@jfunk see these posts.

https://community.tp-link.com/en/home/forum/topic/825110?replyId=1580720

https://community.tp-link.com/en/home/forum/topic/825110?replyId=1580742

  @raven-au 

Oof, that doesn't sound good.  So all these things started breaking as newer firmware was released?

To be clear on your instructions, are you saying that if you do the firmware recovery thing you can choose which version to download OTA or do you need to already have a copy of the old one stored?

jfunk wrote

  @raven-au 

 

Oof, that doesn't sound good.  So all these things started breaking as newer firmware was released?

 

To be clear on your instructions, are you saying that if you do the firmware recovery thing you can choose which version to download OTA or do you need to already have a copy of the old one stored?

  @jfunk sorry, yes, you need to go to the TP Link web site and download the firmware zip file and extract the firmware ready to provide it to the router.

In my case release 1.1.5 was available on the web site for download and 1.1.6 (with the release date I mentioned) was offered as an OTA update and I chose to update.

So you might need to be careful you don't end up with a later firmware if later releases are available, either for download or OTA update, in your region.

  @raven-au 

Thanks, I was able to revert to 116 and the game detection immediately started working again.  Detects games right away (including clients running in the background like Riot Games) and records proper stats for sessions.  Wild that this function (the primary selling point) was just completely broken by newer firmware.  (Note 116 is now on the regular download page instead of 115).  The length of the hold down is definitely on the longer side (I'd say 20 seconds to be sure).  I tried ~15 the first time and it just did a normal settings reset but still booted back into 121.

I do miss the manual fan speed control from 121.  It really does need it, this thing gets scorching hot.

Another question for you.  The sticker on my box says hardware version 1.6 but the GUI says V 1.0, is this normal?

I'm gonna try and put this thing through its paces here, but this experience does have me concerned a bit so far.  Got this on a great deal, but the fact that the primary purpose of this router (gaming prioritization) appears to be broken in newer firmware and a core gaming feature (DDoS protection) is locked behind a paywall that will actually make it MORE expensive than the Asus Rapture 98 Pro over a 3-5 year lifespan has me strongly considering switching.

jfunk wrote 

Another question for you.  The sticker on my box says hardware version 1.6 but the GUI says V 1.0, is this normal?

  @jfunk Usually there's a note on the firmware download page that specifies that. Mostly hardware revision 1.x == firmware revision 1.0.

  @jfunk 

Following this thread.

I also bought the router specifically for my gaming setup mainly for my PS5 Pro and heavy use of GeForce Now with ultimate plan.

Unfortunately, the AI QoS feature isn’t ready yet. I’m experiencing significant packet loss and severe bufferbloat when I activate the Gaming mode and QoS. Without Qos enable my ping is Ok-ish vs my Asus ROG GT-AX6000.

I’m currently in contact with a TP-Link engineer, and they told me they will look into it. Game activity can not see my Geforce Now Seesion etc.

  @Joseph-TP @Bauhaus 

Hi Joseph, thanks for following up.

Yes, I downgraded from 121 to 116 and the game detection seems "better", but here is more information based on my observations since then. 

First, my initial sequence of events was I was on 116 when I booted the router up for the first time.  My son immediately got on his computer and launched Rocket League which I saw pop-up in the game panel right away.  I was prompted to update to 121 OTA so I did that before messing around with too many settings other than turning game boost on.  After updating to 121, Rocket League no longer detected in the Game Panel.  When Rocket League was launched, it would briefly detect "Source Engine" (which RL does not even use, but I suspect was it detecting some kind of traffic to Steam when the game launched as that is the platform we launch RL from), but this "session" would only report being a few Kb and only last 1-2 minutes even if the game were being played for an hour.  I tested a few other games, but rarely did anything show up in the Game Panel.  When it did, it would be something like Source Engine or "Unreal Engine".  This is when I downgraded back to 161, and initially it seemed like it was better as I saw Rocket League detecting again.  However...

After spending some more time with it over the weekend, I can't help but conclude that the QoS simply isn't able to properly identify game traffic in general, let alone prioritize it.  Game detection is still totally hit & miss whether anything at all shows up in Game Panel.  Even when somethign does show up, the length of session and data transmitted is never even close to accurate.  These sessions never show as longer than a few minutes even if the game is on for hours.

In addition, the QoS does not work to help game traffic be prioritized over other traffic on a given PC.  If I open a game, then do some kind of download/upload to produce heavy traffic, the game immediately suffers poor ping and packet loss.  However, what DOES seem to work is the prioritization of "boosted" DEVICES over other devices.  If I saturate my link with traffic from a non-boosted device, my boosted devices do not suffer (my son's PC is plugged into the game port and my PC is manually added to the boosted devices list).  So I can successfully maintain functional gaming perforformance on my & my son's PCs if I use my wife's PC to saturate my bandwidth.  This tells me the QoS is successfully prioritizing our devices over other devices on the network, but it's not successfully prioritizing game traffic on our devices vs other traffic originating from our devices.  Also, if I use one Boosted PC to saturate the link, the other Boosted pc suffers as well (so I can't download something on my boosted PC while my son is playing a game unless I remove my PC from the boosted device list first).  Again, this demonstrates that the QoS is only working at the device level (prioritizing boosted devices vs non-boosted) but not properly differentiating by traffic type.

I don't get any error messages or anything like that.  I can just observe that ping time goes from ~20-40ms in games to 500-1000ms as well as packet loss when I saturate my link with test traffic either in the form of a large download or by a synthetic test like Waveform or Devinia bufferbloat tests.  I have 600/20 internet service.  I've played with the bandwidth settings in QoS but it doesn't really seem to make much difference.  Currently running it set at 580/18. 

Does this not match what the engineers observe in QC testing?  Are PC gaming traffic sessions being properly identified, tracked, and prioritized vs other traffic originating from the same device in your lab environment?  Is your testing environment built with slow enough WAN links that you can actually saturate the link with test data to verify the behavior?

Also, please make note of and report my concern that DDoS protection is pay-walled.  I consider this a core function of a "gaming" router (ddos attacks are indeed a current problem right now, specifically in high level Rocket League which my son plays and trying to mitigate network performance issues from his matches is my primary catalyst for investing in one of these expensive routers right now), so the fact that is costs 55USD per year to activate it kind of hurts tplink's position as a value option vs its direct competition which a has a higher initial price tag but is fully functional out of the box with no further fees.

Not related to my core gaming related concerns here, but I also want to note that the switching from 5Ghz to 2.4Ghz when moving away from the router is also very unreliable on this compared to my old router.  I never even noticed anything happening on my old router, it just worked.  With this one, I frequently find my phone reporting no internet after moving to the other end of the house and have to manually disconnect and reconnect several times to successfully make the switch.  I'm not sure if this is something the router is expected to learn over time and more aggressively push the devices over to the other frequencies.

Thanks for your time, I appreciate any suggestions or feedback you may have.  While device prioritization and the ability to manually set bandwidth limits on individual devices are definitely an improvement over my old router, the seemingly non-functional traffic type based prioritization, poor game detection (which also makes all the statistics reporting void) and lack of DDoS protection out of the box are disappointing.

  @Joseph-TP 

I am leaving today for a vacation but I will definitely follow up when I get back next weekend to provide log files for them.

Thanks again for your time Joseph, it is appreciated.