Question Inconsistent Ping issues with various game servers (tracert output provided) ?

[Z0FT]

Around late August I started having a very weird problem with some of my frequently visited gaming servers where my ping (which was typically ~80 ms to Chicago) would consistently rise to 117-140 between peak usage hours, and still have a unstable connection during the daytime. Some players are experiencing a similar issue, but its only around 3 out of the typical 40 players, and each one of us that is affected is in a different region and uses a different provider. It is not only this server, but others as well (some in New York). However, on other servers in the same regions our ping is perfectly stable and as expected, and there are no issues. Below you will find 2 tracerts, one from a problematic server and one from a perfectly fine server, both of which are in Chicago.

ISP: At&t
Router: 5268AC
Connection Type: Ethernet

Notable PC Specs:
RTX 2080
i7 8700k
EVGA 750 PSU
ASUS ROG Motherboard (Forgot variant)

Problematic Server:
----------------------
2 26 ms 23 ms 24 ms 76-241-8-1.lightspeed.frokca.sbcglobal.net [76.241.8.1]
3 24 ms 24 ms 25 ms 71.147.199.22
4 32 ms 31 ms 30 ms 12.122.160.166
5 33 ms 32 ms 31 ms sffca21crs.ip.att.net [12.122.2.78]
6 28 ms 32 ms 30 ms 12.122.114.53
7 30 ms 29 ms 29 ms 192.205.32.54
8 29 ms 29 ms 29 ms be3669.ccr21.sfo01.atlas.cogentco.com [154.54.43.9]
9 46 ms 44 ms 46 ms be3109.ccr21.slc01.atlas.cogentco.com [154.54.44.138]
10 73 ms 70 ms 77 ms be3037.ccr21.den01.atlas.cogentco.com [154.54.41.146]
11 72 ms 72 ms 72 ms be3035.ccr21.mci01.atlas.cogentco.com [154.54.5.90]
12 73 ms 77 ms 73 ms be2831.ccr41.ord01.atlas.cogentco.com [154.54.42.166]
13 74 ms 74 ms 74 ms be2523.agr22.ord01.atlas.cogentco.com [154.54.81.102]
14 76 ms 78 ms 79 ms be2019.nr01.b056683-0.ord01.atlas.cogentco.com [154.24.70.202]
15 131 ms 125 ms 124 ms 38.122.209.18
16 118 ms 120 ms 127 ms 218-4-88-167.reverse-dns [167.88.4.218]
17 115 ms 112 ms 121 ms 83-12-88-167.reverse-dns [167.88.12.83]
--------------------------

Perfectly Fine Server
--------------------
2 24 ms 23 ms 24 ms 76-241-8-1.lightspeed.frokca.sbcglobal.net [76.241.8.1]
3 24 ms 25 ms 24 ms 71.147.199.22
4 32 ms 32 ms 30 ms 12.122.160.166
5 34 ms 31 ms 31 ms sffca21crs.ip.att.net [12.122.2.78]
6 30 ms 28 ms 30 ms sffca402igs.ip.att.net [12.122.114.29]
7 32 ms 48 ms 28 ms 192.205.32.98
8 74 ms 81 ms 87 ms ae36.cr1-chi1.ip4.gtt.net [89.149.141.165]
9 76 ms 77 ms 76 ms 350.xe2-0-0.br1.ch2.newcontinuum.net [173.205.52.110]
10 * * * Request timed out.
11 * * * Request timed out.
12 78 ms 76 ms 76 ms 104.128.62.207
---------------------

As you can see the server that is perfectly fine seems to have 5 less hops. I left out the first line (since thats my ip 😉 ) but it shows <1ms consistently. Before late August everything was perfectly fine, and unfortunately I dont have tracert results from that time period to compare. I have tried using a VPN as well, but the results were the same. I have tried pretty much everything I can on my end, from flushing my ip, resetting my network, rebooting, and downloading the latest drivers. My speeds are also fine, and are actually slightly above what I'm paying for. Any help in interpreting the results would be appreciated, as well as some help on what to do next.

 

[bill001g]

Number of hops makes no difference it only the latency that matters. Even the pure latency number is not a huge issue as long as it is consistent. Most games have methods so that someone with say a 30ms faster ping does not have a advantage. If the difference gets to say 100ms then it becomes harder to make it even. This feature though can cause a problem if your ping times jump around. Say it is 50ms then 90ms then 30ms. This confuses the prediction software and you can get studdering because of that.

Your big problem is everything you have show you have no issues with "your" network. You have no control over how data travels though the internet. Your only option is to send data out the single path your ISP provides for you and receive data over that same path. How the data travels after that you have no control over.

In your example if we believe the names on the routers your traffic has the same approximate time of 80ms to go from northern California to Chicago. You get the same time even traveling different paths though different ISPs along the way.

