The shitshow called the mobile Internet

Updated: October 21, 2023

Sometimes, I feel like I have too much time on my hands. Now and then, I take on an adventure that seems to bring mostly frustration, and no great practical value. Recently, I did just that. I decided to see how "good" the mobile Internet really is. After all, there are four transport media one can utilize the Web with - cable, phone, satellite, and mobile (radio). For most people in urban settings, the first two are the go-to options. In rural areas, sometimes, satellite and mobile are the only available options.

However, recently, more and more people use their phones as the primary computing device, and often they prance about the net without having a fixed connection at home. In other words, there be no physical cables going underground and into their living space, it's all done over-the-air. Put a SIM card into a device, get a radio signal to the cell tower, boom, Bob's your uncle. The question is, how viable is this for any sort of serious network usage? Well, Dedo the great sufferer decided to undergo several months of hard testing to answer that.

The experiment

My quest to figure out the quality of cellular slash mobile network usage took me across multiple countries. I used a whole bunch of devices, including several dedicated mobile routers, and I tried to be serious about it. There was a slew of important things to check, important things one would do: basic surfing, mail, sustained downloads, multiplayer gaming, VPN, hosting a server, chat and video calls, the stuff you'd expect people to do online. And there are a lot of findings here, so let's break then down.

The so-called 5G connectivity

There's an almost impossible variance of service provision, country to country, place to place. The element of randomness seems so high that there's really no pattern one can find in this whole chaos.

For example, in one particular setting, with an LTE Cat12 smartphone, I was able to get 410 Mbps with a 4G SIM roughly 200 meters from the cell tower, direct line of sight. Just 700 meters away, with a bunch of buildings in the way, the download speed averages 40-60 Mbps on a good day.

Cell tower

ISP companies

This turned out to be another interesting part of the experiment. The beauty here is that you can do a lot of testing really quickly. Get a SIM card, put it in a router, test, replace, repeat. Here, I discovered some rather surprising elements.

Now, let's talk about traffic shaping, throttling, caps, and whatnot:

Here, I was actually pleasantly surprised. This could be part incompetence, part policy, but most mobile network operators do not seem to meddle in user's traffic. Probably because the networks are over-used anyway, so there's no need. But let's notch down on the cynicism, and be optimistic and say this is actually genuine care for the customer. Serious face.

Hardware

This is the wildest part of the whole thing. Buying network equipment for your mobile network. Here, I encountered the highest level of disappointment among all the different usage vectors. If you want to buy a SIM-only router, you're in for a lot of anger.

But wait, it gets worse and worse and worse!

Antenna connection

A mobile router with an SMA coax external antenna connected, which will "hopefully" boost the signal.

Oh, let's crank it up, shall we ...

And that brings me to the end of this fine, multi-month rant.

Conclusion

You will notice I have not mentioned any manufacturer, device or ISP in this article. The reason is, I have not done sufficient testing to lay blame at anyone's door. That requires further testing, a lot more data, and more than one person's setup. The thing is, when you have so many different factors mixed, it's impossible to control and account for every parameter all the time, and therefore, it's all too easy to dismiss results as a glitch in "one of the components" other than the one you suspect. Which is often what happens when you "call" support and tell them about this or that. They immediately send you to one of the other entities in this unholy union of technology and sadness.

Lastly, it's not like people have any real choice. They can often choose only between one, two, maybe three providers if they're lucky, their living location will often dictate the choice, their budget will dictate the mobile plans and the available hardware, and their level of technical skill slash competence and actual needs will narrow the options even further. Most people will go for whatever is cheapest and quietly suffer.

The main purpose of this article isn't to say: X bad, buy Y. It would actually be good if that were the outcome. No. The unfortunate conclusion is that the ENTIRE mobile network landscape is pretty much bad, and way worse than if you have the option of using cable or phone (fiber, whatever). And it's not like the traditional Internet providers are any better. Just margins of awful and slightly more awful.

But what really surprised me is, we have this relatively "new" medium, we're trailblazing with 5G, there's potential to transform the worldwide networks and give people high-quality utility, and instead, it's a morass of mediocrity. I had low expectations, but even these weren't met. I was merely looking for: stable provider, stable connection, okay hardware, okay usability. By and large, I wasn't able to find this anywhere. Not a single country, place, or any which mix of hardware ingredients gave me satisfactory results. At best, I have average service mated to average hardware, average and highly varied throughput, and a great deal of everyday frustrations. You can't take the mobile network for granted, you must always fiddle.

The year is 2023, not 2003, and Dedoimedo is sad. Bye bye.

Cheers.