I’m in Berkeley for 2 days. I’m using my host’s SBC internet connection. Out of curiosity I did a quick traceroute to www.gnu.org (in Boston):
$ traceroute www.gnu.org
traceroute to gnu.org (199.232.41.10), 30 hops max, 40 byte packets
1 10.xx.xx.xx (10.xx.xx.xx) 0.943 ms 0.897 ms 0.851 ms
2 obfuscated.snfc21.sbcglobal.net (xx.xx.xx.xx) 8.286 ms 9.250 ms 8.614 ms
3 dist2-vlan60.snfc21.pbi.net (216.102.187.131) 9.324 ms 8.748 ms 8.856 ms
4 bb2-10g2-0.snfcca.sbcglobal.net (216.102.176.226) 8.360 ms 8.603 ms 9.277 ms
5 bb1-p4-0.snfcca.sbcglobal.net (151.164.190.189) 8.350 ms 8.893 ms 9.321 ms
6 core1-p6-0.crsfca.sbcglobal.net (151.164.243.117) 9.634 ms 9.797 ms 10.065 ms
7 core2-p8-0.crsfca.sbcglobal.net (151.164.242.74) 42.403 ms 150.050 ms 169.998 ms
8 core1-p3-0.crscca.sbcglobal.net (151.164.242.85) 166.216 ms 158.371 ms 199.028 ms
9 core2-p1-0.crscca.sbcglobal.net (151.164.241.234) 10.563 ms 19.842 ms 10.728 ms
10 core1-p11-0.cranca.sbcglobal.net (151.164.242.82) 23.897 ms 30.635 ms 23.132 ms
11 bb1-p1-0.cranca.sbcglobal.net (151.164.40.90) 23.577 ms 24.392 ms 23.095 ms
12 ex1-p15-0.eqlaca.sbcglobal.net (151.164.41.29) 24.640 ms 24.575 ms 24.886 ms
13 lax1-br1-g4-1.gnaps.net (206.223.123.41) 24.355 ms 24.738 ms 24.485 ms
…
Compare with a traceroute from another box in Boston:
$ traceroute www.gnu.org
traceroute to gnu.org (199.232.41.10), 30 hops max, 38 byte packets
1 192.168.xx.xx (192.168.xx.xx) 3.846 ms 1.164 ms 1.208 ms
2 er1.nyc1.speakeasy.net (66.92.70.1) 32.227 ms 24.298 ms 21.421 ms
3 220.ge-0-1-0.cr2.nyc1.speakeasy.net (69.17.83.201) 19.663 ms 20.435 ms 19.566 ms
4 nyiix.ge-2-2-0.gbr1.nyc.nac.net (198.32.160.20) 20.477 ms 19.975 ms 20.885 ms
5 0.ge-3-0-0.gbr1.nwr.nac.net (209.123.11.174) 21.901 ms 21.169 ms 21.166 ms
6 0.so-1-2-0.gbr1.ash.nac.net (209.123.11.18) 31.024 ms 26.838 ms 27.968 ms
7 ash-m20-ge0-0-0.gnaps.net (207.99.39.158) 26.719 ms 26.694 ms 26.339 ms
…
I’ve removed the last part of the traceroute, where it enters the globalnaps network.
The odd thing about the SBC network is the number of hops it requires to go elsewhere. 10 hops within SBC to go from Berkeley to LA? What’s up with that? Doing a traceroute to Europe shows 15 hops within the SBC network, all the way to Amsterdam. I guess I’m spoiled with my (admittedly expensive) Speakeasy connection that takes me from Boston to Amsterdam in 5 hops…
Traceroutes aren’t necessarily a good way to determine route efficiency these days, because it only looks at layer 3 switching. It does nothing to report ATM switching, which can have as much overhead with virtual path rewrites as next-hop information on an IP packet. The Speakeasy path might include just as many ATM switches as you saw hops in California’s SBC routes — you just wouldn’t know it.
-jag