Lots of 2.4Ghz IoTs, will a 3x3 be better than 2x2?

[L&LD]

Huge difference from 1GbE to 2.5GbE cabling connections. Even for home use. The latency reduction (by 2.5x) is noticeable, as is transferring anything over to a NAS (with equivalent Ethernet speeds). Not to mention 'sharing' a backbone cable for higher ISP speeds with APs/AiMesh too.

Using 2.5GbE for over three years is not a regret I have.

 

[coxhaus]

I say backbone with 10 gig and use 1 gig for clients until the heat issues with 10 gig gets better. I think you will have a better overall business network solution. The network is for all devices overall, not testing just 1.

But everybody do their own.

 

[L&LD]

So your solution is to have up to 10 client devices saturate their 1GbE connections over that backbone (consistently, and continually), rather than have all possible clients use the 2.5x faster connection (and therefore finish 2.5x quicker too)?

These aren't test networks. These are networks where work is performed on daily.

2.5x faster is real money/productivity in someone's pocket/bank.

 

[sfx2000]

2.5x faster is real money/productivity in someone's pocket/bank.

2.5Gbe is getting attractive these days - not much more cost per port that Gbe was a couple of years back.

As a personal note - 2.5Gbe is one of the checkbox features I look at for wired access...

Going back to OP's question...

Because of all the 2.4 Ghz devices, would keeping a 3x3 2.4Ghz device be helpful? I see many 6, 6E, even 7 APs that only have 2x2 for 2.4Ghz. Would this reduce the performance of my IoTs?

A 2*2:2 AP for IoT devices in 2.4GHz should be fine - a 3-stream AP might give a bit more range, but not that much more...

The bigger challenge is more at a device level in how many clients per radio it can support - two numbers stand out for most consumer AP's - 32 and 50 clients...

Some AP's have a hard limit of 32 clients, some do a bit more at 50 - thing is airtime per client - most IoT devices don't use so much airtimes, with the exception perhaps of WiFi cameras...

 

[coxhaus]

The bigger challenge is more at a device level in how many clients per radio it can support - two numbers stand out for most consumer AP's - 32 and 50 clients...

Some AP's have a hard limit of 32 clients, some do a bit more at 50 - thing is airtime per client - most IoT devices don't use so much airtimes, with the exception perhaps of WiFi cameras...

Nobody seems to publish this info. Cisco does.

 

[snovvman]

I would suggest exploring SMB gear options and moving away from consumer products.

Are you talking about the likes of Ubiquiti and Cisco? The answer to this is probably "it depends", but how does a typical single SMB node's range compared to an AC/AX86? From what I can understand, there is no substitute for those long external antennas.

 

[L&LD]

A single consumer AIO router like an Asus RT-AX88U Pro has superior range and throughput vs the likes of Ubiquiti and Cisco where a router and at least one AP are required. Further, that AP is not tuned for range, rather the contrary, it is meant to be used with multiple AP units to provide adequate coverage, which in many cases a single Asus router would cover easily (and most likely with higher throughput).

Nor are they tuned for throughput either. Marketing guarantees that (how else can they get you to move to their higher-end products, of course).

Of course, it does depend... on the exact environment we're talking about.

But the link below shows what kind of a jump AX-class equipment took over the previous 'king' of AC-class hardware. And note that a single (entry-level) AX-class router utterly dominated 2x RT-AC86Us.

[ 386.4 alpha Build(s) ] Testing available build(s)

There are some odd regressions being discussed on the dnsmasq-discuss list from the PiHole guy (Dominik). https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015598.html https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015596.html I see Merlin is testing 2.86 out...

www.snbforums.com

 

[snovvman]

A single consumer AIO router like an Asus RT-AX88U Pro has superior range and throughput vs the likes of Ubiquiti and Cisco where a router and at least one AP are required. Further, that AP is not tuned for range, rather the contrary, it is meant to be used with multiple AP units to provide adequate coverage, which in many cases a single Asus router would cover easily (and most likely with higher throughput).

Nor are they tuned for throughput either. Marketing guarantees that (how else can they get you to move to their higher-end products, of course).

Of course, it does depend... on the exact environment we're talking about.

But the link below shows what kind of a jump AX-class equipment took over the previous 'king' of AC-class hardware. And note that a single (entry-level) AX-class router utterly dominated 2x RT-AC86Us.

