SrsRAN: Open-Source 4G/5G | Hacker News Reader

SrsRAN: Open-Source 4G/5G

github.com

195 pointsgballan65 comments

Can I use this to run my own mobile network? Is there something like a blank SIM card which I could use for it? I don't need global coverage but is it possible to create my own BTS on a PC (with some antenna connected to it) and then have my own SIM card which I can insert into regular phone/device and have it connected to my BTS and connect to the Internet?

Yes and yes, although in most parts of the world it won't be legal when done via antennas. You can buy blank SIM cards from vendors like Sysmocom, which are preprogrammed but "writable" SIM cards. The important part is knowing the private key that is used to authenticate end user devices.

Then you'll just need a decent SDR and it actually works fairly well for small test setups.

Nitpick: in most places you need a license to do this with antennas (in some places temporary licenses aren't that hard to get for temporary/noncommercial use)

Do you have an idea what would be my external IP address? On my phone connected to a mobile network, I get assigned mobile IP address which is my external IP address. It's not attached to the SIM card because it changes when I reconnected. Is it handles by the BTS software? Do I get assigned an IP address and BTS communicates on my behalf using that address which comes from the mobile network operators pool?

The LTE core assigns an IP address to your SIM card/UE. In the case of srsRAN, there's a simple CSV file mapping the IMSI to an IP address, but there are no limits on how complicated it can get.

The hard part will be legal. You will need permission to have a radio and that often means certification of the system along with right to use the frequencies phones use.

Yes, is simple enough to get started if you have access to the required hardware. You might able to operate in the n41 band (2.4GHz, like wifi) even without a license.

Phones are nitpicky about network configuration (chippers, emergency calling and so on). I would recommend starting with a USB modem. Also setting your network PLMN to the 00101 (the testing one), as it usually gets preferencial treatment in UEs.

N41 isn't 2.4ghz. It's licensed 2.5ghz. You're thinking N46 which is unlicensed 5ghz.

LTE didn't allow for it to be used as primary carrier, but NR-SA I think does.

This site[1] details setting up an LTE network, including programming a blank SIM.[2]

[1] https://www.quantulum.co.uk/blog/private-lte-with-limesdr-an...

[2] https://www.quantulum.co.uk/blog/private-lte-with-limesdr-an...

Official srsRAN project website:

https://www.srslte.com/

<shill mode> you can run SRSRan on a small self contained SDR such as the Seeve board. Throughput is not going to be good as a custom stack as it's unoptimised, though.

https://rfnm.com/blog/introducing-seeve

> After a NDA with NXP...

It is better than NDA with Broadcom, but still...

It's incredible to see such capability on such a tiny board.

The LA9310 has incredible specs, too, what a beast.

Have you done anything with its Network Listening feature? I'm curious if that's frequency flexible or if it's limited to a specific handful of common bands.

How have you found the NXP SDK, is it reasonably decent to work with?

Could anyone more knowledgeable on the topic explain to what extent common wireless connectivity standards are open and feasible to implement for, say, a medium sized company? Apple has been working on a 5G modem for what feels like a billion years, but other standards seem to be more democratized.

What do you mean by implementing? Make your own radio chips, designed from the ground up? Or merely producing a networking device using chips from suppliers like Intel, TI, Broadcom, Qualcomm etc? Or the software side only?

The availability of hardware seems semi moot, since afaik there's basically no way to get spectrum short of big national auctions.

But now that T-Mobile is renegging their promise & not going to meet the minimum deployment size they promised, they have been saying the FCC should find a way to sell by area some of that spectrum sitting dormant in such a wide wide % of America (personally I think it makes their bid invalid & they should forefeit their bid for such egregious dirty lying). https://www.lightreading.com/5g/t-mobile-relinquishes-mmwave...

I think some of the analog tv spectrum has some precedent for being sold per-area rather than nation wide, but I'm not sure how that's been going.

