Tetherless Computing Blog

Vint Cerf on Internet Challenges

Vint Cerf has some interesting points on the science in computer science and whats missing in networking research. He has a point about networking being weak on theory. In many ways protocol design is an art, that tries to balance conflicting requirements in terms of implementation complexity, flexibility, and minmizing overhead, while trying to achieve good performance. I can't see how one can come up with strong theoretical principles for that. On the other hand, the theoretical foundations for networking in queueing theory, control theory, and game theory are sound, but balancing the assumptions needed by the theory with the facts on the ground is difficult, and sometimes impossible!

posted by S. Keshav @ Thursday, April 21, 2005