A hybrid strategy for load balancing in distributed systems environments
Susana Cecilia Esquivel, Guillermo Leguizamón, Raúl Héctor Gallard · 2002
Load balancing algorithms attempt to improve systems performance through process migration. The article presents a hybrid strategy for load balancing in distributed systems, which exploits the benefits of evolutionary and predictive approaches. In order to decrease the communication traffic typically generated by load balancing schemes, we sought for a reduction in the number of requests done by an overloaded node. The predictive strategy applied to achieve this goal uses the knowledge attained by each node, through its previous experience.