Life on the edge

As ILECs look to stay competitive, access products top their shopping lists

Incumbent local exchange carriers have a tough balancing act. They have to look at the customer side to determine which services the business and residential base will demand first, and prioritize purchases and service rollouts accordingly. At the same time, they have to follow the adoption of new technologies by competitive carriers to ensure they don't get leapfrogged. They do this all while dealing with an aging copper infrastructure, albeit given a new lease on life with DSL technology.

For Supercomm, the ILECs' shopping lists are extensive as usual. Whether these carriers are more urgent to buy is another question. ILECs need to upgrade DSL equipment to support advanced applications and remove deployment roadblocks; squeeze more capacity out of existing in-ground fiber through technologies such as dense wave division multiplexing (DWDM); migrate switching to handle packet-based technologies; and, in general, integrate voice and data access platforms for corporate and residential customers. They must do all of this to handle the high-bandwidth services threatening to burst the seams of the local loop - or pass it by.

DSL access demands

Measured by subscriber uptake, DSL service rollouts by ILECs have been turtle-like. But their network planners are investigating multiple ways to capitalize on the technology such as providing various next generation DSL services off common platforms, reducing the cost to provision services, saving shelf space in a central office (CO) and providing customers cost-efficient, integrated voice and data.

GTE is equipping COs with DSL technology "like gangbusters," says Marion Moskal, director of network planning for GTE. The company has 90,000 subscribers in service and more than 600 offices equipped so far. The target is to have 1000 COs ready by the end of the year.

For the future, the company is looking at common platforms offering plug-and-play line cards, or "blades," that enable different DSL flavors in the same switch fabric. GTE is in the early test stage with common card solutions from various vendors, which it would not divulge. "The integrated card is very key," Moskal says. "Integration will afford us better pricing that the vendor can pass on to us. We'll no longer have two points of failure on the network, and it will increase the number of subscribers using the existing right of way and existing cabinets."

High-speed access also is proving to be a boon for vendors. Alcatel, the vendor with the greatest number of DSL links deployed by the ILECs, in April introduced a successor to its ATM subscriber access multiplexer, the ASAM 1000, called OneStream ASAM.

This multiservice platform is ATM-based and adds support for symmetrical DSL (SDSL) services such as HDSL2, G.shdsl and ISDN DSL, in addition to multiple voice channels over DSL. It also includes support for quality-of-service (QOS) classes and switched virtual circuits.

"Carriers want the flexibility to deploy asymmetrical and symmetrical services in two slots side by side, and even in copper pairs running into the same bundle of phone lines," says Kevin Sheehan, vice president of marketing for Alcatel's DSL product line. Most of the incumbents will have production-quality ASAM 1000s in their labs by June, he says.

Despite Alcatel's overwhelming lead in RBOC implementations, other vendors aren't ceding the market to the company yet. The major ILECs now are ready to perform another evaluation round of DSL equipment because there have been many changes in application potential and QOS in four years, says Bryan Long, vice president of marketing for Copper Mountain. Carriers not only want a second source backing up market share leader Alcatel, but they also need support for new applications such as frame relay, voice over DSL, videoconferencing and IP-based virtual private networks (VPNs) for enterprises.

For example, compelling economics exists to move frame relay onto DSL because one-third the cost of frame relay is the access line, according to a report by Vertical Systems Group.

"CLECs are looking at providing frame relay over a DSL loop, and the ILECs can't afford to stand still," Long says. Most DSL concentrators, however, don't have such intelligence and cannot provide QOS differentiation to give priority to voice traffic over data traffic, Long says. Copper Mountain's Copper Edge DSL concentrators support ATM, IP and frame relay.

U S West says it is looking at Copper Edge DSL platforms, but no incumbents have purchased Copper Mountain products yet.

"There hasn't been any churn," Long says. "The ILECs move in a very deliberate manner and look at things a long time before rolling out [products]."

To solve the distance constraints of DSL, incumbents such as Bell Atlantic and SBC Communications are looking to integrate DSL on digital loop carriers (DLCs).

Bell Atlantic aims to deploy SDSL and asymmetrical DSL (ADSL) on a DLC, says Chuck Dunsey, vice president of network planning for Bell Atlantic. "It will extend the loop of current DSL limitations, although it has been a difficult technology to track."

SBC's Pronto architecture is designed to get ADSL to a large majority of its customers. SBC is installing Alcatel's LiteSpan DLCs into 20,000 widely dispersed remote terminals.

