Blinding them with science

A lot of factors played into the economic boom of the late '90s. The period witnessed generous capital markets, legislation that opened up competition, and of course, significant technological innovation.

The Internet — wired and wireless — presented seemingly endless possibilities for companies with new, innovative ideas. Most of these companies came in the form of vendors, those offering hardware and software-based solutions for the network and for services.

Though the boom has ended, technological innovation marches on. Vendors still play a role in the evolution of services and the public network, but regardless of common perception — Cisco Systems gained such widespread acclaim that it was mentioned in Dilbert comics — service providers have and will continue to play a big role in this evolution.

One of the more noteworthy service provider R&D efforts belongs to Sprint and its Advanced Technology Lab (ATL). The Burlington, Calif.-based facility employs between 70 and 100 scientists at any given time. The lab, led by director Frank DeNap, was created to explore Sprint's network and is best known for its forward-looking innovations, though these applications compose only about 20% of the lab's work.

“That 20% usually attracts a lot of attention,” says DeNap, “but it also helps us prepare for the future and start saying, ‘We need to start thinking differently now. We need to start refocusing our efforts to prepare for a new future,’ which is going to emerge in maybe five years.”

The ATL is divided into five main operating divisions: switching, which handles network solutions for international customers; international, which deals with submarine fiber systems; transmission, the lab's network development group; service architecture, which deals with network development and broadband applications; and IP and Interworking, which according to the company's Web site, has the job of “anticipating information society evolutions.”

Another service provider that runs a noteworthy R&D lab is Verizon Communications. Verizon Laboratories and its 350 to 400 researchers are headed by Mark Wegleitner, senior vice president of technology and chief technology officer for Verizon.

The lab, which handles R&D for Verizon's wireline network, is divided into four groups: the advanced system laboratory, which develops architectures and prototypes of new technologies and services; network infrastructure, which handles the architecture, evolution and security of Verizon's network; operational systems, which handles the company's operations support systems; and services research, whose mission is to create new revenue generating services of all types.

In contrast to Verizon Labs and ATL, AT&T Labs is housed in four main facilities, employs about 2150 researchers and handles network and end-user applications for AT&T's customer-facing business. The work within these units is organized into technology areas such as data mining, communication, access, network and software.

The main work of AT&T Labs is divided into two divisions: research (headed by Larry Rabiner, vice president of research) and development (headed by Bill Leighton, development vice president). And unlike either Sprint or Verizon, AT&T Labs views research and development as two separate functions and is organized to reflect that. The task of research is to shine the headlights months or years down the road and work with a business unit on the creation of a new technology. Once the business unit sees the technology and the marketplace as being acceptable, the project is handed to development for mass application, solving problems of scalability and operational standards.

According to Rabiner, the lab as it is today was created upon the 1996 split-off of Lucent Technologies and NCR. “At the time [AT&T management] realized that the entire structure of telecom was changing radically. AT&T basically decided in order to be a leader in this field we would have to be leading the change, leading the revolution that was changing the network, the access, the operations, the devices and the service.”

Revolution leaders they may be, but R&D labs generally remain hidden in the shadows of their respective companies, usually garnering no more attention than many other divisions. In a way, though, this is fitting, for the roles that each of these labs play within their companies are similar to any other unit's role: make and save money.

“One of our key objectives has always been to grow revenue,” says Verizon's Wegleitner. “Another is to cut costs through the application of new technology. Those are our two overriding principles.”

Creating the future

One of the most striking things about these service provider labs is the sheer variety of work they take on, from testing vendor equipment to improving the company's strategic positioning to cleaning up its operational inefficiencies.

For all the labs, about 20% of the time is spent on long-term applications that won't find their way to market for years. These projects are the ones that seem to get the most attention from the outside world, if for no other reason than that they are futuristic.

Sprint's ATL, for example, is currently working with Silicon Valley-based software developer Headpedal for the development of its intelligent agent, a virtual butler that resides on the Sprint network.

While on a PC, the self-motivated agent, which will be able to recognize people on-site and is able to “learn” user preferences, appears as an individual dubbed “Chase Walker.” In a networked home, Chase will be able to perform tasks as varied as reading e-mail and brewing coffee. While operating via a handset, the agent will also be able to alert the user of traffic tie-ups without any prompting.

While Chase Walker definitely has potential, commercial deployment for the application is still years away. However, Sprint's work on the agent and other broadband applications has a larger purpose than simply devising a new service offering, says Mike O'Brien, manager of the ATL's service architecture group, which is in charge of the intelligent agent's development.

