I work with WiFi all the time and let me tell you that after looking at usage graphs for what feels like forever, you don’t need 1Gbps WiFi.
Most clients aren’t averaging 50Mbps, nevermind 500-1000. What you want is consistent wifi. Something that doesn’t show down because you dumped everything and the kitchen sink onto it. There’s a lot of good ways to ensure this and nobody wants to pay for it.
Simply put, dumping 30-50 client devices, between cellphones, tablets, laptops, TVs, gaming consoles, IoT things (like smart lightbulbs, fridges, etc), and in more cases than I’d like to admit, desktops… Onto a single multifunction wireless router, with little more than dual band WiFi, is generally going to suck.
I usually hear a chorus of responses to this because people don’t really put together that their smart watch, Alexa, smart smoke detectors and thermostats, all count as wifi devices. It usually doesn’t make a huge difference how much each device is actually using the wifi, the fact that all of them are connected at the same time is, in and of itself, a problem with only a single access point where that contact can be made… Dual band or not.
I don’t consider mesh solutions to be solving the core issue since all of the client traffic needs to end up at a single device with all the same problems. The fact that they get filtered through what is essentially, fancy repeaters, isn’t super relevant. The problem still exists. But if you suggest an infrastructure network with multiple wired access points, people generally take one look at the price, then leave and go buy the latest night hawk from Netgear at the nearest electronics store and put it out of their mind, since it’s “good enough” (which it isn’t, in the current WiFi climate).
I want people to have better wifi, but I can’t save you all from yourselves. Now the IEEE is taking on the job, I suppose. Trying to “fix” wifi because most people can’t be arsed to install a reasonable solution for what they actually need. They’d rather spend literally thousands of dollars a year on fast internet service that they don’t need and can’t use because it’s all getting filtered through their sub $300 network that they’ve had (or will have) for two+ years, and then have the gall to complain that their wifi sucks, and they don’t get it because they’re paying $100+ a month for their fancy gigabit or multi-gigabit internet connection.
Oh yeah, having headroom for microbursts is great, and factors into the precieved speed of the connection. It’s a requirement for most users to have that headroom to make the connection feel fast. But a lot more goes into the apparent “speed” of the connection than that. Having quick DNS and high-speed routing to the wireline internet connection is also important, but harder for most to grasp what will actually achieve that goal versus other products.
The main thing is that headroom for burst traffic is mostly shared, since the channel gets used and then freed almost as fast. In this way, others can burst traffic into the channel shortly afterwards, with no detrimental effect.
The headroom doesn’t need to be gigabits of capability in most cases. 300-450mbps is often very sufficient and may be more than what is required, depending on the usage.
You seem to really know your shit. I’m realistic about my WAN speed needs (symmetrical 350 Mbps is more than sufficient) - but I’m pretty tired of my shitty Netgear setup. I’m not really sure what I need LAN wise, and the price tags of Ubiquiti/UniFi systems have me worried about buying more than I really need. I know pfsense/opnsense can be useful alternatives to start with in minimizing prices, but the steep learning curve has me a bit intimidated.
Do you have a suggestion or recommendation on where to start? Is there something that’s functional at a sub ~$500 initial investment, but would be upgradeable/expandable and ultimately more reliable/dependable in the long run? Do I need to wait for this wifi7 gimmick?
350mb/s symmetrical is very adequate for most households. If you have any trouble with speed, it’s likely a priority queue problem or an issue within the network. (Basically, router/firewall or access point - which can be interference or capacity issues mainly)… Provided your equipment isn’t bogged down, eg, CPU isn’t fast enough to keep up. The best way to know on that last point is to check reviews where they test throughput. Having a gigabit ethernet port is nice, but if the router can only handle 200mb/s of traffic, you’re going to have a bad time.