In terms of hardware, there's some fascinating stuff. Facebook's SuperCell large-tower project showed awesome scale out possibility for large towers. Their Terragraph effort is spun out, and seems to have some solid customers using their hardware. Meta spun off their EvenStar 5G system, which has a strong presence at Open compute now. https://www.opencompute.org/projects/evenstar-open-radio-uni...

But it's hard to tell how acquireable such a thing really is. There's plenty of existing nodes out there too. It is unclear to me though how acquireable such things really are- there not being an open market, since there's no usable spectrum feels like a conundrum for the market, even though these are extremely high volume amazingly integrated advanced wireless systems that you'd think would be visibly prolific.

> The availability of hardware seems semi moot, since afaik there's basically no way to get spectrum short of big national auctions.

You can run 5G in the unlicensed spectrum. AWS can rent you hardware for it: https://aws.amazon.com/private5g/ - it's $5k a month per site. I know a plant that switched to that because they couldn't get WiFi to work reliably for them.

But even if you want to run within the licensed spectrum, local licenses for a couple of bands are cheap. I was involved in setting up a private network in the licensed spectrum around 10 years ago (based on https://aviatnetworks.com/ ), and a local site spectrum license was something ridiculously small (in the range of a hundred dollars).

It's expensive if you want to do it nation-wide.

WiFi, Bluetooth and Zigbee has bunch of public specifications and knowledge about it to make it feasible. AFAIK, the specifications for 4G/5G is publicly available but extremely complex + you'd need licensing agreements, pay royalties, etc. So unless this imaginary company of yours have specialized expertise in all that, it seems unlikely to be feasible.

Stuff for GSM/CDMA has been around for years, OpenBTS is the primary example. This is the first I've heard of anything more modern/complicated being implemented. From my understanding, a lot of the hard eng work is in the RF frontend and making it small/low power enough to fit in a phone for example. OpenBTS got around this by using existing SRDs for their RF frontend.

The big problem is patents and copyright. No common wireless standards are open. No wireless standards are feasible to implement. Seriously. It's that bad. Certainly a modern 4G/5G standard is complex from a hardware standpoint to implement - the way you usually do these is using a very powerful embedded DSP, which is also not open (Qualcomm Hexagon is the most reverse-engineered of these if you want to understand what's going on). But the thing that's holding Apple up is purely legal IMO.

If I remember correctly, all the documentation needed to implement a 5G radio approaches 10,000 pages. It’s not only insanely long and complicated but there’s a nasty path dependency with most of 4G which is why Intel and now Apple have such a hard time getting their radios to the finish line. Poaching a few Qualcomm or Broadcom employees with better salaries is one thing but without the cumulative expertise contained within the companies, it’s almost impossible to bootstrap a new radio.

>No common wireless standards are open, No wireless standards are feasible to implement.

What is definition of "Open" here?

The current submission is entirely about Open Source 4G/5G. Fabrice Bellard on top of the crazy amount of other stuff he did also made a LTE/NR Base Station Software [1]. WiFi and Bluetooth are also "Open".

>But the thing that's holding Apple up is purely legal IMO

People constantly mistaken having an open standard regardless of patents and an useable product on the market. There is no reason why you cant have a software modem aka Icera that was acquired by Nvidia in the early 10s. And there are no modem monopoly by Qualcomm which is common misconception across all the threads on HN and wider internet. MediaTek, Samsung, Huawei, Spreadtrum and a few others have been shipping 4G / 5G Modem on the market for years.

The only reason why Apple hasn't released a modem 6 years after they acquired the modem asset from Intel is because having a decent modem, performance / watt comparatively to what on market is Hard. Insanely hard. You have Telecoms from top 50 market each with slightly different hardware software spectrum combination and scenario along with different climate and terrains. It took Mediatek and Samsung years with lots of testing and real world usage at the lower end phone to gain valuable insight. Still not as good as Qualcomm but at least it gets to a point no one is complaining as much.

[1] https://bellard.org/lte/

> What is definition of "Open" here?

