top of page
Search
Writer's pictureHeikki Almay

Agile and affordable 4G/5G

This is a slightly modified version of the Agile and affordable 5G presentation at XCHANGING IDEAS #50 Global 5G Evolution - 4th June, 2024. YouTube at 45min.

The Poutanet journey to consumerized and affordable mobile networks started in 2015 when we still worked at Nokia. Facebook told us in the early days of the Telecom Infra Project that the way mobile networks are built is very old-fashioned.  A lot of expertise and manual work is required. Facebook told us we could do better. So we gave it a try. See Kuha video.


Over time Nokia lost interest in the topic. We decided to continue on our own and founded Poutanet. The only addition to the original ambition is that we now turn the entire mobile network into a consumer product, not just the base stations.


Who needs this kind of networks? At least those who bring payment terminals to events with a lot of people - as public networks get congested uploading all the pictures and videos people take when the rockstars get on stage. Building Wifi is an alternative but it takes a lot of time. With private mobile you can set up your network in 20 minutes and start operating.


So far this has been 4G, as digital payments are not very data hungry – but add video, and we see the use cases for temporary small private 5G networks.


Live on TV - but freezing

Right now broadcasters are driving the topic of temporary small private 5G. If you have to cover an event in a place where there is no fixed cabling for cameras it is time consuming to setup the systems. Wireless is much more convenient. A smaller sports tournament such as an orienteering event in a forest area is a good examples.


Another use case is live TV interviews. In parliaments you have these heated debates over some new legislation. Then all the party leaders come out and want immediately to say something to their favorite media. Usually now the transmission of the live streams is done by packing a number of cellular modems next to the video codec and spraying the traffic into all available mobile networks. This may work fine if you live in a two-party system, but if there are four or more interviews at the same time it is likely that you see the interview freeze mid-sentence and the news anchor apologizing and continuing with some other stories.


In this use case coverage is needed in a limited area indoors, and you could argue that Wifi would do the job as well. The tricky part with Wifi is the unlicensed spectrum. With several camera crews and all the additional journalists around with all sorts of wireless gear it is very likely that you have interference – and again there is a high risk that the live interview is cut short.


Really simple 5G

The simplest incarnation of a private 5G network is what you see in the figure above. This is all you need for the parliament building scenario. Note that in this setup all the 5G functionality is on site. So it is only for the applications that you need an Internet connection. If you want to send several stable 4K video streams, you should make sure that you have some hundreds of Mbit/s capacity reserved for connecting with your studio.

If you are remotely operating a forestry machine with your joystick and large screen in a trailer that parked at the roadside all the video needed for controlling the machine is consumed locally and no Internet connection is needed.


Add QoS to your network - you can go to conferences and talk about it

The figure above is an example of how to stretch it a bit and how to make real use of some of the more advanced 5G features. In a setup with several cameras you have locally the mixer who decides what camera stream is broadcasted. By selecting the stream you invoke a 5G application function that requests a higher priority for that camera. So you make sure the stream that goes out is always 4K disregarding of how many cameras you connect.


How to reduce complexity and cost - Wifi on steroids

So far most private mobile networks have been built in a traditional project mode. Planning and contracting, roll-out and maintenance… and then after five years repeating the process.


In order to make it easier to build small and affordable private networks that can be expanded and modified in an agile way it would be good to have a few new options in the standards.


SIM authentication is a relict . The basic concept is from the 1980s. Operators wanted to have a SIM in GSM for unbundling the subscription from the handsets. Those days some UEs had notoriously poor security and the result was identity thefts and all sort of trouble. Now when eSIMs are soldered to UEs we are to some extent back to the original setup [just waiting for some state sponsored hackers to crack the iPhone eSIM...].


Asymmetric authentication would be a good complement to SIMs. It would allow removing the centralized subscriber database, which is now for small private mobile networks the most important part of the mobile core. This would allow a lot of cool stuff. For example, as a user you could connect to a base station and negotiate with it what kind of service you would like to have today. For this Roam Networks has some working code already.


With SIMs made optional we could in small networks skip the rest of the core as well. For a security camera there is not much use of mobility management or a mobility anchor. For a sensor connected to a 5GLAN the need to run the traffic through a UPF is outright counterproductive. The base station is certainly located in the same factory hall as the target LAN. The UPF is not - unless you place a lot of extra boxes around your factory.


At this point you may wonder, what is left if we remove all the mobile networking features from our private 4G/5G. It is the better radio operating on regulated spectrum. Better coverage and less interference are the most important features. After all, most private mobile customers only need a routed (or in some cases switched) wireless access network. They need Wifi on steroids.


81 views0 comments

Recent Posts

See All

Comments


bottom of page