WO2008009160A1

WO2008009160A1 - An access configuration method for the link aggregation and a link aggregation system

Info

Publication number: WO2008009160A1
Application number: PCT/CN2006/001511
Authority: WO
Inventors: Chunhua Qin
Original assignee: Zte Corporation
Priority date: 2006-06-30
Filing date: 2006-06-30
Publication date: 2008-01-24

Abstract

An access configuration method for the link aggregation and a link aggregation system. The n ports of the access side is set to a link aggregation group, where the unique MAC address is assigned to each port (S102); The m IP addresses are configured to the link aggregation group(S104); The m IP addresses are bound with the n ports so that the difference between the number of the IP address bound with each port and the number of the IP address bound with any other port is less than or equal to 1(S106); Each port sends the ARP message including the corresponding relationship between the assigned IP address of the port and the MAC address of the port(S108). When the connection of the port interrupts, the IP address assigned originally to the interrupted port is floated to the other effective ports. After the interrupted port is normal, the corresponding port is floated to the interrupted port. The floating principle is the difference between the number of the IP address bound with each effective port and the number of the IP address bound with any other effective port is less than or equal to 1. After floating, each effective port sends the ARP message and renovates the corresponding relationship between the IP address and the MAC address.

Description

The present invention relates to a communication network access technology, and in particular, to a link aggregation access configuration method and a link aggregation system. BACKGROUND In a communication system, an external port of a network element is generally interconnected with a switch, and then connected to a peer network device or a router through a switch to complete networking access, and a communication link bandwidth between the network element and the interconnected device is reliable. Sex is the key to ensuring the reliability and efficiency of user information traffic transmission. In order to improve the data transmission bandwidth and reliability, when the network element has multiple ports, it must implement traffic balancing and redundancy backup between multiple ports. Link aggregation aggregates multiple links to form a link aggregation group to improve bandwidth and reliability between interconnected devices. For example, two or three hundred megabit links can be bound together to implement different ports to share network traffic to increase network bandwidth, and ports are mutually backed up to ensure link redundancy. Commonly used link aggregation is applied between switches, between switches and routers, and between switches and servers, usually through the source MAC (Media Access Control) address and destination in data packets. The MAC address is hashed and the hash value is obtained. The forwarding port is matched according to the hash value to achieve link load balancing. A simpler approach is to get the corresponding forwarding port based on the XOR value of the last two bits of the source MAC address and the destination MAC address. In addition, a complex Link Aggregation Control Protocol (LACP) can also be used to bind multiple links between two interconnected devices. For example, CISCO's Layer 3 switch implements this function. Configure load balancing policy to implement traffic balancing. The load balancing policy can be implemented based on attributes such as source MAC address, destination MAC address, IP address, transport layer port number, and protocol. The above link implementation method can implement the functions of traffic redundancy and redundancy backup in some cases, but has certain networking restrictions. For example, the load balancing policy is implemented by using the source MAC address and the destination MAC address. In the communication system, if the network element is connected to only one device and the source MAC address and the destination MAC address are unique, the network element device and the mutual device cannot be implemented. Flow between devices A balanced amount. Even if the device is interconnected with multiple devices, the source MAC address and the destination MAC address are still the same through the hash algorithm or the XOR value. For the networking, the traffic distribution cannot be well planned. Link aggregation is implemented, and traffic balancing functions are implemented through attributes such as IP address, transport layer port number, and protocol. First, LACP implementation has certain complexity. Second, the configuration of each network element in the communication system is different, and it is also possible. In the case where the above attributes are unique, the requirements for traffic balancing are still not well met. The existing patent documents include: Patent Application 'CN No. CN200510076709 Chinese Patent Application "Network Equipment for Link Aggregation Method" and Patent No. GB2376149 International Patent "LINK AGGREGATION" (Link Aggregation). The Chinese patent application No. CN200510076709 has the following deficiencies: 1. The patent realizes load sharing through the generation of network layer equivalent routing, which increases the complexity of network layer processing; 2. The patent is applicable to switch equipment, and the network equipment cannot pass the standard. The case where the switch is interconnected with a standard network device. The international patent No. GB 2376149 has the following deficiencies: In this patent, the port is obtained by the hash algorithm of the source address and the destination address of the data packet, and the problem of equalizing the traffic when the communication system is networked cannot be solved. Therefore, a link aggregation method of the local end is required, which can implement traffic balancing and reliability, and has no special requirements for the interconnection device. The function of accessing the IP network can be conveniently completed when the communication system is networked. SUMMARY OF THE INVENTION A primary object of the present invention is to provide an access configuration method and a link aggregation system for link aggregation, which are used to implement traffic balancing of an aggregated link and enhance reliability of link transmission. In order to achieve the above object, according to a first aspect of the present invention, the present invention provides an access configuration method for link aggregation. The access configuration method includes the following steps: Step S102: Set n ports on the access side to a link aggregation group, where each port is assigned a unique MAC address, where n is a positive integer; Step S104: Configure m IP addresses for the link aggregation group, where m is a positive integer. Step S106, bind m IP addresses to n ports, so that the number of IP addresses bound to each port is the same as any other port. The difference between the number of bound IP addresses is less than or equal to 1; and in step S108, each port sends an ARP (Address Resolution Protocol), and the ARP-text includes the IP address and port assigned by the port. Correspondence of MAC addresses. After one or more ports in the port are disconnected, the IP address originally assigned to the interrupt port is floated to other ports that are connected properly, so that the number of IP addresses bound to each valid port is bound to any other valid port. The difference between the number of IP addresses is less than or equal to 1, and then each valid port sends an ARP 4 message, which refreshes the correspondence between the IP address assigned by the port and the MAC address of the port. After the interrupt port returns to normal, the corresponding IP address is floated back to the interrupt port, so that the difference between the number of IP addresses bound to each valid port and the number of IP addresses bound to any other valid port is less than or equal to 1, and then each valid port Sends an ARP packet to refresh the mapping between the IP address assigned to the port and the MAC address of the port. If m is less than or equal to n, then m ports of the n ports are bound with one IP address, and the remaining (n - m) ports are not bound with an IP address; if m is greater than n, then n ports are bound at least An IP address. If m IP addresses belong to different network segments, the interconnected peer interfaces are configured with m and access sides.

