On the fast track: Carriers migrate from TDM to packet-switched networks

Although the journey to the network of the future may be long and complicated, most incumbent carriers already are beginning their trek toward packet-switched voice networks. Because migrating to a packet-switched core involves a fundamental change in network architecture, the first steps taken have been evolutionary.

Sensing the need to have more efficient networks in the future, incumbents such as Bell Atlantic and SBC Communications are testing equipment that will add packet capability to their existing switches. But for most carriers, concerns still linger. "Two hundred billion dollars worth of installed investment won't go away overnight," said Hilary Mine, executive vice president for Probe Research.

Lucent Technologies markets its 7R/E platform as a migration solution for incumbents and new carriers with networks in metro areas. Nortel Networks' Succession platform is targeted at incumbents, interexchange carriers and competitive local exchange carriers. Both platforms deliver traffic over IP or ATM and allow service providers to maintain current equipment. Both also are in trials with AT&T, MCI WorldCom and SBC. "We want to preserve what we can," said Mark Welsh, senior manager of Succession marketing for Nortel.

The 7R/E platform enables the "reuse of 80% of their existing hardware," addressing a top concern of carriers, said Diane Herr, vice president of 7R/E product management and marketing for Lucent. Service providers can use software to bring forward enhanced voice features, a source of significant revenue for carriers, she added.

In addition to the costs involved with migration to packet, the other great risk for carriers is network stability. "The trunking part of our network is very stable, very reliable. We count on it," said Joe Gansert, director of network architecture planning for Bell Atlantic.

Packet is subject to a lot of scrutiny because standards are being invented as they go along, Gansert said. Besides interoperability concerns between the time division multiplexing (TDM)-based circuit switches at the central office and the packet interface gateway device, questions also remain about maintaining and delivering quality of service (QOS). Bell Atlantic plans to provide capacity with core ATM tandems, then as the technology matures, use it to replace the embedded circuit switches, Gansert said.

"We were gated by the availability of equipment suitable for lab testing," he said. "We didn't want a prototype. We wanted something very close to what we would be using in the field."

On the other hand, it's almost impossible to set up and test the packet equipment in the lab, Gansert said. "You can't simulate the network on this level in a test environment," he said, adding that engineering and operations are some of the unknowns. Even so, Bell Atlantic expects to begin the first application of fast packet ATM in the second half of next year, knowing that it can work out operations details in order to offer relief to its network by 2001, he said.

SBC has been planning the migration for five years, said Sam Sigarto, executive director of ATM distributed network systems/broadband switching and transmission for SBC. Voice networks are at a major junction, he said. "Changes like this happen every 50 to 60 years."

SBC's Project Pronto is the first initiative to address migration and other changes, including pushing out fiber farther. SBC plans to complete tests in 2000 and begin field trials. "Hopefully, the field trials will be successful, and we can cut traffic over," Sigarto said.

SBC's migration will not be like waving a magic wand, Sigarto said. Concerns about delay, availability and network management remain, he said. The incumbent is investing in evaluations to insure that QOS will not be affected. SBC also must demonstrate functionality at the end office before going forward, he said. "The only way that makes good business sense is for the old TDM infrastructure to be taken out," Sigarto said. "But the end-office jury is still out. The business case and the technology might be providing different answers."

The question on everyone's mind, though, is when. "We should see domestic mass market by 2005," Mine said. "Capital investment capping has already occurred. Prices are going down."