Patent unencumbered in a way that someone could make a commercially viable implementation as a "small or midsized" company, as the parent post asked. Open Source proves my point - the issue is not implementation (note - I'm not claiming implementation isn't hard, it is - I certainly know from personal experience that it is and I would never claim to be able to personally build an energy efficient 4G or 5G modem, but I don't think that raw engineering horsepower is what's holding Apple/Intel/NVidia back here).

> MediaTek, Samsung, Huawei, Spreadtrum and a few others have been shipping 4G / 5G Modem on the market for years.

The CCP effectively told Qualcomm to get lost in 2015 and Taiwan settled an antitrust agreement between them and MediaTek in 2018, so MediaTek, Huawei, and Unisoc/Spreadtrum are not good examples here. I believe the South Korean government also intervened on behalf of Samsung. Actually, the list of modem vendors you list pretty much matches exactly the list of governments who prosecuted, fined, and settled with Qualcomm for antitrust.

>Patent unencumbered in a way that someone could make a commercially viable implementation.

Doesn't this exclude all modern cellular standards then?

yes.

Fabrice Bellard open sourced a 4G (LTE) base station.

https://bellard.org/lte/

It doesn't seam open source?

> The LTE/NR eNodeB/gNodeB software is commercialized by Amarisoft.

> A UE simulator is now available. It simulates hundreds of terminals sharing the same antenna. It uses the same hardware configuration as the LTE eNodeB.

> An embebbed NB-IoT modem based on Amarisoft UE software.

Yep, the word "source" never mentioned.

I assume Fabrice Bellard will crack cold fusion as a side project some day.

From my limited understanding, the issue for Apple et al isn’t making a 5g chip, it’s making the chip small, cheap, power efficient enough and capable of having “decent” reception. I’d imagine existing patents by Qualcomm certainly make it a bit more challenging on terms of available (design) options.

> Apple has been working on a 5G modem for what feels like a billion years, but other standards seem to be more democratized.

The main problem is the sheer age of mobile phone networks. A phone has to support everything from top-modern 5G down to 2G to be usable across the world, that's almost as much garbage that a baseband/modem FW/HW has to drag along as Intel has to with the x86 architecture.

And if that isn't complex enough, phones have to be able to deal with quirks of all kinds of misbehaving devices - RF is shared media after all, and there's devices not complying with the standard, the standards containing ambiguous or undefined behavior specs, completely third-party services blasting wholly incompatible signals around (e.g. DVB-T operates on frequencies in some countries that are used for phone service in other countries, and often on much much higher TX power than phone tower sites). If it can't handle that or, worse, disrupts other legitimate RF users, certification won't be possible.

But that experience in dealing with about 35 years worth of history is just one part of the secret sauce - that just makes the costs of entry for FOSS projects really huge (which is why all of these projects I'm aware of support only 4G and afterwards since that generation is the first one to throw away all the legacy garbage).

The other part of why there are so few vendors is patents, and there is a toooooon of patent holders for 5G [1], with the top holders being either Chinese or known for being excessively litigious (Qualcomm). And even assuming you manage to work out deals with all of the patent holders (because of course there is, to my knowledge at least, no "one stop shop" compared to say MPEG), you still have to get a design that fulfills your requirements for raw performance, can coexist peacefully with almost all other users of the RF spectrum to be power efficient at the same time. That is the main challenge for Apple IMHO - they have a lot of experience doing that with "classic" SoCs, but almost none for RF hardware, virtually all of that comes from external vendors.

[1] https://www.statista.com/statistics/1276457/leading-owners-o...

Can someone explain the use case for this project?

Is it intended for device vendors creating phones and tablets or for telecom companies building their own infrastructures?

This is the tower side (RAN) not the phone side (UE). It's not clear who the audience is. I assume it's equipment vendors since running a real carrier on USRPs seems janky.

srsRAN is both. It includes the source for user equipment in its "srsue" directory.

Use cases are partly captured by the list of customers across the bottom of the srsRAN home page (https://www.srsran.com/). It's great for any company/research group smaller than a multinational base station manufacturer, at which size the big players don't want to talk to you.

