Re: [Ekiga-list] IAX2 support (was (Re: [Ekiga-devel-list] Ekiga 3.00 available for WIN32 *only*)


On Fri, 4 Apr 2008, Rene Bartsch wrote:

I think all he wanted to say was: Nobody who does not *want* to use SIP
but would rather prefer IAX2 (for example, because he's behind a NAT
firewall and has problems to make STUN work properly) will not be forced
into SIP anymore.

Yes, I just want to be able to connect to anyone with SIP.

Mmh, I never tried whether IAX clients can work work peer-to-peer (with
some DNS tricks it's possible with any SIP client).

They can work peer to peer. The iax2 protocol does allow for p2p.
Difficulty is that iax2 (as a voip protocol) is part PBX protocol, and there are many PBX things bound into iax2 that (my view) don't belong there.

If IAX2 will marginalize SIP (as SIP did with H.323) is to be seen.

SIP never maginalised h.323 because of better technology. Sip marginalised h.323 cause of better publicity. However, lets us avoid timewasting h.323 vs sip discussions. I will simply quote Craig Southeren
 "I have implemented both protocols. They are both overly complex.
  I hate both of them".

There is some iax2 code in the opal trunk true, and in branch v3_2. I have
just integrated it better into opal's call phase model.

So, Iax2 can be made to work in Ekiga. Anyone interested in fixing the rough edges, and making it work?

- VSPs:             Less horsepowers on the VoIP servers
- Hardware vendors: Much cheaper hardware
- Users:            Much more reliable, only one NAT-port

By the way, about three years ago I did some testing with IAX and SIP.
On a Server (Intel P4, 1024 MB RAM, 100 MBit/s inernet connection) I had
set up an Asterisk installation with facsimile support. The server was
connected to a PSTN gateway of a VSP with SIP and IAX. Then I tried to
send 10 facsimilies from a facsimile machine (FAX (analogue) -> ISDN ->
PSTN -> PSTN Gateway -> Internet -> Asterisk server). The latencies
between PSTN gateway and Asterisk server were about 5 -10 msecs.

With IAX eight facsimilies were successful, SIP failed completely!
This is an implementation issue of the SIP/fax/iax. The packets still take the same time to go over the network. The latency should be the same.

Sip has been correctly described as the largest denial of service attack on the IETF working process.

Derek Smithies Ph.D.
IndraNet Technologies Ltd.
Email: derek indranet co nz
ph +64 3 365 6485

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]