how-to block ads
|
quickcobra
join:2003-05-29
Sacramento, CA
|  Problems w/ Comcast
For the past 5-6 days, I've had pings anywhere from 90ms-600ms to my usual gaming server, which is located in the Dallas, TX area (I am in the Sacramento, CA area). My pings for the past several months were usually in the 40-70ms Range. Here are my findings from a tracert from my computer to the game server:
>tracert 38.113.1.111
Tracing route to ip38-113-1-111.yourhostingaccount.com [38.113.1.111]
over a maximum of 30 hops:
1 36 ms 33 ms 21 ms 10.183.224.1
2 37 ms 24 ms 52 ms 12.244.118.161
3 17 ms 14 ms 15 ms 12.244.73.38
4 15 ms 15 ms 14 ms gbr2-p50.sffca.ip.att.net [12.123.13.62]
5 18 ms 14 ms 32 ms tbr1-p012702.sffca.ip.att.net [12.122.11.69]
6 16 ms 17 ms 15 ms ggr1-p360.sffca.ip.att.net [12.123.13.65]
7 347 ms 379 ms 357 ms p3-0.pr01.sfo01.atlas.psi.net [154.54.11.1]
8 361 ms 355 ms 376 ms p11-0.core02.sfo01.atlas.cogentco.com [154.54.1.
41]
9 361 ms 385 ms 382 ms p15-0.core01.sfo01.atlas.cogentco.com [66.28.4.6
9]
10 442 ms 442 ms 443 ms p14-0.core01.ord01.atlas.cogentco.com [66.28.4.1
86]
11 478 ms 476 ms 478 ms p14-0.core01.bos01.atlas.cogentco.com [66.28.4.1
09]
12 485 ms 491 ms 490 ms g49.ba01.b006523-1.bos01.atlas.cogentco.com [66.
28.6.110]
13 492 ms 500 ms 501 ms Bizland.demarc.cogentco.com [38.112.4.190]
14 497 ms 493 ms 484 ms ip38-113-1-226.yourhostingaccount.com [38.113.1.
226]
15 * * * Request timed out.
16 * * * Request timed out.
17 * * * Request timed out.
18 * * * Request timed out.
19 *
(All times out from there)
So, upon first look, it appears the problem is on the PSI/Cogentco domain...So these are the steps I took to try and solve the problem:
1. Contacted Comcast via a online session. Let representative know of the problem, and how will it be taken care of? Representative did a latency test to MY ip (For the first time), & told me everything is fine on my end. I was told to contact PSI/Cogentco (same company) and let them know of the problem, since the problem is on THEIR end, not Comcast's. Ended session.
2. Contacted Cogentco via phone. Spoke w/ representative who gave me attitude and stated that Comcast is not doing their job correctly, since Comcast is supposed to concact Cogentco-I, as a "subscriber" of Comcast, am not authorized to call Cogentco & file a network issue. Hung up.
3. Contacted Comcast via phone. Spoke w/ representative. Representative stated that he was sorry-I was misinformed by Cogentco employee, who apparently didn't want to do his job and file a network issue. Did a latency test to MY ip (For the 2nd time), & told me everything is fine on my end. He did a tracert to the server ip, and verified the same high latency that I saw once he hit the Cogentco network. Hung up.
4. Contacted Cogentco via phone. Spoke w/ helpful service technician, who did a tracert to my ip, and found that high latency was being encountered on the AT&T domain (which belongs to Comcast, but doesn't REALLY belong to Comcast). He emailed me a copy of the tracert from his end (see below), along with CC'ing the NOC@att.net. He told me I would need to contact Comcast, since it is their domain/hop that is having the problem. Told me good luck, hung up.
5. Contacted Comcast via phone. Spoke w/ helpful service technician, who expressed sorrow at my predicament, since he used to play Quake online, and knows how ping means everything. Representative checked the latency to my computer (3rd time in three calls...think they'd figure out by now that the problem isn't with my computer or routing directly to them?). I explained the 4 previous phone calls to the representative, who basically told me, in so many words, that since the problem was on the AT&T domain, the problem wouldn't be "Looked into", since Comcast "owns" AT&T, but cannot verify problems on the AT&T network at this time. I asked him if the problem was being routed to the necessary technicians so that it could be looked into, he regretfully stated something along the lines of how he was sorry, but he doesn't have the administrative privileges to run this problem to the person who is in charge of networking on the AT&T domain/backbone, and that I should, in so many words, just keep my fingers crossed and hope it gets fixed. Hung up.
Now, does this all sound totally ridiculous or what? Why and how is Comcast able to buy up AT&T, but still not have rights to troubleshoot their backbone?
Anyways, a few minutes later, I received a CC email from the AT&T NOC re: the email that was sent by the Congentco representative.