The problem occurs at the end of the trace and this is likely in the game companies equipment or somehow related to the ISP the game company is using. No ISP unless you buy very special circuits has any type of guarantee of latency even within their own network. The ISP also has no control of traffic once it leaves their network.

It all doesn't matter. The only thing you pay for is the connection between your house and your ISP. Even if your ISP had no connections to other ISP they still technically meet what they sold you. If you were to test to a speedtest server on their network you would likely get the speed they sold.

There is basically nothing you can do. Its not like you can call up a ISP you do not pay for any service and complain about performance. In this case it seems to be a issue with the server. Could be the server itself, the data center it is in , or maybe one of the ISP the data center uses. Unless this is some server you actually lease and pay for nobody has a contractual obligation to you to fix anything.

 

[Z0FT]

Thats unfortunate, but do you have an explanation for why only a small percentage of the playerbase is experiencing issues? Below is another tracert from the same "problematic server" from a player in Arizona, who has no such issues:

2 38ms 38ms 38ms phn4-ds1-gw02.phn4.qwest.net [71.32.112.2]
3 52ms 39ms 39ms 71-32-113-9.phn4.qwest.net [71.32.113.9]
4 77ms 77ms 77ms cer-edge-20.inet.qwest.net [205.171.93.154]
5 77ms 78ms 78ms 63-234-224-178.dia.static.qwest.net [63.234.224.178]
6 77ms 77ms 77ms xe2-0-0.br0.ch2.newcontinuum.net [64.127.129.6]
7 * * * Request timed out
8 79ms 79ms 79ms 83-12-88-167.reverse-dns [167.88.12.83]

 

[bill001g]

Almost impossible to really tell with so many possible paths between different cities and ISP. You can find sites called looking glass that will let you get into ISP routers and issues some commands like ping and trace that will let you get some idea how things are connected. Still it is pretty much just for information purposes, even if you were to find a issue there is nothing you can do to get anything changed or fixed.

Are you sure your problem is purely high latency. In general that does not cause a huge issue. What is much worse is packet loss and/or random variations in the latency.

I would leave a constant ping run to 76.241.8.1. Maybe you get lucky and you see packet loss. This is the connection between your house and the ISP so they should be willing to fix it.

Time of day issues tend to be overloaded paths. It is kinda heavy traffic during rush hour on the roads...you can't fix that either.

 

[Z0FT]

Almost impossible to really tell with so many possible paths between different cities and ISP. You can find sites called looking glass that will let you get into ISP routers and issues some commands like ping and trace that will let you get some idea how things are connected. Still it is pretty much just for information purposes, even if you were to find a issue there is nothing you can do to get anything changed or fixed.

Are you sure your problem is purely high latency. In general that does not cause a huge issue. What is much worse is packet loss and/or random variations in the latency.

I would leave a constant ping run to 76.241.8.1. Maybe you get lucky and you see packet loss. This is the connection between your house and the ISP so they should be willing to fix it.

Time of day issues tend to be overloaded paths. It is kinda heavy traffic during rush hour on the roads...you can't fix that either.

Its primarily latency. I don't feel any packet loss in game (I've dealt with that before), and it feels smooth from that standpoint. There is a lot of variation in latency/ping though, ranging from 96-140 when in game. The changes in ping dont feel jarring either, just a lot less responsive from the normal 80ms. It wouldn't normally be an issue, but the game I am playing is competitive so it makes quite a big difference.

Results from pinging 76.241.8.1:
Ping statistics for 76.241.8.1:
Packets: Sent = 469, Received = 468, Lost = 1 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 21ms, Maximum = 41ms, Average = 23ms

I ran that test 3 times for ~7-10 mins, there was maybe 1 or 2 instances of packet loss per test but they were always within the first 50 tries before stabilizing. Ping was also generally consistently lower the longer the test went on.

I've contacted the server owner and sent him my traceroute, he said he sent it to the server box people to be fixed in conjunction with the database and local ISP's, I'll update this thread if anything comes of it.

 

[Z0FT]

Ye it seems to be a Cogent Communications issue, as 38.122.209.18 is one of their databases and everyone that has a problem runs through that hop. All the players without problems seem to use another database enroute. I even found some reviews of the ISP online detailing:

"Cogent could not provide reliable datacenter internet connection for one of our customers. So he had to migrate to another provider to do not lose his clients."

Hopefully they resolve their issues or do something, as this is a recent occurrence. Given that it started around late August it could even have something to do with the fact that they may not have been prepared to deal with the surges in traffic between 5-9 pacific. Pain.

Thank you for the help and helping me better understand the tracert results.

Edit: I also found this thread (from a different game) with people who had the same problem being routed through cogent - https://us.forums.blizzard.com/en/w...hrough-cogent-communications/214572/87?page=4