I use TP-Link, too, for a few years. I have a wired connection from the router to the home office as we have GB fiber and 7 nodes throughout the house and even in the garage. Each node has 3 copper ethernet ports and can act as a mini-switch at its location. The user interface is near idiot-proof and easy to setup although there is a choice at initialization between passthru (your carrier's router) and router mode where you just take your connection and handle the routing. It is up to you if you trust your carrier or you want to do it yourself. They also have an option to connect via copper if you want more speed than the ~700Mbit over wireless.
We got a TP-Link mesh system in 2022. It was inexpensive, easy to install and set up, and it has worked really well since then.
Our home is wired for Cat-5, state of the art at the time but it doesn't handle long runs very well and now that I have 10G in the home office cluster including 2 Mac Studios and a handful of x86 Linux servers with 10G cards (now pretty reasonable) the Cat-5 cabling has run its course. I have 1 hand-done run of Cat-6a from our router to the home office and that runs at a full 1Gbit.
I used to get in long discussions/arguments about bufferbloat with a late friend of mine. It only affects WAN connections and is usually associated with older inexpensive router hardware with small buffer capacity that is overwhelming the local router. I don't think it is a problem much anymore (I could be wrong) but it has mostly to do with latency and not throughput. I host multiple open-source mirrors like Ubuntu and the Linux Kernel Archives and have symmetric GBit since dumping Comcast for ATT Fiber and my latency is really good at every endpoint, including for gaming. My personal belief is bufferbloat is probably not a problem much anymore because even cheap hardware has pretty good memory capacity so buffers are no longer constrained.I have been in IT since 1980, you can do the math.
We only had DSL until we received a Starlink dish a year and a half ago. When we were using DSL, I worked diligently to optimize our access. We were getting about 3-4MB. With that said, we were able to stream or work from home, but not both at the same time. This taught me to be extremely frugal and optimized everything , everywhere I could.
There are so many variables to consider, as I had as many devices hardwired as possible. Most of this was hard wired with 20+ year old CAT5. For instance, the distance from an access point will proportionally diminish the internet speed. Another thing to consider is an access point can only handle a set number of devices. For us I needed three access points to cover our 5200 sf home (2600 up and 2600 down) for the number of wireless devices we had, not for coverage. Two were in the basement on the corners on side of our house, while the third one was on the main level in the center of the opposite side. This was needed to give good coverage and capacity to handle our 50+ WiFi devices.
Always test internet speed with a hard-wired device. Another thing to consider Is bufferbloat. This is where many routers fail and a handful shine. Waveform’s website is a great resource of information. They have a great test to see how your system is performing. Here is the link to my data, you can click the Test Again button to check your performance. They also have suggestions for routers. I have a Ubiquity EdgeRouter 6P along with three Ubiquiti access points. My system is stable and connectivity is flawless.
Bufferbloat and Internet Speed Test - Waveform
Most households don’t need more than 50MB to cover numerous devices. I have software to monitor the network usage and it rarely goes above 25MB.
So, what I’m trying to get the point across is, you don’t need as much bandwidth as you may think to get reliable internet performance.
Here's mine: Bufferbloat and Internet Speed Test - Waveform
Wow, that's impressive. A little bit more latency when under load, but likely not noticeable. Thanks for sharing.
Was that directly connected bypassing all the old cat5 cable? I'm curious what the H connector is, was it powered?
Supposedly, they upgraded me to 500 mbps some time ago. I don't know why I don't get that. In any event, it didn't cost me anything and what I have is fine for my purposes.
Well, my first foray into mesh routers was unsuccessful. I tested the Netgear MK73S tri-band mesh router kit (router + 2 satellites). It is documented to cover a 4,500 square foot house. While the wired connections and wireless router connections were fine, the satellite speeds seemed to vary a lot with my testing, and the satellite on the top level of our house did not provide any better speed that my AC3600 router.This is a good thread for me to follow. It is something I need to get moving on. Our 3200+ sq house has 3 finished levels, and the router (Netgear AC3600) is in the finished basement because that is where 2 cable runs were done (for TV and router) and the majority of my wired devices are there. While I do get a signal throughout the house and outside, a quck speedtest gives results that are telling and intriguing:
- wired: 480Mbps download
- wireless, basement: 186 Mbps (phone) 90 Mbps (laptop)
- wireless, 1st floor: 125 Mbps (phone), 171Mbps (laptop)
- wireless, 2nd floor: 43Mbps (phone), 31Mbps (laptop)
I'm too lazy/unskilled" to fish wiring between floors; if we need wired speed on the 1st or 2nd floor, I run a long cable up the staircase and along the wall as needed... not aesthetically pleasing, but temporary, out of the way and it does the job.
My main requirement for a mesh network is that I do not want to be "forced" to use a phone app for configuration... call me 'old school', but I find it easier to configure via the web interface, particularly as I use a lot of configuration options.
Time to look again at what is out there. Maybe this will be my Christmas present for myself
.It has been returned.
How long should this test take? My test gets as far as "Warming Up..." and no further.
Same here
