World's First Voice Call From a Free GSM Stack

zycx writes "As Dieter Spaar has pointed out in a mailing list post on the OsmocomBB developer list, he has managed to get a first alpha version of TCH (Traffic Channel) code released, supporting the FR and EFR GSM codecs. What this means, in human readable language: He can actually make voice calls from a mobile phone that runs the Free Software OsmocomBB GSM stack on its baseband processor. This is a major milestone in the history of the project."

GSM Full Rate patent

[tepples]

I thought the GSM voice codecs were patented by Philips, as described in this page about an otherwise Free implementation of GSM FR [quut.com].

Re:

[jack2000]

Think an open source phone that can't have all it's data downloaded by the authorities. Or maybe a phone that you can change everything about. No more vendor lock bullshit.

Re:

[mikiN]

Almost right. You would also need an bunch of open-source cell towers and an open-source backbone network to connect them together and to the rest of the telcos, nevermind a rackstack full of licenses to operate all that.
Leave anything in that list up to chance (read: COTS or outsourced) and you're back to square one: the Feds, foreign spyops and/or crooks _will_ have a way in. ...Oh yeah, you also need a platoon of volunteers to patrol all your cell sites and backbone 24/7 to keep the rodents, copper thiev

Re:

[mikiN]

whoops, think I misread GP's post when I typed up that sarcastic reply.

To prevent your phone's data from being downloaded, you don't necessarily need an open source baseband, you just need to tightly control its communication with the rest of the OS. Turn it into a black box with just the antenna, power, dialing control, audio and a data channel connected to it. Get rid of the tight coupling with the OS already.

Don't get me wrong, IMO open-source baseband software is uber-cool, but the Feds will never willi

Re:

[account_deleted]

Comment removed based on user account deletion

Re:GSM Full Rate patent

[GNUALMAFUERTE]

It is in fact the main codec used in most of my Asterisk systems. The implementation is 100% Free Software. Is it patent-free? No. But nobody has sued anybody so far, mainly because the big guys don't see its use in VoIP as a threat to their big-ass systems used in GSM networks. Now, this might be a different situation.

Got more links about GSM patents?

[ciaran_o_riordan]

If you've any other links, I'd like to add them here:

http://en.swpat.org/wiki/GSM [swpat.org]

Re:

[Yvanhoe]

Codec patents are not legal everywhere. And the places where they are not legal will probably have a monopoly in a lot of awesome things that are to come.

I live in USA, you insensitive clod!

[tepples]

Codec patents are not legal everywhere.

I can't think of any company whose customers all have the finances to emigrate from software patent countries.

Re:

[Thomas Shaddack]

Why emigrate yourself when your phone (or other device) can become an immigrant, even if illegal? I think it is called a mail order.

This is illegal, you know

[tepples]

Why emigrate yourself when your phone (or other device) can become an immigrant, even if illegal?

Because this is illegal, you know [youtube.com]. You're out of pocket $200 for a product that got confiscated at the border.

Legal schmegal...

[Thomas Shaddack]

The customs are a bunch of idiots who cannot recognize an op-amp from an ADC if they'd've stepped on a DIP one pins-up, barefoot. If they weren't, they wouldn't be working as low-level drones. Just get an offshore friend to take the thing apart and ship it declared as spare parts (send the plastic housing shell as a separate shipment, so the recognizable parts won't attract undue attention to the important parts), or get somebody to carry it in a luggage as a personal possession when traveling in. Works vir

Re:

[neonsignal]

Philips might claim that their patent covers it, but only a court can make this determination.

If we are going to talk about software and patents, then just about every bit of software in the world potentially infringes some patent, so it is hardly notable to say that this bit of software does.

What is notable about the project is not the codecs, but the integration of a full GSM stack.

Pardon me, but....

[WED Fan]

Pardon me, but what does this really mean? Does this mean that we could develop our out cell phones, a kind of born unlocked? Would this allow us to create our own devices that include GSM without relying upon the industry providing us feature sets we don't want or need?

Is this really historic, or just a really nerdy, geeky milestone?

In other words: What will this do for me?

Re:

[pyr0r0ck3r]

If I'm reading correctly, yes, eventually you could roll your own phone using completely open source stuff - hw and sw. What I'm less clear on is how the signal is being carried. Granted, I didn't read too in depth, but you still need a carrier to allow your phone on their network, no?

Re:

[tepples]

you still need a carrier to allow your phone on their network, no?

