Larry Roberts

Former chief scientist at the Advanced Research Projects Agency

• Born: December 21, 1937
• Place of Birth: Westport, Connecticut
• Died: December 26, 2018
• Place of Death: Redwood City, California

Primary Company/Organization: Advanced Research Projects Agency

Introduction

Of all of the individuals credited as having contributed to the development of the Internet, Larry Roberts was not only present at the creation but also one of the most critical involved. His participation began with the first small-scale and primitive networks at the Massachusetts Institute of Technology's Lincoln Labs in the early to mid-1960s. His interest in and comprehension of J. C. R. Licklider's concept of an “intergalactic network” informed his eventual role as program manager responsible for all aspects of the management and design of the ARPANET. Like many early developers of ARPANET and its successor, the Internet, Roberts remained deeply involved in making improvements to its capabilities in both infrastructure and the means to communicate on a daily basis. Expressing dissatisfaction with the Internet's capabilities in regard to the demands placed on it, he launched several companies with the intention of making improvements and articulated a vision of how the Internet should develop in the future.

Early Life

Larry Roberts was born in Westport, Connecticut, on December 21, 1937, to Elliot and Elizabeth Roberts. Both of his parents were chemists and both had Ph.D.s. Like many innovators, including George Stibitz and Leonard Kleinrock, Roberts from an early age loved to tinker with electronic and electrical equipment. Among his projects was a small telephone network and a working television set. These interests persisted as he began college. He entered the Massachusetts Institute of Technology (MIT) in 1955, earning his bachelor's degree in 1959, his master's in 1960, and his doctorate in 1963 with concentrations in electrical engineering. One of his classmates in the doctoral program was Kleinrock, with whom he would be involved in later years during the development of ARPANET.

In the early 1960s, Roberts read the series of memoranda written by J. C. R. Licklider that described what Licklider called an “intergalactic network.” Licklider was able to envision what would eventually become ARPANET and ultimately the Internet, but his areas of technical expertise were not in fields that would allow him to design and implement it. Roberts, however, was in such a position, and after reading the memos and discussing them in depth with Licklider in November 1964, he was convinced that creating such a network could be done.

In October 1965, Roberts helped take a step toward the intergalactic network at Lincoln Labs in Massachusetts when he created the first network in which computers communicated with one another. With this working network, Roberts had begun to develop the solution to the problem of finding a way for computers to communicate. In 1966, he published a paper, “Toward a Cooperative Network of Time-Shared Computers,” which not only discussed networking but also addressed another major problem. Computers were becoming more powerful but were expensive, both to acquire and to use, and since George Stibitz's computing machine had gone into operation at Bell Laboratories in 1940, they could be operated by only one person at a time. Roberts saw that this limitation could be removed by the same technology that made networking possible.

In the same year that Roberts published his paper, he began work at the Advanced Research Projects Agency (ARPA, later the Defense Advanced Research Projects Agency, or DARPA), where he would make the network a reality.

Life's Work

Roberts's first job at ARPA was as chief scientist. He was initially unwilling to work on ARPANET, being quite happy with his current position. He eventually accepted the responsibility for heading the program, however, and would not only manage all aspects of contracting and development but also design the system. In June 1967, he published the initial design paper for ARPANET: “Multiple Computer Networks and Intercomputer Communication.” In August 1968, he and his staff prepared the request for proposals (RFP) that would detail the planned project and specified the technical and programmatic (including financial) requirements the winning contractor would have to meet.

Two months after the release of the RFP, Roberts contacted Leonard Kleinrock of the University of California at Los Angeles (UCLA), who headed the Network Measurement Center, there to support measurement and quality functions. In December of that year, Bolt, Beranek, and Newman of Cambridge, Massachusetts, won the contract to build the ARPANET, based on a proposal written, in large part, by Robert Kahn, whose association with ARPANET and similar ARPA projects would be of long standing.

In April 1969, Kahn submitted ARPANET's architecture specification, describing how the host computers would interface with the machines that would pass communications (interface message processors, or IMPs), to Roberts. The specification was approved by ARPA, and in September the first of what would be four nodes was installed at the UCLA Measurement Lab. In May, a second node was installed at Stanford University (which had been chosen because Douglas Engelbart, inventor of the computer mouse and hypertext, worked there). The first message to be carried on ARPANET was transmitted between these two nodes in the same month. The third and fourth nodes of ARPANET were installed in November and December. In the meantime, in addition to remaining involved with the ARPANET project, Roberts was named head of the Information Processing Techniques Office (IPTO), whose first head, Licklider, had convinced Roberts some years before about the possibilities in creating wide-ranging computer communications networks. Upon Roberts's departure, Licklider returned as head of IPTO.

