Netgear XR500 Upload speed...

[Stealth_01]

Hi all,

Been having a re-curring issue with regards to setting up my internet via the wizarf on the xr500.

For some reason it doesnt seem to be able to detect the upload speed automatically.

I constantly have to enter the values manually.

When I first bought the router it worked flawlessly, both upload and download speeds were detected just fine.

Any thoughts??

[Netduma Jack]

28 minutes ago, Stealth_01 said:

Hi all,

Been having a re-curring issue with regards to setting up my internet via the wizarf on the xr500.

For some reason it doesnt seem to be able to detect the upload speed automatically.

I constantly have to enter the values manually.

When I first bought the router it worked flawlessly, both upload and download speeds were detected just fine.

Any thoughts??

Hi Stealth - yeh a few folks have had this issue, but it's nothing to worry about. Manually entering your speeds works just the same - Netgear are aware of this happening in a few cases so there could be a fix in the future

Are you getting your full speeds?

[Stealth_01]

1 hour ago, Netduma Jack said:

Are you getting your full speeds?

Cheers Jack,

Yes I am, just glad a fixed is being worked on in the meantime.

Hopefully we get all the new additional features also with the next firmware update.

Regards

 

[xr500user]

you should send a message to the engineers to probably change the way this feature works..

if you think about it, it's a cool concept - but pretty unreliable.

There's been times where I've manually used the ookla speedtest and I've gotten 200 down - as soon as I see these types of results I instantly hit refresh and choose a different server and then I'm getting 900+ down. Nothing tops the speed test from your own isp (or it shouldn't in theory) they always test out at their max. But even here something can go wrong.

At a minimum it should do 3 tests and do an average (on different servers)  but also throw some sort of a note explaining that this may not reflect your actual speeds and could possibly effect your duma experience due to network congestion on your isp or the chosen testing servers side, and give the option to set your known speeds right then and  there at that time. ps. maybe it does this, I've never used it

Duma does take the results of this initial test as the word of god, so if you happen to have a crappy test, now your anti bufferbloat limiters will limit even more the crappy results, lol.

Preferably you should probably just do away with this and default to 1000MB and let the user decide their speed to start the duma experience in when they click the QoS area.  I know it's for the newbie users,  but most gamers are more techie then not. Seems like more problems then good come from this test, and rather err on the side of max bandwidth instead.  Once you get a feel for what you really get, then you can set the speeds and the sliders manually.

[Stealth_01]

55 minutes ago, xr500user said:

you should send a message to the engineers to probably change the way this feature works..

if you think about it, it's a cool concept - but pretty unreliable.

There's been times where I've manually used the ookla speedtest and I've gotten 200 down - as soon as I see these types of results I instantly hit refresh and choose a different server and then I'm getting 900+ down. Nothing tops the speed test from your own isp (or it shouldn't in theory) they always test out at their max. But even here something can go wrong.

At a minimum it should do 3 tests and do an average (on different servers)  but also throw some sort of a note explaining that this may not reflect your actual speeds and could possibly effect your duma experience due to network congestion on your isp or the chosen testing servers side, and give the option to set your known speeds right then and  there at that time. ps. maybe it does this, I've never used it

Duma does take the results of this initial test as the word of god, so if you happen to have a crappy test, now your anti bufferbloat limiters will limit even more the crappy results, lol.

Preferably you should probably just do away with this and default to 1000MB and let the user decide their speed to start the duma experience in when they click the QoS area.  I know it's for the newbie users,  but most gamers are more techie then not. Seems like more problems then good come from this test, and rather err on the side of max bandwidth instead.  Once you get a feel for what you really get, then you can set the speeds and the sliders manually.

Thanks for your input, some great suggestions.

I tend not to take the speedtest the xr500 detects as gospel anyway. 

I use three differant sites (dslreports/think broadband/ookla) just to make sure the speeds im recording dont fluctuate significantly.

I know u can enter these manually, but so far the xr500 seems to be detecting near enough the average ive been getting from these tests.

Its just annoying that everytime I choose to reset my internet and setup again I have to plug my laptop back into the router just to confirm the speeds again tbh.

Although as simple as this may sound, having to constantly unplug and disconnect all devices is a real pain in the rear.

Whats more concerning to me then all of the above is the fact that netgear still haven't released a firmware to patch certain bugs that have been highlighted by many users of their product.

Regards,

[xr500user]

I'm sure they are working on the new firmware, it's about that time for a release

I didn't mean for you to take it as the gospel, just that duma OS does - it bases its whole functionality around these UL/DL #s. Just to give an example if the built in test detects 50/40 erroneously and you have a 200/200 - if you limit it to 70% with the sliders -- duma is going to cut the ul and dl by 30% (of the 50/40) when anti-buffer bloat is triggered (either always, or when high priority traffic is detected) this could severely impact performance, especially if its a user that has no idea what they are doing.. that's why I rather set the values then let the OS do it on a possibly flawed test. The settings are right there in the QoS area for UL/DL speeds to be set to whenever you want.

