In the past a physics experiment could be set, up in a small laborator
y by a single genius. These clays, physics experiments involve interna
tional collaborations of hundreds of scientists, and tile laboratory o
f tile ''crazy scientist'' has been replaced by kilometers of accelera
tor tunnels, enormous underground caverns, tons of detector apparatus,
hundreds of kilometers of cables, thousands of electronic boards and
dozens of computers. The control of such an experiment is a complex sy
stem involving constraints of reliability, efficiency, reconfigurabili
ty and maintainability. This paper shows that an efficient communicati
on system is an important issue when building such a control system.