Traceroute uses the inherent behavior of the IP routing process to map out each router that a packet is forwarded through, by sending out a series of probe packets which are intended to expire before reaching their final destination, and capturing the resulting ICMP TTL Exceed messages.
Performing a traceroute from a network in Phoenix to a host based in London reveals intermediary hops in Dallas, Atlanta and Ashburn and the use of MPLS as the probes are routed over Telia's fiber:
ryan@phx10:/home/ryan# traceroute -eA --back 46.101.44.214 traceroute to 46.101.44.214 (46.101.44.214), 30 hops max, 60 byte packets (snip) 3 10.140.2.13 (10.140.2.13) [*] 0.536 ms 0.559 ms 10.140.2.1 (10.140.2.1) [*] 0.506 ms 4 phx-b1-link.telia.net (62.115.42.9) [AS1299] 0.684 ms phx-b1-link.telia.net (80.239.194.109) [AS1299] 0.840 ms phx-b1-link.telia.net (62.115.41.253) [AS1299] 0.867 ms 5 phx-b1-link.telia.net (62.115.42.1) [AS1299] '-4' 0.827 ms phx-b1-link.telia.net (80.239.194.109) [AS1299] '-4' 0.877 ms phx-b1-link.telia.net (62.115.42.9) [AS1299] '-4' 0.812 ms 6 dls-b21-link.telia.net (62.115.135.12) [AS1299] '-5' 23.463 ms dls-b22-link.telia.net (62.115.118.246) [AS1299] <MPLS:L=10850,E=0,S=1,T=1> '-5' 32.692 ms dls-b21-link.telia.net (62.115.137.106) [AS1299] '-5' 23.300 ms 7 dls-b21-link.telia.net (62.115.137.106) [AS1299] '-5' 32.474 ms 32.625 ms atl-b22-link.telia.net (80.91.246.74) [AS1299] '-6' 41.248 ms 8 atl-b22-link.telia.net (80.91.246.74) [AS1299] '-6' 50.239 ms 50.358 ms ash-bb3-link.telia.net (62.115.125.190) [AS1299] '-7' 52.116 ms 9 ldn-bb3-link.telia.net (80.91.246.69) [AS1299] '-8' 129.974 ms atl-b22-link.telia.net (80.91.246.74) [AS1299] '-6' 50.181 ms ldn-b4-link.telia.net (62.115.134.139) [AS1299] 129.160 ms (snip)
I assumed that packets over Level 3 fiber are routed more or less using the same path. In fact, a Google search easily produces older traceroute results that reveal Dallas, Atlanta and Washington as hops between Phoenix and London on Level 3 fiber. However, when performing a traceroute from a Level 3-peered network in Phoenix to a same host in London, I noticed that no intermediary hops show up at all, and no indications of MPLS either:
ryan@phx1:/home/ryan# traceroute -eA --back 46.101.44.214 traceroute to 46.101.44.214 (46.101.44.214), 30 hops max, 60 byte packets (snip) 3 lw-dc4-border2-te0-0-0-1.rtr.liquidweb.com (50.28.96.29) [AS19905/AS53824] 0.478 ms 0.510 ms 0.549 ms 4 5-2-34.bear2.Phoenix1.Level3.net (4.28.83.25) [AS3356] 0.615 ms 0.627 ms 0.655 ms 5 ae-116-3502.edge3.London15.Level3.net (4.69.167.78) [AS3356] '-11' 123.048 ms ae-226-3602.edge3.London15.Level3.net (4.69.167.94) [AS3356] '-11' 123.042 ms ae-227-3603.edge3.London15.Level3.net (4.69.167.98) [AS3356] '-11' 122.996 ms (snip)
Why did no TTL decrement happen when the traceroute probes passed through intermediary nodes on Level 3's fiber?
Did the Level 3 router in Phoenix already label every probe with a complete routing decision to London so that intermediary hops could just forward the probes to London based on these labels instead of sending them to their routers for further routing decisions?
Did the intermediary hops send the probes to their routers, but did those routers not decrement the TTL for some reason?
No comments:
Post a Comment