Does anyone have any suggestions on how to handle this situation? I am around 18,000 feet from the nearest DSL station, so that is pretty much out of the question, and Comcast is about all that is available for high-speed internet in my neighborhood. I guess my only option is, as the Comcast representative put it, to just cross my fingers and wait it out (which i think is not cool at all).
Here is the email I received re: the Congentco representative, from AT&T:
----- Original Message -----
From: Terry, Robert W, NPONS
To: Cogent Support ; blackcobra98
Cc: noc
Sent: Monday, June 02, 2003 7:43 PM
Subject: RE: Traceroute - Latency at 12.122.10.42
We see the latency on the return path. Below is a forward and reverse trace. It looks like the latency is on the San Francisco peering link between Cogent and AT&T. The IP address is 154.54.11.1. Please look at this peering link.
Forward trace to 12.246.224.68.
wswdc01ck#tr 12.246.224.68
Type escape sequence to abort.
Tracing the route to 12-246-224-68.client.attbi.com (12.246.224.68)
1 tbr1-p013201.wswdc.ip.att.net (12.122.11.233) 0 msec
tbr1-p013202.wswdc.ip.att.net (12.123.9.50) 0 msec
tbr1-p013201.wswdc.ip.att.net (12.122.11.233) 0 msec
2 tbr1-cl4.sl9mo.ip.att.net (12.122.10.30) 20 msec
tbr1-p013701.sl9mo.ip.att.net (12.122.10.66) 20 msec
tbr1-cl4.sl9mo.ip.att.net (12.122.10.30) 16 msec
3 tbr1-cl2.sffca.ip.att.net (12.122.10.42) 60 msec 60 msec 60 msec
4 gar1-p370.sc1ca.ip.att.net (12.122.2.245) 60 msec 64 msec 64 msec
5 12.244.73.33 64 msec 64 msec 60 msec
6 12.244.118.162 [AS 6478] 64 msec 64 msec 64 msec
Reverse trace to 66.28.3.1.
sc1ca31ck#tr 66.28.3.1
Type escape sequence to abort.
Tracing the route to f1-1.gw01.hq01.atlas.cogentco.com (66.28.3.1)
1 gbr3-p80.sffca.ip.att.net (12.122.2.246) 4 msec 4 msec 4 msec
2 ggr1-p340.sffca.ip.att.net (12.122.11.218) 4 msec 4 msec 0 msec
3 p3-0.pr01.sfo01.atlas.psi.net (154.54.11.1) [AS 174] 460 msec 460 msec 460 msec
4 p11-0.core02.sfo01.atlas.cogentco.com (154.54.1.41) [AS 174] 468 msec 460 msec 464 msec
5 p14-0.core01.dca01.atlas.cogentco.com (66.28.4.210) [AS 16631] 620 msec 560 msec 536 msec
6 g50.ba01.b005944-0.dca01.atlas.cogentco.com (66.28.6.190) [AS 16631] 532 msec 532 msec 520 msec
7 Cogent.hq01.demarc.cogentco.com (66.28.30.110) [AS 16631] 504 msec 500 msec 500 msec
Thanks,
Robert W. Terry
AT&T Common Backbone
Sorry this is so long guys, but I figured I'd put everything on the table. Thanks for your suggestions!
BTW, yes, I have checked router settings, yes, I have power cycled my modem, & yes, I have tried different cabling from the modem to the computer.
[text was edited by author 2003-06-03 00:22:56]
[text was edited by author 2003-06-03 02:29:31] | |
quickcobra
join:2003-05-29
Sacramento, CA
| Here is the email from Cogentco support re: Phone call #4.
-----Original Message-----
From: Cogent Support
Sent: Monday, June 02, 2003 9:29 PM
To: blackcobra98
Cc: 'noc'
Subject: Traceroute - Latency at 12.122.10.42
Hello,
We have a customer that is reporting latency in your network. Please investigate the issue for us and respond with your findings.
