大规模网络管理中的任务分解与调度

集中式的管理模式和简单的基于移动agent的网络管理方法都已不能满足大型网络管理的需要,为此,在给出基于多agent的网络管理框架的基础上,提出了一种基于任务依赖关系的任务分解算法,把网络管理任务分解为具有不同优先级的子任务,处于同一优先层次的子任务可以并行执行,根据子任务的优先级产生结合网络管理特点的调度策略,理论分析和实验表明:在大规模网络管理中采用任务分解技术能够提高时间效率和减轻网络负载。

Task decomposition and scheduling in large-scale network management

With large-scale network management becoming more complicated, neither centralized network management nor generic agent-based network management could satisfy the increasing demands. A three-tier network management framework based on multi-agent was provided, which gave multi-agent the collaboration environment. In this framework, a task decomposition algorithm based on task dependence relationship was proposed, according which scheduling method was produced for multi-agent considering the characteristic of network management function. The theoretic analysis and experimental results reveal that using task decomposition method can decrease tasks execution time and save network bandwidth in large-scale network management.