All things related to multicast

[Tom Pusateri <>]

It would be interesting to see where along the path this happens.
Maybe vbns could try this internally and see if it is repeatable.

65 seconds seems short for any PIM timeouts though.
I'll be happy to help if I can.

Tom

In message 
<>
 you write:
>Hello;
>
>We have noticed an odd pattern of packet loss to nysernet (AS 3754) from 
>Multi
>castTech
>(AS 16517). Although the packet loss is not high over all (~ 2.5%), there is 
>a
> period
>of about 1.5 seconds every 65 seconds where the packet loss is near total.
>
>http://www.multicasttech.com/rr/nysernet.wednesday.ps
>
>shows this with 2 second averages. The repeat period is 65 seconds to pretty 
>h
>igh accuracy
>(65.0056 seconds averaged over a day), so
>this must be some sort of periodic router reset/refresh/etc. This has 
>apparent
>ly repeated this
>way since last Friday (when our data begin).
>
>Now, we also have traces to U Oregon during the same period, and these do 
>NOT 
>show the same
>loss pattern, so the problem is not at our end. However, as the Oregon path 
>go
>es from
>us to Sprint and the Nysernet path goes through vBNS, this does not narrow 
>thi
>ngs down
>much.
>
>Here are traces from Nysernet to us :
>
>   Mtrace from 63.105.122.14 to 199.109.32.33 via group 233.64.133.20
>   Querying full reverse path...
>     0 adelie.nysernet.org (199.109.32.33)
>    -1 c7206-1.nysernet.org (199.109.32.254) PIM/BGP4+ thresh^ 0 Reached 
> RP/Co
>re
>    -2 ? (199.109.2.18) PIM thresh^ 1
>    -3 clev-ipls.abilene.ucaid.edu (199.109.2.2) PIM/BGP4+ thresh^ 0 Reached 
> R
>P/Core
>    -4 ipls-clev.abilene.ucaid.edu (198.32.8.25) PIM/BGP4+ thresh^ 0 Reached 
> R
>P/Core
>    -5 jn1-at1-1-0-2008.dng.vbns.net (204.147.128.13) PIM thresh^ 1
>    -6 jn1-so7-0-0-0.nor.vbns.net (204.147.136.147) PIM thresh^ 1
>    -7 jn1-so7-0-0-0.wor.vbns.net (204.147.136.133) PIM thresh^ 1
>    -8 jn1-so7-0-0-0.wae.vbns.net (204.147.136.136) PIM thresh^ 1
>    -9 Multicast-inc.wae.vbns.net (204.147.129.90) PIM thresh^ 0 Reached 
> RP/Co
>re
>   -10 hendrix.multicasttech.com (63.105.122.14)
>   Round trip time 293 ms; total ttl of 9 required.
>
>   traceroute to 63.105.122.14 (63.105.122.14), 30 hops max, 40 byte packets
>    1 c7206-1.nysernet.org (199.109.32.254) 1 ms 54 ms 1 ms
>    2 199.109.2.18 (199.109.2.18) 52 ms 5 ms 5 ms
>    3 199.109.5.54 (199.109.5.54) 58 ms 14 ms 14 ms
>    4 199.109.5.2 (199.109.5.2) 32 ms 15 ms 14 ms
>    5 wash-nycm.abilene.ucaid.edu (198.32.8.45) 19 ms 19 ms 19 ms
>    6 vbns-abilene.abilene.ucaid.edu (198.32.11.10) 24 ms 23 ms 24 ms
>    7 Multicast-inc.wae.vbns.net (204.147.129.90) 35 ms 35 ms 40 ms
>    8 hendrix.multicasttech.com (63.105.122.14) 34 ms 36 ms 37 ms
>
>
>Anyone out there know why 65 seconds is a magic number ? Anyone have any 
>ideas
> ?
>
>                                   
>                                   Marshall Eubanks (with Hans Kuhn and Bill 
> O
>wens)
>
>
>    Multicast Technologies, Inc.
>   10301 Democracy Lane, Suite 410
>   Fairfax, Virginia 22030
>   Phone : 703-293-9624          Fax     : 703-293-9609     
>   e-mail : 
> 
>      http://www.on-the-i.com
>