IP address corresponding to the network segment of the IP address. The next hop of the route to the destination network segment is the m IP addresses of the peer. The access side balances the traffic with the next hop. Polling is sent. If the m IP addresses belong to the same network segment, the peer end interface is configured with an IP address corresponding to the access side IP address network segment, and the next hop of the access network configured to the destination network segment address route is the peer IP address. Address, the access side polls and sends in the link aggregation group according to the traffic balancing principle when sending packets. In order to enable the access side to receive load sharing, the peer end of the interconnection is configured to multiple next hops of the destination network segment route of the access side, and the multiple next if mega addresses are respectively m IP addresses. The IP address of the same network segment is configured on the link aggregation group. The direct route generated on the same network segment is directed to the link aggregation group. In order to achieve the above i), according to a second aspect of the present invention, the present invention provides a link aggregation system. The link aggregation system includes: a link aggregation group, including at least two links, for performing communication between the interconnection devices, and the link aggregation group is configured with m IP addresses, where m is a positive integer; n ports, located on the access side, are used to set up to link aggregation groups, each port has a unique MAC address, where n is a positive integer. The IP address is bound to the n-ports. The difference between the number of IP addresses bound to each port and the number of IP addresses bound to any other port is less than or equal to 1. Each port sends ARP packets. The ARP packets are included. Correspondence between the IP address assigned by the port and the MAC address of the port. After one or more ports in the port are disconnected, the IP address originally assigned to the interrupt port is floated to other ports that are connected properly, so that the number of IP addresses bound to each valid port is bound to any other valid port. The difference between the number of IP addresses is less than or equal to 1, and then each valid port sends an ARP 4 message, which refreshes the correspondence between the IP address assigned by the port and the MAC address of the port. After the interrupt port is restored, the corresponding IP address is floated back to the interrupt port, and the difference between the number of IP addresses bound to each valid port and the number of IP addresses bound to any other valid port is less than or equal to 1, and then Each valid port sends an ARP packet to refresh the mapping between the IP address assigned by the port and the MAC address of the port. Through the above technical solutions, the present invention achieves the following technical effects. The present invention can achieve traffic balancing on the basis of ensuring reliability. Through the balanced distribution of the IP address and the Layer 2 MAC address mapping, the functions of the links in the link aggregation group are balanced and mutually backed up, which reduces the unnecessary additional processing of the network layer. The interface entity seen by the network layer is only The link aggregation group, its processing does not need to be proud of any special processing. In addition, the method of configuring the IP address of the same network segment on the same link aggregation group saves the IP address configuration of the peer device and simplifies the networking requirements. The same network segment is configured on the same link aggregation group. IP, there is no violation of the routing principle, so there is no additional processing requirement for the configuration. In addition, since the present invention only relates to the implementation of the local end, there is no special requirement for the opposite end, which is advantageous for realizing interconnection of multi-vendor devices. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are set to illustrate,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, In the drawings: FIG. 1 is a flow chart of an access configuration method for link aggregation according to the present invention; FIG. 2 is a schematic diagram of an application according to an embodiment of the present invention; FIG. 3 is a chain diagram according to an embodiment of the present invention. FIG. 4 is a schematic diagram of the local end and the peer learning an ARP entry according to an embodiment of the present invention; FIG. 5 is an IP address after a port connection interruption in a link aggregation group according to an embodiment of the present invention; FIG. 6 is a schematic diagram of an IP address floating process after a connection between two ports in a link aggregation group is interrupted according to an embodiment of the present invention; FIG. 7 is a diagram showing that the number of IP addresses is smaller than that in a link aggregation group according to an embodiment of the present invention; Figure 8 is a diagram showing the IP address floating process after a port in the link aggregation group is interrupted. Figure 9 is a link aggregation group in the case shown in Figure 7. FIG. 10 is a diagram showing an IP address allocation when the number of IP addresses is greater than the number of ports in the link aggregation group according to an embodiment of the present invention; and FIG. 11 is when the local end and the pair are end A schematic diagram of interconnection of different network addresses are respectively disposed. DETAILED DESCRIPTION OF THE INVENTION The present invention will be described in detail below with reference to the accompanying drawings. Referring to FIG. 1, an access configuration method for link aggregation according to the present invention includes the following steps: Step S102: Set n ports on the access side to a link aggregation group, and each port is assigned a MAC only Address, where n is a positive integer; Step S104, configuring m IP addresses for the link aggregation group, where m is a positive integer; Step S106, binding m IP addresses with n ports, so that each port is bound The difference between the number of IP addresses and the number of IP addresses bound to any other port is less than or equal to 1; and in step S108, each port sends an ARP packet, and the ARP 4 message includes the IP address assigned by the port and the MAC address of the port. "It should be related. When one or more ports in the port are disconnected, the IP address originally assigned to the interrupt port is floated to the other port with normal connection, so that the number of IP addresses bound to each valid port is valid. The difference between the number of IP addresses bound to the port is less than or equal to 1. Then, each valid port sends an ARP packet, and the corresponding relationship between the IP address assigned by the port and the MAC address of the port is refreshed. After that, the corresponding IP address is floated back to the interrupt port, so that the difference between the number of IP addresses bound to each valid port and the number of IP addresses bound to any other valid port is less than or equal to 1, and then each valid port sends ARP packets. The relationship between the IP address assigned by the port and the MAC address of the port is refreshed. The steps of the link aggregation method that can implement traffic balancing and reliability provided by the present invention are:

