I'd agree. Even if he's raised the IP Precedence of the VoIP traffic WFQ isn't the best choice. LLQ is the way to go.
There may also be problems with the vendor equipment. Nortel, as an example, has an issue with their ITG-2 product that causes it to use up more bandwidth than is necessary, but that's only in certain call transfer or conference situations, or if you're using centralized voice mail. In a nutshell, it's easy to get redundant call paths that could be dropped but they aren't so they use up bandwidth and increase the total latency unnecessarily. John ""Steven A. Ridder"" wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > LLQ would be his best option, not WFQ. If he is using it, that's probably > his issue. > > -- > > RFC 1149 Compliant. > > > > ""lamb stephen"" wrote in message > [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > > Group, > > Hoping that someone can help me out with a VoIP QoS issue that I am > > currently dealing with. I work for a service provider, and I am currently > > troubleshooting a VoIP over frame relay quality complaint. My end user has > a > > 768K host with four 256K drops dedicated solely to VoIP traffic. My > customer > > states that he experiences intermittent jitter on his calls, but they > follow > > no real pattern. We have had his vendor place test calls, and sometimes 7 > > simultaneous calls can > > go through fine while 3 simultaneous calls will experience poor call > quality > > and excessive jitter. The end user's vendor is of no real help with this > > issue stating that his configurations are fine and the trouble must be > with > > the WAN link. > > I have verified that the entire network is clean, no T1 performance > > monitor errors , no input errors on the customer's serial interfaces, and > no > > input errors to my frame switch. No apparent utilization issues, the host > > averaged 50% port utilization during a 24 hour sniff. We have also > verified > > the drops are not receiving any FECNs or BECNs. I have a copy of the > > customer's router > > configurations and his map-class statements appear to be correct as well. > > His CIR and MINCIR are set to match the frame relay PVC CIR in my network > > (which I believe means that he has configured the statements to prevent > any > > bursting, please correct me if I am wrong). > > On to my question. The only discrepancy I find with this customer's > > configuration is his queuing. On all four of his drop routers he has > > configured WFQ, on his host he has no queuing specified. Could this be the > > cause of all of his problems? Would WFQ be the most desirable method? What > I > > have read in the past led me to believe that a fragment statement in the > > map-class was the most > > desirable because it activated the dual-FIFO feature on the physical > > interface. I do not have a great deal of experience with VoIP so all I > have > > to go on right now are theories. Any direction is greatly appreciated. > > > > Thanks, > > Steve Lamb > > CCDA, CCNA Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=55610&t=55597 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
