Abstract: In order to support large-scale migrating workflow program running on resource-limited mobile devices,a cache management of dock server in code-on-demand migration workflow in wireless environment is introduced.The generation-based algorithm in memory management according to the adaptive thinking is introduced.Considering the different characteristics of the objects in the less-relied generation and the more relied generation,different allocating methods and garbage collecting algorithm are adopted.These methods highly improve the efficiency of the cache management and avoid long-time problem in general garbage collecting algorithms.The experiment indicates that the cache management is reasonable and effectively supports the application of the migrating work-flow in the wireless environment.

Key words: migrating workflow, workflow unit, dock server, cache allocation, cache garbage collection

摘要： 作为基于移动计算范型迁移工作流研究的继续,在无线迁移工作流按需移动中建立工作流单元在停靠站服务器中的存储模型和方法,在移动设备主动或被动地处于断接,即没有网络连接的状态下,有效利用缓存中的工作流单元,以支持大型工作流程序在资源受限的移动设备上更有效地运行。在缓存管理中,引入了分代算法。针对more-relied代和less-relied代中工作流单元的不同特点,采取不同的分配方法和垃圾收回算法,以提高分配效率,在很大程度上避免内存垃圾回收算法带来的长“停顿”。实验表明,无线迁移工作流按需移动中停靠站缓存管理机制模型合理,有效支持迁移工作流在无线环境中的应用。

关键词: 迁移工作流, 工作流单元, 停靠站服务器, 缓存分配, 缓存回收