As a graduate student at UCLA, he was involved in the early design of the ARPANET. He was present when the first IMP was delivered to UCLA. He is called the “father of the Internet”. He earned this nickname as one of the co-authors of TCP/IP - the protocol that allowed ARPA to connect various independent networks together to form one large network of networks - the Internet.
A young man with style
Cerf grew up in Los Angeles. He did very well in school and showed a strong aptitude for math. He had an unusual style of dress for a school kid. He wore a jacket and tie most days. Cerf is still known for his impeccable style. He is usually seen in three-piece suits.
As a child, Cerf began to develop an interest in computers. He attended Stanford and majored in mathematics, bur continued to grow more interested in computing. “There was something amazingly enticing about programing,” said Cerf. “You created your own universe and you were master of it. The computer would do anything you programmed it to do. It was this unbelievable sandbox in which every grain of sand was under your control.” (Cerf in Hafner & Lyon, 139).
When Cerf graduated from Stanford in 1965, he went to work for IBM as a systems engineer, but soon decided to return to school to learn more about computers. He enrolled in UCLA’s computer science department and began pursuing his Ph.D. His thesis was based on work he did on an ARPA-funded project for the “Snuper Computer” - a computer that was designed to remotely observe the execution of programs on another computer.
An interest in networking
The Snuper Computer project got Cerf interested in the field of computer networking. In the fall of 1968, ARPA set up another program at UCLA in anticipation of building the ARPANET. It was called the Network Measurement Center. It was responsible for performance testing and analysis, a sort of testing ground. A man named Len Kleinrock managed about forty students who ran the center. Cerf was one of the senior members of the team.
By the end of 1968, a small group of graduate students from the four schools that were slated to be the first four nodes on the ARPANET (UCLA, Stanford, the University of Utah, and UC Santa Barbara) began meeting regularly to discuss the new network and problems related to its development. They called themselves the Network Working Group (NWG). The NWG proved to be instrumental in solving many of the problems that would arrive during the design and implementation of the ARPANET, but they did not realise their importance at the time. Cerf recalls, “We were just rank amateurs, and we were expecting that some authority would finally come along and say, ‘Here’s how we’re going to do it.’ And nobody ever came along.” (Cerf in Abbate, 73)
One of the main obstacles facing the deployment of ARPA’s network was the problem of getting incompatible host computers to communicate with one another through the IMPs. Bolt Beranek & Newman (BBN) was only responsible for building the IMPs and making sure they could move packets, not devising the methods they and the host computers would use to communicate. Devising standards for communication, what came to be known as a protocol, became one of the NWG’s main tasks.
The NWG implemented a “layered” approach in building a protocol. This means that they created several simple “building block” protocols that could later be joined to oversee network communication as a whole. In 1970, the group released the protocol for basic host-to-host communication called the Network Control Protocol (NCP). They also created several other protocols to work on top of NCP such as Telnet, which allowed for remote logins.
A true Internet
In August 1969, BBN delivered the first IMP to UCLA. A month later the second was delivered to SRI. The ARPANET continued to grow quickly from that point. Cerf was present when the first IMP was delivered to UCLA. He was involved with the IMP immediately performing various tests on the new hardware. It was during this testing that he met Bob Kahn. They enjoyed a good working relationship.
Within a few years of the creation of the ARPANET, other computers networks were deployed. They were all independent self-contained networks. Cerf recalls, “Around this time Bob started saying, ‘Look, my problem is how I get a computer that’s on a satellite and a computer on a radio net and a computer on ARPANET to communicate uniformly with each other without realising what’s going on in between?’” (Cerf in Hafner & Lyon, 235).
They decided that there needed to be a “gateway” computer between each network to route packets. The gateway computers would not care about the various complexities of each network. They would simply be in charge pf passing packets back and forth. But all of the networks transmitted packets in different ways, using their own protocols. A new standard was needed to link ball of the networks and allow inter-network communication.
Cerf and Kahn began working out a plan in 1973. In September, they presented a paper outlining their ideas to the International Networking Group. In May 1974, they complete their paper entitled, “A Protocol for Packet Network Intercommunication.” They described a new protocol they called the transmission-control protocol (TCP). The main idea was to enclose packets in “datagrams.” These datagrams were to act something like envelopes containing letters. The content and format of the letter is not important for its delivery. The information on the envelope is standardised to facilitate delivery. Gateway computers would simply read only the delivery information contained in the datagrams and deliver the contents to host computers. Only the host computers would actually “open” the envelope and read the actual contents of the packet. TCP allowed networks to be joined into a network of networks, or what we now call the Internet.
Cerf continued to refine TCP. In 1976, he accepted a job as a program manager responsible for what was then called the “ARPA Internet” at ARPA. In 1978, Cerf and several of his colleagues made a major refinement. They split TCP into two parts. They took the part of TCP that is responsible for routing packages and formed a separate protocol called the Internet Protocol (IP). TCP would remain responsible for dividing messages into datagrams, reassembling messages, detecting errors, putting packets in the right order, and resending lost packets. The new protocol was called TCP/IP. It went on to become the standard for all Internet communication.
Today Cerf is the chief Internet strategist for MCI WorldCom. His latest pet project is called the Interplanetary Network (IPN). This project, part of NASA’s Jet Propulsion Lab, will basically extend the Internet into outer space. It’s fitting that the “father of the Internet” on Earth should be involved in launching it to new worlds.