Bufferbloat advices

[fds]

Is it normal that I have to significantly reduce my speed in order to achieve an A rating for bufferbloat?

 

Using my maximum speeds of 1 GB and playing with the slayers, it's impossible for me. The only way I've managed to do it is by reducing my speeds.

[DARKNESS]

@fdsYes, that's normal; that's what QoS does. It manages your packets to prevent any queue buildup. What do you mean by "slayers"? You don't need more than 5 Mbps to play a game.

[fds]

I meant ‘sliders,’ sorry.

I was referring to congestion control.

If I try to get an A+ bufferbloat, it's impossible using only congestion control. I have to manually reduce my speeds to get an optimal result.

 

What I've noticed, as you say, is that even when I reduce the speed, everything runs more smoothly and fluidly in games. Therefore, I think I'm heading in the right direction. Any ideas that could help me achieve a good configuration in R3, especially focused on improving fluidity in games, are welcome.

 

I usually play FC26.

 

The truth is that I'm still clumsy with the R3 configuration. Even though I've been using it for a few months now, I feel like I don't know how to get good performance out of it.

 

 

[DARKNESS]

@fdsWhen performing these tests for bufferbloat, ensure you follow the instructions correctly, such as connecting directly to the LAN and ensuring no one is running any background activities, as this could introduce margins of error. It’s not entirely necessary to cap your speeds, but if you want the best experience with no packet loss or latency increase in-game, where it’s most noticeable, it might be worth considering. From my understanding, FC26 isn’t necessarily a hit registration game, plus I don’t play those types due to their “pay-to-win” nature—but to each their own. 

To see this in real-time, perform a bufferbloat test while playing the game. For your specific game, go to Settings, select Visuals, and enable Connection Monitoring. Follow the instructions provided in the video. I mention this because of the placebo effect—you want to ensure you're observing actual results, not influenced perceptions. - https://www.youtube.com/watch?v=4q_-afXn-8U

• Which firmware version are you currently using?
• Are you saying that congestion control keeps you from dropping below 50%? You can easily adjust the speeds in Speedtest Advanced to achieve the same outcome.

[fds]

I have carried out the tests correctly, and although it is not necessary to reduce my speeds, no matter how much I use CC, I cannot achieve decent results. No matter how many adjustments I make, the best I can achieve is a B-C rating. The only way to obtain good results is by lowering my speeds from 1 GB to 500-200. From there, the waveform test starts to give A-A+ results (this is what my query referred to). In my case, I have only achieved this in this way. On the other hand, as I said before, playing this way, I noticed that the matches were smoother and I had better overall response from the actions in DualSense (the skills require a lot of action, and when the network environment is not smooth, the experience becomes very irritating). But anyway, it seems that this could be helping, although EA manages with its servers to give you a terrible experience once again and see if you fix it by paying for items.

 

- What you mention in the video, if I do a bufferbloat test with the game's latency indicators turned on, what do I get out of that? All I'm going to see are bad connection messages... I don't understand what that achieves, what can I check in real time?

 

- The firmware version is the most recent public one, I'm not on the beta at the moment.

 

- The last thing you said is what I think I've done, manually adjusting the speed in speed advanced, which is what has given me the best results.

[DARKNESS]

@fds

By monitoring the game during the bufferbloat test, you can detect abnormal spikes or packet loss. If such issues occur, continue fine-tuning the settings. The goal of SQM (Smart Queue Management) is to optimize the network, prevent overloading, and maintain consistent latency. Every connection is unique; for instance, my connection has 2G download and 400 Mbps upload speeds. To ensure stability without packet loss, I adjust it to 180 Mbps download and 60-80 Mbps upload. This setup consistently earns me an A+ rating, even during ISP congestion. Throughput is less critical for gaming; latency and stability are the key factors. While other elements may influence performance, the most important ones are within your network, which you can control.

It seems that you want full throughput when you're not gaming, which is achievable through the router settings. Simply select the low latency option, and it should only be activated while gaming, if I recall correctly.

[fds]

Okay, I understand what you're saying... I'll continue testing in this direction. Thank you for your input!

[fds]

The results are consistent in monitoring, there is hardly any jitter or fluctuation, and the result is A+.

 

Another question that comes to mind... in the bandwidth reservation options in SmartBoost, does this function act as a limiter? In other words, if I reserve only 20 Mbps, is that all it will offer to the application that applies it, or is that not how it works?

