Upgrading my Networking Gear (2022.1)
I’ve been wanting to get out of the consumer networking gear game for a few years ago — roughly when we got around to getting our own place but the thing about hardware failures is that you can’t really schedule them. A few years ago when we had a sudden device failure, I wasn’t ready to make the leap so I bought what I planned to be my ‘last consumer oriented purchase’
I’ve been using the TP-Link Archer C5400X ‘gaming router’ for the last few years and have been super happy with it but over the last few months, we’ve had some changes in our networking landscape which are better served by improved networking — and now we have our own place to run wires etc. so I guess it’s time to make that leap.
Networking Series
This post is part of a series, check out the other posts!
Quick Links
How we got here
I didn’t really take networking seriously until (relatively) recently: I like to break my journey here into three eras:
The 'who knows’ era — this was the dark ages where I don’t really remember the gear involved. I know I used the legendary Linksys WRT54G and after that died, there was Belkin router briefly and then I moved to the D-Link DGL-4300. There was probably a smattering of ISP-provided modem-routers in there somewhere too
First ‘decent’ router (2015) — driven by the ASUS RT-AC68U, this router was the first to make the bold (at the time) claim of ‘60-second setup’ and by golly it delivered. I really, really abused this router, making it do things that it just wasn’t built to do: at one point, I had a USB hub plugged into the USB-share port and daisy-chained six external USB drives as a makeshift NAS of sorts… Despite the misuse, this router held up superbly - it was a sad day when the NVRAM literally wore out and the unit wouldn’t boot anymore
My ‘last gaming router’ (2019) — the unexpected death of the AC68U meant that I didn’t have the luxury of planning out my replacements so I had to get something on the spot so to speak. One of the concerns at the time (with the AC68U) was the poor WiFi signal two floors away — I was really hoping I could just buy “one thing” and have better reception upstairs without having to mess around with extenders, meshing or access points (we were renting at the time). I settled on the TP-Link Archer C5400X which was a home-run purchase… until my needs/priorities changed recently…
As you can see, the jump from the fantastic RT-AC68U to the C5400X was a flat-upgrade across the board; even when we moved to our current place, the AC5400X has held up spectacularly: it has more than enough power to give us decent coverage two-floors away and the actual performance and functionality has been great.
The catalyst for this was making a transition to 10Gbit with my desktop machines and the NAS — all which can be had for a relatively low investment of a 10GbE NIC for the NAS and each PC and (optionally) a cheap[ish] 10GbE unmanaged switch. Ironically, this setup worked so well that I would have been super content to stay with this for a long time.
Why Upgrade?
In the last few years, we’ve picked up a lot of ‘smart devices’ and I wanted to segregate them into their own network. Neither the ASUS nor TP-Link router could define VLANS but they both allowed for an isolated guest-network which was close enough for what I wanted to do. In the case of the Archer, I could create multiple guest-networks so that I could actually have a guest-network. You know, for actual guests. This was great.
Over the last year, I’ve moved certain devices (old phones, some tablets and e-readers off the ‘main’ network onto the guest network and I also picked up a lot more smart devices. This was when I discovered that with the Archer, there was a limit of about 50-devices that could be registered on the guest network. For any device that tries to connect to the guest-network after the 49th device, it would be a race condition of which device connected first. Since I plan on having more smart devices, a change was needed.
It was also kind of the plan to do a switch away from consumer networking gear once we got our own place and things were more stable.
About that 50-device Limit on the guest network…
To be fair, there’s nothing wrong with the Archer and generally speaking, I’m super happy with it and I totally understand that I’m misusing the whole point of the ‘guest network’ but hey, now I don’t have to worry about those kinds of limits anymore.
With my old network, I never really thought I’d ever hit that cap, after all I was only dealing with current (and old) phones, tablets and a few smart hubs and things like Chromecasts that might need to be segregated. A few years ago, I really went into smart-outlets, in particular those that offer energy monitoring: by having energy monitoring collected at the the device/outlet, I can get a better picture of real-world energy usage over an extended period. Later on, I supplemented my collection of these monitoring-outlets with more basic models that focus on the normal smarts like remote activation, scheduling etc. Generally speaking though, almost every device in the house can be traced with an energy monitoring outlet (and the count adds up in a hurry!)
It starts with a baby step
At first, I didn’t really want to go all-in for this — I wanted to alleviate the guest-network-cap issue with the least amount of cost and effort in mind. I picked up the TP-Link EAP225 (AC1350) access point and plugged it into the Archer. After some fenagling, it was a success - it solved my issue of allowing more than 50 clients to connect to the ‘guest’ network and was relatively straightforward to integrate into my existing setup.
It’s just a coincidence that I happened to get a TP-Link access point — any other (similar) branded access point would likely give me the same results — this model just happened to be on sale.
I had such a good experience directly configuring this access point that I wanted to try it out with a controller (which would be how I would run it in a more extensive environment). Enterprise gear from TP-Link can be managed in three ways:
Direct, manual management - good for very small environments; here you directly connect to each device and manage it via web-portal. There’s zero cost for this
Software controller - you can install a free software controller and manage lots of devices from one central location. There’s no cost for the software, but you do need to provision a machine for it
Hardware controller - you can also buy a device that has the controller built in and running on it. Other than the hardware purchase cost, there’s no ongoing cost for this
I should mention that the controller software is Java based. And Java is not allowed on any machine I in this house — so I ran it on a VM. The experience of managing the access-point via the controller software, even on a VM, even with Java was pretty great. I was hooked.
And down the rabbit hole we go!
