Or when they put it inside or behind a metal file cabinet and wonder why the wifi doesn't work unless you're right next to the file cabinet

Maybe not unifi, but yeah, that's definitely been an issue. One law office type location complained, and the wireless was located under sometimes desk in the corner of the office. There's probably dozens of other examples I could tell, but yes. Absolutely.

It's the reason I got into profiling/modeling of wireless signals, so I could accurately estimate optimal placement and recommend remediation for those situations. Just get a floor plan (or take a few measurements and generate a basic one), and play with the placement until things look good.

It really really really depends on the situation. How big is the home, how fast is the internet, future plans, coverage needs, etc. And budget.

IMO, most people don't spend an appropriate amount on their internet network. They'll spend hundreds per month for internet service from an ISP, but refuse to spend more than $60 on a router to get it from the ISP modem to their $1000 phone. It doesn't make sense, especially when you consider that it's a one-time cost that will serve you for many years; the per month cost of owning it is less than a cup of coffee a month, even if you buy something that's 10x more expensive. People are so resistant to it that is crazy.

Even at $100, you're going to get what you pay for, better than the $60 router, but it's going to fall short of anything that's worth keeping for extended periods of time.

Right now my blind go to, when I don't know anything about someone's situation, is the ubiquiti UDR. It's an all in one device, around $200, with a WiFi 6 access point built in, and four ethernet ports, all 1Gb and two of them have PoE. The limitation is the 1Gbps internet link so anyone with an ISP connection faster than 1Gbps should go with something else. Most don't, so this is my recommendation.

The reason I recommend it is that you can add additional access points to it as needed with basically no additional spend (beyond the ap itself). With two PoE ports, you can add two aps without worrying about whether or not you have power injectors or anything. Adding a small network switch is easy on any of the other two ports if four isn't enough for you, and ubiquiti has some pretty cheap, small switches for that.

The UDR comes with a built in network management device (otherwise known as a "cloud key" in unifi) which puts everything into a single interface. So the access points (both internal and add on), switches, etc are all managed from a single system that's local, and has local sign in, and has a convenient cloud interface which is free to use, and has all the same functionality.

As a starting point, this is excellent. Because having just the UDR you can service a small apartment without any add-ons. If your needs grow, you can build out as needed. If your needs grow beyond the UDR, you can upgrade to the UDM/UDM pro (or similar) and keep using all the add-ons you've purchased. There's room for growth, and ubiquiti has proven themselves to be rather adequate at providing networking for the home. Unless you get into some very advanced features, it should serve whatever needs you have both now and in the future.

My recommendations obviously change depending on the situation. One person I worked with recently, I recommended the UDM pro and a pair of access points, because his ISP connection was in his furnace room. The UDM doesn't have any WiFi built in, and the furnace room is usually a nightmare to get wireless into or out of. The connection will suck. So being able to move his access points (of which he got a couple), out of that area and into more open spaces, was critical. In that case I gifted him a Dell power connect switch and used PoE injectors to power the access points. I was able to provide a very good, very reliable network for him without ongoing difficulty or issues (which was his primary concern). Before this he struggled with the wireless from his ISP modem, and it disconnected and had very poor signal throughout his home.

Since the initial set up, I've heard nothing but good things about it. He's very happy with the system. His situation had some unique challenges, and we even ran ethernet vertically up to his second floor office through what used to be a chimney stack and that had been used to install a forced air duct for the second floor. So the space was only that forced air duct from the furnace. The cable(s) share that space with the air handling duct (but are run outside of the duct for safety), and he has gigabit ethernet jacks in his office to get a reliable connection for his professional systems in the office. He probably doesn't need the Dell switch in the mix, but it gave him extra ports for use for later.

I've done dozens of custom recommendations for people. At this point I'm thinking of making a website to point people to that has all the information about home networking you could need, generalized enough to always be correct. I want to include sections on different manufacturers and why wifi kind of sucks, explanations of different technologies, their benefits, pros and cons, that kind of thing. Etc.... All in one place so I can link people to it and they can learn as much or as little as they want; with pages like recommendations, all time stamped so you know what's current or recent, and deep explainer pages of different technologies and how they work and what benefits and stuff that they have if you want to know more.

It's a huge plan, and it will take me weeks or months to write it all out. I will need to find cheap web hosting for it and get some kind of UI/UX design going for it, and build the whole thing so I can update it without having to work too hard to start the new page entry. Maybe a wiki style? Idk. I'm thinking of calling it something along the lines of "WiFi sucks" or something related... Like wifisucks (dot) com or whatever. I'm sure I'll need help with it, mainly in the UI/UX, but it's something I've been considering doing for a while, just to handle these kinds of questions continually.

