普适计算中基于上下文信息的缓存管理算法  

Cache Management Algorithm Based on Context Information in Ubiquitous Computing Environment

作　　者：陈慕冰[1] 赵季中[2] 郗旻[2] 齐勇[2] 马兆丰[3] 

机构地区：[1]西安交通大学软件学院,陕西西安710049 [2]西安交通大学计算机科学与技术系,陕西西安710049 [3]清华大学计算机科学与技术系,北京100084

出　　处：《小型微型计算机系统》2007年第10期1793-1798,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60573119)资助;国家科技攻关计划项目(2004BA111B01)资助

摘　　要：普适计算的一个常见的难题是断连操作,而移动设备在断连状态下对数据进行操作又是必要的.为了支持断连操作,需要在移动客户端上进行数据缓存.数据收集的目的是在断连前把用户将来可能访问的数据预先存储到本地缓存,因此收集过程的结果将对断连操作的性能产生重大影响.目前针对断连操作的数据收集算法,对缓存命中都有一定效果,为了进一步提高缓存命中率,本文根据上下文信息进行数据收集算法;然后在访问数据时同步建立数据之间的关联,并在数据关联的基础上自动选择要收集的数据集;最后将结果按缓存驻留时间和访问次数进行缓存替换.模拟试验结果表明,此算法对于存储容量小的手持移动设备可以有效地提高断连操作时的缓存命中率,可以更好的支持移动设备的断连操作.A common problem facing ubiquitous computing is disconnected operation, but it＇s necessary for mobile devices to handle the data in disconnected state. Data is stored in mobile clients, which supports the disconnected operation. The purpose of data hoarding is storing the data will be possibly accessed to local cache in advance, so the result of gathering data is significant to the performance of disconnected operation. So far, data hoarding algorithm for disconnected operation have some effect on the hit of cache, for increase the hit ratio of cache more, data are gathered according to context information in this paper, and then, the algorithm synchronously builds the association of data which are accessed and automatically chooses data gather based on the association of data, at last, the chosen data gather replace some date in cache according to resident time and access times. Emluator indicates that the cache management algorithm based on context information can effectively increase the hit ratio of cache in disconnected state for mobile devices which are provided with small memory, or rather, this algorithm supports mobile devices for disconnected operation effectively.

关 键 词：普适计算 上下文感知 缓存管理 断连操作 数据收集 替换算法 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]