Author Topic: Unhappy with usdlls.savvis.net - traceroute included (Read 665 times)

Gator · « **on:** March 09, 2000, 05:41:00 PM »

Did savvis.net change something with their connection between New York and Dallas? Seems like I'm getting slower connects and more lost packets.

Traceroute, 6:40 PM Eastern, 9 March, run over a couple of minutes:

Host Name IP Address Hop Ping Time Ping Avg % Loss Pkts r/s Ping best/worst
fe0-0-0.core1.anp.md.rcn.net 10.65.34.1 2 121ms 107ms 0% 36 / 36 86ms / 193ms
poet0-0-1.core1.col.md.rcn.net 207.172.9.197 3 202ms 176ms 0% 35 / 35 83ms / 399ms
poet6-0-0.core1.blb.md.rcn.net 207.172.19.170 4 93ms 129ms 0% 35 / 35 93ms / 283ms
poet1-0-0.core1.blba.md.rcn.net 207.172.9.53 5 100ms 127ms 0% 35 / 35 94ms / 333ms
poet4-0-1.core1.dcb.dc.rcn.net 207.172.9.49 6 117ms 126ms 0% 35 / 35 91ms / 315ms
pos1-1-0.border1.tcob.va.rcn.net 207.172.19.249 7 113ms 121ms 0% 35 / 35 96ms / 298ms
ge3-0-0.core1.tco.va.rcn.net 207.172.19.213 8 114ms 123ms 0% 35 / 35 92ms / 316ms
fe1-0-0.border2.tco.va.rcn.net 207.172.9.222 9 160ms 125ms 0% 35 / 35 92ms / 259ms
serial2-0-4.hsa1.wdc1.Level3.net 209.244.200.5 10 123ms 123ms 0% 35 / 35 94ms / 248ms
core1.NewYork1.Level3.net 209.244.2.19 11 114ms 122ms 0% 35 / 35 94ms / 266ms
level3-1.CR-1.usnycm.savvis.net 209.83.160.25 12 123ms 115ms 0% 35 / 35 95ms / 185ms
atm9-0-030.CR-1.usdlls.savvis.net 209.83.222.30 13 738ms 740ms 17% 29 / 35 702ms / 872ms
applink-1.usdlls.savvis.net 216.90.2.66 14 813ms 757ms 8% 32 / 35 709ms / 883ms
beta.hitechcreations.com 216.91.192.19 15 744ms 750ms 14% 30 / 35 718ms / 833ms

Gator · « **Reply #1 on:** March 09, 2000, 05:52:00 PM »

BTW, this does not appear very "time dependent". Posted traceroutes from last night taken after 11 PM Eastern in the "Auto trace route on disco" thread, the trend was the same, problems once packets arrived at usdlls.savvis.net.

WB, which I used to have poorer connects to than AH, is currently < 200 ...

Gator · « **Reply #2 on:** March 09, 2000, 07:23:00 PM »

In doing some more checking, the problem is consistent and reproducible, and seems to be caused by "getting on" savvis.net in New York (level3-1.CR-1.usnycm.savvis.net) and then using them (savvis.net) to go to Dallas (atm9-0-030.CR-1.usdlls.savvis.net). If you arrive at Dallas via another route, such as the one posted below (alter.net, UUNET?), then you avoid the bad savvis.net connection and service from New York. Since the problem is within the savvis.net network, seems like they should be able to do something about it, right?

8 159.atm2-0.xr2.dca8.alter.net (152.63.35.202) us.dc.Washington 79.008 64.000 64.001
9 288.atm3-0.tr2.dca8.alter.net (146.188.163.94) us.dc.Washington 79.003 63.008 63.000
10 146.188.138.226 (146.188.138.226) us.il.Springfield 111.007 98.003 99.002
11 198.atm7-0.xr2.dfw4.alter.net (146.188.240.69) us.tx.DFW 108.006 98.004 98.008
12 194.atm8-0-0.gw1.dfw1.alter.net (146.188.240.37) us.tx.DFW 109.008 99.003 100.007
13 savvis-dfw-gw.customer.alter.net (157.130.128.54) us.tx.DFW 109.009 99.000 100.000
14 applink-1.usdlls.savvis.net (216.90.2.66) us.mo.Saint_Louis 112.000 132.000 103.003
15 beta.hitechcreations.com (216.91.192.19) us.tx.Grapevine 106.007 104.007 107.008

When this route was checked, the problem between New York and Dallas on the savvis.net connection was still occurring, so this was not taken during a "window" of good internet traffic for savvis.net in Dallas.

Gator · « **Reply #3 on:** March 09, 2000, 07:38:00 PM »

Okay, last one tonight, I'm off to fly Dawn of Aces, since that's where I've got the good connections. But, I wanted to make a clear comparative post of what I see as the problem.

Both of these were run at the same time, 8:35 PM Eastern, March 9th.

The "good" connection, from Washington, D.C., to HTC, avoiding savvis.net until in Dallas-Ft. Worth:

8 159.atm2-0.xr2.dca8.alter.net (152.63.35.202) us.dc.Washington 75.009 63.002 66.006
9 288.atm3-0.tr2.dca8.alter.net (146.188.163.94) us.dc.Washington 75.002 63.006 62.003
10 146.188.138.226 (146.188.138.226) us.il.Springfield 108.006 98.007 97.007
11 198.atm7-0.xr2.dfw4.alter.net (146.188.240.69) us.tx.DFW 108.003 99.005 98.006
12 194.atm8-0-0.gw1.dfw1.alter.net (146.188.240.37) us.tx.DFW 107.003 101.000 99.008
13 savvis-dfw-gw.customer.alter.net (157.130.128.54) us.tx.DFW 104.004 100.006 101.000
14 applink-1.usdlls.savvis.net (216.90.2.66) us.mo.Saint_Louis 113.009 106.009 104.009
15 beta.hitechcreations.com (216.91.192.19) us.tx.Grapevine 107.003 104.000 106.004

The "bad" connection, also from Washington, D.C., to HTC but using savvis.net out of New York, resulting in poor connections to Dallas:

Host Name IP Address Hop Ping Time Ping Avg % Loss Pkts r/s Ping best/worst
poet5-0-0.core1.dcb.dc.rcn.net 207.172.19.202 6 119ms 122ms 0% 25 / 25 103ms / 210ms
pos1-1-0.border1.tcob.va.rcn.net 207.172.19.249 7 109ms 115ms 0% 25 / 25 100ms / 173ms
ge3-0-0.core1.tco.va.rcn.net 207.172.19.213 8 112ms 117ms 0% 25 / 25 104ms / 262ms
fe1-0-0.border2.tco.va.rcn.net 207.172.9.222 9 112ms 114ms 0% 25 / 25 95ms / 155ms
serial2-0-4.hsa1.wdc1.Level3.net 209.244.200.5 10 125ms 112ms 0% 25 / 25 95ms / 132ms
core1.NewYork1.Level3.net 209.244.2.19 11 104ms 117ms 0% 25 / 25 96ms / 148ms
level3-1.CR-1.usnycm.savvis.net 209.83.160.25 12 119ms 113ms 0% 25 / 25 103ms / 132ms
atm9-0-030.CR-1.usdlls.savvis.net 209.83.222.30 13 716ms 729ms 12% 22 / 25 707ms / 775ms
applink-1.usdlls.savvis.net 216.90.2.66 14 835ms 745ms 16% 21 / 25 712ms / 835ms
beta.hitechcreations.com 216.91.192.19 15 743ms 748ms 12% 22 / 25 717ms / 793ms