[DARKNESS]

@fdsThey no longer use a limiter; previously, DumaOS 3 featured a dedicated limiter, but it proved to be highly inefficient. Instead, they have shifted to using reserves, which, in my opinion, still lack effectiveness. Reserves essentially allocate bandwidth for a specific application or the maximum possible during usage when applied. On a side note, if a game isn't detected by SmartBoost, I recommend switching to "always," as some games are unsupported or have outdated DPI. In such cases, congestion control won’t activate properly since it relies on SmartBoost to enable congestion management, but having it on always essentially mitigates that issue, so it is preferred but it a preference. 

The admin will be available tomorrow if you need further assistance or today.

[Netduma Fraser]

56 minutes ago, fds said:

The results are consistent in monitoring, there is hardly any jitter or fluctuation, and the result is A+.

 

Another question that comes to mind... in the bandwidth reservation options in SmartBoost, does this function act as a limiter? In other words, if I reserve only 20 Mbps, is that all it will offer to the application that applies it, or is that not how it works?

No it doesn't act as a limiter per se but it guarantees that should the device need that amount (it never would for gaming) then it will get it, even if something else is trying to use up all your bandwidth

[fds]

3 minutes ago, DARKNESS said:

@fdsYa no usan un limitador; anteriormente, DumaOS 3 presentaba un limitador dedicado, pero demostró ser altamente ineficiente. En cambio, se han desplazado al uso de reservas, que, en mi opinión, todavía carecen de efectividad. Las reservas esencialmente asignan ancho de banda para una aplicación específica o el máximo posible durante el uso cuando se aplican. En una nota al margen, si SmartBoost no detecta un juego, recomiendo cambiar a "siempre", ya que algunos juegos no están soportados o tienen DPI obsoleto. En tales casos, el control de la congestión no se activará correctamente, ya que depende de SmartBoost para permitir la gestión de la congestión, pero tenerla siempre esencialmente mitiga ese problema, por lo que se prefiere pero es una preferencia. 

El administrador estará disponible mañana si necesita más ayuda.

Thank you for your contributions, they are very helpful.

[fds]

1 minute ago, Netduma Fraser said:

No, no actúa como un limitador per se, pero garantiza que si el dispositivo necesita esa cantidad (nunca lo haría para jugar), entonces lo obtendrá, incluso si algo más está tratando de usar todo su ancho de banda

I understand, it seems like a somewhat different concept from a limiter. On the other hand, it has been mentioned that if this function previously existed in limiter mode, is it possible that it still exists and can be tested with some older firmware?

[Netduma Fraser]

6 minutes ago, fds said:

I understand, it seems like a somewhat different concept from a limiter. On the other hand, it has been mentioned that if this function previously existed in limiter mode, is it possible that it still exists and can be tested with some older firmware?

There has never been a limiter on DumaOS 4 which the R3 launched with right away so no that's not possible. The closest thing to limiting is what you have done by reducing the speed you've told the router you have

[fds]

Ok, I'm understanding, thank,s for contributions!

[fds]

On 10/12/2025 at 4:33 PM, DARKNESS said:

@fds

By monitoring the game during the bufferbloat test, you can detect abnormal spikes or packet loss. If such issues occur, continue fine-tuning the settings. The goal of SQM (Smart Queue Management) is to optimize the network, prevent overloading, and maintain consistent latency. Every connection is unique; for instance, my connection has 2G download and 400 Mbps upload speeds. To ensure stability without packet loss, I adjust it to 180 Mbps download and 60-80 Mbps upload. This setup consistently earns me an A+ rating, even during ISP congestion. Throughput is less critical for gaming; latency and stability are the key factors. While other elements may influence performance, the most important ones are within your network, which you can control.

It seems that you want full throughput when you're not gaming, which is achievable through the router settings. Simply select the low latency option, and it should only be activated while gaming, if I recall correctly.

I have another question, I'll post it here so as not to create another thread.

 

As I often play in party chat on PS5 with friends and usually share my screen in the same party room, I always notice that the matches have more latency and lag on the players I control, even though I have previously tuned the waveform test to get a good A+ rating. Given this, what is the best way to tune the test so that I can improve the conditions while I am in an audio group and sharing my screen?

Can Smartboost detect PS5 party chat in some way (audio calls, etc.)? For screen sharing, the same question applies: what would be the appropriate setting?

And would it be advisable to try to fine-tune the waveform test, not only while I'm playing, but also while I'm in audio and streaming?

 

 

[Netduma Fraser]

1 hour ago, fds said:

I have another question, I'll post it here so as not to create another thread.

As I often play in party chat on PS5 with friends and usually share my screen in the same party room, I always notice that the matches have more latency and lag on the players I control, even though I have previously tuned the waveform test to get a good A+ rating. Given this, what is the best way to tune the test so that I can improve the conditions while I am in an audio group and sharing my screen?

Can Smartboost detect PS5 party chat in some way (audio calls, etc.)? For screen sharing, the same question applies: what would be the appropriate setting?

