一种网格环境下的动态负载均衡机制

Dynamic Resource Load Balancing Mechanism Under Grid Environment

负载均衡机制有利于提高广域分布式环境中资源共享和协同工作的效率。根据网格系统的特点,采用灰色预测方法,设计了一种动态资源负载均衡机制,给出了预测模型和实时预测策略以及基于该机制的负载均衡算法。该资源负载均衡机制具有以下特点:可在较小的开销下取得满意的负载均衡性能,具有网格环境下的可扩展性,能够适应网格资源动态变化的特性,解决资源发现过程中的负载均衡问题。

Aiming at the problems that grids are highly dynamic, and existing load balancing algorithms can not satisfy the unique requirement of grid, a dynamic load balancing algorithm based on grey prediction is presented. According to the characteristics of grid, combining multi-route lookup strategy with grey prediction method, this paper designs dynamic resource selecting strategy, and presents the prediction model, as well as a real-time prediction strategy. Furthermore, on the basis of the dynamic selecting strategy, the paper puts out a load balancing algorithm for grid environments. Simulation results show that the algorithm is valid and effective, and can be used for load balancing under grid environment.