News "Wi-Fi 7 Stomps on the Gas"

[L&LD]

I'm okay with higher throughput = more bandwidth used. That is the promise of 6GHz in any event, correct? I just hope that 8 antennae solutions (4 external, 4 internal) will be more widely available.

 

[sfx2000]

What is AC/AX on 5GHz better than N on 2.4GHz for an IoT device?

In 2.4GHz/ISM, sure would be nice to fully deprecate 802.11b (along with WPA(1)) - the airtime savings alone would make ISM band a lot more efficient than it is with B/G/N in play, and goes without saying that HE mode (11ax) in 2.4 would also benefit.

 

[sfx2000]

You guys know the main way Wi-Fi 7 increases throughput is by using more bandwidth, i.e. 320 MHz channels and 4096 QAM? Which means you get those big numbers only in 6 GHz.

Then we start running into Physics problems - going from 80 to 160MHz channels, we lose 3 dB, going to 320 from 160 is another 3 dB - and 320MHz channels are practical only in the extended 6 GHz block, won't be useful in the 5GHz blocks even now because of DFS...

It's a challenge now to keep good signal in place with 1024QAM, it's going to be even a bigger challenge with 4096QAM.

And then there's the practical side - 2 stream clients will still be dominate for some time to come...

 

[Tech9]

In 2.4GHz/ISM, sure would be nice to fully deprecate 802.11b (along with WPA(1)) - the airtime savings alone would make ISM band a lot more efficient than it is with B/G/N in play, and goes without saying that HE mode (11ax) in 2.4 would also benefit.

Not going to happen soon for cost saving reasons. If something new is 10 cents more expensive, but the old one is good enough - it's out. Amazon sells "smart" plugs for under $10, manufactured for under $2. What technology improvements to expect? Profits come first.

I suspect it will take another upgrade of clients, too.

Another decade and theoretical maximum 40Gbps in specifications only. AC is widely used and sufficient, AX is mainly new routers. Where is AD?

 

[sfx2000]

Not going to happen soon for cost saving reasons. If something new is 10 cents more expensive, but the old one is good enough - it's out. Amazon sells "smart" plugs for under $10, manufactured for under $2. What technology improvements to expect? Profits come first.

OpenWRT already defaults to g/n for 2.4Ghz, removing the 11b support

[OpenWRT] - disables 802.11b legacy rates as default - finally

Nice to see this commit over on OpenWRT master (trunk)... commit ce5bcff304e474d604ac0e25de7b6b124f764351 Author: Nick Lowe Date: Fri Dec 4 17:47:47 2020 +0000 hostapd: Disable 802.11b data rates by default Set legacy_rates to 0 by default to disable 802.11b data rates by default...

www.snbforums.com

It's a relatively minor config change, no software changes needed.

 

[Tech9]

OpenWRT already defaults to g/n for 2.4Ghz, removing the 11b support

That's okay, no B clients around anyway. What I was saying is we may never see AX low bandwidth requirement client. It may be better in theory, but N is good enough and cheaper - is here to stay. Same for AC - it will be around for many years, because it's good enough. Most people mostly browse Internet. Everything >100Mbps feels the same. They don't care what technology is used.

 

[thiggins]

Where is AD

It's still around, used mostly in Dell laptop docks.

Just wasn't worth it to for manufacturers to add the extra radio and antenna(s) for only short range applications, especially on the client side.

 

[RMerlin]

I suspect it will take another upgrade of clients, too. It's unlikely current RF front ends don't have the bandwidth and/or lower EVM to support all those new (and much closer together) constellation points.

And so long that client manufacturers will keep providing clients that barely use the capability of the standard (i.e. using only two streams), these new standards will keep getting knee capped by underperforming clients, and will remain big marketing numbers with limited performance impacts in the real world.

 

[thiggins]

clients that barely use the capability of the standard (i.e. using only two streams),

Two streams will continue to be the norm for mobile devices. That's why wider channels and higher QAM do the heavy lifting. More streams are gravy.

But yes, the BNOB will always be based on best case (most streams, highest link rate, widest channel). The marketing games continue.

 

[sfx2000]

Two streams will continue to be the norm for mobile devices. That's why wider channels and higher QAM do the heavy lifting. More streams are gravy.

But yes, the BNOB will always be based on best case (most streams, highest link rate, widest channel). The marketing games continue.

Apple is always a good benchmark as to what features add value and are worth supporting, esp with the iPhone in particular. As their flagship device and significant moneymaker, they deploy only what makes sense from a consistent user experience, and their devices tend to work pretty well no matter which AP vendor is used.

