dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
3608
share rss forum feed

domiguy85

join:2009-06-02
Warner Robins, GA

3 edits

[GA]WarnerRobins High latency\bad tracerouting to many areas

I just started Cox Premiere Internet Service yesterday and after the technician set everything up (he did a great job by the way), all looked good after he did a speedtest to Atlanta.

However, after playing some online games, I noticed that I was getting horrible pings to many areas in the U.S. I was averaging around 150 ping to places such as Dallas or Chicago. I know there's some decrease in ping the further from my location, but nowhere near to this degree (this is worse than my slower DSL prior to me changing to cable).

I was wondering if there was something currently being worked on by Cox to cause this issue since I'm getting great pings to some areas, but horrid pings to others. Around 11-12 p.m. yesterday, the pings all went down to 50 so I assumed the issue was fixed. However, as of 5 p.m. today, the pings have jumped back up to 150 to select areas. Can someone resolve this issue for me? I really hope it's not a congestion/throttling issue as this was NOT what I was anticipating when I got the Premiere service.

Thanks.

Note: I am currently in the process of doing some traceroutes to display here.

domiguy85

join:2009-06-02
Warner Robins, GA

3 edits

Re: [GA] Very high latency or bad tracerouting to many areas

Here's some straightforward pings:




Here's some traceroutes:




What's up with the 68.1.4.133 node? I seem to be getting packet loss as soon as it hits there or anything in the 68.1.4.x range. It seems to be in Atlanta.




Notice how I get packet loss as soon as I hit the 68.1.4.139 address. Does this narrow down the problem at all? I guess I could call about it if need be as I'm not entirely confident my modem was activated properly (but it wouldn't work at all then would it). I did the online activation method, but noone told me to and the technician seemed surprised about it since the lady on the phone didn't ask me for the numbers.


NormanS
I gave her time to steal my mind away
Premium,MVM
join:2001-02-14
San Jose, CA
kudos:12
Reviews:
·SONIC.NET
·Pacific Bell - SBC
Except for your Yahoo! trace, where your endpoint packet loss starts in the Yahoo! transit network, 100% of your packets are "returned" to you (technically, that represents responses to sent packets; all trace route packets are actually dropped).

But your latency is not consistent. Possibly due to a large asymmetry in the return path. You would need to find a "looking glass" server to trace back to you to see your return path. Best would be to request that looking glass to trace to you, then trace route to the looking glass and compare the results.

Hurricane Electric has them all over the Internet:

»lg.he.net/
--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum


Optimus2357
Premium
join:2010-11-21
West Warwick, RI
kudos:3
Ooh, good link. Im book marking that gem.

domiguy85

join:2009-06-02
Warner Robins, GA
reply to NormanS
Click for full size
Here's a couple results; I first made sure the servers were up by being able to ping them, then put in my ip address. It seemed to fail each time though. Then I ran a traceroute:




merlock
Premium
join:2003-04-05
Warner Robins, GA
reply to domiguy85

Re: [GA]WarnerRobins High latency\bad tracerouting to many areas

No problems for me.


NormanS
I gave her time to steal my mind away
Premium,MVM
join:2001-02-14
San Jose, CA
kudos:12
Reviews:
·SONIC.NET
·Pacific Bell - SBC
reply to domiguy85

Re: [GA] Very high latency or bad tracerouting to many areas

said by domiguy85:

Here's a couple results; I first made sure the servers were up by being able to ping them, then put in my ip address. It seemed to fail each time though.

The failure is consistent with a destination blocking ICMP at the firewall. For a looking glass trace to work the destination has to be "pingable".

It does look as if you go out on Cox transit to Atlanta, Georgia before transition to HE transit, which carries you from Atlanta to Chicago. HE connects to Cox transit directly in Chicago before coming back to Georgia.

158 ms between Chicago and Georgia is high.

There seem to be two separate issues:

• Something going on with Yahoo!.
• Something going on with Cox transit.

