Latency issues lately

[montrealguyhere]

We are starting to noticing latency issues on our Teleworker phones.
Using these phones on MINET not sip yet.
we have 25 in the field.
it comes in to us on a T1 used by no one but the phones, Usage never exceeded 45 % of T1.


calls seem to start out ok but after 5 min we start to notice serious latency, this is really noticeable if callers are calling into our system from a POTS, we do not see this from inter office calls, from ext to ext, this is always clear as a bell.
here is some info from the server

Physical Memory 58% 194.95 MB 272.69 MB 467.64 MB
Disk Swap 0% 258.85 MB 0.00 KB 258.85 MB

I dont like the err or drops on the eth0 interface, its lik eit is acting in a half duplex mode with crashes.

Device Received Sent Err/Drop
lo 202.92 MB 202.92 MB 0/0
eth0 3.30 GB 1.31 GB 483/487
eth1 1.56 GB 1.56 GB 0/0

Mount Type Partition Percent Capacity Free Used Size
/ ext3 /dev/hda3 2% 68.23 GB 1.11 GB 73.04 GB
/boot ext3 /dev/hda1 15% 79.06 MB 14.57 MB 98.72 MB
/dev/shm tmpfs none 0% 233.82 MB 0.00 KB 233.82 MB
Totals : 2% 68.53 GB 1.12 GB 73.37 GB

John In montreal



John - Up in Montreal

It's about who??????
Then make it about us....

 

[TheMitelGuy]

What are your thoughts on replacing the PC w/ the NIC cards?

 

[slapin]

If you see the issue on per-call basis it could be an internal TW server problem.

1. TW server can collect quality stats during the call. Look for it in the tug.log or tugcurrent files in /var/log/tug. Look at Jitter stats. The TW server will try to adjust de-jitter buffer. If it's getting more then 250 ms an average person will start noticing delays.

2. Connection on the romote end. If there is no QoS then data traffic especially ftp can affect voice stream.

 

[montrealguyhere]

I will look into the logs, QOS is not used since it is over the internet and QOS is not supported unles u use MPLS and that is a little to expensive for the end user at this time.
I will check out the jitter stats and maybey post the results of what i see...


John

John - Up in Montreal

It's about who??????
Then make it about us....

 

[montrealguyhere]

Here you are , what are we looking for in turms of what is normal for jitterQ???


2007-06-22 08:19:33.204067500 IN CALL with: 127.0.0.1:20371 Codec: 1 FrmSz(IS/SS): 20/20
2007-06-22 08:19:33.204068500 RxStarted:1 TxStarted:1
2007-06-22 08:19:33.204069500 Dur(Rx/Tx): 102/102
2007-06-22 08:19:33.204070500 Back to Back Teleworker call
2007-06-22 08:19:33.204074500 RxFrames: 5089 TxFrames: 5072
2007-06-22 08:19:33.204075500 From Set:: OOS: 0(0%) Dropped: 0(0%)
2007-06-22 08:19:33.204082500 From IcpSide:: OOS: 1(0.0197%) Dropped: 0(0%)
2007-06-22 08:19:33.204084500 JitterQ: Len: 13 LWM: 13 HWM: 14 AvgQ: 14.2 Under: 21 Over: 0 Shrinks: 15
2007-06-22 08:19:33.204085500 Set->SS0(65.206.22.19:61471->20378)
2007-06-22 08:19:33.204086500 SS0->Set(20378->65.206.22.19:61471)
2007-06-22 08:19:33.204087500 Icp->IS0(127.0.0.1:20371->20379)
2007-06-22 08:19:33.204088500 IS0->Icp(20379->127.0.0.1:20371)


John - Up in Montreal

It's about who??????
Then make it about us....

 

[slapin]

When you took this sample, was the problem presented?
I'm not sure whar does it mean "Back to Back Teleworker call" Let me research. JitterQ and drop rate is pretty good.

 

[montrealguyhere]

this is a sample of a call one of our users had, he is always complaining about lagg


back to back , could that be ip phone to ip phone on same system.



John - Up in Montreal

It's about who??????
Then make it about us....

 

[kwbMitel]

Just a thought, because the issue seems to be active when crossing the E2T barrier. Are you using compression?

*******************************************************
If we don't take care of the customer, Maybe they'll stop bugging us.

 

[montrealguyhere]

Not sure, where do i look ??

John - Up in Montreal

It's about who??????
Then make it about us....