TECH ubiquiti wifi crew

TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
i didn't even bother to read the FAQs, the recommend wired testing, which i wanted to test since the console is hard wired

Should I test bufferbloat over ethernet or WiFi?

Ethernet is generally a better choice.

Here's why:

Bufferbloat happens any time data packets traveling to and from the Internet reach a buffer. Buffers are used to control the flow of data whenever there is a change in the bandwidth of the conenction, or "the size of the pipe." So, for example, a buffer is needed when a fast router is connected to a slow cable modem.

If you're connecting over WiFi, the "size of the pipe" actually changes three times. So there are actually three buffers. Looking at the download path, there is:

A buffer between your cable, DSL, LTE, satellite or fiber modem and your router;
A buffer between your router and the WiFi access point (which may both be in the same physical box);
A buffer between your phone or laptop's WiFi chipset and the actual computer.
By connecting with an ethernet cable, you're eliminating the second and third buffers, allowing you to specifically isolate your "modem to router bufferbloat." This is generally the most important buffer since it impacts every device on the network.

It's also worth testing WiFi bufferbloat on different devices. If you notice high bufferbloat, it may be worth upgrading either your access points and/or your devices.

**********************************

How does bufferbloat affect gaming?

Latency is incredibly important for many online games. When your network is suffering from bufferbloat, the latency will spike, causing noticeable delays, or “lag.” Severe bufferbloat will affect your performance against the other players and your enjoyment of the game. Despite this impact, many routers that advertise themselves as “gaming routers” lack the critical feature to address bufferbloat. Be sure to check out our router recommendations above.
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
ubiquiti edgerouter 4 made the list

What router can I buy to fix bufferbloat?

The quickest route to fixing bufferbloat is buying a router with a fast CPU and a great Smart Queue Management implementation.

Devices we recommend include:

The Eero Pro 5 mesh router with Eero Labs SQM enabled is the most user-friendly option that we've found to date. It can perform SQM on connections up to around 350 Mbps.
(Please note: the new Eero Pro 6 does not support SQM yet, though support is said to be coming soon.)
IQRouter IQRV3 - A router designed specifically to reduce bufferbloat
NETGEAR Nighthawk Pro Gaming XR1000 - Gaming router with DumaOS 3.0 software
Ubiquiti Edgerouter 4 - Enterprise grade router (no WiFi) with a more complex user interface, but excellent SQM performance
 
Last edited:

phat🐄

#meltdowns are NOT based off post count
Oct 9, 2002
43,652
Los Angeles, CA
i don't game a lot, i play 1-2 games with a friend a few times a week usually 1-2 hours before bed. when we play NHL 22 it feels extremely laggy on my end and i can't get any momentum in the game, the best way to explain it is latency. this was before i learned about bufferbloat (also posted the link in whatever's tmobile 5g threaf). maybe 1 in 20 games it feels fine like there is no lag. many years ago it never felt this way and many years ago i wasn't using pfsense where i currently live. anyway, i never thought that it was my pfsense router (or the pfsense OS itself) causing this because i don't notice this lag in COD. many times in COD i have great rounds and finish first it really seems to only be an issue with the NHL games over the last 3 years or so.

the only way to really find out is to change out the router and run this test with a new router to see if the issue really is with pfsense.

i think a few in here use pfsense, if so, can you run this test and tell me your score?

im hesitant to start jumping the gun based off that one test alone. esp since its a simple does latency go up when downloading???. most of the time, yes it does, for a variety of reasons. stupid test imho


waht hardware are you running, and what is the optimal throughput you see based off your filtering framework? do testzone to testzone1 accross interfaces and see what the ideal hardware max is.. do you ahve crypto enabled? Is it using the optmal method?
do you use pfBlocker? if so when do updates happen, and do they correlate with the states reloaded at the end of the updates. What does your hardware look like during a game?

what about QOS and port based flow control?

also drivers for your chipset and the tcp offload settings : did you check that?

MY connection shows a buffer bloat, 1 gig up and down on FIOS, and a core can push about 3gbs with my setup below
take niote of the state tble and buffer size. does that get maxed? is there anything that maxes during the sluggishness?

almost certainly its load you are not aware of if it is intermittent like that.

pK3M19Z.png
 
Last edited:

phat🐄

#meltdowns are NOT based off post count
Oct 9, 2002
43,652
Los Angeles, CA
hold up. is that solution page basically implimenting Qs on your wan interface?
:bowrofl: no fukking shit the throughput drops. youre shaping your traffic.


something is causing excessive load. none of those routers you listed can do what your device is doing at the moment, and id imagine its some cron that is runnign to make it shit itself periodically
so my Bufferbloat grade is D. I turned off Suricata (and pfelk (https://github.com/pfelk/pfelk) and ntopng and i get a B.


its load my man.
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
hold up. is that solution page basically implimenting Qs on your wan interface?
:bowrofl: no fukking shit the throughput drops. youre shaping your traffic.


