MGCP versus SIP versus H.323

[ybpants]

I am currently planning to swap out my 2800 series MGCP voice gateways with 2900 series routers, mainly for CUCM 10 / SRST support (I am currently running CUCM 9.1)

To future proof this deployment I want to build this out that will best support a move from PRI T1/E1 to SIP trunks to the telco and was considering configuring these gateways as SIP in anticipation of that.

From my reading, setting up a SIP gateway is pretty simple but my concern is losing the insight that CUCM has to the gateway that MGCP provides.  Does anyone have any good advice/suggestions as to how best to monitor these devices (active calls, failover, etc...) once they are in place or if there is any reason i would skip SIP and stick with MGCP or move to H.323?

Thanks

[George Thomas · Accepted solution]

If you are going to move to SIP, it would be ideal to change your connectivity from MGCP to SIP at this time. I do a fair number of these and its true you dont get channel status etc but enabling SIP options ping on the SIP trunk profile will let you know if the gateway is up/down and the failover will happen automatically. Also make sure that the "stop routing on unallocated number" service parameter is set to false so that failover happens if the SIP trunk or PRI is down. You can enable options ping on the gateway as well to make sure calls hunt to the subscribers when the primary call control is down. Of course, you will need the appropriate dial-peers.

Also make sure that you enable "Early offer - best support" (parameter in version 10.5) is set and early media is also turned on. I have had best luck with these settings when calling IVRs and audio cut through. I think Cisco is pushing people to move to SIP and in all honesty, i think that is right decision.

I would say instead of moving to H323, its best to move to SIP.

[Yorick Petey]

If you disable the "stop routing on unallocated number" parameter, and if you have multiple SIP trunks or ISDN lines, the CUCM will try to route the call to all lines, even if the number is really unexisting.

Prefer using this IOS parameter:

"no dial-peer outbound status-check pots"

http://www.cisco.com/en/US/docs/ios/12_3t/voice/command/reference/vrht_d1_ps5207_TSD_Products_Command_Reference_Chapter.html#wp1459139

The IOS will then change the cause code in case the physical interface is down.

[ybpants]

Thanks George/Yorick,

How do you handle failback?  In the past when using H.323 and I have failed over to the next device in the route list.  The ping option on the SIP trunk seems like it would handle losing a gateway but what about just losing a T1/E1?

thanks again

[George Thomas]

If you are referring to multiple PRIs on a single gateway, then you need to assign those to a trunkgroup.

[ybpants]

George,

Thanks for your help with this, it is much appreciated.

RIght now with my MGCP set up, Site A and Site B both have TI PRIs.  If the site A's T1 PRI goes down (but the gateway remains up) calls will be routed via Site B's TI PRI which is controlled by the Route Group configuration.  When the PRI is restored for Site A call resume over that.

How can I manage the failover/failback process with SIP?

Thanks again.

[George Thomas]

In the past I have used the "Stop routing on unallocated number" service parameter in conjunction with Route List. It looks like the command that Yorrick mentioned might also work but I havent tried it first hand.

[ybpants]

Thanks both.  I guess the only way to find out for sure is by giving it a try.