Using these terminals, SBC will have an ADSL-capable copper loop within 12,000 feet of customers in rural locations, says Ross Ireland, senior vice president of network planning and engineering for SBC. The carrier also is deploying ADSL-capable DSL access multiplexers that will serve the area within 12,000 feet of about 1500 COs.

"An ADSL line is the POTS line of the future," Ireland says. Within 12,000 feet, ADSL can provide 1.5 Mb/s speeds, or VCR-quality video. "You could actually stream video to the desktop using that technology," he says.

"Every carrier is realizing the problem of reaching 100% of their potential customers," Sheehan says. "If they're not able to, they're opening up the market to competitors."

To further reduce copper expenses, Bell Atlantic is progressing with initial deployments of HDSL2. "Each time I place it that's $375 less worth of copper I chew up," Dunsey says.

Many ILECs will look to use ADSL to deploy multiple lines of voice on a single copper pair, Long says.

A taste for fiber

While getting more out of copper pairs is vital to incumbents, they're not taking their eye off the fiber, to which many see their access networks evolving over time.

The hundreds of millions of dollars a year incumbents spend on placing copper for second lines would be better spent on emerging access technologies, according to many network planners. In addition, ILECs need to increase the capacity of the fiber optic cable that they have in the ground without digging up the streets.

With a Sonet footprint that spans more than 14,000 square miles across major cities in its territory, U S West has some shopping to do. The short-term solution to the coming capacity shortage is a platform that can aggregate traffic of various types and statistically multiplex bandwidth, says Wayne Roiger, director of architecture strategies for U S West's data division.

The other strategy uses DWDM to carve more channels out of a single fiber optic strand. That option is more expensive, but it gives the carrier growing capacity and intelligent provisioning of bandwidth, and is therefore the best long-term strategy, Roiger says. U S West is still examining its strategic options.

Back East, Bell Atlantic is taking on an ambitious project to bring fiber directly to the home. The carrier has "somewhat" frozen its fiber-to-the-curb (FTTC) deployments to take bandwidth to the customer premises, Dunsey says. But that requires placing passive optical networks sans electronics, a costly proposition.

Bell Atlantic currently is using ring DWDM on its Optical Express networks in Manhattan, N.Y.; Washington; Boston; Atlanta; and Philadelphia; which gives it the capability of not having to convert optical to electrical as the signal passes through an interim CO (Figure 1). The carrier is implementing OC-192 for the first time and recently announced a three-year agreement with Inrange Technologies to implement its OptiMux DWDM products to provide managed fiber services - including data center networking, storage area networks and transparent LANs via metro area networks.

BellSouth is looking for cost-effective ways to deploy current services and new ways to supplement those services with high-bandwidth offerings. Although Dan Spears, research director for BellSouth Science and Technology, agrees that access networks will be evolving to more fiber, BellSouth is hedging its bets, putting more bandwidth on copper and going to fiber at the same time. "We can't ignore our embedded base," Spears says. "Most of our customers are still served off copper."

But nearly 500,000 homes are passed by BellSouth's FTTC systems, and the company is slowly deploying an integrated fiber-in-the-loop architecture to offer high-speed Internet and entertainment services (Figure 2). The first integrated fiber deployments were in Atlanta and South Florida in 1999, but so far the company has signed up only about 2000 subscribers. Integrated fiber allows BellSouth to deliver ADSL with Internet access speeds up to 1.5 Mb/s, as well 70 channels of analog TV and 160 channels of digital entertainment.

BellSouth is using DWDM in its interoffice routes, and as the cost comes down, the carrier plans to use it in its metro areas. "The requirements will change - being able to move wavelengths around quickly will become more important," Spears says.

For connecting the CO to the end user, SBC looks at an ATM passive optical network as nirvana, Ireland says, because it removes the electronics from the outside plant. The technology will roll out for businesses later this year, Ireland says, and SBC will begin deploying it, but fiber to the home still isn't ready to scale.

"The technology of scale for the next decade is ADSL," Ireland says. "If we can't compete with ADSL service and fiber to the home isn't ready to serve, we'll look at another technology like VDSL - but that's not a good architecture because it scatters electronics in the outside plant."

There will be no one silver bullet for the network's edge, says Frank Plastina, president of wireless and carrier networks, for Nortel Networks. "Service providers will get more aggressive to mix-and-match types of access to get to the consumer," he says. "The challenge will be to bring on the customer as soon as possible."

