Low cost load balancing algorithms
Weiping Zhu, Andrzej M Goscinski · 2002
Two load-balancing algorithms that significantly reduce network traffic and the latency of the design-making process are presented in this paper. These algorithms take user behavior into account. Based on the intermittent behavior of a user using his/her computer, these algorithms adopt different measures to accelerate decision process and to reduce unnecessary communication. The measures used in these algorithms are beneficial to all type machines, especially to those machines that have bursty usage feature. The simulation shows that even without the intermittent feature on the job stream of each participant machine, these algorithms achieve better performance than other algorithms and have less overhead.>