Why do Asus routers have issues with Apple

[bodean]

Is there a reason asus routers tend to not play nice with apple products?


Sent from my iPhone using Tapatalk

 

[L&LD]

To answer the post's question, I think it is entirely on Apples shoulders to fix issues (they're the proprietary ones).

 

[bodean]

So it's not something a FW would fix on asus router ?


Sent from my iPhone using Tapatalk

 

[RMerlin]

Apple loves to make all kind of weird experiments, which tend to break with various routers. It's not unique to Asus, it also happens with other manufacturers like Netgear. Every time they release a new major iOS update, they have to fix wifi-related issues in the coming point releases.

This has been going on for the last 2 or 3 iOS releases now.

 

[sfx2000]

Apple loves to make all kind of weird experiments, which tend to break with various routers. It's not unique to Asus, it also happens with other manufacturers like Netgear. Every time they release a new major iOS update, they have to fix wifi-related issues in the coming point releases.

This has been going on for the last 2 or 3 iOS releases now.

Actually Apple doesn't do that much odd stuff - it's more about Asus doing an bit of optimization that can cause issues with Apple (and other non-Windows devices)

 

[RMerlin]

Actually Apple doesn't do that much odd stuff - it's more about Asus doing an bit of optimization that can cause issues with Apple (and other non-Windows devices)

It's not specific to Asus. I've heard numerous complains about wifi issues for the past two years from Apple, both related to iOS and OS X, and regardless of the router brand. Overly aggressive power management, reliance on 802.11h/d despite being deprecated, MAC randomization throwing off routers, the network stack change about two years ago which they ended up reverting in an update... Their track record when it comes to interoperability has been pretty bad these past two years. They visibly don't test their stuff against anything but their own products - which is par for the course for Apple. And this extends beyond wifi, with the SMB issues appearing since they decided to reinvent the wheel and replace Samba with their homebrew stuff. Safari website compatibility is also going down the drain recently, as they are somehow scrapping what was a fairly solid web renderer with Webkit by, once again, going down their own path, regardless of the consequences. Sites that render fine under other webkit-browsers are suddenly broken under Safari.

 

[Wutikorn]

I came across a website when search about setting up WiFi in professional part. If I remember correctly, it says iOS has problem connecting to Asus router with beamforming on. I'm not sure if I remember correctly though, and the website may not be a good source either.

 

[wocram]

My wife has a iPhone 6S which doesn't see any 802.11n 2.4GHz SSID's anywhere on any router..
Upgraded iOS, checked/resetting all the settings, Zero result!
Lucky my old (2012) Macbook Pro still has an ethernet port

 

[sfx2000]

It's not specific to Asus. I've heard numerous complains about wifi issues for the past two years from Apple, both related to iOS and OS X, and regardless of the router brand. Overly aggressive power management, reliance on 802.11h/d despite being deprecated, MAC randomization throwing off routers, the network stack change about two years ago which they ended up reverting in an update... Their track record when it comes to interoperability has been pretty bad these past two years. They visibly don't test their stuff against anything but their own products - which is par for the course for Apple. And this extends beyond wifi, with the SMB issues appearing since they decided to reinvent the wheel and replace Samba with their homebrew stuff. Safari website compatibility is also going down the drain recently, as they are somehow scrapping what was a fairly solid web renderer with Webkit by, once again, going down their own path, regardless of the consequences. Sites that render fine under other webkit-browsers are suddenly broken under Safari.

1) mDSNResponder/Avahi/Bonjour - yes, Apple kicked an own goal there with their brief experiment - since fixed - this borked even Airport based networks during that time... between Airprint and Airplay, that was a bit of a mess...

2) MAC Address Randomization - only comes into play when the Client and AP are _not_ associated - if there is a client-AP bind, the MAC address is consistent

3) Testing - it's on the part of the AP to do compat/interop testing, not the client - Asus and Linksys seem to be problematic, Netgear, TP-Link, and D-Link seem to have that sorted. Maybe Asus should buy a couple of Macs, some iPhones and AppleTV's and write some test cases around it - but rather it seems that Asus is more interested in a Windows only world with their Republic of Gamers brand - and we probably should discuss PS4, XBoxOne, and just about every Android and Chrome Device out there...

Seems to me that the QA deficiency is on the part of ASUS, not everyone else...

4) SMB - Apple's in-house stuff was pretty good, and continues to be - I realize that the Samba community is a bit butt-hurt about Apple walking away from Samba - but this is Samba's own goal with GPLv3 - that being said - Asus Samba implementation is fairly broked due to inherited complexity - if Asus did a clean sheet, they'd be better off...

 

[sfx2000]

reliance on 802.11h/d despite being deprecated

I saved this one for last...

1) 11d - this wouldn't be a problem if region definition was mutable - 3rd party guys kinda broke this one to "open" up channels outside of the regulatory domains

For example - TW as a reg domain can be a real problem for any client that observed 11d - as that can and does impact clients doing a service scan, and this is more that MacOS - Linux was impacted as well.. this has also impacted other regulatory domains, and being able to tweak/adjust the RegDomain is pretty much just plain evil for neighboring networks... besides, it breaks the law...

2) 11h - not deprecated at all - this is TPC/DFS, and this opens up spectrum for everyone...

Glass houses and stones - be careful...

 

[RMerlin]

Seems to me that the QA deficiency is on the part of ASUS, not everyone else...

So if a router manufacturer (cause, once again, this is NOT specific to Asus) launches a router, and when Apple issues an iOS update it breaks compatibility, it's the router manufacturer's fault for not traveling in time to test their router against a future change from Apple? Ya, that makes sense...

