Spider Registry Inc., launched earlier this month, is looking to catch every VoIP service provider in the world in its web and create the ultimate address information exchange service to enable universal VoIP peering. It’s a lofty ambition. How well this “registry of registries” succeeds will depend on a number of factors not entirely within its control. But if it does succeed, it will certainly help put the VoIP industry on a more solid footing.

Peering advocates point out that routing VoIP calls on IP fabric from end to end will eliminate a layer of costs—providers currently pay traditional phone companies to carry their calls. It should also improve call quality. Going through IP telephony gateways to the PSTN and then back through a gateway to an IP network introduces latency and jitter.

“I’m pretty confident costs will be lower with peering,” says consultant Lov Kher, a solution partner at BusinessEdge Solutions Inc., who specializes in Tier 1 carrier VoIP implementations. “I’m a little more hesitant about the QoS [quality of service]. If you don’t have a dedicated IP network, if you’re going through the Internet cloud, I’m not sure how much [quality] can be improved.”

Balancing risks and benefits
Most VoIP providers don’t question the value of peering, but they do see risks.

Some are entering into bilateral peering relationships in which they exchange customer telephone numbers and addressing information. When a customer places a call, an address server automatically looks up the number dialed to see if it belongs to one of the provider’s peering partners. If it does, the call is routed over the IP fabric through a network interconnection rather than going out to the PSTN.

Peering partners may charge each other a call completion fee, or there may be no fee. In the case of calls from non-partners, usually delivered via the PSTN, big VoIP providers do typically charge a call completion fee.

Bilateral peering is part of the answer but only part. It works where providers are not competing for the same customers. If they are direct competitors, they’re not going to want to give a rival their customer contact information.

Connecting the dots
“So let’s say you have 10, 20, even 30 of these bilateral peering relationships,” says Steve Heap, chief technology officer at Arbinet-thexchange Inc. and one the driving forces behind Spider. “The question then is, how do you get to the customers of the other 200 or 300 VoIP providers through peering?”

Another partial answer is peering-solutions companies that create registries of phone numbers belonging to many VoIP provider members who do not have or want bilateral relationships. When an end customer dials a number, it’s sent to the peering company’s registry. If the number belongs to one of its other members, it routes the call over its own or leased facilities to an interconnection point. Where there are call completion charges involved, the company provides settlement services—billing or paying members for all calls originating with other members.

The key is that the originating VoIP provider never knows whether the call is going to another VoIP provider or if it is, to which one.

Douglas Ranalli, founder and chief strategy officer at NetNumber Inc., the other prime mover behind Spider and supplier of the addressing infrastructure the registry is using, estimates there are about 20 companies providing such peering services today—mostly to smaller VoIP providers. They include relatively big players such as XConnect Global Networks Ltd., NeuStar Inc., and Stealth Communications Inc., and several smaller ones.

But the commercial peering services are just the tip of the iceberg, Ranalli says. There are also smaller communities of VoIP users—institutions, governments, enterprises—that want to peer to save money and improve quality. “It seems this is the way it’s going to evolve,” he says. “Lots and lots of local communities, cobbling together addressing infrastructures to get started [with peering].”

Enter the arachnid
This is the environment in which Arbinet and NetNumber conceived Spider. NetNumber had the technology to enable a universal addressing information exchange service—one component required to reach what Kher calls the “nirvana” of many-to-many peering. Arbinet, which was already providing PSTN interconnect services, to over 400 customers—of which about 60 percent are VoIP providers—recognized its VoIP customers’ needs for peering services.

But the two companies also recognized that nirvana would remain a distant dream so long as the peering communities created by solutions providers and non-profit groups were disconnected islands. Many if not most VoIP calls would still have to go out on the PSTN to be completed.

No existing peering company was going to abandon its customers to some universal peering aggregator. None could realistically hope to drive all competitors from the field and become such a dominant player itself. And VoIP providers would in any case balk at being locked into such a monopoly service provider.

The solution Arbinet and NetNumber came up with is a neat one. Spider is a non-profit, non-stock corporation headquartered in Wilmington, Delaware. It is controlled not by Arbinet and NetNumber but by an independent board of governors drawn from industry players around the world. It characterizes itself as “the first and only neutral, secure, industry-managed registry of service provider controlled interconnect addressing information.”

Intermediation layer
Spider does not contract directly with VoIP providers or attempt to deliver peering services. Instead it has established a competitive layer of “registrars” who aggregate addressing information from their VoIP provider members and upload it to the Spider database. Registrars, which may and probably will also be peering services providers, continue to manage the peering relationships among their customers and provide routing services.

Some Spider participants will have bilateral peering relationships, in which case they will authorize Spider to download each others’ addressing information to their own servers. More will exchange addressing information anonymously. Registrars that are also peering services providers and already have registries of their own can continue to use them for routing calls within their own communities, or they can use Spider as their registry.

A dialed number might go first to a registrar’s database. If the number belongs to another of that registrar’s customers, the company routes the call accordingly. If it doesn’t appear in the registrar’s internal registry, it goes to Spider. If it appears there, Spider manages the exchange of information and the routing between the two registrars. It will also provide financial settlement services.

Spider charges a very small fee for performing its services. Participating VoIP providers pay Spider only $0.0005 for each query that turns out to be a number in the database, nothing for numbers that are not in the database.

Arbinet, which announced a set of peering services at the same time as Spider was launched, was the first named registrar. A second, Packetexchange, a European-based wide-area IP networking services provider, has also signed up. Eli Katz, CEO of XConnect, was quoted in a Spider press release saying laudatory things about what he referred to as “federated peering”—which gives Spider hope that XConnect and possibly other big players will join.

Build it and they will come
Signing up registrars will clearly be the key to Spider’s success. Can it persuade peering services providers to participate?

The first step was building the infrastructure to make it all possible, Ranalli says. “Now that infrastructure exists, it gives people time and opportunity to evaluate and understand what it means. We may have to be a little bit patient. There’s a certain element here of ‘build it and they will come.’”

Heap says that given all the security and safeguards built in to the way Spider is set up legally and technologically, and given the autonomous role for registrars, there is no reason for peering companies not to participate. “The only barrier is a mental barrier,” he says. “If the company had set itself to being the biggest [peering company] in the world, it’s having to give up a little bit of that aspiration.”

“We’re cautiously quite optimistic,” Ranalli says. They are already in the process of talking to virtually every player in the industry about participating and expect to make announcements of new registrars and VoIP providers over the next three to six months

Kher, meanwhile, points out that Spider needs not only to attract peering companies that aggregate smaller VoIP providers’ information and traffic, but also the big players, including Tier 1 carriers that are beginning to offer VoIP services. “The big players are not part of this right now,” he says. “It sounds very good, but unless they get involved it will all be academic.”