ykram
join:2008-07-16
| [Speed] Comcast Chicago Routing
For the life of me I can't figure this out. Just recently, maybe 2-3 weeks up to now I'm getting unusual latency for servers in my area. For instance, I'm in Chicago and when I visit Chicago game servers, I'm pinging 30-40 ms when it used to be 15-20 ms. Quite unusual..
I'm located in northern Illinois, specifically Gurnee, Illinois. The problem appears to be some bad hops for northern Illinois, anybody else experiencing this? It's constant in that it happens all day, every day and never lets up, always 30-40 ms. I'm using a Motorola Surfboard SB5101:
Software Version: SB5101-2.4.1.6-SCM02-NOSH
Hardware Version: 1
MIB Version: II
GUI Version: 1.0
Our network itself is setup primarily using a Linux router running Linux kernel 2.6 to a linksys 8 port switch which has 3 computers connected to it and a T-Mobile@Home Linksys router which is connected to the switch. The only PC to that router is a separate PC which just browses the web; that pc itself fetches it's IP using DHCP from the T-Mobile@Home router while every other PC uses a static IP, however thats not the cause. I know for sure it's not the cause since I have remote ssh access to a friend's computer in Mount Prospect, Illinois. Here is an example traceroute to a specific Chicago game server which I have the issue to:
% traceroute 8.9.16.3
traceroute to 8.9.16.3 (8.9.16.3), 30 hops max, 38 byte packets
1 * * *
2 ge-1-41-ur01.rollingmdws.il.chicago.comcast.net (68.86.118.113) 9.104 ms 10.764 ms 9.381 ms
3 te-8-3-ur02.rollingmdws.il.chicago.comcast.net (68.87.230.74) 19.162 ms 9.145 ms 9.411 ms
4 te-8-1-ur02.mtprospect.il.chicago.comcast.net (68.87.230.77) 10.326 ms 10.994 ms 9.592 ms
5 te-3-2-ar02.area4.il.chicago.comcast.net (68.87.231.69) 11.800 ms 16.734 ms 11.601 ms
6 pos-0-13-0-0-cr01.chicago.il.ibone.comcast.net (68.86.90.53) 35.002 ms 31.909 ms 36.159 ms
7 xe-10-0-0.edge1.Chicago2.Level3.net (4.71.248.29) 33.113 ms 35.560 ms 33.600 ms
8 ge-7-2.ipcolo1.Chicago1.Level3.net (4.68.101.170) 36.196 ms ge-9-2.ipcolo1.Chicago1.Level3.net (4.68.101.138) 37.141 ms ge-9-0.ipcolo1.Chicago1.Level3.net (4.68.101.10) 34.532 ms
9 8.9.16.3 (8.9.16.3) 36.274 ms 39.168 ms 32.549 ms
Furthermore, here are ping results from his PC:
PING 8.9.16.3 (8.9.16.3) 56(84) bytes of data.
64 bytes from 8.9.16.3: icmp_seq=0 ttl=53 time=33.5 ms
64 bytes from 8.9.16.3: icmp_seq=1 ttl=53 time=31.8 ms
64 bytes from 8.9.16.3: icmp_seq=2 ttl=53 time=32.9 ms
64 bytes from 8.9.16.3: icmp_seq=3 ttl=53 time=33.8 ms
And here are my traceroute and ping results:
traceroute to 8.9.16.3 (8.9.16.3), 30 hops max, 40 byte packets
1 router (192.168.1.1) 0.609 ms 0.117 ms 0.105 ms
2 * * *
3 ge-1-43-ur01.mchenry.il.chicago.comcast.net (68.85.208.1) 9.165 ms 8.161 ms 9.499 ms
4 te-8-4-ur02.mchenry.il.chicago.comcast.net (68.87.231.242) 9.651 ms 8.280 ms 11.626 ms
5 te-3-1-ur01.nchicago.il.chicago.comcast.net (68.87.230.34) 10.570 ms 9.056 ms 9.678 ms
6 te-8-4-ur02.nchicago.il.chicago.comcast.net (68.87.230.226) 10.705 ms 11.049 ms 9.476 ms
7 te-2-1-ar02.area4.il.chicago.comcast.net (68.87.230.41) 12.013 ms 11.295 ms 11.910 ms
8 pos-0-6-0-0-cr01.chicago.il.ibone.comcast.net (68.86.90.49) 33.877 ms 33.730 ms 33.492 ms
9 xe-9-3-0.edge1.Chicago2.Level3.net (4.71.248.21) 33.123 ms 34.054 ms 34.665 ms
10 ge-7-0.ipcolo1.Chicago1.Level3.net (4.68.101.42) 36.555 ms ge-7-1.ipcolo1.Chicago1.Level3.net (4.68.101.106) 35.428 ms ge-7-0.ipcolo1.Chicago1.Level3.net (4.68.101.42) 35.619 ms
11 8.9.16.3 (8.9.16.3) 33.047 ms 34.772 ms 33.871 ms
PING 8.9.16.3 (8.9.16.3) 56(84) bytes of data.
