I've been trying to deploy a L2/L3 switched based infrastructure that will provide the attached IP multicast fault-tolerant environment.
I have sources on one site multicasting a Video stream over dual Eth/IP ports. Each of these ports is connected to a separate Cisco Catalyst L2 switch (2960) connected via Ethernet transparent WAN links to 2 seperate Cisco L3 switches (3750) at a distant site.
I would like to deploy a fault-tolerant/redundant architecture that avoids loss of video stream in the event of a WAN link failure of Catalyst 2960 failure.
At the destination site, I have deployed a separate vlan for destination devices and run IP multicast routing to forward Video streams to them.
Whether using PIM dense mode or sparse mode, I do not succeed in recovering Video within a few seconds if I fail WAN link #1 or Catalyst 2960 #1. At best, it takes 3minutes to recover and I understand that this is the time it takes for the 3750s to clear the multicast route from its table.
Is there an obvious solution to bypass this issue that I might have missed?
To me the issue is not that of incorrect RP settings (I had auto RP configured on both 3750s using /32 IP addresses on loopback interfaces and it looked Ok), but that of having a fixed DR and a large mroute cache.
Clearly, Catalyst 3750 #1 can re-learn/re-discover Multicast streams via the 3750 interco; the question is how quickly can I get this?
Thanks for any replies.