Recently, Yours Truly has been writing a great deal about VoIP Peering. Also known as Voice Peering, normally refers to the forwarding of Voice-over-IP (VoIP) calls from one ITSP (Internet Telephony Service Provider) to another directly using an IP network, thus bypassing the conventional, circuit-switched PSTN. For example, Stealth Communications (News - Alert) (www.stealth.net) founded a private Ethernet-based service called the VPF (Voice Peering Fabric) in 2003 in New York City. It serves as a meeting point for service providers and enterprises to directly connect to each other and exchange voice and telephony related services.
VoIP Peering (News - Alert) can also refer to the way enterprises “peer” VoIP phone connections directly with their trusted partners, suppliers and customers, just like the old “Extranets”
The packets of VoIP peering move over the OSI Layer 2 basis such as a private network, with carriers connecting with it and managing peering between one another, or else on Layer 5 with a central provider handling the routing and signaling for participant providers engaged in multilateral peering, the mechanics of which will be described by the upcoming IETF standard, SPEERMINT (Session PEERing for Multimedia INTerconnect).
Recently Yours Truly was chatting about VoIP peering with BlueNote Networks (News - Alert) (www.bluenotenetworks.com) deploys innovative interactive real-time communication services for global enterprise customers.
BlueNote takes VoIP Peering in Stride
BlueNote was founded in January 2005 and has gained some renown for its SessionSuite family of Business Communications Platforms that deliver a considerable portfolio of of voice, video and other interactive communication services as an integral part of an enterprise IT applications architecture, or Service Oriented Architecture (SOA).
Hal Clark, senior product manager at BlueNote, says, “When it comes to VoIP peering, I think we’ll pass through a series of stages,” says Clark. “Our SessionSuite itself natively does peering, at the SIP [Session Initiation Protocol (News - Alert)] level. We’ve built in the facilities to be able to both enable it as well as to manage the trust relationships. We do a lot of point-to-point peering. It’s certainly in place today. Attempting to peer amongst groups of people has also drawn some interest. Some consortiums are working in this area. At the moment, we don’t have a lot of active participation in those because they seem to be of the ‘walled garden’ variety of peering, and are not all that helpful to most of our customers’ interests. We’re really trying to solve a much larger problem, which can be best described as enterprises in particular participating at the same peering exchange level that you’d might expect service providers to have amongst themselves.”
“We’ve done a fair amount of work in that particular area,” says Clark. “I think that it’s still an immature area in terms of being ready to bring to market. But it holds an interesting opportunity to leverage the public fabric, if you will, for the kind of interoperability you might have expected out of the next iteration of a public infrastructure. For enterprises, it’s interesting to see them trying to get connections with customers that are not just individual connections. These you can do based on a web type of enablement, which is another thing that our SessionSuite actually does as part of the interim steps when running it. SessionSuite is able to communicate directly with customers who come to their websites and want to talk to somebody being able to establish phone calls directly over the Internet with those people on PCs that are provisioned appropriately, to do that happens to be another one of those interim steps.”
“In the long run, what most enterprises that we’ve talked to are really interested in is reaching the consumer who’s already using some VoIP service so they can have have those calls routed directly to the enterprise and not have to undergo all of the transitions up and down and through the PSTN, as it were,” says Clark.
“Many enterprises want something that’s far more universal than a ‘walled garden’ approach,” says Clark. “Which is a sort of updated version of the old ‘extranet’ concept, and they want to do it with a lot less overhead. The ability for the network to discover if an enterprise is actually able to receive a call is certainly a characteristic that’s useful. But it also requires somebody in the public arena to offer those types of translations. That’s where the current public network suppliers are still sort of in control of things, because they possess the public routing databases, captured, as it were, and the only way to get to them is through either TDM trunks or what’s analogous in the IP arena, which is just a SIP trunk. Even so, it’s fairly confined in terms of the routing architectures that the PSTN uses today.”
And Making Money, to Boot
“As far as generating revenues from peering, it all depends on where you think your money is going to come from,” says Clark. “If you think that, as a service provider, the only way you’re going to make sufficient money is by becoming a ‘toll booth’ and control who has access and how long users can use the resources, then they’re basically doing things the same way that they structured their approach to revenue collection in the past.”
“But another perspective involves thinking of bandwidth as being the issue itself,” says Clark, “and perhaps some pieces of services might be useful, but take the public translations for instance, where somebody from one particular VoIP provider is dialing a North American numbering plan, but the enterprise is advertising itself to the interchange fabric with a SIP URI [Uniform Resource Indicator], for instance,” says Clark. “Making those types of translations are the kind of thing that must happen before everybody is using a SIP URI, in which case DNS [Domain Name Service] is the only thing you really need, and that gets built in to the service costs of having bandwidth from that perspective.”
“The other thing is that the current ‘toll booth’ model also separates traffic into individual networks,” says Clark, “and it creates a whole bunch of separate networks, that also have to be interconnected, as opposed to looking at the network itself as a single more or less unified fabric. There are still exchanges between transit networks, if you will, that go on and that obviously still have to be covered by the cost of bandwidth. But I think that’s a simpler model in the long run, as opposed to trying to differentiate what is a telephone service and to have separate routing analogous to what happens in the rest of data communications or with other applications using IP across the network. It creates separate billing scenarios and all kinds of things that are expensive to implement, which is another aspect of all of this billing overhead.”
“If you’re trying to bill for minutes of usage and things like that,” says Clark, “there’s actually a cost to do it that’s eliminated if you’re just selling bandwidth, for example.”
“There’s a critical mass in usage that needs to happen with peering,” says Clark. “There are issues with trust models in terms of how you can send a call to somebody in such a way that they expect the origin of the call is a trusted source and not just somebody trying to spoof or spam the system, for instance, but I guess the process isn’t yet well enough understood for people to ‘jump off the cliff’ so to speak, and do this type of thing on any large scale. It will take the bad guys some time to sort through the issues and have a large enough body of traffic sources to actually be available that will present a sufficiently substantial market to make sense for them to participate in stealing service or spoofing or spamming the system.”
“In any case, the first part of the equation is happening: Enough people are using VoIP services so it’s starting to be of some interest to both users and hackers. But I think that the business relationships that allow the interchange of traffic are still sort of in development right now.”
“I think most of the implementations that we have today of VoIP really weren’t done expecting these types of federations to happen,” says Clark. “Or at least it certainly wasn’t built in to the architectures, particularly those that have evolved from the classic PBX platform, which is a fairly closed environment anyway. So trying to get those older architectures to behave more in the spirit of these federated-type environments is far more difficult as an add-on if you will, to implement. It doesn’t help the progress of all this in the marketplace. But VoIP Peering will become commonplace over time.”
Richard “Zippy” Grigonis is the Executive Editor, IP Communications Group at TMC (News - Alert).