QOS; Is It Necessary for VoIP to VoIP Peering?

VoIP companies need peering, whether they know it or admit it. As time goes on those without peering will lose out, as they are paying unnecessary costs that their competition is not paying.

Some have asked me if two providers are peering over public internet, how do the control QOS. I suppose I look at this somewhat differently.

Let us assume that you have two VoIP subscribers that use two different services. For the explanation we will call these two services abc and xyz. Now John@abc makes calls all day long. One of his contacts is Mary@xyz. But since john is the average VoIP user he pays a flat rate to all calls to his home country. Mary lives in Tasmania and has a DID number from John's country. John's calls to Mary are included in his plan so he does not use or perhaps even know about anything like SIPBroker that may allow him to call Mary over SIP directly.

So John's provider ABC works very hard on QOS on their network. Clearly this is an important issue. However nobody can control QOS over the Internet, so that leg of the call is a crap shoot.

Mary's provider has the same worries, and similar practices. They all worry about QOS, and here again Mary's provider XYZ can not control the QOS on the internet to Mary all the way in Tasmania (perhaps from North America) any more than ABC provider.

Now when John calls Mary, there are these two nagging QOS issues on both ends of the call as it routes through the PSTN, but the call works and they call each other every week. ABC and XYZ do not peer. One of the reasons they may cite is that they are "unable to control QOS to that peer" . But WAIT! Nether leg of the call itself when it routes through the PSTN has any QOS, so what is REALLY the issue? Even if the call needed to be CODEC translated, would that still not be better than routing that call through the PSTN? This is without mentioning that the recipients DID may already be routed over the public Internet.

I have done extensive testing on this, and I can say, I would much rather have a direct SIP to SIP call rather than a PSTN call any day. There are already QOS issues all along the path. If Providers ABC and XYZ have enough bandwidth and low latency there is no real issue. The translation to and from the PSTN is far worse than the amount of jitter that is normal on a SIP to SIP call, not to mention Mary's DID is probably delivered over the public internet from the DID originator in the USA to XYZ in Tasmania.

I would like to point out that I am talking about VoIP to VoIP peering where both endpoints are already connected to the public Internet.

Small to medium VoIP providers need to start participating in ENUM offerings. Unfortunately these existing offerings are mostly geared towards telcos and larger providers, with the single exception of e164.org which to date , has not responded to my emails about entering large numbers of DIDs into their database as they offer on their web site. I think most small to medium VoIP providers would be willing to pay a small annual fee for running and maintaining a small ENUM database. Of course there is always the fear of stale data in truly public offerings such as e164.org , so it would be best if the ENUM database was made available only to qualified parties, who control a minimum quantity of DIDs.