5 minutes
Contemplating 10 Gbit/s internet at home
Revisiting all my running contracts for possibilities to decrease the price or increase the service performance, I came across the announcement of my ISP Digi Mobile that they are extending their own network to my address in the next two months and are therefor able to offer 10 Gbit/s internet to my home.
As a tech enthusiast, I am thrilled to have the possibility to get 10 Gbit/s internet for dirty cheap, because bigger numbers are better, aren’t they. But looking at the whole picture, I am not sure if it worth or even necessary for me. So I wanted to take the decision apart and look at the different aspects.
The current situation
I am currently living in a small apartment alone with my homelab. I currently have the 1 Gbit/s offering vom Digi Mobile and I am highly satisfied with the speed, the stability and the service. They are a competent ISP that offer to get around CG-NAT for 1€ a month and will give you the credentials to run your own router without much hassling.
1 Gbit/s internet is amazing to match the 1Gbit/s LAN speeds in my home and the Wi-Fi 6 speeds of my access points. With commodity hardware today you are able to match and use the full 1 Gbit/s speed with a single client and optimizing bottlenecks is mostly a thing of some configuration changes here or a better cable there. If you are lucky you get a server in the internet that is able to provide you with the full 1 Gbit/s speed for downloads, or you can use a torrent with a lot of seeders, as they already do for Windows Updates, Steam downloads and other game launchers (WoT, WoWs, etc.).
The hardware situation
Upgrading to 10 Gbit/s brings with it the first challenge: The hardware.
While 1 Gbit/s is common in even the cheapest of consumer hardware and 2.5 Gbit/s making it’s way into the consumer space especially marketed to gamers in the mid-price section, 10 Gbits/s is still a enterprise market reserved for companies with reasons to spend it.
Most networking equipment companies (Mikrotik, Ubiquitiy and others…) are already offering 10 Gbit/s hardware and are slowly moving into the prosumer space, those are mostly switches and routers that offer at most one or two 10 Gbit/s ports. Used 10Gbit/s enterprise hardware is available relatively cheap, as you can get used Mellanox cards for 35€ on AliExpress, but that means sparing a PCIe slot that you may not have or want to use for that, completely excluding any mobile devices without a PCIe slot.
The next challenge is the cabling. While 1 Gbit/s can be achieved easily with even the cheapest Cat5e cables, 10 Gbit/s needs Cat6a or better and even then an imperfectly crimped port or plug can already cause problems and you are better off running a certified cable all the way from your router to the client. So you are off scouring the internet for solutions to upgrade your home network to fiber.
The software situation
On the software side of things, 10 Gbit/s is a bit of a mixed bag. Depending on which hardware you use you may end up with different management UIs and OSes to manage your network. When going with MikroTik you are able to use their RouterOS/SwitchOS, while Ubiquity offers the Unifi OS. Mixing and matching hardware is possible but leads to a overhead in management as well as the inability to use the full feature set of for example Unifi OS.
I am a big fan of OpenWRT, as it is a open source project that is able to run on a lot of different hardware. Since the hardware for 10 Gbit/s is still enterprise and prosumer space and OpenWRT mostly focuses on liberating consumer hardware, there probably won’t be any support for 10 Gbit/s hardware in the near future.
The conclusion for myself
Even though my nerd heart tells me to upgrade to 10 Gbit/s just for the numbers, I am pretty sure that for now it is not worth the investment. I am not running a business from home, I am not running a data center from home (yet) and the ability to use 10 Gbit/s on a single device is limited to a few use cases. On the other hand I would have to invest thousands of euros to get my home network to support 10 Gbit/s for more than a few devices and even then the bottleneck starts to be the hard drives in my NAS or on my Hetzner server.
For me, 10 Gbit/s doesn’t make sense if you don’t have at least 10 clients working concurrently on the network, doing some very heavy work depending on the network, like the video editors at LTT working of the same NAS or backing up a massive amount of data each night. For every other use case, 1 Gbit/s is more than enough and probably will be for the foreseeable future.
So as you can see, there are very little reasons in upgrading to 10Gbits/s aside from “I want to toy around with it”, which in itself is a valid reason. While upgrading from whatever you had to 1 Gbit/s was basically a no-brainer, this is a hard sell and I will probably wait until the next upgrade cycle to take a look at the situation again.
Let me know what you think in the Hackernews Discussion
The thoughts of the internet
These are some of the videos I watched to get a better understanding and hear some other opinions on the topic: