2960 Bandwidth and VoIP

[bigdaveh]

I have a small network at a client with 3 2960s. Backbone is gig (C2960G-24TC-L) connected to the gig ports on 2 C2960-48TT-Ls. They are having quality issues with the phones and the phone vendor (small phone reseller with lmited IP phone experience) are pointing to the internal network as a possible cause. When I check the bandwidth I see about 50,000 kbts/sec on the gig switch (this is where the IP phone system is connected) and about 15,000 kbits/sec on the switches down the line from the gig. The phone system is on its own VLAN, but they told me they could not use QOS on the switches because they were not Cisco phones and not a Cisco phone system.

Main question is, is this reasonable traffic on these switches?

Thanks

 

[ISPKing]

First that whole qos thing is bs. You most certainly can run qos. Matching vlan prioritization. And yes 50 mbits is cake for 2960's what you really want to find out tho is the packets per second on both the voice vlan and data, and then the aggregate when it crosses the trunk. That will help in designing a qos scheme.

 

[unclerico]

tell a little more about the environment and how the problem comes up. is it a single site?? does the problem come up only when calling externally, does it happen when people are calling internally, or both?? is the phone system hosted by the vendor or is it in house??

two other things:
1) very rarely will you see the problem prop up because of the LAN, unless of course your topology is incorrect.
2) qos is qos. it doesn't matter if these phone's are cisco or some other vendor. it's ip traffic and you can establish qos policies that fit your environment as needed.

I hate all Uppercase... I don't want my groups to seem angry at me all the time! =)
- ColdFlame (vbscript forum)

 

[bigdaveh]

It is a single site.
Problem is more pronounced on conference calls, but have seen the probelm with only two people on the system.

It does not appear to be an issue with internal calls. It is also intermitant during the call.

System is in house (Allworx) and the line provider is Paetec. 2 T1 lines bonded with about > 512 k used.
The Allworx vendor, Paetec and myself will start trouble shooting on Friday.

 

[VinceWhirlwind]

If it never affects internal calls then that rules out the user side of the LAN.

That leaves the voice gateway and the LAN/switch it sits on to rule out.
Check out the voice gateway interfaces' switchports for errors and performance. Check out the switch they are patched to.

As stated above, the LAN should be a simple beast where you can easily check the basics and therefore unlikely to cause this - it's more likely to be caused by the voice gateway and the monkeys that have set it up.

 

[unclerico]

do the t1's only transport voice or do you have data on them also?? do you have any kind of qos on the multilink interface of the router terminating these t1's?? do you have access to this router or is it managed?? can you post the output of sh int multilink??

I hate all Uppercase... I don't want my groups to seem angry at me all the time! =)
- ColdFlame (vbscript forum)

 

[bigdaveh]

The T1s carry voice and data and are connected to the system via an Adtran TA 908E IAD installed configured and owned by Paetec. There is no Cisco equipment in the network other than the switches and they were configured by Paetec.

Next I will need a good Cisco VoIP tech in the Los Angeles area to take a look at this if Paetec fails.
Over my head, but I'm trying.

 

[vipergg]

Check any interfaces that are connected to the phone system for speed/duplex type issues . Check for errors on the ports, show int count errors , all lines should be clean if operating correctly. You check utilization type stuff with "show controllers utilization .