Starting from the top. Make sure your ISP will let you run in bridged mode. This can significantly reduce the work it needs to do, and increase the latency of requests (aka, ping time). Pfsense is very common, but I find it a bit cumbersome and I prefer opnsense myself. Try one, try the other… You don’t even need to use it as your router, just load it up on a spare or old PC and go poke around the user interface… See if you can find all the settings for your internet connection, LAN connection, NAT/port forwarding, etc. Pretty much anything you would normally want to touch. See what you prefer. I find the user interfaces are some of the biggest differences between the two. For non-open source, there’s some cheap but good vendors like mikrotik, but mikrotik tends to have a pretty steep learning curve, so buyer beware on that ubiquiti has some good options for small home environments too, like the UDR, which has all the fixings of a typical all in one router (switch, access point, etc) plus a ubiquiti controller, and it’s quite capable. I like the UDR because it’s not expensive (around $200) and has two PoE ports (out of the four total ports). This means you can add two access points to it. Throwing that together with a U6 access point for extending your wireless is a good solution and should come in under your $500 budget. The extra access point (or two, if you wish) will join the controller and work with the integrated ap, providing seamless roaming, as long as you turn on the appropriate settings. If you need more ports, there’s a lot of ubiquiti switches to choose from that are well priced, so you can scale out from there if you want.
As far as pfsense/opnsense goes, it can be installed on anything with at least two ethernet ports. You can go pick up a small form factor off-lease workstation for a couple hundred, and drop in a second hand business/server grade ethernet card from eBay, and it will massively outperform the 1Gb/s ethernet interfaces.
If you want to go more piecemeal about it with pfsense/opnsense, there’s plenty of good PoE gigabit switches out there, especially used business switches from HPE (I like the “1920s” model) or Aruba, even Cisco and juniper. Switches are usually pretty bullet proof unless you really abuse them, but companies don’t like to have out of warranty stuff in their network, so the units get flooded onto the secondary markets after a while. PoE is good if you don’t want to run power to access points, though you can also use PoE injectors… Speaking of access points, my favorite non-ubiquiti access point for the home is the Aruba instant-on series. They’re managed from a free cloud portal, and I have yet to find a situation where they would do a great job. The only pinch is that when the internet goes down, you lose any ability to manage them. They’re otherwise perfect. A step up from that (and not to be mistaken for it) is the Aruba instant series; they’re a lot more expensive but locally managed and have a lot more settings… It can be a lot more difficult to get up and running with the instant series if you’re not completely familiar with networking, so buyer beware here too.
If your place is smaller and you only need one access point, the best option I’ve found, for price and performance is enGenius. The locally managed ones… I haven’t kept up with their product line, but I know they were moving to cloud management for most of their stuff. So finding the right access point may be a bit of a challenge. Also, I’ve used their access point systems (for multiple access points) and found it lacking, granted, that was several years ago now, so hopefully things improved, but I would recommend the Aruba instant-on for multiple access point systems over enGenius any day of the week. enGenius has the stand alone access point crown to me though.
For WiFi 6/6e/7, it will depend on your situation. If you’re in a high density environment, you may get benefit from 6e/7… If you’re in the suburbs or in a more rural situation, 6 is going to be fine. Just make sure you’re buying dual band access points. The big benefit for 6e is that you gain access to the 6Ghz band, which is only really helpful if your local 5Ghz band is overcrowded, like in the case of a high-rise apartment (even then, it may not be all too helpful), and WiFi 7 builds on this with some multiple access performance improvements as the key features. Meaning fewer slowdowns from many things on the WiFi at once. WiFi 6 already has some multiple access improvements, most notably MU-MIMO, so, I wouldn’t worry about it too much. 6e is going to simply reduce the chances that you have to share a 5/6Ghz channel with anyone else. The issue with 6ghz is that not all devices support it (laptops, cellphones, etc), and certainly no electronics (IoT stuff, printers, TVs, etc) will support it.
Personally, if you’re not into networking as a hobby or a full time job, then I’d usually point you at the ubiquiti UDR, and add one or two access points depending on coverage requirements. Either going for the U6+ (if you absolutely need to save money), the U6 pro for ceiling mount, or the U6 mesh if you need to plug it in and place it on a surface (The mesh is a bit more money due to it being so compact). For more hardwired ports, if you want more access points or something, maybe the Lite 8 PoE… There’s also the flex mini if you only need a handful of extra ports.
If you want something more challenging, using an opnsense router of custom built off the shelf computer parts is a good option, paired with an eBay find of a HPE 1920s with PoE and a couple of Aruba instant-on AP22’s or something would be spectacular, but a bit more involved, and likely a bit more pricy, since I’m not sure how well the instant-on AP’s fare in secondary markets (I’m not sure if Aruba will let you register a used access point into the cloud controller).