1. Set the n external physical ports of the NE to belong to a link aggregation port group.

2. For the link aggregation port group, configure m IP addresses; the routes related to the link aggregation port group, and the outbound interfaces point to the link aggregation group. 3. Make sure that n ports are connected properly. According to the load sharing principle, the IP addresses of m are bound to n ports. In the case where m is less than or equal to n, where m ports are bound with an IP address, and the remaining (n - m) ports are not bound with an IP address, as a backup of the previous m ports; in the case where m is greater than n The n ports are bound to at least one IP address. It is possible to bind multiple addresses. According to the load sharing principle, the number of IP addresses bound to any port cannot be 2 more than the IP address bound to any other port. 2 or more. Through the above process, at the link layer, each of the m IP addresses is assigned a corresponding MAC address; if some of the n ports are disconnected, in assigning an IP address, only the normal ports of the _n ports need to be considered. .

4. To enable the interconnect switch to learn the correspondence between the port and the MAC address and to obtain the correct IP address and MAC address mapping for the peer device, each physical port needs to send a gratuitous ARP entry to advertise the port to assign the IP address and address. Port MAC correspondence.

5. After the connection of a port in the link aggregation group is interrupted, you need to float the original address assigned to the port to other ports that are connected properly, and send the free ARP as shown in step 4, and refresh. IP address and MAC correspondence.

6. After the port connection in the link aggregation group is interrupted and then resumes normal, you need to float the corresponding address on the port based on the load sharing, and send the free ARP to refresh the IP address as described in step 4. Correspondence with MAC.

