behind the towers

The first public appearance of our products was at the Comnet trade show in Washington, D.C. in January, 1988. As the recently anointed Chief Technology Officer, I was shipped off to the event with the marketing folks. It was an inauspicious beginning as we attempted to demonstrate the products in a modest operational setting. It wasn’t stable, however, and I spent a fair amount of time in the wee hours of setting up trying to get the router to stay up. Eventually we got it working for a few hours at a time, so we could actually show that the product was real if not ready for prime time..

Comnet was not the ideal venue for making our first product announcement. It was a trade show dominated by the traditional telecom vendors – AT&T, MCI, Sprint, Northern Telecom, Siemens, Nokia, etc. It was, however, familiar territory for the FCGs, especially Dick Kendrick. He thought that the timing would be right – we were planning first customer shipment at the end of the first calendar quarter – and he believed that there would be enough of the right set of customers in attendance. We later learned that though that the company names on the business cards of the attendees were some of the right ones, the titles were not. These were telecom dudes, unfamiliar with terms like “MAC addresses” or “TCP/IP” or “DECnet”. So although we all felt pretty good about how the product was behaving, that the foot traffic was brisk, and that we walked away with a stack of customer suspects, the subsequent sales activity was lackluster at best. A swing and a miss, strike one.

Our real initial prospect base, in hindsight, was much smaller. We had actually identified it earlier, in the fall of 1986, when Steve Willis and Asher Waldfogel attended the very first “TCP/IP Interoperability Conference”. Held at the Marriott Hotel in Monterey, California, it was organized by Dan Lynch of SRI and ISI fame, an early implementer of the TCP/IP protocol stack. I believe that there were around 100 or so who attended, each sporting a pin-on button “Geeks By The Bay in Monterey”. There were a total of three vendors in attendance, including, ironically, InterLAN, our previous startup. These were the engineers and technical managers who would set the foundation for what later became today’s Internet.

Dan Lynch is not, in my humble opinion, given enough credit by Internet historians in establishing one of the cornerstones of success in the evolution of the Internet. The unwieldy name, “TCP/IP Interoperability Conference”, eventually was reduced to simply “Interop”, as it became one of the most successful technical conferences and trade shows in the world. Interop moved in subsequent years to San Jose, to Santa Clara, to Moscone Center in San Francisco. By 1991, Interop had outgrown San Francisco and moved to Las Vegas where attendance exceeded 125,000. Every year, Interop had grown in both size and importance to the development of the computer communications industry. Characterized by the openness of the TCP/IP standards and the cooperative spirit among all of the participants from industry, academia and government, Interop became the “glue” that held this effort together.

All of us had to prove that disparate implementations of the TCP/IP protocol stack (and follow-on standards such as SNMP) would indeed interoperate in a live demonstration network constructed at each Interop event. This was all done voluntarily – no federal government mandate required, thank you very much – and necessary if any of us on the vendor side of the equation hoped to sell into this open systems marketplace. Much as the market for local area network devices was driven by the open Digital-Intel-Xerox Ethernet specification, the emerging marketplace for TCP/IP products was driven by its open systems architecture. The activities undertaken by the IAB, the RFC working groups, and widespread participation by so many vendors resulted in standards implemented by hundreds of engineers. At the time, there was no formal testing process or facility where one could verify interoperability with other vendors’ products. Eventually, a testing facility was created at the University of New Hampshire, but until then, if you wanted to test out your product with another’s TCP/IP implementation, Interop was the only game in town. Or in Las Vegas.

And Dan Lynch had created it. It was a seminal effort, critical to the development of The Internet. Grand slam, Dan.