Mesh network for campground - input requested

[Bill Zinn]

Although I'm supposedly "retired", I just got a call from an long-standing client that I 'helped' some five or seven years ago with his requirement to implement a mesh wifi network that fairly well covered his 125 acre commercial campground in the foothills of the Rockies (Colorado Springs area). At the time, we accomplished this with OpenMesh repeaters and APs connected to the only method of accessing the Internet in the area, which was 'bonded DSL' thru CenturyLink (we installed two bonded DSL modems at opposite ends of the campground acting as 'gateways' to achieve something resembling a shared 40MBs download speed, connected thru wireless-AC Openmesh APs and repeaters - arguably not the spiffiest but it worked relatively well, considering the limited means of access to the Internet and the 'terrain' involved). As an aside: it took us nearly two years working with CenturyLink in this rural location to get this system to a 'tolerable' level of performance and "dependability" (and yes, you did hear me snort on this term) which we could "live with", and while not terribly speedy, it did function acceptably (IMHO).
Problem is NOW he has only recently been contacted by the firm that acquired OpenMesh / Cloudtrax in 2016 (Datto Networks), when they informed him that his existing system was gonna disappear in December of this year. So I'm here looking for input on what to suggest to him that he can begin to contemplate for next season. My initial research and reaction to this situation is that Datto can go take a flying leap 'cuz their 'suggested alternatives' leave this guy in a poor situation. Nothing they've suggested to this point is either 'affordable', 'reasonable', or 'acceptable'. This relatively small campground is "family owned" and operates only from mid-May to roughly Labor Day, so not much of what Datto suggests is reasonable or in any way 'affordable'. But unless I miss my guess, it looks like a full-out replacement of all that he has with something that might still use this bonded-DSL Internet connection.
So.. like I said: looking for input from people who might make reasonable suggestions.

 

[Trip]

Hi Bill. Welcome back to the circus.

My condolences regarding Datto. They're an interesting beast; several acquisitions over the past few years and the outcomes haven't always been for the better, as you've experienced. Perhaps it is time to move on. There are certainly other solutions out there, a few of which would likely be a better fit at this point.

First question: any idea on the client's budget? I'm hoping "as cheap as possible!" isn't the answer, but I get where you're coming from.
Second: do you have to run the mesh APs outdoors, or can they run indoors and/or sheltered?

Next, two starting suggestions:

1) If it looks like they'll be keeping DSL for the foreseeable future, and if you don't already have a router that can do SQM-based QoS (using fq_codel, cake, etc.), definitely get a router (or two) that can. It will likely make a world of difference in reducing bufferbloat on the DSL link. For only 40Mb/s, a Ubiquiti EdgeRouter X for $59 or a sub-$100 QCA-based consumer box running OpenWRT with Cake is all you'd need.

2) If they experience higher-than-normal issues with signal loss and/or jitter, I've found shielded RJ-11 patch between the modem and the wall jacks/filter often works wonders for stability.

Interested to help further if I can.

 

[Bill Zinn]

Hey Trip... THX for the rapid reply. Hmm... Yeah well, my "initial interaction" with Datto (just after he called me and I called them) ultimately did not really “bode well' for any further involvement with the firm. 'Nuff said, leave it there ...
As regards your questions – well… I am about to ‘bury’ much of your queries with the "whole enchilada" so you can get a ‘feel’ for what we have:
#1 - budget: Since they are now well aware of their circumstances in the next couple of months, I won't say there's 'panic in the park' but more like acknowledgement of what they're going to have to consider. My guess is that if this can be done for a few thousand (and I DO mean ‘few’ - like under $3K if at all possible) then that's what they'll need to do. I have been doing a bit of research since my IP and presently am looking at Ubiquiti equipment, so an EdgeRouter is arguably a good concept (and here I thought all this IT stuff was behind me after 36 years... heh… go figure).
#2 – Much of the equipment must be outdoor rated, thus my consideration of Ubiquiti mesh systems because I've had good results with Ubiquiti in the past, and I see that several of their “Unifi Mesh” hardware is outdoor rated (thinking the repeaters that are needed). As you say, outdoor rating reduces options tremendously even while ‘dependability’ is absolutely key.
So, now that I have someone like yourself to volley this back and forth with (sorry for that), here's the deal:

