To optimize the flow of traffic down the shared trees in a network that uses a BSR to dynamically advertise candidate RPs, consider the hash function. The BSR uses the hash function to assign multicast group addresses to each C-RP.
The BSR distributes the hash mask used to compute the RP assignment. For example, if two RPs are candidates for the range 239.0.0.0 through 239.0.0.127, and the hash mask is 255.255.255.252, that range of addresses is divided into groups of four consecutive addresses and assigned to one or the other C-RP.
The following figure illustrates a suboptimal design where Router A sends traffic to a group address assigned to RP D. Router B sends traffic assigned to RP C. RP C and RP D serve as backups for each other for those group addresses. To distribute traffic, it is desirable that traffic from Router A use RP C and that traffic from Router B use RP D.
While still providing redundancy in the case of an RP failure, you can ensure that the optimal shared tree is used by using the following methods.
Use the hash algorithm to proactively plan the group-address-to-RP assignment.
Use this information to select the multicast group address for each multicast sender on the network and to ensure optimal traffic flows. This method is helpful for modeling more complex redundancy and failure scenarios, where each group address has three or more C-RPs.
Allow the hash algorithm to assign the blocks of addresses on the network, and then view the results using the command show ip pim active-rp .
Use the command output to assign multicast group addresses to senders that are located near the indicated RP. The limitation to this approach is that while you can easily determine the current RP for a group address, the backup RP is not shown. If more than one backup for a group address exists, the secondary RP is not obvious. In this case, use the hash algorithm to reveal which of the remaining C-RPs take over for a particular group address in the event of primary RP failure.
The hash algorithm works as follows:
For each C-RP router with matching group address ranges, a hash value is calculated according to the formula:
Hash value [G, M, C(i)] = {1 103 515 245 * [(1 103 515245 * (G&M) +12 345) XOR C(i)] + 12 345} mod 2^31
The hash value is a function of the group address (G), the hash mask (M), and the IP address of the C-RP C(i). The expression (G&M) guarantees that blocks of group addresses hash to the same value for each C-RP, and that the size of the block is determined by the hash mask.
For example, if the hash mask is 255.255.255.248, the group addresses 239.0.0.0 through 239.0.0.7 yield the same hash value for a given C-RP. Thus, the block of eight addresses are assigned to the same RP.
The C-RP with the highest resulting hash value is chosen as the RP for the group. In the event of a tie, the C-RP with the highest IP address is chosen.
This algorithm runs independently on all PIM-SM routers so that every router has a consistent view of the group-to-RP mappings.