I never thought about research as a really engrossing activity right until you actually discover or uncover some facts that you never might have thought were there. I could only talk about my thesis because I have just begun analyzing results I had been gathering for a semester or so now already.
It seems that what has been written before about load balancing and processing distribution have a couple of stringent assumptions that may or may not be met in real life: first would be that the processors are the same (setup is homogeneous), and that global balance should be quantifiable. The first assumption is necessary to simplify the task of load balancing -- and stick as much as possible to solving the bin packing problem. However, in reality the bins don't all get emptied at the same rates, and the rate is not constant. The capacity is however, infinite (i.e. a processor can work forever or at least until it can perform as expected) but that hardly changes the setting of the pro
A blog by Dean Michael Berris.