There is no 'central office' or 'management facilities', per se, located within the campground being serviced by the current OpenMesh network (managed via Cloudtrax). The offices for the campground (and the LAN it uses) are located approximately 1500 ft. to the West, and roughly some 2oo feet ‘downhill’ below the campground area. Insofar as there is no ‘hard connection’ between these two networks (he totally balked at my suggestion of a fiber channel between the two networks), there is (to my comprehension) only the one option of ‘cloud management’ to be had, and to accommodate much of this, we’ve installed and provided protected environmental accommodations for all the DSL hardware to provide Internet access for the mesh network.
By this I mean that, currently, the bonded-DSL modems (one at each ‘end’ of the campground area) are mounted inside VERY study steel cases with powered ventilation and AC power conditioning (as well as lightning protection and proven grounding). We’ve accomplished this by driving a minimum of 12 ft. of a 3/4” diameter copper plated grounding rod at the base of each of the two DSL locations (which was no mean feat, let me tell you, there is apparently little else except solid granite rock and ‘boulders’ buried beneath the soil for all of this. Lotsa fun ...). We then provided lightning arresting protection for the enclosures and contents, and installed a significantly over-rated UPS for each of the two locations. Power for these 'enclosures' is delivered via the 120/220 volt power distribution system that serves the entire campground area.
Each of these two 'gateway' locations have an OpenMesh A42 802.11-AC ’MU-MIMO’ AP configured as a 'gateway' which are directly connected to the DSL modem(s) via shielded CAT-5 Ethernet. Maybe not particularly ‘elegant’, I suppose, but functional to this point with no issues related to either power or damage from lightning strikes.
A note: it is CenturyLink that is our largest 'weak point' because DSL is our only option for Internet access for the foreseeable future, and at this distance from the city, it is at best 'not so dependable'. The major Internet access limitation is due to the fact that the campground is located nearly 25 miles outside of any other form of provisioning. We considered satellite but that goes no where quickly when I've tried to suggest it.

However, it is the necessary 'repeater' locations which require outdoor rated equipment, all of which are currently similarly electrically protected, and the OM2P-HS 'repeater' units have a single 'omni' antenna connected to them that are mounted on poles projecting above "facilities buildings" (laundry rooms, picnic shelters, shower facilities, etc.).
BTW: each repeater now shows only a single 'hop' between repeater and 'gateway'. I made certain of this...
These, of course, use that same 110/220 volt power distribution system (which, BTW, was upgraded from the 1940's-era "poles and lines" to a buried power distribution system over the past five years. Couple of bucks invested there, absolutely).
As you probably already know, the OpenMesh OM2P-HS repeaters need only a 12v 1 A "wall-wart" transformer for power (altho Passive PoE is also a possibility), which means currently there is a "wall-wart" transformer plugged into the accompanying UPS's at each location to power these units.
Up till now, there has been little (if any) ‘monitoring’ or metering of network traffic, but this MUST come to an end now (and yes, there is a bit of a horror story regarding this which I won’t get into now). Suffice to say “trust but verify” will be the name of the game as we address these necessary changes to the existing mesh network and 'user access'.
So... From this entire diatribe you can probably now get a bit of understanding of the ‘limitations’ and ‘requirements’ he is facing. Cloud management will most probably be his only alternative insofar as there’s not much hope of ‘local management’, and Ubiquity appears to offer this (still doing research of this, tho..).
So there you are.
Still want to ‘volunteer’ your input on this?

 

[coxhaus]

What was your thinking on DSL on separate ends of the campground. To me it seems better to have both modems at the same location so you can load balance them. But it becomes all or nothing.

I am with you for a long fiber run. It is much more reliable and not that costly.

If you have power poles around the campground you can string fiber for some of the legs.

Then what is left has to be dealt with.

I like structure over wireless.