The M1 Macbook Pros - the intel WiFi5/WiFi4 based devices, along with certain iMacs and Mac Mini's, were some of the few 3-stream radio's available as OEM gear, but the new M1 based Macs are now two-stream radios.

 

[thiggins]

but the new M1 based Macs are now two-stream radios.

I did not know that. Thanks @sfx2000

 

[Tech9]

these new standards will keep getting knee capped by underperforming clients

Most clients just don't need this performance. A phone/tablet with 2-stream AC radio can get 500Mbps. It can't use this speed, it's not needed. Upgrading to AX radios is no improvement, just added cost. Apple/Samsung are doing it for marketing purposes, some of the rest follow. Slow adoption happens when the technology exceeds the needs. Intel knows from experience. Microsoft is now pushing hardware upgrade with Windows 11 to help.

 

[RMerlin]

Two streams will continue to be the norm for mobile devices. That's why wider channels and higher QAM do the heavy lifting. More streams are gravy.

While I agree that 2 streams is enough for a mobile device, I think there is certainly a market for laptops to start supporting three streams (Apple used to if I recall). Even desktop wifi have been largely limited to two streams these past few years (as a lot of them are just reusing the Intel AX200/AX210).

 

[RMerlin]

Most clients just don't need this performance.

Then why sell routers and advertise standards that virtually zero client can take advantage of?

I don't know a single triple or quad-stream Wifi 6 client.

That's why I feel that these newer standards are largely marketing smoke.

 

[avtella]

I suppose marketing can and do claim with MU-MIMO they can serve multiple clients in parallel with extra streams on the AP side boosting performance. Granted in testing it’s a mixed bag with only Qualcomm chipsets doing it well. Broadcom chipsets had issues with it back in the WiFi 5 days, even causing loss in performance at times and some Broadcom based clients like the Galaxy S (whichever model was out at the time) dropping to single stream mode with MU.

As for Apple in a sense @sfx2000 yes they’re conservative in adopting new technologies till they are perfected in a sense, but at times it seems like an excuse to also cheap out with components; like the Intel vs Qualcomm cellular modem iPhone 8 / X instance, where they gimped the QCA variant to perform close to the inferior Intel one, rather than just use QCA for all phones. Granted part of that was due to a spat with Qualcomm. They were also using the BCM4360 based 3x3 stream HT80 chips since like 2013 till the M1 MacBooks last year. The Intel 8265 / 9260 and newer were pretty competitive throughput wise at like a 10-15ft range even at HT80. Now with M1 Macs using 2x2 AX (BCM4378 I believe) and still no HT160 and no real gain in throughput at least from what I tested and a nice drop with AC routers vs the old Intel Macs thanks to 2 streams. Also seems worse than the AX200/210 based in my very limited testing. Granted HT160 may not feasible in many instances still would have been nice to have the option for those who can use it.

 

[Tech9]

That's why I feel that these newer standards are largely marketing smoke.

Even today's AX. The number of AX router models exceeds the number of available AX clients. Average family users with average Internet activities won't notice substantial difference between 5GHz N, AC and AX. Just take a look around how many people upgraded to AX routers because of few new AX mobile devices. And the single more capable AX client is on 160MHz wide channel for unknown reasons.

I suppose marketing can and do claim with MU-MIMO they can serve multiple clients

This was advertised even before the chipsets were capable of doing MU-MIMO. Asus was selling RT-AC88U MU-MIMO router with not working MU-MIMO in BCM4366B1, for example. Do you remember seeing "sorry guys, here is partial refund for you" apology offer? Some early models RT-AC88U users still believe MU-MIMO enabled in Professional settings is actually doing something good.

 

[sfx2000]

As for Apple in a sense @sfx2000 yes they’re conservative in adopting new technologies till they are perfected in a sense, but at times it seems like an excuse to also cheap out with components; like the Intel vs Qualcomm cellular modem iPhone 8 / X instance, where they gimped the QCA variant to perform close to the inferior Intel one, rather than just use QCA for all phones. Granted part of that was due to a spat with Qualcomm.

Just my opinion, but Apple rarely does things without a good reason to do it - so things like MU-MIMO, HT160, etc, while being checkbox features, in the larger scheme of things, don't necessarily provide actual value to their customers.

