Re: [Openbts-discuss] a few questions

[Harald W. <la...@gn...>]

Hi David,

On Sun, Nov 29, 2009 at 06:12:43PM -0800, David A. Burgess wrote:

> The VEA-only operation of the DCT4+ models would appear to be a  
> serious bug.  Since handsets are not required to support VEA (or at  
> least were not always required), it would seem more conservative to  
> run an EA-mode network, but the DCT4+ would not work in such a  
> network.  So, if the DCT4+ really is VEA-only, then it would seem  
> impossible, in principle, to build a network that is guaranteed to  
> support all handsets.  So far, however, I have not actually seen a  
> handset that does not support VEA.  So given what we actually *see*  
> it would seem more prudent to run a VEA network.  

I really don't believe this.  I've never seen an operator doing VEA in a
production network, so this would be catched extremely easy when they first
released/distributed or even tested those phones.

> (BTW, the OpenBSC  solution to this has been to use VEA always and just run
> the SMS  operations on the FACCH.)

This is true, but it was not specifically about working around DCT4+ phones,
but it was because we were taking shortcuts in the development. VEA is much
simpler to implement than any of the other methods.  We right now never have
to send an ASSIGNMENT comamnd and change the channel during any active
transaction.  All that's needed is the IMMEDIATE ASSIGNMENT and that's it.

However, we intend to change that at some point early next year.

> > I can't think of anything else. Since we don't know how Nokia  implemented
> > its GSM stack, I don't see how poking around randomly like this  will help.
> > Ideally, I agree with getting a capture of MOC with an operator  
> > BTS. I'll see if I can find a 3310 somewhere.

I have A-bis traces of MO calls in real networks, but that of course doesn't
include the layer 2 part.

> > 2. I've got a Nokia 3410 with an MBUS cable soldered on, but for some
> > reason it doesn't work properly with gnokii (apparently there's some
> > issues with the firmware on this handset). Has anyone had success  
> > using
> > this handset for GSM packet capture?
> 
> Can't help there.  You might try the airprobe discussion list.

AFAIK the gammu debug trace support is strictly DCT3 only.

-- 
- Harald Welte <la...@gn...>           http://laforge.gnumonks.org/
============================================================================
"Privacy in residential applications is a desirable marketing option."
                                                  (ETSI EN 300 175-7 Ch. A6)