A few applications I could think of:

- Testing of various kinds. You can get deep inside an OSS base station, inspect and debug things, but likely may not even look at the proprietary innards of a commercial BS.

- Emergency connectivity when the backbone is broken, or did not exist. Most people would bring mobile phones anyway.

- Study / teaching, like MINIX OS or trainer aircraft which you can take apart and inspect.

- Maybe very small local cells to improve connectivity in difficult landscapes (mountains, skyscrapers, underground).

> look at the proprietary innards of a commercial BS.

I was wondering a bit whether that means Base Station or perhaps something else...

> Maybe very small local cells to improve connectivity in difficult landscapes (mountains, skyscrapers, underground)

Something like this would've been useful in a couple remote networks that I worked on. Internet was easy to get with some ubiquiti wifi routers and antennae. Getting cell reception was a pain in the ass that involved setting up a repeater in the one spot of the valley where reception was kinda okay. I haven't looked into the project too much, but I imagine I could set it up to connect to a telcom tower and handle the wireless backhaul similarly to how I did the wifi.

Edit: An open implementation like this could lead to cheaper cells for large, concrete/metal buildings that typically have terrible reception.

Femtocell equipment exists; the real issue is whether any of your local carriers support it.

ccc camp et al always have a 3G tower built on the older version of this. I'm no radio person but it's awesome to get a conf sim card and hack away

Cellular comms. are all complex, proprietary, closed standards by big telecoms companies. There are some projects trying to solve this: OpenRAN, OpenAirInterface, O-RAN Alliance.

This project tries to build an open solution following O-RAN standards for RAN software. It is targeting telecom. companies for their Radio Access Network (RAN). Their infrastructure is divided into the Core Network (like a simple internet network) and the Radio Access Network (base stations, antennas and all that stuff). Phones and tablets are User Equipment (UE), nothing to do with srsRAN.

Why can't I create an ad hoc 4/5g network between phones?

Genuine question.

It looks so closed and complex compared to more traditional comms like wifi or even ethernet.

The LTE protocol is asymmetric between towers (eNodeB) and phones (UE).

You could build srsRAN and run a local eNodeB instance to form a local LTE network. Given that it has a radio that operates in the LTE bands, in theory you might be able to build srsRAN for a phone and run the necessary base station software (eNodeB and EPC) on a phone to form an "ad hoc" network. It wouldn't be truly ad hoc in that the network would still have a base station, just that it would be running on phone hardware. srsRAN wouldn't be running as an application, but would be replacing the firmware in the phone's LTE modem.

Edit:

Here is a paper by someone who has built an ad hoc LTE network.

https://arxiv.org/pdf/1802.09262

It does exactly the above: using srsRAN (which used to be called srsLTE) to run local base station software (eNodeB+EPC) to get around the asymmetry in the LTE protocol.

> Why can't I create an ad hoc 4/5g network between phones?

Because billing.

All the complexity in cellular can be traced back to making the system secure the whole way back to SIM cards because there is actual money at stake if the system gets cracked.

What about a "public good" mobile network without any billing, only fair QoS?

There's limited bandwidth available in the space. 5G is "fast" compared to 4G because 4G effectively saturates terrestrial bandwidth available. Public, unauthenticated network would be too easily DDoSable.

The complexity is actually down to mobility. You can turn your phone on anywhere and it'll work. You can have a call/data session while travelling hundreds of miles and it'll work (usually).

What you suggest could be accomplished with mesh networking. GP is correct in saying the reasons are orthogonal to functional reasons. Phone systems have always been centralised like this. Billing is one aspect, but lawful surveillance is another.

With respect, if you think that mesh networking would have worked considering the scale and capacity needed, and relevant type of devices then I am afraid this is a case of not knowing what you don't know...

How do you solve cross talk in congested areas? Made worse by the fact that mesh networks have a significant communication overhead.

Cross talk takes place on PHY and by the time you get up to routing it’s all been eliminated

I was using Mesh networking as an example. Obviously it can’t handle heavy load or longer distances well, however it is much more efficient for local point to point.

