Is the RT-AX88U Pro overkill with PFSense?

[drinkingbird]

Only if the two APs overlap a lot and not placed properly. Two RT-AC68U at -60dBm apart (whatever distance this translates to in the specific environment) will beat any other single AP in range, AX or BE. I read stories around how 2x AC units were replaced with single AX unit, but this only tells me the AC units were not placed in optimal locations. One central AP with better range may cover the house, but what used to be outdoor range will be lost.

Can't beat Wi-Fi physics with imagination.

Not saying AX is a magic pill to overcome wifi physics, just that radios and drivers have improved a lot since the 68U, and along with rethinking placement, one could be enough. Or it may not be, obviously without knowing anything about OPs layout and size, no way to say.

Just always worth considering, I liken it to getting a 50 year old furnace replaced and they just toss in one of the same size, when in reality they're a ton more efficient now and it should have been sized down to run optimally.

 

[Tekomancer]

If Asus, I would get 2x RT-AX86U Pro. More compact design, Pro firmware. Wired in AP Mode, no AiMesh.



Why paying for more 2.5GbE ports on the router you won't use anyway then? Waste of space and money.

Appreciate your patience as my understanding of mesh networks has been devastated in the last 24H If I picked it up correctly, Asus's "AiMesh" just does the same thing that could be done with properly configured WAPs, except properly configured ones will probably do it better. If that's the case, my whole "I should reuse my really old gear" idea goes out the window and, while I think Asus has treated me decently well, I can go with any reliable brand.

I don't need POE as the current AP locations are near both power outlets and ethernet drops. The AX86U is about $210 on Amazon right now, so $420 for two. At that price, I could pick up two or even three TP-Link WAPs as well as the SDN controller (except I don't know what to look for in a dedicated WAP - never bought one before). Let's say I want to keep things under $450 - I want reliable internet in my three story house, but outside isn't an expectation (call it a bonus?) and, if at all possible, I'd prefer not to spend 15 hours tinkering with the settings. Wi-Fi 6/6E is a plus. What would the experts put together?

 

[Tech9]

just that radios and drivers have improved a lot since the 68U

Starting from RT-AC68U you can get about 20-30% better range. This is about it, more sensitive radios. The physics is all the same as well as transmit power. In most places it's not even hardware limited, but regulations limited to 200mW. Starting from RT-AC86U you may get no improvement whatsoever. There is no measurable difference in range and throughput to 2-stream AC client between RT-AC86U and RT-AX86U for example. Both are better than RT-AC68U, but single won't beat 2x RT-AC68U in range. If the routers were placed too close the roaming doesn't work well. If single router can replace two - something was wrong in first place.

At that price, I could pick up two or even three TP-Link WAPs as well as the SDN controller

I would take Omada even at higher price. The controller is a free software, can be run on something else. They have wireless mesh as well and better.

I would test with 2x cheaper EAP610 Ultra-Slim first. They are currently $85 on Amazon. Then expand with the number needed wired or wireless. They come with power adapter. This will save you PoE switch cost. You may have to play with position, they are ceiling mount design, but mostly omnidirectional. Run the controller on whatever you have available initially, see how it goes. If all as expected, get OC200 hardware controller and rebuild the system. Give it some time after setup. It needs time to learn the environment and your devices need time to recognize new AP options.

This is what the controller software looks like:

Omada Controller

emulator.tp-link.com

 

[Tekomancer]

I would take Omada even at higher price. The controller is a free software, can be run on something else. They have wireless mesh as well and better.

Mind blown again. Looks like there's an unRaid Omada Controller docker. I saw pictures of the branded switches in their flows and found a bunch when I checked on Amazon, so I incorrectly assumed that was a mandatory, separate component to make it work. If I don't have to buy a controller, then the TP-Link Omada looks amazing. I was just about to ask if you'd recommend the EAP610's or the EAP650's, but you're a ninja and beat me to it. Thank you so much.

 

[L&LD]

I'd appreciate insights, unfounded opinions, or anything else you have to offer on what makes sense here. Thanks in advance!

More likely than not, a single (well-placed) RT-AX88U Pro will blow away what network 'performance' you have currently with those converted TM '68u's, and at a fraction of the cost of any Omada setup. With more than likely no difference either in performance (except the hit to your pocketbook - see below for the 'why'... i.e. required, wired nodes, on modern cable runs).

Particularly if you can't upgrade those ancient Cat5 cables to at least Cat5E or better.

