ASUS RT-N56U beta firmware

[Mat77]

Laughable. A non-existant problem perpetuated by people with cheap routers and bad ISPs. None of the "experiments" on that site gave me any trouble and nor would they give anyone trouble if they have a properly set up network, a decent ISP and a decent router. Even the Wikipedia entry for this nonsense is full of "citation needed" markups - because it's a load of rubbish.

For info, my setup:

ISP: Virgin Media, 60Mbit/3Mbit
Modem: Netgear VMNG480 in "modem only" mode
Router: ASUS RT-N56U with firmware 3.0.0.4.334, hardware NAT enabled
Switch: Netgear GS108E gigabit switch
PC: Win 7, NVidia nForce gigabit ethernet controller

Pingtest.net reveals 15ms or less average over 50 miles, with 2-3ms jitter.
Downstream max. out at around 62Mbps (7.7MB/s), upstream at 3Mbps (375KB/s)

Similar results to any fast enough server - I can ping all day long and not see this so-called "bufferbloat".

Any bad latency is down to overloaded servers, servers on slow connections, physical distance, or ISP faults ("faults" include deliberate shaping in this case).

If you are having latency issues in games or have trouble streaming a youtube video, you can be assured the problem is not the mythical "bufferbloat". Either your kit, or your ISP, or the server you use, or all of the above, are shoddy or broken - or you are just too far from the server for it to be possible to have a low-latency and stable connection. Physical distance is a barrier you won't defeat with current consumer technology.

maybe you don´t have Bufferbloat.
But how can you say it doesn´t exist?

Can you please make this test? (Java needed)
http://netalyzr.icsi.berkeley.edu/ and tell me the results?

If they say on that test you have Buffebloat on your upload. Is it wrong?

If it doesn´t exist why is something to fight Buffebloat implemented in Linux Kernel 3.3 and above?

and if you read things like that
http://blog.thelifeofkenneth.com/20...he+Life+of+Kenneth)&utm_content=Google+Reader

They are all wrong?

ISP: Kabel BW , 50Mbit/2,5Mbit (Cable Provider) very good ping times all over the world. With this setup I get an Bufferbloat of 809 ms !

So I upgraded to:
ISP: Kabel BW, 100 Mbit/5Mbit (more is not possible at the moment).

The Bufferbloat decreases to 400 ms on uplink. Maybe it´s not my Router or my Modem it can be something on my Provider. But to say this Problem doesn´t exist will not help.

Please do the test and show the results. And thenn tell me if the test is a lie.

 

[Mat77]

IIRC bufferbloat is an effect that occurs when uploading at max speed where your download speed goes down and ping times go up.

Simple solution, use a bandwidth limiter.

fq_codel will help in the case where you're uploading at max speed but that's about all it will do.

with custom Firmware not possible? No QOS Feature inside. What can I do?
Doesn´t QOS add latency?

 

[ceeKayy]

The test is stupid to do (and I will never do it, as I will never use java on my computer)...basically it's like this 1) you're using QoS and have implemented 60% of your upload max so you will never max out your upload speed to keep low latency during uploads or 2) you rather have a faster network by not using QoS and having HW_NAT on.

If you really think bufferbloat is that much of a problem then, by all means, dump the rt-n56u and get a cheap netgear router that is 4x slower than the rt-n56u and put your CeroWRT firmware on it and have the problem with a router with 350Mbps throughput rather than the 56u with 802Mbps throughput...but hey, you'd rather have no "bufferbloat" than a whole network so you can watch your "youtube videos on bufferbloat that will buffer every 5 seconds because you poisoned your connection to improve latency". But to each his own, I guess.

 

[ceeKayy]

ISP: Kabel BW , 50Mbit/2,5Mbit (Cable Provider) very good ping times all over the world. With this setup I get an Bufferbloat of 809 ms !

So I upgraded to:
ISP: Kabel BW, 100 Mbit/5Mbit (more is not possible at the moment).