An even better implementation (after defaulting to full 1000/1000 on initial first time setup) would be to give you an on-demand test in QoS so you can get an idea of what you truly pull, with a warning that it may not be accurate, and also when run it will disable all internet activity except for the LAN'd port/currently connected client (wifi) to the router during the test -- just temporarily pause all other internet traffic during the test (with a warning of course). Then you may get better / closer to true results with just router -> speedtest.net vs. a local device -> router -> speedtest.net with all other devices active possibly using your bandwidth at the time of the test.

It's best to just set what you get UL/DL in the QoS area no matter what anyway - I would not trust this test to be accurate no matter how well they implement it, hence why I've ignored it from the start.

I have a feeling ookla probably charges them something (maybe a few penny) each time the test is run -- so they might not want to give you on-demand testing (cheap?) And if they don't charge them they may be able to know that it's a netgear router test and do some sneaky sneaky to crap the result on you, lol. Just another reason I would not trust it.

[Stealth_01]

6 minutes ago, xr500user said:

An even better implementation (after defaulting to full 1000/1000 on initial first time setup) would be to give you an on-demand test in QoS so you can get an idea of what you truly pull, with a warning that it may not be accurate, and also when run it will disable all internet activity except for the LAN'd port/currently connected client (wifi) to the router during the test -- just temporarily pause all other internet traffic during the test (with a warning of course). Then you may get better / closer to true results with just router -> speedtest.net vs. a local device -> router -> speedtest.net with all other devices active possibly using your bandwidth at the time of the test.

Now that is something that could truly be of use imo. Having the router disconnect all devices temporarily whilst performing the speedtest would be ideal.

13 minutes ago, xr500user said:

I have a feeling ookla probably charges them something (maybe a few penny) each time the test is run -- so they might not want to give you on-demand testing (cheap?) And if they don't charge them they may be able to know that it's a netgear router test and do some sneaky sneaky to crap the result on you, lol. Just another reason I would not trust it.

What puzzles me though with regards to netgear having to pay Ookla a fee, is the fact that the download test performs fine?!? It just happens to be the Upload im having issues with. Which leads me to believe that this may be a firmware bug?!? Although i could easily go back to a previous firmware to check I really cannot be bothered lol. As Jack's already stated, Netgear are aware of this and hopefully the next update patches all the bugs.

Regards,

[xr500user]

I wouldn't worry about it as long as your upload is working and reasonably close to what you pay for.

It could be anything really, there are so many players in the mix. It may be a netgear bug, it may be your isp playing games since they are loving the recent advances in packet inspection. they can tell whatever you are doing for example if its http traffic, vpn traffic, social media, or speed test traffic.  they could if they wanted put a cap on packets that are speed test if they are not going to your isp's own speed test server.  it could be the servers that were selected for the test have a bad relationship with your isp and are capping those packets, it could be network congestion, it could be a mis-configuration on your in house network or the modem the xr500 plugged into activating its QoS, etc, etc. you are sure when you run the test its just pc->xr500->provider modem/ont/cpe connected?  all the more reasons not to put any faith in the test.

for example here verizon definitely knows when a speed test is run and it removes all packet caps when its going to their in-house server to give you the best result each time .. right after the test is done, isp qos starts again, lol.  if your one of those people who run speed test after speed test eventually it's going to limit you for a while somewhere some router AI will pick up on it.  ookla gets their revenue from all the ads around the speed test, so i'm sure they are paying something for it .. and if not that could be the reason why too, ookla said wtf you using our speed test for and not giving us a cut, heres your UL bw: 0. lol.

[Guest Killhippie]

On 12/11/2018 at 2:51 PM, xr500user said:

I have a feeling ookla probably charges them something (maybe a few penny) each time the test is run -- so they might not want to give you on-demand testing (cheap?) And if they don't charge them they may be able to know that it's a netgear router test and do some sneaky sneaky to crap the result on you, lol. Just another reason I would not trust it.

More likely Ookla changed an API and that needs to be amended in new firmware. I wish there was a better way to test it as you said. You can use your actual sync rate and some maths , In the UK if you know your Sync rate and have G.inp on your line (Huawei cab are the only cabs to have G.inp)  )its 96.69% of your Sync speed. You can phone your ISP and ask what your sync rate is, my ISP's app tells me mine, or get it from the info from a modem if that information is available. My Sync rate is on a ECI cab and it just dropped 5Mbps with more crosstalk, so my actual sync rate is 57.57 Mbps and 96.69% of that is 55.57 Mbps and my upstream is 17.18 Mbps so using 96.79% again its 16.61 Mbps up, which is roughly what a speed test gives me. Overheads with equipment can of cause cause variation.

[Netduma Fraser]

There seems to be a new firmware available for the XR500 if you do a check on the router, we've had a report that the speedtest has been fixed on that firmware