 

[Bill Zinn]

Good point, and we started with a 'central distribution' of the original DSL access, but soon found out that due to 'obstacles', this just wasn't what we necessarily wanted.
There were some trees, and some structures that WERE in place back then (buildings now removed but the trees will always be there... Forest Service sees to that ... ), and of course the ever present Godzilla-level Class-A motor homes that perennially show up, meaning the wifi mesh signal suffered massively. (Now, of course, all mesh units are located approximately 20 ft' above ground and obstacles aren't the issues they were before).
So, due to complaints (and it kills me that people b***h about free Internet access at 40 MBs in a campground at close to 8000 ft. elevation which 'comes with' the cost of the RV site ... but I guess that's today's camping world), so it was decided to go with 'dual' bonded-DSL access points because, at the time, it was a simple, fast 'fix' for a better overall signal distribution with minimal 'hops' between repeaters (thus yielding slower performance for those people that pay extra to park those mega-dollar Godzilla-level Class-A units).
Plus, it was the ONLY way that CenturyLink would 'ante-up' the requisite upgrades to the aging telephone service distribution to even support DSL faster than 1 MBs. Like I said, the site is more than 25 miles outside of ANYTHING resembling 'faster'. So at my suggestion, when he upgraded the power distribution, he (we) also installed a 90-pair buried telephone cable right alongside the new power distribution conduits, so we had LOTS of circuits to bring DSL all the way to the North end of the campground (well over 900 ft. distance).
Oh and BTW: Even cell signals are stupidly "hit and miss" at this location, so he bit the bullet and did two bonded-DSL's that CenturyLink (graciously) allows us to 'decommission' during the winter, and then fire-up again in the spring.
From 'end-to-end' the entire campground is nearly 1/2 mile "long" (2426 ft. according to Google maps), so 'central distribution' just ain't gonna cut it.
Plus, Sangre de Cristo Power and Energy (the electric company at this location in Colorado) vehemently prohibits any thing but THEIR wires on any existing pole in this area (remember, the Forest Service is also involved in this, so 'fire hazards' are of huge concern ...), which was yet another reason for the new buried distribution of power throughout the campground.
Prior to the new power distribution plan, they were going to cut off the distributed power until such time he agreed to $900 per linear foot for the new primary wires (and that's for EACH wire on a 3-phase power distribution system) to replace the existing old 1940's-era stuff. And that's all the way from the highway to the campground, which is a distance of over 2 miles. Add to that they wanted $1100 per pole for everything deemed 'insufficient' (which you just KNOW was gonna be all 38 poles...), as well as transformers, cross-ties, service drops, etc. etc. etc. Total estimate soared past $400,000 in a heartbeat.
But, because he owns all his own earth-moving equipment (including a really decent sized excavator), it was decided that it was cheaper to "do-it-ourselves", which they did. Between himself and his two sons, they dug the 5 - 6 ft. deep trenches from the SDCP&E 'demark' point (at least the Forest Service got them to put that on the Forest Service right-of-way) and then into the campground and to his 'private' switch gear. Both the power company and CenturyLink demanded these trenches and conduits "must be in place" for the new power (and telephone cable) to be activated. Cripes, "permitting" through El Paso County was a breeze by comparison.
It's been an 'eventful' seven years, for them ... And I won't even begin to address El Paso County's new water restrictions. That was going to be another $550K to 'meet standards', initially, but thank god for 'arbitrators'... They dodged a bullet there...

 

[Trip]

Thanks for the additional info Bill.

Please confirm, or clarify:
- RV internet is served via 2 separate DSL WANs, each 40Mb bonded, at opposite ends of the park.
- Mesh APs are all outdoor, half being zero to one hops from WAN A, the other half zero to one hops from WAN B.
- All requisite AC power and backup is in place, with overhead room if replacement gear was, say, up to 50% more power-hungry at certain drops.
- Replacement system needs to enough AAA for "trust but verify" access control, or better.
- Office LAN is 1500' west and 200' down-hill, with currently no LAN-to-LAN interconnect to the RV network.
- Upgrade budget is $3K-ish, or less, all-in.

