Through CVOICE... but a question about CUBE

Took two months of straight work, both reading and labbing, but beat this one yesterday. Very detailed questions... surprising the level of detail referred to... but that's Cisco for ya.

So without giving away too much info... cause we just don't want that to happen:

If I set up a CUBE, with a H323 network on one side and a SIP network on the other, and I need to set up dial-peers on the CUBE to route calls between the networks... here, I'll try to illustrate:

CME router 1 (IP of 1.1.1.1)(Ext's 1....) --- [H323] --- CUBE Router --- [SIP] --- CME router 2 (IP address of 2.2.2.2)(Ext's 2....)

I would typically build dial-peers as:
CME Router 1
config t
dial-peer voice 20000 voip
dest-pat 2....
session-tar ipv4:2.2.2.2

CME Router 2
config t
dial-peer voice 10000 voip
dest-pat 1....
session-tar ipv4:1.1.1.1
session protocol sipv2

My questions. Is "codec transparency" required? or is it the default?
IE, if both sides are running G711, then no doubt we're good. But if Router 1 is sending 711 and Router 2 is sending 729, is transcoding automatic?
I had thought that was left to the endpoints to negotiate. If the endpoint routers negotiate, say, 729... what's the CUBE gonna do about that? I have to think 'not a damn thing'. No?

And also... as the signalling changes from 323 to SIP, I imagine that "media flow-through" is also default. I can't find anything that says either way.
What I did find was this link:
Cisco Unified Border Element and Cisco IOS Gatekeeper [Cisco Unified Border Element] - Cisco Systems
which says that "Flow around for H.323-to-H.323 mode is supported as of Cisco IOS Software Release 12.3(1), and flow around for SIP-to-SIP calls is supported as of Cisco IOS Software Release 12.4(9)T." As my analogy has us going from 323 to SIP, I have to assume that "flow-through" is mandatory, right?

Preciate any thoughts.
Mike

Find more posts tagged with

Comments

aaron0011

Mike,

Codec Transparency on CUBE is used so it can pass codec capabilities between end points...basically the end point specifies the codec when the call is setup. It's not turned on by default.

With signaling changes, media flow-through would be needed to be configured on dial peers.

CUBE config example:

voice service voip
allow connections h323-sip
allow connections sip-h323

dial-peer voice 20000 voip
dest-pat 2....
session-tar ipv4:2.2.2.2
codec transparent
media flow-through

mikearama

Gotcha. Makes sense.

A follow up question then... if transparency is not turned on, then what? The local end-point negotiates with the CUBE, which then has to negotiate the codec with the remote end-point?

And both (codec transparency AND media flow-through) would need to be configured on both dial-peers, for both directions, yeah?

aaron0011

"if transparency is not turned on, then what? The local end-point negotiates with the CUBE, which then has to negotiate the codec with the remote end-point?"

Correct.

In your example, yes, codec transparency would be configured on both dial-peers. Media flow-through would be too if the CUBE is connected directly to both CME routers. I think media flow-through is the default but in examples I've seen the command is specified in the config.

shodown

SIP is the devil

mikearama

shodown wrote: »

SIP is the devil

Hahaha... I think so too. But compared to H323... easy peasy.

shodown

SIP is easy on CME's and smaller setups. When I've been doing large corps. I've run into BGP problems, security, MPLS transport and I haven't even gotten into SIP. Then companies who insist on faxing over sip and E911 problems.

DexterPark

I think Cisco just drags their feet when it comes to SIP. I have been playing with Asterisk recently and flashing over Cisco phones can be problematic depending on the version of firmware plus Cisco makes it really hard for using their phones with anything but CUCM </Vendor Blasphemy>