All things related to multicast

[Bill Fenner <>]

Yup, I've been tracking this off and on.  I have a script that
creates output like:

5591 mroutes, 5925 sa-cache entries.
5393 mroutes with 0 pkts, 198 with pkts
177 (S,G) entries with data flowing and SA cache entry.  These are normal.
21 (S,G) entries with data flowing but no associated
        SA cache entry.  This are probably bad.
327 (S,G) entries with no data flowing and no associated
        SA cache entry.  These may not indicate problems.
682 SA cache entries with no associated (S,G) entry.
        These may not indicate problems.
1449 entries in SA cache with 2 different RP's.  (3 with data)  These are bad.
15 entries in SA cache with 3 different RP's.  (0 with data)  These are bad.
Multi-RP's: 204.147.128.141,206.190.40.61
        128.119.3.241,204.147.128.141
        128.182.13.11,204.147.128.141
        128.223.253.7,204.147.128.141
        128.32.0.224,204.147.128.141
        128.39.0.85,204.147.128.141
        129.250.0.4,204.147.128.141
        130.235.72.187,204.147.128.141
        134.55.34.33,204.147.128.141
        134.68.99.1,204.147.128.141
        137.39.3.111,204.147.128.141
        145.41.1.81,204.147.128.141
        164.107.2.49,204.147.128.141
        171.64.0.71,204.147.128.141
        192.108.196.1,192.108.214.254
        193.10.252.5,193.166.5.252
        193.10.252.5,193.166.5.252,204.147.128.141
        193.10.252.5,204.147.128.141
        198.10.80.1,204.147.128.141
        204.147.128.141,206.229.124.2
        204.147.128.141,207.126.96.4


The MSDP spec seems to say that each (S,G) should only be advertised
by a single RP; Dino tells me this is perfectly appropriate.

I'm worried about the state and control message explosion caused by
multiple RPs advertising the same (S,G).

I'll be presenting some of these observations at the MBONED meeting at
the IETF on Wednesday.

  Bill