Walk the last mile

Data and the Internet become top priorities for IXCs as they attempt to go beyond traditional transport service offerings

Seeing the future in the present is a gamble service providers are forced to take each day. While some of those bets may pay off, others may fall into the pile of expenses. In the past, service providers could depend on networks and equipment to last for decades. But today, complacency is a killer, with needs changing and escalating faster than equipment can be developed and deployed.

Now the race is on for carriers to augment their offerings to encapsulate high-bandwidth delivery and a myriad of services in addition to the traditional transport services. Interexchange carriers (IXCs), along with most other carriers, are placing sure bets on data and the Internet as they try to look into the future demand for bandwidth. Most would argue that the odds are good, but because so many service providers are essentially betting on the same horse, the payoff will not necessarily be as high.

The trick for IXCs will be to satisfy customers' current demands while anticipating future needs. How those providers accomplish both depends on the size, scope and scalability of their networks. Some may have roots reaching deep into their service areas or target markets, others may not be so lucky. Naturally, that discrepancy has created a variety of future growth plans.

Despite the fact that equipment would seem to be the greatest changing element for service providers as they attempt to meet future needs, for most, equipment is just one piece of many elements. As customers get smarter, so have service providers. The realization that their competitors can have a network identical to their own has led IXCs to search harder for new markets.

Although the tactics may be varied, the underlying goals are still similar: Meet bandwidth demand and deliver more particular bandwidth services and back it up with reliability and unparalleled customer service. Along with those goals comes the recurring theme of data and the Internet and how those factors are driving the need for change among IXCs.

The track to run on

As most IXCs augment or alter their existing networks, they are changing the way they reach their customers. For most IXCs, that means adding DSL connectivity, metropolitan networks and more extensive or powerful long-haul networks.

Some IXCs have assets in place in addition to their long-haul routes, but to meet the escalating demands, building onto existing footprints is critical. Providers such as Sprint, Alltel and Qwest Communications are tacking DSL onto their service offerings with vamped-up efforts to deploy the technology.

"The [IXCs] are all looking at broadband, and virtually all of them have plans for DSL," says Jay Fausch, senior director of marketing and business development for Alcatel.

For its local builds, Sprint is putting DSL at the forefront of its high-speed data plans. The provider is turning up consumer and business Internet services in several areas. Sprint has gone with Alcatel DSL access multiplexers (DSLAMs) and Lucent Technologies' Stinger DSLAMs, says Greg Crosby, vice president of high-speed data product and project management for Sprint.

To coincide with the DSL rollouts, Sprint is also building enhanced data points of presence (POPs) as a local caching strategy to improve the delivery of data, Crosby says. By caching locally, Sprint theoretically will be able to put more content on the edges of its network.

Sprint also has a partnership and equity interest in EarthLink, which enables the companies to package DSL service. For the added dimension of content, Sprint is partnering with companies such as MSNBC, Fox, ESPN, Disney and StreamSearch.com. "We don't want to just be about pipes," Crosby says.

Sprint's suitor, MCI WorldCom, is banking on the expanding e-commerce and e-business needs as it moves forward with its new "generation d" initiative.

As broadband prolifereates, hosting centers and throughput must be rock solid, says Fred Briggs, chief technology officer for MCI WorldCom. "We plan to build 120 [hosting centers] before year-end," Briggs says.

To better use its assets, the company plans to convert numerous call centers to Web centers.

Pushing more out to the edges is a definite priority for Alltel as well, but the devices at the edge need to beef up to support that migration.

As Broadwing Communications tries to add more metro area connectivity, it is considering technologies that help connect the customers to its POPs as well as metro dense wave division multiplexing (DWDM) technologies to help allocate bandwidth, says Chris Rothlis, vice president of engineering for Broadwing.

In areas where physical connections aren't possible, Broadwing is looking at wireless optical technologies such as those from TeraBeam Networks and Lucent.

Working with the old

Despite the fact that newer, sexier networks and equipment may be more efficient, have greater flexibility and scalability and generally attract more attention, IXCs still are seeking ways to maximize their current investments. Many are looking to find ways to further leverage their current infrastructure without having to completely displace it.

Although providers are aware that changes and augmentations must be made, using engineering and equipment that enables those alterations to be graceful and economically sound is paramount. Augmentations may include improvements to the existing networks to better suit the changing needs. A top concern is making networks even more resilient than they are today.

AT&T uses ATM and IP because of the functions each performs. But reliability and IP are still not synonymous. Vendors need to add more reliability to IP, says Hossein Eslambolchi, vice president of data and Internet network services for AT&T, though "the world is moving to IP."

Alltel is also giving a lot of attention to developing the flexibility of its current network and architecture. "We are going through the engineering process of building resiliency of the network, where it will automatically rearrange if one piece drops off," says Chuck Morse, director of data engineering for Alltel. The provider is doing that with access and with core transport to build levels of resiliency, he says.

Where ATM resiliency is concerned, Alltel models the core network for failure to test how it will rearrange itself, Morse says. The IP layer, he says, "has built-in ability to rearrange, but we need to have additional facilities so each side has more than one route out."

That way, if a failure occurs in one site, the rearrangement can use either ATM or IP, which most core ATM and router equipment has the built-in functionality to do. Those capabilities may not have been used in the past because of economic reasons, Morse says. "How many ways can you economically go out of a building?" he asks. "But you can reach a point where it is economical to do that."

