太交换板皆互联。  4. The device according to claim 1, wherein the first central Ethernet switch board is interconnected with each of the Ethernet switch boards located in the other central switch frame, and the second central Ethernet switch board and the other central switch frame are located. Each Ethernet switch board is interconnected.
5、 根据权利要求 2或 3所述的装置, 其特征在于, 当第一中心以太交 换板本身故障时, 通过互联的第一以太交换板与第二中心以太交换板间的 链路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以太交换板 与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信;  The device according to claim 2 or 3, wherein when the first central Ethernet switch board itself fails, a center is established through a link between the interconnected first Ethernet switch board and the second center Ethernet switch board. The communication between the switch box and the line card frame is established; or the communication between the central switch frame and the line card frame is established through the link between the interconnected second Ethernet switch board and the second center Ethernet switch board;
当第二中心以太交换板本身故障时, 通过互联的第一以太交换板与第 一中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通 过互联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换 框与线卡框的通信。  When the second central Ethernet switch board itself fails, the communication between the central switch frame and the line card frame is established through the link between the interconnected first Ethernet switch board and the first central Ethernet switch board; or, through the interconnected second Ethernet The switch between the switchboard and the first central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
6、 根据权利要求 2或 3所述的装置, 其特征在于, 当第一以太交换板 本身故障时, 通过互联的第二以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过互联的第二以太交换板与第 一中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 当第二以太交换板本身故障时, 通过互联的第一以太交换板与第一中 心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过互 联的第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与 线卡框的通信。 The device according to claim 2 or 3, wherein when the first Ethernet switch board itself fails, a center switch is established through a link between the interconnected second Ethernet switch board and the second center Ethernet switch board. The communication between the frame and the line card frame; or, through the interconnected second Ethernet switch board and the The communication between the central switch fabric and the line card frame is established by the link between the central Ethernet switch boards; and the link between the first Ethernet switch board and the first center Ethernet switch board when the second Ethernet switch board itself fails. The communication between the central switch frame and the line card frame is established. The communication between the central switch frame and the line card frame is established through the link between the first Ethernet switch board and the second center Ethernet switch board.
7、 根据权利要求 2或 3所述的装置, 其特征在于, 当第一中心以太交 换板与第一以太交换板间链路故障时, 通过互联的第一以太交换板与第二 中心以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互 联的第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框与 线卡框的通信; 或者通过互联的第二以太交换板与第二中心以太交换板间 的链路, 建立中心交换框与线卡框的通信;  The device according to claim 2 or 3, wherein when the link between the first central Ethernet switchboard and the first Ethernet switchboard fails, the first Ethernet switchboard is interconnected with the second central Ethernet switch. The communication between the central switch frame and the line card frame is established by the link between the boards; or the communication between the central switch frame and the line card frame is established through the link between the second Ethernet switch board and the first central Ethernet switch board; or Establishing communication between the central switching frame and the line card frame through the link between the interconnected second Ethernet switch board and the second central Ethernet switch board;
当第二中心以太交换板与第二以太交换板间链路故障时, 通过互联的 第一以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡 框的通信; 或者通过互联的第一以太交换板与第二中心以太交换板间的链 路, 建立中心交换框与线卡框的通信; 或者通过互联的第二以太交换板与 第一中心以太交换板间的链路, 建立中心交换框与线卡框的通信。  When the link between the second central Ethernet switchboard and the second Ethernet switchboard is faulty, the communication between the central switch fabric and the line card frame is established through the link between the interconnected first Ethernet switchboard and the first central Ethernet switchboard; Or establishing communication between the central switch frame and the line card frame through the link between the interconnected first Ethernet switch board and the second central Ethernet switch board; or through the interconnection between the second Ethernet switch board and the first center Ethernet switch board Link, establishes communication between the central switching frame and the line card frame.
8、 一种路由器集群中板间互联的方法, 所述方法包括: 将至少两个中 心交换框之间的中心以太交换板相连, 其特征在于, 所述方法还包括: 将 所述线卡框中的以太交换板分别与所述至少两个中心交换框中的中心以太 交换板相连。  A method for interconnecting boards in a router cluster, the method comprising: connecting a central Ethernet switch board between at least two central switch frames, wherein the method further comprises: placing the line card frame The Ethernet switchboards in the middle are respectively connected to the central Ethernet switchboards in the at least two central switch frames.
