MULTICAST-BSR & RP SELECTION

[netops044]

Can anyone confirm as to how we can configure a HSRP LAN IP as BSR & RP source IP?

[Reza Sharifi]

netops044 wrote:
Can anyone confirm as to how we can configure a HSRP LAN IP as BSR & RP source IP?

The RP and the BSR candidate should be your loopback address not HSRP.

HTH

Reza

[Peter Paluch]

Hi,

The IOS command line does not allow to refer to a specific explicit IP address for BSR and RP candidate (the commands ip pim bsr-candidate and ip pim rp-candidate). They can refer only to a particular interface whose IP address is used for these announcements.

In any case, is there a particular need to use the HSRP address as the address for the BSR and RP? Usually, if a stable IP address is requested for, a loopback interface is used for these purposes. If you need to build a redundant RP, you are probably looking for Anycast RP - in this case, I suggest reading this document on Cisco website.

Perhaps you could enlighten us more about what is your intention.

Best regards,

Peter

[netops044]

Definitely:

I have a network which consists of two routers A & B [primary & secondary] running HSRP on the LAN facing interface. Each router has a WAN link termination from an ISP [primary & secondary links]. The downlink of the routers LAN is connected to a pair of ASA 5540 which are on active-standby mode. Router-A is primary BSR & RP while Router-B is secondary BSR & RP. Unfortunately the ASA supports only static RP. Hence I was trying to make the HSRP IP of the routers LAN interface as the BSR & RP source address and define the HSRP IP as the RP statically on the ASA. This is in theory only. Hence exploring the option of making the HSRP IP as BSR & RP source.

NOTE - My multicast source is behind the ASA firewall

[Reza Sharifi]

Do you have another router on the inside (behind the firewall) to act as your multicast RP?

Are the firewall currently deployed as layer-2 devices only?

[netops044]

No, the firewalls are acting as Layer-3 i.e routed mode. Behind them lies the switching layer to which the multicast sources are connected. The default gateway of the source is the firewall

[Peter Paluch]

Hi,

If you can afford specifying the RP address statically on all your devices (which should not be a problem in a small network) then you can refer to your HSRP's IP address as the RP using the command ip pim rp-address and you do not need to run the BSR mechanism. Is there actually a need for you to run an automatic RP discovery if your ASA boxes support only static RP assignment?

Best regards,

Peter

[netops044]

Nops, i am lookin redundant RPs. Hence cannot have a static RP definition. Moreover network is very huge

[Reza Sharifi]

so if the network is very huge, how come you don't have a couple of routers to terminate all your SVIs (user default gateways) and make them also RPs?

[Peter Paluch]

Hello,

The proper solution for RP redundancy is the Anycast RP I have suggested earlier. The principle is that several RP candidates have the same loopback which is advertised into the IGP routing protocol, and this loopback is also used with BSR-candidate and RP-candidate announcements. In effect, the individual routers in the network will use the RP that is, according to the routing protocol, the closest one. However, the individual RP candidates have to exchange information about existing multicast sources using the MSDP (Multicast Source Discovery Protocol), otherwise, a source that whose multicast stream is PIM-REGISTERed towards a particular RP candidate, may not be known by other RP candidates, thus they cannot send a PIM-JOIN towards it and create a multicast distribution tree branch from the source to the RP candidate.

The Anycast RP is usable both with statically set RP addresses and with BSR/AutoRP mechanisms. Please read this article about Anycast RP and come back with any questions you might have. Once again, using HSRP address for RP redundancy is not appropriate - the HSRP is intended primarily for end stations, not for infrastructure devices redundancy. Infrastructure devices speak routing protocols, thus the redundancy can be achieved by other more appropriate means.

Best regards,

Peter