Transitioning to IP

Moving ILECs' traditional time division multiplexing (TDM) networks to "fast-packet" networks is a long-term goal, but one that will crystallize in the next 12 months.

Incumbents will be kicking the tires on ATM switching products such as Nortel's Succession and Lucent Technologies' 7RE to bring them into an era where they can offer carrier-managed services in a reduced-cost network.

"With Succession, the RBOCs will have the ability to offer a richer suite of data services and combine video, voice and data capabilities," says Nortel's Plastina. "The promise of the packetized environment brings the core back together."

"We will no longer invest in a TDM Class 4 network," says Bell Atlantic's Dunsey. Bell Atlantic is evaluating ATM and IP technologies, but Dunsey doesn't see voice over IP coming to fruition for the next 12 to 18 months.

"The powerful thing about IP is that it is not an application-specific network the way the telephone network is - people can invent things," says Larry Lang, vice president of service provider marketing at Cisco Systems.

Offering IP QOS issues will be one of the critical factors to moving to a data centric backbone, BellSouth's Spears says.

Offering telco-like quality data services on an IP network requires encapsulating IP in ATM, says Ireland, who doesn't think the current technology is ready to handle IP over fiber or or IP over Sonet. "The IP protocol has to change so that it becomes more capable of handling real-time events like you would see in a telephone network," Ireland says. "Today, IP assumes that if you see an overload, you drop cells and packets on the floor and you retransmit.

One of the areas vendors are looking into is multiprotocol label switching (MPLS) - a combination of ATM and IP that provisions the actual circuits but is more dynamic. "It is one of the technologies that will help us provide a data-centric network with QOS capabilities in such a way that we can operate it cost-effectively," Spears says.

For the ILECs, the speed of deployment for the alternatives to narrowband switches will come down to the speed of the vendors' development cycles and how fast prices come down. Deploying these platforms is also a big challenge. "These are very large organizations that have back-office systems that have to be retrofitted to have a dialogue with the new digital network," says Brent Bracelin, analyst with Pacific Crest Securities.

"Constantly what we see lacking is the ability to manage these platforms and get reporting from them," U S West's Roiger says.

The good news, Spears says, is that things happens a lot faster than they used to, at least from the incumbents' point of view. "Years ago it took many years to go through a development cycle," Spears says. "It's hard to predict how quickly we'll have products available as alternatives to narrowband switches, but eventually they will all be replaced."

One assumption about the industry that most everyone can agree upon is, as BellSouth's Ron Wojcik says, "it is fundamentally a software-enabled world."

Hardware may make it all happen, but software creates the services.

At the heart of service creation is the Advanced Intelligent Network (AIN). And pumping fresh blood into the AIN is convergence - in particular the emergence of IP as a transport protocol.

As executive director of the advanced technology development center in BellSouth's science and technology group, Wojcik oversees a group responsible for AIN development. He considers Supercomm an opportunity to see "what capabilities vendors can put in the hands of service providers that have historically been more in the province of switch vendors."

Those capabilities revolve around service creation and the promise of one relatively new technology.

"If there is one buzzword that has emerged in this arena over the last year, it is softswitch," Wojcik says. "Even though AIN makes fairly significant service creation available, softswitch gives a more fundamental capability to us."

Supercomm should provide a reality check as to where all the promise stands. "There are a lot of [softswitch] players and it is not clear to me how mature or even real some of these products are," Wojcik says.

Besides softswitch, BellSouth will explore other developments sure to change the world in which it lives. As the carrier moves into the converged space, it will find itself in an SS7 and IP interworked world offering services similar to those available today through AIN, Wojcik says. "But we increasingly see a need for services like unified messaging that might reside on application servers."

That environment creates the need for integrating third-party software to applications already running in the RBOC's network.

Lightweight directory access protocol could play a significant role in this environment, Wojcik says. Citing similarities between directories and databases such as the service control point, Wojcik questions the pursuit of some of these converged services from a pure AIN perspective when there might be an approach that looks to leverage assets or capabilities that are available more in an IP or Internet space.

"We need to work through how we get the networks to cooperate, how we would bridge them together to offer services to customers that could be coming from either a voice direction or a PC direction," Wojcik says.

BellSouth also will be looking at the advancements and capabilities of extensible markup language.

Although one person's opinion - or one service provider's for that matter - does not make a trend, the trends in intelligent networking have become apparent. "We're moving into a world where separating intelligence from access and transport is becoming more feasible," Wojcik says.