9、 根据权利要求 8所述的方法, 其特征在于, 所述方法还包括: 将位 于所述线卡框的第一以太交换板与第一中心以太交换板互联, 将所述第一 以太交换板与第二中心以太交换板互联。  The method according to claim 8, wherein the method further comprises: interconnecting the first Ethernet switch board located in the line card frame with the first center Ethernet switch board, and switching the first Ethernet switch The board is interconnected with the second central Ethernet switch board.
10、 根据权利要求 9所述的方法, 其特征在于, 当第一中心以太交换 板本身故障时, 所述方法还包括: 通过互联的第一以太交换板与第二中心 以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互联的 第二以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡 框的通信; The method according to claim 9, wherein when the first central Ethernet switch board itself fails, the method further includes: interconnecting the first Ethernet switch board and the second center The communication between the central switch chassis and the line card frame is established through the link between the Ethernet switch boards. The communication between the central switch frame and the line card frame is established through the link between the second Ethernet switch board and the second central Ethernet switch board. ;
当第二中心以太交换板本身故障时, 所述方法还包括: 通过互联的第 一以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框 的通信; 或者, 通过互联的第二以太交换板与第一中心以太交换板间的链 路, 建立中心交换框与线卡框的通信。  When the second central switchboard itself is faulty, the method further includes: establishing communication between the central switch frame and the line card frame by using a link between the interconnected first Ethernet switchboard and the first central Ethernet switchboard; or The communication between the central switching frame and the line card frame is established through the link between the interconnected second Ethernet switchboard and the first central Ethernet switchboard.
11、 根据权利要求 9 所述的方法, 其特征在于, 当第一以太交换板本 身故障时, 所述方法还包括: 通过互联的第二以太交换板与第二中心以太 交换板间的链路, 建立中心交换框与线卡框的通信; 或者, 通过互联的第 二以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框 的通信;  The method according to claim 9, wherein when the first Ethernet switch board itself fails, the method further includes: connecting, by the interconnected second Ethernet switch board, the link between the second central Ethernet switch board and the second central Ethernet switch board And establishing communication between the central switching frame and the line card frame; or establishing communication between the central switching frame and the line card frame by using a link between the interconnected second Ethernet switch board and the first central Ethernet switch board;
当第二以太交换板本身故障时, 所述方法还包括: 通过互联的第一以 太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡框的通 信; 或者, 通过互联的第一以太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通信。  When the second Ethernet switch board itself is faulty, the method further includes: establishing, by using a link between the first Ethernet switch board and the first center Ethernet switch board, establishing communication between the central switch frame and the line card frame; or A link between the interconnected first Ethernet switchboard and the second central Ethernet switchboard establishes communication between the central switch frame and the line card frame.
12、 根据权利要求 9所述的方法, 其特征在于, 当第一中心以太交换 板与第一以太交换板间链路故障时, 所述方法还包括: 通过互联的第一以 太交换板与第二中心以太交换板间的链路, 建立中心交换框与线卡框的通 信; 或者通过互联的第二以太交换板与第一中心以太交换板间的链路, 建 立中心交换框与线卡框的通信; 或者通过互联的第二以太交换板与第二中 心以太交换板间的链路, 建立中心交换框与线卡框的通信;  The method according to claim 9, wherein when the link between the first central Ethernet switchboard and the first Ethernet switchboard fails, the method further includes: interconnecting the first Ethernet switchboard and the first The communication between the central switch fabric and the line card frame is established by the link between the two central Ethernet switch boards. The central switch frame and the line card frame are established through the link between the second Ethernet switch board and the first central Ethernet switch board. The communication between the central switch frame and the line card frame is established through the link between the interconnected second Ethernet switch board and the second central Ethernet switch board;
当第二中心以太交换板与第二以太交换板间链路故障时, 所述方法还 包括: 通过互联的第一以太交换板与第一中心以太交换板间的链路, 建立 中心交换框与线卡框的通信; 或者通过互联的第一以太交换板与第二中心 以太交换板间的链路, 建立中心交换框与线卡框的通信; 或者通过互联的 第二以太交换板与第一中心以太交换板间的链路, 建立中心交换框与线卡 框的通信。 When the link between the second central Ethernet switch board and the second Ethernet switch board is faulty, the method further includes: establishing a central switch frame by using a link between the interconnected first Ethernet switch board and the first central Ethernet switch board Communication of the line card frame; or through the interconnected first Ethernet switch board and the second center The communication between the central switch chassis and the line card frame is established by the link between the Ethernet switch boards, or the communication between the central switch frame and the line card frame is established through the link between the second Ethernet switch board and the first center Ethernet switch board. .
PCT/CN2011/072869 2010-08-09 2011-04-15 Inter-plate interconnection device and method in router cluster WO2012019464A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201010251620.4 2010-08-09
CN201010251620.4A CN101917337B (en) 2010-08-09 2010-08-09 Device and method for interconnecting router cluster middle plates