More questions:
- Mesh topology: how many root and secondary mesh nodes on each half of the park, and in total?
- Office-to-RV LAN Interconnect: is there enough line of sight for a PtP link, like a MikroTik Wireless Dish or similar?

Comments:
Just FYI, a central WAN is possible but the kind of mesh required to do it would be miles beyond your client's budget, and moreover totally overkill (carrier-grade gear like ABB/Tropos, etc.).
At $3K max, enterprise campus mesh is likely a no-go (Cisco, Aruba, Ruckus, etc.), unless refurbished, maybe; otherwise, SMB/faux-enterprise like UniFi Mesh, MikroTik or similar is probably where you're headed.

I'll reserve any more comments for now.

 

[Bill Zinn]

Thanks for the additional info Bill.

Please confirm, or clarify:
- RV internet is served via 2 separate DSL WANs, each 40Mb bonded, at opposite ends of the park.
- Mesh APs are all outdoor, half being zero to one hops from WAN A, the other half zero to one hops from WAN B.
- All requisite AC power and backup is in place, with overhead room if replacement gear was, say, up to 50% more power-hungry at certain drops.
- Replacement system needs to enough AAA for "trust but verify" access control, or better.
- Office LAN is 1500' west and 200' down-hill, with currently no LAN-to-LAN interconnect to the RV network.
- Upgrade budget is $3K-ish, or less, all-in.

More questions:
- Mesh topology: how many root and secondary mesh nodes on each half of the park, and in total?
- Office-to-RV LAN Interconnect: is there enough line of sight for a PtP link, like a MikroTik Wireless Dish or similar?

Comments:
Just FYI, a central WAN is possible but the kind of mesh required to do it would be miles beyond your client's budget, and moreover totally overkill (carrier-grade gear like ABB/Tropos, etc.).
At $3K max, enterprise campus mesh is likely a no-go (Cisco, Aruba, Ruckus, etc.), unless refurbished, maybe; otherwise, SMB/faux-enterprise like UniFi Mesh, MikroTik or similar is probably where you're headed.

I'll reserve any more comments for now.

Trip
Addressing your last, point by point:
1. CONFIRM - 2 separate DSL WANs, one WAN at opposite ends of the park - **owner just informed me that CenuryLink CANNOT "guarantee" stable operation of 40Mbs at this distance (2 bonded 20MBs channels) so speed has recently been reduced to 12MBs/channel for a total of 24MBs per WAN**
2. CONFIRM - Mesh APs are all outdoors, 0 to 1 hops from either WAN-A AP or WAN-B AP
3. CONFIRM - All requisite power and backup in place with availability for additional backup power delivery to a maximum of 200 watts each location (limitation would relate to UPS capacity which is, of course, up-gradable)
4. CONFIRM - AAA will now be a requirement due to recent past experience with "bad actors" being present (CenturyLink has contacted them in the recent past with 'evidence' of "illegal Bit Torrent streaming")
5. CONFIRM - Office LAN is 1500 ft. distant and 200 ft. down-hill, no existing LAN-to-LAN interconnection between locations
6. CONFIRM - budget is somewhat 'restricted' by MY estimates, but I'm trying to get confirmation from owners of a "maximum" or "acceptable" limit (my thinking is $3K is pretty capable IF the Ubiquiti Mesh hardware can be implemented in a similar fashion as were the OpenMesh units)
7. Topology of root-secondary mesh nodes is currently only one 'root' per LAN with three secondary 'repeaters' CURRENTLY, however - coverage has been limited and 'spotty' and this would be the time to address "improvements", so the current number of nodes should probably increase. Thus my push for additional funding.
8. NO "line-of-sight' access currently exists between LAN locations due primarily to trees. Forest Service 'frowns' on removal of any more trees beyond "natural losses" (altho, beetle-kill has been a recent problem, every tree in a potential "path" will need approval for removal).