You must have GSM confused with CDMA2000. In GSM, as I understand it, carriers don't allow handsets on their network; they allow SIMs on their network.

Re:Pardon me, but....

[Gruturo]

You must have GSM confused with CDMA2000. In GSM, as I understand it, carriers don't allow handsets on their network; they allow SIMs on their network.

Not quite. The phone has to be allowed as well, its maker and model are sort of embedded in its IMEI and there are blacklist (not just for stolen handsets, but also for models with critical radio flaws which would not work or even disrupt the network in their vicinity while operating).

IMEI blacklisting practices

[tepples]

its maker and model are sort of embedded in its IMEI and there are blacklist (not just for stolen handsets, but also for models with critical radio flaws

Thanks for pointing this out. But every time I searched Google for gsm imei blacklist plus some other keywords, there were so many results about stolen handsets that I couldn't find any related to radio problems. Can you provide a reference or other keywords that would help me learn more about how carriers manage their IMEI blacklists?

Re:

[tepples]

Thanks for mentioning EIR. I learned that the EIR splits IMEIs into three categories: black, gray, and white categories. Black IMEIs are supposed to be stolen or known to cause network problems, gray IMEIs are supposed to be those that may cause minor problems, and white IMEIs are supposed to be those certified by the network operator. What remains is how GSM networks, including the two major U.S. GSM networks, would classify unknown makes and models: would they be gray by default or black by default?

Re:

[J Isaksson]

Not black at least, I've been roaming with various European model cell phones while visiting the US without any problems.

Re:

[dgatwood]

Matters not. The IMEI is typically controlled by the radio firmware, not hardware, AFAIK. If you're writing your own firmware stack, it should be trivial to lie and claim to be supported hardware, and there's really very little the carrier can do about it (except in countries where doing so is illegal, of course).

Re:

[RockDoctor]

If you're writing your own firmware stack, it should be trivial to lie and claim to be supported hardware, and there's really very little the carrier can do about it (except in countries where doing so is illegal, of course).

Brushing aside the legality of the misrepresentation, what could the carriers do about it? The only thing that I can think would be that they would have to implement a whitelist of known-good phones (individual phones, not phone models) and to check for duplicate appearances (so that if

Re:

[HybridJeff]

To my knowledge whitelist phones are normally those branded and sold by the provider, greylist phones would include all unlocked devices ported for other providers and blacklisted devices are usually those reported as stolen or disabled for other reasons (for example if your provider provides a subsidized or free replacement for a phone you claim is broken).

Re:

[mikiN]

Depends on what you mean by 'unknown'.

Unknown to the carrier is a no-brainer, they must be able to admit imported phones on their network (unless they sell only sim-locked contracts and allow no inbound roaming).
I believe (haven't checked fully) that IMEIs for all GSM phones need to be registered with a central database [gsm.org] before they're released into circulation at all.

Depending on strictness, unregistered devices may not be allowed on the network at all (India) or be very restricted in available features.

Re:IMEI blacklisting practices

[Gruturo]

The white/black/grey lists are held in the EIR (Equipment Identity Register), which may or may not exist at all (it's optional, some providers don't have one) and is sometimes integrated within the HLR

This is an explanation (a bit dated, but still) of how to decode manufacturer code, country code, approval code etc from the IMEI: http://www.cellular.co.za/ieminumbers.htm [cellular.co.za]

More info (just relevant stuff which came up googling "imei hlr eir"):
http://www.linkedin.com/answers/technology/wireless/TCH_WIR/612218-35166861 [linkedin.com]
http://www.linkedin.com/answers/technology/wireless/TCH_WIR/608687-35166861 [linkedin.com]
http://www.wordiq.com/definition/HLR#EIR [wordiq.com]

Brief description of the (global?) IMEI DB at the gsmworld site: http://www.gsmworld.com/our-work/programmes-and-initiatives/fraud-and-security/imei_database.htm [gsmworld.com]

Re:

[jack2000]

Your phone could just LIE about it's IMEI, as i see it if it's all open source you can make it return whatever IMEI you want.

Re:

[dave420]

Which is illegal in some places, such as the UK.

Re:

[Thomas Shaddack]

It's illegal only if you get caught. Which, if you don't cause significant billing or technical anomalies, is rather infinitesimally low chance. I'd say well-worth the risk.

Re:

[Shoe Puppet]

I bet you can fake the IMEI if you control the GSM stack.