BTW - it wasn't that they gimped the QCOM modems... they had a performance target to meet, and both Intel and the Qualcomm modems met those targets. To favor Qualcomm would cause other issues, similar to what folks claimed were a difference in performance with the iPhone 6S and the TSMC vs. Samsung fabbed Apple A9 processors (there was no difference other than what unqualified pundits claimed)

 

[avtella]

I didn’t mean they shouldn’t normalize performance between the two variants of course that’s unfair to customers. My point was, instead of dual sourcing with an inferior Intel part (mostly to put pressure on QCA), they should have just used the superior QCA modem for all phones, to it’s full capability as they had in previous models (6S and prior). Even when not at its full capabilities (4x4 MIMO and LAA missing), the QCA modem model was still superior. But I’m diverting from the main topic so my apologies.

Exclusive: Qualcomm's iPhone X Still Outpaces Intel's

With exclusive lab testing from Cellular Insights, PCMag found that the Qualcomm-powered iPhone X has better LTE performance than Intel models, although the gap is narrowing.

www.pcmag.com

 

[GJJ]

Even today's AX. The number of AX router models exceeds the number of available AX clients. Average family users with average Internet activities won't notice substantial difference between 5GHz N, AC and AX. Just take a look around how many people upgraded to AX routers because of few new AX mobile devices. And the single more capable AX client is on 160MHz wide channel for unknown reasons.



This was advertised even before the chipsets were capable of doing MU-MIMO. Asus was selling RT-AC88U MU-MIMO router with not working MU-MIMO in BCM4366B1, for example. Do you remember seeing "sorry guys, here is partial refund for you" apology offer? Some early models RT-AC88U users still believe MU-MIMO enabled in Professional settings is actually doing something good.

I agree.

Your average user is not going to see a speed benefit from a higher speed wifi AP.

(Thinking positive here) Having come through 2 lock downs, with work from home and home schooling (Australia), our old ASUS AC68U was more than up to the task of providing a network connection for our family of 4. The limit that most people would have is their internet connection, because they don't have devices connected within their network that would have need to talk to each other at anything over 100Mbps and even then, the time spent at those kind of speeds is minimal. Unless you have a NAS or some other type of home server, then what 2 clients within your network would need a multi-gig connection? I doubt your average user has either.

Again in reference to COVID, when video conferencing with colleagues, it wasn't speed that was causing issues, it was poor home networks. One colleague has suffered from continued drop outs for the past 2 years and places the blame on their ISP. I'd like to think I'm an above average home user (but not as high as a lot of the people on this forum) and I imagine that my colleague would be your average user. I'm almost 100% positive my colleague has their ISP provided router stuffed somewhere at one end of their 2 storey house and their home office on a different level at the opposite end. There's no way a faster WIFI AP is going to help them over come their drop outs and frustration.

What I think is needed is easier device setup, better sharing of existing bandwidth between your own and other people's devices/networks, fault tolerance, diagnostics and reporting of faults, better concurrent interoperability between devices and better education of the end user of what to expect, not outright speed. Probably, a lot of these features are something that device providers could provide now and I think mesh devices try to aim to solve some of these problems.

Back to using my colleague as an example, they should be able to easily see that the connection between the AP and client is weak and therefore they can relocate 1 or the other to get a better signal. Or, their AP should be able to see that their local area is over crowded with devices and should be able to take steps to mitagate that or to be able to negotiate with other surrounding APs to come up with a better solution. Even if the AP was more content aware, it could prioritise a video conference over general web browsing, file downloads or software updates.

What has surprised me, is that with all the talk of AI and ML over the past few years, it doesn't seem to be making its way into chipsets, or at least that I have seen to be promoted, that would make the user experience better.

Anyway, outright speed isn't needed right now.

Greg

 

[L&LD]

That is a very narrow view of why newer, faster, and better routers are not needed for the 'average user'.

WiFi is a time-sharing medium. With a faster internal connection between the router and the clients, the 'cost' in time is much less for everyone. This allows for higher throughput, in total, for all.

When I installed an RT-AC3100 for a customer that had only 50/10 Mbps speeds, the network experience was greatly improved.

RT-AC3100 Report https://www.snbforums.com/threads/s...-go-with-the-rt-ac1900p-v3.34748/#post-281391

Similarly, when another customer upgraded from 2x RT-AC86U's to 2x RT-AX68U's (note that only one was actually needed to match and surpass what the 2x RT-AC86U's were able to offer, the newer hardware offered a greatly enhanced network experience.

Report - 2x RT-AX68U upgrade over 2x RT-AC86U in wireless backhaul mode


I do not disagree with some of your other points. But in tech, saying something is 'good-enough' is always proven wrong when/if sufficiently new tech is able to be deployed.

Is Wi-Fi 6 Worth It

Wi-Fi 6 Pt 2

Tech moves too fast to say anything is 'good-enough'. It can only be stated with certainty that it is good enough, for now. Any future benefits are mere bonuses, not a given.