Osmocom + USRP

Alexander Chemeris alexander.chemeris at gmail.com
Thu Sep 8 07:50:47 UTC 2011

Hi Harald,

Thank you very much for the detailed description. I'll talk with my
customer and see whether they really want what they're asking for. I
proposed them to use OsmocomBB with Calypso, but their initial
response was "no-no, we want USRP".

On Thu, Sep 8, 2011 at 10:56, Harald Welte <laforge at gnumonks.org> wrote:
> Hi Alexander,
> On Mon, Sep 05, 2011 at 11:32:53PM +0400, Alexander Chemeris wrote:
>> I have a client who is interested in an USRP-based MS side implementation. I
>> think that that OsmocomBB is a natural choice for the higher-level stack,
>> but I have only vague idea about its current state.
> It's fairly complete, you can do FR/EFR/HR voice calls (AMR still
> missing), it does cell (re)selection, encryption, hand-over, etc.
> What's mostly missing is SMS support.
>> As I understand, burst coding/modulation, reference clock management
>> and GSM clock management should be implemented to get a working MS. Is
>> there anything else to implement? How well is API to Calypso DSP is
>> abstracted in OsmocomBB?
> The actual layer1 firmware (on the calypso ARM side) is probably very
> calyps specific, as it uses its various hardware components.
> Nonetheless, there may be things like the TDMA scheduler and multiframe
> scheduler which may make sense in different environments, too.
> The entire L1 is abstracted from L2 using an interface called L1CTL.
> This interface is specified as a protocol, and we normally speak it over
> RS232 to the calypso phone.  It is defined in l1ctl_proto.h
> As the calypso is the only hardware we support so far, there may be some
> leakage into L1CTL, i.e. some things might not be perfect for other L1
> implementations.  But we're happy to make changes to it.
> Please note: there are two alternative solutions
>  * kestrel/range have a proprietary product calledo OpenMS based on
>   OpenBTS.  It's not Open Source, though.
>  * The L1 of the upcoming sysmoBTS already has limited support for
>   behaving like a MS.  It's not a primary feature, as we think it's
>   just too expensive compared to a cheap calypso phone.  However, the
>   L1 interface is definitely better defined as L1CTL.
> I think Andreas Eversberg, Dieter Spaar or Sylvain Munaut would be good
> candidates to take care of the implementation, in case you're looking
> for somebody to take on that project.
> Regards,
>        Harald
> --
> - Harald Welte <laforge at gnumonks.org>           http://laforge.gnumonks.org/
> ============================================================================
> "Privacy in residential applications is a desirable marketing option."
>                                                  (ETSI EN 300 175-7 Ch. A6)

Alexander Chemeris.

More information about the baseband-devel mailing list