I agree, what is needed is beyond their 'acceptance' unless put at the point of a gun (like what was presented to them by both the power company and county water board in other situations). Cisco, Aruba, et al is most probably a 'non-starter' due to costs. Prices of Ubiquiti Mesh units at Amazon seem pretty acceptable at just $86/ea. (Ubiquiti UAP-AC-M Unifi-Access-Point) IF what I beleive is accurate, which is that these could supplant the OpenMesh units. My thought is that with the addition of two units at each 'root' location' (to accommodate better radiation/coverage) this would provide additional signal source and better radiation 'pattern'. Estimated total for nodes would still be under $3K at this cost level. I have no experience with this hardware, so my 'guesstimate' is probably suspect, but it appears to have that capacity. This is my thought, but correct me if I'm wrong.

 

[coxhaus]

You might take a look at the Cisco small business wireless WAP571E. It is an out door unit with a pole mount kit and clustering software built-in as well as the controller software so they are self contained. It clusters 16 units per cluster per network. I could not find how many units you are currently running per network. It offers a rugged housing (IP66 rated), and works even in rain, snow, and extreme high and low temperatures.

We still have a problem for the 1500 ft run.

 

[Trip]

@coxhaus - Correct me if I'm wrong, but Cisco's small biz wifi stuff is not true wireless-mesh capable. "Cluster" capability by wire (which it can do) is a far cry from true mesh capability via pure wireless, nor does it have any of the live intelligence that goes along with it. To do it with Cisco, you'd need to step to full Aironet, at least the 1500 series or higher, and move up to the higher hardware and licensing costs, which puts it out of range for this budget (along with Aruba and Ruckus); dare I say even refurb stuff would still be too spendy from any of those three. I've also looked Aerohive, Mist, and all the firewall vendor solutions, and all that stuff is still too spendy as well.

@Bill Zinn - Thanks for the info.

So no-go on a wireless LAN interconnect. If a fiber run would be too spendy, how about ethernet over coax? Even just a 100Mb/s link would more than suffice for management, even backup masquerade path for park internet (which is now 48Mb/s total). Or, as I'm guessing, is the bulk of the cost the trenching and burial, less so the cabling material...?

As far as gear goes, UniFi Mesh may just be your best bet. It's less customizable than MikroTik, but way more approachable and has a much more slick and mature software ecosystem (cloud-host the controller on AWS, DigitalOcean, Vultr, etc., or do an on-site CloudKey, Raspberry Pi or VM that you could VPN into from the cloud.

As far as the topology and inter-node setups go, it sounds like you have a pretty good handle on broadcast overlapping, power level adjustments, tweaking, etc.

Curious as to your further thoughts.

 

[coxhaus]

The Cisco small business WAPs support WDS bridge and Workgroup bridge which is a wireless connection. I think you can accomplish what you want.

I don't know whether WDS bridge works with Clustering. I will see if I can find anything on it. I would think using a POE+ Ethernet cable would be a better way than running a power cable. But that is just me. And it sounds like he already ran power.

Cisco also has a software product called Findit 2.0 which I run that monitors all small business hardware, wireless, routers and switches. I run it in my house. They support a raspberry Pi as a probe for remote monitoring which I have not run as I have no need for remote monitoring.

I would think the difference in cost of fiber over cable is pretty small. The labor to install is the big cost. I am thinking 4 strands of fiber. Fiber runs even if the pipe fills with water. Copper will go bad over time if you get water in the pipe and it has to be replaced. Lightening is also a big problem with copper in the ground. We were regularly replacing copper cable with near lightening hits at some of our campuses. Once we switched to glass fiber that all stopped.

PS
You cannot enable Single Point Setup if Wireless Distribution System (WDS) is enabled.

 

[Bill Zinn]