something is causing excessive load. none of those routers you listed can do what your device is doing at the moment, and id imagine its some cron that is runnign to make it shit itself periodically
so my Bufferbloat grade is D. I turned off Suricata (and pfelk (https://github.com/pfelk/pfelk) and ntopng and i get a B.


its load my man.

i have 0 issues with pfsense before doing these bloat tests other then what i feel is latency during one online game, that's it.

yes, i'm not a fan of this 'limiter' stuff, but i'm also seeing a lot of people complain about pfsense and gaming.

again, 0 issues when i play COD when using a modern TV. when i play COD on a TV/projector w/o gaming mode there is noticeable latency/lag but i know that's because of the video refresh/etc. i can confirm this because a friend of mine with 1gb fiber NOT running pfsense has the same 'lag' issue in COD and it went away when we stopped using his projector and switched to a modern TV. he still uses the projector for movies, but not for gaming.

i'll post pfsense specs shortly, but i am using a netgate appliance.
 

phat🐄

#meltdowns are NOT based off post count
Oct 9, 2002
43,652
Los Angeles, CA
i have 0 issues with pfsense before doing these bloat tests other then what i feel is latency during one online game, that's it.

yes, i'm not a fan of this 'limiter' stuff, but i'm also seeing a lot of people complain about pfsense and gaming.

again, 0 issues when i play COD when using a modern TV. when i play COD on a TV/projector w/o gaming mode there is noticeable latency/lag but i know that's because of the video refresh/etc. i can confirm this because a friend of mine with 1gb fiber NOT running pfsense has the same 'lag' issue in COD and it went away when we stopped using his projector and switched to a modern TV. he still uses the projector for movies, but not for gaming.

i'll post pfsense specs shortly, but i am using a netgate appliance.
Ahhh then I’d assume the tunables are optimized at least for your hardware


U have the balls to go 2.6? I said Yooloooooo last week and upgraded and my gui lag issues went away. Almost certainly was pfblocker related since I have quite a few custom lists and geoblocked countries.
 

phat🐄

#meltdowns are NOT based off post count
Oct 9, 2002
43,652
Los Angeles, CA
Don’t get me wrong , I love qos and get unreasonably excited when my VoIP traffic is jitter stable and warble free , but to fix “buffer bloat” by adding another software layer on top of it is kind of counter productive.


Is buffer bloat testable from within a network ? Would love to try this with some Firepowers and Palo Altos I have at work in a controlled environment.



Also. Is buffer bloat a technical term ? Is this another way of saying saturation (capacity) ?
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
Ahhh then I’d assume the tunables are optimized at least for your hardware


U have the balls to go 2.6? I said Yooloooooo last week and upgraded and my gui lag issues went away. Almost certainly was pfblocker related since I have quite a few custom lists and geoblocked countries.

i'm on 22.01 and i don't use pfblocker on one of my pfsense boxes.


this pfsense box (stats below) is on 2.4.5 but will be on 22.01 pfsense + when i upgrade soon

System Netgate SG-5100
BIOS Vendor: American Megatrends Inc.
Version: V1.10_5
Release Date: Fri Jun 8 2018
Version 2.4.5-RELEASE-p1 (amd64)
built on Tue Jun 02 17:51:54 EDT 2020
FreeBSD 11.3-STABLE
CPU Type Intel(R) Atom(TM) CPU C3558 @ 2.20GHz
4 CPUs: 1 package(s) x 4 core(s)
AES-NI CPU Crypto: Yes (active)
Hardware crypto AES-CBC,AES-XTS,AES-GCM,AES-ICM
Kernel PTI Disabled
MDS Mitigation Inactive
4gb RAM
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
Don’t get me wrong , I love qos and get unreasonably excited when my VoIP traffic is jitter stable and warble free , but to fix “buffer bloat” by adding another software layer on top of it is kind of counter productive.


Is buffer bloat testable from within a network ? Would love to try this with some Firepowers and Palo Altos I have at work in a controlled environment.



Also. Is buffer bloat a technical term ? Is this another way of saying saturation (capacity) ?

i'm with you, i don't like the idea of having to add an algorithm to 'speed up' a system that has nothing else running at the time that i'm gaming. i do understand qos from the perspective of 'slowing down' nont critical traffic in order to prioritize VOIP traffic, for example.
 

phat🐄

#meltdowns are NOT based off post count
Oct 9, 2002
43,652
Los Angeles, CA
i'm on 22.01 and i don't use pfblocker on one of my pfsense boxes.


this pfsense box (stats below) is on 2.4.5 but will be on 22.01 pfsense + when i upgrade soon