[ 386.4 alpha Build(s) ] Testing available build(s)

There are some odd regressions being discussed on the dnsmasq-discuss list from the PiHole guy (Dominik). https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015598.html https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015596.html I see Merlin is testing 2.86 out...

www.snbforums.com

Excellent explanation. Thanks very much!

 

[coxhaus]

A single consumer AIO router like an Asus RT-AX88U Pro has superior range and throughput vs the likes of Ubiquiti and Cisco where a router and at least one AP are required. Further, that AP is not tuned for range, rather the contrary, it is meant to be used with multiple AP units to provide adequate coverage, which in many cases a single Asus router would cover easily (and most likely with higher throughput).

Nor are they tuned for throughput either. Marketing guarantees that (how else can they get you to move to their higher-end products, of course).

Of course, it does depend... on the exact environment we're talking about.

But the link below shows what kind of a jump AX-class equipment took over the previous 'king' of AC-class hardware. And note that a single (entry-level) AX-class router utterly dominated 2x RT-AC86Us.

[ 386.4 alpha Build(s) ] Testing available build(s)

There are some odd regressions being discussed on the dnsmasq-discuss list from the PiHole guy (Dominik). https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015598.html https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015596.html I see Merlin is testing 2.86 out...

www.snbforums.com

I agree on ASUS's range. But 2 Cisco or probably Ubiquiti well as you said "utterly" destroy 1 ASUS router in range and bandwidth. And you can also have more active talkers using APs vs 1 wireless router. You can buy 2 Cisco 150ax APs for less than 1 RT-AX88U PRO router.

And now that the cost is so high on wireless routers it makes much more sense to buy APs.

 

[L&LD]

Give me a price for a Cisco or Ubiquiti router with 2 APs that will dominate a single RT-AX88U Pro in almost any home environment.

Note that the single RT-AX88U Pro is in the low ~$200 range...

 

[coxhaus]

I have no idea about Ubiquiti as I have not run one but the last Cisco 150ax I bought was $102. I think they were breaking 5 packs up. And of course I bought it from a Cisco partner.

Amazon has the RT-AX88U Pro for $239.

 

[L&LD]

Okay, so $204 and a router is needed. Not a slam dunk by a mile.

 

[coxhaus]

Okay, so $204 and a router is needed. Not a slam dunk by a mile.

It is when it comes to wireless. It provides more bandwidth and more distance.

 

[tgl]

The AIO units are what to buy if your place requires only one. As soon as you need more than one, it's not such an open-and-shut decision. The consumer-gear makers want you to buy multiple AIO units and use them in a "mesh". That's good for them: they get to sell more identical units, which improves their margins, and they get to sell you hardware you do not need, namely multiple routers. Plus, top performance in a mesh setup requires a separate backhaul radio in each unit, so you're buying even more hardware. Ka-ching!

If you need multiple APs and you can possibly run ethernet for backhaul, it'll be way more cost-effective to go with a single, probably wired-only, router plus multiple radio-only APs, and you really have to go to the SMB product lines for that type of gear.

It's a harder call if you just plain can't do wired backhaul. The SMB products can do wireless backhaul, generally, but they're not designed for it so they're unlikely to have a separate backhaul radio. You'll take perhaps a 2x peak-performance hit compared to a full-on mesh setup.

(In the interest of fairness: I believe that Netgear's Orbi line does use cheaper AP-only units for satellites, so at least in that product line they are giving you a break. ASUS not so much.)

 

[L&LD]

It is when it comes to wireless. It provides more bandwidth and more distance.

I think you're in denial. Not only is a router required (which just doubled the cost by itself), APs require wired connections, which most people won't invest in today and would effectively triple the cost to most.

When the $120 RT-AX68U was available (and not EOL yet), it covered a very large 4-story house better than 2x RT-AC86Us.

Give some objective numbers to your statement. Bandwidth and distance. I still think you're wearing Cisco-colored glasses, and drinking too much of their Kool-Aid.


See what an entry-level AX-class router is capable of, and contrast it to your recommendations.

[ 386.4 alpha Build(s) ] Testing available build(s)

There are some odd regressions being discussed on the dnsmasq-discuss list from the PiHole guy (Dominik). https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015598.html https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015596.html I see Merlin is testing 2.86 out...

www.snbforums.com

 

[sfx2000]

I think you're in denial. Not only is a router required (which just doubled the cost by itself), APs require wired connections, which most people won't invest in today and would effectively triple the cost to most.