So you upgrade your internet? How is upgrading your internet going to solve your "problem"? You have the same ISP so upgrading wouldn't really benefit you unless you were upgrading to another ISP, because it's still the same ISP configurations...just different speeds...

and an FYI...your "bufferbloat" problems that everyone "fixes" by using QoS is setting the max upload bandwidth to 60% of your current upload...so your upgrading internet will have to be throttled back to 3Mbps, which is what you had before so the upgrade isn't really in your favor besides costing you more money...and the old speeds your upload would have to be slowed down to 1.5Mbps to cure your "bufferbloat"...and download speeds are irrelevant in this situation.

 

[ceeKayy]

and if you read things like that
http://blog.thelifeofkenneth.com/20...he+Life+of+Kenneth)&utm_content=Google+Reader

They are all wrong?

And if you read that article you would see that "This isn't only happening in home modems. Routers, switches, Ethernet cards, operating systems, wifi drivers, critical ISP infrastructure"

So why are you so set on your router and nothing else, when the majority lies most likely with ISP configurations?

 

[Mat77]

So you upgrade your internet? How is upgrading your internet going to solve your "problem"? You have the same ISP so upgrading wouldn't really benefit you unless you were upgrading to another ISP, because it's still the same ISP configurations...just different speeds...

and an FYI...your "bufferbloat" problems that everyone "fixes" by using QoS is setting the max upload bandwidth to 60% of your current upload...so your upgrading internet will have to be throttled back to 3Mbps, which is what you had before so the upgrade isn't really in your favor besides costing you more money...and the old speeds your upload would have to be slowed down to 1.5Mbps to cure your "bufferbloat"...and download speeds are irrelevant in this situation.

and why does it decrease?

 

[Mat77]

And if you read that article you would see that "This isn't only happening in home modems. Routers, switches, Ethernet cards, operating systems, wifi drivers, critical ISP infrastructure"

So why are you so set on your router and nothing else, when the majority lies most likely with ISP configurations?

because I will be sure, that I have done all I can !

What do you get as a result if you do the test?

 

[ceeKayy]

and why does it decrease?

Because the QoS bufferbloat "fix" is by using QoS to "degrade" your upload to fool the router into thinking you only have 60% of what you actually have...so the router maxes out at that set upload speed rule...thus making your computer appear to have a better latency during that time where your uploading all that data...

And I said before I will not do the test because I will never use Java and never plan to use Java until they fix all their vulnerabilities.

I do alot of upload and multiplayer gaming on both PC and xbox360 as well has video editing and I can upload a 2.5GB file to youtube 1080p and play 360 at the same time on Call of Duty and have no problem with lag NOR BUFFERBLOAT...

 

[ceeKayy]

because I will be sure, that I have done all I can

Frankly, if you are so dead set about bufferbloat your answer will be either setup the router and use it or...get another router, like I said, because bufferbloat isn't introduced into firmwares for specific routers until kernal 3.4.x I believe somewhere around there and I don't see custom firmware Padavan implementing that kernal that high up being as the latest firmware only has linux kernel 3.0.61 and the kernel only really upgrades at a low pace like 3.0.x at a time and I don't see getting kernel 3.4+ anytime soon...

So you can get a cheaper router that supports CeroWRT like some Netgear routers if you feel the router is the culprit.

 

[Mat77]

Because the QoS bufferbloat "fix" is by using QoS to

video editing and I can upload a 2.5GB file to youtube 1080p and play 360 at the same time on Call of Duty and have no problem with lag NOR BUFFERBLOAT...

with Padvans Firmware you can´t use QOS. And if you upload your ping gets worse and for Call of duty this means you get very big advantage because of lag compensation ! And so you prove it, that your ping is very bad while gaming !

Its known !
"MW3 is great now! you just have to turn on some torrents so the LAG fix garbage doesn't ruin you!"

Its because the better your ping the more you get compensated.

Lag compensation explained:
https://www.youtube.com/watch?v=reNPK8XD-k4

 

[icarus]

Have you tried the latest release of Padavan's firmware? There's a thread on this subforum about it, and you can download it here:

https://code.google.com/p/rt-n56u/

Worth a try to see how well it works for you before you go back into the future *smile*.

I'm actually running now as of one hour ago. So far so good minus the sound being pushed to the appletv but that's I think more of an OSX bug. Took a while getting the airplay speaker set up without WPS (No ethernet port and I stupidly didn't save the exact config of the network before... ehh) so I shall report back.