The following steps are required to implement the following: a) The local end configures the link aggregation group with m IP addresses on different network segments, and the peer interface connected through the switch needs to be configured with m. The IP address corresponding to the network segment of the local IP address. The next hop of the local route to the destination network segment is the IP address of the peer. The local device sends multiple packets according to the traffic balancing principle. The polling is sent between the two ends. b) The local end configures the IP address of the link aggregation group to belong to the same network segment. The peer interface connected to the switch needs to be configured with an IP address corresponding to the network segment of the local IP address. The next hop of the route from the end to the destination network segment is the IP address of the peer. The local end sends the polling in the link group according to the traffic principle. c) In the case of a), in order to enable the local end to receive the load sharing, the peer end connected through the switch needs to be configured with multiple next hops to the local destination network segment route, and the multiple next hop addresses are respectively the m IP addresses of the local end. Address: d) Similarly, in the case of b), in order to make the local end receive the load sharing, the peer connected through the switch needs to be configured with multiple next hops to the destination network segment of the local end, and the multiple next hop addresses are respectively m IP addresses of the end; e) The traffic balancing of the peer end can be selected according to the need of multiple next hop polling transmission mode or fixed next crypto mode according to the destination address.

8. In order to implement the traffic balancing described in the following, you need to support multiple IP addresses of the same network segment on the same link aggregation group. The direct routes generated on the same network segment must be directed to the same link aggregation group. Compared with the prior art, the technology is more suitable for the occasion that the network element in the communication system is interconnected through the switch and the peer or accesses the 7-carrier network, and the traffic can be balanced on the basis of ensuring reliability. Through the balanced distribution of the IP address and the Layer 2 MAC address mapping, the functions of the links in the link aggregation group are balanced and mutually backed up, which reduces the unnecessary additional processing of the network layer. The interface entity seen by the network layer is only Link aggregation group, its processing does not require any special processing. In addition, the method of configuring the same network segment IP on the same link aggregation group saves the IP address configuration of the peer device and simplifies the networking requirements. There is no violation of the routing principle, so there are no additional processing requirements for the configuration. Further, since the method only involves the implementation of the local end, there is no special requirement for the opposite end, which is beneficial to realize interconnection of multi-vendor devices. The implementation process of the patented method will be specifically described below with reference to an example. Referring to FIG. 2, the local multi-interface is connected through the switch and the peer device as an example. How the peer device is interconnected with the switch is not related to the discussion of this patent method, but for the sake of simplicity, it is considered that there is a high bandwidth link between the peer device and the switch, and only one interface is provided. As shown in Figure 3, the local end is interconnected by four physical links and switches. The four physical links are aggregated into one link aggregation group, which is identified by GroupPort. The ports corresponding to the four physical links are Ethernet1 and Ethernet2. Ethernet3, Ethemet4 standard · ί only, there is one MAC per physical port Address, identified by Macl, Mac2, Mac3, Mac4. Configure the IP address of the link aggregation group, namely, IP1, IP2, IP3, and IP4. Each address belongs to the same network segment. The port that is generated by the direct route is the link aggregation group. Based on the principle of equalized traffic, mapping each LP address to each physical port forms a mapping relationship between IP address and MAC address, namely IP1 Macl, IP2 Mac2, IP3 Mac3, and IP4 Mac4. As shown in Figure 4, the peer device port is configured with the same network segment as the local address. If the physical port of the local end sends gratuitous ARP packets, the switch binds the MAC address to the interconnected port. The peer device also learns the ARP entry of the local device. The ARP entries of the peer device, that is, the mapping between IPa and MACa, are obtained through the ARP learning. If a packet needs to be sent through the local end and forwarded or terminated to the peer device, the local device needs to configure a static route to the destination network segment. The next f Mbps is IPa. The egress port pointed to by the local end is the link. In the link aggregation group, packets are sent in the polling mode according to the polling mode, so that the traffic is distributed on each physical link in the aggregation link group. When packets are sent through the peer device or from the peer device To the local end, you need to configure static routes on the peer device. The next hops are IP1, IP2, IP3, and IP4. The peer device sends packets according to the polling mode. The destination MAC addresses are Macl and Mac2. , Mac3 and Mac4, the traffic is evenly distributed on each physical link in the aggregation link group. Referring to Figure 5, when a port connection in the link aggregation group is interrupted, the Ethemetl connection is interrupted in this figure. Load