Tech Search

But this wasn't just any guy, it was one of the people who helped invent the Internet -- Vint Cerf. Decades before startin... g his current day job as a "vice president and chief Internet evangelist" for Google, Cerf co-wrote the Internet's TCP/IP language with Bob Kahn.

Cerf was in Tysons for one of Palantir Technologies' monthly Palantir Night Live events, in which the government contractor hosts a tech notable of some sort for a presentation and subsequent Q&A.

(As a historical curiosity, I first heard about these events from a Palantir engineer named Mike Shafer, whom I had met in the late '90s when he -- then a high-schooler and already a budding tech geek -- wrote a few video-game reviews for a monthly Post magazine called Fast Forward.)

Said traffic caused me to arrive late to Cerf's talk (only three more years until I can take Metro to Tysons ... ), so I missed the part where Cerf apparently discussed his Internet Protocol-enabled wine cellar. Here's what I did catch -- note that in some cases, I rely on the Twitter recounts of others for quotes from Cerf, as credited with links back to those tweets:

* Cerf is not thrilled with the state of consumer online security. Password management, for instance, "sucks," too many people still run "naive browsers" that trust strange sites too much (one reason Google built its Chrome browser), and we have inadequate remedies for losing laptops and USB flash drives with sensitive data. Meanwhile, hackers have gotten very good at exploiting these problems for commercial gain: "The hackers don't want to destroy the network. They want to keep it running, so they can keep making money from it."

* As you might expect for a Google employee, Cerf has no problem with "cloud computing," but he sees room for improvement in the ways different clouds can share information. (His slide highlighted this as a "new research area!") How, for example, can one company's server architecture securely and reliably share data with another's? Note the difference in vocabulary between Cerf's talk and what I heard from Microsoft Chief Executive Steve Ballmer on Monday, who referred to "The Cloud" as if it were a monolithic entity.

* Cerf wondered why there isn't a "cyber fire department" -- a recognized, trusted, public entity that companies could call upon when they need help. He suggested that the work done by Sandia National Laboratories in battling the still-afoot Conficker worm was a good analogy.

* He saw a need for greater international cooperation and regulation of the Internet but doesn't expect that to happen quickly. He suggested that a "law of the Internet" would take as long to settle as the United Nations' Law of the Sea treaty (which has acquired the unfortunate acronym "LOST," in part because of its slow progress).

* Cerf sees authentication and identity as problems to be solved online, but not in the way you might think -- with a centralized, single sign-on system. As he clarified to me afterward, he doesn't see why anybody would want such a thing: The risks of accident or abuse would be far too high, so he'd rather see federated identity systems that let us authenticate who we are as needed, but only for limited times and for specific purposes. Nor did he express any regret over leaving those functions out of the TCP/IP specification.

* The fun part of the talk came when he moved to discussing his plans for "InterPlaNetary Internet." This seemingly science-fiction effort aims to solve a genuine problem: the point-to-point communication that has worked acceptably well for individual missions to other planets doesn't scale as we send more hardware Out There and expect more data back. A networked communications system would make more sense, but the Internet's protocols need to be adapted. Specifically, they can't handle the long latency of communication from here to Mars or beyond -- "The speed of light is too slow," he noted -- and they do need to have every packet of data authenticated, given the costs of a compromised system stuck 200 million miles away.

Cerf expanded on many of this ideas in the Q&A afterwards, and one example from that interaction might demonstrate why he's such a fascinating person to hear talk.

A fellow next to me asked a follow-up about interplanetary networking, which led Cerf to observe that what really excites him is the prospect of interstellar Internet. But, he noted, with current technology it would take 50,000 or so years to get to the nearest star, Alpha Centauri. So we first need to figure out how to travel at at least a tenth of the speed of light, to cut the journey down to less than 50 years. That, in turn, led Cerf -- still speaking in the same paragraph -- to speculate about developments in quantum mechanics and particle physics that might enable such a thing.

I'd provide details of those developments, but at that point my brain felt rather small and I could only think about how I'd gotten in over my head in a minute or two of listening to Cerf expound on a technical issue.

At the risk of erasing any ideas you might have about my objectivity with respect to this gentleman, I should note that it quietly astounds me that I've had the chance to talk shop with Cerf, Kahn and Web inventor Tim Berners-Lee. How many reporters get to meet the people who created the things they write about every day? I feel privileged to be reporting on the Internet at this time.