64 bytes from 8.9.16.3: icmp_seq=1 ttl=51 time=33.5 ms
64 bytes from 8.9.16.3: icmp_seq=2 ttl=51 time=34.4 ms
64 bytes from 8.9.16.3: icmp_seq=3 ttl=51 time=33.6 ms
64 bytes from 8.9.16.3: icmp_seq=4 ttl=51 time=35.1 ms
--- 8.9.16.3 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3008ms
rtt min/avg/max/mdev = 33.593/34.210/35.149/0.647 ms
As you can see in both our results, it appears that *-crc01.chicago.il.comcast.net is where the particular lag starts. Is anybody else from Northern Illinois seeing this type of peculiar behavior? I wouldn't normally be concerned with 30-40 ms ping differences but when you pay $70/mo and happen to be a competitive gamer, it tends to get extremely frustrating. |
|
mgmgmg
join:2002-09-08
Minneapolis, MN | same problem here in minneapolis to chicago. ping 100+ should be 15-20. Im routing to san fran? then chicago? Hope it clears up. |
|
NormanS
Premium,MVM
join:2001-02-14
San Jose, CA | reply to ykram
It looks like the increase happens in the transition out of CRAN to the 'ibone'.
--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum |
|
tommyhu
@comcast.net
| reply to mgmgmg
you have to call and complain. I spent over 30mins on the phone, probably have of that was on hold, asking why I'm being routed from minneapolis-omaha-denver and then to chicago instead of just going straight to chicago.
I got the usual run around with "did you restart your modem" and "go to speedtest.com" and tell me your download speed and upload speed so I can tell you everything seems normal. Finally the lady seem to understand my question, which is: Why am I being routed to Omaha and Denver when trying to connect from minneapolis to chicago. She put me on hold to get a "more technical answer" (funny because she never gave me any sort of answer to my question before hand) and came back with this: "That is how our engineers have it setup to meet our business needs" which I responded by saying "at the expense of your customers"
This is a trace to a chicago based server. I wish I has some from a week ago because just recently(within the past 4-5 days) I've been getting these higher that normal pings, on avg about 50ms higher
Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.
C:\Documents and Settings\Administrator>tracert 72.37.231.24
Tracing route to 72.37.231.24 over a maximum of 30 hops
1 1 ms 1 ms 1 ms 192.168.2.1
2 * * * Request timed out.
3 8 ms 8 ms 9 ms ge-9-2-ur01.brooklynpark.mn.minn.comcast.net [68
.85.165.169]
4 8 ms 10 ms 12 ms te-0-1-0-2-ar03.roseville.mn.minn.comcast.net [6
8.87.174.213]
5 8 ms 26 ms 7 ms te-1-4-ar02.roseville.mn.minn.comcast.net [68.87
.174.214]
6 29 ms 28 ms 29 ms te-0-0-0-5-cr01.omaha.ne.ibone.comcast.net [68.8
6.91.153]
7 28 ms 38 ms 27 ms te-4-4.car2.Denver1.Level3.net [4.79.82.57]
8 28 ms 36 ms 36 ms ae-31-55.ebr1.Denver1.Level3.net [4.68.107.158]
9 28 ms 35 ms 36 ms ae-1-100.ebr2.Denver1.Level3.net [4.69.132.38]
10 57 ms 53 ms 53 ms ae-3.ebr3.Chicago1.Level3.net [4.69.132.62]
11 52 ms 52 ms 54 ms ae-78.ebr2.Chicago1.Level3.net [4.69.134.61]
12 68 ms 51 ms 52 ms ge-6-0-0-56.edge2.Chicago1.Level3.net [4.68.101.
174]
13 82 ms 89 ms 89 ms ge0-5.cr01.ord01.mzima.net [4.79.210.2]
14 81 ms 82 ms 91 ms ge0-ubiquity.cust.ord01.mzima.net [72.37.148.158
]
15 84 ms 81 ms 82 ms 72.37.231.24
Trace complete.
I remember doing "tracert"s a few months ago and wondering why I was being routed from Minneapolis to Denver to Chicago which caused a slightly higher, but noticeable, ping, but now that it's going from Minneapolis-Omaha-Denver-Chicago I'm noticing like a 50ping difference. From pinging in the teens to low 20's. Now i'm up to 60's and 70's on chicago servers. |
|
NormanS
Premium,MVM
join:2001-02-14
San Jose, CA
 ·Pacific Bell - SBC
| There should be negligible difference between Minneapolis to Denver, and Minneapolis to Omaha to Denver.
As for why Minneapolis to Denver to Chicago; that likely depends on the kind of agreement that Comcast gets with Level 3. AFAIK, Level 3 could offer a better price to Comcast for sending traffic to connect to the Level 3 transit hub in Denver than to connect to the Level 3 transit hub in Chicago.
Comcast is primarily in business to satisfy their shareholders, and secondarily, to satisfy the 85% of their non-tech-oriented customers; the ones more interested in what they pay for Internet than in what route their packets take between the premises and the site they are going to.
