Abstract: Based on P2P (peer to peer) distributed network architecture, an activity state based distributed task management model is proposed. First, the state of activity, its three characters, change rules of the state, execution process of the task, and the compensation for failed activities are described in detail. Then, based on this model a distributed remote-sensing image processing systemis implemented. Two datasets are used to demonstrate the high performance and its stability of the model. Moreover the model is of strong expandability.

Key words: P2P, distributed network, activity state, distributed task management

摘要： 基于P2P结构的分布式网络架构, 提出了一种基于活动状态的分布式任务管理模型。给出了任务和活动的定义、活动的3个基本约定、活动状态的变化规则, 然后在此模型下详细描述了分布式任务执行的基本过程以及活动出错时的补偿。最后依据该模型, 设计并实现了一个分布式遥感图像处理系统, 在实验中, 采用两组遥感数据进行了测试, 实验结果验证了该模型的可靠性和高效性, 并且该分布式任务管理模型具有很好的扩展性。

关键词: P2P, 分布式网络, 活动状态, 分布式任务管理