C:\>tracert 12.246.224.68
Tracing route to 12-246-224-68.client.attbi.com [12.246.224.68]
over a maximum of 30 hops:
1 10 ms 10 ms 10 ms gw-out.corp.cogentco.com [192.168.168.10]
2 10 ms 10 ms 10 ms f1-1.gw01.hq01.atlas.cogentco.com [66.28.3.1]
3 10 ms 10 ms 10 ms f1.ba01.b005944-0.dca01.atlas.cogentco.com [66.2
8.30.109]
4 10 ms 10 ms 10 ms g0-7.core01.dca01.atlas.cogentco.com [66.28.6.18
9]
5 10 ms 10 ms 10 ms p15-0.core02.dca01.atlas.cogentco.com [66.28.4.2
2]
6 10 ms 10 ms 10 ms p6-0.pr01.dca01.atlas.psi.net [154.54.1.62]
7 10 ms 10 ms 10 ms att.dca01.atlas.psi.net [154.54.10.158]
8 10 ms 10 ms 10 ms tbr1-p013202.wswdc.ip.att.net [12.123.9.50]
9 30 ms 30 ms 30 ms tbr1-cl4.sl9mo.ip.att.net [12.122.10.30]
10 490 ms 481 ms 491 ms tbr1-cl2.sffca.ip.att.net [12.122.10.42]
11 481 ms 491 ms 491 ms gar1-p370.sc1ca.ip.att.net [12.122.2.245]
12 501 ms 501 ms 501 ms 12.244.73.33
13 521 ms 500 ms 501 ms 12.244.118.162
14 * *
[text was edited by author 2003-06-03 02:28:17] | |
CableUZR
Cuidado, Hay Llamas
join:2003-02-04
Mount Holly, NJ
clubs:
| reply to quickcobra
said by quickcobra  :
{snip}
Now, does this all sound totally ridiculous or what? Why and how is Comcast able to buy up AT&T, but still not have rights to troubleshoot their backbone?
{snip}
Actually, yes, it is a bit ridiculous, but only a bit. Comcast did not buy the AT&T backbone, they bought AT&T Broadband Internet (ATTBI). That's why Comcast has no more control over the ATT backbone than they have over any other ISP's devices (like Level3, Sprint etc.).
Anyway, the Comcast support team should be the ones to keep the pressure on ATT here, because they're paying them buckets of ca$h for that connectivity. You should try to get your issue escalated to Tier 3 support.
Good luck,
Uzr | |
Kip patterson
Premium
join:2000-10-23
Columbus, OH
| reply to quickcobra
Short answer: Forget it, you have no hope.
Long answer: Cogentco is bankrupt, although they have not as yet filed for bankruptcy so far as I know. They built a network by buying up other companies at fire sale prices and by selling internet connectivity according to a model which has been proven not to work by Allied Riser (now part of Cogentco) and others. Their problems got worse in January or thereabouts when ATDN terminated their free peering arrangements because the traffic ratio was in excess of 3 to one in Cogentco's favor. Ever since then, all the Cogentco peering points (including PSI, now owned by Cogentco) have been greatly overloaded. They sell 100mb of connectivity for $1000 per month which attracts numerous vendors of naked ladies and gaming companies at a price which is below cost when you have to pay transit to get your traffic accepted by others.
It got worse again in March when they violated their credit agreement with Cisco, making the entire amount due to Cisco callable. They are currently renegotiating the credit agreement. (My dates are approximate)
Their customers are bailing left and right because of the same kind of problems you are experiencing.
Forget it, you have no hope.
[text was edited by author 2003-06-03 09:05:04] | |
quickcobra
join:2003-05-29
Sacramento, CA
| I've received this information from the gaming server's administrator, who confirms that the problem lies with Congent, and that the problem is known, but doesn't give an ETA. So it appears that the problem isn't really on Comcast/AT&T, although it will require their cooperation.
The message went as follows:
There is a known latency issue with AT&T peering. Cogent is aware of the situation and is currently working to resolve it. I assume it will be fixed in the coming days. I am also assured that this is only a recent problem and is to be fixed soon. Although Cogent isnt looked upon as the best provider, the quality of its network in and around the Dallas area hasnt been questioned previously.
We have looked into other bandwith providers however many quote triple the price, this would severly damage us and the cheap prices that give us a strong advantage over our competitors.
Thanks for your replies, all in all, this has definitely been a learning experience for me, and I've also learned how shoddy Congentco/PSI is by doing research over the internet. I'd stay far away from that company if they were providing high-speed internet in my neighborhood. | |
newview
Ex .. Ex .. Exactly
Premium
join:2001-10-01
Parsonsburg, MD
| reply to Kip patterson
said by Kip patterson :
Ever since then, all the Cogentco peering points (including PSI, now owned by Cogentco) have been greatly overloaded. They sell 100mb of connectivity for $1000 per month which attracts numerous vendors of naked ladies and gaming companies at a price which is below cost when you have to pay transit to get your traffic accepted by others.
Another reason PSI.net took a nosedive is the disclosure back in 2000 of their spam support when they owned up to a "pink contract" with spammer Cajunnet. I imagine this revelation resulted in their landing in numerous blacklists.
»www.theregister.co.uk/content/ar···545.html
--
The Rules of Spam | Maryland's New Anti-Spam Law
Where are we going? And what's with the hand basket? | |
Kip patterson
Premium
join:2000-10-23
Columbus, OH | Thanks. That's new info for me. That may be part of what put PSInet on the market at distressed prices. | |
|
