In a previous post, Sparse Mode Multicast Configuration Guide, I mentioned that there is an open standard method of configuring multicast in your network and then a Cisco proprietary method. This post will cover the Cisco proprietary method, autorp. Previously I mentioned that you could specify the RP, or rendezvous point, router. Same goes here. With the open standard method though, you must also have a BSR (bootstrap router). The BSR allows devices to automatically learn who the RP is on the network. With autorp, this is not needed.
For this lab, we will use the same topology that I used in the last example, as shown above.
This will be a basic configuration, to get autorp up and running. There are definitely more advanced options that can be configured, but this will get you up and running with basic functionality.
First off, the thing that needs to be done is to enable multicast routing globally on each router. Very basic and just a single command:
Next, you will need to set IP Pim Sparse Mode on each interfaces you want to receive and or forward the multicast traffic:
R2(config)#int range e0/0-1,lo0 R2(config-if-range)#ip pim sparse-mode
This is where things get different from the open standard, sparse mode config. We will look to R2 and configure it as the RP. Very straight forward and done with a single command:
R2(config)#ip pim send-rp-announce lo0 scope 10
This sets R2 as the RP candidate, and like last time, I am only specifying one candidate. This is configured for a specific interface on the router and is followed by a scope. The scope in the autorp commands is a way to set a TTL value. Be sure to set this large enough to reach the edge of your multicast network.
Next we need to specify a mapping agent. The mapping agent is the device that receives RP mappings from the RP router and then relays that info to the routers in the network that are designated as pim receiving devices. Similar command to the RP command for this one as well:
R2(config)#ip pim send-rp-discovery lo0 scope 10
Lastly, there is an issue here. It is described by so many people as a “chicken and the egg” situation. That is because the mapping agent sends out information via multicast about the RP and groups, but to do this it needs to know the RP. It can’t discover it and send it out if it doesn’t know about the RP already. This can be fixed with a single command as well in our case:
R2(config)#ip pim auto listener
At this point, you have a basic configuration where other routers in the network can begin joining IGMP groups. Just like with the open standard method from my last post. This is a single command:
R1(config)#int e0/0 R1(config-if)#ip igmp join 126.96.36.199
Now if I go to R3 and try my test ping again. I should get a response from R1, since it is joined to that group:
Now since we have covered basic configs of both methods, I will follow this up with one more post of show commands and some debug commands that you can use when configuring a multicast segment in your network. If you had any issues with autorp, somment below and let me know.