about Auto-RP

Unanswered Question
Apr 22nd, 2007

Hi folks,

the topology is like this:

S - RP - MA - PIM R - Hosts

I've two questions about Auto-RP PIM mode:

1) from S (sources) to RP, and from RP to MA (mapping agent), I need to use 'sparse-dense' mode, or is enough 'sparse-mode' (without autorp listener)?

2) the best practice with Auto-RP is to use sparse mode with autorp listener, or sparse-dense mode with a "sink" RP for unknown mGroups?

Any advice will be appreciated



I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Harold Ritter Sun, 04/22/2007 - 16:34

1) Unless all the PIM routers are directly connected to the RP, you absolutely need sparse-dense mode (without autorp listener of course).

2)I would tend toward the first option, although the second one is also commonly seen. In all honesty, my personal preference is to use anycast RP.

Hope this helps,

ariela Sun, 04/22/2007 - 17:19

Thanks for your answer,

, if I well understood, in a word, from S to RP, and from RP to MA, I've to use sparse-dense mode. Then, from MA to all PIM routers, I could use also sparse-mode with autorp listener .. correct?

Anycast RP? I normally use Anycast RP for RP redundancy, then use Auto-RP to propagate the information (rp-announce common-IP on RP, then rp-announce-filter on MA with the common-IP only).

Any advice will be appreciated



Harold Ritter Sun, 04/22/2007 - 18:16


This is not I was eluding to. If you use one startegy you should stick to it all the way.

As I stated, I would rather suggest Anycast RP as it allows you to configure the RP statically on all PIM routers and it offers redundancy at the same time.

Hope this helps,


This Discussion