MCI hangs its future on part of its past

MCI’s focus as it emerges from bankruptcy can be summed up in two letters: I and P. With one of the world’s largest IP core networks, the technology figures prominently in everything MCI is doing, say Joseph Cook, senior vice president of network planning and engineering, and Jack Wimmer, vice president of network architecture and advanced technology.

“Our goal is to leverage that core, to drive all services to it,” Cook says.

The company has committed publicly to migrating all voice traffic to IP by 2005. As part of that effort it has deployed 51 multi-service access gateways that initially can support voice over IP, frame and ATM. Besides making it possible to introduce new integrated IP-based offerings, these gateways let MCI start to step away from legacy gear.

The conversion “won’t happen overnight,” Cook says. “We won’t shut down hundreds of Class 3 switches one night. But we’ll never buy another Class 3 or Class 5 port.”

Another way the company is encouraging migration and leveraging the IP backbone is with its Private IP service, a Multi-protocol Label Switching (MPLS)-based IP VPN service that can support legacy frame and ATM locations. That lets companies begin to benefit from VPNs without having to make a wholesale cutover. Cook says it is MCI’s fastest-growing service.

According to Wimmer, who is responsible for looking at technologies a couple of years out, the next frontier is migrating T-1 access to packet. “All the stuff that is growing is packet-based,” Wimmer says. “We need to get away from TDM grooming and establish native packet environments from the customer into the cloud.”

While an early adopter and proponent of MPLS, Wimmer says MPLS isn’t the answer for this next step because of scaling problems. Although he wouldn’t give specifics, he says it might make more sense to use tunnels over 10G metropolitan Ethernet to reach individual customers in multi-tenant buildings.

Finding a packet answer is important because it could lower costs as MCI continues to build out its local access plant. The pressure to establish more direct customer links is enormous because 50 cents out of every dollar goes to access today, Cook says.

In fact, finding new answers to these and other issues seems to be higher priority than, say, integrating the two frame and two ATM networks that MCI runs today. “We looked and couldn’t find a real business justification to integrate these networks,” Cook says.

Bankruptcy problems aside, MCI seems to have set a meaningful technology direction that could benefit customers long term.