By the way, the Wiki isn't all that great but do you have any further explanation on what Torrent Transmission is?

 

[Dave Trouser]

Unfortunately, I use a lot of apple devices. Apple TV, and an airprint printer, and I'll often mirror the display of my Mac to my Apple TV and both my wife and I print using airprint. For no reason that I can think of, the router will allow us to say, mirror a display for anywhere between 5 minutes up to 3 hours and then *pop*. Everything else works as normal, except anything to do with bonjour or airplay drops.

Disable "IGMP Snooping". Should fix all Apple related woes. Found under Wireless -> Professional settings.

 

[Dave Trouser]

IIRC bufferbloat is an effect that occurs when uploading at max speed where your download speed goes down and ping times go up.

Simple solution, use a bandwidth limiter.

fq_codel will help in the case where you're uploading at max speed but that's about all it will do.

If you upload at maximum speed, your download speeds will naturally decrease because downloading requires packets to be sent upstream at the same time (ACK etc). If you have saturated your upstream, then you aren't going to be able to download at full speed. Likewise, if you have saturated your downstream bandwidth, your upstream is going to be similarly impacted.

There is no "bufferbloat" going on here - it is a non-existent term created by people who don't understand how the Internet or networks in general actually work.

Using QoS will of course help reduce congestion, because it will reserve bandwidth to allow enough capacity for packet overheads and prevent saturation of your up/downstream bandwidth. But if you download a large file at maximum speed, then you will not achieve maximum upload speed at the same time. They will impact on each other because the download will be generating upstream packets at a rate proportional to the download speed, and the upload will generate downstream packets at a rate proportional to the upload speed. This is not some flaw called bufferbloat - it is how things work. It is just more obvious on asymmetrical connections (ie, where your downstream bandwidth is far greater than the upstream, eg 60Mbit down, but only 3Mbit up). If you had symmetrical bandwidth, you may notice it less.

This is why QoS ws created! It is just good sense to make use of it. Anyone expecting maximum sustained download speed at the same time as maximum upload speed is a moron.

 

[Mat77]

If you upload at maximum speed, your download speeds will naturally decrease because downloading requires packets to be sent upstream at the same time (ACK etc). If you have saturated your upstream, then you aren't going to be able to download at full speed. Likewise, if you have saturated your downstream bandwidth, your upstream is going to be similarly impacted.

There is no "bufferbloat" going on here - it is a non-existent term created by people who don't understand how the Internet or networks in general actually work.

Using QoS will of course help reduce congestion, because it will reserve bandwidth to allow enough capacity for packet overheads and prevent saturation of your up/downstream bandwidth. But if you download a large file at maximum speed, then you will not achieve maximum upload speed at the same time. They will impact on each other because the download will be generating upstream packets at a rate proportional to the download speed, and the upload will generate downstream packets at a rate proportional to the upload speed. This is not some flaw called bufferbloat - it is how things work. It is just more obvious on asymmetrical connections (ie, where your downstream bandwidth is far greater than the upstream, eg 60Mbit down, but only 3Mbit up). If you had symmetrical bandwidth, you may notice it less.

This is why QoS ws created! It is just good sense to make use of it. Anyone expecting maximum sustained download speed at the same time as maximum upload speed is a moron.

Like I understand the Problem is not download speed but latency. There is a "hack" around where you can reduce txqueuelen from 1000 to 10. The latency gets better because instead of Buffering your packets they get dropped.

how can you say that?

"There is no "bufferbloat" going on here - it is a non-existent term created by people who don't understand how the Internet or networks in general actually work."

I think Linux Kernel Coders know something about Network and Internet. Also the People from Google who help to reduce the Problem.

When I look on my System/Memory Usage (Padvan Firmware) my Buffer is 2.35 MB. But I don´t know if it fills and if not. with the test Netalyzr the can fill it and and I get a delay of 409 ms whe filled.

So why I´m intersted is: Latency because I´m a Gamer and want to have the best latency ALL the time.