Roberts remained at ARPA for another four years, leaving in 1973 at a time when ARPANET had grown to twenty-three machines (although demonstrations of temporary arrangements of up to forty machines had been given). However, he continued to be actively involved in developing operational hardware and software, improving the infrastructure and shaping the way the Internet would evolve.

Roberts's first post-ARPA position was as chief executive officer (CEO) of Telenet. This carrier, which employed the packet-switching technology used by ARPANET, pioneered methods whereby computers on a network without standard hardware could communicate. That contribution was the X.25 set of protocols, which are still used today. Seven years after taking this position, Roberts left. The year before, Telenet had been purchased by GTE, and it would later become part of Sprint.

Roberts continued to be a major presence in commercial venues, developing and improving communications and networking mostly by concentrating on routers (their design and manufacture), protocols, and packet and Ethernet switches. The companies for which he worked after leaving Telenet included DHL, NetExpress, ATM Systems, Packetcom, Caspian Networks, and Anagran.

Roberts was extremely critical of the current state of the Internet; his main concern was that it was never designed for either the volume or type of traffic that it now supports. One of his projects was to invent a technology that would “create” the bandwidth required for specific types of communication on an as-needed basis. E-mails, for example, require substantially less than the films and videos one downloads from YouTube or other entertainment venues. Roberts noted that the Internet was not designed for watching television and that such activity, with the amount of bandwidth it requires, put a strain on the Internet's infrastructure.

Personal Life

As the Internet has established itself as a major and permanent presence in our lives, widespread recognition was conferred on Roberts. Along with his colleagues in the development of ARPANET (including Kleinrock, Kahn, and Vinton Cerf), Roberts received many awards for his work, including the Association for Computing Machinery (ACM) Special Interest Group on Data Communication (SIGCOMM) Communications Award, the Charles Stark Draper Prize (along with his colleagues) for his contributions to the development of both ARPANET and the Internet, a Meritorious Service Medal from the U.S. Secretary of Defense, and awards from the Institute of Electrical and Electronic Engineers (IEEE). He was inducted into the Internet Hall of Fame in 2012.

Starting in 1999, Roberts became involved in the controversy surrounding the claims of Kleinrock in developing the Internet. Kleinrock, who was brought on board the ARPANET project to support measurement and quality aspects, claimed that he essentially invented the packet-switching technology that was adopted for ARPANET. This claim was disputed rather strongly by individuals who not only developed the theory but also actually built prototypes years before ARPANET was developed. Roberts insisted on the validity of Kleinrock's claims, although some commentators have noted that Roberts did not assign such great importance to Kleinrock's work during the development of ARPANET.

In addition to his continuing work in developing communications, Roberts also became involved in research to extend the typical human life span. He died at the age of eighty-one at his home in Redwood City, California, on December 26, 2018, after suffering a heart attack.

Bibliography

Hafner, Katie. “Lawrence Roberts, Who Helped Design Internet’s Precursor, Dies at 81.” The New York Times, 30 Dec. 2018, www.nytimes.com/2018/12/30/obituaries/lawrence-g-roberts-dies-at-81.html. Accessed 7 Mar. 2024.

Hafner, Katie. Where Wizards Stay Up Late: The Origins of the Internet. New York: Simon, 1996. Print. A good general history of the development of the Internet, covering the development of protocols by Kahn and Cerf.

Mathison, Stuart, Lawrence Roberts, and Philip Walker. “The History of Telenet and the Commercialization of Packet Switching in the U.S.” IEEE Transactions on Communications 50.5 (2012): 28–45. Print. In collaboration with others, Roberts describes how technologies developed for ARPANET were adopted for commercial venues.

Roberts, Larry. “Its Creators Call Internet Outdated, Offer Remedies.” Interview by Bobby White. Wall Street Journal 2 Oct. 2007: n. pag. Print. An interview with Roberts in which he describes shortcomings in the current Internet and how these might be remedied.

---. “Multiple Computer Networks and Intercomputer Communication.” From Gutenberg to the Internet. Ed. Jeremy M. Norman. Novato: Historyofscience.com, 2005. Print. Roberts's 1967 paper is included in this collection of sixty-three seminal readings in the history of computing.

Salus, Peter H. Casting the Net: From ARPANET to Internet and Beyond. Reading: Addison-Wesley, 1995. Print. A thorough history, beginning with the long-distance computing demonstrations at Bell Labs in the 1940s to the full implementation of the Internet. Kahn and the importance of his work in developing protocols is discussed in detail.