If you like, you can consider just plain old VOIP where the central entity is used for discovery but the actual call may then be routed p2p though not in VoLTE and its successors of course because the media plane is also centrally routed.

They are not entirely orthogonal. You need a central base station with scheduling, high dynamic range and power control to maximize performance in an OFDM system. In addition, most cells used to use frequency division multiplexing meaning that the base station and phones send and receive on different frequencies. So lack of point to point capabilities can at least in part be explained by the design goal of optimizing for throughput and user density.

You do if you want to route across a broader area which is a different use case to what we are talking about. Two phones right next to each other are routes through the BTS even though they could easily route directly to one another and that is for non functional reasons.

No, they could literally not send on the frequencies they listen on in case of FDD. Enabling this would require extra radio hardware. Also there would need to be a some kind of encryption key exchange between devices which is not needed in the centralized setup. They could not easily route to one another without adding extra stuff.

You’re talking nonsense we have an abundance of radio devices capable of point to point routing.

AFAIK cellular comms aren’t designed to be ad hoc. They require base stations (whose placement is very studied) to build coverage “cells” (cellular). The network is responsible of delivering a QoS, it’s not a best-effort like WiFi.

However, if I’m not mistaken 5G has V2V comms (Vehicle to Vehicle) specifically designed for self-driving cars, drones and IoT and that is the most “ad hoc” it can get

It's complexity is not completely unnecessary. Scale is a huge bitch.

Networking between phones, without a common base station, is hard. Harder than WI-FI, because identity is managed centrally and the frecs in use require a license to name some "simple" problems. Cellular comms are not designed to handle this kind of scenario.

There has been some work to add inter device connectivity lately, to support many interesting use cases. You can find more looking for "sidelink" or "PC5", but you will not find many phones that implement it.

It seems to be mostly focused in vehicular use cases in V2V and V2X scenarios

Last time I looked is because the chipset makers haven't implemented LTE Direct. There were a handful of phones with LTE Direct that were oriented towards first responders.

5G also has a peer mode, and I'm trying to figure out if it can be implemented in userspace or has to be done in firmware.

My idea is that peer mode would be useful for replacing walkie-talkie or mesh messaging like Meshtastic since everyone has a phone. The range is supposed to be around a mile.

Complex: yes

Proprietary, Closed Standards: no

Here are all the 3GPP 4G/5G standards that govern cellular phone networks:

https://www.3gpp.org/specifications-technologies

Also look under the "Specifications" drop down menu at the top of the page. Every release is listed.

These standards being freely available is why SRS were able to write srsRAN. There is nothing stopping others from doing so either, apart from the amount of information that has to be understood and digested.

Edit: Apart from complexity the other potential barrier is patents, but that's not a technical barrier.

While it's technically true that the specs are public, the implementation not being public is most of the issue being solved

Apart from complexity and patents, the other other barrier is licensing.

Those technologies aren't terribly useful to use in your house like WiFi is, and even if they were, you need more government paperwork to do it. That means less hobbyist interest.

The 3GPP spec is “available”, but the implementation in practice is closed. Stuff like the scheduler is what actually brings value to the company delivering a good service. Being closed, the interoperability of network components from different vendors is just impossible. That’s the point of O-RAN Alliance.

Patents equal proprietary.

They abandoning the UE development was a very sad moment.

Their code base is so much nicer than other projects, and somewhat easy to match with the standards. I am very happy user :)

O-Ran, the closed source that begins with O for open naturally, is really terrifying in terms of spying on users and sharing data.

My personal tinfoil hat says the corporations started the 5G cancer rumors themselves to distract the public from the real terror of 5G that is ORan data sharing.

The tl;dr is any participant in Oran can put software directly on the antena and access your data directly with no need for it to go through a 3rd party like your cell provider. The cell providers charge for access to the antenna. This sidestep any potentially conflict with selling your data - by choosing a 5g cell phone, you have already opted in to your data being accessed. Creepy.

Reply on news.ycombinator.com