It´s the same like HDTV Monitors. They also create Input LAG. Someone started to say this and many peoples say - that´s wrong and so on. Today noone argues about Input LAG in HDTV´s. A Gamer Monitor is much better.

I think here it is the same.
And ASUS uses Linux Kernel for their Firmware and it would be cool (I don´t know if it is possible) if they can implemet a Linux Kernel 3.3 or above. If it helps for the Problem - good. If not - also good.

But to do nothing is bad.

Sorry for my english

 

[icarus]

Disable "IGMP Snooping". Should fix all Apple related woes. Found under Wireless -> Professional settings.

Thanks, I'll give this a go if the custom firmware doesn't yield anything. I'm assuming it just fills up and falls over due to the way Apple sends its data stream?

 

[Dave Trouser]

So why I´m intersted is: Latency because I´m a Gamer and want to have the best latency ALL the time.

Well there's a problem right there. You can never have the lowest latency ALL the time unless you are on the same local network as the server. Anything over the Internet is going to be unpredictable and beyond your control. Either use different servers that are physically closer to you, or get a PC and play games over a LAN. The further you are from a server physically, the higher the latency will get, along with (very likely) more nodes in-between you and the server, each adding their own unknown latency based on equipment spec and local network load.

Seriously, your goal of having the best latency all the time is delusional. If you are getting lag spikes, either you have a bad setup at home, the server is overused, or there is some issue in between, which no amount of tweaking will solve. A simple traceroute would reveal most issues with latency between you and a server. If that shows no problem, then it's either your local setup or the server.
You clearly won't be convinced, so it's time to end this - it's nothing to do with the subject of this thread.

It´s the same like HDTV Monitors.

No it is not. TV input lag is down to image processing - eg interpolating the frames to achieve a 120FPS moving picture from a 60FPS input. Any decent TV has a "game" mode that turns all that off. The input lag is also a fixed delay, and doesn't vary like lag over the Internet.
The fact you made this "comparison" shows that you have no clue what you are talking about.

 

[Dave Trouser]

Thanks, I'll give this a go if the custom firmware doesn't yield anything. I'm assuming it just fills up and falls over due to the way Apple sends its data stream?

I've never looked into the exact cause - I have no Apple products to make me care ;-) Just recall seeing a number of threads on this issue on various forums. From what I recall, it is an Apple flaw. This is unsurprising, since they are terrible with sticking to standards. Eg, the Apple TV never switches off the HDMI "active" signal even when it is in standby. This prevents other HDMI devices from making most TVs auto-switch to their input.

I recommend a Google search for "IGMP Snooping interferes with Airplay" or similar. Should lead you to an explanation if there is one.

My guaranteed solution is: stop using shoddy Apple products ;-)

 

[jhoffy22]

Laughable. A non-existant problem perpetuated by people with cheap routers and bad ISPs. None of the "experiments" on that site gave me any trouble and nor would they give anyone trouble if they have a properly set up network, a decent ISP and a decent router. Even the Wikipedia entry for this nonsense is full of "citation needed" markups - because it's a load of rubbish.

For info, my setup:

ISP: Virgin Media, 60Mbit/3Mbit
Modem: Netgear VMNG480 in "modem only" mode
Router: ASUS RT-N56U with firmware 3.0.0.4.334, hardware NAT enabled
Switch: Netgear GS108E gigabit switch
PC: Win 7, NVidia nForce gigabit ethernet controller

Pingtest.net reveals 15ms or less average over 50 miles, with 2-3ms jitter.
Downstream max. out at around 62Mbps (7.7MB/s), upstream at 3Mbps (375KB/s)

Similar results to any fast enough server - I can ping all day long and not see this so-called "bufferbloat".

Any bad latency is down to overloaded servers, servers on slow connections, physical distance, or ISP faults ("faults" include deliberate shaping in this case).

If you are having latency issues in games or have trouble streaming a youtube video, you can be assured the problem is not the mythical "bufferbloat". Either your kit, or your ISP, or the server you use, or all of the above, are shoddy or broken - or you are just too far from the server for it to be possible to have a low-latency and stable connection. Physical distance is a barrier you won't defeat with current consumer technology.

I'm sure that you're familiar with split testing, correct?

