Gargoyle Forum

Daeron wrote:By adding the IP specific rules you are also entirely circumventing the rest of the setup in question. I'd suggest removing them, testing again and sharing the results, then if you feel like it you can readd them. Although if those upload rules were invalid like fifonik said, you might have tested it "the right way" by accident.

What i wanted is to separete rest of setup from my XBOX ONE X because i notice when i download games at my console the laptop had some latency issue at browsing websites because of the 443 port that was reserved for Console downloading.

Also would like to mention that i notice by disabling the ACC i had better results.
With ACC ON my Fair Link Limit was droping down over time. 3 days i turn it OFF and my console was downloading games during that perion (400 GB+) and i hadnt any issue with surfing or downloading with laptop/smartphone neither playing on-line game with Console.

Although with invalid upload rules didnt notice anything and this is maybe due to i didnt push it to the limit the whole system so cannot confirm.

Hope all this make any sense as i mentioned that my knowledge is limitted and the setup of system and the tests that i perform maybe be wrong and

My system include SMART TV watching YOUTUBE, CONSOLE to play on-line game (World of Tanks) and download some (thanks XBOX GAME PASS), LAPTOP browsing and downloading and Smartphone.
All of them connected via WiFi 5GHz Operation Mode
So when i perform test i open them all and do what i do usualy.

I spent some more time researching and experimenting. These are my current rules:



The only notable change is that I merged the first two download classes into one without losing much functionality, as I theorized earlier. Packet lengths were adjusted slightly too, I'm still trying to settle on exact numbers, but I think they are roughly correct. The only other possible improvement I'm thinking of is seeing how far the middle class (~500 byte packets) can be pushed without getting false positives or degradation of performance ingame.

I also have some graphs below to give context why I ended up with this ruleset (I already explained the reasoning/logic behind it in my previous posts, now I'm just showing the data I used). The two lines on each picture show the minimum and maximum packet sizes during each activity.

Upload side traffic

File downloads/Video streaming/Torrent downloads:
They tend to produce an excessive number of ACKs (usually <78 byte packets). Interestingly there's also periodic spiking to 1500 bytes. I'm not sure why yet.





Gaming:
The overwhelming majority of the packets are under ~500 bytes, however some occasional spikes can go multiple times as high. None of the games I tested ever wen't above ~1300, which is an easy way to differentiate between game traffic and downloads. Some games also tend to use really tiny packets (same range as ACKs), which is somewhat troublesome (although the pings remained fine during my testing).







Light browsing:
Mostly ACKs and spikes to somewhere under ~1300 bytes.



File uploads/livestream upload/torrent upload traffic:
As with any other heavy payload, 1500 is used primarily. The smaller packets are probably due fragmentation and they aren't using much bandwidth.


Download side traffic

File downloads/Video streaming/Torrent download traffic:
They can be easily classified by their tendency of using the largest available packet size (1500).






Light browsing:
Seems to be a combination of medium sized (<500) packets (text/code?) and heavy payloads (1500), the latter presumably being images and other media.



Gaming:
Sizes vary a LOT based on the game and number of players. All the games I tested however seem to cap out at somewhere below ~1300 bytes.







Uploading files/livestreaming:
Just a stready stream of ACKs.

I was going to create a new topic, but the subject is basically the same. These are my settings, maybe it helps someone.