The ubiquiti solution has legs but you’ll be limited by the fact that it only has 1Gbps ports… It’s powerful enough that you can fully use those ports, but it will never go faster than 1G.
I hope that helps. I know it’s a lot of information to process. I don’t know your situation and I can be much more specific about what to use without getting into a lot of detail about your situation. I don’t want you to have to post that information publically on the internet, so I wanted to give you the information to make an informed decision. I hope I have done that, and I wish you the best of luck with your future network.
Yeah this was a big change for my parents who always complained about slow wifi speeds. I spent a summer wiring the walls with cat6 and plugged everything I could directly into a big switch I got for them. Printers, desktops, start tvs, the lot. Then I set them up with multiple access points and voilà, the wifi is suddenly fast again without upgrading their plan
I don’t have the option for multiple access points in my home even though I would like them (I rent). But considering that I just had to explain the difference between Bandwidth and Frequency to someone on another thread because they couldn’t fathom that someone’s experience might be that 5G is unreliable and they have better service with 3G, I don’t think this is going to take off. Don’t even know if most internet providers in the US will do dual access points without it being prohibitively expensive, and I’m not sure the cost is justified.
On top of that, part of the problem with IOT is that those devices are required to run on the same network in order to talk to each other. I had at one point just had my IOT things on 2G band and the other devices (phones and computers) on the 5G band and that caused all kinds of problems on my network. Devices that wouldn’t communicate properly or would drop off randomly.
I do agree though with your solution, dubious though I am about it being implemented.
I had multiple wired access points in a rental apartment with high wife approval factor by running cables in cup hooks along the ceiling lines. They self-tap into the wall and leave little more damage than would be caused by hanging a picture. I put a hook every 18 or so inches for support, and ran a cable under a door, up the door frame, down a hallway and into the living room (from a bedroom where we had the network gear). From there, I only had to mount the access point and connect everything. The cables were always out of the way, and I used white cables to help hide them against the walls. It wasn’t perfect, but the limits of renting are a bit limiting.
I only had two in that place, which was plenty, and we never had bandwidth issues accessing the internet from the wifi.
The fact is, nobody is building homes with wireless access points in mind. Whether rentals, condos, or new homes, they don’t have access point hookups. Many are now being wired with Ethernet in the walls, but nothing in the ceiling. A skilled wireless engineer can easily take a floorplan and build a wireless design with access points in ideal locations for maximum coverage and speed. Simply doing this pre-work and installing ethernet in the ceiling at these locations, is all that is required, yet, I have yet to see any builder do it. The same can be done to retrofit rentals as people vacate units and updates/renovations are done, yet, nobody is doing it.
Those that own their home need only to find placements and pay someone to wire them in. The whole thing only needs to be done once, ever, and the locations should be fine for use for the long term.
Fact is, neither builders, nor property owners, nor homeowners seem to have any interest in the practice. The only time I have personally seen or known of any location that is properly wired for wifi, is when someone has hired such wires to be custom installed. If it was done by default, rather than as an afterthought, retrofit or renovation, then it may be more common that people pursue such solutions. The price problem is another major hurdle, though I stand by the analysis of people paying literally several thousand dollars on internet access from an ISP, compared to a few hundred at most for a router during the same timespan (usually 1-3 years). IMO, that’s a bit like paying $50,000 over 10 years for access to the highway, but never buying a car worth more than $1000. It’s silly.
My main point is, solutions to these problems exist, but people simply won’t pay more than a few months worth of ISP fees for their wifi hardware, then they expect it to perform well and last for many years. They will turn a blind eye to the environmental issues that plague them and instead blame everything on lackluster devices that underperform that they will continue to refuse to pay a reasonable amount to actually get quality access equipment.
I have popcorn ceilings (and those crumble quite a bit), and textured walls. I also have only one cable hookup. But either way, I agree with you. I’m also not really the person complaining my wifi sucks. 3 phones, 4 computers, and a handful of IOT devices. Plus a couple of gaming consoles and none of those things in use all at the same time to cause the kind of strain that would cause wifi drops and so on. I’m just pointing out that it’s not feasible for some people at all but more often than not it’s that nobody wants to up front expense. Thank you for clarifying though.
I work with WiFi all the time and let me tell you that after looking at usage graphs for what feels like forever, you don’t need 1Gbps WiFi.
