Multicast - Rendezvous point (RP) Outage

[ashapd]

Hi All,

Question is relating to Multicast - ASM. I understand that initial multicast flow will be using shared tree via (RP is in forwarding path) after that it will be switched to SPT. Questions - - 

* Does RP /down/outage scenario can bring down the whole multicast traffic in the network ?

Thanks !

[paul driver]

Hello
For the initial MC connection, the FHP/LHP rtrs need to be able to reach the RP, then SPT cutsover unless that has been disabled however I would say there still needs to be a pim rp mappng so if this is lost I would say MC traffic would be lost also.

[ashapd]

Thanks for the update. I guess only new multicast flow will be impacted. The existing one should work ?>

[Giuseppe Larosa · Accepted solution]

Hello @ashapd ,

>> Does RP /down/outage scenario can bring down the whole multicast traffic in the network ?

in ASM PIM SM existing multicast groups that have alreay switched to the source based tree SPT from the shared tree are able to keep going on.

New groups are prevented from reaching the RP so new groups are impacted.

Dense mode fallback can be used here in this case new groups will be flooded every where using PM DM.

The natural design solution for avoiding issues is the use of AnyCast RP

multiple nodes impersonate the RP address at the same time using a loopback address (different from the loop0)

MSDP is used between the nodes to exchange info about active sources SA, SA info also tells what (S,G) pairs are active.

if one node dies the IGP in use OSPF or EIGRP or IS-IS will route packets for the Anycast RP address to the nearest node advertising it in the IGP.

Hope to help

Giuseppe