--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum |
|
espaeth
Digital Plumber
Premium,MVM
join:2001-04-21
Minneapolis, MN
 ·voip.ms
·Vitelity VOIP
·Callcentric
·VoiceStick
·ViaTalk
 ·Comcast
·Embarq
| reply to tommyhu
said by tommyhu :
you have to call and complain. I spent over 30mins on the phone, probably have of that was on hold, asking why I'm being routed from minneapolis-omaha-denver and then to chicago instead of just going straight to chicago. None of the MN netblocks appear to be routing direct to Chicago anymore, which would tend to suggest there is a fiber cut or equipment issue affecting direct connectivity from Chicago to Minneapolis so the alternate path through Denver is being chosen. For example, FDCServers is directly connected to Comcast in Chicago, yet they still route through Omaha out and Denver back.
|
|
ykram
join:2008-07-16
1 edit | reply to ykram
So theres been some developments on this. It seems today that the tech who installed the line a few weeks ago (since we moved) installed it too low and a garbage truck snagged the line today, pulling the cable in half leaving us without internet. Of course Comcast is delaying sending a tech out even though I made it perfectly clear how essential it is that we have availability (for bills, school work, voip, etc). Regardless, I went out earlier and rigged the cable up temporarily till the tech would come. The rig up was essentially easy since the wire was just split, my buddy and I went and temporarily connected them so I could finish my work for today.
I took the opportunity however to change the mac address on the NIC that the modem connects to and voila! I'm getting *NORMAL* pings to that Chicago server. But for how long? Not long at all I imagine. Tech comes this week to replace the line so hopefully we won't have this issue again but it's rather interesting to see that the high latency to Chicago servers from my Chicago home could be solved, possibly temporarily, by spoofing the MAC and forcing a new IP. Perhaps somebody else would care to try and see if they receive the same results?
Here's a sample traceroute showing the routing as of 2:05 PM (CST).
1 router (192.168.1.1) 0.199 ms 0.841 ms 0.137 ms
2 * * *
3 ge-1-43-ur01.mchenry.il.chicago.comcast.net (68.85.208.1) 9.816 ms 10.660 ms 9.418 ms
4 te-3-2-ur02.algonquin.il.chicago.comcast.net (68.87.231.237) 10.430 ms 10.835 ms 10.071 ms
5 te-8-3-ur01.algonquin.il.chicago.comcast.net (68.87.230.21) 9.277 ms 11.548 ms 10.891 ms
6 ge-1-3-ar01.mishawaka.in.sbend.comcast.net (68.87.230.234) 11.829 ms 13.050 ms 11.776 ms
7 ge-1-1-ar03.area4.il.chicago.comcast.net (68.87.230.233) 13.068 ms 11.322 ms 11.811 ms
8 pos-0-6-0-0-cr01.chicago.il.ibone.comcast.net (68.86.90.49) 12.653 ms 14.040 ms 11.760 ms
9 xe-10-1-0.edge1.Chicago2.Level3.net (4.71.248.17) 12.335 ms 12.339 ms 13.442 ms
10 ge-7-0.ipcolo1.Chicago1.Level3.net (4.68.101.42) 12.412 ms ge-9-1.ipcolo1.Chicago1.Level3.net (4.68.101.74) 14.214 ms 13.012 ms
11 8.9.16.3 (8.9.16.3) 12.215 ms 14.516 ms 11.997 ms |
|
espaeth
Digital Plumber
Premium,MVM
join:2001-04-21
Minneapolis, MN | Routing has been going back through Chicago for a few days now. |
|
ykram
join:2008-07-16 | Not for me. Earlier it was still going through Atlanta and still does in fact for a few servers. |
|
pr1mo
Premium
join:2003-11-12
Harwood Heights, IL
1 edit | reply to ykram
I live Harwood Heights. Five minute walk and I'm in Chicago. Game servers which used to give me ~10ms now give me on average, 45-55ms. Terrible. |
|
espaeth
Digital Plumber
Premium,MVM
join:2001-04-21
Minneapolis, MN
·voip.ms
·Vitelity VOIP
·Callcentric
·VoiceStick
·ViaTalk
·Comcast
·Embarq
| said by pr1mo  :I live Harwood Heights. Five minute walk and I'm in Chicago. Game servers which used to give me ~10ms now give me on average, 45-55ms. Terrible. If you are going to game servers hosted at Steadfast, then latency is going to be crap for a little while because the winning bandwidth option between them and Comcast is Savvis currently, which means return traffic needs to route to the NY or San Jose to get back to Comcast. This change is the result of a contract dispute between Steadfast and one of their network vendors: »steadfast.net/blog/index.php/200···-changes
They will be turning up Level(3) and Comcast connections in a few weeks, then latency should return to normal. |
|
cward0625
@comcast.net
| reply to ykram
I have had this issue for 5 days with Chicago. I have a Chicago NFO Gaming server, when I had dsl i pinged 5-10 solid. On comcast its been 10-20 when its good. This past week, about 10% of the time its been 10-20, the rest of the time i ping anywhere from 70 - 280. I have had Cust. Serv run multiple tests, and a tech come out today, they can't find anything wrong. Today its got better with 14 ping, but at around 430-530 it went to 100ping again.
comcast seems far more inconsistent than what I had with ATT Dsl, thinking of going back to dsl at this point. |
|
