Adaptive Approach to Information Dissemination in Self-Organizing Grids

摘要: The size, complexity, heterogeneity, and dynamism of largescale computational grids make autonomic grid services solutions necessary. In particular, schedulers must map applications onto resources whose state (1) influences the effectiveness scheduling choices, (2) changes frequently considerably. A resource information dissemination service negotiate inherent tradeoff between covering a large portion (so that all can informed decisions with largest number options), limiting protocol’s overhead (i.e. packets sent). This paper argues probabilistic forwarding protocols adapt to changes, because static assignments probabilities lead excessive or lower-than-possible query satisfaction rates in some scenarios. We introduce an approach compares node’s local utilization generation corresponding vicinity, as whole. These comparisons, turn, produce score is used adjust probabilities. show even this simple initial adaptive work better than probability assignments.