Most clients aren’t averaging 50Mbps, nevermind 500-1000. What you want is consistent wifi. Something that doesn’t show down because you dumped everything and the kitchen sink onto it. There’s a lot of good ways to ensure this and nobody wants to pay for it.
Simply put, dumping 30-50 client devices, between cellphones, tablets, laptops, TVs, gaming consoles, IoT things (like smart lightbulbs, fridges, etc), and in more cases than I’d like to admit, desktops… Onto a single multifunction wireless router, with little more than dual band WiFi, is generally going to suck.
I usually hear a chorus of responses to this because people don’t really put together that their smart watch, Alexa, smart smoke detectors and thermostats, all count as wifi devices. It usually doesn’t make a huge difference how much each device is actually using the wifi, the fact that all of them are connected at the same time is, in and of itself, a problem with only a single access point where that contact can be made… Dual band or not.
I don’t consider mesh solutions to be solving the core issue since all of the client traffic needs to end up at a single device with all the same problems. The fact that they get filtered through what is essentially, fancy repeaters, isn’t super relevant. The problem still exists. But if you suggest an infrastructure network with multiple wired access points, people generally take one look at the price, then leave and go buy the latest night hawk from Netgear at the nearest electronics store and put it out of their mind, since it’s “good enough” (which it isn’t, in the current WiFi climate).
I want people to have better wifi, but I can’t save you all from yourselves. Now the IEEE is taking on the job, I suppose. Trying to “fix” wifi because most people can’t be arsed to install a reasonable solution for what they actually need. They’d rather spend literally thousands of dollars a year on fast internet service that they don’t need and can’t use because it’s all getting filtered through their sub $300 network that they’ve had (or will have) for two+ years, and then have the gall to complain that their wifi sucks, and they don’t get it because they’re paying $100+ a month for their fancy gigabit or multi-gigabit internet connection.
I agree when it comes to average usage, but having >=1Gbps headroom for bursty traffic, e.g., when moving files locally between devices, is awesome.
Oh yeah, having headroom for microbursts is great, and factors into the precieved speed of the connection. It’s a requirement for most users to have that headroom to make the connection feel fast. But a lot more goes into the apparent “speed” of the connection than that. Having quick DNS and high-speed routing to the wireline internet connection is also important, but harder for most to grasp what will actually achieve that goal versus other products.
The main thing is that headroom for burst traffic is mostly shared, since the channel gets used and then freed almost as fast. In this way, others can burst traffic into the channel shortly afterwards, with no detrimental effect.
The headroom doesn’t need to be gigabits of capability in most cases. 300-450mbps is often very sufficient and may be more than what is required, depending on the usage.
You seem to really know your shit. I’m realistic about my WAN speed needs (symmetrical 350 Mbps is more than sufficient) - but I’m pretty tired of my shitty Netgear setup. I’m not really sure what I need LAN wise, and the price tags of Ubiquiti/UniFi systems have me worried about buying more than I really need. I know pfsense/opnsense can be useful alternatives to start with in minimizing prices, but the steep learning curve has me a bit intimidated.
Do you have a suggestion or recommendation on where to start? Is there something that’s functional at a sub ~$500 initial investment, but would be upgradeable/expandable and ultimately more reliable/dependable in the long run? Do I need to wait for this wifi7 gimmick?
Thanks mate.
350mb/s symmetrical is very adequate for most households. If you have any trouble with speed, it’s likely a priority queue problem or an issue within the network. (Basically, router/firewall or access point - which can be interference or capacity issues mainly)… Provided your equipment isn’t bogged down, eg, CPU isn’t fast enough to keep up. The best way to know on that last point is to check reviews where they test throughput. Having a gigabit ethernet port is nice, but if the router can only handle 200mb/s of traffic, you’re going to have a bad time.