I am guessing that in your home, WiFi is what you want to optimize (and that is exactly what an RT-AX88U Pro (or two) will do). Not a semi-professional setup that will need cables ripped out and replaced too, on top of more expensive base gear.

Testing the above is what makes the most sense to me. Your move.

 

[drinkingbird]

If the routers were placed too close the roaming doesn't work well. If single router can replace two - something was wrong in first place.

Not exactly an uncommon scenario. I have one poor spot so I grab a second router and place it with heavy overlap of the first etc.

I'm still waiting (but losing hope) for one of these new seamless roaming technologies like MLO, 802.11k/r/v, something, anything to finally take hold. One with client participation that will all but eliminate these issues.

 

[drinkingbird]

More likely than not, a single (well-placed) RT-AX88U Pro will blow away what network 'performance' you have currently with those converted TM '68u's, and at a fraction of the cost of any Omada setup. With more than likely no difference either in performance (except the hit to your pocketbook - see below for the 'why'... i.e. required, wired nodes, on modern cable runs).

Particularly if you can't upgrade those ancient Cat5 cables to at least Cat5E or better.

I am guessing that in your home, WiFi is what you want to optimize (and that is exactly what an RT-AX88U Pro (or two) will do). Not a semi-professional setup that will need cables ripped out and replaced too, on top of more expensive base gear.

Testing the above is what makes the most sense to me. Your move.

I'm not a proponent of running higher than a cable is rated for but if it is already in the walls and can't be replaced, and aren't long runs next to electrical wire etc, it most likely will be fine for APs at gig and maybe even more.

Given they already have a pfsense setup personally I think a full blown "home router" especially a Pro line Asus is not the right choice and a good high end AP for less money would probably be a better fit, but obviously they have options to choose from depending on their needs.

 

[Tech9]

802.11k/r/v

Omada supports all 802.11k/v/r, but the controller has to be running. I personally don't use Omada, but have experimented with it for some time and it was one of the best systems around for fast roaming. There is no "seamless roaming" with any system. Only in advertisements. I'm happy with what I get from Ruckus and Cisco, but can't recommend for home because of the price.

 

[Tech9]

I was just about to ask if you'd recommend the EAP610's

No one knows your environment. In this case I can only give suggestions. Good Wi-Fi starts with good planning. What I know for sure is Omada will give you multiple options for AP settings not available in home routers or "mesh" systems. If you go with AiMesh you have to agree to "it is what it is" first.

 

[Tekomancer]

I thought the cables were cat5, but my wired internet tests at 900/800 so either I'm amazingly lucky with electrical interference or it's cat5e

 

[drinkingbird]

Omada supports all 802.11k/v/r, but the controller has to be running. I personally don't use Omada, but have experimented with it for some time and it was one of the best systems around for fast roaming. There is no "seamless roaming" with any system. Only in advertisements. I'm happy with what I get from Ruckus and Cisco, but can't recommend for home because of the price.

Yeah the main issue right now is clients. Lots of APs support (or claim to support) the standards but without the client participating, it's going to be somewhat useless. I have not seen widespread support for any of them in clients so far. MLO is promising as it is "part" of the BE standard but still no guarantee that every BE device will implement it.

MLO is really promising as it is essentially how your cell phone works, always keeping track of multiple towers and ready to switch over instantly as soon as needed. But I'm sure if it consumes extra battery cell phones will either not implement it or have it off by default. They already screw you with scan throttling which makes roaming even worse.

 

[drinkingbird]

I thought the cables were cat5, but my wired internet tests at 900/800 so either I'm amazingly lucky with electrical interference or it's cat5e

As long as the CAT5 has all 4 pairs wired (which in wall bulk cable usually does, only cheap patch cables omitted 4) then it could still be cat5. If you saturated it at 1G or especially 2.5 or more, you might see a few errors/drops, or you might be totally lucky and have no issues. If you can get a peek at one of the ends it should be printed right on the cable what it is, may have to check a couple cables as it is printed every 1-2 feet so depending how much you can see and where it happened to be cut, may or may not be on every one. If it is 30 years old it definitely isn't 5e, actually even 5 wasn't out then. If it was wired later then it might be 5e. Would be very rare for a 30 year old house to have ethernet wiring at all so I suspect it was wired later.

Distance and proximity to AC wiring are the two main factors affecting how much you can "over drive" a cable.

 

[Tech9]

Yeah the main issue right now is clients.