Hopefully Cox can figure out the latter for you; but Yahoo!?
Tracing route to ir2.fp.vip.bf1.yahoo.com [98.139.183.24]
over a maximum of 30 hops:
 
  1    <1 ms    <1 ms     1 ms  homeportal [192.168.42.1]
  2    26 ms    26 ms    25 ms  173-228-7-1.dsl.static.sonic.net [173.228.7.1]
  3    24 ms    25 ms    24 ms  gig1-4.cr1.lsatca11.sonic.net [70.36.243.13]
  4    24 ms    25 ms    24 ms  0.xe-5-1-0.gw.pao1.sonic.net [69.12.211.1]
  5    24 ms    25 ms    24 ms  0.xe-6-0-0.gw3.equinix-sj.sonic.net [64.142.0.185]
  6    25 ms    25 ms    26 ms  PAT1.SJC.yahoo.com [206.223.116.150]
  7    74 ms    54 ms    74 ms  as-0.pat1.da3.yahoo.com [216.115.100.67]
  8    80 ms    76 ms    78 ms  ae-3.pat2.che.yahoo.com [216.115.96.54]
  9    91 ms    91 ms    94 ms  xe-1-2-0.pat1.bfy.yahoo.com [216.115.97.196]
 10   107 ms   109 ms   128 ms  ae-5.pat2.bfz.yahoo.com [216.115.96.67]
 11   128 ms   130 ms   108 ms  ae-4.msr1.bf1.yahoo.com [216.115.100.25]
 12   110 ms   109 ms    92 ms  UNKNOWN-98-139-129-X.yahoo.com [98.139.129.167]
 13    93 ms   113 ms   112 ms  et17-1.fab8-1-sat.bf1.yahoo.com [98.139.128.91]
 14   112 ms   110 ms   102 ms  po-16.bas2-7-prd.bf1.yahoo.com [98.139.130.3]
 15   383 ms   413 ms   360 ms  ir2.fp.vip.bf1.yahoo.com [98.139.183.24]
 
There is no site in the continental U.S. which should be more than 120 ms, or so, from San Jose, California.

--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85

Re: [GA]WarnerRobins High latency\bad tracerouting to many areas

I tried to put in a number of areas in case of specific sites having bad servers in general, but yeah, I agree with you that the highest I should be getting is something in California. No way I'll be fine with high pings in other areas though.


Savvis

@cox.net
reply to domiguy85
Evening,

I'm noticing the same issues with Cox in Warner Robins too. Here are my results for the local area:

Speed Test:

»testmy.net/fuNvVcz.B6AjdXV.png

Ping Test (Atlanta):

»www.pingtest.net/result/79745814.png

Ping Test 2 (Atlanta):

The results drastically change after 6 minutes.

»www.pingtest.net/result/79746054.png

Ping Test (Orlando):

»www.pingtest.net/result/79745850.png

Ping Test (Ashburn):

»www.pingtest.net/result/79745998.png

Ping Test (Dallas):

The results return back to normal.

»www.pingtest.net/result/79746199.png

domiguy85

join:2009-06-02
Warner Robins, GA

2 edits
reply to domiguy85
It's still bad for me as of 11:20 p.m.

Annnnnnnd, as of 11:30 p.m my pings are all at acceptable ranges. This is ridiculous and suspicious how it magically fixes itself at a certain time.




Note that I am not ok with this inconsistent performance, especially considering my service just started.

domiguy85

join:2009-06-02
Warner Robins, GA

4 edits
reply to domiguy85
Just a reminder that I will be constantly posting here and calling Cox technical support until I feel confident that this issue is resolved. I also have DSL service as a backup, so I'm under no obligation to keep this noncontracted (monthly only) Premiere service if I feel it's inadequate.

--

Pings are fine as of 6:30 a.m 4/12. I have a traceroute continually running so I'll be able to tell when it goes bad again -- which I expect it will during those 'peak' hours.

Update:

11:30 a.m 4/13 - pings still about ~50 on avg

1:20 p.m. 4/13 - pings still about ~50 on avg

3:45 p.m 4/13 - pings still about ~50 on avg

5:40 p.m 4/13 - pings averaging about ~70 or so; some congestion is understandable since this is probably peak time -- as long as it's not anywhere as bad as before.

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
At about 7:00 p.m the pings started jumping to unacceptable levels. Congestion aside, there should not be that poor of a performance hit, especially for higher tier services. I will be calling to see what can be done about it; if nothing, I will be canceling/getting a refund for the poor service.



domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
So I called tech support again and was able to get through to a level 2 support guy named Basil I believe.

I basically gave him all the information you guys see here and we did notice how the latency shoots up after it hits certain hops (usually 68.1.4.133 and 68.1.8.1 depending on server). He recorded all the information I sent to him in an email and said he'd contact his supervisors about the issue. I will be getting a guaranteed call (his words) within 72 hours about it.

Based on that call, it'll likely be the make or break point of me deciding to continue the service or not since I really can't do anything on my end to alleviate the issue (that I know of).