System Netgate SG-5100
BIOS Vendor: American Megatrends Inc.
Version: V1.10_5
Release Date: Fri Jun 8 2018
Version 2.4.5-RELEASE-p1 (amd64)
built on Tue Jun 02 17:51:54 EDT 2020
FreeBSD 11.3-STABLE
CPU Type Intel(R) Atom(TM) CPU C3558 @ 2.20GHz
4 CPUs: 1 package(s) x 4 core(s)
AES-NI CPU Crypto: Yes (active)
Hardware crypto AES-CBC,AES-XTS,AES-GCM,AES-ICM
Kernel PTI Disabled
MDS Mitigation Inactive
4gb RAM
the new plus version has some intel crypto accelerator thing that exists in the 5100. I dont have plus since i dont have a netgate appliance (i have a Supermicro 1U with 10G cards)
get your sheit current first imo.
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
the new plus version has some intel crypto accelerator thing that exists in the 5100. I dont have plus since i dont have a netgate appliance (i have a Supermicro 1U with 10G cards)
get your sheit current first imo.

i will, my other box is on pfsense plus and it failed the test until i made those rules. the good news is that it is easy to disable the rule, not much to it, just check it and disable it, then it was like it was never there.
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
@phat🐄

the box that i added the limiter rule to

System Netgate 3100
Version 22.01-RELEASE (arm)
built on Mon Feb 07 16:39:01 UTC 2022
FreeBSD 12.3-STABLE
CPU Type ARM Cortex-A9 r4p1 (ECO: 0x00000000)
Multiprocessing, Thumb2, Security, VMSAv7, Coherent Walk
2 CPUs:
SOC: Marvell 88F6820, TClock 250MHz, Frequency 1600MHz
Crypto: Marvell Cryptographic Engine and Security Accelerator
Hardware crypto AES-CBC,SHA1,SHA256
2gb Ram

i suppose it is possible that the 'lag' is in my head, but it certainly doesn't seem that way, but who knows.....
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
@phat🐄

i don't have gigabit internet, but pfsense has no problem showing higher speeds than what i paid for, no concerns there. i do iperf tests over VPN but since my uploa is limited to 10mb, those test aren't very useful to me.

my friend has a rackmount pfsense appliance (netgate) (way overkill for his setup, but her wanted 1U to fit in his rack) and w/o doing anything to his network he gets a C on that test. he has gigabit and gets gigabit speeds (download) with his test, but his wifi blows. this is the friend that has too many APs in his setup and that's another issue i need to help him with, but that's not pfsense related.
 

phat🐄

#meltdowns are NOT based off post count
Oct 9, 2002
43,652
Los Angeles, CA
@phat🐄

the box that i added the limiter rule to

System Netgate 3100
Version 22.01-RELEASE (arm)
built on Mon Feb 07 16:39:01 UTC 2022
FreeBSD 12.3-STABLE
CPU Type ARM Cortex-A9 r4p1 (ECO: 0x00000000)
Multiprocessing, Thumb2, Security, VMSAv7, Coherent Walk
2 CPUs:
SOC: Marvell 88F6820, TClock 250MHz, Frequency 1600MHz
Crypto: Marvell Cryptographic Engine and Security Accelerator
Hardware crypto AES-CBC,SHA1,SHA256
2gb Ram

i suppose it is possible that the 'lag' is in my head, but it certainly doesn't seem that way, but who knows.....
what you got running under services. when you say lag, do you mean typical browsing? how do you do DNS? most of the time, the lag is caused by slow resolution


i use Unbound and Bind on PFSENSE, unbound simply is to do the PFSENSE rules, and BIND is what my LAN clinets hit, since i can do slaves up the wazoo.
 
TS
TS
TomDlgns

TomDlgns

OT Supporter
Oct 24, 2004
67,818
what you got running under services. when you say lag, do you mean typical browsing? how do you do DNS? most of the time, the lag is caused by slow resolution


i use Unbound and Bind on PFSENSE, unbound simply is to do the PFSENSE rules, and BIND is what my LAN clinets hit, since i can do slaves up the wazoo.

i use pihole, but the xbox uses 8.8.8.8 and 9.9.9.9 directly. however, i don't see how DNS resolution could cause lag in a game.

service are nothing special, just whatever comes by default. i use pihole, but pihole points to pfsense because the DNS resolver in pfsense has some specific entries. although, i could remove those entries from pfsense since the pihole now has the ability to have specific/custom DNS entires (intial pihole release didn't allow that).

client DNS goes to pihole, pihole blocks/allows domain then sends to pfsense for the handful of custom lookups pfsense knows about.
 

Users who are viewing this thread

  • About Us

    • Please do not post anything that violates any Local, State, Federal or International Laws. Your privacy is protected. You have the right to be forgotten. Site funded by advertising, link monetization and member support.
    OT v15.9.0 Copyright © 2000-2022 Offtopic.com
    Served by fu.offtopic.com

    Online statistics

    Members online
    433
    Guests online
    83
    Total visitors
    516

    Forum statistics

    Threads
    74,334
    Messages
    7,222,409
    Members
    86,944
    Latest member
    lisasalon