No amount of math is going to make multicast happen.<p>Mathematicians always say that their work has applications in network optimization, but it seems to me that the math itself is largely inspirational; Jain's CUTE scheme (precursor of TCP/IP congestion control, the most important distributed algorithm on the Internet) had fluid dynamics inspiration, but trivial implementation.<p>I don't like the connotation that mathematics is a prerequisite to study and improve distributed systems. I can name more far-reaching advances, even in routing, from non-mathematicians.