Now the company is starting to use what was built in by rearranging the network and building additional facilities. Alltel also is engineering to move the resiliency out further into its network. "To do that, we need the alternate paths," says Morse.

The best way to accomplish that may be by leveraging existing technologies such as the time division multiplexing (TDM) networks that provide resiliency for other networks. Alltel is working on a network design that uses the stability of its existing long-distance network to provide added dial service backup and stability, in addition to packetizing traffic.

Is convergence in the cards?

To achieve resiliency and dependability, IXCs have turned to Sonet-based networks because of their survivability and reliability. But as the needs and service demands of the IXCs progress in Internet time, those traditional-style networks must be revamped.

"There seems to be an architectural or technical shift occurring," says Ken Wirth, vice president of global emerging markets for Lucent's optical networking group. In most cases, that shift is going from the use of Sonet rings to the implementation of optical mesh networks.

"We are moving away from Sonet rings fundamentally," says AT&T's Eslambolchi. "We need to head to an optical core mesh [network]. That way we will be able to deliver services much more rapidly than before."

To bring optical mesh networks to reality, vendors such as Cisco Systems, with its wavelength router via an acquisition of Monterey Networks last year, and Lucent , with its LambdaRouter, are making strong plays for the use of mesh networks.

Lucent's LambdaRouter provides protection switching for the traffic flowing over mesh networks, Wirth says. "Rings are nice, but they are not necessarily required."

The idea is to overlay services instead of using basic transport systems. Therefore, different protection paths can be set up. The router then theoretically takes care of the protection, which frees up more capacity for services.

"Networks need a reduction in complexity," says Joe Bass, vice president and general manager of Cisco's core optical transport unit and former CEO of Monterey Networks. "And when you reduce complexity, you can bring the network costs down."

Nortel Networks also is working to give its clients a road map for mesh architectures. Through better planning for the future, Nortel hopes to make that migration to mesh networks more fluid and long-lasting.

"The evolution of the network is mesh," says Vivian Hudson, vice president of high-capacity optical networks for Nortel. "We are looking at how we can make the network forecast-tolerant because no one can predict where traffic will come from.'

The OPTera Connect cross-connects different pieces of traffic throughout the network, Hudson said. And to help shed more electrical elements in the mesh network, Nortel plans to leverage its acquisition of Xros.

"Xros eliminates the [optical-electrical-optical] conversion while being able to do wavelength switching," Hudson says.

MCI WorldCom is considering using ring and mesh network configurations. The provider is deploying an IP-over-optical infrastructure, but once the software becomes available, it will evolve the architecture to encapsulate optical cross-connects and mesh architecture.

In addition to finding ways to better customize networks for growth and efficient management, IXCs are searching for methods to converge voice and data offerings to gain even more efficiencies and streamline costs even further.

"[AT&T's Integrated Network Connection] customers can get a single physical path, and they really don't need two separate paths," Eslambolchi says. But what is suitable for enterprises differs from what carriers can deploy. Although converged technologies may work for enterprises today, the technology has not reached the carrier-class level.

"As time moves on, the telephone network will begin to evolve, and [companies] like Nortel and Lucent are evolving that [network] to where it can take the data directly in rather than have to reconvert it back to TDM," Alltel's Morse says.

That is still several years away from going end-to-end. "There are a number of technologies out there that work extremely well for enterprises," Morse says. "The technology just isn't there yet to meet my standards for customer service, which are critical."

But for providers adding DSL to their offerings, voice over DSL creates an inroad to more converged services, although that technology is not currently being deployed by most of the larger vendors because of its maturity level (see sidebar on page 88).

Stacking the odds

As service providers pressure vendors to quickly create mature products to help them meet their service goals, many are realizing that success does not depend solely on the equipment. Winning in the long run means placing winning bets on several different elements.

Time to market is just as critical as the prowess of a provider's network, but considering that service providers usually have access to equipment at the same time, the providers' own implementation becomes a heightened variable. Their streamlined processes combined with engineering, foresight and leading-edge technology will ensure that they secure the best odds.

"Technology is available to all providers, and we can't sustain an advantage based on that," says Broadwing's Rothlis. "You have to be the best at timely and effective implementation."

The concept of voice over DSL is enticing to most service providers with DSL in their portfolios. Incumbents may be grappling with how to prevent the technology from cannibalizing their current offerings, but as a whole, interexchange carriers are performing lab or field trials of voice-over-DSL equipment.

Sprint is working with vendors to deploy the technology by year-end, says Greg Crosby, vice president of high-speed data product and project management for Sprint. However, the standardization of the technology needs to escalate to make it more approachable for deployment into networks, he adds.

"These companies will take the time to make sure solutions don't diminish the brands they have worked to create," says Stephen Gleave, vice president of marketing for Jetstream Communications, a developer of voice-over-DSL solutions (see related story on page 58).

Although the products still need tweaking to support greater scalability and reliability suitable for larger networks, the values are hard to ignore.

"Voice over DSL gives [IXCs] a way to make an attractive business model," says Jennifer Stagnaro, vice president of marketing for CopperCom. "What they are seeing is a way to come up with a creative new service."

And as with most technologies, service providers don't necessarily want to be tied too tightly to one particular technology. Broadwing Communications is looking at several different voice-over-IP vendors as it seeks to converge services.

"We are testing voice over IP, voice over DSL, voice over ATM or whatever flavor," says Chris Rothlis, vice president of engineering for Broadwing. "We really try to stay agnostic to support all [of the technologies] in the end."