Omada controller makes a difference in roaming. It's a proprietary software and no much details about inner workings, but it found some say it tracks the clients and promotes AP switching with fast dynamic tx power adjustment. I don't know if this is the case for all Omada compatible APs though. The fact is it works better with existing clients. Wi-Fi planning still apply. Too many or incorrectly placed APs will affect the performance negatively.

 

[drinkingbird]

Omada controller makes a difference in roaming. It's a proprietary software and no much details about inner workings, but it found some say it tracks the clients and promotes AP switching with fast dynamic tx power adjustment. I don't know if this is the case for all Omada compatible APs though. The fact is it works better with existing clients. Wi-Fi planning still apply. Too many or incorrectly placed APs will affect the performance negatively.

Yeah they've all been doing that, Ubiquiti for years (starting with N) has had an implementation that works great as long as everything is spaced properly. Too far or too close and it was problematic. Also required the controller to be running to function. I had 3 of their N APs and for the most part it worked very well but sometimes was a headache. I ended up disabling it and turning clients up to aggressive roaming and using minRSSI, a combo I found to be more reliable (but still a bit stickier than I would have liked). Even AiMesh has a crude implementation, we know how well that typically works.

That's my point though, I want to see the day when it is no longer some algorithm trying to force clients to do something, rather the client and APs negotiating and working together.

 

[Tech9]

Yes, perhaps Ubiquiti is doing the same tricks. One of the reasons higher class systems cost more - there is more engineering involved. Clients will stay the same for quite long time. Most clients are still AC-class, most IoT are still N-class. Whatever is cheaper. The market wants cheap.

 

[Tekomancer]

As long as the CAT5 has all 4 pairs wired (which in wall bulk cable usually does, only cheap patch cables omitted 4) then it could still be cat5. If you saturated it at 1G or especially 2.5 or more, you might see a few errors/drops, or you might be totally lucky and have no issues. If you can get a peek at one of the ends it should be printed right on the cable what it is, may have to check a couple cables as it is printed every 1-2 feet so depending how much you can see and where it happened to be cut, may or may not be on every one. If it is 30 years old it definitely isn't 5e, actually even 5 wasn't out then. If it was wired later then it might be 5e. Would be very rare for a 30 year old house to have ethernet wiring at all so I suspect it was wired later.

Distance and proximity to AC wiring are the two main factors affecting how much you can "over drive" a cable.

My parents built the house. Dad was a CIO and insisted on wiring the whole house with internet. Everyone thought he was crazy at the time and I'm thankful he did it. It hasn't been exactly 30 years, but getting close. It must be Cat5 because it's still original cabling. One of my drops used to be half internet, half phone but I re-punched that as soon as I found it. Good times.

 

[drinkingbird]

My parents built the house. Dad was a CIO and insisted on wiring the whole house with internet. Everyone thought he was crazy at the time and I'm thankful he did it. It hasn't been exactly 30 years, but getting close. It must be Cat5 because it's still original cabling. One of my drops used to be half internet, half phone but I re-punched that as soon as I found it. Good times.

1995 is when CAT5 became readily available for the most part.

Yes was very common to have 100M ethernet and a phone line (analog or digital) sharing the same cat 5, since there were 2 extra pairs.

From what you describe you most likely have CAT 5 but it will likely be fine for gig. If you were wiring it new, I'd say no don't save a few bucks, but where it is already there, likely will be no issue.

If your dad was a CIO it may even be shielded. If it is, and you properly ground it (only one end should be grounded, the other not), you can probably push it over a gig with no drops (crosstalk will start to factor in before you hit 10 gig but 2.5 or even 5 might work).

 

[drinkingbird]

Yes, perhaps Ubiquiti is doing the same tricks. One of the reasons higher class systems cost more - there is more engineering involved. Clients will stay the same for quite long time. Most clients are still AC-class, most IoT are still N-class. Whatever is cheaper. The market wants cheap.

I would hope that laptops and especially phones will be on board. Don't care about a wifi light bulb, but something that is going to be moving around and using higher bandwidth is worthwhile. Like I said though, if it sacrifices battery life, phones probably will be reluctant. Guess we'll see.

Maybe we'll all be using Li-Fi by then. No roaming issues because you can't move

 

[Tech9]

Maybe we'll all be using Li-Fi by then.

Me and you probably will be using this technology by then:

 

[drinkingbird]

Me and you probably will be using this technology by then:

View attachment 53160

Nah ours will have ethernet ports.