A student asked me recently if our lab was part of a start-up company. (She hadn’t heard that Telcordia Technologies is the new name for what was formerly called Bellcore.) I told her "No." Indeed, we have nearly 7,000 employees in the company and $1.5 billion in revenues—not the usual parameters of a fledgling .com operation! Upon reflection, however, I think I could have given a more thoughtful answer.

For researchers in Telcordia’s 120-person Information and Computer Science Research Lab (ICSRL), there is definitely the atmosphere of excitement and opportunity that one expects to find in start-ups. Perhaps the main differences are the diversity of problems we are tackling, the variety of opportunities we have for research impact, the critical mass we have to form cross-disciplinary teams, and the momentum we have from 15 years of leading-edge research in the computer, mathematical, and information sciences.

The ICSRL is located primarily in Morristown, New Jersey, the corporate headquarters. It is one of three major labs in Telcordia’s research organization. Telcordia itself is one of the world’s largest providers of operations support systems, network software, and consulting and engineering services to the telecommunication industry. Our primary research mission is to provide research support for existing products and services and to spawn new ones.

Most of the people in our lab have a background in computer or mathematical science. Areas of disciplinary focus include software architecture, engineering, quality, testing and understanding; databases; distributed and fault tolerant computing systems; programming languages; cable systems; performance analysis; information management, modeling, security, and survivability; cryptography; data analysis, modeling, and mining; and network mathematics. Some of our research targets rapidly developing areas such as mobile communications, cable-related services, and electronic commerce. Others aim to assure that Telcordia’s more than 100 million lines of software work properly in the field.

Virtually every staff member in our lab has either a Ph.D. or M. S. degree. Many got their start at Telcordia by working here on a summer internship. (The ICSRL runs a vigorous summer program for about 25 undergraduate and graduate students each year.) Several members have come to Telcordia after working in academia, including the lab's Vice President and General Manager, Rich DeMillo. A few, like Fan Chung Graham and Dick Lipton, even hold joint appointments with universities. There is also an "old-timers" list (which includes me!) of people who came from Bell Labs in 1984. We were part of the core group that formed Bellcore at the time of the Bell System's divestiture. At the other extreme, one of our scientists, Scott Stornetta, took the unusual path of university physics Ph.D. to Telcordia computer science researcher to start-up company (formed around his co-invention of a digital time-stamping technique), and then back to Telcordia.

While Stornetta’s case is unusual by traditional standards, it may prove to be a more common model in the future. We are often faced with an embarrassment of technology riches, not all of which can find a logical home in the current internal business plans. The value of some of these nuggets may be best realized by following non-traditional paths.

The ICSRL emphasizes rock-solid ties with the external scientific community. Our researchers are expected to be active in professional activities, including publishing and presenting conference papers. We strongly support a variety of research institutes that focus their energies on computer and mathematical science problems of importance to us. In 1988, we were one of four New Jersey institutions to join and found the Center for Discrete Mathematics and Theoretical Computer Science, DIMACS, which is located at nearby Rutgers University.

We also have long-standing relationships with the Software Engineering Research Center (SERC) housed at Purdue University, reflecting Telcordia’s emphasis on software quality and productivity; and the Center for Advanced Computing and Communications (CACC) located at Duke and North Carolina State Universities because of our interest in high-performance and highly reliable distributed systems. Telcordia has been a member of the Participating Corporations Program of the Institute for Mathematics and Its Applications (IMA) at the University of Minnesota since its inception in 1986. This year, when the National Institute of Statistical Sciences (NISS) in North Carolina established an Affiliates Program centered on cross-disciplinary research involving statistics, we became one of its charter members. We also work closely with many university colleagues and departments in virtually every section of the country. All of these connections help us to multiply the impact of our own research investment.

Internally, the ICSRL is organized into the Software Technology Research Department, headed by Mark Segal; the Information Analysis and Services Research Department, headed by Sid Dalal; the Cable Access Technologies and Services Research Center, recently created by Peter Bates; and the Mathematical Sciences Research Center, which I head. Within this structure we have about 15 focused research groups where all the serious research work gets done.

A scan of some of our current research projects will, I hope, give a sense of the research portfolio and the types of problems we work on.

• Mark Segal, in collaboration with researchers from our lab and SUNY Stony Brook, is developing techniques to detect and respond to computer system intrusions. Their approach matches patterns of system calls and network packets against event-based behavioral specifications. It provides protection without requiring access to source code and allows mission-critical systems to keep running if attacked.

• Josephine Micallef leads a research team that develops and applies tools and technology for building and maintaining large, enterprise-critical, distributed applications. Major thrusts include Web architectural models, component-based software development techniques, workflow technology to manage and automate business processes, and language technology for finding industrial-scale solutions to hard optimization problems.

• A team of software engineering researchers, led by Bob Horgan, has developed a suite of methods and tools for analyzing and visualizing the reliability of software in different ways. While most approaches to this problem ignore the internal structure of the software under study, this one exploits it by capturing exact execution traces. (The technology is available for use by university researchers.)

• Ernie Cohen is working on an automatic verifier for crypto protocols and systems. The basic idea is to generate a special kind of invariant to reduce verification to first-order theorem proving, and then to blast it with a resolution theorem prover. Common authentication protocols can be checked in seconds.

• Raj Rajagopalan heads a team of researchers who are investigating "smart data" approaches to managing networks of firewalls. The goal is to develop tools that will determine if global security properties are satisfied by firewall configurations and service specifications and, if not, to figure out how to reconfigure the firewall parameters so they are.

• Amjad Umar and Paolo Missier are developing an adaptive knowledge-based decision support system. It is designed to behave like an expert system for routine problems and an intelligent assistant for harder problems. As part of this research, they are exploring Web mining using "Knowbots" that will recognize knowledge patterns and periodically update a knowledge base.

• David Shallcross, Fan Chung Graham, and others are developing mathematical algorithms and theory for network tomography. The basic problem is to infer the interior topology of a network from a series of delay measurements taken across its edges.

• Cliff Behrens and his team are building a next generation information retrieval system that supports concept-based queries on massive multilingual collections of documents.

There are numerous other examples ranging from analysis of sailboat performance data from an America’s Cup syndicate to a software communications infrastructure for monitoring compliance with the comprehensive (nuclear) test ban treaty. We are not at a loss for things to do!