With that great experience in hand, I went right down the rabbit hole:
Router: TP-Link ER7206. The TP-Link website lists two models, the ER7206 and the ER605. Their respective datasheets don’t go into too much detail but there seems to be a substantial performance uptick with the ER7206 so for an extra $50 at the time, it made sense to just get the ‘better’ model and not second guess any bottlenecks down the road
Main Switch: TP-Link TL-SX3008F. I wanted a 10-gig switch to branch everything off of and there were two choices, this 8-port and a 16-port model. Since I have an unmanaged 10-gig switch already, it made sense to save a bit of money and go with 8-ports. Also, the 8-port model is fanless (one less component to fail). Time will tell if going with 8-port was a good or bad call
Bulk switch: TP-Link TL-SG3428X. I waffled a bit on whether or not I needed/wanted this since I already had a 24-port unmanaged switch that I was using, but I wanted some more granular control of VLANs so I bit the bullet and replaced my old bulk-switch with this. This model is pricy because of the 10Gig upstream ports.
Controller: TP-Link OC300. I didn’t really want to provision a permanent VM to run the controller software (and there was no chance I was installing Java on a real PC), so I got a hardware controller who’s only job is to run the controller management software. TP-Link has two models, the OC200 and OC300. I didn’t to deal with slow restart times for the controller so I got the more performant model
Access Point: as mentioned, I got a cheap TP-Link EAP225. I didn’t want to invest in a much more expensive model since I knew that high-performance models were going to be launched in 2022 (and I wanted to wait for Wi-Fi 6E before I really looked at wireless). The nice thing about access-points is that they don’t really ‘go bad’ and I can always add some more to improve my coverage area
We're not done yet!
Before I can really explore any enhancements, I’d like to get Skype Station in the upstairs office properly wired up with a hardline (right now we have both Wi-Fi and a powerline networking drop upstairs). I’m hoping to run two cables upstairs which will give me some options for 1GbE, 2.5GbE or 10GbE.
Device Upgrades
Obviously, this wont be a truly final configuration - newer, better, faster stuff will continue to roll out — I saw on a few product roadmaps of a 10GbE router, the ER8411 (which popped up on their UK site). It would be nice to have a 10Gig stack top to bottom…
Multi-gig
For 2022, I’d really love to move past the 1GbE barrier for all the computers in the house. In a magical world of unlimited resources, moving straight to ‘10GbE for everything’ would be awesome but truthfully, certain machines simply wouldn’t be able to make appreciable use of all the extra bandwidth (— non-DTR, ULV-laptops, I’m lookin’ at you). By adding a 2.5GbE switch to the network, I can segregate my devices into fast (10GbE), medium (2.5GbE) and slow (1GbE or less) nodes — what a time to be alive when 1GbE is classed as ‘slow’.
I’ve been watching the TP-Link TL-SG3210XHP-M2. I haven’t bought this yet because I still need to get ethernet run upstairs to the upstairs office for Skype Station and possibly to the living room for the TV. Oh and because it’s expensive. This switch has POE (+$$$) and 10Gig uplink (+$$$) and all of the ports are 2.5Gig (+$$$) so I won’t tackle this until I need it.
Access Points
At some point in the next year, I’d like to cut-over all the major devices to Wi-Fi 6E although this is dependent on what future devices I get (and when I get them) — support for Wi-Fi 6E on laptops and phones will be a criteria I look out for. Once we get an ethernet drop upstairs, I will likely run a few access points to get more signal saturation. As my ‘primary AP’, I think I want to get a TP-Link EAP690E HD to replace my current AP in the basement. I would like to get an outdoor and in-wall access points, but I would really like them all to support Wi-Fi 6E before I commit to them.
Why TP-Link?
Let me rephrase this to what we’re all thinking anyways:
Why not Ubiquiti?
My first encounter with Ubiquiti left me with a really bad impression: it was awhile back with their G3 Pro camera (before I even knew they were knee deep in the networking space). At a glance I saw:
1080P capture (standard at the time)
30FPS (reasonable)
1/2.8” sensor
H264 encoded stream
$400 price tag when I saw it, $300 today, at time of writing!
I fully recognize that there is supposed to be premium for a POE, networked camera that has everything nice and neat and all together. But I still can’t get over that price tag for what amounts to less-performance than Go-Pro. Let’s be honest, those specs put it closer to a Logitech webcam. And that point, I wrote the entire company off as overpriced froo-froo. Even checking the current price, my blood boils at the $300 list price.
More recently, a good friend of mine who happened to have some Ubiquiti networking gear gave me a rundown of their [impressive] networking side and that their networking division is pretty cool and it’s just their camera division that is out of touch with pricing. Funny enough, Ubiquiti did come out with some pretty awesome surveillance cameras genuinely offer some novel features (like the Camera AI 360).
At the end of the day though, the pricing didn’t match the hardware specs for me (and I’m a hardware junkie); and I wasn’t about to play the ‘constantly being out of stock’ game.
Closing Thoughts
A coworker of mine warned me that this was going to be a rabbit hole and at the time, I only half-heartedly believed him. In hindsight, he was totally right and while this is fun and all, it was why I stuck with high-performance consumer-grade networking gear for so long. That being said, as much as the consumer-networking experience has grown in the last decade, I’m a little upset that I didn’t make the jump to business-grade gear sooner.
It might be a bit naïve but man, I’m really hoping that after I fill in the remaining pieces that I want (2.5Gbit, some more APs and maybe a router upgrade), that I’ll be done for a long-ish time. Ha! Who am I kidding…
Network Upgrade (2022.2)
If something isn’t broken, that’s a good time to upgrade it right? I can’t leave well enough alone so follow along for the next step in my network’s evolution.
Build Log: Network Rack Mk 2
With this networking change, this was a good opportunity to revisit the enclosure for everything.
Product links may be affiliate links: MinMaxGeek may earn a commission on any purchases made via said links without any additional cost to you.