And would it be advisable to try to fine-tune the waveform test, not only while I'm playing, but also while I'm in audio and streaming?

The bufferbloat test isn't the best way to determine values, follow this guide https://support.netduma.com/frequently-asked-questions/legacyfaqs/test-your-ping/ while downloading & start with 95% for Congestion Control (set to Always), check results, decrease by 10%, check, decrease by 10% etc, until you get to a value that is pretty good & then try 5% either side of that value to see if it can be improved. Download & Upload on Congestion Control don't have to be the same value & you may have a better experience with differing values.

Doing the screensharing is more bandwidth intensive, that's why following this guide and instructions is better as you make your network essentially as bad as it could possibly get and then apply settings incrementally to determine what best brings your network as close as possible to its baseline.

[fds]

33 minutes ago, Netduma Fraser said:

La prueba de bufferbloat no es la mejor manera de determinar valores, siga esta guía https://support.netduma.com/frecuently-asked-questions/legacyfaqs/test-your-ping/ mientras descarga y comienza con 95% para Control de Congestión (configurado en Always), comprobar los resultados, disminuir en un 10%, comprobar, disminuir en un 10%, etc, hasta que llegue a un valor que es bastante bueno y luego probar el 5% de cada lado de ese valor para ver si se puede mejorar. Descargar y cargar en Control de congestión no tiene que ser el mismo valor y puede tener una mejor experiencia con diferentes valores.

Hacer el uso compartido de pantalla es más intensivo en ancho de banda, es por eso que seguir esta guía e instrucciones es mejor, ya que hace que su red sea esencialmente tan mala como podría ser posible y luego aplicar configuraciones gradualmente para determinar qué mejor lo que lleva su red lo más cerca posible de su línea de base.

Okay, I'll try to follow this guide and see what results I can get. So you're saying that I should change the way I test bufferbloat while gaming and screen sharing, for this broader saturation test, and stick with the most optimal values I get. Will this have a better effect while I'm doing several things at once?

 

On the other hand, in SmartBoost, can it detect and improve something in that case if I set priorities or some bandwidth reserve for sharing and audio?

 

I don't know, I think about the variety of paths to follow and it's a bit complex to determine which one is the most optimal.

 

On the other hand, I also have to do (when SmartBoost detects FC26 as a game) search in statistics for experts and see the IP (I suppose it will not be the same as the server's) but where I connect while playing and do a tracert to see how my company communicates with the servers. I've done some with the current data and in jump 3 there's always an error like: * * * and I'm not quite sure how to interpret it or where the problem lies.

 

 

[Netduma Fraser]

49 minutes ago, fds said:

Okay, I'll try to follow this guide and see what results I can get. So you're saying that I should change the way I test bufferbloat while gaming and screen sharing, for this broader saturation test, and stick with the most optimal values I get. Will this have a better effect while I'm doing several things at once?

On the other hand, in SmartBoost, can it detect and improve something in that case if I set priorities or some bandwidth reserve for sharing and audio?

I don't know, I think about the variety of paths to follow and it's a bit complex to determine which one is the most optimal.

On the other hand, I also have to do (when SmartBoost detects FC26 as a game) search in statistics for experts and see the IP (I suppose it will not be the same as the server's) but where I connect while playing and do a tracert to see how my company communicates with the servers. I've done some with the current data and in jump 3 there's always an error like: * * * and I'm not quite sure how to interpret it or where the problem lies.

Yes absolutely. 

It could potentially but if you're already prioritizing the gaming device then it would already be covered.

Game servers don't always respond to pings and like game servers not every hop along the route to the server you're testing will respond to pings either, that's what you're seeing there, it isn't an issue or anything that needs to be resolved, you can just ignore that entry. 

[fds]

6 hours ago, Netduma Fraser said:

Sí absolutamente.

Podría potencialmente, pero si ya está priorizando el dispositivo de juego, entonces ya estaría cubierto.

Los servidores de juegos no siempre responden a los pings y, como los servidores de juegos, no todos los saltos a lo largo de la ruta hacia el servidor que está probando, responden a los pings, eso es lo que está viendo allí, no es un problema ni nada que deba resolverse, puede ignorar esa entrada.

Yes, I'm prioritising the PS5 on devices, but there's a big difference when I play games on my own compared to when I do so with audio chat and streaming (I suppose the cause is more to do with streaming, although I notice it overall), so I'm trying to find a way to mitigate this and be able to play with the same fluidity as when I play on my own. That would be a good point.

[Netduma Fraser]

I think fine tuning Congestion Control in the way I mentioned would get you a better result, try that out and if not we can try to dive deeper into it