Starting from the top. Make sure your ISP will let you run in bridged mode. This can significantly reduce the work it needs to do, and increase the latency of requests (aka, ping time). Pfsense is very common, but I find it a bit cumbersome and I prefer opnsense myself. Try one, try the other… You don’t even need to use it as your router, just load it up on a spare or old PC and go poke around the user interface… See if you can find all the settings for your internet connection, LAN connection, NAT/port forwarding, etc. Pretty much anything you would normally want to touch. See what you prefer. I find the user interfaces are some of the biggest differences between the two. For non-open source, there’s some cheap but good vendors like mikrotik, but mikrotik tends to have a pretty steep learning curve, so buyer beware on that ubiquiti has some good options for small home environments too, like the UDR, which has all the fixings of a typical all in one router (switch, access point, etc) plus a ubiquiti controller, and it’s quite capable. I like the UDR because it’s not expensive (around $200) and has two PoE ports (out of the four total ports). This means you can add two access points to it. Throwing that together with a U6 access point for extending your wireless is a good solution and should come in under your $500 budget. The extra access point (or two, if you wish) will join the controller and work with the integrated ap, providing seamless roaming, as long as you turn on the appropriate settings. If you need more ports, there’s a lot of ubiquiti switches to choose from that are well priced, so you can scale out from there if you want.
As far as pfsense/opnsense goes, it can be installed on anything with at least two ethernet ports. You can go pick up a small form factor off-lease workstation for a couple hundred, and drop in a second hand business/server grade ethernet card from eBay, and it will massively outperform the 1Gb/s ethernet interfaces.
If you want to go more piecemeal about it with pfsense/opnsense, there’s plenty of good PoE gigabit switches out there, especially used business switches from HPE (I like the “1920s” model) or Aruba, even Cisco and juniper. Switches are usually pretty bullet proof unless you really abuse them, but companies don’t like to have out of warranty stuff in their network, so the units get flooded onto the secondary markets after a while. PoE is good if you don’t want to run power to access points, though you can also use PoE injectors… Speaking of access points, my favorite non-ubiquiti access point for the home is the Aruba instant-on series. They’re managed from a free cloud portal, and I have yet to find a situation where they would do a great job. The only pinch is that when the internet goes down, you lose any ability to manage them. They’re otherwise perfect. A step up from that (and not to be mistaken for it) is the Aruba instant series; they’re a lot more expensive but locally managed and have a lot more settings… It can be a lot more difficult to get up and running with the instant series if you’re not completely familiar with networking, so buyer beware here too.
If your place is smaller and you only need one access point, the best option I’ve found, for price and performance is enGenius. The locally managed ones… I haven’t kept up with their product line, but I know they were moving to cloud management for most of their stuff. So finding the right access point may be a bit of a challenge. Also, I’ve used their access point systems (for multiple access points) and found it lacking, granted, that was several years ago now, so hopefully things improved, but I would recommend the Aruba instant-on for multiple access point systems over enGenius any day of the week. enGenius has the stand alone access point crown to me though.
For WiFi 6/6e/7, it will depend on your situation. If you’re in a high density environment, you may get benefit from 6e/7… If you’re in the suburbs or in a more rural situation, 6 is going to be fine. Just make sure you’re buying dual band access points. The big benefit for 6e is that you gain access to the 6Ghz band, which is only really helpful if your local 5Ghz band is overcrowded, like in the case of a high-rise apartment (even then, it may not be all too helpful), and WiFi 7 builds on this with some multiple access performance improvements as the key features. Meaning fewer slowdowns from many things on the WiFi at once. WiFi 6 already has some multiple access improvements, most notably MU-MIMO, so, I wouldn’t worry about it too much. 6e is going to simply reduce the chances that you have to share a 5/6Ghz channel with anyone else. The issue with 6ghz is that not all devices support it (laptops, cellphones, etc), and certainly no electronics (IoT stuff, printers, TVs, etc) will support it.
Personally, if you’re not into networking as a hobby or a full time job, then I’d usually point you at the ubiquiti UDR, and add one or two access points depending on coverage requirements. Either going for the U6+ (if you absolutely need to save money), the U6 pro for ceiling mount, or the U6 mesh if you need to plug it in and place it on a surface (The mesh is a bit more money due to it being so compact). For more hardwired ports, if you want more access points or something, maybe the Lite 8 PoE… There’s also the flex mini if you only need a handful of extra ports.
If you want something more challenging, using an opnsense router of custom built off the shelf computer parts is a good option, paired with an eBay find of a HPE 1920s with PoE and a couple of Aruba instant-on AP22’s or something would be spectacular, but a bit more involved, and likely a bit more pricy, since I’m not sure how well the instant-on AP’s fare in secondary markets (I’m not sure if Aruba will let you register a used access point into the cloud controller).
