A conversation with AT&T Labs’ Clayton Lockhart

CHICAGO--

This spring, AT&T steered the voice-over-IP market into a new phase by launching its CallVantage VoIP service for the residential market. Though VoIP has matured gradually over the last few years, the service launch still represents a remarkable change in attitude and shift in strategy by the mother of all old-school telcos. Clayton Lockhart, vice president for global network planning and development at AT&T Labs, sat down with Telephony's Dan O'Shea at the Supercomm 2004 trade show Thursday morning to discuss how AT&T got to this market turning point and what to expect in the future.

Two years ago, AT&T had deemed VoIP too immature for a broad network deployment. What changed your minds?

It was a combination of things. If you look at the marketplace, you've got to hand it to companies like Vonage for getting this kick-started. With our share of the long-distance market going down, we had to ask ourselves what new kinds of product we could offer to create new types of business. Also, for more than a year, we have been working in research and development with the architectural elements, the gateways and application servers and so forth, and we have gotten to a point of having much higher-density equipment. The technology has matured to a point where we can really scale it throughout the whole network. It took more than a year for that to happen, but you're not just sticking this equipment in a garage--you're putting it in a live, real network.

Did AT&T consider using a hosted VoIP service from another carrier, similar to how it plans to return to the wireless business through the mobile virtual network operator model?

No, we wouldn't consider hosted VoIP. We think that in the future people can wholesale from us. We can take this technology out of the long-distance world and host it for other companies as a way to give us greater reach into a new business.

Will VoIP over wireless LANs play a role in your new VoIP service strategy?

We will move to that, though I don't know how quickly. There are signaling issues, and we have to make that transparent to the user. AT&T doesn't need to own the wire, so in the long run, we can see Wi-Fi and cable coming together to support voice. A lot of the stuff that's flaky and unpredictable about operating this service is right inside the home, and we have millions of homes to serve. We had to look at where we were going to put the terminal adapter, and we put it between the router and the cable modem. Some companies will do a it a different way. These issues are part of why it took us a year to get where we're at now.

Is there still room for the VoIP cost-of-provisioning structure to improve? In what other ways will VoIP continue to mature?

A: Gateways, chipsets, terminal adapters and all these other elements will get cheaper. We're going to see Moore's Law take hold here, and if the stuff gets cheaper, it will certainly help us scale the service. Also, we're driving our vendors to be more efficient in the software code running their boxes. The benefit could be improvements in the number of busy hour call attempts the equipment can handle. There's no common reference models between all these vendors, and that's something that would help. As this becomes an enterprise service, we can do more to leverage the intelligence of IP PBXs to do auto-configuration and performance monitoring. As this becomes a managed service, we don't want it to be labor intensive.