Method and router for realizing virtual connection supported optimum restart

Abstract

The invention provides a method that virtual connection supports grace restart (GR) and a router. When the active control plate works normally, virtual connection parameter information can be backed up to a standby control plate, after the active / standby changeover happens, the standby control plate can obtain the virtual connection parameter information directly, and the virtual connection interface the state of which is UP and the common interface can take part in the in-phase of link state database (LSDB) and the reconstruction of neighbor relation together, therefore, complete LSDB is obtained; by utilizing the in-phase LSDB, after routing calculation is carried out, the LSDB is distributed to a repeater board, and virtual connection supporting GR is realized. The speed for completing GR is greatly improved, and the probability of GR failure is reduced.

Description

technical field [0001] The invention relates to network communication technology, in particular to a method and a router for realizing a virtual connection supporting graceful restart. Background technique [0002] Open Shortest Path First (OSPF, Open Shortest Path First) and OSPFv3 are link-state-based interior gateway protocols that divide an autonomous system into a backbone area and multiple non-backbone areas, which are identified by area IDs (Area IDs). figure 1 It is a schematic diagram of areas of an autonomous system in the prior art, Area 0 is a backbone area, and the others are non-backbone areas, wherein the router connecting the backbone area and the non-backbone area is called an Area Border Router (ABR, Area Border Router). The backbone area is responsible for the routing between areas. The routing information between non-backbone areas must be forwarded through the backbone area. Therefore, all non-backbone areas must be connected to the backbone area, and th...

AI Technical Summary

Problems solved by technology

Although this method in the prior art can obtain a complete LSDB, since two LSDB synchronizations and two routing calculations are performed, this multi-stage serial operation is not only complicated to implement, but also greatly increases the duration of GR. Time, if the surrounding topology changes during the GR process, it is easy to cause the old routing and forwarding table in the forwarding board to be inconsistent with the new network topology, resulting in forwarding errors; when the duration of GR is too long or even exceeds the preset maximum duration of GR , will cause GR to fail

In addition, since the virtual link can only be started after the LSDB synchronization with the common neighbor is completed and the route calculation is performed, and the hello message can be sent to the neighbor at the peer end of the virtual link to restore the neighbor relationship after the virtual link is started. Therefore, when the common neighbor When the number and number of LSDBs are large, it takes a long time from the active / standby switchover of the router to the start of the virtual link and sending the first hello message to the neighbor at the peer end of the virtual link, which may exceed the neighbor keep-alive time of the virtual link , causing the neighbor router at the opposite end of the virtual connection to disconnect from the neighbor relationship with the router where the active / standby switchover occurred, resulting in GR failure

Images