Administrator Guide

First Packet Forwarding for Lossless Multicast
Beginning with the Dell Networking OS version version 8.3.1.0, all initial multicast packets are forwarded to receivers to achieve
lossless multicast.
In previous versions, when the Dell Networking system is an RP, all initial packets are dropped until PIM creates an (S,G) entry. When
the system is an RP and a Source DR, these initial packet drops represent a loss of native data, and when the system is an RP only,
the initial packets drops represent a loss of register packets.
Both scenarios might be unacceptable depending on the multicast application. Beginning with the Dell Networking OS versions
noted here, when the system is the RP, and has receivers for a group G, it forwards all initial multicast packets for the group based
on the (*,G) entry rather than discarding them until the (S,G) entry is created, making Dell Networking systems suitable for
applications sensitive to multicast packet loss.
NOTE: When a source begins sending trac, the Source DR forwards the initial packets to the RP as encapsulated
registered packets. These packets are forwarded via the soft path at a maximum rate of 70 packets/second. Incoming
packets beyond this rate are dropped.
Multicast Policies
The Dell Networking OS oers parallel multicast features for IPv4.
IPv4 Multicast Policies
The following sections describe IPv4 multicast policies.
• Limiting the Number of Multicast Routes
• Preventing a Host from Joining a Group
• Rate Limiting IGMP Join Requests
• Preventing a PIM Router from Forming an Adjacency
• Preventing a Source from Registering with the RP
• Preventing a PIM Router from Processing a Join
Limiting the Number of Multicast Routes
When the total number of multicast routes on a system limit is reached, the Dell Networking OS does not process any IGMP or
multicast listener discovery protocol (MLD) joins to PIM — though it still processes leave messages — until the number of entries
decreases below 95% of the limit.
When the limit falls below 95% after hitting the maximum, the system begins relearning route entries through IGMP, MLD, and
MSDP.
• If the limit is increased after it is reached, join subsequent join requests are accepted. In this case, increase the limit by at least
10% for IGMP and MLD to resume.
• If the limit is decreased after it is reached, the system does not clear the existing sessions. Entries are cleared after a timeout
(you may also clear entries using clear ip mroute).
NOTE: The Dell Networking OS waits at least 30 seconds between stopping and starting IGMP join processing. You may
experience this delay when manipulating the limit after it is reached.
When the multicast route limit is reached, the system displays the following:
3w1d13h: %RPM0-P:RP2 %PIM-3-PIM_TIB_LIMIT: PIM TIB limit reached. No new routes will
be learnt until TIB level falls below low watermark.
3w1d13h: %RPM0-P:RP2 %PIM-3-PIM_TIB_LIMIT: PIM TIB below low watermark. Route learning
will begin.
To limit the number of multicast routes, use the following command.
Multicast Features
527