From: ccie2be (ccie2be@nyc.rr.com)
Date: Mon Apr 04 2005 - 17:06:24 GMT-3
Hi guys,
Since stumbling over this mcast issue while doing NMC lab 5, I've been
spending a fair amount of time going over this and similar mcast issues.
Here's a summary of what I've found (or believe):
When 2 or more mcast routers share a lan segment where receivers exist, 3
types of elections are possible:
1. IGMPv2 Querier - To determine which router queries for mcast receivers.
2. Forwarder - To determine which router forwards mcast traffic onto lan
segment.
3. DR - To determine which router sends joins/prunes and/or register
messages to RP
I believe the querier election is independent of whether sparse or dense
mode is used and election outcome has no impact on either SM or DM
operations. If sparse mode is being used, and the igmpv2 querier is a
different router than the DR, the DR will still know when to send
joins/prunes because the DR hears all igmp traffic even if it's not the
router doing the queriering. The router with the LOWEST ip address wins.
Forwarder: This only applies to Dense mode and is determined by AD, then
metric, then highest ip address. If running sparse mode, there is no issue
of which router will forward mcast traffic as it will always be the router
that sends the joins up the tree which in the case of sparse is always the
DR.
DR: Has no effect if running Dense mode and igmpv2 which is the default.
(If running igmpv1 then DR is also the querier but that still has no effect
on dense mode operation.) However, when running Sparse mode, the DR
election can be "fixed" by using the interface command, ip pim dr-priority #
where highest priority wins. If DR-priority is equal, then highest ip
address wins.
Re: ip multicast multipath. I don't know for sure but my guess is that this
command will allow load balancing across multiple equal cost paths by
overriding the default rule of only having 1 rpf interface. I also guess
that it doesn't matter if sparse or dense mode is being used to use this
command.
Please let me know what you think of this summary - especially if you think
I've said something incorrect.
TIA, Tim
This archive was generated by hypermail 2.1.4 : Tue May 03 2005 - 07:54:52 GMT-3