GRE CE to CE for multicast MPLS

[nihit-tandon]

Hi Team,

Can you please let me know the way to enable GRE for running mulicast traffic. CE to CE ,

In this case since tunnel is to be established from CE to CE dnt think any configs needs to be done at the PE.

CE ---> PE---> P---->PE----> CE

Last mile protocol CE----> PE  BGP

[Peter Paluch]

Hello Nihit,

No special configuration on PE routers is required indeed. You should simply create a GRE tunnel between CE routers and activate multicast routing on it.

You will, however, need to modify the multicast routing table for the RPF check to pass - as you are not performing your ordinary unicast routing over the GRE tunnel, multicast packets received over the GRE tunnel would be dropped as they would be received on a different interface (GRE tunnel) than the one pointing back to the source of the stream (the physical interface towards PE) and hence fail the RPF check. On both CE routers, you would need to use the command

ip mroute 0.0.0.0 0.0.0.0 X.X.X.X

where X.X.X.X is the IP address of your PE router (or the next-hop router in general towards the senders on the opposite location of your VPN).

EDIT: this was incorrect. The X.X.X.X should either point to the Tunnel interface itself or to the tunnel IP address on the other end of the tunnel. Thanks goes to Marwan for correcting me!

Best regards,

Peter

[nihit-tandon]

Hello Peter,

Thanks for the revert !!!

Scenario :- I hve 6 locations out of these one location is Hub and remaining all are spoke , so i am planning to run a GRE tunnel from one of the spoke location to the Hub location

In case ill give the  ip mroute 0.0.0.0 0.0.0.0 X.X.X.X at the hub location does it mean that all the traffic will flow from

GRE tunnel only,

Also my PE -CE is BGP do i need to advertise tunnel network also

Regards,

Nihit Tandon

[Marwan ALshawi]

Hello Peter

About the multicast static route I believe it has to be either:

Ip mroute 0.0.0.0 0.0.0.0 x.x.x.x.  Where xxxx is the next hop/end of the gre tunnel on configured on the receiver side

Or

Ip mroute 0.0.0.0 0.0.0.0 tunnelx.

Regards,

[Peter Paluch]

Marwan,

Good catch! I don't know what I was thinking ... You are correct, the next hop with the ip mroute should either be the Tunnel interface itself or the IP address on the other end of the tunnel.

Thank you once again!

Best regards,

Peter

[nihit-tandon]

Hi Peter,

The multicast senders are located at the headquarters i.e at the hub location  only.

Thanks,

nihit

[Peter Paluch]

Hi Nihit,

In that case, the ip mroute 0.0.0.0 0.0.0.0 TunnelX is necessary only on the spoke routers. Do not configure it on your hub router.

Best regards,

Peter

[nihit-tandon]

Thanks Peter,  Let me try and implement

One more query  after configuring the GRE does my hub bandwidth will get dedicated  for a particular spoke site  or only the specific spoke location traffic will flow via GRE tunnel  rest of the other spoke sites traffic will not take via GRE.

Thanks,

nihit

[Peter Paluch]

Hello Nihit,

I am not sure if I understand your query about bandwidth in particular. In any case, with simply creating the GRE tunnels and running PIM-DM or PIM-SM over them, only the multicast traffic will be carried by GRE tunnels. Unicast traffic between your branches will flow natively without going through tunnels.

Best regards,

Peter

[Peter Paluch]

Hello Nihit,

I believe it is not necessary to advertise the tunnel IP addresses into the BGP. You may do it for management purposes so that you can ping the tunnels from each router in your VPN but otherwise, the tunnels in your case are used only as transit links and therefore they are - in my opinion - not interesting as endpoints, hence no reason to have them advertised throughout your VPN.

Allow me a question: where do you expect the multicast senders to be located? Are they going to be located at the headquarters, or can they be present also on spoke locations?

Best regards,

Peter

[Marwan ALshawi]

You welcome and thanks for rating :)