I've ran tests using a direct connection to my router vs a direct connection to my modem (wired for both).

The modem visibly out-performs the router every time. Guess what? I've hooked up my 5 year old WRT-54G router and it performs on par with my modem.

All three tests were ran with no other devices connected to the network and during the same time period during the day over the course of a few days.

There IS a clear problem with Xbox live connectivity and latency. You may not notice it as much in games such as COD because of the lag compensation filter and the fact that because the game plays a lot different than a skill based shooter such as Halo. I am a Halo player. The router is the only variable in this equation. If I use my old router or go directly through the modem, the lag issues are non-existent. There is no arguing this.

Another issue with xbox live and this router - if you're using a computer at the same time that you're on xbox live and you decide to shut off your computer, it will kick you off of xbox live.

I've contacted Asus wireless support both on the telephone and online. They get the same complaint all the time about there being latency issues with xbox live. I wish people would stop acting like the problem doesn't exist.

 

[Mat77]

I'm sure that you're familiar with split testing, correct?

I've ran tests using a direct connection to my router vs a direct connection to my modem (wired for both).

The modem visibly out-performs the router every time. Guess what? I've hooked up my 5 year old WRT-54G router and it performs on par with my modem.

All three tests were ran with no other devices connected to the network and during the same time period during the day over the course of a few days.

There IS a clear problem with Xbox live connectivity and latency. You may not notice it as much in games such as COD because of the lag compensation filter and the fact that because the game plays a lot different than a skill based shooter such as Halo. I am a Halo player. The router is the only variable in this equation. If I use my old router or go directly through the modem, the lag issues are non-existent. There is no arguing this.

Another issue with xbox live and this router - if you're using a computer at the same time that you're on xbox live and you decide to shut off your computer, it will kick you off of xbox live.

I've contacted Asus wireless support both on the telephone and online. They get the same complaint all the time about there being latency issues with xbox live. I wish people would stop acting like the problem doesn't exist.

yes that´s it. But it´s not only the ASUS Router. I had an D-Link DIR655 before and because I felt that something is wrong I bought the ASUS Router. In earlier times as I had not so fast Internet the game experience was much much better. If you use your 5 year old Router game is better you said and I belive it. Unfortunately I can´t use an older Router. So I search the net for the problems. The only thing makes sense to me is the description of Bufferbloat. Like I said in earlier times the game feeling was much better - and now there are so much complaints of people with very fast Internet that their games lag - and it´s not "normal" lag - so it must have something to do with high speed Routers. And if I read that there is a Solution (Linux Kernel 3.3 or above) I want to try that solution. And because I have an ASUS Router and ASUS uses also a Linux Kernel I ask here if the Linux Kernel could be implemented. Or if something is implemented where I can set my Buffer for the Ethernetports.

Only to say it doesn´t exist is wrong ! I know that also the Routers of my Provider could be the Problem but i repeat me: I know that it was much better with non high performance routers and less speed.

 

[jhoffy22]

yes that´s it. But it´s not only the ASUS Router. I had an D-Link DIR655 before and because I felt that something is wrong I bought the ASUS Router. In earlier times as I had not so fast Internet the game experience was much much better. If you use your 5 year old Router game is better you said and I belive it. Unfortunately I can´t use an older Router. So I search the net for the problems. The only thing makes sense to me is the description of Bufferbloat. Like I said in earlier times the game feeling was much better - and now there are so much complaints of people with very fast Internet that their games lag - and it´s not "normal" lag - so it must have something to do with high speed Routers. And if I read that there is a Solution (Linux Kernel 3.3 or above) I want to try that solution. And because I have an ASUS Router and ASUS uses also a Linux Kernel I ask here if the Linux Kernel could be implemented. Or if something is implemented where I can set my Buffer for the Ethernetports.

Only to say it doesn´t exist is wrong ! I know that also the Routers of my Provider could be the Problem but i repeat me: I know that it was much better with non high performance routers and less speed.

1) There is only an issue with THIS router for me. No other router that I have used has given me issues. I've tried both an old Linksys and an old Netgear router.

2) It is not a problem with my ISP. Everything works fine and dandy directly through the modem, on a very consistent basis.