You're transiting two NAPs (MAE-WEST and the DEC exchange (?)) which
both provide service that's essential on one hand (wide access to
the greater Internet) and lousy on the other hand (focussed access
to a particular site).
This whole thread seems to have taken a wrong turn somewhere, or ignored
the fact that this is really goofy routing.
Best (TLG) and BBNplanet are both at mae-west, yet velvet.com packets
traverse Alternet to get to cisco.com. Cisco also appears to have a
direct connection to Alternet which is also at mae-west, yet the packets
traverse BBNplanet.
I would prefer to think there is an 'Oops' somewhere. I don't know why
any provider would purposely do this to their customer's traffic.
==>Best (TLG) and BBNplanet are both at mae-west, yet velvet.com packets
==>traverse Alternet to get to cisco.com. Cisco also appears to have a
==>direct connection to Alternet which is also at mae-west, yet the packets
==>traverse BBNplanet.
This morning, a message went out to the Cisco population that there was a
problem with the circuit to Alternet. This was corrected later this
morning. Below is the correct path taken to cisco when everything is
loaded correctly.
quad:~>traceroute <machine>
traceroute to <machine>, 30 hops max, 40 byte packets
1 best-gw-e0.quadrunner.net (205.166.195.1) 6 ms 5 ms 5 ms
2 b00m.best.net (205.149.160.126) 34 ms 34 ms 34 ms
3 core-ether12-1.mv.best.net (205.149.160.65) 41 ms 38 ms 45 ms
4 905.Hssi3-0.GW1.SCL1.ALTER.NET (137.39.133.89) 40 ms 44 ms 38 ms
5 Fddi0-0.CR1.SCL1.Alter.Net (137.39.19.5) 60 ms 39 ms 39 ms
6 Hssi3-0.San-Jose3.CA.Alter.Net (137.39.100.1) 48 ms 43 ms 61 ms
7 Fddi0-0.San-Jose5.CA.Alter.Net (137.39.27.6) 42 ms 41 ms 41 ms
8 cisco-gw.ALTER.NET (137.39.170.110) 40 ms 46 ms 41 ms
9 sj-wall-2.cisco.com (192.31.7.34) 44 ms 61 ms 46 ms
...deleted...
/cah