sharing principle 4B IP address IP1 bound to Ethernet1 floats to Ethernet2, the mapping relationship between IP address and MAC address changes, that is, from IP1 - Macl to IP1 - Mac2, send free ARP through Ethemet2, switch change port and MAC ground The mapping relationship, and the peer device changes the ARP entry, that is, the mapping between IP1 Mac2, IP2 Mac2, IP3 Mac3 and IP4 Mac4 is learned on the peer device. In this way, Ethernet2 will be subject to more traffic load. As shown in Figure 6, when the connection between the two ports in the link aggregation group is interrupted, the connection between Ethemetl and Ethernet2 is interrupted in this figure. According to the load sharing principle, the IP address IP1 bound to Ethemetl is floated to Ethemet3 on Ethernet2. The bound IP address IP2 floats to Ethernet4, and the mapping between IP address and MAC address changes, that is, IP1 - Macl becomes IP 1 - Mac3, from IP2 - Mac2 to IP2 - Mac4, free ARP is sent through Ethernet3 and Ethemet4 The switch changes the mapping between the port and the MAC address, and the peer device changes the ARP entry, that is, the peer device. Learn about the mapping between IP 1 Mac3, IP2 Mac4, IP3 Mac3 and IP4 Mac4. In this way, traffic will be evenly distributed across Ethernet3 and Ethernet4. When the port that is connected to the interrupt is restored, as shown in Figure 5, the Ethernet1 connection is interrupted and then restored to normal. Then, the IP1 that is floating on Ethernet2 is bound to Ethernet1, and the mapping between the IP address and the MAC address is based on the load sharing principle. Change, from IP1 - Mac2 back to IP1 - Macl, send free ARP through Ethernetl, the switch changes the port and MAC address mapping, and the peer device changes the ARP entry, that is, the peer device learns IP1 Macl, IP2 Mac2, Mapping relationship between IP3 Mac3 and IP4 Mac4. Traffic 4 is balanced across physical ports. After multiple connections are interrupted, they return to normal. The processing is similar to the processing after a connection is interrupted. As shown in Figure 7, when the IP address is less than the number of ports, the IP address is bound to some of the physical ports. The excess port has no traffic, but only exists as the port backup used. When the port connection being used is interrupted, The IP address bound to the port is floated to the redundant port to achieve reliability guarantee, and the traffic balance is also considered. The method is similar to the foregoing. Figure 8 depicts the address floating process in the event of a connection interruption in this case; Figure 9 depicts the address floating process when two connections are interrupted in this case. As shown in Figure 10, when the IP address is greater than the number of ports, each IP address is bound to a specific physical port. Some ports have more IP addresses than other ports. In this case, the traffic is not completely balanced. Therefore, in a specific configuration, it is not recommended to configure the IP address to be larger than the number of ports. The above-mentioned load sharing principle is to ensure that multiple IP addresses are balanced and the ports that are connected to each other are bound to each other. That is, the number of IP addresses bound to a port cannot be more than the number of IP addresses bound to any other port. More than two. As shown in Figure 11, if the local end is configured with different network segment addresses IP1, IP2, IP3, and IP4, the peer end configures different network segment addresses IPa, IPb, IPc, and IPd. IP1 and IPa are on the same network segment. IP2 and IPb are on the same network segment. IP3 and IPc are on the same network segment. IP4 and IPd are on the same network segment. The IP address of the local end is configured on the link aggregation group. The generated direct route also points to the link aggregation group. The IP address is bound to each physical port according to the load balancing principle. When a packet needs to be sent through the local end and forwarded or terminated to the peer device, the local device needs to configure a static route to the destination network segment. The next hops are IPa, IPb, IPc, and IPd. The outbound port is the link aggregation group, and the physical port bound to the address directly connected to the local end is found according to the next hop, and the packet is sent out from the physical port. The polling is implemented according to the load sharing principle between the multiple next hops, so that traffic balancing of each port in the link aggregation group is achieved. If a packet needs to be sent to the local device or sent from the peer to the local device, you need to configure a static route on the peer device. The next hops are IP1, IP2, IP3, and IP4. The peer device is polled. Sending packets, the destination MAC addresses are Macl, Mac2, Mac3, and Mac4, and the traffic is evenly distributed on each physical link in the aggregation link group. To simplify the networking requirements, you are advised to configure multiple IP addresses of the same network segment on the same port. Of course, when this function is not implemented, the reliability and traffic balance can be achieved through proper configuration. In summary, the link aggregation method implements the functions of reliability guarantee and traffic balancing in the link aggregation group by binding and floating the IP address to the physical port in the link aggregation group. To simplify network configuration, the local end implements Configure the same network segment address for the same link aggregation group. The method can be interconnected with a standard device, and the reliability guarantee and traffic balancing function in the link aggregation group can be implemented without implementing a complex link aggregation protocol. The above is only the preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any alterations, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Claims