Re:

[Doc Ruby]

So does that mean that this "open phone" still needs a SIM issued from the carrier, like AT&T or T-Mobile?

The phone also has to have an IMEI that the carrier recognizes, which means it can't be generated arbitrarily by this open phone, right? If so, can I clone the IMEI from a phone the carrier issued to me over to the open phone?

And will this stack run on an Android phone?

Other than those two dependencies, could I just switch over to an open phone I install on an Android phone available today?

Re:

[Anonymous Coward]

The GSM frequency bands are licensed spectrum. You can not legally operate a DIY device on these frequencies without prior approval of the license holders. (There is a loophole: some of the frequencies used for GSM in Europe overlap with amateur radio frequencies in the US, but then you have to operate your own "network" as well, not just the handset.)

Re:

[tepples]

You can not legally operate a DIY device on these frequencies without prior approval of the license holders

As I understand it, the license holders are the carriers. I searched Google for get handset approved on t-mobile, but I couldn't find anything relevant.

Re:

[schnell]

Here's AT&T's handset approval and certification [att.com] process as an example, and here is Verizon's [verizon.com]. Nearly all carriers around the globe have them - some are very rigorous and demanding, while others are not much more than checking your CTIA and [your country's version of the FCC] radio performance certifications.

Regarding your specific example of T-Mobile USA - their certification process is known to be really easy, which makes things less onerous for handset developers but also doesn't catch sometimes seri

Re:

[CyberDragon777]

You subscribe to a carrier and put the SIM card in your unlocked phone.
At least this is how it works in Europe.

Re:Pardon me, but....

[drinkypoo]

What this means is that it is now theoretically possible to have a phone with zero closed source code. So far all phones have had at least proprietary radio module code.

Re:

[sznupi]

All those data & authorisation are accessed via applications running on what is really a smart card.

Re:

[JuniorJack]

According to the wiki, it could be used only with openbts due to legal obstacles - FCC licensing and so on

Re:

[BorgDrone]

If you start working on that phone now, you might even get a few months use out of it once you're finished. Don't expect too many GSM networks to still be around in a few years time.

Re:

[tepples]

If you start working on that phone now, you might even get a few months use out of it once you're finished. Don't expect too many GSM networks to still be around in a few years time.

Keeping GSM turned on is the only way that can AT&T's Christo-inspired TV commercial can claim 97% coverage. There's no way that AT&T will have UMTS everywhere and that even the cheap GoPhone handsets will support UMTS by the time it deploys LTE.

Re:

[nurb432]

Does this mean that we could develop our out cell phones, a kind of born unlocked?

Which you can get all day long now from china wholesalers, who can afford to build the devices. ( setting up to produce devices like a cell phone isn't exactly trivial ) Sounds more like just 'cool factor' to me unless I'm misunderstanding something.

Congratulations!

[gweihir]

Sounds like a pretty impressive feat. Shows that talented, dedicated individuals collaborating in a small group are still by far the most effective way to create software. All that "process" and "management" BS can do is decrease the performance of talented people. And with untalented ones, the final product will always suck, no matter what "process" or "management method" is used.

Re:

[Anonymous Coward]

WTF are you smoking? This is open source catching up with what proprietary software has done for YEARS. And sorry but,

All that "process" and "management" BS can do is decrease the performance of talented people.

[citation-required]

so agreed. he was just pandering to an attitude very common on slashdot and got rewarded with plus mod points

Re:

[LingNoi]

I love how whenever someone does something fricken awesome, like makes a way for you to set up your own phone network for free minus hardware there's always some dumb shit on slashdot ready to poop poop the whole thing.

Re:

[sznupi]

I would be a bit surprised if GSM goes away anytime soon. Most of the world is still on it, not even on 3G/UMTS - I expcet this one to go away sooner, replaced with LTE when all handsets (used by those who care about bandwith) will have it.

But GSM...it looks like one of those "good enough" things, especially if you want to maintain wide coverage.

Re:

[silanea]

With an open source GSM access point [wikipedia.org] to go with your open source GSM stack you can create your own GSM network, both towers and handsets - if you can afford to ignore frequency spectrum regulations and and patent issues. Think developing countries that want to roll their own gear and not let foreign companies take their market. And GSM will be around for a long time. You would not believe what kinds of devices use it to communicate. Even if all handsets switched to UMTS today there should be sufficient dema

First Call

[Nerdfest]

"Mr Stallman, come here, I want to see you!".