The ubiquiti solution has legs but you’ll be limited by the fact that it only has 1Gbps ports… It’s powerful enough that you can fully use those ports, but it will never go faster than 1G.
I hope that helps. I know it’s a lot of information to process. I don’t know your situation and I can be much more specific about what to use without getting into a lot of detail about your situation. I don’t want you to have to post that information publically on the internet, so I wanted to give you the information to make an informed decision. I hope I have done that, and I wish you the best of luck with your future network.
Yeah this was a big change for my parents who always complained about slow wifi speeds. I spent a summer wiring the walls with cat6 and plugged everything I could directly into a big switch I got for them. Printers, desktops, start tvs, the lot. Then I set them up with multiple access points and voilà, the wifi is suddenly fast again without upgrading their plan
This is the way.
I don’t have the option for multiple access points in my home even though I would like them (I rent). But considering that I just had to explain the difference between Bandwidth and Frequency to someone on another thread because they couldn’t fathom that someone’s experience might be that 5G is unreliable and they have better service with 3G, I don’t think this is going to take off. Don’t even know if most internet providers in the US will do dual access points without it being prohibitively expensive, and I’m not sure the cost is justified.
On top of that, part of the problem with IOT is that those devices are required to run on the same network in order to talk to each other. I had at one point just had my IOT things on 2G band and the other devices (phones and computers) on the 5G band and that caused all kinds of problems on my network. Devices that wouldn’t communicate properly or would drop off randomly.
I do agree though with your solution, dubious though I am about it being implemented.
I had multiple wired access points in a rental apartment with high wife approval factor by running cables in cup hooks along the ceiling lines. They self-tap into the wall and leave little more damage than would be caused by hanging a picture. I put a hook every 18 or so inches for support, and ran a cable under a door, up the door frame, down a hallway and into the living room (from a bedroom where we had the network gear). From there, I only had to mount the access point and connect everything. The cables were always out of the way, and I used white cables to help hide them against the walls. It wasn’t perfect, but the limits of renting are a bit limiting.
I only had two in that place, which was plenty, and we never had bandwidth issues accessing the internet from the wifi.
The fact is, nobody is building homes with wireless access points in mind. Whether rentals, condos, or new homes, they don’t have access point hookups. Many are now being wired with Ethernet in the walls, but nothing in the ceiling. A skilled wireless engineer can easily take a floorplan and build a wireless design with access points in ideal locations for maximum coverage and speed. Simply doing this pre-work and installing ethernet in the ceiling at these locations, is all that is required, yet, I have yet to see any builder do it. The same can be done to retrofit rentals as people vacate units and updates/renovations are done, yet, nobody is doing it.
Those that own their home need only to find placements and pay someone to wire them in. The whole thing only needs to be done once, ever, and the locations should be fine for use for the long term.
Fact is, neither builders, nor property owners, nor homeowners seem to have any interest in the practice. The only time I have personally seen or known of any location that is properly wired for wifi, is when someone has hired such wires to be custom installed. If it was done by default, rather than as an afterthought, retrofit or renovation, then it may be more common that people pursue such solutions. The price problem is another major hurdle, though I stand by the analysis of people paying literally several thousand dollars on internet access from an ISP, compared to a few hundred at most for a router during the same timespan (usually 1-3 years). IMO, that’s a bit like paying $50,000 over 10 years for access to the highway, but never buying a car worth more than $1000. It’s silly.
My main point is, solutions to these problems exist, but people simply won’t pay more than a few months worth of ISP fees for their wifi hardware, then they expect it to perform well and last for many years. They will turn a blind eye to the environmental issues that plague them and instead blame everything on lackluster devices that underperform that they will continue to refuse to pay a reasonable amount to actually get quality access equipment.
I have popcorn ceilings (and those crumble quite a bit), and textured walls. I also have only one cable hookup. But either way, I agree with you. I’m also not really the person complaining my wifi sucks. 3 phones, 4 computers, and a handful of IOT devices. Plus a couple of gaming consoles and none of those things in use all at the same time to cause the kind of strain that would cause wifi drops and so on. I’m just pointing out that it’s not feasible for some people at all but more often than not it’s that nobody wants to up front expense. Thank you for clarifying though.