Claim

An access configuration method for link aggregation, which includes the following steps:

Step S102: Set n ports on the access side to a link aggregation group, where each port is assigned a unique MAC address, where n is a positive integer;

Step S104, configuring m IP addresses for the link aggregation group, where m is a positive integer;

Step S106: Bind the m IP addresses to the n ports, so that the difference between the number of IP addresses bound to each port and the number of IP addresses bound to any other port is less than or equal to 1;

Step S108: Each port sends an ARP packet, where the ARP packet includes a correspondence between an IP address allocated by the port and a MAC address of the port.

The access configuration method according to claim 1, wherein after the one or more ports in the port are disconnected, the IP address originally allocated to the interrupt port is floated to other connections. The port is such that the difference between the number of IP addresses bound to each valid port and the number of IP addresses bound to any other valid port is less than or equal to 1, and then each valid port sends an ARP-text to refresh the allocated port. The correspondence between the IP address and the MAC address of the port.

The access configuration method according to claim 2, wherein after the interrupt port returns to normal, the corresponding IP address is floated back to the interrupt port, so that the number of IP addresses bound to each valid port is The difference between the number of IP addresses bound to any other valid port is less than or equal to 1. Then, each valid port sends an ARP packet, and the corresponding relationship between the IP address allocated by the port and the MAC address of the port is refreshed.

The access configuration method according to claim 1, wherein if _{m is} less than or equal to n, each of the n ports is bound with one IP address, and the remaining (n - m) The port is not bound to an IP address. If m is greater than n, the n ports are bound to at least one IP address. The access configuration method according to claim 1, wherein if the m IP addresses belong to different network segments, the interconnected peer interfaces are configured with m corresponding to the access side IP address network segment. The address, the next hop of the route configured on the access side to the destination network segment is the m IP addresses of the peer end, and the access side polls and sends between multiple next hops when sending packets. The access configuration method according to claim 1, wherein if the m IP addresses belong to the same network segment, the peer interface of the interconnection is configured with an IP address corresponding to the network segment of the access side IP address. The next hop of the route configured on the access side to the destination network segment is the IP address of the peer. The access side sends the polling packet in the link aggregation group when sending the packet. The access configuration method according to claim 5 or 6, wherein, in order to enable the access side to receive load sharing, the peer end of the interconnection is configured to multiple next hops of the destination network segment route of the access side, and more The mega addresses are respectively the m IP addresses. The access configuration method according to claim 5 or 6, wherein the link aggregation group is configured with multiple IP addresses of the same network segment, and the direct route generated by the same network segment points to the link. Aggregation group. A link aggregation system, comprising:

The link aggregation group includes at least two links for performing communication transmission between the interconnection devices, where the link aggregation group is configured with m IP addresses, where m is a positive integer;

n ports, located on the access side, are configured to be set to the link aggregation group, each port having a unique MAC address, where n is a positive integer, where

The m IP addresses are bound to the n ports, so that the difference between the number of IP addresses bound to each port and the number of IP addresses bound to any other port is less than or equal to 1; The ARP packet includes a correspondence between an IP address allocated by the port and a MAC address of the port. The link aggregation system according to claim 9, wherein after the one or more ports in the port are disconnected, the IP address originally allocated to the interrupt port is floated to another port that is normally connected. The difference between the number of IP addresses bound to each valid port and the number of IP addresses bound to any other valid port is less than or equal to 1, and then each of the valid The port sends an ARP packet to refresh the mapping between the IP address assigned to the port and the MAC address of the port.

11. The link aggregation system according to claim 10, wherein after the interrupt port returns to normal, the corresponding IP address is floated back to the interrupt port, so that the number of IP addresses bound to each valid port is The difference between the number of IP addresses bound to any other valid port is less than or equal to 1. Then, each valid port sends an ARP packet, and the corresponding relationship between the IP address allocated by the port and the MAC address of the port is refreshed.