Re: [Openbts-discuss] Effort estimation for GPRS support

[Alexander C. <ale...@gm...>]

Hi all,

Seems something went wrong and David's e-mail hasn't went through.
Here it is:

======================================
We have previously estimated the effort to add GPRS to OpenBTS at
around 9 labor months. It is a LOT of work.  The problem is that GPRS
is not so much a new feature within the GSM as a separate service
stack, running parallel to GSM.  It will not require a new GMSK modem,
but will require new L1 FEC and completely new L2 and L3.  The radio
resource management is also much more complex.

When I say "9 labor months", I mean an experienced engineer or
developer working on this one project full-time for 9 months.  You
would know your student and your program well enough to judge if that
is a reasonable expectation.

Circuit-switched data (CSD) is not quite as exciting as GPRS, but
would useful and lot less work.  In this bearer service, two peer
devices can form a dedicated data link at up to 13 kb/s.  This data
link interfaces to outside applications on the high side of L1 and can
be connected to a public IP network with PPP as its L2.
(Functionally, it is like a dial-up modem application and is accessed
with the "AT DT" command on standard GSM serial interfaces.)  CSD
might also be a good intermediate step in the development of GPRS,
since it will require similar solutions to IP interfacing problems and
similar additions to L1, but with the radio channel management that is
already written into OpenBTS.
======================================

-- 
Regards,
Alexander Chemeris.