Here's some pings of those two ip addresses by the way as of 9:42 p.m 4/12 -- these are cox owned (I guess they're the routers or something) and a majority of connections to out of state servers will go through these hops:

>ping 68.1.4.133 -n 10

Pinging 68.1.4.133 with 32 bytes of data:
Reply from 68.1.4.133: bytes=32 time=28ms TTL=60
Reply from 68.1.4.133: bytes=32 time=141ms TTL=60
Reply from 68.1.4.133: bytes=32 time=28ms TTL=60
Reply from 68.1.4.133: bytes=32 time=152ms TTL=60
Reply from 68.1.4.133: bytes=32 time=155ms TTL=60
Reply from 68.1.4.133: bytes=32 time=44ms TTL=60
Reply from 68.1.4.133: bytes=32 time=148ms TTL=60
Reply from 68.1.4.133: bytes=32 time=56ms TTL=60
Reply from 68.1.4.133: bytes=32 time=153ms TTL=60
Reply from 68.1.4.133: bytes=32 time=31ms TTL=60

Ping statistics for 68.1.4.133:
Packets: Sent = 10, Received = 10, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 28ms, Maximum = 155ms, Average = 93ms

>ping 68.1.8.1 -n 10

Pinging 68.1.8.1 with 32 bytes of data:
Reply from 68.1.8.1: bytes=32 time=136ms TTL=254
Reply from 68.1.8.1: bytes=32 time=139ms TTL=254
Reply from 68.1.8.1: bytes=32 time=125ms TTL=254
Reply from 68.1.8.1: bytes=32 time=118ms TTL=254
Reply from 68.1.8.1: bytes=32 time=121ms TTL=254
Reply from 68.1.8.1: bytes=32 time=130ms TTL=254
Reply from 68.1.8.1: bytes=32 time=156ms TTL=254
Reply from 68.1.8.1: bytes=32 time=104ms TTL=254
Reply from 68.1.8.1: bytes=32 time=119ms TTL=254
Reply from 68.1.8.1: bytes=32 time=121ms TTL=254

Ping statistics for 68.1.8.1:
Packets: Sent = 10, Received = 10, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 104ms, Maximum = 156ms, Average = 126ms

>

As you can tell, it's impossible to get a decent ping if I'm already hosed before even leaving the state.

domiguy85

join:2009-06-02
Warner Robins, GA

1 edit
reply to domiguy85
Good example of how the latency is erratic then dies down during off hours. I was told this was NOT normal whatsoever.




If I had to guess, there's some hardware that's struggling to handle the load at those router locations I mentioned above.

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
Bad latency again as of 8:57 p.m 4/13.

I haven't received the call about the issue yet (I think).

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
Still bad as of 11:44 p.m 4/13


.


Savvis

@cox.net
reply to domiguy85
Thanks for keeping us updated. I'd be curious what they say when they get back to you.

domiguy85

join:2009-06-02
Warner Robins, GA

3 edits
reply to domiguy85
Another example of latency being terrible then slowly reverting to normal once Cox's hardware can take it at about 12:50 a.m 4/14. I sat here and watched the latency of their 68.1.8.x addy go from 150 to the numbers you see now:




Basically there's a only a couple things that'll happen from here. I'll be contacted by or call level 2 support and get a yes, we're fixing it, or no we don't know what to do, and then decide whether I'll continue with their services or not. I should know by Monday or Tuesday at the latest. I'll cancel if nothing good comes out of it. Due to the poor service and the fact that I wasn't a customer for even a week, I'll obviously not be paying for anything. I am tempted to try and test other modems and/or the Ultimate tier just to see if there's any differences. I'd be extremely surprised though.

What's a shame is the technician who came by when my connection was setup did a great job running the cable, going under the crawlspace, checking everything was working fine, and he was really friendly during his time here. It's disheartening to know he wasted his time in a way. I made sure to give him credit in the feedback though.

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
Summary of this issue so far:

Modem: Motorola SB6141 out of the box directly connected to PC
Service: Premiere
Length of issue: Since I was a customer (4/10/13)
Ticket Number: 1538941

Latency issues during peak congestion hours (usually 6 p.m to 11 p.m).

In-state tests are usually fine. Once my connection is routed out of state, notably by the ip addresses 68.1.8.1 and 68.1.4.133, latency jumps to 120+. These are cox owned addys.

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
1:44 p.m 4/14



domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
Update 2:28 4/14 p.m.:

Called Level 2 tech support and talked to a rep named Bill. He looked over the ticket information and said that the engineers couldn't find any issues on their network and the problem seemed to stem from general routing issues after the fact. However, he said that he did see the high latency himself and that it was definitely unusual behavior that's happened before and is temporary.

