NGN bearer network four NE40 are connected. Each link of two NE40 is interconnected with the same CN2 SR cross-transmission. When one link to CN2 interrupts, BGP route does not switch over slave link. TG is off line about one minute and the service interrupts.
Make ping and traceroute test on NE40. After trace reaches SR router, it starts to drop packet and recovers one minute later.
When NE40 reaches SR router of CN2, it passes transmission device. After the link between NE40 and CN2 is disconnected, NE40 as local line has detected link disconnection and switches the route to local slave link. However, CN2 SR is directly connected to transmission device and cannot sense remote link down. It does not enable route switchover at once. It waits for HOLDTIME. In HOLDTIME, if it does not receive KEEPALIVE packet of the peer, it will regard BGP neighbor DOWN and enable route switchover. There is one minute of TG interruption.
Handling result: Change BGP KEEPALIVE as 3S. Fast convergence of routing protocol reduces route convergence time and link interruption to ensure TG not to be offline.
1. Check the configuration on SR and NE40 and they are not problematic.
2. Check the link and it is found that NE40 and CN2 SR are connected through transmission link and they do not deploy link detection protocol such as BFD.
3. Routing protocol is problematic. Debug and it is found that Keepalive time of both peers is 30S and BGP convergence requires one minute or so.
1. Deploy BFD path detection protocol. But the protocol cannot be deployed because it is not mature.
2. Change BGP KEEPALIVE time. Fast convergence of routing protocol reduces route convergence time and link interruption to ensure TG not to be offline. It is suggested to change BGP KEEPALIVE time of both peers as 3S.