PIM-SM builds a shared multicast distribution tree within each domain, and the RP router is at the root of this shared tree. Although you can physically locate the RP anywhere on the network, it must be as close to the source as possible. Only one active RP router exists for a multicast group.
At the RP router, receivers meet new sources. Sources use the RP to identify themselves to other routers on the network; receivers use the RP to learn about new sources.
The RP performs the following tasks:
registers a source that wants to announce itself and send data to group members
joins a receiver that wants to receive data for the group
forwards data to group
You can configure a set of routers as C-RP routers that serve as backup to the RP router. If an RP fails, all the routers in the domain apply the same algorithm to elect a new RP from the group of C-RP routers. To make sure that the routers use a complete list of C-RP routers, the C-RP router periodically sends unicast advertisement messages to the BSR. The most common implementation is to configure a PIM-SM router as both a C-RP router and a C-BSR.
The switch devices use the hash function defined in the PIM-SM standard to elect the active RP.
You can configure a static entry for an RP router with static RP. This feature avoids the process of selecting an active RP from the list of candidate RPs and dynamically learning about RPs through the BSR mechanism. Static RP-enabled switches cannot learn about RPs through the BSR because the switch loses all dynamically learned BSR information and ignores BSR messages. After you configure static RP entries, the switch adds them to the RP set as if they were learned through the BSR.
Important
In a PIM domain with both static and dynamic RP switches, the static RP switches cannot use a local interface configured as an RP.
When you configure a PIM static RP in a switch, the next hop of the unicast route toward the PIM static RP must be a PIM neighbor. The PIM protocol fails to work, due to a route change, if the next hop toward an already configured static RP becomes a non-PIM neighbor. If a PIM neighbor cannot reach the configured RP, the RP does not activate and its state remains invalid.
A static RP-enabled switch can communicate with switches from other vendors that do not use the BSR mechanism. Some vendors use either early implementations of PIM-SM v1 that do not support the BSR or proprietary mechanisms like the Cisco Auto-RP. For a network to work properly with static RP, you must map all the switches in the network (including switches from other vendors) to the same RP or RPs, if several RPs exist in the network.
To avoid a single point of failure, you can also configure redundant static RPs.
Use the static RP feature when you do not need dynamic learning mode, typically in small networks, or for security reasons, where RPs are forced to devices in the network so that they do not learn other RPs.
Before you can configure a static RP, you must enable PIM-SM and enable static RP.
After you meet these prerequisites, keep in mind the following configuration considerations:
You cannot configure a static RP-enabled switch as a BSR or as a C-RP router.
All dynamically learned BSR information is lost. However, if you disable static RP, the switch loses the static RP information and regains the BSR functionality.
Static RPs do not age, that is, they cannot time out.
Switches do not advertise static RPs, so, if a new PIM neighbor joins the network, it does not know about the static RP unless you configure it with that static RP.
Configure all the switches in the network (including switches from other vendors) to map to the same RP.
In a PIM domain with both static and dynamic RP switches, the static RP switches cannot use a local interfaces configured as an RP.
To avoid a single point of failure, you can configure redundant static RPs for the same group prefix. If you use a mix of vendor switches across the network, you must ensure that all switches and routers use the same active RP because other vendors can use different algorithms to elect the active RP. The switch devices use the hash function defined in the PIM-SM standard to elect the active RP; other vendors can use the lowest IP address to elect the RP.
Important
To reduce convergence times, create only one static RP for each group. The more static RPs you configure for redundancy, the more time PIM requires to rebuild the mroute table and associate RPs.
Static RP configured on the switch is active as long as the switch uses a unicast route to the static RP network. If the switch loses this route, the static RP is invalidated and the hash algorithm remaps all affected groups. If the switch regains this route, the static RP is validated and the hash algorithm remaps the affected groups.