He said I should be able to get a full refund if it's like that after two weeks, so I decided to just switch between using my current DSL service and cable service depending on the time of day. They've been updating my ticket information with all this.

It kind of annoys me I'm giving them time, but at least they admitted it's not on my end and it's not normal. I'll update here when that times comes as it's the last chance I'm giving.

Target date: 4/28/13

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
Comparison of my Cox 'Premier' Service vs the DSL line I have. I switched back and forth to see when the cable's latency went down. It usually isn't till around 11:30 p.m. or so after all.


domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
Here's a better picture:




So the DSL currently has superior routing over COX's lines. Or at least they have much better contingency plans.


Cox_tiff

join:2013-01-12
Chesapeake, VA
reply to domiguy85
Hi domiguy85,

Looking at your traceroute results, it appears to be packet depriorization and not a latency issue. The packets after that all show lower response times which would not be possible in a true latency issue. Each packet would have to be at least as high or higher than the point for there to be latency. If you continue to receive low speeds with the router bypassed, we may want to schedule a service call to look at your signal.

Tiffany
Cox Social Media and Forums moderator

domiguy85

join:2009-06-02
Warner Robins, GA

1 edit
said by Cox_tiff:

Hi domiguy85,

If you continue to receive low speeds with the router bypassed, we may want to schedule a service call to look at your signal.

Tiffany
Cox Social Media and Forums moderator

said by domiguy85:

Summary of this issue so far:

Modem: Motorola SB6141 out of the box directly connected to PC

I appreciate the quick comments Tiffany (please look over my other posts carefully), but is there implication that it is ok since it's 'packet prioritization'? The end result is I get higher latency than I should be getting, so saying as such seems very much applicable to me. Regardless of the term, it's ongoing as shown here:




I called Tier 2 today at 5:00 p.m and once again I was told that a tech would likely find nothing since the signals appear green on their end. However, if that's needed to escalate the issue even more, I'll go ahead and do it since I am getting conflicting answers. There doesn't seem to be much coordination on the issue.


Cox_tiff

join:2013-01-12
Chesapeake, VA

1 edit
Hi Domiguy85,

The packet depriorization wouldn't affect your speed when surfing the internet, playing online games, transferring files, etc. Most likely ICMP packets (Ping and Traceroute) are assigned a low priority. If the device is configured to prioritize other traffic first, then all ICMP traffic will have to wait till all other traffic has been processed before they will respond to ICMP traffic.This will appear as a false positive, but if non-ICMP traffic is routed through the same server, it would not experience the same depriorization. The reason why this would not cause any latency issue, is because the type of traffic you would be sending through the server is not ICMP.

I would be happy to have our escalations team look at it again for you, since it does appear in your later traceroutes to maintain those ping rates.

Tiffany
Cox Social Media and Forums moderator

domiguy85

join:2009-06-02
Warner Robins, GA

1 edit
Well, whatever it is, it's very apparent when playing online games. The pings are all 120+ or higher than usual to servers. The traceroutes very much show this when I run it to those same servers.

Anyway, I called tech support to schedule an appointment regardless of what was told to me. Apparently this is needed to escalate the issue from the technician to maintenance if no problems can be found at my end. I did not know this (and I would've scheduled an appointment sooner if I had known).

They're coming by in a couple days, so I'll see where it goes from there. I'm glad I didn't wait the week or two as I originally planned as the Level 2 guy I got off the phone with seemed confident about it being fixable.

domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
Another traceroute for documentation sake 4/15 at 8:08 p.m:



domiguy85

join:2009-06-02
Warner Robins, GA
reply to domiguy85
4/16 8:19 p.m



NormanS
I gave her time to steal my mind away
Premium,MVM
join:2001-02-14
San Jose, CA
kudos:12
Reviews:
·SONIC.NET
·Pacific Bell - SBC
reply to Cox_tiff
said by Cox_tiff:

The packet depriorization wouldn't affect your speed when surfing the internet, playing online games, transferring files, etc.

domiguy85 See Profile is showing unusually high latency on his first Cox IP hop. I have run tracert on AT&T (ADSL Fastpath), Charter (DOCSIS 2.0), Comcast (DOCSIS 2.0), and Sonic.net "Fusion" (ADSL2+ Interleaved) connections. I've never seen latency higher than 25 ms (ADSL2+ Interleaved) on the first IP hop; the others being 9 ms (ADSL Fastpath) to 15 ms (DOCSIS 2.0).

Even I consider 100+ ms on the first IP hop as unacceptable, and latency isn't all that critical to my operations.
--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum