Ip Multicast Layer 3 Switching Overview; Multicast Layer 3 Switching Cache - Cisco 6500 Series Software Configuration Manual

Understanding How IP Multicast Layer 3 Switching Works

IP Multicast Layer 3 Switching Overview

Policy Feature Card 2 (PFC2) provides Layer 3 switching for IP multicast flows using the hardware
replication table and hardware Cisco Express Forwarding (CEF), which uses the forwarding information
base (FIB) and the adjacency table on the PFC2. In systems with Distributed Forwarding Cards (DFCs),
IP multicast flows are Layer 3 switched locally using Multicast Distributed Hardware Switching
(MDHS). MDHS uses local hardware CEF and replication tables on each DFC to perform Layer 3 switching
and rate limiting of reverse path forwarding (RPF) failures locally on each DFC-equipped switching module.
The PFC2 and the DFCs support hardware switching of (*,G) state flows. PFC1, PFC2, and the DFCs
support rate limiting of non-RPF traffic.
Policy Feature Card 1 (PFC1) provides Layer 3 switching of IP multicast flows with Multilayer
Switching (MLS) using the NetFlow and hardware replication tables.
Multicast Layer 3 switching forwards IP multicast data packet flows between IP subnets using advanced
application-specific integrated circuit (ASIC) switching hardware, offloading processor-intensive
multicast forwarding and replication from network routers.
Layer 3 flows that cannot be hardware switched are still forwarded in software by routers. Protocol
Independent Multicast (PIM) is used for route determination.
PFC1, PFC2, and the DFCs all use the Layer 2 multicast forwarding table to determine on which ports
Layer 2 multicast traffic should be forwarded (if any). The multicast forwarding table entries are
populated in conjunction with Internet Group Management Protocol (IGMP) snooping (see
"Configuring IGMP

Multicast Layer 3 Switching Cache

PFC1, PFC2, and the DFCs maintain Layer 3 switching information in one or more hardware tables as
follows:
•
•
In systems with PFC1, the maximum switching cache size is 128K entries and is shared by all Layer 3
switching processes on the switch (such as IP unicast MLS and Internetwork Packet Exchange [IPX]
MLS). However, a cache exceeding 32K entries increases the probability that a flow will not be switched
by the PFC and will get forwarded to the MSFC.
In systems with PFC1 or PFC2, the MSFC updates its multicast routing table and forwards the new
information to the PFC whenever it receives traffic for a new flow. In addition, if an entry in the multicast
routing table on the MSFC ages out, the MSFC deletes the entry and forwards the updated information
to the PFC. In systems with DFCs, flows are populated symmetrically on all DFCs and on PFC2.
Catalyst 6500 Series Switch Cisco IOS Software Configuration Guide—Release 12.1 E
18-2
Snooping").
PFC1 populates the Layer 3 flow as {source IP, IP group, ingress-interface/VLAN} in the NetFlow
cache. It also stores the Layer 3 rewrite information and a pointer to a list of outgoing interfaces
(such as replication entries) for the flow. If a flow does not match these parameters, it is considered
a NetFlow miss and is bridged on the incoming port based on the Layer 2 lookup.
PFC2 and DFC populate the (S,G) or (*,G) flows in the hardware FIB table with the appropriate
masks; for example, (S/32, G/32) and (*/0, G/32). The RPF interface and the adjacency pointer
information is also stored in each entry. The adjacency table contains the rewrite and a pointer to the
replication entries. If a flow matches a FIB entry, the RPF check compares the incoming
interface/VLAN with the entry. A mismatch is an RPF failure, which can be rate limited if this
feature is enabled.
Chapter 18 Configuring IP Multicast Layer 3 Switching
Chapter 21,
78-14099-04