Publications (1)

Publication Number Publication Date
WO2012019464A1 true WO2012019464A1 (en) 2012-02-16

Family

ID=43324731

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/072869 WO2012019464A1 (en) 2010-08-09 2011-04-15 Inter-plate interconnection device and method in router cluster

Country Status (2)

Country Link
CN (1) CN101917337B (en)
WO (1) WO2012019464A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108632060A (en) * 2017-03-20 2018-10-09 中兴通讯股份有限公司 A kind of cascade method and device of router

Families Citing this family (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101917337B (en) * 2010-08-09 2015-06-03 中兴通讯股份有限公司 Device and method for interconnecting router cluster middle plates
CN102752190B (en) * 2011-04-22 2017-02-08 中兴通讯股份有限公司 Method and device for preventing loop from forming
CN102497308B (en) * 2011-11-30 2014-08-27 福建星网锐捷网络有限公司 Multiple slot switch
CN103379361A (en) * 2012-04-26 2013-10-30 东方有线网络有限公司 Method based on main information source and standby information source for achieving time shifting and repeated-watching program redundancy recording
CN102710451A (en) * 2012-06-21 2012-10-03 迈普通信技术股份有限公司 BASE-bus-based board card device and control method thereof
CN103269285B (en) * 2013-05-24 2016-08-17 杭州华三通信技术有限公司 Network service Extension arrangement and the method being used for realizing network service cluster
CN103336708B (en) * 2013-05-31 2016-12-28 华为技术有限公司 The polymerization method of network interface card resource and relevant device and computer system
CN104363181B (en) * 2014-08-28 2018-05-08 新华三技术有限公司 Flow transfer control method and device
CN104468389B (en) * 2014-11-27 2018-02-06 华为技术有限公司 Processing method, server and the server system of message
CN104822101B (en) * 2015-04-08 2018-08-24 新华三技术有限公司 The method and device of Fabric Network Synchronizations port diagnostic message
CN104902510B (en) * 2015-05-21 2018-10-30 华为技术有限公司 The loop detection method and base station controller system of base station controller system
CN106789753B (en) 2015-11-24 2020-06-26 新华三技术有限公司 Line card frame, multi-frame cluster router and message processing method
CN106789679B (en) 2015-11-24 2020-02-21 新华三技术有限公司 Line card frame, multi-frame cluster router, route selection method and message processing method
CN106533771B (en) * 2016-11-24 2019-12-06 新华三技术有限公司 Network equipment and control information transmission method
CN107547255B (en) * 2017-06-30 2020-12-01 新华三技术有限公司 Cluster system upgrading method, main control board and cluster system
CN108667509B (en) * 2018-03-22 2020-04-14 新华三技术有限公司 Cluster joining method and device and cluster equipment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030200330A1 (en) * 2002-04-22 2003-10-23 Maxxan Systems, Inc. System and method for load-sharing computer network switch
CN101355430A (en) * 2007-07-27 2009-01-28 华为技术有限公司 Exchange frame, cluster router
CN101394309A (en) * 2008-10-29 2009-03-25 华为技术有限公司 Cluster system expanding method, apparatus and cluster system
CN101917337A (en) * 2010-08-09 2010-12-15 中兴通讯股份有限公司 Device and method for interconnecting router cluster middle plates

