Dave Clark Today I spoke to
Dave Clark about my thesis. We knew it was going to happen eventually.
Dave Clark was a guest in
Jonathan Zittrain's
cyber law class, and I can tell you with assurance that the clarity and precision with which he presented the issues surrounding Internet
governance,
quality assurance,
market efficiency, and
filtering policy are unparalleled. He spent time articulating how the flow of bits is generally orthogonal to the flow of value (i.e. why
peering arrangements without cash flow might sometimes be the best we can do), how assumptions about last-mile network utilization are essential to ISP business models (i.e. why
Verizon perceives danger in
Google reducing the
burstiness of the traffic of its customers), and why some issues of Internet structure are really
important policy issues rather than simply
technical issues.
I told him that I study
network neutrality and that I am working on a
peer-to-peer overlay network for sharing network perspectives, that our network performs the necessary routing and directory services to allow endpoints to choose how they see the world.
"That's interesting. Won't this make some people angry?" he said, with a bit of a twinkle in his eye. "It is not exactly value-neutral technology. It clearly promotes the vision of network openness [and
crypto-anarchy, presumably]."
I explained that my thesis was primarily an experiment to
characterize and
quantify the extent to which such a network is feasible, and that it is entirely possible that by providing this service, we may actually promote localization of services, since we provide a means of specifying alternate locations from which we want to view the set of services. In this sense, it is not entirely clear that we are promoting an open Internet at all, but perhaps instead a set of small, closed Internets that are able to persist precisely because we have built an
escape-hatch for the privileged. As technologists, we have no reason to make a prediction about which of these possible results will occur, right?
At this point he pointed at me, and flatly declared, "that's cowardly." What? Had I heard correctly that this man whom I had met only minutes ago -- nay, seconds ago -- had reduced my years of thesis work to a single derogatory adjective?
He continued by explaining that my work takes a position, and that I had chosen to
promote Western values of ubiquitous exposure to culture and speech. "That's a good thing," he added, "but you can't get away with arguing that the issues surrounding the design of your system are only technical in nature." He declared that he would consider himself personally and intellectually irresponsible if he had chosen to build or deploy a system without addressing these fundamental issues. Similarly, he said, it would be a real shame if technical researchers did not work in this space, since there is so much to be done.
"The CS research community needs to grow up," he said, with reference to
numerous technical conferences featuring
disruptive technologies in the form of papers and research descriptions that hide behind empirical measurement and theoretical performance issues rather than addressing practical problems associated with deployment, policy, and social effects. For example, one might design a
peer-to-peer system for the sort of scalability that implies popular use, but if we describe it in terms that downplay its social effects, then we are in some sense choosing a path of intellectual dishonesty. "If you develop these technologies, then you are participating in a war. You should come out and say so."
"That doesn't sound very civilized," I responded.
"Of course not, but we know the other side certainly isn't."
So at this point our conversation ended, and it seems to me that if he were reading my
SIGCOMM submission, he would probably want some serious treatment of filtering and policy in addition to empirical measurements that demonstrate that the system works. At any rate, I think my work is cut out for me.