Read an article this morning in the Techday Telco review and almost choked on my coffee, Truenet have carried out some "VoIP Testing" and put the results in this months Telco Review as a feature article.
This follows on to the discussion I had with John Butt from Truenet a while ago and a follow up meeting in our offices, it seems that Truenet have decided to go ahead anyway so I will raise the same concerns again that I made in the original thread.
Original discussion here where I take issue with Truenet attempting to provide Stats on VoIP and the reasons why http://www.geekzone.co.nz/forums.asp?forumid=65&topicid=100014
Since its not online here is a copy of the article.
Now I take a huge amount of issue with this John, if you are reading this I wonder why you are attempting to put articles like this in the press, can you care to explain your testing methodology around these figures and in my opinion is incredibly misleading, I had some other words for it but am being semi nice here.
Can you explain how you arrived at the conclusion that VoIP is "pretty good across most providers" when all you are doing is sending some automated test RTP stream from your probes to a server located in your test location (fell free to correct me here if I'm wrong,) this just testing access !!, how many hops from each probe to get to your test server from each probe location John, pretty sure that path and QOS configurations are going to be a hell of lot different to the companies VoIP offerings... your comparing apples and penguins.
How does this test VoIP for these listed providers, your article reads like you have tested the VoIP of these companies and are now telling the industry the performance of the said companies VoIP offerings.
I interconnect with a number of these carriers over dedicated QOS circuits and use RFactor for quality measurements as it gives true end to end I can tell you some of those listed your not even in the ball park, so here we are doing real life VoIP traffic with some of these carriers and in some cases have done million and millions of minutes with these carries you have mentioned here as providing "pretty good quality" I disagree based on real information so again I ask what do you base your information on ?
I am looking at report here on a VoIP provider you list as good yet I have over a million minutes sent from them where there Rfactor barely gets above 80 in some case and below 80 is pretty common.
So I have a problem with articles like this as it seems to postion yourself as experts when it comes to the VoIP monitoring for providers, really you actually have no real way to do substantive testing of a VoIP providers network to back these statements up do you ?, I stand by the comments that all you are doing is send test packets over a access pipe to a test server, this is no way a true reflection of VoIP quality for service providers.