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030200330A1 (en) * 2002-04-22 2003-10-23 Maxxan Systems, Inc. System and method for load-sharing computer network switch
CN101355430A (en) * 2007-07-27 2009-01-28 华为技术有限公司 Exchange frame, cluster router
CN101394309A (en) * 2008-10-29 2009-03-25 华为技术有限公司 Cluster system expanding method, apparatus and cluster system
CN101917337A (en) * 2010-08-09 2010-12-15 中兴通讯股份有限公司 Device and method for interconnecting router cluster middle plates

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108632060A (en) * 2017-03-20 2018-10-09 中兴通讯股份有限公司 A kind of cascade method and device of router

Also Published As

Publication number Publication date
CN101917337A (en) 2010-12-15
CN101917337B (en) 2015-06-03

Similar Documents

Publication Publication Date Title
WO2012019464A1 (en) Inter-plate interconnection device and method in router cluster
CN110912780B (en) High-availability cluster detection method, system and controlled terminal
CN111865779B (en) Route synchronization method and cross-device link aggregation group
EP3036873B1 (en) Dedicated control path architecture for stacked packet switches
US7639605B2 (en) System and method for detecting and recovering from virtual switch link failures
US9264346B2 (en) Resilient duplicate link aggregation emulation
EP2617165B1 (en) System and method for providing ethernet over infiniband virtual hub scalability in a middleware machine environment
US9300574B2 (en) Link aggregation emulation for virtual NICs in a cluster server
US6470013B1 (en) Use of enhanced ethernet link—loop packets to automate configuration of intelligent linecards attached to a router
Sidki et al. Fault tolerant mechanisms for SDN controllers
EP2961112B1 (en) Message forwarding system, method and device
WO2009023996A1 (en) Method for implementing network interconnect via link aggregation
WO2010048875A1 (en) Method, device of cluster system extension and cluster system
JP5609995B2 (en) COMMUNICATION SYSTEM, COMMUNICATION METHOD, AND COMMUNICATION DEVICE
JP2005354362A (en) Network repeater and its controlling method
US20120057453A1 (en) Method, apparatus and system for two-node cluster hot backup
US8625407B2 (en) Highly available virtual packet network device
WO2011003337A1 (en) Method, system and apparatus for protecting link data
CN115174468A (en) Route synchronization method, cross-device link aggregation group, electronic device and medium
CN111343019A (en) Reconnection method, reconnection device and storage medium
CN105763442A (en) PON system and method avoiding interruption of LACP aggregation link in main-standby switching process
JP4883317B2 (en) COMMUNICATION SYSTEM, NODE, TERMINAL, PROGRAM, AND COMMUNICATION METHOD
JP6118464B2 (en) Port status synchronization method, related device, and system
WO2013004115A1 (en) Method for message interaction among single boards in network device, and network device
KR100456461B1 (en) A Packet Transmission Method for Dual Redundancy of Network Layer and Transport Layer Protocol

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11816023

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11816023

Country of ref document: EP

Kind code of ref document: A1