Need little help with a VoIP issue that doesnt seem to have a simple cause. I am primarily a route/switch guy, so please bear with me as I try to explain this issue. I have several remote branch locations that are serviced by a single T-1 for all data and voice. One of these sites is complaining of broken audio during their calls, both internal branch to branch and external calls via a PRI in another location. The problem is reported to be intermiitent and I have been able to verify some of the users complaints. Inbound calls to the branch frequently fail after a single ring and most always have poor call quality (broken and low volume). I have always understood volume issues to be associated with a bad PRI but the issue is present with some calls that do not use the PRI. I have tested the T-1 and no issues have been found. I am still waiting for the service provider to verify that jitter is within standards and have also asked them to ensure that the T-1 has been properly provisioned with QoS - all of our devices have good QoS configs. The broken audio is heard on both ends of the conversation (when the issue occurs) so that's why I currently suspect that the service provider has not configured this circuit with QoS. Any advice on what to look at next would be greatly appreciated...
Sorry -- I should have been more specific in my question...I'm just looking for some general info as to a direction that I can look in. I am not exactly certain what the call flow is. All routers involved are 2911's. I should have more details from our telecom group in a day or two...I am just trying to help them out by veryifying that network path is good. The problematic site is one of four that are configured exactly alike.
Guessing is not good in this industry, however the problematic site may have a congested WAN circuit, and no or ineffective QoS. To confirm, as you have bad quality call on the phone press ? twice and you will see statistics showing lost packets.
I couldn't agree with you more but as I stated earlier, I am just looking for some general guidance - looking at the lost packet stats on the phone is a great suggestion and just what I am looking for. I am not a VoIP guy but I am always willing to learn new technologies. I will take a look at the phones in the affected branch tomorrow morning...
You might want to verify that you dont have slips on your t1 by doing a show cont t1 x/x/x , if you do verify that you are accepting timing from the provider by the command network clock select t1 x x Check the cable lenght is also set properly on the cont t1. Are you using mgcp or h323 for the call leg to the cucm? Post your config and we can verify your setup. When you talk about voip, ensure qos is set properly if you can or setup locations within cucm to limit the amount of voice calls over the wan. Also use the g.729 codec for wan and g.711 for lan in your regions, you might have to setup an mtp when you change your codec based on your setup.
Do you have solarwinds or somthing similar to monitor the network and voice sla?
Does this help?
Sent from Cisco Technical Support iPhone App
Thanks Nick, that gives me a couple of ideas. Wouldn't T-1 slips also affect data performance? I realize UDP traffic is much more susceptible to circuit issues but the affected branch is not complaining of data issues. Also, is the call setup traffic TCP or UDP based?
Miss read yor original post a bit, you have a data T1 and voice is all IP over that, got it. If you only 1.5Mbps from a site then I would absolutly run QOS leaving that external interface on each loacation. This is most likely your issue because voice will be good and then a user starts watching you tube and voice gets choppy. The G.729 codec @ ~14kbps per call instead of ~80kbps (g.711) will also help but I think you just dont have the BW you need when users are requesting more data. Can you capture stats of the external interface of the router throughout the day to verfiy this?
Call setup in UDP or TCP can vary depending on what protocol and then what stage of the setup.
We have QoS configured for voice traffic on all router WAN interfaces - this is why I suspect that the service provider may not have configured their switch to honor our markings or to prioritize the voice traffic leaving their switch to the branch. Definitely seems QoS related given the sporadic nature but I have never seen call setup issues because of QoS config issues like this. I do see the receive traffic at the branch reaching 1.5Mbps at times, so it would seem logical that they would have issues if the service provider was not prioritizing the voice traffic destined for the branch...
If the provider is not honiring the markings , that may not be an issue unless the provider is also over subscribed. I would think that the provider has a large pipe on the other side of the router that is serviceing your T1. show your service policy when you have an active voice call and make sure your RTP voice is in the priority queue. This will ensure that packets leaving your oversubscribed external interface are prioritized which I believe is more important that the ISP honoring your QoS
Looks like my inital suspicions were correct - the service provider did not provision any COS on this MPLS circuit. As their network is very congested, this is most likely the root cause. I will update once this issue has been corrected. Also, I was finally able to capture some stats from one of the afftected users...not good:-)
If your not running the lower bandwidth codec I would configure this to use G.729 if possible to save BW. This would be good to do eiether way