Plagued by Lag Spikes/Stuttering

From a week past or so i'm having frequent lag spikes and stuttering during prime hours in Europe (7-11pm GMT).
At first i thought it was my ISP but they couldnt detect any issues with the signal, so i started running WinMTR.

FYI, i'm on a wired connection, did a clean format and install (to be safe).

Seems to me the issue is in the last hop (with significant packet loss), on both the Paris Amsterdam servers (London is completely unplayable for me).

WinMTR results:

Paris (no issues, non peak hour):
https://pastebin.com/0pFwibUr

Paris (light spiking/stuttering):
https://pastebin.com/S13MFJJT

Paris (heavy spiking/stuttering):
https://pastebin.com/PsCNVqeL


London (heavy spiking/stuttering):
https://pastebin.com/uESigKrK


I appreciate all the help you can give me.
Last bumped on Sep 25, 2017, 1:49:59 AM
In each of the Paris traces your packet loss starts anywhere from 1-4 hops before the last. London is 2nd to last hop where PL happens.

Email techsupport@grindinggear.com. Send them your traces. They can holler at their service provider to see if something can be done.
"Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."

"Your mother was a hamster and your father smelt of elderberries!"
Thankyou for your response.

I have emailed tech support in the meantime.



Fyi, i was disregarding the Packet Loss (PL) in "ae-2-3201.ear1.Paris1.Level3.net" as it seems to not carry forward and was only taking into account the last PL "5.71.089f.ip4.static.sl-reverse.com".

is this an incorrect reasoning?

And regarding the last node on the WinMTR ("5.71.089f.ip4.static.sl-reverse.com"), what sort of node is it? seems to be (at least from the different designation) something already in GGG's server network.

Ty
"
severinopt wrote:
Thankyou for your response.

I have emailed tech support in the meantime.



Fyi, i was disregarding the Packet Loss (PL) in "ae-2-3201.ear1.Paris1.Level3.net" as it seems to not carry forward and was only taking into account the last PL "5.71.089f.ip4.static.sl-reverse.com".

is this an incorrect reasoning?


The 27 is an aberration, but there is still PL that is carrying forward from that node. (I am at work and can only look at things quickly)

"
And regarding the last node on the WinMTR ("5.71.089f.ip4.static.sl-reverse.com"), what sort of node is it? seems to be (at least from the different designation) something already in GGG's server network.

Ty


I don't know their server structure or layout. It could be or it couldn't be, I do not know.
"Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."

"Your mother was a hamster and your father smelt of elderberries!"
Last edited by Aragorn14 on Sep 19, 2017, 10:53:26 AM
Been in contact with support (going through the "is your connection wired?" dialogue atm).

Meanwhile, seems to me (from further testing on various gateways) that the issue of PL is consistent on the last (or previous to last) hop.

IS GGG aware of issues like this?

https://pastebin.com/h5CjiPcD

https://pastebin.com/7gudeLeC
I don't know what sort of telemetry they run on the backend to know whether they can see these types of things or not. So, maybe? :)

Best thing to do though, to make it better for everyone, is to make sure and send those MTR's to them when you can so that it gets into the right hands.
"Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."

"Your mother was a hamster and your father smelt of elderberries!"
I can't reproduce this problem at the moment: https://pastebin.com/Rske6UjL

I think it's about 12am GMT currently, so it may have returned to normal after peak hours.

I'll run a capture to see if this changes over the next day.
"
apexmateria1 wrote:


It's not an FPS issue. It seems to be a Packet Loss issue.

I just want to add that I've been having similar experiences on most european realms during prime time (9pm to midnight GMT). After a few days of trying to diagnose my routing I have resigned myself to fishing for an acceptable realm once prime time hits.

I normally play on Paris, which is pretty stable and has the best ping for me, but once I start seeing the spikes at around 9pm GMT, I run concurrent winmtr traces on all european realms and pick the one that looks the best. Amsterdam and London tipically handle prime time load better but generally speaking all euro realms are showing packet loss on the last and before last hop (softlayer on most realms although London's realm apparently uses a different backbone provider).

I don't believe this to be a problem within PoE's server providers network but rather their backbone providers struggling to keep up with prime time load. I've had my fair share of dealings with Softlayer over the years (I'm a network engineer) and I'm fairly certain they are aware of the problem and trying to fix it. Backbone load increases steadily every year and it peaks up after summer break every year which is when most providers upgrade their infrastructure. Hopefully it will be fixed soon.

Report Forum Post

Report Account:

Report Type

Additional Info