Rob Huffman Sun, 01/25/2009 - 08:21

Hi Youssef,


Here are some related docs. I have included some MGCP references as well;


Cisco IOS H.323 Gateway Configuration for Use with Cisco CallManager


http://www.cisco.com/en/US/products/sw/voicesw/ps556/products_tech_note09186a0080094636.shtml


SIP Versus H.323


http://www.cisco.com/en/US/tech/tk652/tk701/technologies_configuration_guide_chapter09186a00800eadee.html#93799


MGCP and H.323 Voice Gateway Protocol Comparison


http://www.cisco.com/en/US/tech/tk1077/technologies_tech_note09186a00806fedbe.shtml



Voice/Data Integration Technologies


Comparison and Contrast of the Various VoIP Signaling Alternatives

The various signaling alternatives each offer advantages and disadvantages for system designers. A few highlights are presented here.


First, regarding MGCP and H.323, the scope of the protocols is different. MGCP is a simple device-control protocol, while H.323 is a full-featured multimedia conferencing protocol. H.323 is currently approved up to version 3, while MGCP has not been and may never be fully ratified; it is merely a de facto standard adopted by some manufacturers. As such, MGCP interoperability has been demonstrated, but not industry-wide. Likewise, the complexity of H.323 has inhibited interoperability as well.


MGCP can set up a call in as few as two round-trips, while H.323 typically requires seven or eight round-trips. (Note: H.323v2 provides for a fast start process to set up some calls in only two round-trips, but this is not widely implemented.) Call control is little more than device control for MGCP, while H.323 derives call flow from Q.931 ISDN signaling as a media control protocol. This control information is transmitted over UDP for MGCP, and over TCP for H.323.


SIP and H.323 are more direct competitors. They are both peer-to-peer, full-featured multimedia protocols. SIP is an IETF RFC, while H.323v3 has been approved by the ITU. Interoperability of both protocols has been demonstrated. SIP is more efficient than H.323, allowing some call setups in as little as a single round-trip. In addition, SIP uses existing Internet-type protocols, while H.323 continues to evolve new elements to fit into the Q.931 ISDN model.


Comparison of SIP to MGCP is similar to the comparison of H.323 to MGCP, in that SIP (like H.323) is a media-control protocol and MGCP is a device-control protocol. The same differences emerge as before between client/server and peer-to-peer. The fundamental difference is that peer-to-peer protocols such as H.323 and SIP tend to scale more gracefully, but client/server protocols such as MGCP are easier to design and maintain.



http://www.cisco.com/en/US/docs/internetworking/technology/handbook/Voice_Data.html#wp1022130


Hope this helps!

Rob

wilson_1234_2 Sun, 01/25/2009 - 18:34

How does the 711 and 729 codec fit into this?


Are SIP calls 711 or 729 or neither?

Nicholas Matthews Sun, 01/25/2009 - 21:07

The only place where codecs come into play between the two is this:


The H323 specification mandates that every device that supports H323 to support G711.


SIP doesn't have that mandate.


Both protocols are able to support the same codecs, at least from what I have seen. H323 is more popular for video, so there may be some additional video support in H323.


In short:

H323 is more mature and has more interoperability support.

SIP is newer, and there is still a lot of 'fighting' between different vendors because the specification is much more open. A lot issues due to one vendor being slightly in the wrong due to some wording, etc. SIP has less rules, and is more flexible. A lot of applications use SIP, and it's much easier to read/troubleshoot for beginners.



hth,

nick

wilson_1234_2 Mon, 01/26/2009 - 17:58

So, Verizon could send me both g711 and 728 via SIP?


Id so, haow can I tell if an inbound call is 711 or 729 on the gateway that has the sip trunk termininated on it?

Nicholas Matthews Mon, 01/26/2009 - 19:11

You would look at the SDP of SIP.


This is a sample SDP:

v=0

o=Cisco-SIPUA 20472 49800 IN IP4 1.1.1.1

s=SIP Call

c=IN IP4 1.1.1.1

t=0 0

m=audio 16384 RTP/AVP 0 8 18 101

a=rtpmap:0 PCMU/8000

a=rtpmap:8 PCMA/8000

a=rtpmap:18 G729/8000

a=fmtp:18 annexb=no

a=rtpmap:101 telephone-event/8000

a=fmtp:101 0-15

a=sendrecv



Here in this line:

m=audio 16384 RTP/AVP 0 8 18 101


We see the codecs advertised. A few lines below, some of them are defined. You can look up the specs on these. Generally 0 is G.711ulaw, 8 is G.711alaw, and 18 is G.729. 101 in this example is RTP-NTE, or RFC 2833.



You would want to look at the SDP in the INVITE and ACK of an incoming call, or the incoming 200 OK for an outgoing call.



hth,

nick

wilson_1234_2 Tue, 01/27/2009 - 16:13

what commands would show me this?


is it a debug?


We have a SIP trunk that Verizon has designed configured. It is supposed to be three logical trunks with the SIP trunk to our gateway:


one for our customers toll free @ g729

one for our customers local @ g729

one for all other calls @ g711


the 729 trunks were going to a CAT phone banking system


the 711 were going to an IPCC system.


My understanding is the 729 in uncompressed and the 711 is compressed.


Also, the phone banking is going to IPCC as well.


Does this all sound feasable?

Nicholas Matthews Tue, 01/27/2009 - 16:53

A little backwards - G711 is uncompress and G729 is compressed.


The debug you would use to see the SDP is 'debug ccsip messages'


This sounds feasible, yes.


hth,

nick

wilson_1234_2 Tue, 01/27/2009 - 18:14

Thanks,


One last question:


IPCC can take inbound calls ing g711 or g729?


or is only g711?

Actions

This Discussion