#
# Stefan Knoblich <s.knoblich@axsentis.de>
# Date: 2008-07-08
#

Since Wordpress / Seamonkey deceided to eat my blogpost, a short summary in good old ascii style:

What has been changed:

Q.921, lots of it. I've changed a lot of code, so you could nearly call it a rewrite.
The new version is closely following the processes outlined in the Q.921 '97 spec [1].

Additionally some of the Q.931 had to be changed too (and i'll change more of it while
working on the other bits).

Q.921 changes in short:

	- Timers (those that are in the base spec (== T200 - T203))
	- Queueing (needs some more work, not really happy with the two-queue approach for I frames)
	- Retransmission of I frames (untested, probably broken)
	- Retransmission counters and limits
	- Can be stopped now (Q921Stop, nothing uses this -> untested)
	- Logging & dumping of packet information in debug loglevel (testing sucks without this)
	- Code cleanup, a couple of bugfixes (some already in OpenZAP trunk)

Q.931 changes:

	- Bugfixes
	- Ability to send and receive DL-* indications/requests to Q.921 (needs more work)
	- Logging (should add some nice IE debug output too)


How i've been testing:

FreeSWITCH trunk + OpenZAP trunk + bristuffed zaptel-1.2.x in PTMP TE mode connected
to a couple of "blackboxes" running in NT mode:

	AVM FRITZ!box fon 5012
	D-Link HorstBox DVA-G3342SD

	Bristuffed asterisk-1.2.x + zaptel-1.2.x, libpri-1.2.x (+ HFC-S based PCI card)


Other hardware:

	ISDN Phone for PTMP testing

	Single port HFC-S based PCI card(s)



What works:

	BRI PTMP TE mode seems to work so far, there are still some rough edges and missing pieces

	The following scenarios have been tested:

		FreeSWITCH --- isdn --->  FRITZ!box  --- sip ----> Asterisk (echo app)
		ISDN Phone --- isdn ---> (FRITZ!box) --- isdn ---> FreeSWITCH (echo app) *

		(* opposite direction will be tested tonight, i hope the Q.931 layer is smart enough to ignore
		   it-'s own SETUP message, since it will be re-broadcasted to all endpoints after we send it to the NT
		   using a P2P message)


	Everything else is untested and may be broken (PRI / BRI PTP mode)

	PTMP NT mode doesn't work!



What's next:

	Add missing pieces, refine current code.
	Improve error condition handling (e.g. invalid frame sizes / types)
	Q.931 needs a little love too
	More testing.
	NT mode...
	Even more testing.

	Get my employer to switch to FS...




[1]	http://www.itu.int/rec/T-REC-Q.921/en