“The idea is to be able to see what these services imply about what the Sprint network architecture is going to look like,” he says. “What we end up doing is we take a look at new technologies that are on the horizon and what technologies are going to be available. We end up prototyping those services so we can find out what they require of the network.”

Gaining such an understanding of the network is in fact one of the main reasons service providers have labs, says William Hurley, program manager with The Yankee Group. “The real benefit is for them to think through how to monetize services. What that means is understanding machine behavior, infrastructure performance and then putting together [service level agreements] and making a margin delivering a service,” he says.

In fact, influencing network design is a task undertaken by all the labs. Verizon Labs, for example, is currently developing vertical services for DSL such as streaming media and video-on-demand. Through this work, says Wegleitner, “we're evolving DSL architecture on a day-to-day, month-to-month, week-to-week basis.”

Scientists and housekeepers

Not all the work done by the labs directly involves a new service. Some of the projects focus on maximizing the provider's revenue potential by streamlining its operations.

AT&T Labs, for example, is currently working on perfecting the company's order-to-billing process. According to Rabiner, most estimates predict that the average company looses from 2.5% to 5.5% of its revenue from orders that are never billed properly.

“We're probably no better or no worse than any other company,” he says. “Our guess is in the 4% to 4.5% range of our revenue that we could potentially put back on the table. Take 4.5% of $67 billion (approximately AT&T's annual revenue). That comes to about $2.7 billion dollars. That alone could make or break a huge number of corporations.”

The labs are able to improve the bottom line by taking some of the providers' technologies out of house as well. Using various contractual methods, all three of the labs discussed in this article license their technologies to interested outside businesses, providing yet another financial benefit.

“We created the world's best document image compressing technology,” says Rabiner. “AT&T recognized that and didn't doubt it for a minute. Their answer was, ‘I don't care. It's not important to me.’ We licensed that technology outside, and the company that's got it is going gangbusters.”

Licensing out technologies also helps the labs keep talent in house for a simple reason: Researchers like to see their work put into use.

According to Rabiner, before AT&T Labs began licensing out technology, one of the most common reasons researchers left the lab was their frustration with not seeing their work put into practice.

Show me the love

The scientists at all three service provider labs seem to share an unbridled passion for their work, which is an almost necessary quality, according to their respective leaders.

Because of the enthusiasm many scientists have for the technologies they develop, they are frequently the perfect candidates to push that technology in-house — an occasionally necessary function. For Verizon, it was the provider's lab that first pushed DSL as a high-speed access technology, says Wegleitner.

More often than not, though, such evangelization is not needed because of the close working relationships between the labs and the business units.

The method by which AT&T Labs is funded differs from Verizon and Sprint and perhaps creates the closest working relationship between the carrier's lab and business units.

Like Sprint and Verizon, AT&T Labs used to be funded by a simple budget that was handed down from the corporate level. But with the ongoing breakup of AT&T, a new financial structure was needed. Now each of AT&T's customer-facing units contract out with the research team for work to be done over the course of a year, while the lab's development unit gets paid by the customer-facing units on a project-by-project basis.

This simple payment method makes the units more aware of — and hence more involved with — the lab, says Rabiner. “The bottom line always is, if you pay directly then you know it,” Rabiner says. “If you go to a doctor and you physically pay him, you think real carefully about what he is doing for you and what you want done and how that money is spent. If you're taxed for it, you say, ‘Oh well, I can't do anything about it…. It's an awareness kind of thing.”

Technologies licensed out by AT&T Labs

Company	Technology	Date
Digigram	Next-generation digital music compression	February 2001
Excalibur Technologies	Digital video indexing and archiving	November 1999
LizardTech	Document imaging	March 2000
SpeechWorks International	Speech processing	June 2000
Tellium	Intellectual property for mesh restoration software	October 2000
Source: AT&T

Indeed, every division in every company seems ever more aware of financials. R&D labs are no different.

As in any business unit during any downturn — especially following a boom — the attention of the labs seems to have turned to squeezing every dollar out of their current offerings. According to Wegleitner, the downturn has led Verizon labs to focus more heavily on revenue generating opportunities from mature technologies such as advanced intelligent networking.

In the case of AT&T, extra projects are also being taken on due to the downturn, says Leighton of AT&T Labs. “Perfectly good businesses died because they couldn't get funding. There are some things we may have to go do ourselves that we had originally hoped to buy a solution to.”

And like every other unit, the labs themselves are also feeling the pressure to watch every dollar. Still, research and development remain their focus.

“Other than being very careful controlling our expenses,” says Leighton, “we're moving forward aggressively in almost everything we started out to do. We want to get these services in the marketplace.”