Dropped Packets

[slvrsi]

I'm getting quite a few dropped packets that makes a call un-usable. The old routes a week or two ago was through cogent. This gave me ping rates in the 130ms range. I'm on DSL so i'm stuck with at least 40ms through the first ping. This seems bad but the phone line was very usable with few dropped packets through the cognet route. Recently, this changed to torontointernetxchange.net and although the ping times have improved(50-60ms), Dropped packets are common place. for a 1min call, I'm getting about 500dropped packets. Just a coincidence? I'm on TEKSAVVY for my isp.

[FONGO_kris]

Please send call trace logs to the email bellow - be sure to include your H323 Conf-ID from your Porta customer interface.

[slvrsi]

I've sent an email to admin but for the benfit of others:
I found in the evenings the dropped packets are few and far in
between, 1-2 for a 5 minute call. In the day it is mostly affected.
Its not as bad as I mentioned in the post this time around (could be
time of day) but still not as good as in the evening. What do you mean
by call trace logs?

Latest results for a day / morning call:

Time Of Call: 9:48AM
Call Length:7:29
Packet Lost: 516

Compare with a 10 minute call with 3-5 packet loss on an evening call.

TraceRt w pings:
IP MIN MAX AVG DIFFERENCE
2 2120.ae0.bdr02.tor.packetflow.ca (69.196.136.66) 42.83 57.85 47.95 1.75
3 border2.10ge7-3.teksavvy-5.tor001.pnap.net (70.42.26.113) 41.94 70.40 60.59 12.63
4 core1.te2-2-bbnet2.tor001.pnap.net (70.42.24.193) 41.62 44.70 42.99 -17.60
5 cr1.e3-2.tor001-tor003-796.tor003.internap.net (66.79.153.33) 41.66 48.15 43.85 0.86
6 gw-295ca.torontointernetxchange.net (198.32.245.78) 77.31 78.17 77.75 33.90
7 host-208-65-240-142.static.295.ca (208.65.240.142) 56.89 88.43 68.25 -9.51

[asht]

yes , I am facing same problem .
till yesterday it was good , however now its breaking lot with drop packets .
my isp speed is ok .

[redshift]

I'm noticing the same problem too. Incoming calls, sound fine. But on outgoing calls the person I'm calling sounds a bit choppy. No changes on my end. Tested with a back up voip provider, and there weren't any problems.

[FONGO_kris]

I'll be sure to notify the programmers, it looks like a global issue, a new release of the softphone is coming out soon so that will fix that problem on there, as for the many ATA/SIP Phone users I'll see whats going on there.

In the mean time check your settings especially your codecs and RTP (Real-time Transfer Protocol -- Voice) Packet sizes -- Preferably G729a and Packet size 0.020.

[slvrsi]

It looks like sometime over the weekend the routes were changed back to go through cognetco. I haven't had time to do a full test yet but it seems that the packet drops are not excessive anymore during the day. At night, it still seems great as well. Ping times have also been reduced from when it was previously on cognetco, but I have noticed lag reduction with cognetco in the last few weeks to freephoneline.ca. Seems to have gone from 130ms to 60ms (full path) for me. On par to the previous internetxchange routes previously. Thanks admins. I'll report back once I've had time to do some detail tests...

[slvrsi]

Spoke too early. still getting packet drops during the day

I am using G711u at 10ms frames. I could increase it to 20ms frams and G729a, but I think that will make packet drops more noticeable? Anyways, some data to analyze:

Call 1 Duration: 00:11:51 Call 2 Duration:
Call 1 Decode Latency: 30 ms Call 2 Decode Latency:
Call 1 Jitter: 4 ms Call 2 Jitter:
Call 1 Packets Lost: 5576 Call 2 Packets Lost:

Call 1 Duration: 00:12:33 Call 2 Duration:
Call 1 Decode Latency: 40 ms Call 2 Decode Latency:
Call 1 Jitter: 4 ms Call 2 Jitter:
Call 1 Packets Lost: 5989 Call 2 Packets Lost:
~40 sec 400 packets lost.

Call 1 Duration: 00:13:37 Call 2 Duration:
Call 1 Decode Latency: 50 ms Call 2 Decode Latency:
Call 1 Jitter: 4 ms Call 2 Jitter:
Call 1 Packets Lost: 6402
~60 sec 400 packets lost.

Call 1 Duration: 00:14:35 Call 2 Duration:
Call 1 Decode Latency: 50 ms Call 2 Decode Latency:
Call 1 Jitter: 8 ms Call 2 Jitter:
Call 1 Packets Lost: 6619 Call 2 Packets Lost:
~60 sec 200 packets lost.

Call 1 Duration: 00:15:23 Call 2 Duration:
Call 1 Decode Latency: 40 ms Call 2 Decode Latency:
Call 1 Jitter: 6 ms Call 2 Jitter:
Call 1 Packets Lost: 6745 Call 2 Packets Lost:
~60 sec 100 packets lost.

The load on the network was extremely light. This on a 6M DSL connection.

[magicray]

I have have FPL account for about a year now. Using a ATA2 for my main line, solved a couple of problems in the beginning by tweaking the ATA, but it's a month now that packets are being dropped and the calls are choppy. I doesn't matter if I use G729 or G711 codec, packets are still getting dropped just like the user above me. In a 10 minute conversion about 1000 packets are dropped which is alot. Has something changed? why isn't any one from FPL answering this thread and trying to come with a solution with the users?

[FONGO_steve]

We're not sure what's causing this for a small handful of users right now - I commented on it in another thread but forgot to do so here.
If you're experiencing this problem then please log the exact date, time and phone number of the conversations you experience it with. We will need a good handful or more of conversations logged to have ample data for our system technicians to compare. Once you have this data, PM me your email address and I will send you an email for you to respond to me with the data collected.

[volenin]

Have similar issues - dropped packets during the day, evenings works quite well. I'm experiencing the problem for most numbers, but some numbers work ok.

Please PM or post here what data I should collect and how I should collect this data - will be happy to provide it.

Vlad

[FONGO_kris]

We've disabled some scripting on the our backend which should lighten things up and improve packet loss and call quality.

[magicray]

I'm going to try the line when I get home and report back on this.

[FONGO_kris]

Awesome thanks, the call quality issue in freephoneline does still seem to be occurring but was lessened for some this morning when the Logmaster scripts were turned off and thus lowered the HDD usage in Porta (FPL backend).

However, it is still a very common issue that is ongoing with a very large percentage of our users. Steve and I have tried to duplicate the problem by configuring an ATA with the accounts of some of those affected but we were unable to replicate the issue in our office.

We'd like to get this resolved as fast as we can, but since we are unable to duplicate the problem here we can't get any further attention from our head programming office. When we have lots of logs to show, then they will start to look into it.

All info we recieve will be collected over a period of one week then aggregated into a spreadsheet document and sent away to Sofia, Bulgaria - Our head programming office, which manages our routing tables etc...

[magicray]

i tested the system late tonight and things have seem to have gone back to normal. Packet loss is still there but it's ALOT better than before.

Before: 5 Min Call - 500packet loss (711u or 729a codec)
This Evening: 5 Min Call - 150PACKET LOSS (this was the norm before if I remember correctly - G711)
5 Min Call - 50Pack LOSS (G729a)

Everything has gone back to normal and working properly thanks to all the admins.

Some of the numbers that i use to test FPL:

Google411 - 1.800.466.4411 (talk to the recognition sotware nad hear back the resilts)
Voicemail (even if not active) - *98 (go through the menus to hear her voice)