All things related to multicast

["Michael H. Lambert" <>]

Thanks for all the responses I've received.  Now for a little more
explanation on my part--I'm still not sure what's happening here.

Our topology (the rest of our GigaPoP does it "right"--ie, MSDP/MBGP
everywhere; because of layer 8/9 constraints we don't have a true mcast
border router):

Interior mcast router:  Cisco 7507, IOS 12.1(5)T9.  GRE tunnel (.100/30)
to Juniper (which is .101).  Static RP of .102.  Static default mroute
over tunnel.  MSDP peering with .101.  Network behind this router is
128.182.0.0/16.

Intermediate router:  Juniper M40, JunOS 4.4R2.3.  GRE tunnels with
interior router (.100/30) and GigaPoP mcast border router (.84/30).
Static RP of .85 (our end of tunnel).  MBGP with .86.  Static mroute to
128.182.0.0/16 over tunnel.  MSDP peering with .86 and .101.  Direct
routes for tunnels are in inet.2 rib.

Mcast border router:  Cisco 12000, IOS 12.0(15)S1.  In same AS as Juniper.
GRE tunnel with Juniper (.84/30).  MBGP peering with Abilene, Juniper.
Static RP of Abilene interface on box.  MSDP peering with Abilene,
Juniper.  Static mroutes to 128.182.0.0/16, .100/30 over tunnel.

All RPF tests on 12000, Juniper look right to me.  What I see as an oddity
is that on the 12000, "show ip rpf 192.88.115.102" (on the 7500) shows
"static mroute" while RPF checks on the 7500 show "static" rather than
"static mroute".  Just IOS version difference?

An additional item:  (128.182.61.42, 233.2.171.1) shows up on the Abilene
core routers.  It is sourced from behind our 7500.  However, the 7500 is
only seeing (*, 233.2.171.1) entries from internal sources.

Hence, it appears we can send but not receive (not the usual multicast
problem!).

I hope this explains the problem a little better.

Michael