dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
6
share rss forum feed


bluepoint

join:2001-03-24
reply to DrDrew

Re: Abnormal Pings With Roadrunner

Again it depends on what's acceptable to you. The OP is a gamer that wants a clean route to the server. At primetime, when the gateway is overwhelmed of requests the overall latency of gamers are affected. Gamers needed the fastest response it could get. You can tell us there is nothing wrong when the destination is reached in accepted time but it's not as fast as a clean route.
See you keep on telling us equipment can be configured to response to certain request but you are not even sure how it is configured, in other words you just want to divert the problem away from TWC. If you remember my traces right after Sandy? My gateway shows it's timing out then but now at anytime it responds within 30ms. Can you still say it's configured not to respond?


DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:16

1 edit
said by bluepoint:

Again it depends on what's acceptable to you. The OP is a gamer that wants a clean route to the server. At primetime, when the gateway is overwhelmed of requests the overall latency of gamers are affected. Gamers needed the fastest response it could get. You can tell us there is nothing wrong when the destination is reached in accepted time but it's not as fast as a clean route.

Read the OPs original complaint again:
said by Xtreme2damax:

It's not that my ping is high, just higher than it should be. I game online and I've been noticing lag and I've been told that my ping is higher than it should be considering the location of the server I am playing on. I live in Auburn, NY and when I ran a speed test my ping to Syracuse, NY is 5ms which is fine I believe. But then I tried running a speed test and connected to the Rochester, NY server where I was getting over 100ms ping.

The game server I connect to is down by New York City and my ping to that is anywhere from 35ms - 50ms, someone that lives in Florida and also plays on the same server gets around the same ping as me. I ping around 10ms - 15ms to a server in Chicago, IL, 90ms to Los Angeles, CA, 55ms to Miami yet my ping connecting to a NY server is higher than it should be. I should be getting around 5 - 15ms.

Pings to SOME NY servers are higher than he thinks they should be because pings to Chicago and other NY sites are lower. Pings to other sites around the country seem good too. To me that seems like more of a routing issue, especially through a specific router or path to that location, that may or not be a TWC issue. Really more example traces to a variety of sites need to be shown. If it were a gateway issue pings to all sites would be affected, but that's not what the OP is complaining of.

said by bluepoint:

See you keep on telling us equipment can be configured to response to certain request but you are not even sure how it is configured, in other words you just want to divert the problem away from TWC. If you remember my traces right after Sandy? My gateway shows it's timing out then but now at anytime it responds within 30ms. Can you still say it's configured not to respond?

You don't know how it's configured either, but you keep latching on to something you don't even understand and ignoring the majority of the other information which points to something else entirely.

You're can't see the forest for the trees...
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.

harald

join:2010-10-22
Columbus, OH
kudos:2
If you drive your car to Poughkeepsie or Kankakee, you are near every location in those cities. The internet doesn't work that way.

Routing is everything. Destination is nothing. In most cities there is very little interconnection between local sites.

If you run trace routes, run them to the sites of interest to you. Comparisons make no sense otherwise.


DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:16
said by harald:

If you drive your car to Poughkeepsie or Kankakee, you are near every location in those cities. The internet doesn't work that way.

Routing is everything. Destination is nothing. In most cities there is very little interconnection between local sites.

If you run trace routes, run them to the sites of interest to you. Comparisons make no sense otherwise.

The reason for comparisons is to find the common path and common choke points. You can trace to the same destination from different start points, often there's a common destination path that may show similar issues. You can trace to different destinations from the same start point and check to see if and where in the common path an issue appears.

The goal is to find the problem by running data multiple ways to eliminate as many single data points as you can. Another PC with the same issue, usually eliminates the PC. Another neighbor with the same issue, usually eliminates the home wiring. Another area with the same issue, usually eliminates the local last mile network. Another ISP with same issue, usually eliminates the ISP. etc, etc.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


bluepoint

join:2001-03-24
reply to DrDrew
said by DrDrew:

Read the OPs original complaint again:

You are ignoring the second hop of his tracerts. That's the reason he's having high latency, it's a sign of an oversold node.

said by DrDrew:

You don't know how it's configured either, but you keep latching on to something you don't even understand and ignoring the majority of the other information which points to something else entirely.

You're can't see the forest for the trees...

You are a funny guy, If I'm not mistaken networking is part of your job at TWC. You are expecting me to know how your network is configured? I can not tell you for sure but I have an idea that it's configured not to respond to ICMP packets when it reach capacity because that's how it reacts. I know you know the answer but you ignore to tell us, instead you want to blame it on anything but TWC.
Hobs knows the nodes are overloaded, what else can you add?


DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:16

4 edits
said by bluepoint:

You are ignoring the second hop of his tracerts. That's the reason he's having high latency, it's a sign of an oversold node.

I'm not ignoring it, I'm using that information in context of what the reported issue is and experience troubleshooting cable ISP connections.

Why isn't he having high latency EVERYWHERE if his node is "oversold"? Why not to Chicago or other parts of NY state? Why is the latency complaint just about the route to his game server?

said by bluepoint:

You are a funny guy, If I'm not mistaken networking is part of your job at TWC. You are expecting me to know how your network is configured?

Not my network, not part of my job to design, install, configure, maintain, or repair. I've troubleshot network connections for a variety of providers mostly from the end user side over a dozen years though and seen CMTSs not respond to pings for years on TWC, Comcast, Charter, Adelphia and others. Again, just because the CMTS isn't responding to ICMP ping packets full time doesn't mean it's overloaded or having issues routing end user data. You can try other packets to see if it's just ICMP ping packets, try UDP or TCP pings. See if those are answered the same way.

I've learned to see through what seems to be a problem to the uninformed, narrowed down the actual issue, and referred them to the companies and departments that can handle the issues. I'm not here to deflect problems away from TWC or any other companies' responsibility, I'm here to direct people with problems to solutions. If this showed signs of being an overloaded anything I'd say it, right now it doesn't look like it.

What I can add is, more data about the situation from the OP is needed. Pings and tracerts from the outside world back to the user's IP or neighbors would show an overloaded gateway more than a single non responding hop would. It could also show odd routing and/or a variety of other possible issues.
said by bluepoint:

I can not tell you for sure but I have an idea that it's configured not to respond to ICMP packets when it reach capacity because that's how it reacts. I know you know the answer but you ignore to tell us, instead you want to blame it on anything but TWC.

Capacity of WHAT exactly? Capacity of the DOCSIS channel, capacity of the route processor, capacity of the configured ICMP limits, etc., etc.?

I have no clue how it's actually configured. I'm just using previous experience with similar conditions on several ISPs to guess at what is actually going on. More information is needed before blaming anyone for the cause of problem.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


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 bluepoint
said by bluepoint:

You are ignoring the second hop of his tracerts. That's the reason he's having high latency, it's a sign of an oversold node.

Having actually been on a congested "node", I can say that trace route does not look congested. In my case, the "node" was an aggregation router, first IP hop after the DSLAM. The SBC techs referred to it as an, "exhausted router". I saw high latency from that IP hop to the end. OP's trace does not exhibit that characteristic.

What baffles me is that nobody has brought up return path issues. If there is a wide disparity between the forward path (which we can see from the OP's trace) and the return path (which requires a trace from the remote connection back to the OP), that can cause anomalous results.

And if the performance is caused by routing, it may be beyond Road Runner's control.
--
Norman
~Oh Lord, why have you come
~To Konnyu, with the Lion and the Drum