We might disagree about why the change is occurring, but it does seem that discussions of "voice peering" are moving in a different direction. Early on, there might have been more emphasis on how electronic numbering or native IP interconnection could save providers money, disintermediate legacy carriers or disrupt the voice business. If recent discussions are any indicator, there now is much more emphasis on solving basic interconnection tasks in a world of IP traffic, as well as creating a platform for introducing new services.
That isn't to say all peering supporters dismiss advantages of the disintermediating sort. There is no question but that cable companies as an industry segment are anxious to avoid interconnection payments to telephone companies whenever possible, as GSM-based mobile carriers likewise are interested in avoiding transit costs where possible.
The point is that there is a new practicality about the issues. Arbinet CTO Steve Heap, for example, points out that "peering is interconnection between two or more service providers to preserve quality, lower costs and create new services." In fact, Heap points to new problems created by number porting as a mundane but important problem peering can address. "In the Belgian market, for example, 18 percent of mobile numbers are ported," Heap notes.
And since every operator has different termination rates, peering can help service providers determine what the settlement rate ought to be when a mobile call is terminated, where to send a call and make those sorts of decisions in real time. Peering can also help with the time-consuming but relatively mundane issues of negotiating termination agreements with hundreds of discrete carriers. "Not every carrier has a relationship with every mobile operator, so maybe you want to route to provider who does have a relationship," says Heap.
One measure of how the discussion is changed is that a major service provider such as Tata views peering as a simple matter of ensuring call quality under conditions of increased routing complexity. "It isn't just about free calling," says Christian Michaud, Tata SVP.
In fact, routing complexity now appears to be a problem in its own right. "There are more choices of endpoints in the IP world," says Georges Smine, Nomin um senior director. There also are codec transcoding issues that will grow as more voice traffic shifts to IP origination.
In a business increasingly using IP transmission, "what we actually deliver changes as well," says Sarina Tu, Telcordia senior director. "These days, you really don't know where to send a call, what the class and quality of service are supposed to be or what the business relationship is between the originating and terminating networks."
Then there is a growing class of "presence" information that has to be exchanged, not simply the bearer traffic and signaling.
Then there's the matter of supporting all sorts of new services and applications over discrete physical networks, says Shrihari Pandit, Stealth Communications CEO. In many cases there will be advantages to terminating traffic without touching the public switched telephone network, especially when some features simply cannot be passed between networks based on PSTN switches.
The general notion of application-aware networks also applies to voice communications. "Types of calls are more diverse" and peering fabrics can provide the intelligence to support that diversity, Smine argues.
"Who can access and control your information and preferences," Tu asks, especially when that information might be scattered among any number of discrete databases?
"Who would be the central repository for the various databases?" asks Heap. "What do you do about conflicting returns if multiple databases provide different results when a query is made?"
"Service providers want all routing information processed internally, not by a third party," says Heap. "The issue is how all that information gets there."
Nor "do we want to create a new monopoly," Michaud adds.