When the $120 RT-AX68U was available (and not EOL yet), it covered a very large 4-story house better than 2x RT-AC86Us.

Give some objective numbers to your statement. Bandwidth and distance. I still think you're wearing Cisco-colored glasses, and drinking too much of their Kool-Aid.

Let's take a deep breath and take a step back... not sure why you're picking a fight with @coxhaus

BHR's always have a place, as to Mesh Packs, as well as a discrete solution with a dedicated GW/Edge device, managed switches, and purpose built Access Points.

each has a different use case - going with purpose built AP's for example, it's not about max speed, but overall capacity - BHR's put a lot more emphasis on individual client speed, but perhaps capacity takes a back seat...

Go back and look at the original question by OP - is a 3*3:3 AP beneficial over a 2*2:2 AP - it doesn't matter if it's mesh node, dedicated AP, or a big-honking router from a radio and MAC layer, as most of the 2.4GHz IoT devices are single-stream, and 20MHz only...

so when balancing capacity vs. speed - IoT says capacity, as they don't need much speed.

 

[L&LD]

I'm not picking a fight with anyone. This is a conversation as far as I can tell.

@coxhaus is stating that Cisco is cheaper than a single Asus router, to cover an average/larger home.

It's not. And I've given my reasons to defend my position. @coxhaus has given facts to back up my assertions, not his.

Give me a price for a Cisco or Ubiquiti router with 2 APs that will dominate a single RT-AX88U Pro in almost any home environment.

Note that the single RT-AX88U Pro is in the low ~$200 range...

I have no idea about Ubiquiti as I have not run one but the last Cisco 150ax I bought was $102. I think they were breaking 5 packs up. And of course I bought it from a Cisco partner.

Amazon has the RT-AX88U Pro for $239.

Okay, so $204 and a router is needed. Not a slam dunk by a mile.

It is when it comes to wireless. It provides more bandwidth and more distance.

I think you're in denial. Not only is a router required (which just doubled the cost by itself), APs require wired connections, which most people won't invest in today and would effectively triple the cost to most.

When the $120 RT-AX68U was available (and not EOL yet), it covered a very large 4-story house better than 2x RT-AC86Us.

Give some objective numbers to your statement. Bandwidth and distance. I still think you're wearing Cisco-colored glasses, and drinking too much of their Kool-Aid.


See what an entry-level AX-class router is capable of, and contrast it to your recommendations.

[ 386.4 alpha Build(s) ] Testing available build(s)

There are some odd regressions being discussed on the dnsmasq-discuss list from the PiHole guy (Dominik). https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015598.html https://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2021q3/015596.html I see Merlin is testing 2.86 out...

www.snbforums.com

 

[coxhaus]

Another thing about using 2 APs that people keep forgetting is that you can use 2 radio channels for IOT devices at the same time whereas a single wireless router can only have 1 2.4GHz active talker at a time so you have a 50% performance boost in active talkers. This also beats mesh as mesh uses the same channels not different ones. If you limit bandwidth, you can even go higher on active, parallel transmitting using APs vs a wireless router or mesh system. It also works the same with 5GHz radios to where you have increased active talkers vs any wireless router or mesh system. Bandwidth is much better on APs.

 

[L&LD]

Yes, that is true for current wireless backhaul AiMesh systems. But there is no bandwidth drop with wired AiMesh, even with the same Control Channels used (and the main and node are far enough apart).

I suspect 3.0.0.6.xxx level firmware for Asus WiFi 7 routers may change that too, for both wired and wireless AiMesh.

The reasons for not using an Asus-powered RMerlin router are getting smaller and smaller all the time. And will always be considerably cheaper with more direct control too.

 

[coxhaus]

Yes, that is true for current wireless backhaul AiMesh systems. But there is no bandwidth drop with wired AiMesh, even with the same Control Channels used (and the main and node are far enough apart).

I suspect 3.0.0.6.xxx level firmware for Asus WiFi 7 routers may change that too, for both wired and wireless AiMesh.

The reasons for not using an Asus-powered RMerlin router are getting smaller and smaller all the time. And will always be considerably cheaper with more direct control too.

What it sounds like to me is the ASUS is trying to match small business APs that we have had all along. They are just trying to come up to specs to match business APs. And they are still not there today, maybe in the future.