Throughout his early career, Yogen Dalal was a key contributor to the development of internetworking protocols. As a graduate student at Stanford, he was a member of Vint Cerf’s team that published “Specification of Internet Transmission Control Program” in 1974, the first formal documentation of Cerf and Kahn’s design for TCP. He continued to stay active in the development of TCP/IP after receiving his PhD in Electrical Engineering. His first job after leaving Stanford was with Xerox PARC, where he worked on the 10 Mbps Ethernet Specification with DEC and Intel that led to the IEEE 802.3 LAN standard. Dalal also made important contributions to Xerox’s own internetworking protocol suite, XNS.
Dalal, originally from India, came to Stanford in ’72, to pursue his PhD in Electrical Engineering. His interest in the subject of data communications led him to the office of Vint Cerf, who had recently joined the faculty as an assistant professor. Dalal’s eagerness and persistent pestering of Cerf to take him on as one of his research assistants finally paid off when Cerf made him an assistant in the summer of ’73. Cerf had just started working with Robert Kahn at BBN to develop a new internetworking protocol for the Arpanet, based on Louis Pouzin’s CYCLADES network. Dalal recalls early discussions with Cerf, Kahn and Robert Metcalfe that led to TCP.
In 1977 Dalal went to work at Xerox PARC, where he researched their existing PUP networking protocol to inform the development of XNS. It was during this time that Dalal and colleagues John Shoch and David Boggs realized the coming importance of local area networking and the needs that an exponential increase in the number of LAN’s would require of internetworking protocols. To that end, they did what they could to convince Cerf and his team of the need for both a datagram and a session protocol. This led eventually to the splitting of TCP into TCP and IP protocols.
Dalal passed on Metcalfe’s offer to join him at his startup, 3Com, because Dalal felt like he still had work to do at PARC to help see XNS through to completion. Dalal eventually left Xerox in 1982 to join fellow PARC colleague David Liddle’s startup, Metaphor Computer Systems. They had both become frustrated with Xerox’s unwillingness to release upper levels of XNS. After leaving Metaphor, Dalal co-founded the Apple spinoff, Claris.
A friend asked to walk me to Dalal’s cubicle at Claris, which was located at the far end the floor, with the other software developers. We reached Dalal’s office just as a meeting was breaking up. The camaraderie between Dalal and his team was immediately clear. I explained why I wanted to interview him and we briefly discussed people we knew in common. Once we began, Dalal spoke clearly and thoroughly, remembering fondly the stimulating work environments and amazing amount of talent at Stanford, PARC and the early startups of the day. He was an easy interview, and I could see why he had such an excellent reputation as a manger. By the time we ended, virtually everyone had left and he walked me over to my friend’s office.
Keywords: TCP, TCP/IP, Xerox, Ethernet, PUP, XNS