Developing the Internet: A Generation of Openness

Burt Kaliski | Jan 31, 2013


Why has the Internet been so successful?

The key, Internet pioneer Vint Cerf explained to an audience of more than 200 at his recent Verisign Labs Distinguished Speaker Series lecture, is its openness.

"The openness of the net has allowed so many other people to contribute that it's sort of like creating a giant tidal wave," Cerf stated.  "We've been surfing (no pun intended) on that tidal wave for 30 to 40 years."

The tidal wave started when Vint and Bob Kahn invented the TCP/IP protocol for sending packets of data reliably from one location to another.  TCP/IP is the "lingua franca" of Internet communications, the common language by which computer systems exchange data of any kind.  How the data is exchanged -- dial-up, broadband, wired, wireless, or carrier pigeon -- is up to the network implementation.  What the data means is up to the application.  But the format of the exchange is standard:  a source address, a destination address, and data.

A standard interface between layers can often be the catalyst for innovation in both.  By reducing the diversity of interfaces, a standard can increase the diversity of parties on both sides of the service interface -- applications and network implementations in this case, service consumers and service providers more generally.  Instead of a polyglot of point-to-point interfaces within specific protocol "stacks", developers could master one interface, and focus instead on the applications above or the network implementations below.  This architecture of "loose coupling" would naturally attract more applications, because they could be used with more network implementations, and vice versa, a symbiotic relationship that would ultimately make the architecture dominant.

Open architecture takes the principle a step further, by inviting contributors to share their work with one another, and relying on their engagement in the ongoing evolution of standards, a point also made by IETF chair Russ Housley in his lecture in this series.  The parts that are built may have proprietary aspects, but the growing body of publicly available specifications and reference material on how they work together strengthens the system overall.  Vint described how similar openness benefited another layer of today's Internet stack, HTTP with its HTML format, where developers learned how to write HTML by copying one another's pages.  "Openness [created] new opportunities," he remarked, as "everybody learned from everybody else."

A generation of technologists has grown up surfing the Internet wave, learning along the way as a result of the essential openness that Vint has personally championed throughout his career.  More than the technical merits of TCP/IP itself, it's the open, collaborative environment around its development that has made the difference.  And that environment itself is shaped by the Internet:  the more easily technologists have been able to connect with one another, the more rapidly and confidently they've been able to develop TCP/IP and the Internet.  (Keeping in mind, of course, the ongoing value of meeting in person, in addition to the online interactions, whether at conferences or in this series.)

My observations here are distilled from my reflections on just the first 10 minutes or so of Vint's lecture, the rest of which was equally well thought and presented.  Another thread he pursued was a three-fold classification of security responses (or more broadly, "safety", a phrase he prefers).  The best response, he said, is "a technical one that prevents the bad thing from happening."  The next best is "post hoc enforcement":  detecting a security incident and responding after the fact.  The third alternative, "social morays", has merits as well:  the admonition "don't do that."  Some actions, if not easily preventable or detectable, can simply be discouraged.

In the world of cybersecurity -- which would be a story of its own -- I'm not sure how much we can count on the last approach.  But in the world of Internet development, the converse has energized a transformation.  No technical action ensured the "good thing" would happen, but the social moray of "request comments" and "collaborate" has turned a protocol designed in the lab into a part of the lives of 3 billion people and counting, as well as an entire industry.

How has the open architecture of the Internet made a difference for you?