Manual

[phone] www.qtech.ru
Москва, Новозаводская ул., 18, стр. 1
265
correctness of unicast route shall be guaranteed above all. If all attempts fail to solve the
problems on PIM-DM, then use debug commands such as debug ipv6 pim, copy DEBUG
information in 3 minutes and send to Technology Service Center.
29.2 PIM-SM6
29.2.1 Introduction to PIM-SM6
PIM-SM6(Protocol Independent Multicast, Sparse Mode) is the IPv6 version of Protocol
Independent Multicast Sparse Mode. It is a multicast routing protocol in sparse mode and
mainly used in large network with group members distributed relatively sparse and wide. It is
no difference from the IPv4 version PIM-SM except the addresses it uses are IPv6 addresses.
Thus we don’t differentiate between PIM-SM and PIM-SM6 in this chapter. All PIM-SM in the
text without specific explanation is IPv6 version PIM-SM. Unlike the Flooding-Prune of Dense
Mode, PIM-SM Protocol assumes no host needs receiving multicast data packets. PIM-SM
router forwards multicast data packets to a host only on definite request.
By setting RP (Rendezvous Point) and BSR (Bootstrap Router), PIM-SM announce multicast
packet to all PIM-SM routers and establish, using Join/Prune message of routers, RPT (RP-
rooted shared tree) based on RP. Consequently the network bandwidth occupied by data
packets and control messages is cut down and the transaction cost of routers is reduced.
Multicast data get to the network segment where the multicast group members are located
along the shared tree flow. When the data traffic reaches a certain amount, multicast data
stream can be switched to source-based SPT (Shortest Path Tree) to shorten network delay.
PIM-SM doesn’t rely on any specific unicast routing protocol but make RPF examination using
existing unicast routing table.
1. PIM-SM Working Principle
The working process of PIM-SM mainly includes neighbor discovery, creation of RPT,
registration of multicast source, SPT switch and so on. The neighbor discovery mechanism is
the same with the mechanism of PIM-DM. We won’t introduce any more.
(1) Creation of RP Shared Tree (RPT)
When a host joins a multicast group G, the leaf router directly connected with the host finds out
through IGMP message that there is a receiver of multicast group G, then it works out the
corresponding Rendezvous Point RP for multicast group G, and send join message to upper
level nodes in RP direction. Every router on the way from the leaf router to RP will create a (*,
G) table item, indicating the message from any source to multicast group G is suitable for this
item. When RP receives the message sent to multicast group G, the message will get to the
leaf router along the established path and then reach the host. In this way, the RPT with RP as
root is created.
(2) Multicast Source Registration