I have a blog about it, called untangle the tubes, IIRC, and I go on long, in-depth rants about stuff. But it's highly disorganized and random.

3COM was picked up by HPE networking, which eventually became Aruba.

The last 3COM I worked with was a 48 port gigabit switch.

My collection is mainly Cisco stuff. Some of their early gigabit equipment and late FE stuff. Everything from ISR 1841 and even older 2600 router stuff, and 3750 stacking switches at 100mbps with and without poe, and a bunch of newer stuff too. I recently picked up a 3750X 48 port PoE, and a catalyst 4948.

I also have other vendors like HPE, juniper, even some of D-Link's "enterprise" stuff (spoiler, it sucks). I have firewalls from Cisco, watchguard, sonicwall, etc.

I have wireless stuff too. Mainly Cisco, a couple of WLC 2504, with aironet 3602, 2802, a bunch of 1141 and 1142.... A mix of other one-off versions.

Yeah. I have a bit of everything. From firewalls to switches to wireless to routers. Active in my network is a sonicwall 2650(? I believe), the Cisco 4948, and 3750X PoE, one of my WLC 2504's and a pair of 2802i access points. I also built and run my brother's network, which is almost entirely unifi, a UDM pro, enterprise 48 PoE switch and a few U6 pro access points. It works for him and it's not so difficult to manage, so if he needs to update his WiFi password or something, he is able to do that.

I realized I didn't address the garden issue you have. Ha. I'm dumb sometimes.

It's going to depend on what you have in mind and what capabilities you have to install the equipment, as well as what you're currently using.

I'll discuss two situations, in the first, you have a preferred vendor for all of your wifi stuff. Someone like ubiquiti. From there you have two choices, either find an outdoor unit to install on the side of your home near the garden(something with waterproofing), and install an indoor unit as close to the garden as possible. In either case it may be a good idea to get a directional or semi-directional unit that can "point" the wireless in the desired direction. It doesn't need to be a very directional unit (like a dish antenna), and something like that may actually not improve things very much. So look for something like a panel antenna style unit. It may be a good idea to get a point to point panel-like semi directional unit. If you're uncertain what I mean, Google panel antenna and it should be a bit more clear. If no such type of device exists, I would see if there's a unit that can use external antennas and go that way, buying a compatible panel antenna for the job.

If the garden is not very large, you may be able to get by with a normal "Omni directional" unit.

Installing outside has the difficulty of getting it wired into the network. The most common solution is to run a cable through the soffit, into the attic, and to the rest of the network. I would highly recommend outdoor cable that is UV rated, otherwise it will fall over time. Failure will take many years with a non-outdoor rated cable, but it will happen. So a bit of extra investment in cabling is going to be worth it in the long run. Nobody wants to climb through the attic to run ethernet, and nobody wants to do it twice for the same network run. So save the time and effort of doing it, and get the UV rated cable.

The other scenario is that you have random network gear or no preferred vendor. In this case the network will be disjointed, where you have to disconnect from your home wifi, and connect to your garden wifi, either manually or from saved profiles automatically. This is a bit less ideal, but usually cheaper. Finding a wireless bridge access point for outdoor install, or simply putting an access point as close to the garden inside the home, may be sufficient.

I'm planning on doing the same, but it's winter right now, where I am, so I'm at a stand still. The network in question is ubiquiti based, so I got a U6-Mesh, which I'm planning to install in the soffit (the unit is water resistant and has a "ceiling mount" option which will embed it into the soffit part way). The biggest challenge I have is getting the wire out there. In that case the wire will not be outside, so I don't need UV resistant cable, but I'll likely get it anyways for that unit. I'll skip the specifics of my difficulty running this network cable. While the unit is called the U6-MESH, it does not need to be running in a mesh operational mode, which I will not be running it in a mesh mode. The U6-MESH is small, its a bit larger than a redbull can, so I shouldn't have any issue putting it in. It's omnidirectional, so I'm planning to place it around the midpoint of my home, in the soffit, to give fair and adequate converage of the area. I want to repeat this for the front of my home.