THIS (right here) is why I came here... Input like this is invaluable.
Trip, I had NOT thought of coax (at least not in this decade), but you're right... Any 'capable' speed would be adequate simply for local management. Possibly I may need to traipse down there again and do another 'site survey' just so I can remind myself of obstacles vs accessible pathways as measured against 'risks'. And by 'risks' I have to keep reminding myself this is "publicly accessible", and thus prone to vandalism, which is perennially a consideration for them.
But... I gotta agree that cost differences of coax vs fiber is not all that significant at these lengths unless one already had this coax 'in hand'... Coxhaus is correct in saying anything copper is definitely going to be 'fragile' by comparison to fiber. We've already seen how some copper has performed in a couple of locations for them in just these past six or seven years, and to this point he's been fantastically lucky because lightning strikes "on the mountain" are frequent and widespread, but he's somehow managed to avoid that in these past years, even while I can't help but believe they're "living on borrowed time" and it WILL happen at some point.
But, there is yet another 'advantage' to be had, which now occurs to me, that a fiber interconnect could bring, as well:
The offices "downhill" CANNOT be serviced with bonded-DSL because that building (1509 ft away and 243 feet lower, I've just confirmed) is on a different muti-pair telephone cable which is in even worse condition than what is serving the 'new' LAN up in the campground (the lower cable having been installed in 1946, as I now understand). By comparison, the campground has a virtually 'brand new' POTS connection (installed in 2010) while the office is connected to this old cable that was installed in 1946. (Welcome to the rural Rockies serviced by CenturyLink, y'all..).
As of right now, the offices are being serviced with a single 12 MBs DSL connection, through which they do all management of reservations, accounting, credit card processing, and the day-to-day standard Internet access for all other purposes. And yet, this DSL is about as "trustworthy" or "stable" as the worst of the DSL connections in the campground ever were.
So he's actually paying for three DSL connections, currently, none of which are as capable or 'stable' as we might experience in any other 'average' location. But .... with the possibility of interconnecting "upper" with "lower", then the office could also access a better Internet connection than they now have, and most probably would experience no worse of any sort of 'dependability' than they do now.
But none of this is my decision to make. If it were my decision, I'd already have his two sons starting a narrow trench right at the side of the access road leading to the campground and throw a 2" irrigation tube in the bottom of it (which he seemingly has MILES of ...), and pull quad-fiber through it. As a former electrician in my 'youth', I've done worse... But, IMHO termination and interconnect has gotten to be a lot less of an issue than it once was, and fits with a nifty deal I just stumbled across where a "brand new" 'pre-terminated' 4-fiber cable 1540 ft. (475.5 meters) in length was offered for less than $800, as compared to some numbers we were quoted a couple of years ago that ran close to $2400 for two fibers. It was that last quote that stopped any further consideration for him, but most assuredly if I could lay fiber between the two LANs, we'd certainly be in business and this discussion would be 'moot'.
But without interconnection, the only method of 'management' or monitoring would be via something like the 'Cloudtrax' management service that is still in operation, and which has been free of cost to this point with the purchase of any of the OpenMesh hardware (in the event this was not known) - although it appears Datto has every intention of killing this at year's end.
And which is why I am here now looking for alternatives that might provide similar service and capabilities.
So general question and observation: Is there anyone "out there" that has had any experience or exposure to the Ubiquiti UAP-AC-M Unifi Access Point that Amazon is offering for that low price?

 

[coxhaus]

Make sure you match the fiber size and plug to the fiber module which goes into the switch. The fiber module will have a distance listed on it so you can tell whether it work or not. I have only used Cisco with fiber. We had 1000s of fiber connections so we standardized on connections so equipment could be moved around. One gig connections should be really cheap now.

 

[Trip]

While you're researching, you may also want to check out TP-Link Omada Mesh. Omada is their more-or-less equivalent to UniFi, albeit with a bit less software polish and not quite the feature set. Not sure what level of 802.11 standard you're aiming for, but they make an N300-class model for around $55/unit. and an AC1200 (roughly equivalent to the UAP-AC-M) for $75. Price is so close it's probably just worth sticking with UniFi for the platform maturity and ecosystem, but just though I'd mention that as well.

I was also going to Engenius Neutron Mesh, but their mesh-capable EWS hardware is priced into enterprise territory, where Cisco Aironet, Aruba or Ruckus make way more sense anyways.

 

[coxhaus]

The problem I have with TP-Link is they do not support their hardware for very long. You will be replacing it real soon.

 

[Bill Zinn]

Make sure you match the fiber size and plug to the fiber module which goes into the switch. The fiber module will have a distance listed on it so you can tell whether it work or not. I have only used Cisco with fiber. We had 1000s of fiber connections so we standardized on connections so equipment could be moved around. One gig connections should be really cheap now.