Re:

[Anonymous Coward]

As far as I know, the primary reason why rms doesn't have a cell phone is because he doesn't want to be tracked [stallman.org], and not because it contains non-free software (though that may be an additional reason...):

Police in the US use cell phones to track people's movements, real time. They can collect records of your past movements without meeting even the usual standard for a search warrant. Now courts are considering whether they must meet that standard for real-time tracking.

This is why I do not have a cell phone: I don't want to give the police a record of everywhere I go. It's not that I have something specific to hide; rather, it's my duty as a citizen to resist the total surveillance state.

Re:

[TheRaven64]

Meh. I have a mobile phone and mainly leave it in my house when I go out, unless I am actively expecting a call. The police are welcome to track it - maybe if I decide to become a criminal I can use it as an alibi...

Re:

[sznupi]

You can't be tracked if the phone is turned off... (heck, it doesn't even need to have a SIM card; still useful in emergency situations)

Re:

[GrumblyStuff]

Is this the same phone that still dials 911 when off? Or do you remove the battery when you don't want to use it? (And looked inside it for another.)

Re:

[sznupi]

Turning it on for 112 is not so hard. And don't believe in too many conspiracy stories...

Re:

[OutputLogic]

Somebody is calling from a stack of free GSM

Re:

[sznupi]

I don't see how this way of bootstrapping Stallman into phones could work - how will ge answer the first call?

(yes, yes, POTS - but I imagine you have to search nowadays to find a phone which doesn't have some closed firmware)

Re:

[OldeClegg]

rofl wild applause!

Congrats!!

[udippel]

Not only Free and Open Source Software, you also beat a lefthanded iPhone hands down with TWENTY minutes of call time!

Well

[ledow]

According to the all-knowing Wiki: "phase I of the GSM specifications were published in 1990"

So, depending on your point of view:

- it's taken 20 years to implement something that had a published standard and worldwide, cheap hardware examples used by millions of people every day.

- the standards took 20 years for an outsider to be able to implement them independently.

And we're still only talking alpha code with specialised hardware.

What could cause a 20 year delay ...

[Anonymous Coward]

... perhaps the first patents expired.

Re:Well

[blair1q]

Actually, it's been done dozens of times before.

By people who had proprietary knowledge enabling them to use the hardware properly, and hardware to do it on.

The software is not that special, and the system isn't either.

It's constructing the electronics that are capable of doing all the things needed to get the job done that slows you down.

Big companies have $billions to invest in making complex micro-gadgets that they can sell for a $thousand each other big companies who can find millions of little people to rent them for a $hundred a month to send sexts and tweets. You expect things to get done in that business model.

People with the word "free" in their corporate charter, not so much.

Besides, there were other things we wanted to get done [gnu.org].

Re:

[Man On Pink Corner]

One of the reasons it took 20 years is that for most of that time, you had to be (or pay) both a hardcore software dev guy and a hardcore RF guy to even think about trying. Now, GNU Radio and other low-cost SDR platforms have largely taken care of the RF side. That is something that will remain true no matter what kind of obscure protocols the carriers adopt for their next generation phones.

Put another way, it's now just another software problem, and we all know how much that changes the development pictu

New Famous Words?

[The Wild Norseman]

'Herr Watson--hergekommen--Ich möchte Sie sehen.'

Mr Watson

[PPH]

Can you hear me now?

Unleash, the patent lawyers!

[gnarlin]

You can bet your sweet ass that ass soon ass this software stack gets stable and usable for normal people there will be swarms of lawyers from the big telecoms attacking this project.

Re:

[TheRaven64]

The GSM spec was published 20 years ago, and will count as prior art to anything in it that wasn't patented before it was published, so it's unlikely that there are any patents outstanding in GSM...

Re:

[zeropointburn]

You mean the GPL? That would be free software for the end user, no restrictions. If you become a distributor, then there are certain reasonable restrictions. So long as you don't distribute the code, you can do anything else you want with it anywhere, any time. Certainly that doesn't qualify as 'many limits', nor does it even apply to the user.
Users are better served by the GPL. Distributors (particularly commercial distributors) arguably are better served by BSD/Apache-like licenses such as you re

I would like to say:

[tigerbody1]

Congralations to the team!

Phreaking

[uvajed_ekil]

A new era of phreaking is just around the corner, with commodity hardware, free software, and the will to continue to hack service networks. We're not there yet, but it is looking more and more like we may get there. Not a modern day equivalent of the good ol' blue box yet, but we'll keep trying.