The last step for me, and something I would recommend, is to have someone with a WiFi device that can report signal strength, at the very end of the area (at the back fence) about as far as possible from the access point as anyone should reasonably be for access, then "turn down" the broadcast power of the unit until that area is just barely covered enough to stay connected. Outdoor wireless can go far with no obstacles in the way, so I want to limit coverage so I'm not blasting my neighbors with the power of the unit at full, all day and all night, possibly disrupting their wireless in the process. My main motivator for this is that if they can "hear" my network, they can connect, and if they can connect, then it's possible for them to illegitimately use my internet, or "hack" into access that they have not been given. By limiting the signal strength, I can reduce this attack vector. As long as the signal can be picked up outside, however, it will never be eliminated. The point is to make it harder for the layperson or casual hacker to try. Any professional trying to gain access will overcome this with ease, so I don't want to invest too much effort into it.

I hope that helps. I have no idea if it will. Either way, good luck.

I'm a qualified amateur operator. I can operate on any ham bands up to 190W EIRP if I recall my countries regulations correctly.

The issue with doing something like wifi on ham bands technical issues finding radio chips capable of signalling at such a high rate of speed, on bands that are able to be used by hams. There's also the requirement for hams to identify themselves on air, and the general use of AM/FM and derivative technology on ham bands and general resistance to the OFDM used as the main signal encoding for WiFi. So finding an OFDM capable radio transmitter/receiver for use in... say, the 2m band (144Mhz .... ish) is basically impossible, and there's no way to identify. You would have to build a new protocol and standard from the ground up and use very modified or rare/expensive radio chips, and likely build the drivers/firmware for it entirely yourself. People with the required hardware, software, baseband, radio, and firmware experience that are hams who want a product like WiFi for ham radio channels is extraordinarily rare.

As for city-wide WiFi/mesh networks: it has been attempted, and has seen some limited success, but doesn't scale well with the usual protocols. Routing protocols like BGP, OSPF and IS-IS are meant for much larger IP blocks being routed between interfaces. A wireless mesh system would use a single interface (one radio) for both send and receive, which most protocols don't support, and each "hop" or station on the mesh would only be advertising a single IP (or an extremely small set of IPs) per participating node.

Most routing protocols assume that every node on an interface can talk to every other node on the same interface and thus there's no need to repeat or relay messages from an interface to the same interface.

There's also no standards that allow wifi to use multiple channels/frequencies for tx/rx, eg, send on 5.45 GHz, and receive on 5.65Ghz. it simply isn't something that any WiFi chip is capable of. So full duplex isn't possible right now.

The common wifi frequencies are also extremely power limited and on bands that are prone to interruption. In the wild, there's plenty of things that can disrupt 2.4Ghz and 5/6Ghz transmissions. With the power limits, to go any significant distance, you need directional antennas that limit free space path loss so the signal travels further. In the case of wireless internet service providers (WISP, not to be confused with the mobile carriers), they generally use panel or dish antennas to extend the range. For power output, at the high end, some bands allow for upwards of 5W of directional power, or 1W of omnidirectional power (in EIRP). On the low end, handheld ham radio units start at 5W of power, and can usually attenuate their transmitter to 1W or lower as an option. Household WiFi is usually around 0.1W of power per radio. Even cranking that up to the maximum legally allowed wattage won't result in covering more than a few blocks of a city with a fairly poor signal overall; that signal is going to be fairly easily blocked, absorbed, reflected, or otherwise attenuated by just about everything, including, but not limited to the structure of your house.

Meanwhile, standing in my home with a 5W handheld transceiver operating on 70cm (440Mhz), I can hit a repeater that's something like 10 miles away with a nearly perfect signal over FM, without assistance. OFDM signals would likely be scrambled beyond recognition at 2.4 or 5 GHz across that much of a gap, or even one that's 1/10th as far, with only 1W of legal power, without using some kind of directional antenna or antenna array.

Don't get me wrong, well configured wireless can go so far that you have to account for the curvature of the earth, but they're always very very directional, using dish antennas or similar.

Don't get me wrong, the ideas are great, but the challenges faced are enormous. It can quickly turn into a lifelong project to get something functional, and even then, there's no guarantee that it will ever catch on as a product. The limitation for ham operators regarding encryption is problematic when it comes to data communication as well, since just about everything that's data-driven on the internet implements SSL. Computers and systems expect encryption all over the place and bluntly, those messages cannot be sent over ham bands. There's nuance to that regulation, at least in my country, but I won't get into the fine print here.

Even so, there are some crude digital modes used by ham operators which are normally voice encoding or plain text encoding. Uses are limited on purpose. If you're interested in longer distance emergency communications you could look into LoRA, which is relatively new.

There's a lot more to say on this, but bluntly, I've said enough. It's all interconnected, and I love it, but I'm just ranting now.