I play mainly in Europe, and the best server gaming ping in my country is 8 ms. Unfortunately I don't get any benefit from this 8 ms ping ...which seems rather strange! When I choose a remote game server like usa I feel the latency because the server is very far but my permanent latency is added to that! When we play offline compared to online the difference is obvious!
Is it better in off hours? (eg: Tuesday 4 AM)
I would like to insist on a precise point, step 1 consists in finding the reason why it remains whatever I do, this latency! Step 2 would be the remedy but it would be a miracle! So I concentrate on step 1, the reason of this latency and this unimaginable delay!
Unfortunately the schedule doesn't matter... I have 3 fluid games in 365 days and it can happen at 4:00 AM or 04:00 PM
I still think you should have persevered with cake-autorate's logging capability to expose RTTs to your gaming servers at the times you see lag. A few hours worth of data from that during your gameplay would be very useful and help move the needle forward.
You just need usb stick and service running in background capturing the data like a flight recorder on a plane.
From what I'm reading, we have the same problem, we play with European servers and it's unplayable, a constant lag and a feeling of brutal heaviness.
1 Like
Thank you very much for insisting, I will have more time and work hard to set this up... I hope you will be available that day !
1 Like
Regarding the flight recorder as on a plane, you think of a switch that allows mirroring on the wan with wireshark?
Hello guys,
Im from germany and suffering from this too and want to share my expierences.
I suffering since 2017 and i had this phenomal on 2 adresses with VDSL and cable internet and was able to fix this 2 times for a short period of time.
On my first adresse i had cable and tried everything what you can imagine of, with no sucess. I was aware that it has to do something with the internet and its not on the own home network. It has to be something outside, something unknown.
I feel the delay in every online game which i play. I have no time to react, enemys are ice skating, bullets not hitting, mouse feels sluggish. The whole gameplay is bad and heavy and you are like always a half of a second behind. Everyone besides you have an advantage.
The first time i fixed it on cable internet with a VPN and it worked like wonder. Unfortunatly my ISP got bought from another one since that it was not working anymore. Im assume that they have changed something on the routing policy. No matter which VPN i choose, i had immidiately gameplay improvement and everything worked as intendet for 2 hours. After that i gain over 50% of packet loss and had to choose another one, but it worked.
Then i moved out to a new location where only VDSL was avaible. I had the same issue there and no VPN helped. Than i was reading many threads on blurbusters where most of the people expierences the same but thinking that it has to be power related. For me its clearly that it is a network issue. Some people expierences the issues only at the afternone and night and in the mornings its battery smooth. Many people suffer from this only on specific times. This might be think that the whole area where you living is overused by other customers.
I got it fixed on my new adresse for 3 days, when i called my ISP and forced them to reset my line. After that i had 3 days long smooth gameplay till they restarted my router on 3 am, which never happened before and the delay came back. The logs was reading that the ISP forced a restart so i assume that they are the cause of the issue. They had implemented something on my line which causes the lag.
My suggestion is DLM Dynamic line Management with ASSIA profile. I think that this profile got reseted the first time, when they reseted my line and the software needed time to collect my datas, error rates etc. After 3 days it collected enough data to inject my old profile to my line and since that, everything is sluggish. So for me its clearly that the ISP is throttling my connection with DLM. Other linereset was not helping anymore and the only fix for is, to move out to a new location, where the issue is not occure. The ISPs are fault in my opinion. I only gained improvement when they changed something and let me tell you, it was like night and day.... when i take my pc to friends, i have no lags so yes. It is 100% related to internet due ISPs throttling.
DLM is a software which runs on the DSLAM where you connected to and runs automaticly so changing IPS will not help, since its always the same DSLAM Port. Thats why changing ISP wont help either.
There is also video of this, where you clearly see the difference between a player who had perfect connection and someone who suffer from this. Even the uploader of that video says, that he has an advantage over the player who got killed and that no pro gamer can react to that.
This is not how DLM works. In the network of Deutsche Telekom the statistics for each end user line are aggregated in 15 minute intervals (from the DSLAM/MSAN side) and sent to higher up analysis nodes. There something close to Assias DSLExpresse runs and tries to optimize each line's settings individually (but in a way that does not necessarily makes each line as fast as possible, but over the whole "bundle") in regards to maximal sync rates as well as interleaving depth and transmitter power.
These analyses run every night and if a specufic link is deemed ripe for an intervention it will experience a resync somewhen in the night, IIRC between 1 and 5AM or in that rough period.
If DLM intervened on your link you might be able to note this by seeing DSLAM max rates of even values (see here for a collection of rates observed after DLM intervention).
After that DLM will continue to monitor your link's statistics and might change things for the faster or the slower depending on the statistics, but DLM will only ever change the DSL profile which then the DSLAM will enforce on your link, DLM itself does not throttle your link!. Your ISP however will also employ an additional traffic shaper to limit your link to slightly below the sync rate (DSLAMs are basically switches with weird L2 interfaces and are not that well suited to enforce traffic rates).
Mind you, I am not saying your ISP might not be responsible for your bad gaming experience (I have too little data to build an opinion on that) just that "DLM" is very unlikely to be the explanation.
No it does not, it collects statistics data from the DSLAMs/MSANs (that these collect anyways) and then tries to generate optimal* DSL profiles for each individual link. These profiles then get moved back to the DSLAM/MSANs and are employed for the respective lines.
