Routing through non-addressable space???

Simon Higgs simon at
Sat Jan 26 09:04:39 UTC 2002

At 09:35 PM 1/25/02 -0600, you wrote:

Mmmmmm... this trace is from an @Work/Excite circuit that is about to go 
dark. The "cable modem" is (as far as I can tell) the Cisco 2600 at the 
customer site. Either that or @Work's customer gateway address.

>Cable modems actually "have" two IP's: one rfc1918 that they use for
>provisioning purposes and in the actual IP of the cable modem device,
>one real IP that they pass on to the end-user's device.  Many of the
>newer cable modems are now showing the private IP in traces where
>previously they'd pass and not respond to ICMP.  As for why, who knows?
>My best guess is to ease tracking down offending IP's vs. users.
>Michael Graham
>-----Original Message-----
>From: owner-nanog at [mailto:owner-nanog at] On Behalf Of
>Turpin, Mark
>Sent: Friday, January 25, 2002 8:43 PM
>To: 'Simon Higgs'; 'nanog at'
>Subject: RE: Routing through non-addressable space???
>Routing through a cable interface that has a 10.x address as its primary
>address would be my first guess, esp. due to the 24.x address.
>int cable3/0
>ip address 10.x.x.x. x.x.x.x.x
>ip address public.for.cpe.equip mask.mask.mask.mask secondary
>Mark Turpin
>Charter NOC
>mturpin at
>-----Original Message-----
>From: Simon Higgs [mailto:simon at]
>Sent: Friday, January 25, 2002 8:47 PM
>To: nanog at
>Subject: Routing through non-addressable space???
>Can someone explain this trace to me. How on earth did the
>get inserted into the last hop?
>#    IP address       Host name                                Round
>4 64 ms
>5 71 ms
>6     Unavailable                         79 ms
>7          94 ms
>8                     No response
>9          110 ms
>10                  112 ms
>11        120 ms
>12    Unavailable                         134 ms
>13   Unavailable                         141 ms

Best Regards,



More information about the NANOG mailing list