A load sharing method and device

Abstract

The present invention proposes a load sharing method, which is applied to a local switch of a link aggregation group, including: creating a link aggregation symmetric group, and adding a physical member port to the link aggregation symmetric group; wherein the physical member port The traffic between them is isolated from each other; the main port is elected among the physical member ports, and the ARP message sent by the local host is broadcast to the peer IRF device through the master port; the ARP message or ARP message of the peer host is received. The reply message independently learns the MAC address of the peer host through the physical member port that receives the ARP message or the ARP reply message, and forwards the message based on the learned MAC address. The present invention guarantees the symmetry of the traffic of the same session using the physical link in the uplink and downlink, avoids the invalid forwarding of the traffic of the same session inside the IRF device, improves the forwarding efficiency of the device, and reduces the IRF device at the same time The overhead of system MAC resources.

Description

technical field [0001] The present invention relates to the technical field of communications, in particular to a load sharing method and device. Background technique [0002] With the development of cloud computing, the data center network gradually tends to be flat, and the gateway is often installed at the core layer of the network to facilitate the migration of virtual machines; at the same time, IRF (Intelligent Resilient For a large number of applications in the center, distributed aggregation is often used between the IRF device and the uplink device to load balance the data and perform link backup at the same time. For distributed aggregation, due to various reasons such as the load sharing algorithm and connection sequence at both ends of the aggregation link, the session traffic carried by the aggregation link is often distributed on different physical links in the uplink and downlink. . [0003] For example, see figure 1 ,when figure 1 When the upstream host P...

AI Technical Summary

Problems solved by technology

[0003] For example, see figure 1 ,when figure 1 When the upstream host PC1 communicates with the downstream host PC2, the links calculated by the uplink device and the downlink device in the link aggregation group may be different according to the load sharing algorithm. As a result, the packets sent from PC1 to PC2 pass through slot 4 on the IRF device. Invalid forwarding by member devices four times, and when the message is forwarded by the IRF device, the member devices of each hop need to learn the MAC address of the message, which increases the overhead of system MAC resources

[0004] In addition, when the IRF device is a traditional distributed device, there are similar problems

Images