For the past few years, every time Apple issues a new iOS/ MacOS major update, I see a bunch of complains on the web about wifi issues, for all brands of routers. I know you got into your head that this is an Asus-specific issue, but it isn't. Could be it's a Broadcom-specific issue, since Asus is heavily Broadcom-based, while other manufacturers tend to have a larger mixture of SoC providers. But the fact that those complains start appearing after NEW iOS releases make me suspect that Apple doesn't test much outside of their own, private ecosystem.

We can discuss it forever and constantly run into circles - this is pointless. I'm just mentioning what I see reported left and right, stating that it's NOT a single manufacturer's problem.

 

[vdemarco]

I have nothing but apple stuff connected to my ac66, and it all works perfectly. I don't have 802.11h/d turned on, and my ac66 does nothing but routining and wifi.

 

[sfx2000]

So if a router manufacturer (cause, once again, this is NOT specific to Asus) launches a router, and when Apple issues an iOS update it breaks compatibility, it's the router manufacturer's fault for not traveling in time to test their router against a future change from Apple? Ya, that makes sense...

It absolutely makes sense - there are 1 billion IOS devices out there - not to do some level of QA is insane...

 

[dlandiss]

For the past few years, every time Apple issues a new iOS/ MacOS major update, I see a bunch of complains on the web about wifi issues

I have seen recommendations, especially at routerguide.net, that Apple support is eased somewhat with WMM Support = ON and IGMP Snooping = ON. It was also recommended to set WMM APSD = OFF if mobile devices were dropping off.

Do those recommendations still make sense?

Thank you!

 

[L&LD]

I have seen recommendations, especially at routerguide.net, that Apple support is eased somewhat with WMM Support = ON and IGMP Snooping = ON. It was also recommended to set WMM APSD = OFF if mobile devices were dropping off.

Do those recommendations still make sense?

Thank you!

From what RMerlin has posted and what I have experience with; it makes sense as long as Apple doesn't change anything in the future (future - from the posts you were reading).

 

[RMerlin]

I have seen recommendations, especially at routerguide.net, that Apple support is eased somewhat with WMM Support = ON and IGMP Snooping = ON. It was also recommended to set WMM APSD = OFF if mobile devices were dropping off.

Do those recommendations still make sense?

Thank you!

WMM defaults to enabled. I believe someone posted a few years ago that disabling this will have some negative impact on wifi performance.

IGMP snooping shouldn't be required unless you use an application that specifically requires it.

The only setting I advise changing in general is disabling Beamforming, as compatibility can vary between clients. The rest should be fine on their default values. The usual recommendations still apply: avoid fancy SSID and WPA2 keys, manually pick the best channel, avoid DFS channel unless you have no alternative, etc...

 

[john9527]

I'm personally also wary about the WMM APSD setting. When I was browsing the spec for it, I started to get a headache thinking of the ways something could possibly go wrong If everything is working, great, but if you are having problems, I'd look at this setting as well.

 

[sfx2000]

WMM defaults to enabled. I believe someone posted a few years ago that disabling this will have some negative impact on wifi performance.

WMM should always be enabled - it's what makes 802.11 frame aggregation work for HT/VHT...

IGMP snooping shouldn't be required unless you use an application that specifically requires it.

This is a switching function - IGMP snooping basically turns multicast packets into unicast packets - This is in the broadcom SDK layer, so the switch there is a boolean - fairly harmless as all clients support unicast, and it saves some bandwidth...

(FWIW - this is also available on Apple's Airport AP's, and it can help)

IGMP Proxy however - this is not at the switch level, and this can cause problems for many clients - so consider disabling that one...

The only setting I advise changing in general is disabling Beamforming, as compatibility can vary between clients. The rest should be fine on their default values. The usual recommendations still apply: avoid fancy SSID and WPA2 keys, manually pick the best channel, avoid DFS channel unless you have no alternative, etc...

I recall we dug into this sometime back - it's all in the broadcom drivers - but the controls for it are in the Asus side of the house that's really a challenge to do right -- it can be done, but it would be hell to manage unless Asus did it directly because of complexity between the WebGUI and the IOCTL's sent down to the 4360 (I was looking at the 68U specifically) - so Beamforming turns into a bit of a mess as it tries to do implicit and explicit for 11n, and the explicit also controls 11ac for the 5GHz radio - so the thought at the time was to disable implicit, enable explicit - but the WebGUI logic there is a bit rough...

The biggest issues I've seen with Asus WRT is that they still do advertise the RegDomain (11d) and they don't do 11h (DFS/TPC) - I'm ok without 11h if they turn off the 11d stanza - that helps out with many clients actually, not just Apple based.

the other issue that I've seen with Asus, and it's not just Asus, is properly handling local domain multicast traffic for service discovery and traffic - basically Bonjour/Avahi - which basically breaks AirPlay and AirPrint - this isn't proprietary stuff... Android and desktop Linux also depend on this, and Windows 10 is starting to make much more use of it.

And it's broken at a low level (Broadcom) - some vendors that use Broadcom with later versions of the SDK (including newer Kernels) have sorted this, but bringing in a later SDK/Kernel into AsusWRT is a huge risk that Asus would have to take on..

 

[N66uASUS]

what exactly is the issue? sometimes, my safari takes forever to connect to reddit on my iphone

 

[dlandiss]

WMM should always be enabled - it's what makes 802.11 frame aggregation work for HT/VHT...

I'm confused. is this the AMPDU RTS and AMPDU aggregation that you recommend be enabled? But WMM APSD still off?

Or when you say "WMM enabled" do you mean WMM APSD?

Thanks!