Intermittent dropped calls after 11 minutes 11 seconds

[netpcs]

Hi,

A client of mine has been having ongoing intermittent call drop outs for a few years now and whilst a little random as to when they happen, there are 2 consistencies in the dropped calls. She can and has had calls that last longer to these types of numbers though - such is the intermittent nature of the problem. I had posted this problem some years ago when it first happened but it was ignored! Now she has emailed me again with another dropped call. The last one prior to this that I’m aware of was on the 13/1/14 to 1300 150 300 (Macquarie bank) for 11 minutes and 11 seconds.

Whilst this is annoying the worst part is that as these calls go into a call centre, it takes 10 minutes to wade through the IVR system and to be on hold before speaking to a earth creature, then once the call gets dropped you have to start the call all over again – this is when my client gets very annoyed and I get the resultant email!

The nature of the dropped call is as follows:

1. The call is dropped after 11 minutes 10 or 11 seconds – exactly.
2. The called number is to a 13 xx xx, 1300 xxx xxx or 1800 xxx xxx, typically these numbers will belong to the call centre of an Australian business with the call being made from Melbourne, although the geographical location of the call centre may not be in Australia, although most banks etc. would be. The number in this instance is to Westpac bank

The call statistics page in her Quadro 2x is shown as getting a “bye” message to end the call from the called party – which is when the call is dropped.

If there are any more clues in the SIP UA log attached I would be very interested to know as I would a fix for the problem too. Her network setup is very basic and orthodox, with just the WAN port being used on the 2x going through the Cisco RV042 router and a bridged Netcomm NB6 modem, with a business nDSL and VOIP service provided by iiNet.net.au

Attached is the SIP UA log a screen dump image of the current call stats would be attached but for the upload size limit!

If anyone can shed some light on this I would be very grateful.

Thanks, Francois.

PS. The first person to suggest a firmware upgrade will have their goolies cut off!

[KSComs]

The IP address in the SIP UA doesnt appear to be from IINET so I am unsure as to the validity of the call path unless you are hiding it as well as the customers number for security reasons.

May I suggest you follow up this problem and flag it with Alloy the can be reached on 03 8562-9000.

K.

[netpcs]

Hi Kevin,

Thanks for the reply, if the IP address you’re referring to is 203.216.42.123, then yes, it’s not my client’s WAN IP but some Japanese address! All other personally identifiable information has also been obfuscated; the other 2 iiNet addresses are legitimate.

Would it help if I provided the real IP address?

I’ve been made aware that this has happened a few more times since January but I was not informed, so it is a bit more frequent than just twice a year.

Francois

[KSComs]

Hi Francois,

If it is only a few calls a year then that is totally acceptable - because it will more than likely be externally from your customer and out of your control to fix this problem. Even Telstra can not definitively say they have an uptime of more than 98.9% - there is always someone working in the street, placing a test telephone across the wrong pair of wires, for SIP that is enough to break it and drop calls.

If you were to amortise it - 20 calls a day, 5 days a week, for 48 weeks per year is 4,800 calls. if we were to make it 1 % failures as per Telstra's accepted uptime that would equate to 48 calls that would fail per year.

How many of them have actually failed for your customer ? if it is under the 48 mark then that is acceptable isnt it?

Just another slant on things for you to ponder.

Kev

[netpcs]

Hi Kevin,

Thanks for the insight, whilst I can go along with the theoretical uptime for call services, especially when you take into account maintenance work performed at an exchange or in a pit, I think my client won’t readily accept the explanation, although her issue is not the result of that type of human intervention – I’ll just use that as the excuse anyway.

Somewhat interesting is that I too run a 2x over iiNet nDSL and whilst I make a point of never ringing 13 / 1300 numbers – it really is a no-brainer! I will ring the proper number so in the case of Westpac I call 02 9293 9272 – it still goes through the same IVR system – it just doesn’t cost me for the call and I cannot remember ever having a dropped call to any number local or national.

Quite some time ago I setup a call route on her 2x so that if she prefixes a number with 2 it will push the call out over her Optus PSTN line and when done so, she has never reported to me that she has had the same problem.

Apart from your welcomed input I don’t think this issue is drawing any further interest so I won’t bother pursuing this any further. Once again thanks for your input.

Just one last thing as I’ve got your attention, is there an Epygi document that explains what all the various entries in the SIP UA log mean, without some reference to interpreting the log it means nothing to me.

Francois

[KSComs]

How about the following link does that help?

https://telconotes.wordpress.com/2013/03/13/sip-transactions-vs-dialogs/

[netpcs]

Hi Kevin,

It most certainly does and the site as a whole is a good source of information too and something I can learn from, thanks.

Francois