*) Optimal from the perspective of your ISP, so likely aiming for stability over all lines and probably allowing the highest aggregate speed over a bundle as compared to all individual lines.
Out of curiosity do reboots of the router affect the process? Is it ever in the users interest to try to force more frequent resyncs by e.g. powering off the modem for 1 hour every night or week?
I can not tell for certain, but it was/is rumored that DLM can not really distinguish between unenforced resyncs (which are diagnostic for an unstable link where intervention can help to increase stability) and a user that simply restarts/reboots the modem frequently. The recommendation was to try not to have more than one resync per 15 minutes (so to only accumulate one resync in that 15 minute bin). But when people tried to test it some reported harsh intervention by DLM as response to frequent resyncs, while others reported no intervention even for days on end with forced resyncs. (It does not help that the ISP rolling out DLM in Germany has not done so with a big flag day but bit by bit, so the two different behaviours described above might be caused by the former having DLM activated, while the latter not yet). There were also rumors that some modems might tell the DSLAM properly when they where shut down (I think this was called dying gasp) and that DLM might exclude such resyncs from its stability observations. But these were always just rumors by those sitting at the end-user side of the DLM while those in the know from the ISP side did not really talk...
Personally I never purposefully tested the boundaries of the DLM system because at that time my OpenWrt modem was quite unstable with frequent resyncs, so I was already in DLM's doghouse all of the time anyway... Since this is fixed I got the maximum profile for download (116.7 Mbs) but my uplink is still limited to 37Mbps (the reported estimated line capacity is ~40 Mbps); but I do not complain about these handful of Mbps I loose in upload direction, as I value a stable reliable link over 'top-speed'.
I haven't done dedicated 'testing' either, but in my experience (DTAG, 2017-2021, VDSL2+vectoring/ profile 17b, Broadcom DSLAM <--> Draytek Vigor 130b, 100/40 MBit/s) I could reboot (or replace) the router (PPPoE endpoint) as often as liked, just don't dare to update/ reboot the modem - or I got penalized for up to a week (single modem reboot as part of a firmware upgrade, -20%-25%).
1 Like
In Switzerland ISPs have 32 different profiles, depending on customer complaints they may choose a profile down for a customer who is experiencing disconnections. We have tested all 32 profiles with ISP and no profile is able to stabilize a line except to reduce the bandwidth... turn off the router all night at startup the next day it is very likely that the public IP address is new. I have found that depending on the IP you inherit, some do not generate the same ping on a test like Speedtest.
The part I find most interesting is how people say that they get relief from this problem by going to a friends house... where everything is fine. This is what makes me think it's packet loss or delay in the ISP routing infrastructure. the weird part is that there isn't clear evidence of this in packet capture. I suppose it could be one-way (ie. on upload only)
Yeah i guys i dont know. I mean, i was always aware hat is has to be an network issue and tried so many things. Upgraded to bussiness line with a static IP adresse. Switched from 3 pay to 2 pay in the hope, that the interleaving depth will change, since fastpath is not a viable option anymore. I also thought that my the error correction is fault, why i wanted to have fastpath to turn that off and it is possible on VDSL.
I had contact to a employee from my ISP, where i told him this and he told me that everyone has the same interleaving depth between 2 pay and 3pay contract with TV.
Before i tried VPN on cable, i upgraded it with an static IP and wanted to ask if its matter ? Like do you gain better routing with a static one between to a dynamic one ? Like is there difference or is it necesarry ?
Is there probably a way to force somehow fastpath ? To the contributor threw telnet/SSH ?
I bought once a asus router because it has more options to turn things on and off, like G.INP. i tried that and gameplay was the same. Ping got higher and my bandwith reduced by a big ammount.
No, that is decided by the DSLAM.
1 Like
Yes, the DLM system operates on lower level DSL statistics and DSLAM and modem chip are blissfully unaware of the PPPoE state, however resyncs/retrains of the DSL link do happen (among other reasons like a scheduled modem reboot by the end-user) if the link accumulates severe enough errors that the sync is lost (after which the CPE typically will automatically reconnect). So IMHO interpreting resyncs as indicators of potential stability issues is not a bad idea per se. Especially given that almost all users will notice a link that resyncs like crazy (breaking all on-going connections while doing so, and since most ISPs assign new IP addresses for every PPPoE re-connect, all of these connections will time out and will not automatically recover after sync and pppoe connection are re-established) while a (modest) reduction in link capacity will not be as obvious. (And to be clear as end-user I agree to that interpretation, a stable link is clearly superior to a faster but flaky one; however if an ISP promises rate X and DLM slows a link down significantly below X I would expect the ISP to either pro-actively try to debug and fix the issue causing the rate-downgrade or in case that is impossible/unfeasible to automatically offer the affected users a price rebate proportional to the lost fraction of capacity).
This has been observed by some users, extreme sensitivity, while others needed to collect a few resyncs in a short duration to tickle DLM into action, unclear to me what caused the difference. However deutsche telekom (the biggest operator of DSL links in Germany that also operates the DLM system, other ISPs mostly re-package such links so most german DSL links inherit DLM) has been completely silent on the topic and there is no way to figure out how close a given link is deemed by DLM to requiring intervention, could be that your link was only "marginally" intervention free, while others requiring multiple resyncs had a larger "stability" margin (as seen by DLM). But that is mostly speculation
1 Like