Yeah.. matching fiber to module adapter is a given... I just thought that the price was pretty freaking amazing for the length... But I have a source here locally that can make this happen - if only he'd agree to this...
I've also reviewed Engenius, but as you say, for what's needed this cost approaches others that he'll discard "out of hand"..

 

[Trip]

The Cisco small business WAPs support WDS bridge and Workgroup bridge which is a wireless connection. I think you can accomplish what you want. I don't know whether WDS bridge works with Clustering. I will see if I can find anything on it. I would think using a POE+ Ethernet cable would be a better way than running a power cable. But that is just me. And it sounds like he already ran power.

Bill has already confirmed he cannot run data cabling through the entire campground to the APs. That being understood, WDS or Workgroup bridging is no where near as optimized as a proper mesh code base for access and backhaul on a purely wireless layout. With UniFi Mesh being in the same ballpark for cost, I'd lean towards Ubiquiti over Cisco small biz stuff -- for this use-case.

Cisco also has a software product called Findit 2.0 which I run that monitors all small business hardware, wireless, routers and switches.

FindIt is actually not too bad; it's nice that they embed the Probe into the hardware itself so you don't need a separate appliance or VM instance. That said, it's fairly basic in its feature set when looking to centrally administrate the network from a single control panel.

On the LAN interconnect cabling, I agree regarding interference and would second the move for fiber if it's at all in the realm of possibility.

@Bill Zinn - I haven't personally rolled out UniFi Mesh, but I have a few colleagues who have and they've nothing but good things to say in general. I think you will also find this general theme among Amazon reviews, reddit/r/networking, etc. That said, I would definitely purchase a few spares of the APs, for if there's one achilles' heel to Ubiquiti it's that they do tend to have slightly higher hardware failure rates on certain model serial runs, and don't offer the same level of direct coverage/replacement as many peers (a large reason why their stuff is so cheap to begin with).

 

[Trip]

Another option you might want to look into is Cisco Meraki Go, which is basically the small-business "baby" version of full Meraki. Like all Meraki, it does require an active license for all gateways and APs (otherwise they stop passing traffic), but presuming your client keeps good tabs are kept on their billing practices, it could be an option. A bit more costly than UniFi, especially if you end up having to deploy a lot of APs, but the hosting and cloud piece is all handled for you and the support and code base stability are probably a bit better than UniFi. If cost remains the primary driver, then perhaps still a no-go, but something worth looking into.

 

[coxhaus]

How many APs in each network in the campground?

 

[Bill Zinn]

Currently, there are only three "AP's" per 'gateway' unit, but I believe there is a need for more "AP's" per 'gateway'... So my focus will be on 'MU-MIMO' capability to accommodate better 'throughput' per connection.
(and I'm stating the following solely for the 'edification' of the owner of this network who has been a bit boggled by this discussion as I have printed these all out and shared them with him):

Per his suggestion:
"Bearing in mind that ALL of the existing OpenMesh hardware units can be either an 'AP' or a 'gateway' depending solely on whether it is physically connected directly to the DSL modem (AKA 'gateway unit') or if it is 'repeating' the signal offered from the 'gateway' unit (AKA 'AP unit' or 'repeater') ."

Sheesh... a somewhat 'contentious' discussion of 'terminology' - IE: 'AP' vs. 'gateway' vs. 'repeater' ...

 

[Bill Zinn]

Another option you might want to look into is Cisco Meraki Go, which is basically the small-business "baby" version of full Meraki. Like all Meraki, it does require an active license for all gateways and APs (otherwise they stop passing traffic), but presuming your client keeps good tabs are kept on their billing practices, it could be an option. A bit more costly than UniFi, especially if you end up having to deploy a lot of APs, but the hosting and cloud piece is all handled for you and the support and code base stability are probably a bit better than UniFi. If cost remains the primary driver, then perhaps still a no-go, but something worth looking into.

Yeah... been peeking at stuff like this but... cost will be the determinant, so....
Looks more and more like Ubiquiti throughout...