Bell Labs eyes broadband's future

When Jeff Jaffe talks about the shape of things to come, he's usually not thinking in terms of weeks or months. He's often talking in terms of decades--or even longer.

But that's his job as president of Lucent Technologies' Bell Labs, a 77-year-old institution that Federal Communications Commission Chairman Michael Powell recently called one of America's technological crown jewels. Hyperbole aside, the Labs have produced sundry seminal technology inventions such as the transistor, the laser and the Unix operating system.

Jaffe has been at the helm of the laboratories for three years, taking over just as the telephone industry began to go into a prolonged financial tailspin. He's been battling the mysteries of networking on a tight budget ever since.

Along with thousands of other telephone executives, Jaffe attended last week's Supercomm 2003, the largest telephone network equipment show every year, when he took time out to talk with CNET News.com about the future of broadband and telecommunications.

Q: What's at the top of Bell Labs' project agenda these days?
A: The most exciting thing is probably securing high-speed wireless data. Here's an interesting problem. Let's assume that a service provider wants to provide new services for end users. If the package with which it's supposed to deal is encrypted and it cannot inspect what's inside, that's a problem. So we had IPsec (a wireless security standard that's short for "Internet protocol security"), and we've since developed an approach called IP SuperSEC. It's not the best name, but what can you do?

That's not bad. What is IP SuperSEC?
It essentially is a standardized way of encrypting, to the edge of the network, that gives the service provider choices and access. We'd only use it once it's standardized. Security is not something that you do ad hoc.

How long might it be before Lucent offers IP SuperSEC?
The standards process takes two years. It started two or three months ago, so it's fresh off the presses. The main supporters at this point are all the service providers. They are the ones that will really get the advantages.

Why secure data on a cell phone network? Nobody's really using it right now.
I think it's a real worry. Most large companies are not going to allow you to access corporate data unless you're going through some virtual private network, firewall or something like that. At the consumer level or at a small company, protection isn't necessary. But a large company that has access to corporate databases does need security.

What else is hot at Bell Labs these days?
What's very exciting is the triple-play services--TV, phone and data. How do you leverage services that integrate all those capabilities into one? And not only compressed video, but also high-definition TV--things like that. When it comes to entertainment, people have said quite clearly that they are willing to spend the extra buck to get high-quality video. So delivering that on fiber is probably going to be a demanding application, but one which fiber can easily do.

Can the copper networks now used by telephone companies do that?
No.

Not with any of the new enhancements being introduced at Supercomm?
Copper is great for voice. For data, it's great for compressed video and that meets a lot of applications' needs.

Security is not something that you do ad hoc.

But if you want to watch something like real-time sports, which has a lot of movement and action, you really want high-definition capability. It's silly to go through all the effort to create a high-definition standard and then compress it so that you don't see any of the high quality.

Intel, Alcatel and Hewlett-Packard are now pushing modularity, the idea that all telephone network equipment should share some of the same building blocks. How do you feel about that push?
There are places in the system where there's room for innovation, and there are places in the system where all you care about is the cost. The whole movement of our industry is to standardize those things in which you need low cost and to focus investing on innovation. So we're very big supporters of standardization, not only at the protocol level but also at the system level. We're placing better software for better management capabilities on top of that. We're achieving some differentiations for each vendor that comes in so that we can do a better job than the next guy.

Isn't that what happened to computers?
A computer's value used to be all in the hardware, and now you can buy a whole PC basically on the street and put sheet metal around it. The value's now in the software.

In tough economic times, with budgets getting slashed, it must be difficult to innovate. There's less money and lousy morale. Am I wrong?
Researchers are an interesting breed. They are most stimulated when they have a ready source of exciting new technical problems to work on. That's what researchers do: They try to create the next generation. From a research organization standpoint, getting a steady stream of new problems to solve is what's most important. How do I recover stranded bandwidth? How do I recover revenue in a network? I always find that most researchers who are presented with potential problems can do a very good job. The hard aspect of research is finding that problem.

With less money, there must be more emphasis on profit-making projects. Has the amount of pure research versus market-driven projects changed as a result of the economy?
I can't give you its whole 77-year history. After the three years that I've been here, about 15 percent to 20 percent of Bell Labs researchers are still working on fundamental science. But fundamental physics doesn't particularly impact products or Lucent projects in five years, ten years--maybe ever. So we've maintained a traditional focus on scientific research.

People want the next hot box, but they want such that it can be incorporated into an overall end-to-end solution. Not only should it be cool, but it should solve a customer's problems.

What did you find interesting at the show?
Frankly, a lot of folks are copying the stuff that we've been talking about.

Like what?
Lucent Technologies was the first equipment manufacturer to declare openly and publicly that it's not only about the next hot box, but that it's going to provide integrated customer solutions that it'd back up with services like consulting or management.

The thing that stimulates researchers the most is having a ready source of exciting new technical problems to work on.

I'm seeing a lot of that.

So, what's new is actually old? What's the most disruptive technology you've seen?
The fiber for the premises technology of Verizon, SBC and Bellsouth will be extremely disruptive on multiple levels. For starters, it's going to move the ability we have to get high bandwidth to the home another order of magnitude. Just think of the difference that we experienced when we moved from 56 kilobit modems to DSL and cable--it was huge. Now we get fiber to the premises. The immediacy that one will have to data gets greatly improved. It's going to then put the next generation of load on a telephone network's core. That means a doubling or tripling every year of traffic on the network. And when that continues, the carriers can only go so far with their existing equipment before they have to rebuild and re-engineer the metro Ethernet and even, ultimately, the core.

It sounds, from your description, that the carriers are creating a problem by solving one.
That's true from a strictly technical point of view. They are creating a major new revenue source for themselves. In order to support that revenue opportunity, they have to make some investments.

When might this problem arise?
That depends on the speed of the deployment of the fiber to the premises. There are lots of regulatory issues, so I can't really comment. That's really hard for a technologist to predict.

Fiber to the home is meant to deliver high-speed Web access--even cable TV. But it could provide a way to deliver voice services such as Internet telephony, aka voice over IP (VoIP). That raises the question: Will VoIP be the only way that telcos deliver a voice call, rendering the copper networks that they've spent 125 years building moot?
Yes. I think it will happen. There is efficiency to the packet approach, which does make it the ultimate technology. But it's got to be done reliably. There's a tremendous amount of copper-installed base and there has to be a value proposition to move from the old to the new. That's going to take a long time.

It seems that the latest standard that everyone is embracing is MPLS, which lets carriers offer a multitude of services on the same connection. What's Lucent's position on the standard?
We're partnering with Juniper Networks. They have an outstanding MPLS switch, and we're going to surround it with all the network management, provision and traffic engineering, along with our quality service.

What changes have you seen in the ways that telephone companies sell their networks? It seems that they are going beyond just selling airtime, now offering more services.
A lot of those carriers are saying that transmitting bits of information themselves is not what they want to do. They want to add valued-added services.